2011fluids nephsap

Volume 10 • Number 2 • March 2011

Fluid, Electrolyte, andAcid-Base DisturbancesCo-Editors:

Richard H. Sterns, MD,

and Michael Emmett, MD

■ Editor-in-Chief: Stanley Goldfarb, MD

■ Deputy Editor: Raymond R. Townsend, MD

NephSAPNephrology Self-Assessment Program

®

EDITOR-IN-CHIEFStanley Goldfarb, MDUniversity of Pennsylvania Medical SchoolPhiladelphia, PA

DEPUTY EDITORRaymond R. Townsend, MDUniversity of Pennsylvania Medical SchoolPhiladelphia, PA

MANAGING EDITORGisela Deuter, BSN, MSAWashington, DC

ASSOCIATE EDITORSRajiv Agarwal, MDIndiana University School of MedicineIndianapolis, IN

David J. Cohen, MDColumbia UniversityNew York, NY

Michael J. Choi, MDJohns Hopkins University School of MedicineBaltimore, MD

Michael Emmett, MDBaylor UniversityDallas, TX

Linda F. Fried, MD, MPHUniversity of PittsburghPittsburgh, PA

Richard J. Glassock, MDProfessor Emeritus, The David Geffen Schoolof Medicine at the University of CaliforniaLos Angeles, CA

Kathleen D. Liu, MDUniversity of California San FranciscoSan Francisco, CA

Kevin J. Martin, MBBChSt. Louis University School of MedicineSt. Louis, MO

Rajnish Mehrotra, MDHarbor UCLA Research and Education InstituteTorrance, CA

Patrick T. Murray, MDUniversity College DublinDublin, Ireland

Patrick H. Nachman, MDUniversity of North CarolinaChapel Hill, NC

Aldo J. Peixoto, MDYale UniversityWest Haven, CT

Richard H. Sterns, MDUniversity of Rochester School of Medicineand DentistryRochester, NY

John P. Vella, MDMaine Medical CenterPortland, ME

FOUNDING EDITORSRichard J. Glassock, MD, MACPEditor-in-Chief Emeritus

Robert G. Narins, MD, MACP

PrefaceNephSAP® is one of the three major publications of the American Society of Nephrology(ASN). Its primary goals are self-assessment, education, and the provision of ContinuingMedical Education (CME) credits and Maintenance of Certification (MOC) credits forindividuals certified by the American Board of Internal Medicine. Members of the ASNautomatically receive NephSAP with their monthly issue of The Journal of the AmericanSociety of Nephrology (JASN).

EDUCATION: Medical and Nephrologic information continually accrues at a rapid pace.Bombarded from all sides with demands on their time, busy practitioners, academicians, andtrainees at all levels are increasingly challenged to review and understand all this new material.

Each bimonthly issue of NephSAP is dedicated to a specific theme, i.e., to a specific areaof clinical nephrology, hypertension, dialysis, and transplantation, and consists of an Editorial,a Syllabus, a Commentary on the Syllabus, and self-assessment questions. Over the course of24 months, all clinically relevant and key elements of nephrology will be reviewed and updated.The authors of each issue digest, assimilate, and interpret key publications from the previousissues of other years and integrate this new material with the body of existing information.

SELF-ASSESSMENT: Twenty-five single-best-answer questions will follow the 50 to 75 pagesof Syllabus text. The examination is available online with immediate feedback. Those answer-ing �75% correctly will receive CME credit, and receive the answers to all the questions alongwith brief discussions and an updated bibliography. To help answer the questions, readers maygo to the ASN web site, where relevant material from UpToDate in nephrology will be posted.Thus, members will find a new area reviewed every 2 months, and they will be able to test theirunderstanding with our quiz. This format will help readers stay abreast of developing areas ofclinical nephrology, hypertension, dialysis, and transplantation, and the review and update willsupport those taking certification and recertification examinations.

CONTINUING MEDICAL EDUCATION: Most state and local medical agencies as well ashospitals are demanding documentation of requisite CME credits for licensure and for staffappointments. A maximum of 48 credits annually can be obtained by successfully completingthe NephSAP examination. In addition, individuals certified by the American Board of InternalMedicine may obtain credits towards Maintenance of Certification (MOC) by successfullycompleting the self-assessment portion of NephSAP.

BOARD CERTIFICATION AND INSERVICE EXAMINATION PREPARATION: Each issuewill also contain 5 questions and answers examining core topics in the particular disciplinereviewed in the Syllabus. These questions are designed to provide trainees with challengingquestions to test their knowledge of key areas of nephrology.

� This paper meets the requirements of ANSI/NISO Z39.48-1921 (Permanence of Paper),effective with July 2002, Vol. 1, No. 1.

NephSAP®

©2011 by The American Society of Nephrology

NephSAP®

Fluid, Electrolyte, and Acid-Base Disturbances

Editorial 91We Come to Bury “Contraction Alkalosis,” Not to Praise

It—John H. Galla, MD, and Robert G. Luke, MD

Syllabus 96Fluid, Electrolyte, and Acid-Base Disturbances—Richard H.

Sterns, MD, and Michael Emmett, MD

Acid-Base Disorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .96

Physiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .96

Bicarbonate Levels and Chronic Renal Failure . . . . . . . . . . .96

Acquired Metabolic Acidosis . . . . . . . . . . . . . . . . . . . . . . . . .98

Lactic Acidosis and Septic Shock . . . . . . . . . . . . . . . . . . .98

Lactic Acidosis and Hematologic Malignancy . . . . . . . . . . .100

Metformin and Lactic Acidosis . . . . . . . . . . . . . . . . . . . .100

Anion Gap Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . .103

Propofol-Related Infusion Syndrome and Lactic Acidosis .104

Toxic Alcohols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .105

Asthma: Complex Interaction of Respiratory Alkalosis,Metabolic Acidosis, and Respiratory Acidosis . . . . . . . . .107

Epidemiology of the Anion Gap in “Normal Populations” .109

Topiramate and Hyperchloremic Metabolic Acidosis . . . . .110

Acquired Forms of Metabolic Alkalosis . . . . . . . . . . . . . . .110

Metabolic Alkalosis: The Milk Alkali Syndrome (CalciumAlkali Syndrome) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .110

Pseudo-Bartter (or Bartter-Like) Syndrome . . . . . . . . . . .112

NephSAP®

Volume 10, Number 2, March 2011

Mixed Metabolic Alkalosis and Respiratory Acidosis: Rolefor Acetazolamide? . . . . . . . . . . . . . . . . . . . . . . . . . . . .112

Metabolic Alkalosis (Pseudo-Bartter Syndrome) in CysticFibrosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .114

Pendred Syndrome: Pendrin (SLC26A4) Defects . . . . . . . .115

General Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .117

Hypokalemia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .117

Inherited Forms of Hypokalemia . . . . . . . . . . . . . . . . . . . . .117

Familial Hypokalemic Periodic Paralysis . . . . . . . . . . . . .118

Thyrotoxic Hypokalemic Periodic Paralysis . . . . . . . . . .120

Acquired Forms of Hypokalemia . . . . . . . . . . . . . . . . . . . . .121

Acetaminophen Poisoning . . . . . . . . . . . . . . . . . . . . . . . . .121

High-Dosage Penicillin, Hypokalemia, and MetabolicAlkalosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .122

Hypokalemia from Intestinal Pseudo-obstruction OgilvieSyndrome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .122

Thiazides, Hypertension, and Hypokalemia . . . . . . . . . . .123

Hypokalemia as a Risk Factor in Patients with ChronicKidney Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .124

Hypokalemia-Induced Hyponatremia and Correction ofHyponatremia with K� Replacement. . . . . . . . . . . . .124

Hyperkalemia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .125

Physiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .125

Pseudohyperkalemia . . . . . . . . . . . . . . . . . . . . . . . . . . . . .126

Excess K� Intake . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .127

Internal K� Shifts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .128

Hyperglycemia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .128

Octreotide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .128

Hyperkalemic Periodic Paralysis . . . . . . . . . . . . . . . . . .128

Succinylcholine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .128

NephSAP®


Cardiac Glycosides . . . . . . . . . . . . . . . . . . . . . . . . . . . . .129

Impaired K� Excretion . . . . . . . . . . . . . . . . . . . . . . . . . . .130

Decreased Mineralocorticoid Levels or Activity. . . . . .130

Addison Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .130

Hypoaldosteronism . . . . . . . . . . . . . . . . . . . . . . . . . . . . .130

Pseudohypoaldosteronism. . . . . . . . . . . . . . . . . . . . . . . .130

Drug-Induced Hyperkalemia . . . . . . . . . . . . . . . . . . . . . . .130

Decreased Aldosterone Levels . . . . . . . . . . . . . . . . . . . .130

Aldosterone Receptor Antagonism. . . . . . . . . . . . . . . . .131

Collecting Tubule Sodium Channel Blockade. . . . . . . .132

Consequences of Hyperkalemia . . . . . . . . . . . . . . . . . . . .133

Muscle Weakness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .133

ECG Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .133

Mortality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .133

Metabolic Acidosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . .133

Treatment of Hyperkalemia . . . . . . . . . . . . . . . . . . . . . . .133

Sodium Polystyrene Sulfonate . . . . . . . . . . . . . . . . . . . .134

Fludrocortisone and Glycyrrhetinic Acid. . . . . . . . . . . .136

Nonhypotonic Hyponatremia . . . . . . . . . . . . . . . . . . . . . . . .137

Pseudohyponatremia . . . . . . . . . . . . . . . . . . . . . . . . . . . . .137

Hyperglycemia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .138

Exogenous Solutes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .139

Intravenous Mannitol . . . . . . . . . . . . . . . . . . . . . . . . . . .139

Irrigant Absorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . .139

Hypotonic Hyponatremia: Pathophysiology . . . . . . . . . . . . .140

Acute Hypotonic Hyponatremia . . . . . . . . . . . . . . . . . . . . . .143

Psychotic Polydipsia . . . . . . . . . . . . . . . . . . . . . . . . . . . . .144

NephSAP®


Exercise Hyponatremia . . . . . . . . . . . . . . . . . . . . . . . . . . .144

Obstetric Hyponatremia . . . . . . . . . . . . . . . . . . . . . . . . . . .145

Postoperative Hyponatremia . . . . . . . . . . . . . . . . . . . . . . .145

Neurosurgical Hyponatremia . . . . . . . . . . . . . . . . . . . . . . .146

Treatment of Acute Hyponatremic Emergencies . . . . . . .146

Chronic Hypotonic Hyponatremia . . . . . . . . . . . . . . . . . . . .148

Differential Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . .148

Causes of Renal Salt Wasting . . . . . . . . . . . . . . . . . . . . . .148

Addison Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .148

Congenital Adrenal Hyperplasia . . . . . . . . . . . . . . . . . .148

Cisplatin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .149

Cerebral Salt Wasting. . . . . . . . . . . . . . . . . . . . . . . . . . .149

Causes of Euvolemic Hyponatremia . . . . . . . . . . . . . . . . .150

Tumor-Associated SIADH . . . . . . . . . . . . . . . . . . . . . . .150

Hyponatremic Hypertensive Syndrome . . . . . . . . . . . . .150

Drug-Induced Euvolemic Hyponatremia . . . . . . . . . . . . .150

Cyclophosphamide . . . . . . . . . . . . . . . . . . . . . . . . . . . . .150

Carbamazepine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .151

Selective Serotonin Reuptake Inhibitors . . . . . . . . . . . .151

Other Drugs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .151

Nephrogenic Syndrome of Inappropriate Antidiuresis . . .152

Clinical Outcomes of Chronic Hyponatremia . . . . . . . . . . .153

Mortality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .153

Hospital Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .155

Falls and Fractures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .155

Rhabdomyolysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .157

Osmotic Demyelination Syndrome . . . . . . . . . . . . . . . . . .157

NephSAP®


Treatment of Chronic Hyponatremia . . . . . . . . . . . . . . . . . .160

Renal Replacement Therapy . . . . . . . . . . . . . . . . . . . . . . .161

Vasopressin Antagonists (Vaptans) . . . . . . . . . . . . . . . . .161

Conivaptan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .162

Tolvaptan. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .164

Hypernatremia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .166

Physiology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .166

Essential Hypernatremia . . . . . . . . . . . . . . . . . . . . . . . . . .168

Neurogenic DI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .169

Familial Neurogenic DI . . . . . . . . . . . . . . . . . . . . . . . . .169

Lymphocytic Hypophysitis. . . . . . . . . . . . . . . . . . . . . . .169

IgG4 Disease. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .170

Neurosurgical DI. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .170

Congenital Nephrogenic DI . . . . . . . . . . . . . . . . . . . . . . .170

Secondary NDI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .170

Acquired NDI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .171

Gestational DI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .172

Consequences of Hypernatremia . . . . . . . . . . . . . . . . . . . .172

Mortality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .172

Brain Injury from Hypernatremia . . . . . . . . . . . . . . . . . . .174

Treatment of Hypernatremia . . . . . . . . . . . . . . . . . . . . . . .174

Edema and Diuretics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .176

Cirrhosis and Ascites . . . . . . . . . . . . . . . . . . . . . . . . . . . . .176

Heart Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .179

Nephrotic Syndrome . . . . . . . . . . . . . . . . . . . . . . . . . . . . .180

CME Self-Assessment Questions . . . . . . . . . . . . . . . . . . . . . 182Questions Linked to UpToDate in Green

NephSAP®


Core Knowledge Questions. . . . . . . . . . . . . . . . . . . . . . . . . . 190

Upcoming Issues

Acute Kidney Injury and Critical Care Nephrology—

Patrick T. Murray, MD, and Kathleen D. Liu, MD . . . . . . . . . .May 2011

Renal Pathology—

Glen S. Markowitz, MD, Barry Stokes, MD, Neeraja Kambham, MD, Leal

C. Herlitz, MD, and Vivette D. D’Agati, MD . . . . . . . . . . . . . .July 2011

Chronic Kidney Disease and Progression—

Linda F. Fried, MD, and Michael J. Choi, MD . . . . . . .September 2011

Transplantation—

John P. Vella, MD, and David J. Cohen, MD . . . . . . . .November 2011

NephSAP®


The Editorial Board of NephSAP extends its sincere appreciation to the following reviewers. Their efforts and insights have helped toimprove the quality of this postgraduate education offering.

NephSAP Review PanelNihal Y. Abosaif, MBBChSt. James University HospitalLeeds, United Kingdom

Georgi Abraham, MBBSPondicherry Institute of Medical SciencesMadras Medical MissionChennai, India

Pablo H. Abrego, MD, FASNMarshfield ClinicWausau, WI

Anil K. Agarwal, MD, FASNOhio State University Medical CenterColumbus, OH

Mustafa Ahmad, MDKing Fahad Medical CityRiyadh, Saudi Arabia

Jafar Al-Said, MD, FASNBahrain Specialist HospitalManama, Bahrain

Dante Amato-Martinez, MD, PhDUniversidad Nacional Autonoma de MexicoTlalnepantla, Mexico

Anis U. Ansari, MDMedical AssociatesClinton, IA

Akhtar Ashfaq, MD, FASNNorth Shore University HospitalGreat Neck, NY

Azra Bihorac, MD, FASNUniversity of FloridaGainesville, FL

Mona B. Brake, MDRobert J. Dole VA Medical CenterWichita, KS

Mauro Braun, MDCleveland Clinic FloridaWeston, FL

Chokchai Chareandee, MD, FASNRegions HospitalSaint Paul, MN

W. James Chon, MD, FASNUniversity of Chicago Medical CenterChicago, IL

Devasmita Choudhury, MDUniversity of Texas SouthwesternMedical SchoolDallas, TX

Bulent Cuhaci, MD, FASNDrexel University College of MedicinePhiladelphia, PA

Rajiv Dhamija, MDWalk in Medical CareArtesia, CA

Susan R. DiGiovanni, MDVirginia Commonwealth UniversityRichmond, VA

Francis Dumler, MDWilliam Beaumont HospitalRoyal Oak, MI

Mahmoud T. El-Khatib, MD, PhD, FASNUniversity of Cincinnati Medical CenterCincinnati, OH

Lynda A. Frassetto, MD, FASNUniversity of California at San FranciscoSan Francisco, CA

Duvuru Geetha, MDJohns Hopkins UniversityBaltimore, MD

Carl S. Goldstein, MDRobert Wood Johnson Medical SchoolNew Brunswick, NJ

Nabil G. Guirguis, MDKidney Dialysis and Transplant GroupBridgeport, WV

Pawan K. Gupta, MDAltoona Regional Health SystemAltoona, PA

Carsten Hafer, MDUniversity of HannoverHannover, Germany

Richard N. Hellman, MDIndiana University School of MedicineIndianapolis, IN

Ekambaram M. Ilamathi, MD, FASNSuffolk Nephrology ConsultantsStony Brook, NY

Viswanathan S. Iyer, MD, FASNAKD-HTN LLCHarrisburg, PA

Bernard G. Jaar, MDJohns Hopkins Medical Institutions andNephrology Center of MarylandBaltimore, MD

Avanelle V. Jack, MDLouisiana State University HealthSciences CenterNew Orleans, LA

Sharon L. Karp, MDIndiana University School of MedicineIndianapolis, IN

Pranay Kathuria, MD, FASNUniversity of Oklahoma College of MedicineTulsa, OK

Quresh T. Khairullah, MD, FASNSt. Clair Specialty PhysiciansDetroit, MI

Apurv Khanna, MDSUNY Upstate Medical UniversitySyracuse, NY

Ramesh Khanna, MDUniversity of Missouri at ColumbiaSchool of MedicineColumbia, MO

Edgar V. Lerma, MD, FASNUniversity of Illinois at ChicagoCollege of MedicineChicago, IL

Meyer D. Lifschitz, MDShaare Zedek Medical CenterJerusalem, Israel

Philippe S. Madhoun, MDChu CharleroiCharleroi, Belgium

Jolanta Malyszko, MD, PhD, FASNMedical UniversityBialystok, Poland

Naveed N. Masani, MDWinthrop University HospitalMineola, NY

Hanna W. Mawad, MD, FASNUniversity of Kentucky Medical CenterLexington, KY

Pascal Meier, MD, FASNCentre Hospitalier Universitaire VaudoisLausanne, Switzerland

Beckie Michael, DO, FASNMarlton Nephrology and HypertensionMarlton, NJ

Shahriar Moossavi, MD, PhDWake Forest University BaptistMedical CenterWinston-Salem, NC

Scott R. Mullaney, MDUniversity of California at San DiegoSan Diego, CA

Quaid J. Nadri, MD, FASNKing Faisal Specialist Hospital andResearch CenterRiiyadh, Saudi Arabia

NephSAP®


Suzanne M. Norby, MD, FASNMayo ClinicRochester, MN

Michal Nowicki, MDMedical University of ŁodzŁodz, Poland

Macaulay A. Onuigbo, MD, FASNMayo ClinicEau Claire, WI

Than N. Oo, MDNephrology CenterKalamazoo, MI

Kevin P. O’Reilly, MDColumbus Nephrology, Inc.Columbus, OH

Malvinder S. Parmar, MB, MS, FASNNorthern Ontario School of MedicineTimmins, ON, Canada

Pairach Pintavorn, MD, FASNEast Georgia Kidney and HypertensionAugusta, GA

Paul H. Pronovost, MD, FASNYale University School of MedicineWaterbury, CA

Mohammad A. Quasem, MD, FASNState University of New YorkBinghamton, NY

Wajeh Y. Qunibi, MDUniversity of Texas Health Sciences CenterSan Antonio, TX

Venkat Ramanathan, MD, FASNBaylor College of MedicineHouston, TX

Karthik M. Ranganna, MDDrexel University College of MedicinePhiladelphia, PA

Pawan K. Rao, MD, FASNSt. Joseph’s Hospital Health CenterSyracuse, NY

Joel C. Reynolds, MD, FASNBrooke Army Medical CenterSan Antonio, TX

Robert M.A. Richardson, MDUniversity of TorontoToronto, ON, Canada

Bijan Roshan, MDJoslin Diabetes CenterHarvard Medical SchoolBoston, MA

Abinash C. Roy, MDUniversity of Utah School of MedicineSaint George, UT

Mario F Rubin, MDMassachusetts General HospitalBoston, MA

Ehab R. Saad, MD, FASNMedical College of WisconsinMilwaukee, WI

Mohammad G. Saklayen, MDWright State University Medical SchoolDayton, OH

Ramesh Saxena, MD, PhDUniversity of Texas SouthwesternMedical CenterDallas, TX

Gaurang M. Shah, MDLong Beach VA Healthcare SystemLong Beach, CA

Robert J. Shay, MD, FASNEast Georgia Kidney andHypertension GroupAugusta, GA

Bhupinder Singh, MD, FASNSouthwest Kidney InstituteTempe, AZ

Rolf A.K. Stahl, MDUniversity of HamburgHamburg, Germany

Harold M. Szerlip, MD, FASNMedical College of GeorgiaAugusta, GA

Bekir Tanriover, MDDialysis Nephrology AssociatesDallas, TX

Tushar J. Vachharajani, MD, FASNWake Forest UniversitySchool of MedicineWinston-Salem, NC

Allen W. Vander, MD, FASNKidney Center of South LouisianaThibodaux, LA

Luigi Vernaglione, MDM. Giannuzzi HospitalManduria, Italy

Shefali Vyas, MDSaint Barnabas Medical CenterLivingston, NJ

Alexander Woywodt, MD, FASNLancashire Teaching Hospitals NHSFoundation TrustPreston, United Kingdom

Program Mission and ObjectivesThe mission of the Nephrology Self-Assessment Program (NephSAP) is to regularly provide a vehicle that will be useful for clinicalnephrologists who seek to renew and refresh their clinical knowledge and diagnostic and therapeutic skills. This Journal consists of aseries of challenging, clinically oriented questions based on case vignettes, a detailed Syllabus that reviews recent publications,and an Editorial on an important and evolving topic. Taken together, these parts should assist individual clinicians under-taking a rigorous self-assessment of their strengths and weaknesses in the broad domain of nephrology.

Accreditation and Credit DesignationThe American Society of Nephrology is accredited by the Accreditation Council for Continuing Medical Education to provide con-tinuing medical education for physicians.

The ASN designates this journal-based activity for a maximum of 8.0 AMA PRA Category 1 Credits™. Physicians should only claimcredit commensurate with the extent of their participation in the activity.

Continuing Medical Education (CME) Information

CME Credit: 8.0 AMA PRA Category 1 Credits™

Date of Original Release: March 2011Examination Available Online: on or before Monday, March 7, 2011Audio Files Available: On or before Tuesday, March 15, 2011. A notice will be posted on the ASN website when the audiofiles become available.

CME Credit Eligible Through: February 29, 2012

Answers: Correct answers with explanations will be posted on the ASN website in March 2012 when the issue is archived.UpToDate Links Active: March and April 2011

Core Nephrology question links active: March, April, and May 2011

Target Audience: Nephrology Board and recertification candidates, practicing nephrologists, and internists.

Method of Participation:● Read the syllabus that is supplemented by original articles in the reference lists, and complete the online self-assessment

examination.● Examinations are available online only after the first week of the publication month. There is no fee. Each participant is

allowed two attempts to pass the examination (�75% correct) for CME credit.● Upon completion, review your score and incorrect answers.● Your CME certificate can be printed immediately after completion.● Answers and explanations are provided with a passing score and/or after the second attempt.● CME Credit will be posted to your transcript within 48 hours after checking the attestation box.

Instructions to access the Online Examination and Evaluation:● Go to the ASN website: www.asn-online.org.● Click the CME tab at the top of the homepage.● Click the ASN CME Center button on the left side of the page.● Click on to the ASN CME Center icon.● Login to the ASN website.● Select Claim Credits for the NephSAP topic-activity you would like to complete.● Complete the NephSAP examination.● Complete the evaluation.● Enter the number of CME credits commensurate with your participation in the activity.● Check the box attesting that you have completed this activity.● You can print your CME certificate immediately.● CME credit will be posted to your transcript within 48 hours.● View or print your full transcript anytime at “My CME Center.”

NephSAP®


Instructions to Obtain American Board of Internal Medicine (ABIM) Maintenance of Certification(MOC) Points:

Each issue of NephSAP provides 10 MOC points. Respondents must meet the following criteria:● Be certified by ABIM in internal medicine and/or nephrology and must be enrolled in the ABIM–MOC program

via the ABIM website (www.abim.org).● Take the self-assessment examination within the timeframe specified in this issue of NephSAP.● Designate the issue for MOC points by clicking on the MOC link on the CME certificate page after passing the examination.

You will be leaving the ASN site and transferring the information directly to the ABIM in real-time.● Provide your ABIM Certificate ID number and your date of birth.● You will receive a confirmation message from the ABIM indicating the receipt of your information.

MOC points will be applied to only those ABIM candidates who have enrolled in the program. It is your responsibility to completethe ABIM MOC enrollment process.

Instructions to access the ASN website, NephSAP, and the UpToDate link

Compatible Browser: The ASN website (asn-online.org) has been formatted for cross-browser functionality, and shoulddisplay correctly in all modern web browsers. We recommend that you use Internet Explorer.

Monitor Settings: The ASN website was designed to be viewed in a 1024 � 768 or higher resolution.

Technical Support: If you are having difficulty viewing any of the pages, please refer to the ASN technical support pagefor possible solutions. If you continue having problems, contact Hal Nesbitt at [email protected] provides an additional source of information that should help you answer up to 5 selected NephSAPquestions from each issue. The link is free and will remain active for the first 60 days after publication for eachissue.

● On the ASN home page, double click on the NephSAP link (NephSAP cover) on the bottom side of the page.● On the NephSAP page, click on the UpToDate button on the left hand side to access the current links.

NephSAP®


Disclosure InformationThe ASN is responsible for identifying and resolving all conflicts of interest prior to presenting any educational activity to learners to ensure thatASN CME activities promote quality and safety, are effective in improving medical practice, are based on valid contents, and are independent of thecontrol from commercial interests and free of commercial bias. All faculty are instructed to provide balanced, scientifically rigorous and evidence-based presentations. In accordance with the disclosure policies of the Accreditation Council for Continuing Medical Education (ACCME) as well asguidelines of the Food and Drug Administration (FDA), individuals who are in a position to control the content of an educational activity are re-quired to disclose relationships with a commercial interest if (a) the relation is financial and occurred within the past 12 months; and (b) the individ-ual had the opportunity to affect the content of continuing medical education with regard to that commercial interest. For this purpose, ASN considerthe relationships of the person involved in the CME activity to include financial relationships of a spouse or partner. Peer reviewers are asked toabstain from reviewing topics if they have a conflict of interest. Disclosure information is made available to learners prior to the start of any ASNeducational activity.

Agarwal, Rajiv—Research funding: Abbott; Consultant/scientific advisor: Rockwell Medical, Watson Pharma; Honoraria: Abbott, Astra-Zeneca, MerckCohen, David J.—Research funding: Life Cycle Pharma, Novartis, Roche, Wyeth; Honoraria: Bristol-Myers-Squibb, Novartis, RocheEmmett, Michael—Honoraria: Braintree Laboratories Fresenuis; Editorial board membership: American Journal of Cardiology, Clinical

NephrologyFried, Linda F.—Research funding: Merck, Reata; Honoraria: PfizerFuchs, Elissa (Medical Editor)—noneGlassock, Richard J.—Consultant: Bio-Marin (inactive), Eli Lilly (active), FibroGen (inactive), Genentech (active), Lighthouse Learning (active),

Novartis (active), QuestCor (active), Wyeth (inactive); Ownership interests: LaJolla Pharmaceutical, Reata Inc.; Honoraria: American Society ofNephrology, various medical schools for lectures and/or visiting professor; Membership board of directors/scientific advisor: American RenalAssociates, Los Angeles Biomedical Institute, University Kidney Research Associates (UKRO), Wyeth; Editorial board: UpToDate, AmericanJournal of Nephrology; Royalties: Oxford University Press; Paid expert testimony: Various legal firms regarding product liability

Goldfarb, Stanley—Consultant: Bayer, GE Healthcare; Ownership interests: Polymedix; Honoraria: GE Healthcare, FreseniusLiu, Kathleen D.—Ownership interest: AmgenMartin, Kevin J.—Consultant: Abbott, Cytochroma, Kai, Shire; Honoraria: Abbott, Genzyme, Kai, Shire; Scientific advisor: Abbott,

Cytochroma, KaiMehrotra, Rajnish—Research funding: Amgen, Baxter, Shire; Consultant: Novartis; Honoraria: AMAG, Baxter, Healthcare ShireMurray, Patrick T.—Employment: spouse, Merck, Sharpe, & Dohme (Europe); Consultant/honoraria/research funding/

scientific advisor: Abbott Laboratories (USA), Argutus Medical (UK), FAST Diagnostics (USA); NxStage Medical (USA).Nachman, Patrick H.—Honoraria: QuestCor; Multicenter clinical trial participation: OtsukaPeixoto, Aldo J.—Consultant: Abbott, Sanofi-Aventis; Research funding: Pulsemetric; Honoraria: Boehringer-Ingelheim, Merck, Novartis,

Takeda; Scientific advisor/membership: Associate Editor–Blood Pressure Monitoring; Editorial Board: American Journal of Nephrology,Brazilian Journal of Nephrology

Sterns, Richard H.—Honoraria: Astellas, Otsuka, MerckTownsend, Raymond R.—Consultant: Daiichi-Sankyo, GlaxoSmithKline, Merck, Nicox, Novartis, Roche; Research funding: Novartis;

Honoraria: American Society of Hypertension, National Kidney FoundationVella, John P.—noneEditorial authors:Luke, Robert G.—Scientific advisor; UpToDate; Other; American College of Physicians, Chair, Board of RegentsGalla, John H.—none

Commercial SupportThere is no commercial support for this issue.

NephSAP®


EditorialWe Come to Bury “Contraction Alkalosis,” Not to Praise It

John H. Galla, MD, FACP and Robert G. Luke, MD, MACPUniversity Cincinnati College of Medicine

Metabolic alkalosis (MA) has multiple causes,some associated with volume contraction and somewith volume expansion (1). Moreover, volume con-traction, particularly when severe, is usually associ-ated with lactic acidosis (2). Thus, general clinicalobservations show that volume contraction does notnecessarily lead to MA unless associated with selec-tive chloride (Cl) depletion. MA is associated to vari-able degrees with abnormalities in sodium, potassium(K�), Cl�, and fluid volume depending on the natureof the primary stimulus, the availability of correctivefactors, and further abnormalities generated by theprimary stimulus.

In a seminal article, “contraction” alkalosis byethacrynic acid described in humans gave rise to thehypothesis that extracellular fluid (ECF) volume con-traction produces alkalosis (3). The authors concludedthat the abrupt change in ECF volume was the primaryevent while acknowledging that Cl depletion mightinfluence renal bicarbonate retention.

Our goal is to show that the hypothesis thatmaintenance of MA that is caused by Cl� depletion bythe kidney is dependent on ECF or plasma volumecontraction is not supported by experimental evidence.We argue that the term “contraction alkalosis” ismisleading in both a pathophysiologic and therapeuticsense and suggest that a better term is Cl depletionmetabolic alkalosis (CDMA). This concept stressesthe use of urinary Cl� concentration to categorize themechanism for MA and the therapeutic need for Cl�

containing fluids to correct CDMA.There are basically three important causes of

MA: The most common is Cl� depletion secondary tovomiting or nasogastric aspiration or chloruretic di-uretics; K� depletion (not hypokalemia secondary toan intracellular shift of K� caused by MA); and, leastcommonly, base loading in severe chronic renal failureor in ESRD. In true K� depletion states, the major

mechanism contributing to renal maintenance of MAcaused by K� depletion is increased reabsorption ofHCO3

� in proximal and distal tubules related to in-tracellular acidosis and H�-K�-ATPase–induced K�

conservation in the collecting duct.The normal kidney rapidly restores HCO3

� lev-els to normal after oral or intravenous bicarbonateloading. Metabolic alkalosis, regardless of mecha-nism, is now conveniently divided into three phases:induction, maintenance, and correction. In the steady-state maintenance phase of MA, the kidney, unlike thenormal kidney, fails to correct the serum HCO3

� tonormal by increased urinary HCO3

� excretion. Asnoted, some still claim that the maintenance phase instates of MA associated with Cl� depletion is main-tained by volume contraction. The reasoning behindthis follows.

Historically, in the 1960s and 1970s, Cl� wasregarded as “the mendicant anion,” passively follow-ing sodium across epithelial membranes or the para-cellular pathway. During transepithelial transport, pro-ton secretion and renal tubular HCO3

� absorption wasbelieved to occur via Na�-H� exchange throughoutthe nephron. It was thus logical at that time to believethat maintenance of renal bicarbonate reabsorption inCDMA was driven by enhanced Na�-H� exchange instates of ECF volume contraction, especially in thepresence of reduced Na� reabsorption with Cl�. Itwas concluded that the demands of maintenance ofECF volume in the presence of reduced reabsorptionof NaCl must enhance Na�-H� exchange and thusmaintain NaHCO3 absorption and MA.

This now outmoded concept has continued tobe the view expressed in NephSAP since at least2006 and is restated in this issue in discussion of theneed for “volume expansion” to correct MA com-plicating chronic respiratory acidosis in the inten-sive care unit (page 55). A similar view is expressed

Nephrology Self-Assessment Program - Vol 10, No 2, March 2011

91

in regard to the mechanism of generation and main-tenance of MA in some patients with cystic fibrosiseven though equimolar NaCl (and thus enhancedrelative Cl� depletion) occurs in the sweat glands ofsuch patients (page 58). There are now a plethora ofknown Cl�-HCO3

� anion exchangers and Cl�

channels involved in transepithelial solute transport(4). Furthermore, mutations in these account forvarious congenital electrolyte and acid-base syn-dromes and renal diseases. These specialized tubulesegments, cells, and transport mechanisms in thedistal nephron offer attractive explanations of renaladjustments to acid-base perturbations, independentof sodium balance, and are discussed next.

The pathogenesis of CDMA was studied bySchwartz’s group in carefully done balance studies ofmen and dogs (5). CDMA was produced by gastricaspiration, chloruretic diuretics, NaNO3 infusion (aneffect of “unreabsorbable” anions), and previous hy-percapnia (“post-hypercapnic CDMA”). These studiesestablished unequivocally that Cl� repletion by NaClor KCl—but not replacement of Na� and K losseswithout Cl�—fully corrected CDMA in the mainte-nance phase. The issue of the specific role of ECF

volume depletion was assumed but not resolved at thistime, although it was established that Na� repletionwithout Cl� repletion, regardless of the method ofgenesis of CDMA, clearly did not correct CDMAduring its maintenance phase.

Subsequent studies of men and rats with acuteand chronic CDMA, using careful measurement ofplasma volume changes during the correction phase,established that correction of Na� balance and volumedeficits was not necessary for correction of CDMAand that Cl� repletion was the single, sufficient, andnecessary effect to correct CDMA.

CDMA induced by the combination of furo-semide; administration of Na�, K�, citrate; and di-etary Cl� restriction in normal men was completelycorrected in the maintenance phase of CDMA by oralKCl, with continued restriction of Na�, despite thepresence of maintained negative Na� balance andplasma volume contraction (measured by I131 albuminspace and plasma albumin concentrations), and persis-tently lowered GFR and estimated renal plasma flow(6) (Figure 1).

During correction, net acid excretion decreasedwith HCO3

� diuresis. Sodium phosphate given instead

F id 40 d O l KCl

Na,K citrate Oral NaPO4

FFFFF idddddddd 444444440000000 ddddddd

Na,K citrate Oral NaPO4

Furosemide 40mg qd

35

OOOOOOOOOO l KCKCKKCKCCCCCKCCClOral KCl

Low Cl diet

P HCO3

(meq/L) 30

25

0[0]

Day 1 2 3 4 5 6 7

(-)100

Cumulative Net Cl

Balance (meq)

(-)200

(-)300

Figure 1. Diuretic-induced MA in men. Plasma HCO3 and cumulative Cl balance are shown for each day of the study; K�

balance was neutral (data not shown). Oral KCl started in the middle of day 6 for six men corrected alkalosis at the end ofday 7 with quantitative Cl repletion, whereas oral NaPO4 in two men worsened alkalosis. Reprinted from Galla JH:Chloride-depletion alkalosis. In: Acid-Base Disorders and Their Treatment, edited by Gennari FJ, Adrogue HJ, Galla JH,Madias NE, Boca Raton, Taylor & Francis, 2005, p 523, with permission.

92 Nephrology Self-Assessment Program - Vol 10, No 2, March 2011

of KCl replacement was associated with increasedserum HCO3

� concentration despite increased plasmavolume. In control subjects, furosemide administrationwithout Cl� restriction did not cause CDMA, andserum electrolytes and net acid excretion did notchange with the same amount of KCl administration.

In rat studies, using Cl� depletion produced byperitoneal dialysis (PD) against dialysate containingisotonic NaHCO3 and a K� concentration equivalentto a normal serum K� level, intravenous infusion of an80 mM Cl� solution (the same concentration as in thehypochloremic rat) containing several cations except-ing Na� resulted in the complete correction of CDMA(7). Serial measures of GFR, plasma volume, and Na�

balance showed that this Cl� infusion, without Na� oran increase in plasma volume from its depressed stateand likewise for GFR, corrected CDMA promptly andcompletely. In controls, which received only 5% dex-trose intravenously, bicarbonate excretion continuedto fall, whereas in the CDMA rats, with maintainedplasma volume contraction, negative sodium balance,and low GFR, there was a brisk HCO3

� diuresis(Figure 2).

In other rat studies, we noted that renal Cl con-

servation continued during this Cl infusion untilplasma Cl� normalized (8).

In segmental nephron studies of the PD rat modelof CDMA, we observed that the change in HCO3

�

reabsorption in response to Cl� infusion occurred inthe distal nephron during correction without a signif-icant increase in delivery to these sites as comparedwith rats with maintained CDMA (9). Further supportfor the importance of adaptation of the distal nephronin maintenance and correction was provided by Wes-son (10) and Levine et al. (11).

In vitro studies of the cortical collecting duct(CCD) in rats with acute CDMA showed vigorousHCO3

� secretion, abolished by zero Cl� luminal per-fusion (12). The degree of HCO3

� secretion in vitrowas directly proportional to the degree of Cl� deple-tion and CDMA produced in vivo. Proton ATPase wasdiminished in the apical membrane of the A cell andincreased in the basolateral membrane of the B cell(13). The B-intercalated cell in the CCD seemed to be“poised to secrete” HCO3

� as soon as Cl� is provided,and the proton-secreting A cell activity seemed to beinhibited. Further details of our in vivo and in vitro ratstudies are provided in the original publications andare reviewed in detail elsewhere (14).

During our studies, the molecular nature of theluminal neutral Cl�-HCO3 exchanger in B cells hadnot been identified. Numerous studies have nowshown that it is pendrin, as noted elsewhere in thisissue (page 60). Pendrin is rapidly activated in statesof renal Cl conservation and in MA and inhibited bychloruresis and metabolic and respiratory acidosis(15–17). In CDMA, B cells are activated along withthe relevant transporters, pendrin and basolateral H�-ATPase, and proton secretion by A cells is reduced.Clearly, the increased HCO3

� secretion in the CCDmust be accompanied by reduced HCO3

� reabsorptionin more distal segments of the CCD for HCO3

�

diuresis to occur (18).That the B cell and the CCD may be important in

normal acid-base balance is also supported by thedevelopment of MA in rats subject to oral NaHCO3

loading on a low Cl� diet (11). Feeding babies alow-Cl�, high-alkaline diet has also caused MA, againsuggesting that collecting duct Cl–HCO3 exchangemay be important for excreting HCO3

� load (19).How would the substitution of CDMA for “con-

traction alkalosis” contribute to an approach to treat-ment? First, understanding the true pathophysiology of

Figure 2. Urinary HCO3 excretion in rats with persistentvolume depletion, Na� depletion, and decreased GFR. Ratsreceiving Cl (CC) increased urinary HCO3 excretion asalkalosis was corrected (data not shown), whereas thosereceiving only glucose (DX) had a further decrease whilealkalosis was maintained (data not shown). Reprinted fromGalla JH: Chloride-depletion alkalosis. In: Acid-Base Dis-orders and Their Treatment, edited by Gennari FJ, Ad-rogue, Galla JH, Madias NE, Boca Raton, Taylor & Francis,2005, p 530, with permission.

Nephrology Self-Assessment Program - Vol 10, No 2, March 2011 93

disease usually helps to improve treatment. Second, itfocuses the clinician on the diagnostic value of mech-anisms and results of Cl� loss in the history, clinicalfindings, and laboratory data. Third, it emphasizes thereplacement of the Cl� deficit rather than just expan-sion of ECF and plasma volume. Volume contractionis commonly an associated finding, and infusion ofNaCl is then always appropriate. CDMA often causesshifts of K� into cells and enhanced K� loss in thedistal nephron so that KCl replacement is often alsoneeded. Indeed, in protracted vomiting, continued uri-nary K� losses may eventually lead to true K� deple-tion, which may contribute to maintenance of MA insuch circumstances. Volume contraction is not alwayspresent in CDMA, and then administration of KCl,HCl, and, especially when KCl or HCl is not feasible,acetazolamide plus KCl as needed should be consid-ered. This is discussed further on page 55 in theconsideration of the mixed acid-base disturbance ofchronic respiratory acidosis and MA; the latter isusually CDMA because of chloruretic treatment forcongestive heart failure. All of these treatments aresubject to the limitation as to whether renal function isor becomes adequate to respond to these various ther-apies. In ESRD or preexisting stage 5 chronic kidneydisease, dialysis with a Cl�-rich dialysate may berequired to correct the CDMA.

Other possible benefits of replacement of Cl� inCDMA include an increase in GFR by suppressingtubuloglomerular feedback via the macula densa sig-nal (20). We interpret this reduction of GFR as anattempt to reduce the initial sodium wasting phase ofacute CDMA. Renal concentrating ability may alsoimprove because the delivery of Cl� is rate limitingfor NaCl absorption in the thick ascending limb of theloop of Henle (20). These phenomena, however, havenot been studied in humans. Improvement in responseto loop diuretics has been shown in humans withcorrection of CDMA (21).

In summary, CDMA is corrected by selectiveCl� repletion:(A) Despite

Y Maintained or increasingly negative Na� or K� bal-ance

Y Continued HCO3� loading

Y Continued high levels of angiotensin II or aldosterone(10,22,23)

(B) And is not corrected

Y By Na� or K� repletion without Cl� repletionY By Cl� repletion in the absence of renal function (24)Y By plasma volume expansion by as much as 25% or by

restoring baseline GFR without Cl� repletion

(C) The adaptive corrective response is in

Y The distal nephron by an integrated response in theCCD to ensure a bicarbonate diuresis in response to Clreplacement

We believe that we have made an adequaterationale to finally bury the term “contraction alkalo-sis” and replace it with CDMA!

References1. Luke RG: Metabolic alkalosis: General considerations. In: Acid-Base

Disorders and Their Treatment, edited by Gennari FJ, Adrogue, GallaJH, Madias NE, Boca Raton, Taylor & Francis, 2005, pp 501–518

2. Hood VL: Lactic acidosis. In Acid-Base Disorders and Their Treat-ment, edited by Gennari FJ, Adrogue, Galla JH, Madias NE, BocaRaton, Taylor & Francis, 2005, pp 351–382

3. Cannon PJ, Heinemann HO, Albert MS, Laragh JH, Winters RW:Contraction alkalosis after diuresis of edematous patients withethacrynic acid. Ann Intern Med 62: 979–990, 1965

4. Dorwort M, Shcheynikov N, Yang D, Muallem S: The solute carrier26 family of proteins in epithelial ion transport. Physiology(Bethesda) 23: 104–114, 2008

5. Schwartz WB, van Ypersele de Strihou, Kassirer JP: Role of anionsin metabolic alkalosis and potassium deficiency. N Engl J Med 279:630–639, 1968

6. Rosen RA, Julian BA, Dubovsky EV, Galla JH, Luke RG: On themechanism by which chloride corrects metabolic alkalosis in man.Am J Med 84: 449–458, 1988

7. Galla JH, Bonduris DN, Luke RG: Effects of chloride and extracel-lular fluid volume on bicarbonate reabsorption along the nephron inmetabolic alkalosis in the rat. J Clin Invest 80: 41–50, 1987

8. Galla JH, Bonduris DN, Dumbauld SL, Luke RG: Segmental chlorideand fluid handling during correction of chloride-depletion alkalosiswithout volume expansion in the rat. J Clin Invest 73: 96–106, 1984

9. Galla JH, Bonduris DN, Luke RG: Superficial and deep nephrons in thecorrection of metabolic alkalosis. Am J Physiol 249: F485–F489, 1989

10. Wesson DE: Augmented bicarbonate reabsorption by both the prox-imal and distal nephron maintains chloride-deplete metabolic alkalo-sis in rats. J Clin Invest 84: 1460–1469, 1989

11. Levine DZ, Iacovitti M, Harrison V: Bicarbonate secretion in vivo byrat distal tubules during alkalosis induced by dietary chloride restric-tion and alkali loading. J Clin Invest 87: 1513–1518, 1991

12. Gifford JD, Ware MW, Luke RG, Galla JH: HCO3 transport in ratCCD: Rapid adaptation by in vivo but not in vitro alkalosis. Am JPhysiol 264: F435–F440, 1993

13. Verlander JW, Madsen KM, Galla JH, Luke RG, Tisher CC: Re-sponse of intercalated cells to chloride depletion metabolic alkalosis.Am J Physiol 262: F309–F319, 1992

14. Galla JH: Chloride-depletion alkalosis. In: Acid-Base Disorders andTheir Treatment, edited by Gennari FJ, Adrogue, Galla JH, MadiasNE, Boca Raton, Taylor & Francis, 2005, pp 519–551

15. Amlal H, Petrovic S, Xu J, Wang Z, Sun X, Barone S, Soleimani M:Deletion of the anion exchanger Slc26a4 (pendrin) decreases apicalCl(�)/HCO3(�) exchanger activity and impairs bicarbonate secre-tion in kidney collecting duct. Am J Physiol Cell Physiol 299:C33–C41, 2010

16. de Seigneux S, Malte H, Dimke H, Frøkiaer J, Nielsen S, Frische S:Renal compensation to chronic hypoxic hypercapnia: Downregula-


tion of pendrin and adaptation of the proximal tubule. Am J PhysiolRenal Physiol 292: F1256–F1266, 2007

17. Frische S, Kwon TH, Frøkiaer J, Madsen KM, Nielsen S: Regulatedexpression of pendrin in rat kidney in response to chronic NH4Cl orNaHCO3 loading. Am J Physiol Renal Physiol 284: F584–F593, 2003

18. Galla JH, Rome R, Luke RG: Bicarbonate transport in collecting ductsegments during chloride-depletion alkalosis. Kidney Int 48: 52–55, 1995

19. Roy S III, Arant BS Jr: Alkalosis from chloride-deficient Neo-Mull-Soy. N Engl J Med 301: 615, 1979

20. Galla JH, Bonduris DN, Sanders PW, Luke RG: Volume-independentreductions in glomerular filtration rate in acute chloride-depletionalkalosis in the rat. J Clin Invest 74: 2002–2008, 1984

21. Loon NR, Wilcox CS: Mild metabolic alkalosis impairs the natri-uretic response to bumetanide in normal human subjects [erratumappears in Clin Sci (Colch) 94: 687, 1998]. Clin Sci (Lond) 94:287–292, 1998

22. Walters EA, Rome L, Luke RG, Galla JH: Absence of a regulatoryrole of angiotensin II in acute chloride-depletion alkalosis in the rat.Am J Physiol 261: F741–F745, 1991

23. Kassirer JP, Appleton FM, Chazan JA, Schwartz WB: Aldosterone inmetabolic alkalosis. J Clin Invest 46: 1558–1571, 1967

24. Craig DM, Galla JH, Bonduris DN, Luke RG: Importance of thekidney in the correction of chloride-depletion alkalosis in the rat.Am J Physiol 230: F54–F57, 1986


SyllabusFluid, Electrolyte, and Acid-Base Disturbances

Richard H. Sterns, MD, MACP* and Michael Emmett, MD†

*University of Rochester School of Medicine and Dentistry, Rochester, New York; and†Department of Internal Medicine, Baylor University Medical Center, Dallas, Texas

Learning Objectives:1. To examine recent scientific advances in our un-

derstanding of the pathophysiology of disordersof potassium, acid-base, sodium, and water bal-ance

2. To review how our understanding of pathophysi-ology can be applied to the care of patients

3. To understand how recent clinical trials related tofluid, electrolyte, and acid-base disorders can beapplied to clinical decision-making

Acid-Base Disorders

PhysiologyThe physiology of acid-base disorders was well

reviewed in a previous issue of NephSAP (1), and thereader is referred to that review for general informa-tion. New discoveries in acid-base physiology in thepast few years are discussed in the context of specific,relevant clinical syndromes; more in-depth informa-tion about recent advances in the understanding ofacid-base physiology is available in recent reviews.An update of kidney acid-base regulation was pub-lished by Koeppen in 2009 (2), and the physiology ofkidney ion transport, including acid-base, was re-viewed extensively (3).

References1. Sterns RH, Palmer BF (eds): Fluid, electrolyte, and acid-base distur-

bances. NephSAP 6: 210–272, 20072. Koeppen BM: The kidney and acid-base regulation. Adv Physiol Educ

33: 275–281, 20093. Pflugers Arch 458: 1–222, 2009

Bicarbonate Levels and Chronic Renal FailureOne of the most important recent developments

in the field of metabolic acidosis is the accumulatingevidence that the treatment of metabolic acidosis mayslow the progressive deterioration of kidney function,

often a seemingly inevitable aspect of the chronicrenal disease syndrome. More than 25 years ago, Nathet al. (1) showed that high ammonia/ammonium con-centrations in the kidney could be toxic and thatammonium-related nephrotoxicity is a potential resultof the activation of the alternative complement cas-cade via amidination of C3 (2,3). Although the totaldaily quantity of excreted urine net acid and ammo-nium falls as kidney function decreases, the ammo-nium concentration within each nephron unit increasesand systemic acidosis further increases the ammoniumconcentration in the kidney tissue. Nath et al. studiedrats subjected to 13⁄4 nephrectomy, which then expe-rienced progressive kidney dysfunction. The adminis-tration of sodium bicarbonate supplements to rats withchronic renal failure reduced renal ammonia synthesis,ammonium excretion, and ammonium concentrationwithin the renal tubules. Such findings had been pre-viously described. However, the discovery that sodiumbicarbonate supplements also reduced urine proteinexcretion was unexpected. In addition, other markersof renal tubule dysfunction improved (1). Subse-quently, Torres et al. (3) studied a strain of rats thatdeveloped an inherited form of progressive renal cys-tic disease. Feeding NH4Cl to these rats worsened thecystic disease process and accelerated the progressionrate of kidney dysfunction, but the administration ofsodium bicarbonate supplements decreased urine am-monium excretion and markedly reduced progressionof the cystic disease and the severity of interstitialinflammation.

Studies of other rat models of progressive kidneydisease have not always been consistent. Some havenot confirmed the beneficial effects of sodium bicar-bonate supplementation; in some experiments, meta-bolic acidosis actually seemed to be protective andslowed the progression rate of kidney failure (4,5).


96

The mechanism of acidosis protection has been hy-pothesized to be related to increased solubility ofcalcium phosphate salts in an acid environment. Thus,metabolic acidosis could potentially have both adverseand beneficial effects. Nonetheless, the current con-sensus is that the adverse effects of metabolic acidosispredominate.

In a recent issue of the Journal of the AmericanSociety of Nephrology, de Brito-Ashurst et al. (6) ad-dressed this question in humans with chronic kidneydisease (CKD). They randomly assigned 134 adult pa-tients with CKD (measured creatinine clearance [CrCl]between 15 and 30 ml/min per 1.73 m2) and metabolicacidosis (serum bicarbonate level between 16 and 20mEq/L) to receive either standard care only or standardcare plus oral sodium bicarbonate supplements for a2-year period. The final dose of oral sodium bicarbonateaveraged 1.82 � 0.80 g/d (approximately 22 mEq/d).Patients who received sodium bicarbonate developedhigher serum bicarbonate and lower serum potassiumlevels. The most important findings were that the bicar-bonate-supplemented group had a slower decline in theirmeasured CrCl (1.9 versus 5.9 ml/min/1.73 m2), wereless likely to experience a rapid deterioration of GFR (9versus 45%), and fewer of them developed end-stagekidney disease (6.5 versus 33.0%). A number of nutri-tional parameters, including dietary protein intake, nor-malized protein nitrogen appearance, serum albumin, andmid-arm muscle circumference, all were improved in thebicarbonate-supplemented group. The bicarbonate ther-apy was moderate in dosage and was very well tolerated.It has been known that patients with renal disease andthose with hypertension tolerate exogenous sodium bi-carbonate much better than equimolar exogenous NaClsupplementation; less weight gain, edema, and BP ele-vation occur in bicarbonate-supplemented groups, com-pared with NaCl-supplemented groups (7,8,9).

Similar conclusions were reached in a smallerstudy by Phisitkul et al. (10), in which GFR was notdirectly measured. Phisitkul et al. found that basesupplementation (with sodium citrate) slowed the rateof GFR decline as estimated from both the serumcreatinine concentrations (with the Modification ofDiet in Renal Disease [MDRD] equation) and thecystatin C levels. The yearly decline of the MDRD-derived estimated GFR (eGFR) was 1.60 � 0.13ml/min per 1.73 m2 in the treated group versus 3.79 �0.30 ml/min per 1.73 m2 in the control group. TheGFR decline rates estimated from the cystatin C level

showed similar results: 1.82 � 0.08 ml/min in thetreated group versus 4.38 � 0.98 in the control. Twomarkers of tubulointerstitial disease, urine endothe-lin-1 excretion (a surrogate of kidney endothelin pro-duction) and N-acetyl-beta-D-glucosaminidase, werealso measured; excretion rates fell in patients whowere treated with sodium citrate. Shah et al. (11)provided additional evidence for the adverse impact ofmetabolic acidosis on kidney function deteriorationrate, retrospectively analyzing 5422 outpatients witheGFRs of �60 ml/min per 1.73 m2. Compared with areference bicarbonate baseline level of 25 to 26mEq/L, patients with lower or higher bicarbonatelevels were more likely to have a major reduction inkidney function, defined by an eGFR decrease of 30%or a doubling of serum creatinine level.

Independent of the question, “Can oral bicarbon-ate salts and bicarbonate precursors slow the rate ofrenal deterioration?” there is ample evidence that al-kalinizing salts have multiple beneficial effects onprotein metabolism, endocrine function, muscle func-tion, skeletal structure, bone density, and other phys-iologic parameters in patients with chronic metabolicacidosis of various causes. This is likely to be true forpatients with moderate degrees of CKD as well aspatients with more severe kidney disease, and forthose who require dialysis.

Interpretation of the serum bicarbonate is con-founded by multiple factors, some beneficial and oth-ers harmful, that affect bicarbonate concentration.These factors can have opposite effects on survivaland morbidity. From the research described, it is clearthat metabolic acidosis has multiple negative effects,but lower bicarbonate levels could also reflect excel-lent nutrition because higher protein intake generateslarger acid loads. For patients with renal disease and alower bicarbonate level as a result of higher proteinintake, this may actually predict better survival, al-though lower bicarbonate levels could also reflectsystemic inflammation, which would adversely affectmorbidity and survival. Conversely, higher bicarbon-ate levels may have a number of beneficial effects, butin cases in which a higher level reflects poor proteinintake or acid loss as a result of vomiting, higherbicarbonate levels may portend a poor outcome. In2006, Wu et al. (12) evaluated baseline predialysisserum bicarbonate levels (averaged over the first 3months of dialysis treatments) to determine whetherthey could predict 2-year mortality rates in 56,385


maintenance hemodialysis patients. The study con-cluded that the “optimal” predialysis serum bicarbon-ate level associated with the lowest unadjusted mor-tality rate in these patients was between 17 and 23mEq/L. After adjustment for the confounding benefi-cial acidifying impact of nutrition and the adverseimpact of inflammation, a bicarbonate level of �23mEq/L predicted the highest survival. A subsequentstudy directed by the same senior author, Kalantar-Zadeh, evaluated this question in patients with predi-alysis CKD (13); patients with serum bicarbonatelevels in the range of 26 to 29 mEq/L had the lowestmortality rate, whereas those with lower or higherbicarbonate levels had higher mortality rates. In 2010,Kraut and Madias (14) reviewed the consequences ofmetabolic acidosis in patients with renal disease andthe risks and benefits of therapy. They agree with therecommendation of the National Kidney FoundationKidney Disease Outcomes Quality Initiative (NKF-KDOQI) that serum HCO3 be increased to �22mEq/L and/or the recommendation of the Caring forAustralians with Renal Impairment (CARI) panel thatHCO3 be increased to �22 mEq/L. Their report ap-propriately cautions that HCO3 levels should not beincreased above the normal range, at the risk of in-creased fatality (14).

de Brito-Ashurst et al. (6) reported that oralsodium bicarbonate supplements at a dos-age of approximately 22 mEq/d (and stan-dard therapy) slowed the decline in mea-sured CrCl from 5.9 to 1.9 ml/min per 1.73m2. Fewer patients developed end-stagekidney disease (6.5 versus 33.0%). Nutri-tional parameters, including dietary intake,normalized protein nitrogen appearance,serum albumin, and mid-arm muscle cir-cumference, improved in the bicarbonate-supplemented group.

References1. Nath KA, Hostetter MK, Hostetter TH: Pathophysiology of chronic

tubulo-interstitial disease in rats: Interactions of dietary acid load,ammonia, and complement component C3. J Clin Invest 76: 667–675, 1985

2. Tang Z, Sheerin N: Complement activation and progression ofchronic kidney disease. Hong Kong J Nephrol 11: 41–46, 2009

3. Torres VE, Mujwid DK, Wilson DM, Holley KH: Renal cystic

disease and ammoniagenesis in Han:SPRD rats. J Am Soc Nephrol 5:1193–1200, 1994

4. Throssell D, Brown J, Harris KP, Walls J: Metabolic acidosis doesnot contribute to chronic renal injury in the rat. Clin Sci (Lond) 89:643–650, 1995

5. Jara A, Felsenfeld AJ, Bover J, Kleeman CR: Chronic metabolicacidosis in azotemic rats on a high-phosphate diet halts the progres-sion of renal disease. Kidney Int 58: 1023–1032, 2000

6. de Brito-Ashurst I, Varagunam M, Raftery MJ, Yaqoob MM: Bicar-bonate supplementation slows progression of CKD and improvesnutritional status. J Am Soc Nephrol 20: 2075–2084, 2009

7. Husted FC, Nolph KD: NaHCO3 and NaCl tolerance in chronic renalfailure II. Clin Nephrol 7: 21–25, 1977

8. Husted FC, Nolph KD, Maher JF: NaHCO3 and NaC1 tolerance inchronic renal failure. J Clin Invest 56: 414–419, 1975

9. Luft FC, Zemel MB, Sowers JA, Fineberg NS, Weinberger MH:Sodium bicarbonate and sodium chloride: Effects on blood pressureand electrolyte homeostasis in normal and hypertensive man. J Hy-pertens 8: 663–670, 1990

10. Phisitkul S, Khanna A, Simoni J, Broglio K, Sheather S, Rajab MH,Wesson DE: Amelioration of metabolic acidosis in patients with lowGFR reduced kidney endothelin production and kidney injury, andbetter preserved GFR. Kidney Int 77: 617–623, 2010

11. Shah SN, Abramowitz M, Hostetter TH, Melamed ML: Serumbicarbonate levels and the progression of kidney disease: A cohortstudy. Am J Kidney Dis 54: 270–277, 2009

12. Wu DY, Shinaberger CS, Regidor DL, McAllister CJ, Kopple JD,Kalantar-Zadeh K: Association between serum bicarbonate and deathin hemodialysis patients: Is it better to be acidotic or alkalotic? ClinJ Am Soc Nephrol 1: 70–78, 2006

13. Kovesdy CP, Anderson JE, Kalantar-Zadeh K: Association of serumbicarbonate levels with mortality in patients with non-dialysis-depen-dent CKD. Nephrol Dial Transplant 24: 1232–1237, 2009

14. Kraut JA, Madias NE: Consequences and therapy of the metabolicacidosis of chronic kidney disease. Pediatr Nephrol 26: 19–28, 2011

Acquired Metabolic Acidosis

Lactic Acidosis and Septic ShockIn 2001, Rivers et al. (1) published a study of

“early goal-directive therapy in the treatment of severesepsis and septic shock.” This study and the subse-quent Surviving Sepsis Campaign (2) have had aprofound impact on emergency treatment of patientswith possible sepsis. The diagnosis of severe sepsisrequires evidence of infection and two or more Sys-temic Inflammatory Response Syndrome (SIRS) cri-teria:

1. Temperature �38 or �36°C2. Heart rate �90 beats/min3. Respiratory rate �20 breaths/min or PCO2 �32

mmHg4. WBC greater than 12,000/cc or less than 4,000/cc,

or greater than 10% bands.A diagnosis of septic shock requires either a meanarterial BP �65 or a systolic BP �90 mmHg thatpersists after IV expansion, or a blood lactate level �4mEq/L regardless of the blood pressure. Although


some controversy has developed regarding the abso-lute beneficial impact on patient outcomes of treat-ment based on these guidelines, there is no doubt thatthis campaign has generated an enormous number of“stat” blood lactic acid measurements.

Chawla et al. (3,4) investigated whether an in-creased anion gap level could substitute for the directmeasurement of lactic acid in critically ill patients,concluding that the inherent variability of the aniongap calculation proved to be too large to predictreliably relatively small lactic acid increases. Criti-cally ill patients often have hypoalbuminemia, whichconfounds the anion gap calculation. Although theanion gap corrected for albumin is a better lactic acidlevel predictor, the correction only marginally im-proved the predictive value of the anion gap. Berkmanet al. (5) studied 1419 emergency department patientsand reached similar conclusions. An anion gap of �12mEq/L in a patient who presented to the emergencydepartment with clinically suspected infection had asensitivity of 80% and a specificity of 69% for pre-dicting a lactate level of �4 mEq/L. They concluded,“This information may be somewhat helpful to emer-gency physicians to risk-stratify their patients to pro-vide more aggressive early resuscitation” (5). Thus,the anion gap cannot substitute for direct blood lacticacid measurements when relatively small increases areof critical importance (3–6).

After the diagnosis of septic shock has beenestablished, a related issue is whether serial monitor-ing of lactate levels provides any clinical benefits.Jansen et al. (7) reviewed the literature and concludedthat although lactate levels are a helpful risk-stratifi-cation tool for critically ill patients, they are not usefulas “resuscitation end points.” This question was stud-ied directly by Jones et al. (8), who conducted amulticenter, randomized, noninferiority trial to exam-ine sequential lactate levels as an index of therapeuticadequacy and to determine whether lactate levels mayhelp to direct subsequent therapy. Three hundred pa-tients with early severe sepsis or septic shock wererandomly assigned to either (1) a group in whichresuscitation efforts were directed at normalization ofcentral venous pressure, mean arterial pressure, and anScvO2 of at least 70% or (2) a group that receivedidentical treatments and also used each patient’s “lac-tate clearance” measurement to direct therapy. Lactateclearance is calculated as 100 � (initial lactate �delayed lactate)/initial lactate, where the initial lactate

is the measurement obtained at the start of resuscita-tion and delayed lactate is a repeat measurement ob-tained at least 2 hours after the start of resuscitation.The therapeutic goal was a lactate clearance of at least10%. Lower lactate clearance rates required the addi-tion of sequentially added interventions, includingblood transfusions, dobutamine infusion, etc. Effortsdirected at improving lactate clearance did not reducein-hospital mortality.

When patients with sepsis syndrome developlactic acidosis, is lactate the primary or entire cause ofthe metabolic acidosis and the increased anion gap? In1991, Mecher et al. (9) evaluated the anion gap ele-vation in patients with severe sepsis and, after consid-ering the impact of elevated phosphate and urateconcentrations and changes in serum protein concen-trations on the anion gap, concluded that lactate couldaccount for only approximately 50% of the residualincrease. More recently in 2008, Moviat et al. (10)attempted to identify all of the acid anions that accu-mulated in 31 critically ill patients with metabolicacidosis. They used multiple analytic methods includ-ing standard clinical laboratory methods, ion-ex-change column chromatography, reverse-phase HPLC,and gas chromatography–mass spectrometry. Again, alarge fraction of the anion gap remained unidentified;additional discussion of the “missing component” ofthe gap is provided in the Anion Gap Compositionsection.

In addition to anion gap acidosis, which is soprevalent in the patient with severe sepsis and septicshock, the frequency and severity of hyperchloremicmetabolic acidosis seems to have increased, probablyreflecting therapeutic changes. This acid-base disorderwas emphasized in a study of patients with severesepsis and septic shock at the time they were admittedto the intensive care unit (11). Hyperchloremic acido-sis accounted for between 5 and 10 mEq/L of thereduction in bicarbonate and was most likely gener-ated by the aggressive use of intravenous isotonicsaline during the patient’s emergency departmenttreatment. As noted, large-volume isotonic saline isnow standard of care for most patients with severesepsis and septic shock. Furthermore, the patients whohad sepsis and survived had developed less severehyperchloremic acidosis than the patients who died.Although it seems likely that these sicker patientswould have received more intravenous fluid, differ-ences in the volume of intravenous isotonic saline


between these groups could not entirely be attributedto the severity of the illness or entirely explain thedevelopment of hyperchloremic acidosis (11).

References1. Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B,

Peterson E, Tomlanovich M, Early Goal-Directed Therapy Collabor-ative Group: Early goal-directed therapy in the treatment of severesepsis and septic shock. N Engl J Med 345: 1368–1377, 2001

2. Dellinger RP, Levy MM, Carlet JM, Bion J, Parker MM, Jaeschke R,Reinhart K, Angus DC, Brun-Buisson C, Beale R, Calandra T,Dhainaut JF, Gerlach H, Harvey M, Marini JJ, Marshall J, Ranieri M,Ramsay G, Sevransky J, Thompson BT, Townsend S, Vender JS,Zimmerman JL, Vincent JL: International Surviving Sepsis Cam-paign Guidelines Committee; American Association of Critical-CareNurses; American College of Chest Physicians; American College ofEmergency Physicians; Canadian Critical Care Society; EuropeanSociety of Clinical Microbiology and Infectious Diseases; EuropeanSociety of Intensive Care Medicine; European Respiratory Society;International Sepsis Forum; Japanese Association for Acute Medi-cine; Japanese Society of Intensive Care Medicine; Society of CriticalCare Medicine; Society of Hospital Medicine; Surgical InfectionSociety; World Federation of Societies of Intensive and Critical CareMedicine. Surviving Sepsis Campaign: International guidelines formanagement of severe sepsis and septic shock: 2008. Crit Care Med36: 296–327, 2008

3. Chawla LS, Shih S, Davison D, Junker C, Seneff MG: Anion gap,anion gap corrected for albumin, base deficit and unmeasured anionsin critically ill patients: Implications on the assessment of metabolicacidosis and the diagnosis of hyperlactatemia. BMC Emerg Med 8:18, 2008

4. Chawla LS, Jagasia D, Abell LM, Seneff MG, Egan M, Danino N,Nguyen A, Ally M, Kimmel PL, Junker C: Anion gap, anion gapcorrected for albumin, and base deficit fail to accurately diagnoseclinically significant hyperlactatemia in critically ill patients. J In-tensive Care Med 23: 122–127, 2008

5. Berkman M, Ufberg J, Nathanson LA, Shapiro NI: Anion gap as ascreening tool for elevated lactate in patients with an increased risk ofdeveloping sepsis in the Emergency Department. J Emerg Med 36:391–394, 2009

6. Emmett M: Anion gap, anion gap corrected for albumin, and basedeficit fail to accurately diagnose clinically significant hyperlac-tatemia in critically ill patients. J Intensive Care Med 23: 350, 2008

7. Jansen TC, van Bommel J, Bakker J: Blood lactate monitoring incritically ill patients: A systematic health technology assessment. CritCare Med 37: 2827–2839, 2009

8. Jones AE, Shapiro NI, Trzeciak S, Arnold RC, Claremont HA, KlineJA, Emergency Medicine Shock Research Network: Lactate clear-ance vs central venous oxygen saturation as goals of early sepsistherapy: A randomized clinical trial. JAMA 303: 739–746, 2010

9. Mecher C, Rackow EC, Astiz ME, Weil MH: Unaccounted for anionin metabolic acidosis during severe sepsis in humans. Crit Care Med19: 705–711, 1991

10. Moviat M, Terpstra AM, Ruitenbeek W, Kluijtmans LA, Pickkers P,van der Hoeven JG: Contribution of various metabolites to the“unmeasured” anions in critically ill patients with metabolic acidosis.Crit Care Med 36: 752–758, 2008

11. Noritomi DT, Soriano FG, Kellum JA, Cappi SB, Biselli PJ, LiborioAB, Park M: Metabolic acidosis in patients with severe sepsis andseptic shock: A longitudinal quantitative study. Crit Care Med 37:2733–2739, 2009

Lactic Acidosis and Hematologic MalignancyPatients with extensive lymphoma and leukemia

may develop persistent lactic acidosis that is not read-ily explained by overt tissue hypoxia. In 2007, Frie-denberg et al. (1) described this disorder in sevenpatients, five with lymphomas and two with chroniclymphocytic leukemia; this association has been re-ported in multiple studies in the past few years (2–8).In some, the liver was extensively infiltrated withtumor, and reduced hepatic lactate uptake probablyplayed a major role. In other studies, however, liverstatus was normal. It is likely that tumor cells wererapidly synthesizing lactic acid. Tumor cell glycolyticenzyme activity may be increased and high levels ofIGF and/or TNF have been implicated. In many ofthese cases, the lactic acidosis is associated with hy-poglycemia that is refractory to glucose infusion(2,4,8). The prognosis of these patients is generallyvery poor. Thiamine and/or riboflavin deficiency alsocan contribute to development of lactic acidosis, andFriedenberg et al. (1) suggested that these vitamins beadministered to patients with this condition.

References1. Friedenberg AS, Brandoff DE, Schiffman FJ: Type B lactic acidosis as

a severe metabolic complication in lymphoma and leukemia: A caseseries from a single institution and literature review. Medicine (Balti-more) 86: 225–232, 2007

2. Luscri N, Mauer M, Sarafoglou K, Moran A, Tolar J: Lactic acidosisand hypoglycemia with ALL relapse following engrafted bone marrowtransplant. Pediatr Blood Cancer 53: 223–225, 2009

3. Jung B, Zoric L, Chanques G, Konate A, Nocca D, Jaber S: Acuteabdomen and severe lactic acidosis can lead to a surprising diagnosis.Intensive Care Med 36: 169–170, 2010

4. Diaz J, Antoine J, Azad N: A case of hypoglycemia, lactic acidosis,and hematologic malignancy. Endocr Pract 16: 241–243, 2010

5. Kestler MH, Gardner EM, Cohn DL: Hepatic non-Hodgkin’s lym-phoma with lactic acidosis in with HIV infected patients: Report of 2cases. J Int Assoc Physicians AIDS Care (Chic) 9: 301–305, 2010

6. Chan FH, Carl D, Lyckholm LJ: Severe lactic acidosis in a patient withB-cell lymphoma: A case report and review of the literature. CaseReport Med January 4, 2010 [epub ahead of print]

7. Lee HS, Kim HJ, Choi S, Kim CK, Lee NS, Lee KT, Won JH, ParkHS, Hong DS: A case of type B lactic acidosis in acute leukemia.Yonsei Med J 51: 460–462, 2010

8. Keller BC, Nussensveig D, Dowell JE: Diffuse large B-cell lymphomain a hepatitis C virus-infected patient presenting with lactic acidosisand hypoglycemia. Am J Med Sci 339: 202–204, 2010

Metformin and Lactic AcidosisMetformin is an extremely effective drug for

type 2 diabetes and has become a cornerstone oftherapy for these patients. However, its use is associ-ated with the development of severe and often fatallactic acidosis, especially when the drug is given to


patients with renal insufficiency. Under those circum-stances, it can accumulate to toxic levels and generatemetformin associated lactic acidosis (MALA), an ex-tremely rare condition (approximately 0.03 cases/1000patient-years, with approximately 0.015 fatal cases/1000 patient-years) (1). When lactic acidosis occurs ina patient who has diabetes and is taking the drug, he orshe almost always has multiple comorbidities, is tak-ing multiple medications, and is usually critically ill.Therefore, the exact etiologic role played by met-formin in the development of the lactic acidosis oftenremains uncertain (hence the acronym is MALA andnot MILA for “metformin induced lactic acidosis”).Several recent publications have addressed the issue ofcausality and relative frequency of this complication.Bodner et al. (2) performed a nested case-controlanalysis using the UK General Practice Research Da-tabase and confirmed a lactic acidosis rate of approx-imately 0.03 cases/1000 patient-years for metformin-treated patients, but they also found an even higherrate of lactic acidosis (0.048 cases/1000 patient-years)among sulfonylurea-treated patients. Salpeter et al. (3)published a Cochrane Database Systematic Review ofpooled data from 347 comparative trials and cohortstudies, which compared 70,490 patient-years of met-formin use with 55,451 patients-years of treatmentwith non-metformin antidiabetic drugs. No case oflactic acidosis occurred in either group. They con-cluded that there was no statistical difference in thelactic acidosis rate between the groups. Kamber et al.(4) reported data from the Fremantle Diabetes Study(FDS) in Australia, where 1279 patients with type 2diabetes were observed over 12,466 patient-years.Five confirmed cases of lactic acidosis were identified,and four of five patients had other derangements thatcould have caused the acidosis (e.g., cardiogenicshock). Again, there was no significant differencebetween the occurrence rate of lactic acidosis in pa-tients who were and were not taking metformin. How-ever, the relatively small number of patients followedand the very low rate of lactic acidosis development inthis study means that a small increased risk may havebeen missed. Lalau (5) reviewed the literature through2009 and came to the same conclusion.

Metformin is very efficiently excreted in theurine of patients with normal kidney function. Itsclearance is 300 to 400% of the GFR, and 90% of anoral dose is excreted by the kidneys (1). Consequently,high systemic levels develop only when renal function

is impaired or major overdose occurs. Concern aboutmetformin accumulation, and the potential for devel-opment of lactic acidosis caused the Food and DrugAdministration to mandate a “black box” warningabout this complication. The drug is currently contra-indicated for men with a serum creatinine level of�1.5 mg/dl and for women with a serum creatininelevel of �1.4 mg/dl. It also must be stopped for 48hours whenever an iodinated contrast study is donebecause of concern that acute renal damage may de-velop in patients with diabetes. Significant heart fail-ure and/or liver disease are additional strong relativecontraindications to use of the drug.

Despite these very clear formal contraindications(which may be too stringent), many patients withunequivocal contraindications are still treated, eitherbecause physician and patient believe the benefit ofmetformin therapy outweighs the risk or because ther-apeutic errors have occurred. For example, in the FDS(4), up to 38% of patients who were taking the drughad contraindications. The percentage of patients whohad contraindications and were treated with metforminwas even higher in a recent study from Thailand (6); ofpatients who had type 2 diabetes and had clear-cutcontraindications to metformin, 84% still received thedrug. It is worth noting that despite the frequent clearviolation of the manufacturer’s recommendations theincidence of lactic acidosis was not increased abovethe level which was seen in the non-metformin usingcontrol patients in the FMS (4). Pongwecharak et al.(6) also could not identify any case of lactic acidosis.These accumulating findings have led to a number ofproposals that the strict GFR restrictions on metforminbe modified and liberalized. For example, Shaw et al.(7) proposed that metformin prescribing be based onestimated GFR (eGFR) results and not serum creati-nine concentration. They further propose that met-formin use should be permitted for NKF stage IIkidney disease (eGFR between 30 and 59 ml/min/1.73M2) and that the drug should only be contraindi-cated for patients with an eGFR less than 30 ml/min/1.73M2. Herrington and Levy (8) advanced similarrecommendations but suggested that the starting andmaximal metformin doses be decreased by 50% whenthe GFR is between 30 and 60 ml/min per 1.73 m2

and/or when treating elderly patients. Philbrick et al.(9) also concurred, emphasizing that the benefit ofusing metformin for any patient must always be bal-anced with the risk, thereby deeming absolute rules


that are based on creatinine concentration inappropri-ate and harmful. Tahrani et al. (10) also recommendedliberalization of other treatment contraindications,such as heart failure.

Despite all of these data regarding the safety ofmetformin use and lack of toxicity in patients withnormal renal function or modestly reduced renalfunction, most researchers believe that metformincan produce lactic acidosis when it reaches veryhigh systemic concentrations. High concentrationsinhibit mitochondrial uptake and metabolism of py-ruvate, bind to mitochondrial membranes, and blockoxidative phosphorylation, so the association is sci-entifically plausible. Very high metformin levelsgenerally occur in one of three settings: The drug issometimes inappropriately prescribed for patientswith severe or end-stage kidney disease (11), pa-tients with relatively normal kidney function maydevelop acute kidney injury (e.g., volume depletionfrom nausea and vomiting) and continue to take thedrug (12), and there are cases of accidental (13) andintentional overdose (14 –18). When massive over-doses of metformin are ingested, measured met-formin levels are extremely high; some but not all ofthese patients develop lactic acidosis, and manydevelop hemodynamic instability.

When patients present with presumed met-formin toxicity, lactic acidosis, and/or hypotensionand the metformin levels are believed to be veryhigh, efforts to remove metformin are indicated.The high plasma levels must be presumed because

results of measurement will be delayed by manydays or more. Usually, renal function is markedlyreduced or else high metformin levels would nothave developed; therefore, extracorporeal removalmodalities are often used. Metformin is minimallybound to proteins and is readily dialyzable, but aprolonged period of dialysis will be required be-cause this drug has a very large distribution volumeand is distributed to multiple compartments (17–19). Seidowsky et al. (18) have accumulated a greatdeal of data on this topic; Table 1 and Figure 1 fromtheir article are shown. Important points from thisstudy are that patients with intentional overdose areyounger and have higher metformin levels but lesssevere lactic acidosis and excellent survival rates.Those who accidentally overdose, by definitionhave diabetes, and are also older and sicker andhave worse kidney, and liver dysfunction. Theydevelop more severe lactic acidosis and are morelikely to die (58%), despite aggressive intervention(18). Figure 1 shows that prolonged hemodialysis(�15 hours) is required to reduce the metforminlevels by 90%. A postdialysis rebound will occur. Ifthe patient is too unstable for hemodialysis, then acontinuous slow form of hemofiltration or dialysiscan be used (15). Although individual cases oftreatment with plasma exchange and charcoal he-moperfusion have been published, the efficacy ofthese modalities is unclear, and hemodialysis con-tinues to be the standard method of extracorporealremoval (15,17).

Table 1. Comparison between the intentional overdose group (group 1) and the accidental overdose group(group 2)

Parameter Group 1 (n � 13) Group 2 (n � 29) P

Age (y) 45 (10; 16–81) 69 (19; 49–81) �0.0001Total LODS score (points) 3 (3; 0–13) 9 (4; 0–14) �0.0001Probability of mortality (%) 3 (3; 3–88) 58 (39; 3–92) �0.0001pH 7.34 (0.07; 7.15–7.39) 6.90 (0.22; 6.60–7.35) �0.0001Plasma bicarbonate (mmol/L) 18 (6; 8–21) 5 (5; 2–21) �0.0001Blood arterial lactate (mmol/L) 6 (7; 2–18) 12 (8; 4–30) 0.001Plasma anion gap (mmol/L) 20 (5; 13–34) 41 (12; 19–50) �0.0001Blood urea (mmol/L) 4 (6; 0.5–20.0) 22 (13; 1–83) 0.0003Serum creatinine (�mol/L) 79 (48; 53–500) 523 (366; 35–1126) �0.0001Plasma metformin (mg/L) 33 (43; 3–83) 9 (5; 3–16) 0.0002Plasma potassium (mmol/L) 4.1 (0.7; 3.4–8.7) 5.8 (1.6; 3.3–7.7) �0.0001Prothrombin activity (%) 95 (10; 32–100) 50 (41; 10–109) �0.0001

Reprinted from reference 18 (Seidowsky A, Nseir S, Houdret N, Fourrier F: Metformin-associated lactic acidosis: A prognostic and therapeutic study. Crit Care Med 37:2191–2196, 2009), with permission of Lippincott Williams & Wilkins, Inc. Data are medians (interquartile range; extreme values) and are compared with the Mann-WhitneyU test. LODS, logistic organ dysfunction system.


References1. Glucophage (metformin hydrochloride) [package insert], Princeton,

NJ, Bristol-Myers Squibb, Revised Jan 20092. Bodmer M, Meier C, Krahenbuhl S, Jick SS, Meier CR: Metformin,

sulfonylureas, or other antidiabetes drugs and the risk of lacticacidosis or hypoglycemia: A nested case-control analysis. DiabetesCare 31: 2086–2091, 2008

3. Salpeter SR, Greyber E, Pasternak GA, Salpeter EE: Risk of fatal andnonfatal lactic acidosis with metformin use in type 2 diabetes melli-tus. Cochrane Database Syst Rev 4: CD002967, 2010

4. Kamber N, Davis WA, Bruce DG, Davis TM: Metformin and lacticacidosis in an Australian community setting: The Fremantle DiabetesStudy. Med J Aust 188: 446–449, 2008

5. Lalau JD: Lactic acidosis induced by metformin: Incidence, manage-ment and prevention. Drug Saf 33: 727–740, 2010

6. Pongwecharak J, Tengmeesri N, Malanusorn N, Panthong M,Pawangkapin N: Prescribing metformin in type 2 diabetes with acontraindication: Prevalence and outcome. Pharm World Sci 31:481–486, 2009

7. Shaw JS, Wilmot RL, Kilpatrick ES: Establishing pragmatic esti-mated GFR thresholds to guide metformin prescribing. Diabet Med24: 1160–1163, 2007

8. Herrington WG, Levy JB: Metformin: Effective and safe in renaldisease? Int Urol Nephrol 40: 411–417, 2008

9. Philbrick AM, Ernst ME, McDanel DL, Ross MB, Moores KG:Metformin use in renal dysfunction: Is a serum creatinine thresholdappropriate? Am J Health Syst Pharm 66: 2017–2023, 2009

10. Tahrani AA, Varughese GI, Scarpello JH, Hanna FW: Metformin,heart failure, and lactic acidosis: Is metformin absolutely contraindi-cated? BMJ 335: 508–512, 2007

11. Perrone J, Phillips C, Gaieski D: Occult metformin toxicity in threepatients with profound lactic acidosis. J Emerg Med June 18, 2008[epub ahead of print]

12. Bruijstens LA, van Luin M, Buscher-Jungerhans PM, Bosch FH:

Reality of severe metformin-induced lactic acidosis in the absence ofchronic renal impairment. Neth J Med 66: 185–190, 2008

13. Spiller HA, Weber JA, Winter ML, Klein-Schwartz W, Hofman M,Gorman SE, Stork CM, Krenzelok EP: Multicenter case series ofpediatric metformin ingestion. Ann Pharmacother 34: 1385–1388,2000

14. Dell’Aglio DM, Perino LJ, Kazzi Z, Abramson J, Schwartz MD,Morgan BW: Acute metformin overdose: Examining serum pH,lactate level, and metformin concentrations in survivors versus non-survivors—A systematic review of the literature. Ann Emerg Med 54:818–823, 2009

15. Arroyo AM, Walroth TA, Mowry JB, Kao LW: The MALAdy ofmetformin poisoning: Is CVVH the cure? Am J Ther 17: 96–100,2010

16. Turkcuer I, Erdur B, Sari I, Yuksel A, Tura P, Yuksel S: Severemetformin intoxication treated with prolonged haemodialyses andplasma exchange. Eur J Emerg Med 16: 11–13, 2009

17. Guo PY, Storsley LJ, Finkle SN: Severe lactic acidosis treated withprolonged hemodialysis: Recovery after massive overdoses of met-formin. Semin Dial 19: 80–83, 2006

18. Seidowsky A, Nseir S, Houdret N, Fourrier F: Metformin-associatedlactic acidosis: A prognostic and therapeutic study. Crit Care Med37: 2191–2196, 2009

19. Pearlman BL, Fenves AZ, Emmett M: Metformin-associated lacticacidosis. Am J Med 101: 109–110, 1996

Anion Gap CompositionWhen a patient is found to have anion gap

metabolic acidosis, the physician must always seekto identify the “unmeasured” anions. Often, theprincipal cause is readily apparent, as with lacticacidosis, ketoacidosis, or uremic acidosis. However,in many patients, a large component of the “miss-ing” anions cannot be identified (see previous sec-tion). Thirty years ago, Gabow et al. (1,2) inten-sively searched for the “missing” anions using avariety of chemical analytic techniques includingenzymatic analysis for organic acids, gas chroma-tography–mass spectrometry (GC-MS) of blood andurine, and calculation of the contribution of serumproteins and inorganic phosphate. Despite thesemultifaceted identification efforts, a large compo-nent of “missing” anions remained elusive in manypatients. More recently, several groups continuedthe search for the identity of the “missing anions,”which presumably were added to the blood as acids.Moviat et al. (3) attempted to identify all of the acidanions in 31 critically ill patients with metabolicacidosis, using standard analytic methods, ion-exchange column chromatography, reverse-phaseHPLC, and GC-MS. Again, a large fraction of the“anion gap” remained unknown. McKinnon et al.(4) published a series of articles describing theirsearch for the missing anions in patients with met-abolic acidosis using new techniques such as liquid

0

20

40

60

80

100

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33

Cumulative duration of hemodialysis (h)

Dec

reas

e o

f m

etfo

rmin

leve

ls

(%)

Figure 1. The effect of hemodialysis on plasma metforminlevels in 20 patients. The x-axis shows the cumulativedialysis duration. All predialytic metformin levels wereconsidered as the 100% value and the decrease in metforminlevel was calculated as the percentage of this value. Meandecrease and SD of the mean are depicted. Clinical toxicol-ogy by American Academy of Clinical Toxicology. Copy-right 2009. Reprinted from reference 18 (Seidowsky A,Nseir S, Houdret N, Fourrier F: Metformin-associated lacticacidosis: A prognostic and therapeutic study. Crit Care Med37: 2191–2196, 2009), with permission of Lippincott Wil-liams & Wilkins, Inc.


chromatography coupled to electrospray ionizationmass spectrometry. They emphasized the contribu-tion of Kreb’s cycle acids, such as isocitric, alphaketo-butyric, malic, and other acids (5– 8). In total,these acids can explain 3 to 5 mEq/L of the aniongap in many patients with metabolic acidosis, but alarge component of the anion gap remains missing.Either relatively high concentrations of several an-ions are still unable to be identified, large numbersof anions in low concentration are not being recog-nized, or the conceptual framework about thesedisorders needs revision. For example, Kamel et al.(9) suggested that a component of anion gapchanges that occur in a variety of pathologic con-ditions may be generated by major valence changesof polyvalent anions, including proteins.

References1. Gabow PA, Kaehny WD, Fennessey PV, Goodman SI, Gross PA,

Schrier RW: Diagnostic importance of an increased serum anion gap.N Engl J Med 303: 854–858, 1980

2. Gabow PA: Disorders associated with an altered anion gap. Kidney Int27: 472–483, 1985

3. Moviat M, Terpstra AM, Ruitenbeek W, Kluijtmans LA, Pickkers P,van der Hoeven JG: Contribution of various metabolites to the “un-measured” anions in critically ill patients with metabolic acidosis. CritCare Med 36: 752–758, 2008

4. McKinnon W, Lord GA, Forni LG, Peron JM, Hilton PJ: A rapidLC-MS method for determination of plasma anion profiles of acidoticpatients. J Chromatogr B Analyt Technol Biomed Life Sci 833: 179–185, 2006

5. McKinnon W, Pentecost C, Lord GA, Forni LG, Peron JM, Hilton PJ:Elevation of anions in exercise-induced acidosis: A study by ion-exchange chromatography/mass spectrometry. Biomed Chromatogr22: 301–305, 2008

6. Hilton PJ, McKinnon W, Lord GA, Peron JM, Forni LG: Unexplainedacidosis of malnutrition: A study by ion-exchange chromatography/mass spectrometry. Biomed Chromatogr 20: 1386–1389, 2006

7. Forni LG, McKinnon W, Hilton PJ: Unmeasured anions in metabolicacidosis: Unravelling the mystery. Crit Care 10: 220, 2006

8. Forni LG, McKinnon W, Lord GA, Treacher DF, Peron JM, Hilton PJ:Circulating anions usually associated with the Krebs cycle in patientswith metabolic acidosis. Crit Care 9: R591–R595, 2005

9. Kamel KS, Cheema-Dhadli S, Halperin FA, Vasudevan S, HalperinML: Anion gap: May the anions restricted to the intravascular spaceundergo modification in their valence? Nephron 73: 382–389, 1996

Propofol-Related Infusion Syndrome and LacticAcidosis

Propofol is a short-acting, intravenously ad-ministered hypnotic agent, unrelated to barbiturates,that is used for induction and maintenance of sur-gical general anesthesia, outpatient procedures suchas colonoscopy, and sedation of mechanically ven-tilated patients in the intensive care unit. It hasmany attributes that have made it a very popular

drug. Its onset of action and the recovery from itseffects are very rapid. Patients do not feel “groggy”after its use and indeed may awake feeling refreshedand with a sense of euphoria. Propofol has neuro-protective effects, including inactivation of GABAreceptors and blockade of excitatory neurotransmit-ters. It reduces cerebral oxygen consumption andintracranial pressure. It also has antioxidant, anti-inflammatory, and bronchodilator properties. How-ever the drug also has a rare but often fatal risk,recently called the propofol related infusion syndrome(PRIS). This was first described in 1992 by Parke et al.(1) in five very critically ill children; all received highdoses of the drug. They developed metabolic acidosis(lactic), hyperlipidemia, bradyarrhythmias, and pro-gressive cardiac failure. For several years, it wasunclear whether their reaction to propofol marked atruly distinct syndrome because the patients who de-veloped it all were critically ill and there were multiplepotential explanations for these findings. However,PRIS is now generally accepted as a true complicationof propofol infusion that can occur in both childrenand adults (2–4). Additional features include rhabdo-myolysis, hyperlipidemia, kidney failure, and electro-cardiographic (ECG) changes including Bruggadasymptom–like abnormalities. (Bruggada syndrome isan inherited channelopathy that causes a spectrum ofECG abnormalities; the most common characteristic isextreme J-point elevation and a coved ST segment.)Animal studies have shown that propofol may inter-fere with mitochondrial metabolism, including oxygenutilization, oxidative phosphorylation, fatty acid up-take, and fat oxidation, but the specific defects thatcause PRIS in humans remain obscure; some patientsmay have a genetic predisposition. PRIS is more likelyto develop when the drug is used in critically illpatients at doses above 4 mg/kg per hour for longerthan 48 hours. When recognized the infusion must bediscontinued as soon as possible.

Roberts et al. (5) reported the results of a pro-spective multicenter study (11 academic medical cen-ter intensive care units) of monitoring critically illadults who were given propofol. In that report, PRISwas defined as metabolic acidosis plus cardiac dys-function and one or more of the following: rhabdomy-olysis, hypertriglyceridemia, or renal failure occurringafter initiation of the propofol infusion. Among 1017patients who received propofol, PRIS developed in1.1% after an average of 3 days of treatment; 91% of


patients who developed PRIS were receiving vaso-pressors. Most patients with PRIS in this study sur-vived; the mortality rate of 18% was much lower thanreported from previous studies. Children, who have ahigher mortality rate with PRIS, were excluded fromthe study. Fong et al. (6) evaluated presenting charac-teristics associated with mortality in PRIS. They re-viewed all published cases, as well as oral case reportssubmitted to the Food and Drug Administration’sMedwatch Adverse Event Reporting System Databasebetween 1989 and 2005; information was analyzedusing a multivariate logistic regression model. A totalof 1139 patients with suspected PRIS were identified,342 (30%) of whom died. Likelihood of death in-creased when the patient was male; was �18 years ofage; required vasopressor use; and/or developed met-abolic acidosis, cardiac or renal failure, hypotension,rhabdomyolysis, or dyslipidemia.

References1. Parke TJ, Stevens JE, Rice AS, Greenaway CL, Bray RJ, Smith PJ,

Waldmann CS, Verghese C: Metabolic acidosis and fatal myocardialfailure after propofol infusion in children: Five case reports. BMJ 305:613–616, 1992

2. Cremer OL: The propofol infusion syndrome: More puzzling evidenceon a complex and poorly characterized disorder. Crit Care 13: 1012,2009

3. Fodale V, La Monaca E: Propofol infusion syndrome: An overview ofa perplexing disease. Drug Saf 31: 293–303, 2008

4. Corbett SM, Montoya ID, Moore FA: Propofol-related infusion syn-drome in intensive care patients. Pharmacotherapy 28: 250–258, 2008

5. Roberts RJ, Barletta JF, Fong JJ, Schumaker G, Kuper PJ, Papado-poulos S, Yogaratnam D, Kendall E, Xamplas R, Gerlach AT, SzumitaPM, Anger KE, Arpino PA, Voils SA, Grgurich P, Ruthazer R, DevlinJW: Incidence of propofol-related infusion syndrome in critically illadults: A prospective, multicenter study. Crit Care 13: R169, 2009

6. Fong JJ, Sylvia L, Ruthazer R, Schumaker G, Kcomt M, Devlin JW:Predictors of mortality in patients with suspected propofol infusionsyndrome. Crit Care Med 36: 2281–2287, 2008

Toxic AlcoholsEthylene glycol, the major ingredient in radiator

antifreeze, is a widely recognized poison with a wellunderstood pathologic profile and treatment regimen.Recent articles have addressed the effects of two othertoxic glycols (glycols are double alcohols, or diols,which contain two hydroxyl groups) that have beenunderrecognized as causes of metabolic acidosis andreviewed the spectrums of clinical pathology that theygenerate. Diethylene glycol poisoning was first iden-tified �70 years ago. This clear, colorless, viscous,sweet-tasting solvent has been associated with severalmajor epidemics. The most famous US epidemic oc-curred in 1937 and is called the “Massengill Incident”

(1). The S.E. Massengill Company was one of theearly manufacturers of the newly discovered anti-infectious agent sulfanilamide. The company dis-solved sulfanilamide in diethylene glycol with rasp-berry flavoring to create a pleasant-tasting liquidmedication that was especially easy to administer tochildren. During a 2-month period, more than 100people who ingested this medication died. This epi-sode was the major event leading to the creation of theFederal Food Drug and Cosmetic Act of 1938, a majorregulatory, authority-enabling law for the Food andDrug Administration.

Despite documentation of this drug’s lethality,many diethylene glycol poisoning epidemics occurredsince the Massengill Incident and continue to occuraround the world. In most cases, diethylene glycol hasbeen substituted for glycerin, either accidentally orpurposefully by disreputable companies, in the prepa-ration of various medications, toothpastes, etc. Awidely publicized epidemic occurred in 2006, when aChinese pharmaceutical supply company sold diethyl-ene glycol as a “glycerin substitute” to a Panamanianpharmaceutical company that then used the compoundto prepare several different cough syrups sold in Pan-ama (2,3). Depending on the source of information,this event caused between 78 and 365 deaths; a muchlarger number of patients sustained severe and perma-nent neurologic damage. Of special interest to neph-rologists is that many of the victims had underlyingdiabetes and hypertension and were using the coughsyrup to treat a cough related to their use of angioten-sin-converting enzyme inhibitors. When an epidemicof severe neurologic dysfunction began, the initialinvestigation focused on the angiotensin-convertingenzyme inhibitors as the possible toxin.

The pathogenic sequence and clinical symptom-atic time course generated by diethylene glycol inges-tion is similar to what follows ingestion of other, morecommonly encountered toxic alcohols. Soon after in-gestion, a period of inebriation ensues, with or withoutgastrointestinal symptoms. Initially, the glycol gener-ates an osmolal gap, but because diethylene glycol is arelatively large molecule, the molar concentration andosmolal gap produced by any given level measured inmg/dl will be less impressive than similar concentra-tions generated by smaller alcohols, such as methanol.The parent compound is then oxidized, mainly by theenzyme alcohol dehydrogenase, followed by aldehydedehydrogenase, to a series of highly toxic metabolites


that include multiple organic acids; the metabolicpathway for diethylene glycol is shown in the Figure 2from Schep et al. (4).

After oxidation, anion gap metabolic acidosis;acute renal failure; and severe neurologic dysfunction,including peripheral neuropathies, cranial nerve pal-sies (facial diplegia occurs commonly, in up to 50% ofthe Panamanian epidemic [5]), and central nervoussystem dysfunction, ensue. Although there is littlepublished experience, antagonists of alcohol dehydro-genase—ethanol and fomepizole—should be protec-tive. Dialysis should also be effective, but, again,because the diagnoses have often been markedly de-layed, the published experience regarding interventionis scant. Schep et al. (4) recently wrote a comprehen-sive review of diethylene glycol poisoning.

The generation of 2-hydroxyethoxyacetic acid(HEAA) is the principal cause of the anion gap met-abolic acidosis. HEAA is also a very toxic moleculeand may be the major neurotoxin. In animal models,co-ingestion of diethylene glycol with fomepizole pre-

vented the development of both the metabolic acidosisand neurotoxicity (6).

The ingestion of diethylene glycol resultsin an anion gap metabolic acidosis (theglycol is oxidized to HEAA and other or-ganic acids); acute renal failure; and severeneurologic dysfunction, including peripheralneuropathies, cranial nerve palsies (facial di-plegia is particularly common), and centralnervous system dysfunction. Treatment withantagonists of alcohol dehydrogenase, suchas fomepizole, should be protective.

Propylene glycol is a pharmaceutical solventused to dissolve a variety of intravenous, oral, andtopical medications and can be toxic in large doses. Itis metabolized to several toxic products includinglactic acid (both L- and D-lactate can be generated; seeFigure 3). Most cases of propylene glycol poisoningoccur when high doses of lorazepam or phenobarbitalare used either to treat patients with alcohol with-drawal or to induce a therapeutic coma. Other medi-cations dissolved in this solvent are shown in Table 2(7). Figure 4 shows the relationship between loraz-epam infusion rates and the levels of propylene glycoland the osmolal gap (7). When the osmolal gap ex-ceeds 10 mosmol/L in patients who receive thesedrugs, toxic propylene glycol levels should be consid-ered (7,8). Propylene glycol is excreted by the kidney,so a reduced GFR will contribute to its accumulation.Case reports of this syndrome continue to appear andrecently have emphasized its occurrence in the setting

Figure 3. The metabolic pathway for propylene glycol.

O

CH2

CH2 CH2OH

CH2OH

O

CH2

CH2 C

CH2OH

O

H

O

CH2

CH2 C

CH2OH

O

OH

NAD+

NADH

Lactate

Pyruvate

NAD+

NADH

Lactate

Pyruvate

Alcoholdehydrogenase

Aldehydedehydrogenase

Diethylene glycol

2-Hydroxyethoxyacetaldehyde

2-Hydroxyethoxyacetic acid

Figure 2. Pathway of diethylene glycol metabolism by alcoholdehydrogenase oxidation to form 2-hydroxyethoxyacetalde-hyde followed by aldehyde dehydrogenase oxidation to form2-hydroxyethoxyacetic acid. Clinical toxicology by AmericanAcademy of Clinical Toxicology. Copyright 2009. Reprintedfrom reference 4 (Schep LJ, Slaughter RJ, Temple WA, Bea-sley DM: Diethylene glycol poisoning. Clin Toxicol (Phila)47: 525–535, 2009).


of therapeutic drug coma (9,10). Zosel et al. (11)recently described the case of an accidental therapeuticoverdose with the development of extreme toxicity.The patient was given 2 mg/min lorazepam instead of2 mg/h. After 10 hours of infusion, the propyleneglycol level was 659 mg/dl, the arterial pH was 6.9,serum bicarbonate level was 5 mEq/L, and lactatelevel was 18.6 mmol/L. Treatment with fomepizoleand continuous venovenous hemofiltration led to res-

olution of the acidosis, and his propylene glycol levelfell to 45 mg/dl. Unfortunately, the patient died, but hehad sustained severe anoxic brain damage before themedication accident and the fomepizole and continu-ous venovenous hemofiltration seemed to reverse themetabolic effects of the poisoning effectively. Zar etal. (12) published a review of propylene glycol poi-soning, and Kraut and Kurtz (13) published a compre-hensive review of the toxic alcohol poisoning syn-dromes and their treatment.

References1. Calvery HO, Klumpp TG: The toxicity for human beings of dieth-

ylene glycol with sulfanilamide. South Med J 32: 1105–1109, 19392. Bogdanich W, Hooker J: From China to Panama, a Trail of Poisoned

Medicine. New York Times May 6, 2007 Available at: http://www.nytimes.com/2007/05/06/world/americas/06poison.html?_r�1&ref�waltbogdanich. Accessed January 23, 2011

3. Rentz ED, Lewis L, Mujica OJ, Barr DB, Schier JG, Weerasekera G,Kuklenyik P, McGeehin M, Osterloh J, Wamsley J, Lum W, AlleyneC, Sosa N, Motta J, Rubin C: Outbreak of acute renal failure inPanama in 2006: A case-control study. Bull World Health Organ 86:749–756, 2008

4. Schep LJ, Slaughter RJ, Temple WA, Beasley DM: Diethylene glycolpoisoning. Clin Toxicol (Phila) 47: 525–535, 2009

5. Morelle J, Kanaan N, Hantson P: The case: Cranial nerve palsy andacute renal failure after a ‘special drink.’ Kidney Int 77: 559–560,2010

6. Besenhofer LM, Adegboyega PA, Bartels M, Filary MJ, Perala AW,McLaren MC, McMartin KE: Inhibition of metabolism of diethyleneglycol prevents target organ toxicity in rats. Toxicol Sci June 7, 2010[epub ahead of print]

7. Arroliga AC, Shehab N, McCarthy K, Gonzales JP: Relationship ofcontinuous infusion lorazepam to serum propylene glycol concentra-tion in critically ill adults. Crit Care Med 32: 1709–1714, 2004

8. Yahwak JA, Riker RR, Fraser GL, Subak-Sharpe S: Determination ofa lorazepam dose threshold for using the osmolal gap to monitor forpropylene glycol toxicity. Pharmacotherapy 28: 984–991, 2008

9. Miller MA, Forni A, Yogaratnam D: Propylene glycol-induced lacticacidosis in a patient receiving continuous infusion pentobarbital. AnnPharmacother 42: 1502–1506, 2008

10. Bledsoe KA, Kramer AH: Propylene glycol toxicity complicating useof barbiturate coma. Neurocrit Care 9: 122–124, 2008

11. Zosel A, Egelhoff E, Heard K: Severe lactic acidosis after an iatrogenicpropylene glycol overdose. Pharmacotherapy 30: 219, 2010

12. Zar T, Graeber C, Perazella MA: Recognition, treatment, and pre-vention of propylene glycol toxicity. Semin Dial 20: 217–219, 2007

13. Kraut JA, Kurtz I: Toxic alcohol ingestions: Clinical features, diag-nosis, and management. Clin J Am Soc Nephrol 3: 208–225, 2008

Asthma: Complex Interaction of RespiratoryAlkalosis, Metabolic Acidosis, and RespiratoryAcidosis

Acute asthma increases respiratory drive andcommonly generates respiratory alkalosis. Approxi-mately half of all adult patients hospitalized withsevere asthma have respiratory alkalosis (1). If thePCO2 begins to increase, then this may indicate that

Figure 4. Correlation between the lorazepam infusion rateand serum propylene glycol (PG) concentrations and osmo-lal gap. Adapted from reference 7 (Arroliga AC, Shehab N,McCarthy K, Gonzales JP: Relationship of continuous in-fusion lorazepam to serum propylene glycol concentrationin critically ill adults. Crit Care Med 32: 1709–1714, 2004),with permission of Lippincott Williams & Wilkins, Inc.

Table 2. Commonly used intravenous drugs thatcontain propylene glycol

Drug

Amount ofPropylene Glycol

(% vol/vol)

Lorazepam, 2 mg/ml 80Phenobarbital, 30–130 mg/ml 68–75Diazepam, 5 mg/ml 40Pentobarbital, 50 mg/ml 20–40Phenytoin, 50 mg/ml 40Trimethoprim-sulfamethoxazole

(16:80)40

Etomidate, 2 mg/ml 35Nitroglycerin, 5 mg/ml 30Esmolol, 250 mg/ml 25

Reprinted from reference 7 (Arroliga AC, Shehab N, McCarthy K, Gonzales JP:Relationship of continuous infusion lorazepam to serum propylene glycol con-centration in critically ill adults. Crit Care Med 32: 1709 –1714, 2004), withpermission of Lippincott Williams & Wilkins, Inc.


very severe obstruction exists and/or that respiratorymuscle fatigue is developing. Overt respiratory acido-sis in this setting is an ominous sign that generallyindicates that bronchodilator and anti-inflammatorytherapy should be intensified and mechanical ventila-tion considered.

Metabolic acidosis also develops in many pa-tients who are hospitalized for treatment. It is presentin 12 to 38% of hospitalized adults and in up to 28%of hospitalized children with asthma (2–4). The met-abolic acidosis in these patients is usually due to lacticacidosis (4). There has been increased interest inidentifying the cause and clinical significance of thismetabolic acidosis (3,4). Multiple factors may contrib-ute to the development of lactic acidosis in thesepatients, including extreme muscle exertion and fa-tigue, underperfusion of vascular beds, systemic hyp-oxia, and reduced hepatic lactate metabolism. Severalstudies have also proposed that, in some cases, lacticacidosis is caused by overly aggressive treatment with�2-adrenergic agonists and glucocorticoids (3–8). �2-Adrenergic agonists accelerate glycolysis and therebyincrease the generation of pyruvic acid. They alsoincrease lipolysis, and the resulting delivery of freefatty acids to hepatic mitochondria reduces mitochon-drial uptake and metabolism of the pyruvic acid. Theseeffects combine to increase lactic acid generation andaccumulation. Accelerated lipolysis also contributes toketogenesis, and overt ketoacidosis has been describedin some patients who were treated with �2-adrenergicagonists, especially pregnant women. Glucocorticoidswill exacerbate any tendency to hyperglycemia andalso contribute to the development of ketoacidosis insome patients.

It is extremely important to recognize the devel-opment of metabolic acidosis in these patients becausethey will compensate with additional hyperventilationand may develop symptomatic air hunger. Sometimes,these findings are mistakenly attributed to worseningreversible airway disease, and this misdiagnosis leadsto intensification of the patient’s �2-adrenergic ago-nist therapy, which is an inappropriate therapeuticresponse. In fact, the �2-adrenergic agonists may ac-tually be the cause of the metabolic acidosis, and inthose cases, the dose should be reduced. Furthermore,Bohn (5) believed that the higher rates of metabolicacidosis that are now recognized in patients withasthma may be the result of overly aggressive therapywith �2-adrenergic agonists and glucocorticoids.

Hyperchloremic metabolic acidosis also occursin some patients with asthma. Rashid et al. (9) re-ported that 32 of 109 adults hospitalized with acuteasthma exacerbation had an element of hyperchlor-emic metabolic acidosis. This is more common thananion gap metabolic acidosis, which was seen in only11 of 109 patients. Why does hyperchloremic acidosisdevelop in these patients? Most likely, this represents“post hypocapnic” metabolic acidosis. If these patientshad been hyperventilating for �1 day before theiradmission, as a result of worsening bronchospasm,then they would have compensated for the respiratoryalkalosis by reducing their blood HCO3 concentration,simultaneously increasing their blood chloride concen-tration. In this scenario, after admission and aggressivetreatment of their asthma, the respiratory alkalosiswould improve or completely resolve, and the patientswould be left with a residual “hyperchloremic acido-sis.” With time and adequate salt intake, metabolicacidosis would correct as their kidneys excretedNH4Cl. However, it should be noted that in the studyby Rashid et al. (9), there is another explanation forthe surprisingly high proportion of patients with hy-perchloremic acidosis compared with anion gap aci-dosis. These investigators used a relatively high upperlimit of 12 mEq/L for their anion gap normal range; inaddition, they did not report albumin levels. If some ofthe patients had had hypoalbuminemia (and this infor-mation were included in the anion gap calculation) andif a lower upper normal limit for the anion gap hadbeen used, then a number of patients with “normalanion gap” metabolic acidosis would instead havebeen classified as having “elevated anion gap” meta-bolic acidosis.

Regardless of whether the metabolic acidosis isanion gap (mainly lactic), hyperchloremic, or a com-bination of both and regardless of the pathophysio-logic explanation for the acidosis, another questionmust be addressed: Does treatment of metabolic aci-dosis with parenteral sodium bicarbonate have a ther-apeutic role in these patients with asthma? The bene-fits of sodium bicarbonate therapy for patients withasthma have been touted for more than 40 years (10).More recently, Buysse et al. (11) published a retro-spective, observational study of the use of sodiumbicarbonate in 17 children with life-threatening refrac-tory asthma. Virtually all of these patients also hadmarked respiratory acidosis; the PCO2 was �70 mmHgin most, and many had PCO2 levels �100 mmHg. A


number of patients who were given sodium bicarbon-ate demonstrated overall clinical improvement, andmost of them showed improvement in PCO2. However,there was no contemporaneous matched control group.The benefit of bicarbonate therapy for patients withsevere asthma and metabolic acidosis remains contro-versial (12).

References1. Mountain RD, Heffner JE, Brackett NC Jr, Sahn SA: Acid-base

disturbances in acute asthma. Chest 98: 651–655, 19902. Roncoroni AJ, Adrougue HJ, De Obrutsky CW, Marchisio ML,

Herrera MR: Metabolic acidosis in status asthmaticus. Respiration33: 85–94, 1976

3. Meert KL, Clark J, Sarnaik AP: Metabolic acidosis as an underlyingmechanism of respiratory distress in children with severe acuteasthma. Pediatr Crit Care Med 8: 519–523, 2007

4. Koul PB, Minarik M, Totapally BR: Lactic acidosis in children withacute exacerbation of severe asthma. Eur J Emerg Med 14: 56–58,2007

5. Bohn D: Metabolic acidosis in severe asthma: Is it the disease or isit the doctor? Pediatr Crit Care Med 8: 582–583, 2007

6. Ramnarayan P, Chhabra R, Maheshwari P: Metabolic acidosis, re-spiratory distress, and children with severe acute asthma. Pediatr CritCare Med 10: 142–143, 2009

7. Creagh-Brown BC, Ball J: An under-recognized complication oftreatment of acute severe asthma. Am J Emerg Med 26: 514.e1–514.e3, 2008

8. Veenith TV, Pearce A: A case of lactic acidosis complicating assess-ment and management of asthma. Int Arch Med 1: 3, 2008

9. Rashid AO, Azam HM, DeBari VA, Blamoun AI, Moammar MQ,Khan MA: Non-anion gap acidosis in asthma: Clinical and laboratoryfeatures and outcomes for hospitalized patients. Ann Clin Lab Sci 38:228–234, 2008

10. Mithoefer JC, Runser RH, Karetzky MS: The use of sodium bicar-bonate in the treatment of acute bronchial asthma. N Engl J Med272:1200–1203, 1965

11. Buysse CM, de Jongste JC, de Hoog M: Life-threatening asthma inchildren: Treatment with sodium bicarbonate reduces PCO2. Chest127: 866–870, 2005

12. Agarwal R, Gupta D: Sodium bicarbonate in life-threatening asthma:Not so soon! Chest 128: 1890–1891, 2005

Epidemiology of the Anion Gap in “NormalPopulations”

Farwell and Taylor (1–3) have published threestudies that analyzed several National Health and Nu-trition Examination Surveys (NHANES) to determinewhether there is a relationship between the anion gapand hypertension, markers of inflammation, and insu-lin resistance. NHANES is a cross-sectional databaseof the general population of the United States. Forthese three studies, individuals with chronic diseases,including hypertension, kidney disease, diabetes, andcancer, were excluded. The anion gap was found to beindependently associated with higher BP, and thegroup with the highest quintile of anion gap (16.2 �

0.6 mEq/L) had the lowest bicarbonate concentration(22.2 � 0.2 mEq/L), and their systolic BPs were 3.73mmHg higher than those in the lowest anion gapquintile (9.0 � 0.11 mEq/L; HCO3 25.2 � 0.2 mEq/L)(1). Similar confirmatory results were obtained from asmaller, different data base that consisted of 1057members of Harvard Vanguard Medical Associates. Inthat study, each 1 mEq/L increase of anion gap wasassociated with a 0.22-mmHg (P � 0.04) higher meanarterial pressure (4). In analogous studies, a higher aniongap and lower bicarbonate level were associated withhigher leukocyte counts and C-reactive protein levels.Patients in the in the highest anion gap quartile (15.1 �0.1 mEq/L) and with lowest bicarbonate levels (22.9 �0.2 mEq/L) had an average leukocyte count that was1.0 � 109/L higher and a C-reactive protein level thatwas 10.9 nmol/L higher (P � 0.01) than those in thelowest anion gap quartile (8.5 � 0.0 mEq/L; bicarbonate25.4 � 0.1 mEq/L) (2). The third article from this groupdetermined that higher anion gap and lower bicarbonatelevels are associated with lower insulin sensitivity (esti-mated by the MFFM index, which predicts glucosedisposal rates (M) corrected for fat-free mass (FFM)from fasting insulin and triglyceride levels in individualswith normoglycemia) (3). All of these parameters weremeasured in ambulatory “healthy” individuals, and themeasured parameters were within their respective normalranges. Because these are cross-sectional studies, it isimpossible to determine causations, leaving researchersto wonder whether subtle metabolic acidosis and aniongap increase (within the respective normal range) elevateBP, inflammatory markers, and insulin resistance or ifone of these factors, such as higher BP, leads to themetabolic acidosis and anion gap increase. These areprovocative results and generate multiple hypotheses.Does the low-grade accumulation of some organic acidor acids lead to these pathologies?

References1. Taylor EN, Forman JP, Farwell WR: Serum anion gap and blood

pressure in the national health and nutrition examination survey.Hypertension 50: 320–324, 2007

2. Farwell WR, Taylor EN: Serum anion gap, bicarbonate and biomark-ers of inflammation in healthy individuals in a national survey. CMAJ182: 137–141, 2010

3. Farwell WR, Taylor EN: Serum bicarbonate, anion gap and insulinresistance in the National Health and Nutrition Examination Survey.Diabet Med 25: 798–804, 2008

4. Forman JP, Rifas-Shiman SL, Taylor EN, Lane K, Gillman MW:Association between the serum anion gap and blood pressure amongpatients at Harvard Vanguard Medical Associates. J Hum Hypertens22: 122–125, 2008


Topiramate and Hyperchloremic MetabolicAcidosis

Topiramate is a drug that is approved to be usedas monotherapy or adjunctive therapy for partial-onsetseizures and primary generalized tonic-clonic seizures,for adjunctive treatment of seizures associated withLennox-Gastaut syndrome, and for prophylaxis of mi-graine headaches. However, it is also very widely usedfor many “off label” indications. This drug is a car-bonic anhydrase inhibitor and therefore produces mul-tiple renal acidification defects, including proximaltubule bicarbonate wasting and impaired distal acidi-fication (carbonic anhydrase inhibition and metabolicacidosis may play a role in the drug’s neurologictherapeutic efficacy as well). Metabolic acidosis re-duces urine citrate levels, and this combines with thealkaline urine pH to promote precipitation of calciumphosphate. Soon after acetazolamide was marketed, itwas discovered that the drug could lead to calciumphosphate stones and intrarenal calcium phosphateprecipitates (1). This adverse effect has been seen withvirtually all carbonic anhydrase inhibitors, and topira-mate is no exception. The risk for kidney stones isapproximately 2 to 4 times that expected in the back-ground population (2). Welch et al. (3) evaluated urinestone risk chemistry profiles of 32 patients who weretaking topiramate and also evaluated these parametersbefore and after treatment with topiramate in sevenpatients. Serum bicarbonate levels were lower in thoseon topiramate, and their urinary pH, absolute bicar-bonate excretion rate, and fractional excretion of bi-carbonate all were higher than in the control group.Urinary citrate excretion was reduced to 278 � 226compared with 737 � 329 mg/d in the control group,and the relative saturation ratio for brushite (calciumphosphate) doubled compared with the control group(3). Certain populations will have even higher risk; forexample, nonambulatory patients who take the drugare very likely to form stones because it develops inmore than half of these individuals (4). Also, thesimultaneous use of a ketogenic diet, which itself isalso lithogenic, with topiramate will generate an in-creased risk for developing urine stones (5). Otherrecently introduced antiseizure drugs, such as Zoni-samide, are also carbonic anhydrase inhibitors andwould be expected to have a similar effect on urinecalcification tendency (6). Vega et al. (7) reviewed theincreased propensity for calcium phosphate kidneystones caused by topiramate.

The hyperchloremic metabolic acidosis gener-ated by topiramate and all carbonic anhydrase inhibi-tors will cause compensatory hyperventilation. This isusually asymptomatic but sometimes produces a sen-sation of dyspnea. Some patients will present with thisas their chief complaint (8,9), and it is important toremember this possibility when patients who use thisdrug complain of dyspnea. Massive topiramate over-dose also generates severe hyperchloremic metabolicacidosis, which resolves over several days (10,11).

References1. Harrison HE, Harrison HC: Inhibition of urine citrate excretion and

the production of renal calcinosis in the rat by acetazoleamide(Diamox®) administration. J Clin Invest 34: 1662–1670, 1955

2. Lamb EJ, Stevens PE, Nashef L: Topiramate increases biochemicalrisk of nephrolithiasis. Ann Clin Biochem 41: 166–169, 2004

3. Welch BJ, Graybeal D, Moe OW, Maalouf NM, Sakhaee K: Bio-chemical and stone-risk profiles with topiramate treatment. Am JKidney Dis 48: 555–563, 2006

4. Goyal M, Grossberg RI, O’Riordan MA, Davis ID: Urolithiasis withtopiramate in nonambulatory children and young adults. PediatrNeurol 40: 289–294, 2009

5. Paul E, Conant KD, Dunne IE, Pfeifer HH, Lyczkowski DA, LinshawMA, Thiele EA: Urolithiasis on the ketogenic diet with concurrenttopiramate or zonisamide therapy. Epilepsy Res 90: 151–156, 2010

6. Wroe S: Zonisamide and renal calculi in patients with epilepsy: Howbig an issue? Curr Med Res Opin 23: 1765–1773, 2007

7. Vega D, Maalouf NM, Sakhaee K: Increased propensity for calciumphosphate kidney stones with topiramate use. Expert Opin Drug Saf6: 547–557, 2007

8. Delpirou-Nouh C, Gelisse P, Chanez P, Carlander B: Migraine andtopiramate induced dyspnea. Headache 47: 1453–1455, 2007

9. Shiber JR: Severe non-anion gap metabolic acidosis induced bytopiramate: A case report. J Emerg Med 38: 494–496, 2010

10. Lynch MJ, Pizon AF, Siam MG, Krasowski MD: Clinical effects andtoxicokinetic evaluation following massive topiramate ingestion.J Med Toxicol 6: 135–138, 2010

11. Wisniewski M, Łukasik-Głebocka M, Anand JS: Acute topiramateoverdose: Clinical manifestations. Clin Toxicol (Phila) 47: 317–320,2009

Acquired Forms of Metabolic Alkalosis

Metabolic Alkalosis: The Milk Alkali Syndrome(Calcium Alkali Syndrome)

In 1915, Sippy (1) introduced a new regimen,later named for him, to treat peptic ulcer disease thatincluded an hourly diet of milk and cream, as well aseggs and cereals, combined with the ingestion of largeamounts of three antacids: Sodium bicarbonate, mag-nesium carbonate, and bismuth subcarbonate (“SippyPowders”). The “Sippy Regimen” proved to be veryeffective, but within a decade, a spectrum of toxicmanifestations, which included metabolic alkalosis,were described (2).

In 1936, Cope (3) described what is now called the


acute milk alkali syndrome. The findings included hy-percalcemia, hyperphosphatemia, hypermagnesemia, in-creased bicarbonate levels, and kidney damage; abnor-malities usually resolved with discontinuation of theSippy Regimen. In 1949, Burnett et al. (4) described amore chronic and persistent form of the same diseaseprocess.

The introduction of histamine 2 receptor block-ers (1976) and proton pump inhibitors (1989) to blockacid secretion, as wall as treatments directed at erad-icating Helicobacter pylori eliminated the use of theSippy Regimen and its subsequent antacid modifica-tions. It was believed that the milk-alkali syndromehad become a disease of purely historical interest, buta modern variant of this disorder has returned with avengeance.

Patel and Goldfarb (5) recently reviewed thepathophysiology of the more recent disorder andproposed that we call the current version of milk-alkali syndrome “calcium-alkali syndrome” becauseit is now rarely associated with the ingestion oflarge amounts of milk and other dairy products.Instead, it is generated by the ingestion of largeamounts of calcium carbonate, sometimes togetherwith vitamin D and occasionally with thiazide di-uretics; these drugs are often used for osteoporosistherapy or prevention. The syndrome is now thethird leading cause of hypercalcemia in hospitalizedpatients (after hyperparathyroidism and malig-nancy) and is often responsible for severe levels ofhypercalcemia (6).

There are some important differences between thehistorical classic varieties of milk-alkali syndrome andthe modern form of calcium-alkali syndrome. First, hy-perphosphatemia was very common with the originaldisorders (because kidney dysfunction was combinedwith large phosphorus loads from dairy products) but ismuch less common now because of the lower oral phos-phorous load and the phosphorous-binding capacity ofcalcium carbonate. Second, although ingestion of cal-cium is excessive in all forms of the syndrome, calciumcarbonate does not provide nearly as much absorbablebase as sodium bicarbonate and perhaps less than themagnesium and bismuth salts used in the past. Conse-quently, the degree of alkalosis is now moderate unlessvomiting has developed. Third, vitamin D supplementa-tion is now common and contributes to the spectrum ofpathology, especially hypercalcemia.

Hypercalcemia, metabolic alkalosis, and renal

insufficiency have complex interactions that may syn-ergize and reinforce the severity of each disorder.Hypercalcemia causes renal vasoconstriction and in-terstitial damage, contributing to the renal dysfunction.Hypercalcemia also has complex systemic and renalacid-base effects, likely directly increasing renal pro-ton secretion and, to the extent that parathyroid hor-mone is suppressed, increasing renal bicarbonate re-absorption. Metabolic alkalosis also has complexeffects on calcium metabolism and increases renaltubule calcium reabsorption (7). Reduced renal func-tion will blunt the patient’s ability to excrete bothcalcium and bicarbonate.

In recent years, our understanding of this com-plex syndrome has been enhanced by integrating theroles of the calcium-sensing receptor (CaSR) and thecalcium selective channel called transport transientreceptor potential vanilloid member 5 (TRPV5). In thethick ascending loop of Henle, the CaSR is presentprimarily on the basolateral cell membrane. Whenthese receptors are activated by high blood calciumlevels, they reduce the open state of the renal outermedullary kidney (ROMK) channel on the apicalmembrane, which acts to blunt activity of the Na-K-Cl-Cl co-transporter (NKCC) on the apical membrane(8). CaSR activation may also directly reduce NKCCactivity. Therefore, hypercalcemia generates a loopdiuretic–like effect that increases sodium and calciumdelivery out of the thick limb and blunts the kidney’sconcentrating capacity. Further down the length of thetubule, in the distal convolution, activation of apicalCaSR by high urine calcium levels increases calciumreabsorption via TRVP5 channels (9). In the collectingduct, high urine calcium levels then activate CaSR onthe apical membrane, and in these tubule segments,CaSR activation reduces expression of aquaporin 2water channels. This reduces water reabsorption andcauses excretion of more dilute urine (10). In addition,CaSR activation at this site stimulates proton secretionvia H-ATPase (11). In summary, the net effect ofhypercalcemia on these various renal CaSRs leads to aloop diuretic–like effect in the thick ascending limb ofHenle, which acts to increase urine calcium and so-dium excretion. Some moderation of the hypercalci-uria occurs in the distal tubule via stimulation ofTRVP5-mediated calcium reabsorption and, furtherdown the tubule, dilution and acidification of theurine, should reduce the likelihood of precipitation ofcalcium salts, especially calcium phosphate. Finally,


all of these renal CaSR effects are enhanced by sys-temic metabolic alkalosis, which clearly increases thesensitivity of the CaSR and the activity of the TRVP5channel (12). Much of this new information is derivedfrom animal studies and from cell/tissue preparations.We still must determine its relevance to intact humanswith this disease process (13). The fascinating inter-action of these complex systems both under physio-logic conditions and in response to pathophysiologicdisorders such as the milk-alkali (or calcium-alkali)syndrome continue to be elucidated. Figure 5 showsthe current understanding of the interplay of blood andurine calcium, systemic alkalosis, and urine bicarbon-ate on renal tubule calcium transport.

Pseudo-Bartter (or Bartter-Like) SyndromeMany hereditary and acquired disorders bear some

similarity to Bartter syndrome. In some cases, the simi-

larities may be very close (e.g., continuous surreptitiousloop diuretic use) whereas others may be less so. Treat-ment with aminoglycosides sometimes generates a renaltubule disorder syndrome that mimics Bartter syndrome.These drugs have many toxic renal effects, but thisunusual complication results from the fact that the CaSRcan be stimulated by these antibiotics. The result of thisstimulation in the thick ascending limb of Henle isinhibition of the ROMK channel on the apical membraneand perhaps the NKCC co-transporter (see previous sec-tion - The Milk Alkali Syndrome [Calcium Alkali Syn-drome]). As discussed, this generates a furosemide-likeeffect causing NaCl diuresis, a less positive electricalcharge in the lumen and reduced calcium and magnesiumreabsorption via the paracellular pathway. The clinicalpicture is very similar to that seen when inherited Bartter-like syndrome is due to a gain-of-function mutation ofthe CaSR (14). Chen et al. (15) recently reported a caseof pseudo-Bartter syndrome caused by gentamicin andreviewed the literature. Other, similar cases have beenpreviously reported (16). This syndrome was also re-cently described in a patient receiving another aminogyl-coside antibiotic, amikacin (17).

Mixed Metabolic Alkalosis and RespiratoryAcidosis: Role for Acetazolamide?

When metabolic alkalosis develops in the inten-sive care unit, it is usually due to nasogastric suctionor diuretic therapy. With gastric drainage, the meta-bolic alkalosis is obviously generated by removal ofgastric HCl and is then maintained primarily by intra-vascular volume contraction with a contribution fromthe potassium depletion that almost always develops.Gastric alkalosis is usually readily corrected with in-travenous saline and potassium salts, which expandeffective intra-arterial volume and allow the kidney toexcrete the excess bicarbonate. In other patients, themetabolic alkalosis may be generated by the aggres-sive use of diuretics, mainly loop diuretics or combi-nations of loop and thiazide diuretics. In this disorder,the kidney is the site of bicarbonate generation, and,again, effective intra-arterial volume contraction is thereason the metabolic alkalosis is maintained. Volumeexpansion with intravenous isotonic saline is usuallycontraindicated (these patients usually have beengiven the diuretics because of clinical evidence ofvolume overload) or would be ineffective, as in pa-tients with hepatic cirrhosis, heart failure, etc. It is notuncommon for these forms of metabolic alkalosis to be

B

A

Figure 5. Mechanisms for renal calcium transport dependon the location of the CaSR and TRPV5 channel. (A)Thick ascending loop of Henle. (B) Distal convolutedtubule. NKCC, sodium potassium-2-chloride co-transporter; ROM-K, renal outer medullary potassiumchannel; NaKATPase, sodium-potassium ATPase; �stimulates; � inhibits; NCC, sodium chloride co-transporter; NCX, sodium-calcium exchanger. Reprintedfrom reference 5, with permission.


superimposed on chronic respiratory acidosis. Patientswith chronic respiratory acidosis should already havea high blood bicarbonate concentration as a result ofmetabolic compensation. This mixed disorder existswhen the bicarbonate level is “too high” for the coex-istent PCO2 level; this may occur when patients withchronic respiratory acidosis require gastric drainage ordiuresis. The use of steroids, which have some min-eralocorticoid effect, can contribute to this mixeddisorder. Bicarbonate loads from anticoagulated (withsodium citrate) blood products may be another alkalisource. Furthermore, when patients with chronic re-spiratory acidosis require artificial ventilation, theyoften develop a “post-hypercapnic” metabolic alkalo-sis. In these cases, PCO2 is reduced by the ventilator,but the serum bicarbonate level remains at its previ-ously appropriate high level, which is now too high.Should these various forms of metabolic alkalosis beaggressively corrected, and, if so, does such correctionhelp the patient’s overall clinical status? It has beentaught that the development of metabolic alkalosis inpatients with chronic respiratory acidosis raises theblood pH to levels that may further depress respiratorydrive and thereby promote higher PCO2 levels. Thismixed disorder may also make it more difficult towean these patients from mechanical ventilation.Banga and Khilnani (18) recently confirmed this im-pression in a retrospective study of 84 patients whohad chronic obstructive pulmonary disease and re-quired mechanical ventilation. Twenty percent ofthese patients developed posthypercapnic metabolicalkalosis. Glucocorticoid use for �10 days was anindependent risk factor for this complication. Thesepatients had increased ventilator dependence (64.7versus 37.3%) and a longer intensive care unit stay(14.7 � 6.7 versus 9.5 � 5.9 days). Consequently,common wisdom is that metabolic alkalosis in thisgroup of critically ill patients should be aggressivelycorrected. This can most readily be accomplished withvolume expansion when that maneuver is indicatedand kidney function is adequate. It is also possible tocorrect the alkalosis by administering strong acid oracid precursors, such as intravenous HCl or enteralNH4Cl. If additional diuresis is required, then the useof acetazolamide will produce a brisk bicarbonatediuresis (if kidney function is adequate). A number ofstudies have confirmed many of these clinical impres-sions (19–23). Those reports suggest that such therapyimproves gas exchange, and PCO2 levels usually fall.

However, Faisy et al. (24) could not confirm thisimpression. They treated 26 mechanically ventilatedpatients who had chronic obstructive pulmonary dis-ease and this mixed acid-base disorder (serum bicar-bonate �26 mmol/L and arterial pH �7.38) withacetazolamide 500 mg intravenously. Unfortunately,they did not have a contemporaneous control groupand instead compared their results with a historicalcontrol group matched for serum bicarbonate, arterialpH, age, and severity of illness. The bicarbonate con-centration fell as expected after acetazolamide treat-ment. They could not demonstrate any reduction in thelength of the weaning period or a statistical improve-ment in extubation success compared with the histor-ical control. The lack of a contemporaneous controlgroup and the small group size weaken these con-clusions. Furthermore, the metabolic alkalosis com-ponent of this mixed disorder was mild (bloodbicarbonate averaged 34 mEq/L), and this was re-duced only to 31.5 mEq/L by the time of extubation.It still seems very reasonable to treat the metaboliccomponent of this mixed disorder, especially whenthe metabolic alkalosis is severe. If diuresis isrequired, then acetazolamide is appropriate, al-though it must be emphasized that this diureticcauses brisk potassium excretion, as well as a so-dium bicarbonate diuresis. Usually, aggressive po-tassium replacement is necessary.

References1. Sippy BW: Gastric and duodenal ulcer: Medical cure by an efficient

removal of gastric juice corrosion. JAMA 64: 1625–1630, 19152. Hardt LL, Rivers AB: Toxic manifestations following the alkaline

treatment of peptic ulcer. Arch Intern Med 31: 171–180, 19233. Cope CL: Base changes in the alkalosis produced by the treatment of

gastric ulcer with alkalies. Clin Sci 2: 287–300, 19364. Burnett CH, Commons RR, Albright F, Howare JE: Hypercalcemia

without hypercalciuria or hyperphosphatemia, calcinosis and renalinsufficiency: Syndrome following the prolonged intake of milk andalkali. N Engl J Med 240: 787–794, 1949

5. Patel AM, Goldfarb S: Got calcium? Welcome to the calcium-alkalisyndrome. J Am Soc Nephrol 21: 1440–1443, 2010

6. Picolos MK, Lavis VR, Orlander PR: Milk-alkali syndrome is amajor cause of hypercalcaemia among non-end-stage renal disease(non-ESRD) inpatients. Clin Endocrinol (Oxf) 63: 566–576, 2005

7. Peraino RA, Suki WN: Urine HCO3 augments renal Ca2� absorp-tion independent of systemic acid-base changes. Am J Physiol 238:F394–F398, 1980

8. Riccardi D, Brown EM: Physiology and pathophysiology of thecalcium-sensing receptor in the kidney. Am J Physiol Renal Physiol298: F485–F499, 2010

9. Topala CN, Schoeber JP, Searchfield LE, Riccardi D, Hoenderop JG,Bindels RJ: Activation of the Ca2-sensing receptor stimulates theactivity of the epithelial Ca2 channel TRPV5. Cell Calcium 45:331–339, 2009

10. Bustamante M, Hasler U, Leroy V, de Seigneux S, Dimitrov M,


Mordasini D, Rousselot M, Martin PY, Feraille E: Calcium-sensingreceptor attenuates AVP-induced aquaporin-2 expression via a calm-odulin-dependent mechanism. J Am Soc Nephrol 19: 109–116, 2008

11. Renkema KY, Velic A, Dijkman HB, Verkaart S, Kemp AW, NowikM, Timmermans K, Doucet A, Wagner CA, Bindels RJ, HoenderopJG: The calcium-sensing receptor promotes urinary acidification toprevent nephrolithiasis. J Am Soc Nephrol 20: 1705–1713, 2009

12. Quinn SJ, Bai M, Brown EM: pH sensing by the calcium-sensingreceptor. J Biol Chem 279: 37241–37249, 2004

13. Bergsland KJ, Coe FL, Gillen DL, Worcester EM: A test of thehypothesis that the collecting duct calcium-sensing receptor limitsrise of urine calcium molarity in hypercalciuric calcium kidney stoneformers. Am J Physiol Renal Physiol 297: F1017–F1023, 2009

14. Huang C, Miller RT: Regulation of renal ion transport by thecalcium-sensing receptor: An update. Curr Opin Nephrol Hypertens16: 437–443, 2007

15. Chen YS, Fang HC, Chou KJ, Lee PT, Hsu CY, Huang WC, ChungHM, Chen CL: Gentamicin-induced Bartter-like syndrome. Am JKidney Dis 54: 1158–1161, 2009

16. Chou CL, Chen YH, Chau T, Lin SH: Acquired Bartter-like syn-drome associated with gentamicin administration. Am J Med Sci 329:144–149, 2005

17. Chrispal A, Boorugu H, Prabhakar AT, Moses V: Amikacin-inducedtype 5 Bartter-like syndrome with severe hypocalcemia. J PostgradMed 55: 208–210, 2009

18. Banga A, Khilnani GC: Post-hypercapnic alkalosis is associated withventilator dependence and increased ICU stay. COPD 6: 437–440, 2009

19. Bear R, Goldstein M, Phillipson E, Ho M, Hammeke M, Feldman R,Handelsman S, Halperin M: Effect of metabolic alkalosis on respi-ratory function in patients with chronic obstructive lung disease. CanMed Assoc J 117: 900–903, 1977

20. Miller PD, Berns AS: Acute metabolic alkalosis perpetuating hyper-carbia: A role for acetazolamide in chronic obstructive pulmonarydisease. JAMA 238: 2400–2401, 1977

21. Holland AE, Wilson JW, Kotsimbos TC, Naughton MT: Metabolicalkalosis contributes to acute hypercapnic respiratory failure in adultcystic fibrosis. Chest 124: 490–493, 2003

22. Brimioulle S, Berre J, Dufaye P, Vincent JL, Degaute JP, Kahn RJ:Hydrochloric acid infusion for treatment of metabolic alkalosis as-sociated with respiratory acidosis. Crit Care Med 17: 232–236, 1989

23. Dickinson GE, Myers ML, Goldbach M, Sibbald W: Acetazolamidein the treatment of ventilatory failure complicating acute metabolicalkalosis. Anesth Analg 60: 608–610, 1981

24. Faisy C, Mokline A, Sanchez O, Tadie JM, Fagon JY: Effectivenessof acetazolamide for reversal of metabolic alkalosis in weaningCOPD patients from mechanical ventilation. Intensive Care Med 36:859–863, 2010

Metabolic Alkalosis (Pseudo-Bartter Syndrome)in Cystic Fibrosis

Cystic fibrosis (CF) is an autosomal recessive,often fatal disease. In Caucasian populations, it occursat a rate of approximately one in 3000 live births andis less common in Hispanic, African-American, andAsian populations. The genetic defect results in abnor-mal synthesis, transport, and/or function of the cysticfibrosis transmembrane regulator (CFTR). This is aregulated chloride channel that also modulates otherepithelial ion channels. Although CF is usually diag-nosed in childhood, milder forms are increasinglyrecognized in adult patients who often have nonclas-

sical presentations. It has been known for many de-cades that children with CF can become severelyvolume depleted during hot summer months as a resultof excessive loss of NaCl in sweat. This early obser-vation was a major clue leading to the development ofthe sweat chloride test for diagnosis of this disease.Children with known CF may develop what has beencalled “pseudo-Bartter syndrome” from environmentalheat exposure. Yalcon et al. (1) reported that 12% of241 children with CF developed this disorder. Duringone particularly hot 2-week period in August, ninesuch patients were admitted to a pediatric unit inAnkara, Turkey (2). Occasionally, pseudo-Bartter syn-drome is the presenting symptom complex in youngchildren.

Adults with known CF can also develop severehyponatremia, hypokalemia, and metabolic alkalosiswhen exposed to a hot environment and again thesefindings may also be their initial presenting symptomcomplex (3). Smith et al. (4) described this presentationin 1995 but emphasized only the hyponatremia, hypoka-lemia, and volume contraction. Although it is likely thatthis patient also had marked metabolic alkalosis, acid-base data were not included in that brief report. Bates etal. (5) described a 17-year-old boy whose present-ing findings were metabolic alkalosis, hypokalemia,azotemia, and modest hyponatremia. Several other, sim-ilar adult cases have been reported (6,7), and Priou-Guesdon et al. (8) recently reported this presentation inthree adults.

It seems obvious why these patients would bepredisposed to salt and volume depletion. They losehigh concentrations of NaCl in their sweat, so large-volume sweating on a hot day can generate hypovo-lemia and, when extreme, cardiovascular collapse. Thesweat also contains relatively high K� concentrationsto account for hypokalemia. If the patients drink water,it will be retained as a result of reduced renal functionand high ADH levels and generate hyponatremia.They will also develop secondary hyperaldosteronism.All of these factors will reduce renal excretion ofbicarbonate and thereby contribute to maintenance ofthe metabolic alkalosis. But why do these patientsdevelop metabolic alkalosis? Where is the generationsite of the bicarbonate? Although some invoked sec-ondary hyperaldosteronism as the generation mecha-nism, this cannot be the explanation unless some otherfactor increases the distal tubule delivery of NaCl inthe face of overt extracellular fluid volume contrac-


tion. Bates et al. (5) suggested that the CFTR mutationmay impair both proximal and distal renal tubule NaClreabsorption in some patients. If that occurs, then arenal origin for the alkalosis becomes plausible. How-ever, a gastric origin for bicarbonate generation ismuch more likely because almost all of the casesreported with this syndrome have vomiting as a prom-inent feature. To the extent that occurs, it will worsen thevolume contraction, generate bicarbonate, and increaserenal potassium excretion. Once generated, the metabolicalkalosis would be maintained as a result of the factorsdescribed already. Although this syndrome is sometimescalled pseudo-Bartter syndrome, the urine chloride con-centration should readily distinguish between true Barttersyndrome (high) and this disorder (low).

When adults present with otherwise unex-plained metabolic alkalosis and signs andfindings consistent with volume contrac-tion, consider the possibility of a mild formof CF. This presentation is much more com-mon during heat waves. Hyponatremia alsocommonly occurs.

References1. Yalcin E, Kiper N, Dogru D, Ozcelik U, Aslan AT: Clinical features

and treatment approaches in cystic fibrosis with pseudo-Bartter syn-drome. Ann Trop Paediatr 25: 119–124, 2005

2. Kose M, Pekcan S, Ozcelik U, Cobanoglu N, Yalcin E, Dogru D,Kiper N: An epidemic of pseudo-Bartter syndrome in cystic fibrosispatients. Eur J Pediatr 167: 115–116, 2008

3. Ballestero Y, Hernandez MI, Rojo P, Manzanares J, Nebreda V,Carbajosa H, Infante E, Baro M: Hyponatremic dehydration as apresentation of cystic fibrosis. Pediatr Emerg Care 22: 725–727, 2006

4. Smith HR, Dhatt GS, Melia WM, Dickinson JC: Lesson of the week:Cystic fibrosis presenting as hyponatraemic heat exhaustion. BMJ 310:579–580, 1995

5. Bates CM, Baum M, Quigley R: Cystic fibrosis presenting withhypokalemia and metabolic alkalosis in a previously healthy adoles-cent. J Am Soc Nephrol 8: 352–355, 1997

6. Augusto JF, Sayegh J, Malinge MC, Illouz F, Subra JF, Ducluzeau PH:Severe episodes of extra cellular dehydration: An atypical adult pre-sentation of cystic fibrosis. Clin Nephrol 69: 302–305, 2008

7. Dave S, Honney S, Raymond J, Flume PA: An unusual presentation ofcystic fibrosis in an adult. Am J Kidney Dis 45: e41–e44, 2005

8. Priou-Guesdon M, Malinge MC, Augusto JF, Rodien P, Subra JF,Bonneau D, Rohmer V: Hypochloremia and hyponatremia as theinitial presentation of cystic fibrosis in three adults. Ann Endocrinol(Paris) 71: 46–50, 2010

Pendred Syndrome: Pendrin (SLC26A4) DefectsThe term “pendrin” is derived from a disease

described in 1896 by Vaughan Pendred. Pendred ob-

served a relationship among congenital bilateral deaf-ness, goiter, and hypothyroidism, which was laterdubbed Pendred syndrome (1). This was subsequentlyfound to be caused by an autosomal recessive disorderlinked to mutations on chromosome 7. Next, it wasdiscovered that the affected gene directed the synthesisof the SLC26A4 transporter, later named pendrin. Thisneutral anion exchanger is present in the cochlea, thethyroid (where it functions mainly as a chloride/iodideexchanger), and the cortical collecting tubule of thekidney (mainly in intercalated cells) (2).

Intercalated cells are a family of proton-secretingcells located in the late distal convolved tubule, theconnecting tubule, and the collecting duct. All inter-calated cells use carbonic anhydrase II (CA-II) tocatalyze the generation of bicarbonate and protonsfrom CO2 and H2O, and they all express V-typeH�-ATPase pumps to secrete the protons. The vecto-rial directions in which the proton is pumped and thebicarbonate exits determine whether the intercalatedcell is acid secreting (a type A intercalated cell), basesecreting (a type B intercalated cell), or acid-baseneutral (a non–type A/non–type B intercalated cell).See figure 6. The V-type H�-ATPases may be locatedon either the luminal membrane (in type A cells and innon–type A/non–type B cells) or the basolateral mem-brane (type B cells). The bicarbonate exit step occursvia one of several anion exchangers. The two bestcharacterized are anion exchanger 1 (AE1), a memberof the SLC4 (solute carrier family 4) transporter fam-ily, and “pendrin,” which is the common name forSLC26A4 (the A4 member of solute carrier family 26)(3). Although these two anion exchangers seeminglyserve the same function in opposite membranes of theintercalated cell (they each exchange one bicarbonateion for one chloride ion), they are distinctively differ-ent proteins. AE1 is present in type A intercalatedcells, whereas pendrin is present in both type B inter-calated cells and non–type A/non–type B intercalatedcells. ATP-energized proton secretion is what drivesthe bicarbonate/chloride exchangers in each of thesecells. Note that the secretion of both the proton andbicarbonate ion into the lumen by the non–typeA/non–type B intercalated cells is acid-base neutral.However, a chloride ion is reabsorbed from the lumen.The relative density and activity of each of theseintercalated cells are probably modulated by the pa-tient’s acid-base status and avidity of renal chloridereabsorption.


Figure 6. Three types of intercalated cells are shown. They are expressed from the late distal convoluted tubule to the initialthird of the inner medullary collecting duct. In all intercalated cells HCO3 and H� are produced from CO2 and H2O catalyzedby the enzyme carbonic anhydrase II. Type A intercalated cells secrete H� across the apical (luminal) membrane mainly viaV-type H�-ATPase. The apical membrane also expresses H�/K�-ATPase, which may function mainly to reclaim potassiumwhen potassium deficiency occurs. The basolateral (interstitial) membrane transporters include the anion exchanger AE-1,which is a HCO3/Cl exchanger and the KCC4 KCl co-transporters that may have an important role in the maintenance of lowintracellular chloride concentrations. Type B intercalated cells express the pendrin HCO3/Cl exchanger on the apicalmembrane and V-type H�-ATPase on the basolateral membrane. The net effect of ion transport by these two processes resultsin HCO3 secretion and Cl reabsorption (i.e., HCl addition to the interstitium). Non–type A/non–type B intercalated cells alsohave the pendrin HCO3/Cl exchanger on the apical membrane. However, in these cells, the V-type H�-ATPase is also insertedin the apical membrane instead of the basolateral membrane as in the type B intercalated cells. Consequently, both H� andHCO3 are secreted into the lumen. No net acid-base secretion or absorption occurs, but chloride is reabsorbed from the lumen.The chloride exits the cell into the interstitium through chloride channels consisting of ClC-kb and Barttin subunits. Type-Aintercalated cells are found from the late distal convoluted tubule to the initial portion of the inner medullary collecting duct.In contrast, the type B intercalated cells and the non–type-A/non–type-B intercalated cells are mainly expressed in the distalconvoluted tubule and the connecting tubule. Adapted from reference 3 (with kind permission from SpringerScience�Business Media Wagner, CA, Devuyst O, Bourgeois S, Mohebbi N: Regulated acid-base transport in the collectingduct. Pflugers Arch 458: 137–156, 2009).


In mice, the abundance of this protein increaseswith alkalosis and decreases with acidosis (4). In vitroexperiments show that the collecting ducts of mice withpendrin gene knockout cannot secrete bicarbonate (5).

Despite this bicarbonate secretory defect, patients withPendred syndrome generally do not manifest any kidney oracid-base disorder. However, we would expect these pa-tients to be more susceptible to the development of severemetabolic alkalosis. Indeed, Pela et al. (6) described a childwho had Pendred syndrome and developed severe meta-bolic alkalosis when treated with thiazide diuretics. We arejust beginning to understand how the pendrin exchangernormally contributes to salt, volume, and BP regulation.Aldosterone probably increases pendrin activity, which in-creases to the chloride reabsorption that must accompanysodium reabsorption. Pendrin activity is required for up-regulation of epithelial sodium channel (7).

Pendrin is a chloride/bicarbonate ex-changer on the luminal membrane of typeB and non–type A/non–type B intercalatedcells. In type B intercalated cells, it playsan important role in net bicarbonate secre-tion. In non–type A/non–type B interca-lated cells, both protons and bicarbonateare secreted into the lumen. Hence, thereis no net acid or base secretion. However,chloride is reabsorbed.

References1. Pendred V: Deaf mutism and goitre. Lancet II: 532, 18962. Royaux IE, Wall SM, Karniski LP, Everett LA, Suzuki K, Knepper

MA, Green ED: Pendrin, encoded by the Pendred syndrome gene,resides in the apical region of renal intercalated cells and mediatesbicarbonate secretion. Proc Natl Acad Sci U S A 98: 4221–4226, 2001

3. Wagner CA, Devuyst O, Bourgeois S, Mohebbi N: Regulated acid-base transport in the collecting duct. Pflugers Arch 458: 137–156,2009

4. Wagner CA, Finberg KE, Stehberger PA, Lifton RP, Giebisch GH,Aronson PS, Geibel JP: Regulation of the expression of the Cl�/anionexchanger pendrin in mouse kidney by acid–base status. Kidney Int62: 2109–2117, 2002

5. Amlal H, Petrovic S, Xu J, Wang Z, Sun X, Barone S, Soleimani M:Deletion of the anion exchanger Slc26a4 (pendrin) decreases apicalCl(�)/HCO3(�) exchanger activity and impairs bicarbonate secretionin kidney collecting duct. Am J Physiol Cell Physiol 299: C33–C41,2010

6. Pela I, Bigozzi M, Bianchi B: Profound hypokalemia and hypochlor-emic metabolic alkalosis during thiazide therapy in a child withPendred syndrome. Clin Nephrol 69: 450–453, 2008

7. Wall SM, Pech V: The interaction of pendrin and the epithelial sodiumchannel in blood pressure regulation. Curr Opin Nephrol Hypertens17: 18–24, 2008

General Principles

HypokalemiaHypokalemia, generally defined as plasma potas-

sium (K�) �3.5 mEq/L, is usually caused by the lossof K� from the body. Less commonly, it is the resultof a redistribution of K� from the extracellular fluidspace to the intracellular space. It may have majorcardiovascular, neuromuscular, renal, and/or meta-bolic consequences.

The physiology of K� transport and its implica-tions in hypokalemic disorders has been well reviewedin a previous issue of NephSAP devoted to fluid,electrolyte, and acid-base disturbances (1), so thereader is referred to that review for general informa-tion. Regarding new discoveries in K� physiology thathave occurred in the past few years, they are discussedin the context of specific clinical syndromes whenrelevant; more in-depth information is available inseveral excellent recent reviews (2–5).

References1. Sterns RH, Palmer BF (eds): Fluid, electrolyte, and acid-base distur-

bances. NephSAP 6: 210–227, 20072. Sorensen MV, Matos JE, Praetorius HA, Leipziger J: Colonic potas-

sium handling. Pflugers Arch 459: 645–656, 20103. Welling PA, Ho K: A comprehensive guide to the ROMK potassium

channel: Form and function in health and disease. Am J Physiol RenalPhysiol 297: F849–F863, 2009

4. Greenlee M, Wingo CS, McDonough AA, Youn JH, Kone BC:Narrative review: Evolving concepts in potassium homeostasis andhypokalemia [published erratum appears in Ann Intern Med 151:143–144, 2009]. Ann Intern Med 150: 619–625, 2009

5. Wang WH, Giebisch G: Regulation of potassium (K) handling in therenal collecting duct. Pflugers Arch 458: 157–168, 2009

Inherited Forms of HypokalemiaA number of transport mechanisms are respon-

sible for maintaining the distribution of potassium(K�) between the intracellular fluid and extracellularfluid (ECF) spaces and for regulating K� excretion viathe kidney and gastrointestinal tract. Of greatest im-portance is energy utilizing Na�-K�-ATPase, a trans-port protein that moves K� into cells and Na� out ofcells. Also of critical importance are several familiesof K� channels. K� channels are membrane-spanningproteins that create water-filled permeation poresthrough which K� can be selectively transferredacross the cell membrane; K� always flows throughthese channels down its electrochemical gradient.Many different K� channels have been identified andcharacterized; they have various gating mechanismsthat switch the channels into open or closed confor-


mations and different rates of K� transit. The openversus closed probability of various K� channels is acritically important determinant of a cell’s restingtransmembrane voltage. Sometimes, even when anidentical electrochemical gradient for K� is artificiallycreated, the ion can flow more readily in one directionversus the other. Channels with this property are called“rectified.” The rectification can be either inward oroutward, depending on whether the flow of the ion, inthis case K�, is more rapid going into or out of thecell. One of the best characterized rectified K� chan-nels is the renal outer medullary kidney channel(ROMK channel). The physiology and pathophysiol-ogy of renal K� channels were reviewed by the Yalegroup (1,2).

Familial Hypokalemic Periodic ParalysisHypokalemic periodic paralysis (HypoKPP) is

an autosomal dominant disorder that is transmittedwith incomplete penetrance. Patients with this diseasedevelop episodes of flaccid paralysis associated withmarked hypokalemia. The prevalence of HypoKPP isthought to be approximately one in 100,000. Althoughinherited as an autosomal dominant genetic disease, itoccurs more frequently and is more severe in men, inwhom the symptoms often develop during the first 2decades of life. When symptoms of the disease de-velop in women, they are generally milder and usuallybegin later in life. The paralytic attacks often occurduring the night or in the early morning hours but canalso be precipitated by strenuous exercise (in this case,developing after the actual period of exercise); theingestion of carbohydrate-rich meals; cold exposure;or the administration of glucose, insulin, or glucocor-ticoids. Each of these stimuli or factors may generatean intracellular K� shift. Paraparesis or tetraparesismay develop. Although respiratory and cardiac mus-cles are generally spared, profound hypokalemia mayresult in respiratory muscle weakness and cardiacarrhythmias. A discrete attack may last hours to days.In addition, a chronic myopathic form of HypoKPPdevelops in approximately one quarter of affectedindividuals and can result in progressive fixed muscleweakness.

The initial search for the genetic molecularmechanism responsible for this disease obviously fo-cused on K� channels and other K� transport muta-tions. Therefore, it was surprising when familial ge-netic analyses instead identified the most common

mutation to be in a gene coding for a muscle calciumchannel. This is found in 50 to 70% of patients withHypoKPP. The second most common mutation, occur-ring in approximately 10% of patients with HypoKPP,is in a gene coding for a muscle sodium channel. Theaffected calcium channel gene, CACNA1S (more spe-cifically, the affected gene codes for the �1 subunitof the dihydropyridine-sensitive voltage-gated Ca2�

channel), and affected sodium channel gene, SCN4A(more specifically the affected gene codes for the �subunit of the tetrodotoxin-sensitive voltage-gatedNa� channel), direct the synthesis of critical compo-nents of their respective sarcolemmal ion channels.These two channels (Cav1.1 and Nav1.4 is the nomen-clature used to identify the channel proteins) have agreat deal of molecular homology and may haveevolved from a common precursor gene. Each of thesechannels is “voltage gated,” meaning that they openand close in response to changes in the intracellularelectrical charge generated by cellular polarization ordepolarization events. How does a change in a cell’sinternal electrical charge affect the permeation char-acteristics of a channel, and how does the mutationalter the channel properties? The answers to thesequestions were explored in a recent article by Mat-thews et al. (3); Figures 7 and 8 shows the proposed

Figure 7. Diagrammatic representation of the alpha subunitof CACNA1S with S4 segments and arginine residuesknown to be mutated in hypokalemic periodic paralysishighlighted. The novel arginine substitution reported here isboxed in gray. � indicates the number of positively chargedresidues in each S4 segment. R, arginine. Reproduced withpermission of Taylor & Francis Group, LLC from reference3 (Matthews E, Labrum R, Sweeney MG, Sud R, HaworthA, Chinnery PF, Meola G, Schorge S, Kullmann DM, DavisMB, Hanna MG: Voltage sensor charge loss accounts formost cases of hypokalemic periodic paralysis. Neurology72: 1544–1547, 2009); permission conveyed through Copy-right Clearance Center, Inc.


structure of the two implicated ion channels. The �1subunits of each of these channels have four repeateddiscrete domains, and each of the domains is com-posed of six �-helical transmembrane segments. Thefourth segment (S4) of each domain is comprised ofmany charged amino acids with every third one being anarginine or a lysine. These multiple arginine or lysinemoieties will either release or accept a proton in responseto changes in cell charge and/or pH. These ionic alter-ations cause the S4 segments to physically changeshape and move in response to the cell voltagechanges. These movements and shape changes mod-ify the permeability characteristics of the channel.Therefore, these S4 segments are now consideredthe “voltage sensors” of these channels.

It has been determined that virtually all of themutations responsible for HypoKPP result in the re-placement of one of the arginine molecules in the S4segment by a less charged, or neutral, amino acid.Consequently, these mutations alter the way thesemuscle cell Ca� (or Na�) voltage gated channelsrespond to muscle cell depolarization–repolarizationcycles. This new information affords a better under-standing of how the gene defects discovered in pa-tients with this disorder result in the characteristicparalytic episodes. Although we now have a much

better understanding of why the muscles become par-alyzed, it remains unknown why the paralytic attacksare associated with profound hypokalemia.

Acetazolamide is a carbonic anhydrase inhibitorthat generates metabolic acidosis and hypokalemia asa result of NaHCO3 diuresis and secondary hyperal-dosteronism. This drug was first empirically used totreat patients with hyperkalemic periodic paralysis,and it was found to be effective in these patients (4).Because it generates hypokalemia, its potential effi-cacy for hyperkalemic conditions seems apparent.Similarly, its use in patients with HypoKPP seems tobe counterintuitive. Nonetheless, it was tried in thesepatients and again found to be extremely effective;acetazolamide is now recognized as one of the mosteffective prophylactic drugs for patients with Hypo-KPP. Not only does it reduce the frequency andseverity of acute paralytic attacks, but it also improvesthe muscle strength of these patients between attacks(5,6). However, not all patients with HypoKPP re-spond. How does acetazolamide work in this disease?Do the recent genetic/molecular discoveries help ex-plain this drug’s efficacy in some patients but notothers? It is likely that acetazolamide’s efficacy isprimarily related to the chronic metabolic acidosisgenerated by the drug. Acidosis causes movement ofprotons into muscle cells, and this seems to amelioratethe disease process; inhibition of carbonic anhydrasewithin the muscle cells may also play a role (7). Whatis clear is that patients with calcium channel mutationsrespond to acetazolamide therapy much more reliablythan do those with a mutated sodium channel (7).Indeed, some patients with the sodium channel muta-tions become worse when treated with acetazolamide.

It has been suggested that a high-K�, low-Na�,and low-carbohydrate diet may alleviate the symptomsof HypoKPP and reduce the frequency of attacks. Ifattacks occur at a specific time, such as during sleep orearly morning, then bedtime prophylactic doses of oralKCl may prove helpful. The acute paralytic attacks aretreated with oral or parenteral potassium salts. Thiscertainly increases the plasma K� concentration butoften may improve the muscle weakness only to asmall degree. The recommended oral K� dosage foracute attacks is approximately 0.2 to 0.4 mmol/kgevery 15 to 30 minutes over several hours. If thepatient cannot be treated with oral potassium, thenKCl can be given intravenously. It is recommendedthat 20 to 40 mEq of KCl be diluted in 1 L of 5%

Figure 8. Diagrammatic representation of the alpha subunitof SCN4A with S4 segments and arginine residues known tobe mutated highlighted. Novel arginine mutations reportedhere are boxed in gray. �, Number of positively chargedresidues in each S4 segment; *substitutions of R1448 causea phenotype of paramyotonia congenita. R, arginine. Repro-duced with permission of Taylor & Francis Group, LLCfrom reference 3 (Matthews E, Labrum R, Sweeney MG,Sud R, Haworth A, Chinnery PF, Meola G, Schorge S,Kullmann DM, Davis MB, Hanna MG: Voltage sensorcharge loss accounts for most cases of hypokalemic periodicparalysis. Neurology 72: 1544–1547, 2009); permissionconveyed through Copyright Clearance Center, Inc.


mannitol but not in 5% dextrose (8). Infusion ofdextrose can shift additional K� into cells and pre-sumably worsen the paralytic attack. In addition, it iscritical to remember that the hypokalemia in Hypo-KPP is entirely generated by a shift of K� from theECF into cells. Therefore, significant risk is associatedwith the administration of large amounts of exogenouspotassium because there is no true total body potas-sium deficit. The administration of KCl must bestopped when the serum potassium concentration ap-proaches normal, even if muscle weakness persists.Cases of fatal hyperkalemia after treatment of hy-pokalemic periodic paralysis continue to be reportedboth with this disease and with the related disorder ofthyrotoxic hypokalemic periodic paralysis (9).

The most common cause of familial hy-pokalemic periodic paralysis is a mutation inthe �1-subunit of the dihydropyridine-sensi-tive voltage-gated Ca2�channel Cav1.1. Themutations reduce the number of posi-tively charged (arginine or lysine) aminoacids which are critically important forresponse to voltage changes – they arepart of the channel’s “voltage sensor.”

Thyrotoxic Hypokalemic Periodic ParalysisThyrotoxic hypokalemic periodic paralysis is

much more common than the inherited forms of Hypo-KPP. By definition, this form of hypokalemic periodicflaccid paralysis always occurs in patients with hyper-thyroidism and is eliminated (cured) when the thyroiddisease is effectively treated. It is a very commondisorder in Asian men who develop hyperthyroidism(approximately 10% of Asian men with hyperthyroid-ism are thus affected) and is also relatively common inhyperthyroid Latin American men. It occurs in ap-proximately 0.1% of American Caucasian men withhyperthyroidism. A genetic component has beenstrongly suspected in view of the differential incidencein these various ethnic groups. The emergency treat-ment of the acute paralytic attacks is identical to thatdescribed for HypoKPP, plus �-adrenergic blockade isused to antagonize the sympathetic drive generated byhyperthyroidism. Of course, definitive therapy re-quires reestablishment of a euthyroid state. In view ofthe strong clinical similarity with HypoKPP, once

investigators recognized the molecular etiology of thatdisease, they evaluated the calcium channel geneCACNA1S and sodium channel gene SCN4A of pa-tients with thyrotoxic hypokalemic periodic paralysisfor mutations (10); however, none were identified.

Recently, Ryan et al. (11) reported mutations ina gene that codes for a newly described muscle K�

channel. This muscle cell, K� channel is anotherinwardly rectifying channel (Kir 2.6) that, very inter-estingly, has a thyroid hormone response element in itspromoter region. Therefore, thyroid hormone wouldnormally be expected to “turn on” this gene. Ryan etal. discovered that this gene was mutated in approxi-mately one third of Caucasian patients with thyrotoxichypokalemic periodic paralysis. A variety of differentmissense, frameshift, and stop codon mutations of thisgene were identified in various patients. When thisgene was probed in Asian patients with thyrotoxichypokalemic periodic paralysis, mutations were iden-tified in seven of 27 patients from Singapore but inonly one of 83 patients from Hong Kong and none of31 patients from Thailand. Clearly, a variety of muta-tions and genes must be involved in this disorder, andtheir frequency must vary in different ethnic popula-tions. Venance et al. (12) reviewed the diagnosis,pathogenesis, and treatment of all of the periodicparalysis syndromes.

References1. Wang WH, Giebisch G: Regulation of potassium (K) handling in the

renal collecting duct. Pflugers Arch 458: 157–168, 20092. Hebert SC, Desir G, Giebisch G, Wang W: Molecular diversity and

regulation of renal potassium channels. Physiol Rev 85: 319–371,2005

3. Matthews E, Labrum R, Sweeney MG, Sud R, Haworth A, ChinneryPF, Meola G, Schorge S, Kullmann DM, Davis MB, Hanna MG:Voltage sensor charge loss accounts for most cases of hypokalemicperiodic paralysis. Neurology 72: 1544–1547, 2009

4. McArdle B: Adynamia episodica hereditaria and its treatment. Brain85: 121–148, 1962

5. Resnick JS, Engel WK, Griggs RC, Stam AC: Acetazolamide pro-phylaxis in hypokalemic periodic paralysis. N Engl J Med 278:582–586, 1968

6. Griggs RC, Engel WK, Resnick JS: Acetazolamide treatment ofhypokalemic periodic paralysis: Prevention of attacks and improve-ment of persistent weakness. Ann Intern Med 73: 39–48, 1970

7. Matthews E, Hanna MG: Muscle channelopathies: Does the predictedchannel gating pore offer new treatment insights for hypokalaemicperiodic paralysis? Physiology 588: 1879–1886, 2010

8. Griggs RC, Resnick J, Engel WK: Intravenous treatment of hy-pokalemic periodic paralysis. Arch Neurol 40: 539–540, 1983

9. Ahmed I, Chilimuri SS: Fatal dysrhythmia following potassiumreplacement for hypokalemic periodic paralysis. West J Emerg Med11: 57–59, 2010

10. Ng WY, Lui KF, Thai AC, Cheah JS: Absence of ion channels


CACN1AS and SCN4A mutations in thyrotoxic hypokalemic peri-odic paralysis. Thyroid 14: 187–190, 2004

11. Ryan DP, da Silva MR, Soong TW, Fontaine B, Donaldson MR,Kung AW, Jongjaroenprasert W, Liang MC, Khoo DH, Cheah JS, HoSC, Bernstein HS, Maciel RM, Brown RH Jr, Ptacek LJ: Mutationsin potassium channel Kir2.6 cause susceptibility to thyrotoxic hy-pokalemic periodic paralysis. Cell 140: 88–98, 2010

12. Venance SL, Cannon SC, Fialho D, Fontaine B, Hanna MG, PtacekLJ, Tristani-Firouzi M, Tawil R, Griggs RC, CINCH investigators:The primary periodic paralyses: Diagnosis, pathogenesis and treat-ment. Brain 129: 8–17, 2006

Acquired Forms of Hypokalemia

Acetaminophen PoisoningSeveral recent reports have described a very

high frequency of hypokalemia among patients whoare admitted for treatment of acute acetaminophenpoisoning (1–5). The fall in serum K� is clearlyrelated to the severity of the acetaminophen poison-ing and to the presenting acetaminophen level (1,5).Hypokalemia develops in up to 80% of those withsevere poisoning. Of note, the K� concentration isoften normal or only slightly reduced at the time ofadmission and then falls during the first day ofhospitalization. This sequence was highlighted by arecent report of two teenage girls who attemptedsuicide with acetaminophen poisoning. Their K�

concentrations were 3.9 and 3.2 mEq/L at admissionand fell to 2.3 and 2.6 mEq/L, respectively, withinthe first 30 hours of hospitalization (3). Althoughmultiple factors may contribute, including vomiting,the administration of intravenous dextrose and hy-perventilation, none of them can completely accountfor the development of hypokalemia. Pakravan et al.(5) reported that the change in serum K� at 4 hoursand the severity of hypokalemia at 24 hours bothwere correlated with the acetaminophen serum levelat 4 hours after admission. In addition, they foundthat the renal fractional excretion of K� was increased at4 and 12 hours to approximately 16% despite the sharpfall in serum K� and that the renal K� excretion fellsharply by 24 hours. The transtubular potassium gradient(TTKG) was also increased at 4 and 12 hours andcorrelated with initial acetaminophen level; it fell sharplyat 24 hours after admission. Increased renal fractional K�

excretion, with a similar time course, has been reportedin a rat model of acetaminophen toxicity (6).

Increased renal fractional excretion of K� hasalso been reported in patients on the day of admissionfor ibuprofen poisoning (7). In that study, the admis-sion ibuprofen level also correlated with increased

fractional K� excretion (7). It remains unclear whetherinappropriate renal potassium loss after these drugs isthe result of renal tubule toxicity or is hormonallydriven. In many aspects, these findings of inappro-priate kaliuresis parallel those of inappropriatephosphaturia and hypophosphatemia after acetamino-phen poisoning (8). The study by Pakravan et al. alsoreported that acetaminophen levels 4 hours after ad-mission correlated with a fall in the renal tubularmaximum reabsorption of phosphate, or TmPO4/GFR.

Of note, phosphaturia and hypophosphatemiahave been shown to be associated with a better prog-nosis after acetaminophen poisoning (9). Perhaps thesickest patients present to hospital with lactic acidosisand/or an acute fall in GFR. These complicationscould raise serum phosphate levels and reduce phos-phate excretion. Impressive, self-resolving hypophos-phatemia as a result of renal phosphate excretion hasalso been well described in the immediate postopera-tive period after patients undergo partial hepatic resec-tion (10,11). Although it would be attractive to invokeincreased levels of a phosphatonin or parathyroid hor-mone as the cause, recent studies were not consistentwith that hypothesis (11). The possible mechanismresponsible for phosphaturia after acetaminophen poi-soning has not been defined.

Hypokalemia occurs commonly after se-vere acetaminophen poisoning and is inlarge part the result of inappropriate kali-uresis. This may be the result of acetamin-ophen-related tubule toxicity.

References1. Waring WS, Stephen AF, Malkowska AM, Robinson OD: Acute

acetaminophen overdose is associated with dose-dependent hypoka-laemia: A prospective study of 331 patients. Basic Clin PharmacolToxicol 102: 325–328, 2008

2. Zyoud S, Awang R, Syed Sulaiman SA, Al-Jabi S: High prevalenceof hypokalemia after acute acetaminophen overdose: Impact of psy-chiatric illness. Hum Exp Toxicol 29: 773–778, 2010

3. Godber IM, Jarvis SJ, Maguire D: Hypokalaemia following parac-etamol overdose in two teenage girls. Ann Clin Biochem 44: 403–405, 2007

4. Pakravan N, Bateman DN, Goddard J: Effect of acute paracetamoloverdose on changes in serum and urine electrolytes. Br J ClinPharmacol 64: 824–832, 2007

5. Pakravan N, Goddard J, Bateman DN: Hypokalaemia followingparacetamol overdose. Ann Clin Biochem 45: 111, 2008

6. Trumper L, Girardi G, Elias MM: Acetaminophen nephrotoxicity inmale Wistar rats. Arch Toxicol 66: 107–111, 1992

7. Goddard J, Stachan FE, Bateman DN: Urinary sodium and potassium


excretion as measures of ibuprofen nephrotoxicity. J Clin Toxicol 41:747, 2003

8. Jones AF, Harvey JM, Vale JA: Hypophosphataemia and phospha-turia in paracetamol poisoning. Lancet 2: 608–609, 1989

9. Schmidt LE, Dalhoff K: Serum phosphate is an early predictor ofoutcome in severe acetaminophen-induced hepatotoxicity. Hepatol-ogy 36: 659–665, 2002

10. Lee HW, Suh KS, Kim J, Shin WY, Cho EH, Yi NJ, Lee KU:Hypophosphatemia after live donor right hepatectomy. Surgery 144:448–453, 2008

11. Nafidi O, Lapointe RW, Lepage R, Kumar R, D’Amour P: Mecha-nisms of renal phosphate loss in liver resection-associated hypophos-phatemia. Ann Surg 249: 824–827, 2009

High-Dosage Penicillin, Hypokalemia, andMetabolic Alkalosis

The use of massive doses of penicillin in the1960s led to the recognition that sodium penicillindelivers a large sodium load and that the penicillinmolecule can act as a “poorly absorbed” anion in thedistal renal tubule. Under certain conditions, this cangenerate inappropriate kaliuresis and renal acid excre-tion as well as an osmotic diuresis. Brunner and Frick(1) reported a series of such patients who developedhypernatremia, hypokalemia, and metabolic alkalosis(the most extreme case developed Na of 165, K of 2.0,and HCO3 of 37). These patients were receiving 100“mega” units (100 million units) of penicillin eachday, which delivered 170 mEq of sodium. Subse-quently, similar electrolyte abnormalities were re-ported in patients who were given high-dosage car-benicillin (2,3). Again, the sodium and penicillin-likeanion loads were enormous. For example, 30 to 60 g/dcarbenicillin provides 4.7 mEq Na/g � 141 to 282mEq/d Na and carbenicillin. Lippner et al. (4) pub-lished in vivo and in vitro animal studies (rats and toadbladders) that were entirely consistent with the “poorlyabsorbed anion” mechanism. In subsequent years, hy-pokalemia and metabolic alkalosis have been reportedin patients who were given a number of other penicil-lin derivatives, including ticarcillin, oxacillin, dicloxa-cillin, flucloxacillin, and, most recently, meropenem(5–11).

Although the same mechanisms are sometimesinvoked, it is unlikely that this explanation is correct.Meropenem, for example, delivers approximately 4mEq Na/g of antibiotic. Therefore, even a large mero-penem dose of 3 g/d will provide only 12 mEq ofNa-Meropenem, probably far too little to generatehypokalemia via the mechanism described above.Other mechanisms must contribute when hypokalemiaand/or metabolic alkalosis develops in patients who

are treated with the lower doses generally used withcurrent penicillin derivatives. This point is furtheremphasized in the case reported by Hoorn and Zietse(7). Their patient received 12 g/d ticarcillin (approxi-mately 26 mEq/d sodium ticarcillin). Although shedeveloped inappropriate kaliuresis, there was no evi-dence of volume contraction; her urine chloride con-centration was high; her renin and aldosterone levelswere low; and she did not improve with triamtereneadministration, which should have blocked the “clas-sic” mechanism described in the past. Therefore, thehypokalemia and metabolic alkalosis that develop withcurrently used penicillin-class antibiotics may becaused by some form of direct tubule damage, whichleads to obligatory renal potassium wasting. Zietse etal. (11) recently published a very nice review of thefluid, electrolyte, and acid-base disorders associatedwith antibiotics.

References1. Brunner FP, Frick PG: Hypokalaemia, metabolic alkalosis, and hy-

pernatraemia due to massive sodium penicillin therapy. Br Med J 4:550–552, 1968

2. Hoffbrand BI, Stewart JD: Carbenicillin and hypokalaemia. Br Med J4: 746, 1970

3. Klastersky J, Vanderklen B, Daneau D, Mathiew M: Carbenicillinand hypokalemia. Ann Intern Med 78: 774–775, 1973

4. Lipner HI, Ruzany F, Dasgupta M, Lief PD, Bank N: The behaviorof carbenicillin as a nonreabsorbable anion. J Lab Clin Med 86:183–194, 1975

5. Nanji AA, Lindsay J: Ticarcillin associated hypokalemia. ClinBiochem 15: 118–119, 1982

6. Schlaeffer F: Oxacillin-associated hypokalemia. Drug Intell ClinPharm 22: 695–696, 1988

7. Hoorn EJ, Zietse R: Severe hypokalaemia caused by flucloxacillin. JAntimicrob Chemother 61: 1396–1398, 2008

8. Johnson DW, Kay TD, Hawley CM: Severe hypokalaemia secondaryto dicloxacillin. Intern Med J 32: 357–358, 2002

9. Bhagwat A, Goel N, Sharma R, Jain S, Dua K: Meropenem: Aunusual cause of metabolic alkalosis in critical care patients. AnaesthIntensive Care 36: 745–746, 2008

10. Bharti R, Gombar S, Khanna AK: Meropenem in critical care:Uncovering the truths behind weaning failure. J Anaesth Clin Phar-macol 26: 99–101, 2010

11. Zietse R, Zoutendijk R, Hoorn EJ: Fluid, electrolyte and acid-basedisorders associated with antibiotic therapy. Nat Rev Nephrol 5:193–202, 2009

Hypokalemia from Intestinal Pseudo-obstructionOgilvie Syndrome

It is generally accepted that patients with chronickidney failure develop adaptive mechanisms thatmarkedly increase stool K� secretion. However, theoriginal reports by Hayes et al. (1,2) that describedstool K� excretion in the �50-mEq/d range in multi-ple patients have never been confirmed. Although


stool K� excretion does increase in comparison withnormal individuals, the increment is generally verymodest; stool K� excretion with advanced kidneydisease is usually in the 10- to 15-mEq/d range (3,4).Agarwal et al. (5) reviewed the literature and con-cluded that colon adaptation in chronic kidney failurewas not a major contributor to total K� balance.Another common assumption is that watery diarrheagenerates major K� loss as a result of high stool K�

concentration. The K� concentration in the waterphase of normal stool is very high—on the order of 85to 95 mEq/L. However, when diarrhea develops, thestool K� concentration falls markedly and with volu-minous watery diarrhea often barely exceeds theplasma K� concentration (6). Consequently, the majorcation lost in patients with secretory diarrhea is Na�.Against this background, the discovery of a colonicsecretory condition in which K� was the overwhelm-ing predominant cation was a great surprise. In the lastissue of NephSAP (7), the case of a 78-year-oldwoman who developed colonic pseudo-obstruction(Ogilvie syndrome) after surgical repair of a hip frac-ture was discussed (6). The colonic fluid K� concen-tration ranged between 130 and 170 mEq/L, whereasthe Na� concentration was between 4 and 15 mEq/L.The electrochemical data (colon lumen PD was �13mV) was entirely consistent with active K� secretion.

Active K� secretion had not previously beendescribed as a mechanism for secretory diarrhea. Morerecently, extremely high colon K� concentration inpatients with colonic pseudo-obstruction was con-firmed in five consecutive patients with this disorder(8). It is likely that high-conductance K� channels(BK channels) play an important role in this disorderand other, related colon pathologies. The physiologyand pathophysiology of these channels was recentlyreviewed by Sandle et al. (9). Of great interest andpotential clinical importance is the discovery that theseBK channels were overexpressed in the colon of apatient who developed pseudo-obstruction and secre-tory diarrhea after an episode of hemorrhagic shock(10). Her stool Na� concentration was only 11 mEq/L,and her stool K� concentration was 143 mEq/L. Over-expressed BK channels may also account for colon K�

losses in some patients with ulcerative colitis (11). Theopen state of these K channels is increased by stretch,so this may play an important role in disorders withextreme colon dilation. Sorensen et al. (12) recentlyreviewed the topic of colonic potassium secretion.

References1. Hayes CP Jr, Robinson RR: Fecal potassium excretion in patients on

chronic intermittent hemodialysis. Trans Am Soc Artif Int Organs 11:242–246, 1965

2. Hayes CP Jr, McLeod ME, Robinson RR: An extravenal mechanismfor the maintenance of potassium balance in severe chronic renalfailure. Trans Assoc Am Physicians 80: 207–216, 1967

3. Schrier RW, Regal EM: Influence of aldosterone on sodium, waterand potassium metabolism in chronic renal disease. Kidney Int 1:156–168, 1972

4. Kopple JD, Coburn JW: Metabolic studies of low protein diets inuremia. I. Nitrogen and potassium. Medicine (Baltimore) 52: 583–595, 1973

5. Agarwal R, Afzalpurkar R, Fordtran JS: Pathophysiology of potas-sium absorption and secretion by the human intestine. Gastroenter-ology 107: 548–571, 1994

6. van Dinter TG Jr, Fuerst FC, Richardson CT, Ana CA, Polter DE,Fordtran JS, Binder HJ: Stimulated active potassium secretion in apatient with colonic pseudo-obstruction: A new mechanism of secre-tory diarrhea. Gastroenterology 129: 1268–1273, 2005

7. Sterns RH, Palmer BF (eds): Fluid, electrolyte, and acid-base distur-bances. NephSAP 6: 219–220, 2007

8. Blondon H, Bechade D, Desrame J, Algayres JP: Secretory diarrhoeawith high faecal potassium concentrations: A new mechanism ofdiarrhoea associated with colonic pseudo-obstruction? Report of fivepatients. Gastroenterol Clin Biol 32: 401–404, 2008

9. Sandle GI, Hunter M: Apical potassium (BK) channels and enhancedpotassium secretion in human colon. QJM 103: 85–89, 2010

10. Simon M, Duong JP, Mallet V, Jian R, MacLennan KA, Sandle GI,Marteau P: Over-expression of colonic K� channels associated withsevere potassium secretory diarrhoea after haemorrhagic shock.Nephrol Dial Transplant 23: 3350–3352, 2008

11. Sandle GI, Perry MD, Mathialahan T, Linley JE, Robinson P, HunterM, MacLennan KA: Altered cryptal expression of luminal potassium(BK) channels in ulcerative colitis. J Pathol 212: 66–73, 2007

12. Sorensen MV, Matos JE, Praetorius HA, Leipziger J: Colonic potas-sium handling. Pflugers Arch 459: 645–656, 2010

Thiazides, Hypertension, and HypokalemiaThe seventh report of the Joint National Com-

mittee on Prevention, Detection, Evaluation, andTreatment of High Blood Pressure (JNC 7) states that“thiazide-type diuretics should be used in drug treat-ment for most patients with uncomplicated hyperten-sion, either alone or combined with drugs from otherclasses” (1). This recommendation was in part basedon the Systolic Hypertension in the Elderly Program(SHEP) and the Antihypertensive and Lipid Loweringtreatment to prevent Heart Attack Trial (ALLHAT)studies, which demonstrated the efficacy of thesedrugs (2,3).

However, thiazide diuretics are associated withthe development of hypokalemia, hyperuricemia, anddiabetes. Although reducing BP certainly has a majorimpact on cardiovascular risk, the benefit of the BPreduction may be blunted in some patients by thesemetabolic complications of therapy. Shafi et al. (4)reviewed and discussed these issues and the relation-ship between developing hypokalemia and diabetes in


two previous trials. The risk of incident diabetes was50% higher compared with placebo in SHEP andbetween 39 and 48% higher than with amlodipine orlisinopril, respectively, in ALLHAT. The diabetes riskincreases early (generally in the first year) and duringthis period is closely linked to the development ofhypokalemia.

Agarwal (5) discussed these results in an editorialentitled “Hypertension, Hypokalemia, and Thiazide-In-duced Diabetes: A 3-Way Connection.” Figure 9 fromthis editorial shows how these three factors may inter-act. Although it is not yet proved that a combination ofpotassium (K�) loss and hypokalemia is the initiatingderangement that leads to the development of diabetesand the worsening of BP control, there is much indi-rect evidence that this is the case. High K� intake mayreduce BP as a result of salt excretion and beneficialneurohumoral effects, but not all studies are consistent(6,7). Diuretic-induced hypokalemia is also associatedwith reduced � cell response to glucose. Agarwal (5)hypothesized that K� depletion might have an adverseimpact on glucose metabolism by reducing the perfu-sion of skeletal muscle. If diuretic-related K� lossesdo contribute to the development of diabetes, thentreatment with K� supplements, K�-sparing diuretics,angiotensin-converting enzyme inhibitors, or angio-tensin receptor blockers may not only prevent hypo-kalemia but also decrease the risk for developingdiabetes and also improve BP control.Hypokalemia as a Risk Factor in Patients withChronic Kidney Disease. The U-shaped relation-

ship between mortality or morbidity and many physi-ologic and biochemical parameters is not surprising ornew. Humans evolved to maintain these parameterswithin a certain range, and extremes on either sideusually cause physiologic malfunction, morbidity, anda greater risk for death. This U-shaped relationshipwas recently confirmed for serum K� concentration inpatients with chronic kidney disease. In a prospective,observational study of adult patients with stages 3through 5 chronic kidney disease, conducted at fourUS outpatient nephrology clinics, 834 patients werefollowed for an average of 2.6 years each (8). Theiraverage age was 60.5 years, average GFR was 25.4ml/min per 1.73 m2, and average K� concentrationwas 4.6 mmol/L. Mortality was significantly greaterwhen the serum K� was �4 mmol/l, and serum K�

�5.5 mmol/L increased the risk for reaching thecompound end point of “death or cardiovascularevent.” Although the risk for death as an isolated endpoint did not reach significance for the group withhyperkalemia, this may be explained by the fact therewere very few patients in this group; only 65 patientshad hyperkalemia at enrollment. In addition, the use ofdialysis as “rescue” therapy for severe hyperkalemiaprobably blunted its mortality impact.

Chan et al. (9) evaluated the safety of digoxintherapy in patients with ESRD. In an observationalcohort analysis, digoxin use among 120,864 incidenthemodialysis patients increased the risk for death by28%. This risk was greatest in patients with lowerpredialysis serum K� concentrations (�4.3 mEq/L),in whom the hazard ratio for death was 2.53 comparedwith patients who had higher predialysis serum K�

concentrations. The hazard ratio for death among pa-tients with predialysis K� concentration �4.6 mEq/Lwas not statistically different from 1.0.Hypokalemia-Induced Hyponatremia and Correc-tion of Hyponatremia with K� Replacement. Berland Rastegar (10) recently discussed the case of awoman who was taking thiazide diuretics and non-steroidal anti-inflammatory drugs long term and thenbecame ill with a cough and sinusitis. At presentationto the hospital, she had severe hyponatremia (96 mEq/L), hypokalemia (1.6 mEq/L), and metabolic alkalosis(38 mEq/L). After what her physicians believed wasvery conservative treatment of her hyponatremia, hersodium (Na�) concentration increased 18 mEq/L over20 hours, and she developed a “locked-in syndrome”with radiologic evidence of the osmotic demyelination

Thiazide

Hypertension

HypokalemiaDiabetesmellitus

Figure 9. Hypertension-hypokalemia-diabetes: A three-wayrelationship. Hypertension treated with thiazides, especiallyin higher dosages, can cause hypokalemia. Hypokalemia, inturn, can aggravate hypertension and also lead to diabetes.Diabetes, in turn, can cause hypertension, and people withhypertension are more likely to get diabetes. Correcting K�

stores may, therefore, be beneficial for both diabetes andhypertension. Reprinted with permission from reference 5(Agarwal R: Hypertension, hypokalemia, and thiazide-in-duced diabetes: A 3-way connection. Hypertension 52:1012–1013, 2008).


syndrome. Two important and underrecognized pointsthat the authors lucidly discussed in this publicationare reemphasized. First, when K� salts are given toreplace K� losses, they are osmotically active saltsand, to the extent that they are retained in the body,will increase the plasma osmolality and the Na� con-centration as much as equimolar quantities of retainedNa� salts (11). The physicians ordered water restric-tion and only 300 ml of isotonic saline over the first 24hours, which by itself should not have raised the Na�

concentration very rapidly. However, they also treatedher K� deficit with 430 mEq of KCl over the first 24hours. This K� replacement and a 3-L electrolyte-freewater diuresis that ensued resulted in her Na concen-tration increasing from 96 to 114 mEq/L, or 18 mEq/L,over 20 hours. The other issue relates to the impact onbrain cell swelling or shrinking when hyponatremia iscaused mainly by water gain, by Na� loss, or by K�

loss. In general, acute hyponatremia means that braincells have swollen. The degree of swelling graduallydecreases as the brain cells adapt with time. Con-versely, when the serum Na� concentration increasesas a result of the hypertonic saline infusion or a waterdiuresis water moves out of the brain cells and theyshrink. If the Na� concentration increases too rapidlyand/or too much, osmotic demyelination syndromemay occur. The exact mechanism of the demyelinatingdamage remains obscure but too-rapid cell shrinkageis thought to contribute. However, these generaliza-tions may not apply when the mechanism for thehyponatremia is mainly K� depletion. Analogously,they may not apply when the increase in Na� concen-tration is largely due to K� repletion. Under thesecircumstances, the development of hyponatremiawould be associated with brain cell shrinkage as aresult of brain cell K� depletion. A fall in intracellularosmotic solute will cause water to move from the cellsinto the extracellular fluid (ECF). Expansion of theECF dilutes the Na� concentration while the intracel-lular space simultaneously contracts. Analogously, tothe extent that correction of hyponatremia is the resultof K� repletion, it occurs because water moves fromthe ECF into cells and they should therefore swell. Ifthese changes had occurred in the patient described inthe article, it would be difficult to attribute the devel-opment of demyelination to shrinking of her braincells. Nonetheless, despite these considerations, ep-idemiologic studies have found repeatedly that hy-pokalemia is a risk factor for the development of the

osmotic demyelination syndrome in patients withhyponatremia with treatment (12,13). Undoubtedly,the brain cell swelling/shrinking hypothesis for thedemyelination process is an oversimplification. It issuggested that the interested reader carefully reviewthe Berl and Rastegar (10) publication as well as anin depth discussion of these issues by Nguyen andKurtz (11).

References1. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA,

Izzo JL Jr, Jones DW, Materson BJ, Oparil S, Wright JT Jr, RoccellaEJ: Seventh report of the Joint National Committee on Prevention,Detection, Evaluation, and Treatment of High Blood Pressure. Hy-pertension 42: 1206–1252, 2003

2. ALLHAT Collaborative Research Group: Major cardiovascularevents in hypertensive patients randomized to doxazosin vs chlortha-lidone: The antihypertensive and lipid-lowering treatment to preventheart attack trial (ALLHAT). JAMA 283: 1967–1975, 2000

3. Hulley SB, Furberg CD, Gurland B, McDonald R, Perry HM,Schnaper HW, Schoenberger JA, Smith WM, Vogt TM: SystolicHypertension in the Elderly Program (SHEP): Antihypertensive ef-ficacy of chlorthalidone. Am J Cardiol 56: 913–920, 1985

4. Shafi T, Appel LJ, Miller ER 3rd, Klag MJ, Parekh RS: Changes inserum potassium mediate thiazide-induced diabetes. Hypertension52: 1022–1029, 2008

5. Agarwal R: Hypertension, hypokalemia, and thiazide-induced diabe-tes: A 3-way connection. Hypertension 52: 1012–1013, 2008

6. Barri YM, Wingo CS: The effects of potassium depletion andsupplementation on blood pressure: A clinical review. Am J Med Sci314: 37–40, 1997

7. Dickinson HO, Nicolson DJ, Campbell F, Beyer FR, Mason J:Potassium supplementation for the management of primary hyper-tension in adults. Cochrane Database Syst Rev 3: CD004641, 2006

8. Korgaonkar S, Tilea A, Gillespie BW, Kiser M, Eisele G, FinkelsteinF, Kotanko P, Pitt B, Saran R: Serum potassium and outcomes inCKD: Insights from the RRI-CKD cohort study. Clin J Am SocNephrol 5: 762–769, 2010

9. Chan KE, Lazarus JM, Hakim RM: Digoxin associates with mortalityin ESRD. J Am Soc Nephrol 21: 1550–1559, 2010

10. Berl T, Rastegar A: A patient with severe hyponatremia and hypo-kalemia: Osmotic demyelination following potassium repletion. Am JKidney Dis 55: 742–748, 2010

11. Nguyen MK, Kurtz I: Role of potassium in hypokalemia-inducedhyponatremia: Lessons learned from the Edelman equation. Clin ExpNephrol 8: 98–102, 2004

12. Lohr JW: Osmotic demyelination syndrome following correction ofhyponatremia: Association with hypokalemia. Am J Med 96: 408–413, 1994

13. Soupart A, Decaux G: Therapeutic recommendations for manage-ment of severe hyponatremia: Current concepts on pathogenesis andprevention of neurologic complications. Clin Nephrol 46: 149–169,1996

Hyperkalemia

PhysiologyAldosterone plays an important role in the de-

fense against hyperkalemia. Aldosterone stimulateselectrogenic sodium (Na�) reabsorption through the


epithelial Na� channel (ENaC), creating a lumen-negative potential that serves as a driving force forCl� reabsorption through the paracellular pathway andsecretion of potassium (K�) and hydrogen (H�) intothe collecting duct lumen.

Aldosterone secretion is mediated by a directstimulatory effect of angiotensin II on cells in the zonaglomerulosa of the adrenal gland and through a directeffect of K� on the zona glomerulosa. As extensivelyreviewed in the last fluid and electrolyte issue of Neph-SAP, two members of the “with-no-lysine” (WNK) fam-ily of kinases (a name derived from the atypicalplacement of the catalytic lysine as compared withother types of kinases) play an important role inmodulating K� secretion, and mutations of these pro-teins are responsible for congenital abnormalities thatresult in hyperkalemia.

In adults, dietary K� intake is matched by K�

excretion primarily by the kidney, with a lesser con-tribution by the gastrointestinal tract. K� is freelyfiltered by the glomerulus and then reabsorbed in theproximal tubule and loop of Henle such that only 10%of the filtered load reaches the distal nephron. K�

delivery to the distal nephron is usually small and isfairly constant. Most of the K� in the final urinederives from K� secreted by the distal nephron (pri-marily the cortical collecting duct) in response tophysiologic need.

K� enters the tubular lumen of the collectingduct through K� channels. The ROMK channel,which has a low conductance and a high probability ofbeing open under physiologic conditions, is the majorK�-secretory pathway. The maxi-K� channel, foundmostly in intercalated cells, is relatively quiescent inthe basal state but plays a major role in flow-stimu-lated K� secretion. A high K� intake stimulates K�

secretion mediated by both aldosterone-dependent andaldosterone-independent mechanisms that are summa-rized in a recent review (1).

A typical Western meal provides a K� load thatequals the total K� content of extracellular fluid.Thus, every meal represents a life-threatening K�

challenge, and the organism must quickly adapt to it.Aldosterone is an unlikely candidate to mediate thisadaptation because the effects of the hormone requiretranscription, requiring too much time, and becausealdosterone also effects sodium and acid-base balance.The serine protease tissue kallikrein (TK) was recentlyidentified as an aldosterone-independent mechanism

that protects against postfeeding hyperkalemia by pro-moting increased renal K� excretion (2). TK is syn-thesized in large amounts by connecting tubule cellsand secreted into the tubular lumen; its secretion ispromoted by K� intake. TK activates the ENaC,which promotes K� secretion and inhibits K� reab-sorption by H�-K�-ATPase. Knockout mice lackingTK become more hyperkalemic than wild-type miceafter a large feeding, and cortical collecting ductsisolated from TK null mice absorb potassium ratherthan secrete it; K� absorption is associated with in-creased H�-K�-ATPase.

PseudohyperkalemiaPseudohyperkalemia is an in vitro phenomenon

that results from mechanical release of K� from cellsduring phlebotomy or specimen handling. Commoncauses include fist clenching, use of small-bore nee-dles during phlebotomy, and exposure of blood sam-ples to lower ambient temperatures either during trans-port or by placing the sample on ice.

Inappropriate phlebotomy techniques in whichpatients are requested to fist clench to facilitate blooddraws is known to cause pseudohyperkalemia, but theincidence of the problem is difficult to determine. Aprogram to retrain phlebotomists to eliminate thispractice in a primary care population in the UnitedKingdom reduced the percentage of K� values abovethe reference range (�5.2 mmol/L) from 9 to 6% andreduced the percentage of K� values �5.8 mmol/Lfrom 0.9 to 0.5% (3). Similarly, when a standardprotocol was introduced in a university hospital inJapan that mandated avoidance of fist clenching dur-ing phlebotomy and nonuse of the first blood samplefor electrolyte measurements when multiple speci-mens were obtained from a single patient, the numberof cases of pseudohyperkalemia fell from eight to oneper year (4).

A spurious increase in plasma K� concentrationcaused by in vitro contamination with K-EDTA,sometimes used as an anticoagulant in sampling tubes,should be considered when unsuspected hyperkalemiais accompanied by a very low plasma Ca2� concen-tration (5). Most other cases of pseudohyperkalemiaare seen in patients with thrombocytosis and pro-nounced leukocytosis (6,7).

In vitro elevation of the serum K� caused byrelease of K� from platelets during clot formation inthe specimen tube was first reported in 1955 by Hart-


mann and Mellinkoff in patients with thrombocytosis;plasma K� level (sampled from anticoagulated bloodin heparinized tubes) was normal. To determine theincidence of pseudohyperkalemia in myeloprolifera-tive disorders, a 6-year retrospective audit was con-ducted on all patients who had thrombocytosis andwere referred to the hematology department of a largedistrict general hospital in Ireland (8). Ninety patientswith thrombocytosis (platelet count range of 413,000to 2,997,000) as a result of reactive thrombocytosis,primary thrombocythemia, polycythemia vera, andmyelofibrosis were studied. The median serum K�

during the peak of thrombocytosis was 5.3 mmol/L(range 3.3 to 8.7 mmol/L). Pseudohyperkalemia wasmost frequent among patients with primary thrombo-cythemia (75.7%) and polycythemia rubra vera (75%),followed by myelofibrosis (50%) and reactive throm-bocytosis (34.5%). Among the 90 patients, a highlysignificant positive correlation was found between theplatelet count and the serum K� level (r � 0.998).Above a platelet count of 870,000, the highest re-corded serum K� levels always exceeded 5.5 mmol/L,and all patients with a platelet count of �1,192,000had at least one serum K� level of �6 mmol/L; serumK� levels as high as 8.7 with plasma K� in the normalrange were recorded. Pseudohyperkalemia led to unnec-essary admissions, repeated venipunctures, and inappro-priate therapy with intravenous calcium and insulin, andadministration of K�-binding resins. Investigators sug-gested that clinical laboratories could help cliniciansavoid these errors by checking platelet counts in patientswith unexplained serum K� levels of �5.5 mmol/L; aplatelet count �800,000 should prompt a request for aplasma K� in the appropriate heparinized tube.

Patients with high platelet counts and coexistentK� depletion may not exhibit the expected differencebetween serum and plasma K� values (9). Presum-ably, K� that has been released by platelets into theserum during clotting is subsequently taken up byK�-depleted erythrocytes, masking the pseudohyper-kalemia phenomenon; for example, one K�-depletedpatient with a myeloproliferative disorder and throm-bocytosis redeveloped pseudohyperkalemia after hisK� deficit was replaced (9).

In patients with isolated erythrocytosis fromchronic obstructive pulmonary disease and congestivecyanotic heart disease, the difference between serum andplasma K� level exceeded that in normal control sub-jects, but it was less than the discrepancy found in

patients with thrombocytosis or mixed-type disorders (9).In patients with erythrocytosis, K� released fromplatelets during the clotting process enters a smallervolume of serum than in patients with a normalhematocrit level; however, even with extreme iso-lated erythrocytosis, serum K� is no more than 0.7mmol/L higher than plasma K�.

High white blood cell (WBC) counts (�120,000)caused by chronic lymphocytic leukemia can also leadto falsely increased K� concentrations, promptingunnecessary therapeutic interventions (6,7,10). How-ever, unlike thrombocytosis, spurious hyperkalemiaoccurs in both serum and plasma samples and mayactually be more prominent when blood is sampled inheparinized tubes (10). For plasma collection, theblood specimen is centrifuged immediately, and thismay lead to in vitro cell destruction and release of K�

as these cells are freely suspended in the plasma.WBCs in patients with chronic lymphocytic leukemiaare fragile, and coupled with higher WBC counts,more cells are destroyed during the direct centrifuga-tion process, leading to a falsely increased K� value.This phenomenon can be visualized by the presence ofa thin layer of cells above the gel in the plasma tubeafter centrifugation in patients with a very high WBCcount. In contrast, this layer of cells is not visiblypresent in a serum tube after centrifugation. Serum(plasma without the clotting factors) is collected byfirst incubating a blood sample at room temperature toallow clotting to separate the serum from the cellsbefore centrifugation. During the clotting process, afibrin clot is formed, generating a matrix that entrapsand protects fragile leukemic WBCs, minimizing celllysis and in vitro release of K� into the serum. Mea-surement of whole-blood K� in an uncentrifugedspecimen using direct potentiometry is another option,and this agrees well with values obtained in serum(10). Pseudohyperkalemia has also been reported aftermechanical disruption of leukocytes during the transportof blood samples by pneumatic tube systems; in onereported case, spurious hyperkalemia led to unnecessarydialysis, which resulted in hypokalemia (11).

Excess K� IntakeWith normal renal function, it is difficult to

ingest enough K� to cause hyperkalemia. However,excess K� intake is an important contributing cause ofhyperkalemia in patients with impaired kidney func-tion. Melons, citrus juice, and salt substitutes are


widely recognized sources of K� intake, whereasmore exotic sources of excess K� can include rawcoconut juice (K� concentration 44 mmol/L), Nonijuice (56 mmol/L), river bed clay (100 mmol/100 gclay), and burnt match heads, a hazard to those proneto eating them (cautopyreiophagia) (12).

Internal K� ShiftsCellular redistribution of K� is an important

cause of hyperkalemia; a shift of as little as 2% ofbody K� stores can increase the plasma K� concen-tration to 8 mmol/L. When renal function is intact, K�

shifts cause only transient hyperkalemia because theexcess K� is normally excreted in the urine.Hyperglycemia. Hyperglycemia has multiple ef-fects on internal K� balance. Low insulin levels pro-voke translocation of intracellular K� to the extracel-lular compartment. In addition, hypertonicity and thecatabolic state associated with hyperglycemia causeegress of K� from cells. In patients with anuria,changes in the serum K� concentration caused byhyperglycemia and its correction are reflections ofinternal K� balance. Serum K� values of �5.5mmol/L were encountered in one third of episodes ofnonketotic hyperglycemia and three fourths of epi-sodes of diabetic ketoacidosis in hemodialysis pa-tients. Insulin corrects the hyperkalemia of dialysis-associated hyperkalemia and is usually the onlytreatment needed. Lethal hyperkalemia has been re-ported, and emergency dialysis is often considered forpatients with severe hyperkalemia accompanied byelectrocardiographic (ECG) abnormalities; it is un-known whether this intervention is really needed. In arecent case report, a 33-year-old woman who hadESRD and poorly controlled diabetes and presentedwith hypertension, blood sugar level of 1884 mg/dl(104.7 mmol/L), K� level of 7.2 mmol/L, serumbicarbonate level of 4 mmol/L, and T-wave elevationon electrocardiogram was treated with an insulin dripfor 2.5 hours before starting on dialysis because ofpersistent coma and severe hyperkalemia at presenta-tion (13). Predialysis chemistries showed improve-ment of blood glucose to 1782 mg/dl (a decrease of 40mg/dl per h) and a 2-mmol/L fall in K� to 5.2 mmol/L.Dialysis was stopped early because of concerns thatthe rapid fall in tonicity could cause cerebral edema.Octreotide. Octreotide, a long-acting synthetic oc-tapeptide analog of the human hormone somatostatin,has been reported to be useful in the treatment of

sulfonylurea-induced hypoglycemia that is refractoryto conventional therapy, especially in patients withrenal disease. Similar to somatostatin, octreotide di-rectly inhibits the release of insulin from the pancreas,and it also suppresses the secretion of glucagon,growth hormone, vasoactive intestinal peptide, andgastrin. Suppression of insulin release impairs cellularuptake of K� and can result in hyperkalemia if K�

excretion is impaired. A 48-year-old patient who hadESRD and was on maintenance hemodialysis wastreated with octreotide for sulfonylurea-induced hypo-glycemia and developed severe hyperkalemia (7.3mEq/L) after three doses (14). There have only beenthree other reports of the use of octreotide in patientswith ESRD, and one of these patients developed hy-perkalemia; there have been two other reported casesof octreotide-induced hyperkalemia in patients withoutESRD (14).Hyperkalemic Periodic Paralysis. Hyperkalemicperiodic paralysis is an inherited disorder caused bymutations in SCN4A encoding the voltage-gated skel-etal muscle sodium channel; an extensive review ofthe disorder, including the various allelic variantsassociated with the disease is available online(www.ncbi.nlm.nih.gov/omim/17-500). The disease ischaracterized by attacks of flaccid limb weakness andsometimes weakness of the muscles of the eyes, throat,and trunk; hyperkalemia (�5 mmol/L or an increase ofat least 1.5 mmol/L in plasma K�); and provocation ofattacks by K�-rich food or K� supplements. Sponta-neous attacks typically begin before breakfast, last for15 minutes to 1 hour, and then resolve. Serum K�

concentrations are normal between attacks. Attackscan be aborted by intake of carbohydrates, inhalationof �2-adrenergic agents, or intravenous calcium glu-conate. Continuous use of thiazide diuretics or acet-azolamide and avoidance of fasting and K�-rich foodscan prevent attacks.Succinylcholine. Hyperkalemia from succinylcho-line was first recognized by Gronert in burn units inthe 1960s and was studied in a controlled trial thatcompared serial electrocardiograms and arterial K�

values in 89 burn victims who were given anestheticagents, as compared with 16 normal control subjects.Subsequent studies have shown that in addition topatients with burns, succinylcholine-induced hyperka-lemia occurs in patients with muscle trauma and inpatients with neurologic disorders that result in motormuscle defects. Hyperkalemia results from extrajunc-


tional acetylcholine receptor spread along the musclemembrane in conditions that undergo prolonged depo-larization in response to succinylcholine with releaseof K�. In addition, myopathies that weaken skeletalmuscle membranes are subject to rhabdomyolysis withstress. Gronert (15) recently published a fascinatingbiographical recounting of his studies related to theepidemiology and pathophysiology of succinylcho-line-induced hyperkalemia that includes ECG record-ings of a burn patient whose serum K� increased from4.7 to 9.1, accompanied by a widened QRS complexand loss of P waves within 5 minutes of receivingsuccinylcholine under close observation. Remarkably,no patient experienced a cardiac arrest in these studiesdespite the extreme hyperkalemia that was observed.Rebound hyperkalemia may occur after K� chloridetreatment of thyrotoxic hypokalemic periodic paralysis(TPP) beginning 2 to 3 hours after recovery (16).During the recovery phase, K� release from muscleoccurs at rates up to 15 mmol/L per h. Reboundhyperkalemia has been reported to occur in 40 to 70%of patients who are treated with intravenous K�. Afatal case of iatrogenic hyperkalemia was recentlyreported in a 40-year-old Hispanic woman who hadTPP and was treated with 240 mmol of KCl (at 30mmol/L per h) for a serum K� level of 1.9 mmol/L(16). K� therapy was stopped 8 hours after the start oftherapy, when the serum potassium rose to 6.6mmol/L. However, despite the administration of 50 mlof 50% dextrose and 10 U of insulin, serum K�

continued to increase, reaching 10.1 mmol/L at thetime of death. Oral and intravenous propranolol havebeen effective in reversing TPP hypokalemia withoutrisk for rebound hyperkalemia. A similar phenomenonhas been described in patients with drug-induced in-ternal shifts of K�. Barbiturates, which are often usedas a neuroprotective agent in patients with head traumaand other neurosurgical interventions, cause hypoka-lemia by shifting K� into cells. Once the administra-tion of barbiturate is stopped, a rebound hyperkalemiamay ensue, which has been fatal in several reportedcases (17). K� replacement to correct severe barbitu-rate-induced hypokalemia is likely to increase theseverity of hyperkalemia once the drug is stopped. Ina recent case report, dramatic, severe fluctuations ofthe serum K� every few hours between 1.8 and 8.3mmol/L developed in a patient who had head traumaand was treated with therapeutic hypothermia (34.5°C)and norepinephrine (0.1 �g/kg per min) to maintain

BP but no barbiturates (17). He was treated withparenteral K� supplements when hypokalemic andthen, hours later, with K�-lowering agents (insulin,furosemide, and Kayexalate) when his K� rose toohigh. The serum potassium ultimately increased abruptlyto 8.9 mmol/L, resulting in cardiac arrest and treatmentwith venovenous hemofiltration. Presumably, endog-enous catecholamine discharge after head trauma re-sulted in a transient �2-adrenergically mediated shiftof K� into cells augmented by exogenous norepineph-rine and hypothermia.Cardiac Glycosides. Inhibition of Na-K-ATPaseby cardiac glycosides results in migration of K� out ofcells, and digitalis poisoning can cause clinically sig-nificant hyperkalemia. The stems, leaves, flowers,roots, and seeds of two common species of Oleander(Nerium oleander, or common oleander, and Thevetiaperuviana, or yellow oleander) contain high concen-trations of cardiac glycosides, and they have become acommon cause of poisoning in tropical and subtropicalparts of the world (18,19). Deliberate self-harmthrough ingestion of yellow oleander is very popular inSouth Asia, and several thousand cases have beenreported each year with a fatality rate ranging between4 and 10%. Symptoms are those of digitalis intoxica-tion and include nausea, vomiting, abdominal pain,diarrhea, cardiac arrhythmias, and hyperkalemia. Themanagement of serum K� is difficult because hypo-kalemia can worsen digitalis toxicity and predispose todangerous arrhythmias, but severe toxicity results inhyperkalemia. The treatment of hyperkalemia andother complications of oleander poisoning was re-cently reviewed (19). Intravenous calcium is generallyavoided because intracellular calcium concentrationsare already high in the setting of digoxin toxicity andcalcium administration may worsen arrhythmias. In-sulin-dextrose infusion was shown to be cardioprotec-tive in an animal model; in addition to the K�-lowering effect of insulin, which drives K� back intocells, insulin may modify access of digoxin to theNa-K-ATPase–binding site, reducing its toxicity.Treatment with digoxin-specific Fab antibody frag-ments is very effective in the correction of hyperkale-mia and is the preferred therapy; however, their highcost and lack of availability limit their use in countrieswhere yellow oleander poisoning is common. Treat-ment with K�-binding resins or other measures aimedat eliminating potassium is ill advised because thecause of hyperkalemia is a shift of K� out of cells


rather than excess body K�; K�-depleting therapiesmay precipitate hypokalemia when given with digoxin-specific antibody fragments.

Impaired K� ExcretionAlthough redistribution of K� can cause hyper-

kalemia, a sustained increase in the plasma K� con-centration is indicative of defective K� excretioncaused by oliguric renal failure, low aldosterone lev-els, or aldosterone resistance. Calculation of the tran-stubular K� concentration (TTKG) may be helpful indistinguishing between these two possibilities:

TTKG � (urine [K�]/plasma [K�])/(urine osmola-lity/plasma osmolality)

The TTKG is intended to estimate the tubularfluid K� concentration in the cortical collecting duct,the site responsible for most excreted K�, where theluminal fluid is isosmotic to the plasma. The calcula-tion assumes that changes in the K� concentrationbetween that point and the final urine result from waterreabsorption, which raises urine osmolality aboveplasma and increases the concentration of K� in thefinal urine above that in the collecting duct lumen.Impaired K� excretion as a result of defects in tubularK� secretion results in a low TTKG, whereas impairedexcretion attributable to decreased delivery of fluid tothe secretory site results in a higher TTKG. There is apositive correlation between mineralocorticoid activityand the TTKG, but values have varied greatly in thevarious published studies (20). The greatest clinicalutility of the TTKG may be to differentiate betweenhyperkalemia caused by low aldosterone levels andhyperkalemia caused by aldosterone resistance and togauge the effectiveness of mineralocorticoid replace-ment therapy. Administration of physiologic doses ofmineralocorticoid (0.05 mg of fludrocortisone) to pa-tients with adrenal insufficiency increases the TTKGto �6 within 4 hours, whereas in patients with aldo-sterone resistance, the TTKG remained �6 with avariable or delayed response after 24 hours to phar-macologic doses (0.2 mg) (20).Decreased Mineralocorticoid Levels or Activity. De-creased mineralocorticoid activity can be caused bydisturbances anywhere along the renin-angiotensin-aldosterone system (RAAS) arising from disease ordrug effects.Addison Disease. When Addison disease was firstreported in 1855, most of the cases were caused bydisseminated tuberculosis. Today, autoimmune adren-

alitis counts for 70 to 90% of the cases; tuberculosis isresponsible for 7 to 20%; and the remaining cases arecaused by adrenal hemorrhage or infarction, drugs,and malignancy (21,22). Infiltration of the adrenalglands by metastatic cancer is found at autopsy in 40to 60% of patients with disseminated lung or breastcancer, probably because of their rich blood supply.Although most adrenal metastases are thought to befunctionally unimportant, adrenal insufficiency can bemissed in patients with cancer because hyperkalemiacan be ascribed to other comorbidities such as im-paired renal function and K�-sparing diuretics. Hyper-kalemia can be the first sign of adrenal metastases, aswas the case in a patient who had adenocarcinoma ofthe lung and presented with tetraparesis and widenedQRS complexes on electrocardiogram caused by aserum K� of 8.8 mmol/L (22).Hypoaldosteronism. Functional hypoaldosteron-ism with hyperkalemia may occur after successfulresection of aldosterone-producing adenomas (23).Long-term suppression of contralateral aldosteronesynthesis by the adenoma or by chronic untreatedhypokalemia can result in hyperkalemia with serumK� concentrations of �6 mmol/L and postural hypo-tension for several weeks after surgery; administrationof fludrocortisones may be necessary.Pseudohypoaldosteronism. Pseudohypoaldoste-ronism type 2, also known as Gordon syndrome, is adominantly inherited hyperkalemic hypertension withan associated (usually mild) acidosis, in which re-moval of the distal NaCl co-transporter from the distalconvoluted tubule apical surface and insertion ofROMK into the collecting duct membrane is impairedbecause of mutation in one of their regulators, WNK1or WNK4 (with-no-lysine kinases). The pathogenesisof this disorder was discussed in the previous fluid andelectrolyte issue of NephSAP. Patients with the WNK4mutation have hypercalciuria, whereas patients withthe WNK1 mutation do not (24).

Drug-Induced HyperkalemiaInhibition of the RAAS is a key strategy in

treating hypertension and cardiovascular and renaldiseases. However, RAAS inhibitors (angiotensin-converting enzyme inhibitors [ACEIs], angiotensinreceptor blockers [ARBs], aldosterone receptor antag-onists, and direct renin inhibitors) increase the risk forhyperkalemia, particularly when they are prescribed incombination (25). Common drugs that are associated


with hyperkalemia and their mechanisms of action arelisted in Table 3.Decreased Aldosterone Levels. Several studieshave shown that ACEIs can cause hyperkalemia byinterfering with the production and/or secretion ofaldosterone. An analysis of the African AmericanStudy of Kidney Disease and Hypertension (AASK)database showed that the incidence of severe hyper-kalemia (defined as a serum potassium �5.5 mmol/L)was significantly higher among 1094 African Ameri-can patients who did not have diabetes, had hyperten-sion, had a GFR of 20 to 65 ml/min, and wererandomly assigned to treatment with ACEIs as com-pared with patients who were treated with calciumchannel blockers or a � blocker (26). However, therewas no increased risk for hyperkalemia when the GFRwas �40 ml/min. Diuretic use decreased the risk forhyperkalemia by 59%. Patients with a body massindex (BMI) of �25 were more likely to develophyperkalemia on an ACEI, possibly because the low-BMI group might have had a lower GFR than whatwas estimated. The results of the study suggest thatmore frequent monitoring of the serum K� is neces-

sary for patients who are treated with an ACEI andwhose GFR is �40 ml/min; patients who have a lowBMI; all patients with a GFR �30 ml/min, regardlessof BMI; older patients; patients with higher levels ofmicroalbuminuria; and patients who are not receivingconcurrent treatment with diuretics.

A review of the PubMed database of clinicaltrials published through December 2008 found that therisk for developing a serum potassium �6 mmol/L is�2% with RAAS inhibitor monotherapy. The risk forhyperkalemia with monotherapy remains low in pa-tients who have chronic kidney disease (CKD) and aGFR of 30 to 60 ml/min. The risk for hyperkalemiawith ACEI/ARB combination therapy is higher thanwith monotherapy; however, the number of patientswho experience “hyperkalemia events” (discontinua-tion or adverse event of hyperkalemia or serum K�

�6.0 mmol/L) was similar for combination therapyand lisinopril monotherapy. The addition of an aldo-sterone receptor antagonist is associated with a smallbut statistically significant 0.3-mmol/L increase inserum K� levels, and the incidence of severe hyper-kalemia (serum K� �6.0 mmol/L) is also greater.Aldosterone Receptor Antagonism. Spironolac-tone blocks the effect of aldosterone on the mineralo-corticoid receptor. The Randomized Aldactone Eval-uation Study (RALES) showed benefit in treating withspironolactone patients who had heart failure. In Can-ada, an increase in hospital admissions and deathsfrom hyperkalemia occurred concurrently with thepublication of RALES. A study of a stable populationin a town in Scotland (population 400,000) found thatthe number of prescriptions for spironolactone and thenumber of ordered laboratory tests for creatinine andserum K� doubled after the release of the RALESresults in 1999 (27). Similarly, among patients whowere taking ACEIs and had been recently admitted tothe hospital for heart failure, use of spironolactoneincreased from 19.8% before the publication ofRALES to 70.1% after the trial. Although the rate ofmild hyperkalemia (�5 mmol/L but �6 mmol) in-creased as the number of spironolactone prescriptionsand laboratory K� determinations increased, the rateof severe hyperkalemia (�6 mmol/L) did not. Thenumber of hospital admissions for hyperkalemia re-mained at three or less per quarter before and afterRALES. During the 6-year study, a total of 578 pa-tients with spironolactone-associated severe hyperka-lemia (K� �6 mmol/L) were identified, including 172

Table 3. Drugs that cause hyperkalemia

Mechanism Drugs

Exogenous potassium Potassium supplementsSalt substitutesHerbal medicines (e.g.,

alfalfa, dandelion)Redistribution of intracellular

potassiumHypertonic mannitol� blockersSuccinylcholineSomatostatinOctreotideDigitalis glycosides

Decreased aldosterone levels ACEIsARBsHeparinNSAIDsTacrolimus

Aldosterone receptor antagonism SpironolactoneEplerenone

Collecting tubule sodiumchannel blockade

TrimethoprimPentamidineAmilorideTriamterene

Collecting tubule Na�-K�-ATPase blockade

CyclosporineTacrolimusx


patients with heart failure, 45 with cirrhosis, and 361with hypertension. Among those with hyperkalemia,75% were older than 65 years. A serum creatinine�2.5 mg/dl (�220 �mol/L) preceded hyperkalemia in76% of patients with heart failure, 71% of patientswith cirrhosis, and 53% of patients with hypertension.The authors took this to mean that spironolactone is asafe drug when it is used with appropriate monitoring(presumably more diligent in Scotland than it was inCanada!) and that closer monitoring is indicated forolder patients and for patients with impaired renalfunction.

There is growing evidence that aldosterone con-tributes to hypertension and insulin resistance in obe-sity (28,29). Adipokines released by the adipocyteinduce aldosterone secretion from human adrenocor-tical cells and sensitize the cells to stimulation byangiotensin II. The resulting hyperaldosteronism pro-motes insulin resistance, inflammation, oxidativestress, and sodium retention, which contribute to thedevelopment of resistant hypertension. Mineralocorti-coid receptor blockade seems to be useful in treatingpatients who have both the metabolic syndrome andresistant hypertension (30,31). Use of spironolactonefor the treatment of hypertension in patients withcoexistent renal disease carries a risk for hyperkale-mia. An observational study conducted in two univer-sity-based hypertension clinics identified 46 patients(38 of them African Americans) with resistant hyper-tension and stage 2 or 3 CKD and evaluated the safetyand efficacy of addition of aldosterone blockade topreexisting BP-lowering therapy that included a di-uretic and a renin-angiotensin system (RAS) blocker(31). Addition of an aldosterone receptor antagonistfurther decreased systolic BP by an average of 14.7mmHg. The mean increase in serum K� was 0.4mEq/L, with 17% of the patients having a serum K� of�5.5 mEq/L. Of the eight patients with hyperkalemia,only one had a serum K� of �6 mEq/L. The odds ratiofor developing hyperkalemia was significantly in-creased when the baseline estimated GFR (eGFR) was�45 ml/min, baseline serum K� was �4.5 mEq/L, orthe eGFR fell by �30% after the introduction of thealdosterone antagonist. The study suggests that pa-tients who are given aldosterone antagonists for resis-tant hypertension should have their serum creatinineand K� checked within 1 to 2 weeks after initiation oftherapy (similar to the monitoring after the initiation ofRAS blockade). Hyperkalemia seems to be a manage-

able risk when it is limited to patients with an eGFR of�45 ml/min and a baseline serum K� of �4.5 mEq/Lwhile on optimal diuretic and RAS-blocker therapy.

The non–testosterone-derived progestin dro-spirenone, used as an oral contraceptive agent, blocksthe mineralocorticoid receptor and in the dosage usedfor contraception is roughly equivalent to 25 mg ofspironolactone (32). Monitoring of the serum K� con-centration has been recommended when these drugsare prescribed to patients who are treated with K�

supplements, ACEIs, angiotensin receptor antagonists,or nonsteroidal anti-inflammatory drugs (12). How-ever, a study of a drospirenone-containing contracep-tive identified only one case of hyperkalemia among22,429 women who began ethinyl estradiol 0.03 mg/drospirenone and none among women who were iden-tified as having preexisting adrenal, renal, or hepaticinsufficiency (32).Collecting Tubule Sodium Channel Blockade. Treat-ment of Pneumocystis jiroveci pneumonia (PCP) withtrimethoprim-sulfamethoxazole results in hyperkale-mia in as many of 53% of patients with HIV infectionbecause of an amiloride-like effect of trimethoprim onsodium channels in the distal nephron. Corticosteroidsare often used as adjunctive therapy in patients withmoderate to severe PCP, and they might be expectedto ameliorate trimethoprim-associated hyperkalemiaby increasing GFR and by exerting a mineralocorti-coid effect. However, a single-center study of 30patients who had PCP and were treated with trim-ethoprim-sulfamethoxazole found that the addition ofglucocorticoids potentiates the risk for hyperkalemia(33). Patients with serum creatinine �1.5 mg/dl andthose who were taking medications that are known toinfluence serum K� levels were excluded from thestudy. Hyperkalemia developed in seven of the 18patients who were treated with trimethoprim-sulfame-thoxazole plus prednisolone and none of the 12 pa-tients treated with trimethoprim-sulfamethoxazolewithout prednisolone; release of K� caused by steroid-induced catabolism was thought to be the best expla-nation for this finding because each gram of nitrogenreleased from muscle is accompanied by 2.7 mEq ofK�. Although the investigators did not measure uri-nary K� losses in the two groups, the blood ureanitrogen concentration on the eighth day of therapywas significantly higher in patients who were receiv-ing steroids (30.0 versus 11.8 mg/dl).

Treatment with trimethoprim-sulfamethoxazole


is a potential risk factor for hyperkalemia in ambula-tory patients who are treated with ACEIs or ARBs.The risk of trimethoprim-sulfamethoxazole in thesepatients was explored in a population-based nestedcontrol study of elderly patients who were aged � 66years and resided in Ontario, Canada (34). Comparedwith amoxicillin, the use of trimethoprim-sulfame-thoxazole was associated with nearly a sevenfold in-creased risk for hospitalization for hyperkalemia. Noincreased risk was seen with any of the comparatorantibiotics that are used to treat urinary tract infec-tions. The investigators could not identify outpatienthyperkalemia (and they did not have access to serumK� values), so the study probably underestimates theclinical consequences of the drug interaction betweenRAS blockade and trimethoprim. The findings suggestthat the drug should be avoided in elderly patients whoreceive RAS blockade and/or aldosterone antagonistswhen other options exist.

Consequences of HyperkalemiaMuscle Weakness. Neuromuscular manifestationsof hyperkalemia range from paresthesias, mild motorparalysis, paraparesis, ascending paraplegia, and even-tually quadriplegia. Paralysis usually begins distallywith an ascending course that may mimic Guillain-Barre syndrome. The diaphragm, cranial nerves, andsensory functions are usually unaffected; muscleweakness from hyperkalemia usually reflects a serumK� �7 mmol/L (35).ECG Findings. In experimental settings, hyper-kalemia has been associated with a defined series ofECG findings: Shortening of the QT interval, peak-ing of T waves, QRS prolongation, shortening of thePR interval, reduction in P wave amplitude, loss ofsinoatrial conduction with onset of a wide-complex“sine-wave” ventricular rhythm, and ultimatelyasystole. The most severe cardiac manifestationshave been shown to occur with serum K� concen-trations �9 mEq/L (36). As discussed in the previ-ous fluid and electrolyte issue of NephSAP, clinicalstudies have shown a poor correlation between theserum K� concentration and cardiac manifestations;in a recent series, abnormal ECG findings weredocumented for only 12% of patients with serumK� values of 6.0 to 7.1 and in 39% of patients withvalues of 7.2 to 9.3 (37).Mortality. A prospective, observational study ofadult patients with stages 3 to 5 CKD conducted at

four outpatient nephrology clinics in the United Statesidentified 86 deaths before reaching ESRD among 820patients (38). The investigators found that pre-ESRDmortality was significantly higher for a serum K�

�4.0 mmol/L compared with K� values �4.0 and�5.5 mmol/L, but surprisingly a higher serum K�

(�5.5 mmol/L) was not associated with elevated mor-tality. A multivariable analysis showed that the low-est risk for mortality was in patients with serum K�

values of 4.1 to 5.5 mmol/L. Even at serum potas-sium levels of 5.5 to 5.9 mmol/L, a range that oftenprovokes therapeutic intervention, no increase inmortality was observed. This modest level of hyper-kalemia seemed to be well tolerated in this patientpopulation from the perspective of predicting mor-tality risk.Metabolic Acidosis. Normal men fed a high-K�

diet decrease their urine pH and ammonium and netacid excretion; in experimental animals, K� loadinghas been shown to decrease ammonia generation in theproximal tubule. The precise mechanism is not firmlyestablished, but presumably hyperkalemia raises theintracellular pH of the tubular cell by exchanging withprotons, impairing enzymes that promote the deami-dation of filtered or secreted glutamine (39). In hu-mans, despite reduced renal ammonia production, ad-ministration of K� does not cause acidosis whenadrenal and renal functions are intact because upregu-lation of aldosterone accelerates hydrogen ion secre-tion. When aldosterone is deficient or in the presenceof genetic or acquired defects in function of the epi-thelial sodium channel, hyperkalemia is accompaniedby acidosis.

Treatment of HyperkalemiaGuidelines for the treatment of hyperkalemia are

based on consensus or expert opinion because of a lackof controlled clinical trials; most sources recommendemergency interventions for serum K� concentrations�6 mmol/L if there are ECG changes and for K�

values �6.5 mmol/L regardless of the electrocardio-gram (40). Severe ECG findings can be reversed byinfusion of calcium gluconate in a volume of 10 ml of10% solution infused over 3 to 5 minutes; calciuminfusion does not affect the serum K� level, butimprovement in the electrocardiogram can be seenwithin 1 to 3 minutes and its effect can last 30 to 60minutes. Calcium is contraindicated for patients whoare taking digoxin because it potentiates digoxin tox-


icity. A shift of K� into cells can be achieved withadministration of insulin or a �2-adrenergic agonist.Insulin is given as an intravenous bolus with enoughglucose to prevent hyperglycemia; the most commonlyrecommended regimen is a bolus injection of 10 U ofinsulin combined with 50 ml of 50% glucose givenover 5 minutes. Insulin’s hypokalemic effect can beseen within 20 minutes, peaking between 30 and 60minutes, and lasting for 6 hours. Salbutamol (theinternational nonproprietary name), known as albu-terol in the United States, is the most commonly used�2-adrenergic agonist for hyperkalemia, either as asingle agent or in combination with insulin. It can begiven by nebulizer (10 to 20 mg in 4 ml of saline); itseffect may be seen in 30 minutes, with maximumeffect at 90 to 120 minutes. Patients with acidosis canbe given isotonic bicarbonate, but the benefit is uncer-tain and routine use of bicarbonate for the treatment ofhyperkalemia is controversial. With the exception ofpatients with hyperkalemia caused by a shift of K� outof cells, definitive therapy of hyperkalemia requiresthat excess K� be eliminated. K� intake should berestricted, and medications that impair K� excretionshould be discontinued. Loop diuretics are often rec-ommended to enhance K� excretion (12,40) becauseof the known effect of flow in the distal nephron, but,surprisingly, no studies have documented their effec-tiveness in the acute treatment of hyperkalemia. Dial-ysis is the most effective means to remove K� frompatients with compromised renal function, but cationexchange resins, which are intended to eliminate K�

from the colon in exchange for K�, are often used asa temporizing maneuver; as discussed next, theirsafety and effectiveness have recently been challenged(41,42).Sodium Polystyrene Sulfonate. Sodium polysty-rene sulfonate (SPS; also known by its trade name,Kayexalate) has been used for 40 years to treat hyper-kalemia. In its early use, SPS was given suspended inwater, but constipation and fecal impactions associ-ated with the drug led to the recommendation that it begiven in a suspension with hypertonic sorbitol, anosmotic cathartic. Several reports of intestinal necrosiscaused by SPS in sorbitol ensued, initially in the renaltransplant literature. Most of the cases occurred in apostoperative setting or in critically ill patients withESRD. Until recently, bowel necrosis was thought tobe an extremely rare complication affecting very sickpatients who were given K�-binding resin enemas.

McGowan et al. (43) conducted a 9-year retrospective,single-center study at a university hospital and found29 patients with reports of SPS crystals in surgicalspecimens. Of these, 11 patients were identified withconfirmed intestinal ischemia (four of them fatal)temporally related to the oral administration of SPS-sorbitol (mean dosage 92 g, range 30 to 170 g). Onlytwo of these patients had been admitted for surgicalprocedures (both of them orthopedic); four (36%) ofthe 11 patients had ESRD that required dialysis, andthree had been treated with SPS-sorbitol despite nor-mal renal function. Three of the four patients with fatalbowel necrosis associated with SPS-sorbitol had beenadmitted to the hospital for noncritical illness, includ-ing one with normal renal function. Since McGowan’sstudy, there have additional reports of bowel necrosisassociated with K�-binding resins. A 64-year-oldwoman with severe heart disease and renal failuredeveloped severe abdominal cramping, abdominal dis-tention, and colonic dilatation a few hours after receiv-ing the first of three doses of SPS (30 g) in sorbitol; 2weeks later, massive hematochezia developed and acolonic biopsy showed ulcerated mucosa and promi-nent granulation tissue with eosinophilic, angulatedcrystals embedded in mucosal ulcers (44). A 56-year-old woman with stage 3 renal disease was given asingle dose of 15 g of SPS orally for a serum K� of 5.8mmol/L (45). She developed a large sessile mass in themidtransverse colon; when examined on biopsy, themass was shown to contain crystals, consistent withSPS-induced colonic injury. The authors noted thatshe really did not need the SPS that was given to herbecause she had no changes on her electrocardiogramand her mild hyperkalemia was already resolving inresponse to intravenous fluids. A 24-year-old womanwho had acute renal failure and a K� of 6.8 mmol/Land was treated with 30 g of the premixed SPS-sorbitol preparation and a 50-g SPS-sorbitol enemadeveloped patchy transmural small bowel necrosis 12days after exposure to the binding agents, resulting inpurulent peritonitis that required bowel resection; SPScrystal deposition was found within the intestinal mu-cosa. The patient had been receiving thiopental toinduce barbituric coma for epilepsy before SPS-sorbi-tol, and the resulting ileus may have increased hersusceptibility to this complication, and norepineph-rine, given to counteract the hypotensive effects ofthiopental for 10 days before the recognition of peri-tonitis, was likely to have reduced splanchnic blood


flow (46). On the basis of experiments on rats, sorbitolis thought to be the offending agent in this complica-tion. However, there has been one report of colonicmucosal necrosis with ulceration but not infarction ofthe bowel after administration of calcium polystyrenesulfonate, an analog of sodium polystyrene sulfonatewithout sorbitol; the patient had been given the drugboth orally and as an enema suspended in 20% dex-trose (47). Colonic biopsies showed the crystals withinthe ulcer bed tissue and within the necroinflammatorydebris.

In response to reports of these complications, inSeptember 2009, the US Food and Drug Administra-tion issued an advisory against concomitant adminis-tration of Kayexalate powder with sorbitol. Because itwas believed at the time that toxicity was related to70% but not 30% sorbitol, a premixed preparation ofSPS in 30% sorbitol was allowed to remain on themarket. However, several of the more recent reports ofcolon necrosis followed use of the premixed prepara-tion containing 30% sorbitol (43,44,46). These dis-turbing reports prompted an invited commentary in theAmerican Society of Nephrology’s official journal toreview the evidence showing that K�-binding resinsare safe and effective in the treatment of hyperkalemia(42). Synthetic cation-exchange resins are insolublepolymers with a molecular structure resembling acrystal lattice to which reactive carboxylic or sulfonicgroups are attached. When placed in a solvent, theresin’s reactive groups, which can be preloaded toform sodium, K�, calcium, or ammonium salts, ex-change their bound cations with cations dissolved inthe solvent. However, although hydrogen-cycled res-ins markedly increase fecal K� losses (albeit, at theexpense of acidosis from absorption of hydrogen ionsin exchange for K�), administration of a sodium-cycled carboxylic acid resin to experimental animalsshowed no increase K� excretion. This is not surpris-ing given both the limited amount of K� available forexchange in the colon and the competition with thehigh concentrations of ammonium in the colonic lu-men, which compete with K� for binding (41). Thereare no published data showing increased K� elimina-tion in the stool after administration of SPS to animals;the only evidence that it increases K� eliminationcomes from uncontrolled data showing K� binding inthe stool and a hypokalemic effect in four patients withrenal failure and a single normal volunteer. The bestevidence available that SPS lowers the serum K�

concentration is an uncontrolled study in 1961 show-ing that in 23 of 30 patients with hyperkalemia andrenal failure, the plasma K� fell by at least 0.4 mEq/Lin the first 24 hours after treatment with SPS sus-pended in water. Because of concurrent treatment withextremely low-K� diets and administration of largequantities of dextrose, it is uncertain that the hy-pokalemic effect was actually due to the resin. Be-cause most K� enters the bowel in the rectum, speed-ing delivery of the oral resin to the distal colon withcathartics would be expected to enhance its effective-ness, but adding sorbitol to SPS does not seem to makeit more effective in correcting hyperkalemia. Studiesof patients with normokalemia and mild hyperkalemiaand with ESRD found that the serum K� concentra-tion rose slightly (0.4 mEq/L) on placebo and did notchange during the course of 12 hours in response to asingle dose of 30 g of resin in water, 30 g of resin in60 g of sorbitol, or 60 g of sorbitol alone. The JASNcommentary concluded that “clinicians must weighuncontrolled studies showing benefit against uncon-trolled studies showing harm” as well as stating that it“would be wise to exhaust other alternatives for man-aging hyperkalemia before turning to these largelyunproven and potentially harmful therapies” (42). Aneditorial in the society’s clinical journal (CJASN)called these conclusions “immoderate,” arguing that(1) “the majority clinical consensus is that they dowork,” a conclusion supported by preliminary findingsfrom an observational study demonstrating a dose-dependent decrease in plasma K� within 8 hours oftreatment with SPS/sorbitol (48); (2) “the only other‘excretory’ modalities available are dialysis and loopdiuretics, both of which have limitations and potentialside effects, and like SPS, may take hours to have aneffect”; (3) “kaliuresis induced by loop diuretics(which have not, per se, been studied for the treatmentof hyperkalemia) requires adequate kidney function,which is often not present, and at high doses maycause azotemia and ototoxicity”; (4) “the estimatedincidence of colonic necrosis would be less than 0.1%per dose”; and (5) “demonstrating SPS crystals in thenecrotic lesion does not prove causality.” The authorsalso cited settings such as the Haitian earthquake andother disasters in which dialysis availability is limitedand SPS may be the only option for K� removal.

No one would argue that treatment with SPS insorbitol is the best available choice for patients withESRD in a disaster area. However, it is doubtful that


the 5 million doses of premixed SPS in sorbitol soldeach year are reserved for such patients. Most patientswho are given SPS have some renal function andserum K� values �6 mmol/L (48); it is difficult tojustify even a small risk for a drug that is probablyunnecessary. We clearly need more data on both thebenefit and the harm associated with SPS and sorbitol.We also need data on alternative therapies: Diureticsalone and in combination with bicarbonate and high-dosage mineralocorticoids and carthartics other thansorbitol.

In September 2009, the US Food and DrugAdministration issued an advisory againstconcomitant administration of Kayexalatepowder with sorbitol. Because it was be-lieved at the time that toxicity was relatedto 70% but not 30% sorbitol, a premixedpreparation of SPS in 30% sorbitol wasallowed to remain on the market. However,several of the more recent reports of colonnecrosis followed use of the premixedpreparation containing 30% sorbitol.

Fludrocortisone and Glycyrrhetinic Acid. SerumK� levels �6 mmol/L have been reported to occur inup to 12% of patients who receive long-term hemodi-alysis and is a frequent reason for emergency dialysis.Because of concerns about the efficacy and safety ofcation exchange resins, alternative agents to enhanceK� elimination in the colon have been sought for thispopulation. Because K� secretion in the distal colon isregulated by the mineralocorticoid receptor, mineralo-corticoids might be expected to be effective in man-aging hyperkalemia. In uncontrolled studies, the syn-thetic mineralocorticoid fludrocortisone has beenreported to lower the plasma K� concentration inpatients on hemodialysis. However, not all studieshave confirmed the effectiveness of fludrocortisone,possibly because of differences in dosing. High dosesof fludrocortisone may be more effective, but becausethe drug is eliminated in the urine and has a higherglucocorticoid potency than cortisol, there is concernthat accumulation of fludrocortisone may cause ad-verse effects (49). In a 2-week preliminary proof-of-principle study, inhibition of 11�-hydroxysteroid de-hydrogenase (11�-HSD2), by glycyrrhetinic acid, the

active ingredient of licorice, was shown to decreasethe serum K� concentration in dialysis patients (49).The enzyme 11�-HSD2, which is present in mineralo-corticoid target tissues including colonic epithelialcells, converts endogenous cortisol into cortisone,which has a much lower affinity for the receptor;endogenous cortisol can activate the mineralocorticoidreceptor. The efficacy and safety of glycyrrhetinic acidwas subsequently studied in 20 patients who were onmaintenance hemodialysis in a prospective, placebo-controlled 6-month crossover study. Glycyrrhetinicacid reduced the frequency of severe hyperkalemia(�6 mmol/L) when compared with placebo (0.6 ver-sus 9.0%; P � 0.001), and within days, it produced asustained decline in predialysis K� concentrations,prompting the use of higher dialysate K� concentra-tions during treatment with glycyrrhetinic acid thanduring treatment with placebo (3.2 � 0.4 versus 2.8 �0.4 mmol/L) to avoid hypokalemia. During treatmentwith glycyrrhetinic acid, the ratio of plasma cortisol/cortisone increased in all patients and plasma aldoste-rone and the aldosterone/renin ratio decreased. Treat-ment with glycyrrhetinic acid was well tolerated andno differences in BP or interdialytic weight gain oc-curred, but there was a modest increase in liver en-zymes during treatment. The study indicates that gly-cyrrhetinic acid is effective as a K�-lowering agent inpatients whose ability to excrete K� in the urine islimited, but more data on pharmacokinetics and ontoxicity are needed before the drug can be adopted forlong-term use.

References1. Wang WH, Giebisch G: Regulation of potassium (K) handling in the

renal collecting duct. Pflugers Arch 458: 157–168, 20092. El Moghrabi S, Houillier P, Picard N, Sohet F, Wootla B, Bloch-

Faure M, Leviel F, Cheval L, Frische S, Meneton P, Eladari D,Chambrey R: Tissue kallikrein permits early renal adaptation topotassium load. Proc Natl Acad Sci U S A 107: 13526–13531, 2010

3. Bailey IR, Thurlow VR: Is suboptimal phlebotomy technique impact-ing on potassium results for primary care? Ann Clin Biochem 45:266–269, 2008

4. Seimiya M, Yoshida T, Sawabe Y, Sogawa K, Umemura H, Matsu-shita K, Nomura F: Reducing the incidence of pseudohyperkalemiaby avoiding making a fist during phlebotomy: A quality improvementreport. Am J Kidney Dis 56: 686- 692, 2010

5. Han SW, Park JI, Yi JH, Park IK, Kim HJ: Fictitious pseudohyper-kalemia due to the dilution of plasma by anticoagulant. Clin Nephrol72: 156–157, 2009

6. Kim A, Biteman B, Malik UF, Siddique S, Martin MR, Ali SA,Maboud N, Raja S, Zachry A, Mahmoud A: A case of pseudohyper-kalemia in a patient presenting with leucocytosis and high potassiumlevel: A case report. Cases J 3: 73, 2010

7. Ruddy KJ, Wu D, Brown JR: Pseudohyperkalemia in chronic lym-phocytic leukemia. J Clin Oncol 26: 2781–2782, 2008


8. Ong YL, Deore R, El-Agnaf M: Pseudohyperkalaemia is a commonfinding in myeloproliferative disorders that may lead to inappropriatemanagement of patients. Int J Lab Hematol 32: e151–e157, 2010

9. Sevastos N, Theodossiades G, Archimandritis AJ: Pseudohyperkal-emia in serum: A new insight into an old phenomenon. Clin Med Res6: 30–32, 2008

10. Lee HK, Brough TJ, Curtis MB, Polito FA, Yeo KT: Pseudohyper-kalemia: Is serum or whole blood a better specimen type thanplasma? Clin Chim Acta 396: 95–96, 2008

11. Chawla NR, Shapiro J, Sham RL: Pneumatic tube “pseudo tumorlysis syndrome” in chronic lymphocytic leukemia. Am J Hematol 84:613–614, 2009

12. Palmer BF: A physiologic-based approach to the evaluation of apatient with hyperkalemia. Am J Kidney Dis 56: 387–393, 2010

13. Gupta A, Rohrscheib M, Tzamaloukas AH: Extreme hyperglycemiawith ketoacidosis and hyperkalemia in a patient on chronic hemodi-alysis. Hemodial Int 12[Suppl 2]: S43–S47, 2008

14. Adabala M, Jhaveri KD, Gitman M: Severe hyperkalaemia resultingfrom octreotide use in a haemodialysis patient. Nephrol Dial Trans-plant 25: 3439–3442, 2010

15. Gronert GA: Succinylcholine-induced hyperkalemia and beyond.1975. Anesthesiology 111: 1372–1377, 2009

16. Ahmed I, Chilimuri SS: Fatal dysrhythmia following potassiumreplacement for hypokalemic periodic paralysis. West J Emerg Med11: 57–59, 2010

17. Wiesel O, Szold O, Bentov I, Sorkin P, Nimrod A, Biderman P:Dyskalemia following head trauma: Case report and review of theliterature. J Trauma 67: E149–E151, 2009

18. Bandara V, Weinstein SA, White J, Eddleston M: A review of thenatural history, toxinology, diagnosis and clinical management ofNerium oleander (common oleander) and Thevetia peruviana (yellowoleander) poisoning. Toxicon 56: 273–281, 2010

19. Rajapakse S: Management of yellow oleander poisoning. Clin Toxi-col (Phila) 47: 206–212, 2009

20. Choi MJ, Ziyadeh FN: The utility of the transtubular potassiumgradient in the evaluation of hyperkalemia. J Am Soc Nephrol 19:424–426, 2008

21. Chakera AJ, Vaidya B: Addison disease in adults: Diagnosis andmanagement. Am J Med 123: 409–413, 2010

22. Nagler M, Muller B, Briner V, Winterhalder R: Severe hyperkalemiaand bilateral adrenal metastasis. J Oncol 2009: 831979, 2009

23. Huang WT, Chau T, Wu ST, Lin SH: Prolonged hyperkalemiafollowing unilateral adrenalectomy for primary hyperaldosteronism.Clin Nephrol 73: 392–397, 2010

24. Mayan H, Melnikov S, Novikov I, Holtzman EJ, Farfel Z: Familialhyperkalemia and hypertension: Pathogenetic insights based on lith-ium clearance. J Clin Endocrinol Metab 94: 3010–3016, 2009

25. Weir MR, Rolfe M: Potassium homeostasis and renin-angiotensin-aldosterone system inhibitors. Clin J Am Soc Nephrol 5: 531–548,2010

26. Weinberg JM, Appel LJ, Bakris G, Gassman JJ, Greene T, KendrickCA, Wang X, Lash J, Lewis JA, Pogue V, Thornley-Brown D,Phillips RA, African American Study of Hypertension and KidneyDisease Collaborative Research Group: Risk of hyperkalemia innondiabetic patients with chronic kidney disease receiving antihyper-tensive therapy. Arch Intern Med 169: 1587–1594, 2009

27. Wei L, Struthers AD, Fahey T, Watson AD, Macdonald TM: Spi-ronolactone use and renal toxicity: Population based longitudinalanalysis. BMJ 340: c1768, 2010

28. Sowers JR, Whaley-Connell A, Epstein M: Narrative review: Theemerging clinical implications of the role of aldosterone in themetabolic syndrome and resistant hypertension. Ann Intern Med 150:776–783, 2009

29. Bogaert YE, Linas S: The role of obesity in the pathogenesis ofhypertension. Nat Clin Pract Nephrol 5: 101–111, 2009

30. de Souza F, Muxfeldt E, Fiszman R, Salles G: Efficacy of spirono-

lactone therapy in patients with true resistant hypertension. Hyper-tension 55: 147–152, 2010

31. Khosla N, Kalaitzidis R, Bakris GL: Predictors of hyperkalemia riskfollowing hypertension control with aldosterone blockade. Am JNephrol 30: 418–424, 2009

32. Loughlin J, Seeger JD, Eng PM, Foegh M, Clifford CR, Cutone J,Walker AM: Risk of hyperkalemia in women taking ethinylestradiol/drospirenone and other oral contraceptives. Contraception 78: 377–383, 2008

33. Mohan S, Jaitly M, Pogue VA, Cheng JT: Influence of concomitantprednisolone on trimethoprim-associated hyperkalaemia. J Antimi-crob Chemother 64: 850–852, 2009

34. Antoniou T, Gomes T, Juurlink DN, Loutfy MR, Glazier RH,Mamdani MM: Trimethoprim-sulfamethoxazole-induced hyperkale-mia in patients receiving inhibitors of the renin-angiotensin system: Apopulation-based study. Arch Intern Med 170: 1045–1049, 2010

35. Wang CC, Wu CC, Shiang JC, Tsai MK, Chen IH: Acute paralysis ina uremic patient. Am J Med 123: e7–e8, 2010

36. Cornelius BG, Cornelius A, Desai B: Identification of sine wave inearly suspicion of hyperkalemia. West J Emerg Med 11: 94, 2010

37. Montague BT, Ouellette JR, Buller GK: Retrospective review of thefrequency of ECG changes in hyperkalemia. Clin J Am Soc Nephrol3: 324–330, 2008

38. Korgaonkar S, Tilea A, Gillespie BW, Kiser M, Eisele G, FinkelsteinF, Kotanko P, Pitt B, Saran R: Serum potassium and outcomes inCKD: Insights from the RRI-CKD cohort study. Clin J Am SocNephrol 5: 762–769, 2010

39. Karet FE: Mechanisms in hyperkalemic renal tubular acidosis. J AmSoc Nephrol 20: 251–254, 2009

40. Nyirenda MJ, Tang JI, Padfield PL, Seckl JR: Hyperkalaemia. BMJ339: b4114, 2009

41. Kamel KS, Wei C: Controversial issues in the treatment of hyperka-laemia. Nephrol Dial Transplant 18: 2215–2218, 2003

42. Sterns RH, Rojas M, Bernstein P, Chennupati S: Ion-exchange resinsfor the treatment of hyperkalemia: Are they safe and effective? J AmSoc Nephrol 21: 733–735, 2010

43. McGowan CE, Saha S, Chu G, Resnick MB, Moss SF: Intestinalnecrosis due to sodium polystyrene sulfonate (Kayexalate) in sorbi-tol. South Med J 102: 493–497, 2009

44. Thomas A, James BR, Landsberg D: Colonic necrosis due to oralkayexalate in a critically-ill patient. Am J Med Sci 337: 305–306,2009

45. Bomback AS, Woosley JT, Kshirsagar AV: Colonic necrosis due tosodium polystyrene sulfate (Kayexalate). Am J Emerg Med 27: 753e1–e2, 2009

46. Trottier V, Drolet S, Morcos MW: Ileocolic perforation secondary tosodium polystyrene sulfonate in sorbitol use: A case report. Can JGastroenterol 23: 689–690, 2009

47. Joo M, Bae WK, Kim NH, Han SR: Colonic mucosal necrosisfollowing administration of calcium polystyrene sulfonate (Kalimate)in a uremic patient. J Korean Med Sci 24: 1207–1211, 2009

48. Joshi P, Beaullieu J, Shemin D: The effect of a single dose of sodiumpolystyrene sulfonate (SPS) and sorbitol in hyperkalemic patientswith kidney disease [Abstract]. J Am Soc Nephrol 19: 355A, 2008

49. Farese S, Kruse A, Pasch A, Dick B, Frey BM, Uehlinger DE, FreyFJ: Glycyrrhetinic acid food supplementation lowers serum potas-sium concentration in chronic hemodialysis patients. Kidney Int 76:877–884, 2009

Nonhypotonic Hyponatremia

Pseudohyponatremia“True hyponatremia,” or “hypotonic hyponatre-

mia,” is characterized by a decreased concentration of


sodium in the aqueous phase of plasma. Plasma isnormally 93% water and 7% proteins and lipids. Hy-perproteinemia or hyperlipidemia decreases the aque-ous fraction of the plasma sample and causes anartifactually low plasma sodium concentration despitea normal sodium concentration in plasma water—aphenomenon called “pseudohyponatremia.” Autoana-lyzers for routine chemical analysis dilute the plasmasample before the actual measurement is obtained;correction for the dilution factor (which assumes thatplasma water equals 93% of the total sample volume)results in an artificially low result when the plasmawater content is actually �93%. Instruments for mea-suring arterial blood gases use direct ion-specific elec-trodes (ISE) without any dilution and measure theactivity of sodium in the water phase only. Thus,pseudohyponatremia does not occur when the mea-surement is made with direct ISE (sometimes called“direct potentiometry”). Although many autoanalyzersuse ISE to measure the sodium concentration, thedilution step (sometimes called “indirect potentiom-etry”) allows pseudohyponatremia to occur. Thesedifferences in methodology should be taken into ac-count to explain discrepancies between results ob-tained with classical biochemistry analyzers and withblood gas apparatus, and they can be exploited inevaluating patients in whom pseudohyponatremia issuspected.

Hyperlipemia that causes hyponatremia is usu-ally due to the presence of triglyceride-rich chylomi-crons, which cause specimen turbidity. When lipemiais observed, the laboratory can avoid spurious resultsby either measuring the sodium by a direct potentio-metric method or by ultracentrifuging the specimenbefore using an indirect sodium method. However,specimen turbidity, even when objectively determined,has a poor correlation with actual triglyceride levelsand the subsequent potential for various interferences.It is commonly believed, incorrectly, that pseudohy-ponatremia that is attributable to hyperlipemia is seenonly in specimens that show obvious lactescence.Unlike hypertriglyceridemia, high cholesterol levelsdo not cause the blood to be visibly lipemic, and, insome cases, hypercholesterolemia can be associatedwith pseudohyponatremia. High concentrations of li-poprotein X (Lp-X), found in patients with intrahe-patic and extrahepatic cholestasis and in the plasma ofpatients with lecithin cholesterol acyltransferase(LCAT) deficiency, cause pseudohyponatremia with

little or no specimen turbidity (1). Lp-X is an abnor-mal lipoprotein that has a density in the LDL fraction.In cholestasis, it is thought to form from reflux ofunesterified cholesterol and phospholipids into thecirculation from cholestatic bile ductules. In LCATdeficiency, a condition in which cholestasis is not afeature, Lp-X may form in the plasma from the accu-mulation of phospholipids and free cholesterol. UnlikeLDL, HDL, and VLDL, the particles of Lp-X choles-terol are not soluble in plasma water and thus increasethe solid fraction of plasma (and decrease its watercontent). Levels of Lp-X can be as high as severalthousand milligrams per deciliter of cholesterol.

HyperglycemiaHyperglycemia is the most common cause of

nonhypotonic hyponatremia. In patients with diabetes,accumulation of glucose in the extracellular spacedraws water out of cells, lowering the serum sodiumconcentration. On theoretical grounds, Katz (2) pre-dicted that serum sodium concentration would be ex-pected to decrease by 1.6 mmol/L for every 100 mg/dlin blood glucose. Similarly, as blood sugar decreases,serum sodium would be expected to increase by 1.6mmol/L for every 100-mg/dl fall in blood glucose. Onthe basis of data from acute somatatostatin-inducedhyperglycemia in six healthy subjects, Hillier et al. (3)concluded that the mean decrease in serum sodiumconcentration averaged 2.5 mmol/L for every 100-mg/dl increase in glucose concentration and that theassociation between sodium and glucose concentra-tions was nonlinear; up to 400 mg/dl, Katz’s correc-tion factor of 1.6 worked well, but if the glucoseconcentration was �400 mg/dl, then a correction fac-tor of 4.0 was better. However, neither of these pre-dictions can be applied to patients with intact renalfunction because serum sodium concentration in-creases during therapy for two reasons: (1) Watershifts back from the extracellular space to cells as theplasma glucose concentration falls, and (2) glucoseacts as an osmotic diuretic eliminating electrolyte-freewater in the urine, which also raises the serum sodiumconcentration. In patients with oligoanuria and ESRD,hyperglycemia is corrected by metabolism rather thanexcretion. A review of six published cases of dialysis-associated hyperglycemia found that the averagechange in serum sodium concentration during correc-tion of hyperglycemia was almost identical to Katz’spredicted correction factor of �1.6 mmol/L per 100


mg/dl; however, the �[Na]/�[glucose] values in indi-vidual studies ranged between �2.48 and �1.44 (4).The data did not support Hillier’s observation that ahigher correction factor is necessary when the bloodglucose concentration is �400 mg/dl. Edematousstates would be expected to blunt the effect of glucoseon the serum sodium concentration because the extra-cellular volume is larger than that assumed in thecalculation. The mean value of �[Na]/�[glucose] was�1.04 � 0.23 mmol/L per 100 mg/dl in four dialysispatients with severe edema and was �1.65 � 0.08mmol/L per 100 mg/dl in five patients who were closeto their dry weight (4).

A review of six published cases of dialysis-associated hyperglycemia found that theaverage change in serum sodium concen-tration during correction of hyperglycemiawas almost identical to Katz’s predictedcorrection factor of �1.6 mmol/L per 100mg/dl.

Exogenous SolutesIntravenous Mannitol. Similar to hyperglycemia,infusion of hypertonic mannitol draws water out ofcells, transiently lowering the extracellular sodiumconcentration. Mannitol acts as an osmotic diuretic,and its eventual excretion in the urine results in freewater losses that raise the serum sodium concentrationcausing hypernatremia when water losses are not re-placed. However, high doses of mannitol can causerenal tubular injury and acute renal failure. In thiscase, accumulation of mannitol results in sustainedhypertonic hyponatremia that can be distinguishedfrom true or hypo-osmotic hyponatremia by demon-strating the presence of an osmolar gap (5), a discrep-ancy between the measured plasma osmolality and theosmolality calculated by multiplying the serum so-dium by 2 (to include accompanying anions) andadding the osmotic contributions of glucose and ureain mmol/L (values reported in mg/dl can be convertedto mmol/L by dividing the glucose concentration by 18and the blood urea nitrogen by 2.8).Irrigant Absorption. Many endoscopic surgicalprocedures require an irrigating fluid to dilate theoperating field and to wash away debris and blood.Systemic absorption of the irrigating fluid can lower

the serum sodium concentration by expanding theextracellular space with sodium-free irrigant. Systemicabsorption of isosmotic or hypo-osmotic irrigants dur-ing endoscopic prostate, bladder, or intrauterine sur-gery lowers the plasma sodium concentration withlittle or no change in plasma osmolality. In contrast tohyponatremia that is caused by hyperglycemia or hy-pertonic mannitol infusion, no shift of water from cellsoccurs; rather, serum sodium concentration falls be-cause the extracellular volume is expanded with fluidcontaining nonelectrolyte solutes: Glycine, mannitol,sorbitol, or glucose. Immediately upon absorption,regardless of the solution used, serum sodium concen-tration falls to very low levels because the solute isconfined to the extracellular fluid, but plasma osmo-lality remains constant or, if a hypo-osmotic solutionis used, falls slightly. Shortly after surgery, regardlessof the solution, serum sodium concentration begins toincrease but for different reasons depending on whichsolute is used. If isosmotic (and isotonic) mannitol isabsorbed, then the absorbed solute remains confined tothe extracellular fluid and the plasma osmolality isunaltered; the serum sodium concentration increasesas mannitol, an osmotic diuretic, is excreted in theurine with water. Changes in serum sodium concen-tration after absorption of isosmotic (5%) dextrose orsorbitol are more complex than those produced bymannitol. Serum sodium concentration increases asthese sugars, also osmotic diuretics, are excreted in theurine, but, in addition, some of the absorbed solute ismetabolized to carbon dioxide and water, allowing theplasma osmolality to fall over time. Some of thepostabsorptive increase in serum sodium occurs aswater leaves the extracellular fluid and enters cells.After initially distributing in the extracellular space,absorbed glycine, an amino acid, is able to entermuscle cells, which increases serum sodium above itsnadir level. However, the absorbed glycine, along withserine and other osmotically active metabolites, per-sists in both the circulation and intracellular fluid formany hours. The amino acids are metabolized gradu-ally to urea, another solute that is measured by theosmometer; therefore, the plasma osmolality remainsstable as the serum sodium concentration rises toreflect distribution of the absorbed solution in totalbody water. Only a few studies have reported bothserum sodium concentration and osmolarity after irri-gant absorption. In a series of 72 patients who under-went transurethral resection of the prostate (TURP),


serum sodium concentrations decreased in 26% of thepatients, with a decrement ranging from 10 to 54mmol/L, whereas plasma osmolarity changed in onlytwo (3%) cases (6).

Several irrigating fluids are available commer-cially, and the choice among them tends to be gov-erned by price, stickiness and transparency, personalpreference, and tradition. Solutions of glycine, anendogenous amino acid that is transparent and reason-ably inexpensive, have become the most popular, butsystemic absorption of the solution commonly resultsin nausea, vomiting, confusion, arterial hypotension,and, more rarely, coma and death. Hyperammonemicencephalopathy may develop because ammonia is anintermediate product in glycine metabolism. Labora-tory studies of animals showed that glycine has directand indirect cardiotoxic effects, and high plasma lev-els of glycine have been associated with echocardio-graphic abnormalities and increased troponin levels(7). More physiologic alternatives to glycine havebeen used, but limited data demonstrating their supe-riority are available. A single-center, randomizedstudy of 360 patients who underwent TURP for benignhypertrophy randomly allocated patients to use glycine1.5% solution (n � 120), glucose 5% (n � 120), ornormal saline 0.9% solution (n � 120) as irrigatingfluid during and immediately after surgery and com-pared perioperative morbidity associated with the so-lutions (8). Patients who irrigated with glycine orglucose had lower postoperative serum sodium con-centrations than those who irrigated with saline, butonly those in the glycine group experienced symptomsor signs of TURP syndrome such as nausea, vomiting,bradycardia, hypotension, chest pain, mental confu-sion, anxiety, parasthesia, and visual disturbance;symptoms of the TURP syndrome developed in 17glycine-irrigated patients (14%), and these patientshad the highest postoperative plasma glycine concen-trations. Six patients in the glycine group developedischemic electrocardiographic changes, and three de-veloped elevated troponin I.

As an isotonic electrolyte medium, normal salineis the most physiologic irrigant for TURP, but itselectrical conducting properties prohibit its use withconventional monopolar cautery. The development ofbipolar resection systems now permits the use ofnormal saline as an irrigant. Use of bipolar cautery hasbeen reported to be associated with less collateral andpenetrative tissue damage, lower incidence of TURP

syndrome, shorter catheter indwelling times, and ear-lier hospital discharge. New bipolar resection systemspermit normal saline to be used as irrigant duringendoscopic surgical procedures. A single-center, ran-domized, controlled trial compared monopolar cauteryusing 1.5% glycine as an irrigant (n � 30) with bipolarcautery using 0.9% saline as irrigant (n � 30). Serumsodium concentration fell by 1.3 mmol/L in the salinebipolar group and by 4.12 mmol/L in the glycinemonopolar group, but these differences did not reachstatistical significance in this small study.

References1. Klinke JA, Shapira SC, Akbari E, Holmes DT: Quetiapine-associated

cholestasis causing lipoprotein-X and pseudohyponatraemia. J ClinPathol 63: 741–743, 2010

2. Katz MA: Hyperglycemia-induced hyponatremia: Calculation of ex-pected serum sodium depression. N Engl J Med 289: 843–844, 1973

3. Hillier TA, Abbott RD, Barrett EJ: Hyponatremia: Evaluating thecorrection factor for hyperglycemia. Am J Med 106: 399–403, 1999

4. Tzamaloukas AH, Ing TS, Siamopoulos KC, Rohrscheib M, ElisafMS, Raj DS, Murata GH: Body fluid abnormalities in severe hyper-glycemia in patients on chronic dialysis: review of published reports.J Diabetes Complications 22: 29–37, 2008

5. Tsai SF, Shu KH: Mannitol-induced acute renal failure. Clin Nephrol74: 70–73, 2010

6. Gravenstein D: Transurethral resection of the prostate (TURP) syn-drome: A review of the pathophysiology and management. AnesthAnalg 84: 438–446, 1997

7. Collins JW, Macdermott S, Bradbrook RA, Drake B, Keeley FX,Timoney AG: The effect of the choice of irrigation fluid on cardiacstress during transurethral resection of the prostate: A comparisonbetween 1.5% glycine and 5% glucose. J Urol 177: 1369–1373, 2007

8. Yousef AA, Suliman GA, Elashry OM, Elsharaby MD, Elgamasy AelN: A randomized comparison between three types of irrigating fluidsduring transurethral resection in benign prostatic hyperplasia. BMCAnesthesiol 10: 7, 2010

Hypotonic Hyponatremia: PathophysiologyIn 1959, Edelman described a relationship among

the plasma sodium (Na) concentration (PNa), isotopi-cally measured exchangeable sodium (Nae), exchange-able potassium (Ke), and total body water (TBW) in 98heterogeneous patients with stable PNas:

PNa � � (Nae � Ke)/TBW � �Edelman’s equation has been applied to clinical

situations in which the PNa is changing, assuming thatchanges in the PNa can be predicted or explained byexternal balances of Na, K, and water (1). However,although this “tonicity balance” calculation hasworked well in individual cases (2–4), its assumptionshave not been rigorously tested. A study of acutehyponatremia in pigs (with plasma Na reduced from136 to 120 mmol/L over 480 minutes) found thatplasma Na values calculated from external balances of


cations and water on the basis of Edelman’s equationfit the observed plasma Na values extremely well (r �0.98) with a significantly better fit than Na valuescalculated from water retention alone (Figure 10) (5).The calculated values for the slope of the equation (�)and its intercept (�) in pigs were quite close to thosederived from Edelman’s original data.

The findings of this experiment have importantimplications. Individual cells and some tissues(most notably the brain) are known to volume reg-ulate when subjected to a hypotonic environment;for example, minimizing cell swelling by extrudingK and chloride within minutes and organic os-molytes over several hours (6). If muscles, contain-ing 80% of body water in pigs, were to respondrapidly to acute water retention with a volumeregulatory response, then the decrease in PNa wouldbe much greater (more of the retained water wouldbe confined to extracellular space) and hyperkale-mia would result from the efflux of K from musclecells. External K losses did not exceed controllevels during the experiment, and the plasma K did

not increase, suggesting that there is no globalvolume regulatory response to acute water intoxica-tion. Furthermore, muscle water content, as mea-sured in vivo by magnetic resonance imaging andvalidated by direct measurements of tissue watercontent, increased as would be expected for perfectosmotic behavior, and there was no change in tissueK content (7).

The observation that changes in plasma Na aredetermined by external balances of Na, K, and water isconsistent with our classic understandings of Na andwater balance. In this traditional model, Na and itsaccompanying anions are the principal extracellulareffective osmoles, and K salts account for almost all ofthe intracellular effective osmoles. These ions act aseffective osmoles maintaining the volume of the ex-tracellular and intracellular spaces because they re-main osmotically active regardless of external bal-ances and because they are restricted to theirrespective compartments by the activity of the Na�-K�-ATPase pump in the cell membrane. The PNa isthus considered to be an accurate measure of theconcentration of Na throughout body fluid becauseconcentration gradients cannot exist between the in-travascular, interstitial space, and the cellular fluidcompartments beyond the very small differences ex-plained by the Gibbs-Donnan equilibrium. However,several observations in humans have challenged the tra-ditional model and have increased interest in the possi-bility that pools of osmotically inactive Na are able tostore or release Na: (1) Balance studies have shown thatlarge amounts of Na can accumulate on a high-salt dietwithout accompanying water retention; (2) marked dis-crepancies have been found between changes inwater and Na balance during weight loss induced bydieting; and (3) discrepancies between changes inbody weight and changes in plasma Na have beenreported after intense exercise (8,9). Mixed connec-tive tissues are known to contain large amounts ofNa, and in contrast to most animal cells, chondro-cytes are surrounded by an extracellular fluid withan ionic composition that is altered by the negativecharge density of glycosaminoglycans in the extra-cellular matrix. Glycosaminoglycans (GAGs) arenegatively charged polyanions that attract cationsand repel anions, resulting in extracellular cartilageNa concentration of 250 to 350 mmol/L and extra-cellular osmolality of 350 to 450 mmol/L. Thisosmotic strength maintains a high water content in

Figure 10. Plasma Na concentrations predicted by the Edel-man equation agree well with measured plasma Na values(solid circles represent pigs with acute hyponatremia andsolid circles are normonatremic controls). Copyright 2011by American Physiological Society. Reproduced from ref-erence 5 (Overgaard-Steensen C, Larsson A, Bluhme H,Tonnesen E, Frokiaer J, Ring T: Edelman’s equation is validin acute hyponatremia in a porcine model: Plasma sodiumconcentration is determined by external balances of waterand cations. Am J Physiol Regul Integr Comp Physiol 298:R120–R129, 2010), with permission of American Physio-logical Society.


cartilage, which provides its biomechanical proper-ties. Conversely, the high electrolyte concentrationsof cartilage indicate “osmotically inactive” Na stor-age. Studies in rats subjected to very-high- or very-low-salt diets for several weeks showed that skin behaves ina similar manner acting as an adaptable pool for osmot-ically inactive Na storage (8). Increases in the polyan-ionic character of the extracellular matrix on a high-saltdiet were accompanied by Na storage in skin, whereasreduced GAG polymerization and GAG sulfation and asubsequent reduction in the polyanionic character of theextracellular matrix were associated with the mobiliza-tion of Na from the skin reservoir with dietary saltscarcity. The previously mentioned pig studies suggestedthat, at least in the short term, osmotically inactive Nastorage pools do not alter the traditionally expectedrelationships between Na, K, and water.

Studies of rats that were fed very-high- orvery-low-salt diets for several weeksshowed that skin acts as an adaptablepool for osmotically inactive Na storage.Increases in the polyanionic character ofthe extracellular matrix on a high-salt dietwere accompanied by Na storage in skin,whereas reduced GAG polymerization andsulfation and a subsequent reduction inthe polyanionic character of the extracel-lular matrix were associated with the mo-bilization of Na from the skin reservoir withdietary salt scarcity.

In contrast to skeletal muscle, the brain does notswell as much as would be expected for perfect os-motic behavior, even in acute hyponatremia; an exten-sive review of the response of the brain to hyponatre-mia was recently published (10). Previous studies haverelied on measurements of water and solute contents inthe whole brain. In the pig model described already,magnetic resonance imaging was used to measurebrain water content in vivo at 70-minute intervalsduring the induction of hyponatremia (plasma Na 123mmol/L) over 7 hours. These measurements were thenvalidated with direct measurements of water content asdetermined by tissue drying and deuterium dilution.Major regional differences in the brain were found.Surprisingly, brain water content in the region of the

cell-rich thalamus increased as would be expected forperfect osmotic behavior, swelling by 12%. In whitematter, brain water content increased by only 3%,possibly reflecting a regulatory volume decrease as aresult of extrusion of solutes or, alternatively, slowwater equilibrium in hydrophobic regions of the brain.Consistent with these findings, measurements of tissueNa content showed that brain Na content decreasedsignificantly, but, in contrast to rodents subjected tohyponatremia over a similar time course, brain Kcontent did not change. The decrease in Na contentcan be explained by a decrease in extracellular Na-richfluid as interstitial fluid, cerebrospinal fluid, and bloodare squeezed out of the skull (11).

In rats with hyponatremia, intracerebroventricu-lar administration of benzamil, a specific Na channelblocker, completely abolished brain swelling in re-sponse to acute hyponatremia (a decrease in plasmaNa from 146 to 110 mmol/L over 2 hours); conversely,intracerebroventricular administration of arginine va-sopressin (AVP) exacerbated osmotic brain swelling(12). There is some evidence that AVP stimulates Nauptake by the brain by activating V1B receptors,which are widely distributed in the central nervoussystem, and the authors speculated that an early cel-lular Na uptake through benzamil-sensitive activatedNa channels may play a role in brain swelling that iscaused by hyponatremia. However, the mechanism forthese intriguing observations is not entirely clear.Brain tissue Na content (likely reflecting extrusion ofextracellular fluid from the brain) decreased in re-sponse to hyponatremia to a similar degree in allgroups with hyponatremia, including those with exac-erbated brain swelling induced by AVP and those withameliorated brain swelling induced by benzamil.

Hyponatremia occurs when water intake exceedswater losses. Discussions of the pathogenesis of hy-ponatremia have focused on the regulation of waterexcretion, but we should ask why patients who areunable to eliminate water normally do not compensatefor that abnormality. Similar to how patients withdiabetes insipidus–related defective water conserva-tion avoid hypernatremia by drinking more, patientswho cannot eliminate water normally might avoidhyponatremia by decreasing water intake. The tran-sient receptor potential vanilloid 4 (TRPV4) channel,expressed in the osmosensing nuclei of the brain, isactivated by hypotonic stress. Targeted deletion of theTRPV4 gene causes aberrant osmoregulation in murine


models, causing them to drink more and to becomemore hyponatremic than wild-type mice when givenexogenous vasopressin. A polymorphism in the TRPV4gene resulting in a proline-to-serine substitution atresidue 19 (TRPV4P194) was found to be significantlyassociated with the serum Na concentration in twomale non-Hispanic Caucasian healthy aging cohorts(one cohort participating in a study on the geneticdeterminants of Alzheimer disease and one participat-ing in a study to assess the genetic determinants ofosteoporotic fractures) (13). Male participants with theallele were 2.30 to 6.45 times as likely to exhibithyponatremia as male participants without the allele. Itwas inferred that the presence of the TRPV4P194

allele may act synergistically with conditions ofvasopressin excess, making participants drink morethan those without the allele, thereby promotinghyponatremia. Consistent with this hypothesis, cellstransfected with the mutated TRPV4P194 channelexhibited diminished whole-cell cationic currentswhen exposed to mild hypotonic shock (correspond-ing to a 15% reduction in osmolality), as comparedwith cells transfected with the wild-type TRPV4gene.

References1. Nguyen MK: Quantitative approaches to the analysis and treatment of

the dysnatremias. Semin Nephrol 29: 216–226, 20092. Berl T, Rastegar A: A patient with severe hyponatremia and hypo-

kalemia: Osmotic demyelination following potassium repletion. Am JKidney Dis 55: 742–748, 2010

3. Sterns RH, Hix JK, Silver S: Treating profound hyponatremia: Astrategy for controlled correction. Am J Kidney Dis 56: 774–779,2010

4. Lindner G, Kneidinger N, Holzinger U, Druml W, Schwarz C:Tonicity balance in patients with hypernatremia acquired in theintensive care unit. Am J Kidney Dis 54: 674–679, 2009

5. Overgaard-Steensen C, Larsson A, Bluhme H, Tonnesen E, FrokiaerJ, Ring T: Edelman’s equation is valid in acute hyponatremia in aporcine model: Plasma sodium concentration is determined by exter-nal balances of water and cations. Am J Physiol Regul Integr CompPhysiol 298: R120–R129, 2010

6. Hoffmann EK, Lambert IH, Pedersen SF: Physiology of cell volumeregulation in vertebrates. Physiol Rev 89: 193–277, 2009

7. Overgaard-Steensen C, Stødkilde-Jørgensen H, Larsson A, Broch-Lips M, Tønnesen E, Frøkiaer J, Ring T: Regional differences inosmotic behavior in brain during acute hyponatremia: An in vivoMRI-study of brain and skeletal muscle in pigs. Am J Physiol RegulIntegr Comp Physiol 299: R521–R532, 2010

8. Titze J, Machnik A: Sodium sensing in the interstitium and relation-ship to hypertension. Curr Opin Nephrol Hypertens 19: 385–392,2010

9. Rosner MH: Exercise-associated hyponatremia. Semin Nephrol 29:271–281, 2009

10. Mount DB: The brain in hyponatremia: Both culprit and victim.Semin Nephrol 29: 196–215, 2009

11. Mount DB, Krahn TA: Hyponatremia: Case vignettes. Semin Nephrol29: 300–317, 2009

12. Sulyok E, Pal J, Vajda Z, Steier R, Doczi T: Benzamil prevents brainwater accumulation in hyponatraemic rats. Acta Neurochir (Wien)151: 1121–1125, 2009

13. Tian W, Fu Y, Garcia-Elias A, Fernandez-Fernandez JM, VicenteR, Kramer PL, Klein RF, Hitzemann R, Orwoll ES, Wilmot B,McWeeney S, Valverde MA, Cohen DM: A loss-of-function nonsyn-onymous polymorphism in the osmoregulatory TRPV4 gene is asso-ciated with human hyponatremia. Proc Natl Acad Sci U S A 106:14034–14039, 2009

Acute Hypotonic HyponatremiaSince the 1920s and 1930s, it has been under-

stood that “acute water intoxication” can cause fatalcerebral edema and that brain swelling and death canbe prevented by administering hypertonic saline.There should be no doubt that when the plasma so-dium concentration decreases more rapidly than thebrain can adapt to it, patients are at risk for seriousneurologic complications. Very rapid onset of hypo-natremia is most likely to occur in only a few settings:(1) Self-induced water intoxication, in psychotic pa-tients with marked polydipsia; (2) excess water con-sumption by marathon runners; (3) use of the illicitdrug “ecstasy” (3,4-methylenedioxymethamphetamine),which provokes both thirst and vasopressin release;and (4) excessive hypotonic fluid administration topatients with an impaired ability to excrete an acutewater load, most commonly in the postoperative pe-riod (1). However, there are occasional reports ofacute water intoxication in other settings. For example,a 34-year-old man developed seizures, requiring me-chanical ventilation, associated with a serum sodiumof 123 mmol/L after drinking approximately 40glasses of water (approximately 8 L) over a few hours,during a domino game that required the loser of eachmatch to drink a full glass of water (2). Two previouslyhealthy young women developed acute hyponatremicencephalopathy (serum sodium 122 and 126 mmol/L)after drinking 3 to 6 L of water in preparation for a pelvicultrasound (in response to advice to drink “as much aspossible” because “the more you drink, the better the testresults”); both survived without sequelae (3).

Brain edema has seldom been documented be-fore death in patients with acute hyponatremia. In arecent report, an elderly woman developed acutesymptomatic hyponatremia (vomiting, stupor, Babin-ski signs and mydriasis of one pupil) after experienc-ing a decrease in plasma sodium from 133 to 109mmol/L over 26 hours in the hospital. A computedtomography (CT) scan showed near absence of cere-


bral sulci, compression of the ventricles and sylvianfissures, and decreased cerebrospinal fluid in the basalcisterns and the cerebral sulci (4). Despite a 4-hourdelay, during which her plasma sodium fell to 108mmol/L, she survived the experience without neuro-logic sequelae. A repeat CT scan obtained 36 hourslater, when plasma sodium was 120 mmol/L, showedimprovement in cerebral edema; however, there wasno improvement in a scan done at 12 hours, when theplasma sodium was 116 mmol/L. In the two youngwomen who developed hyponatremia preparing fortheir pelvic ultrasounds, CT scans 2 and 8 hours aftersymptom onset showed diffuse sulcal effacement andcompression of lateral ventricles consistent with brainedema. Magnetic resonance imaging (MRI) obtained10 and 14 hours after onset revealed cortical sulcalnarrowing and restricted diffusion; findings suggesteddiffuse cerebral cortical cytotoxic edema and blood-brain barrier breakdown, although a repeat scan in oneof these cases showed the findings to be reversible (3).The authors noted that with T1- and T2-weightedMRI modalities, detection of mild to moderate ce-rebral edema is difficult because the mass effect isfaint and can be overlooked. Diffusion-weightedimages show easily noticeable findings analogous tothe cerebral edema found in ischemic stroke. It isnot clear whether the evidence of blood-brain bar-rier breakdown observed in these MR images reflectthe acute cerebral edema or rapid correction ofhyponatremia.

Psychotic PolydipsiaPolydipsia is very common among chronic psychi-

atric patients, although the actual prevalence of the dis-order is still unclear. Objective measures to identifypatients, such as urine specific gravity and diurnal weightgains, have not been applied in a consistent manner. Astudy of psychiatric inpatients from Cuba used the samecriteria that were used in two previously published stud-ies from the United States and Spain (5). Urine specificgravity �1.009 in an afternoon urine sample was used todefine polydipsia, and a 4% diurnal weight gain wastaken to identify the population at serious risk for waterintoxication. In the Cuban patients, the risk for primarypolydipsia and water intoxication were 47% (n � 189)and 6.7% (n � 27), respectively. These results are similarto those reported for patients in the United States (34.0and 3.7%) and Spain (25 and 7%) for the total populationof psychiatric inpatients. Pooling findings from studies in

the United States, Spain, and Cuba using the samecriteria showed that the prevalence of polydipsia is twiceas high in schizophrenic inpatients as in institutionalizedpsychiatric patients with other diagnoses (41.5 versus25.1%, respectively; odd ratio 2.12; P � 0.001).

The anterior hippocampal volume is smaller inpatients with schizophrenia and a history of hyponatre-mia compared with similar patients with schizophreniaand polydipsia but without a history of hyponatremia,patients without polydipsia, and healthy control subjects(6,7). There is evidence that this brain region normallyconstrains hypothalamic-pituitary-adrenal and vasopres-sin responses to psychological stressors, and some au-thors have hypothesized that abnormalities of vasopres-sin release are central to pathogenesis of waterintoxication. However, pathologic antidiuresis is not aprerequisite for self-induced water intoxication becausepatients with severe polydipsia have been observed todevelop hyponatremia despite excretion of maximallydilute urine. Thus, it is unclear whether the reset osmostatthat has been found in the recovery phase of acute waterintoxication represents the cause or consequence of waterretention. Similarly, cognitive defects have been shownin patients with schizophrenia and a history of waterintoxication; although these findings could reflect pri-mary pathology within the anterior hippocampus andassociated prefrontal/limbic brain regions, they couldalso represent neurologic morbidity from repeated epi-sodes of water intoxication (8).

Hyponatremia is sometimes attributed to the use ofantipsychotic drugs, but a cause-and-effect relationship isnot as well established as that of antidepressants, partic-ularly selective serotonin reuptake inhibitors (SSRIs).Using a standard drug reaction probability rating system,Meulendijks et al. (9) reviewed all published reports ofhyponatremia attributed to antipsychotic drugs and iden-tified 91 published case reports and case series describing120 patients with one or more hyponatremic events; 19%of these were rated as showing a “probable” relationship,and in 7% of cases, hyponatremia recurred on rechal-lenge with the drug. In the majority of cases, it was notpossible to exclude reliably other possible causes forhyponatremia including psychosis itself.

Exercise HyponatremiaExercise-associated hyponatremia (EAH) has

emerged as a serious problem among endurance athletes,affecting as many as 7% of participants. Hyponatremiadevelops primarily because of excess water or hypotonic


fluid intake, causing athletes to gain weight despite lossof tissue glycogen. Ingested water is retained because ofthe nonosmotic release of vasopressin, and hyponatremiais augmented by sweat losses of sodium. Since 1993, atleast four otherwise healthy women and one male runnerhave died from EAH after following advice to consumeas much fluid as possible during exercise; the subject wasdiscussed extensively in previous fluid and electrolyteissues of NephSAP, and a review was published recently(10). On-site whole-blood sodium testing was introducedat the finish-line tent of the Boston marathon beginningin 2001 to provide diagnosis and to facilitate triage forappropriate treatment (11). Hypernatremia was identifiedin 27.7% of collapsed runners and was treated with 0.9%saline when runners were unable to drink fluids. Hypo-natremia was found in 63 (4.8%) runners, including 26(2.0%) with values �130 mmol/L; those with seizures orcoma received intravenous 3% saline, and 16 runnerswho were able to drink were given a concentrated oralsolution composed of 4 bouillon cubes in 4 oz of water.The treatment increased the serum sodium by 6 to 7mmol/L within 16 to 26 minutes in three patients withdelirium, resulting in complete resolution of symptoms.

Obstetric HyponatremiaRare cases of water intoxication have been reported

during labor. In a recent case report, a 37-year-old privi-gravida with a history of habitual water drinking devel-oped multiple seizures associated with a serum sodium of111 mmol/L (12); her newborn had a serum sodium of108 mmol/L. The patient’s serum sodium was increasedto 123 mmol/L on the second hospital day, and she wastreated with 3% saline but likely autocorrected becauseurine osolalities of 67 and 80 were documented in thefirst 2 days. Infusion of oxytocin, which acts as anantidiuretic hormone, in the peripartum period mayresult in hyponatremia if the drug is administered inhypotonic fluids. In a recent case report, a youngwoman who received oxytocin in 5% dextrose waterexperienced seizures 24 hours postpartum when herserum sodium concentration fell to 113 mmol/L; sherecovered consciousness 12 hours later, when herserum sodium had increased to 119 mmol/L (13).

Postoperative HyponatremiaThe first reported case of fatal cerebral edema

caused by water intoxication occurred in a postop-erative patient. Surgery almost invariably results inthe nonosmotic release of vasopressin, and admin-

istration of hypotonic fluid in the postoperativeperiod predictably causes hyponatremia. Manydeaths have been recorded, mostly in young womenand children (1,14). Children exhibit more symp-toms than adults in response to abnormal sodiumlevels because there is less room for brain cells toswell (the brain reaches its adult size by the time thechild is 6 years old, but the skull does not reachadult size until a person is 16 years old). Fourpediatric deaths from cerebral edema with brainherniation as a result of acute hyponatremia associ-ated with intravenous administration of hypotonicsolutions—three in a postsurgical setting, one in amedical setting—were voluntarily reported (two tothe Institute for Safe Medication Practices Canadaand two to the Institute for Safe Medication Prac-tices in the United States) (15). Details of theclinical course of these patients is provided in apublished report: (1) A 4-year-old who was given3.3% dextrose and 0.3% sodium chloride solution(“2⁄3 and 1⁄3,” containing 50 mmol/L NaCl) at 55ml/h and 200 ml of oral fluids after a routinetonsillectomy developed vomiting and seizures;when the serum sodium was found to be “less than120 mmol/L,” 3% saline was begun, but the patientdied of herniation; and (2) a previously healthy3-year-old who was admitted for vomiting and di-arrhea and was given “2⁄3 and 1⁄3” at 130 ml/hintravenously for a total of �1.5 L; less than 24hours later, the serum sodium fell from 138 to “lessthan 120 mmol/L,” and the child became rigid,developed seizures and oxygen desaturation, andthen cardiac arrest despite the administration ofhypertonic saline; (3) a 6-year-old child who wasgiven 5% dextrose at 200 ml/h after an outpatienttonsillectomy because of a prescription error diedafter her serum sodium concentration fell to 107mmol/L over 12 hours; death was preceded byseveral hours of vomiting, jerking motions, rigidextremities, and rolled back eyes, erroneously at-tributed to a dystonic reaction to promethazine; and(4) a child of unspecified age who became somno-lent, with vomiting and seizure-like activity forseveral hours after the serum sodium fell to anunspecified level and after unspecified fluids 2 daysafter surgery for coarctation of the aorta. The Na-tional Patient Safety Agency in the United Kingdomhas identified hospital-acquired hyponatremia inchildren as a major patient safety issue. Safety alerts


and guidelines for the administration of fluids tochildren have been published as a result, includingthe requirement for sodium chloride 0.18% withglucose 4% intravenous solution to be removedfrom hospitals. A provincial coroner identified sixpediatric deaths related to acute hyponatremia inhospital settings in a 10-year period and provided aguideline for practitioners who administer paren-teral fluids to children.

These tragic experiences have caused many inves-tigators to question the widespread use of hypotonicsolutions for parenteral fluid maintenance therapy inchildren, a practice based on a formula that was devel-oped more than 50 years ago (16). The formula is derivedfrom minimum free water requirements based on caloricexpenditure per kilogram of body weight in normalchildren. These estimates are unlikely to be valid duringfasting and without physical activity; more important, theformula presumes normal excretion of free water by thekidneys because it was developed before release of ADHin response to nonosmotic stimuli was widely appreci-ated. A study of 124 children who were admitted forelective surgery documented vasopressin levels at orabove the range of maximal antidiuresis (3 to 5 pg/ml),before the induction of anesthesia (presumably reflectingthe effects of pain, anxiety, or nausea); during and aftersurgery, most patients had vasopressin levels of �3pg/ml that remained in many patients 24 hours aftersurgery (17). A number of studies, including randomizedtrials, have shown that fluid replacement with isotonicsaline carries a lower risk for hyponatremia withoutcausing hypernatremia (17–19).

Neurosurgical HyponatremiaHyponatremia is common in patients with acute

neurologic and neurosurgical conditions. Although theonset of the electrolyte disturbance is variable in theseconditions, patients with hyponatremia and intracranialpathology are at serious risk because even mild degreesof brain swelling are poorly tolerated. Hyponatremia isparticularly common in subarachnoid hemorrhage(SAH), and it augers a poor prognosis. Thus, there isgeneral agreement that hyponatremia in patients withintracranial pathology regardless of the duration of thedisturbance should be treated when the serum sodiumlevel is �131 mmol/L (20) and, if symptomatic, aggres-sively with hypertonic saline (21,22). A sliding-scalehypertonic saline infusion protocol, to be started once theserum sodium falls to �133 mmol/L or after a serum fall

in serum sodium concentration by �6 mmol/L over 24 to48 hours, was found to be effective in avoiding progres-sion of mild acute hyponatremia in critically ill patientswith neurologic and neurosurgical diseases (23). Theprotocol calls for administration of NaCl tablets 3 g every6 hours orally or per nasogatric tube combined withintravenous 3% NaCl, starting at an initial dosage of 20ml/h. The dosage of 3% NaCl is then escalated every 6hours by 10 to 20 ml/h, depending on the serum sodiumconcentration, to a maximum dosage of 80 ml/h; the rateis held constant for serum sodium values of 136 to 140mmol/L and held for 6 hours for serum sodium concen-trations �140 mmol/L. In contrast to use of hypertonicsaline maintenance fluid without sliding-scale adjust-ment, which has been reported to result in a 52% inci-dence of serum sodium concentrations �155 mmol/Land 34% ��160 mmol/L (24), �1% of the sodiumvalues exceeded 145 mmol/L with the sliding-scale ap-proach, whereas 0.06% of the time was spent withsodium values �130 mmol/L and 84% of the time wasspent in the goal range (Na 136 to 145 mol/L). Theregimen was well tolerated with no neurologic adverseeffects and no incidents of heart failure.

A sliding-scale hypertonic saline infusion pro-tocol, to be started once the serum sodiumfalls to < 133 mmol/L or after a serum fall inserum sodium concentration by >6 mmol/Lover 24 to 48 hours, was found to be effectivein avoiding progression of mild acute hypona-tremia in critically ill patients with neurologicand neurosurgical diseases.

Treatment of Acute Hyponatremic EmergenciesSelf-induced water intoxication and symptomatic hos-

pital-acquired hyponatremia are true emergencies that de-mand prompt and aggressive intervention with hypertonicsaline. In these conditions, the risks of cerebral edemacaused by hyponatremia exceed the risks of excess therapy.Similar recommendations apply to patients who have hypo-natremia and intracranial pathology and develop neurologicsymptoms and to patients who have hyponatremia andactive seizures regardless of the duration of the electrolytedisturbance. If hypertonic saline is withheld from patientswho have self-induced water intoxication and whose urineis not maximally dilute, there is a risk that serum concen-tration may decrease spontaneously because of delayed


absorption of water from the gastrointestinal tract. If hyper-tonic saline therapy is delayed or if isotonic saline is giveninstead to patients with acute postoperative hyponatremia,then excretion of concentrated urine can further reduce theserum sodium concentration with disastrous results. A re-view of the literature to identify reports of patients withhyponatremia and seizures or coma concluded that a 4- to6-mEq/L increase in the serum sodium concentration wasenough correction to rescue the most severely affectedpatients from complications of acute hyponatremia (1). Aconsensus has emerged that this goal is best achieved with100 ml or 2-ml/kg bolus infusions of 3% saline, repeated upto two times if necessary (1,25). Use of hypertonic saline totreat life-threatening intracranial hypertension in patientswith normonatremia and neurosurgical emergencies hasshown that an increase in serum sodium concentration ofthis magnitude can successfully reduce intracranial pressureand reverse impending herniation (1,26,27). Some cliniciansmay be reluctant to prescribe bolus infusions of 3% saline,in the erroneous belief that this violates an old in recom-mendation that correction not exceed 0.5 mEq/L per h. Suchreluctance reflects the confusion that was created when aproposed limit of 12 mEq/L per day was expressed as anhourly rate. In fact, there is no evidence that a rapid hourlyrate of correction is harmful as long as the total increase inserum sodium concentration over a 24-hour period is notexcessive.

A consensus has emerged that treatmentgoals for patients with seizures and coma asa result of acute hyponatremia are bestachieved with 100 ml or 2-ml/kg bolus infu-sions of 3% saline, repeated up to two timesif necessary.

Confusion was created when a proposedlimit of 12 mEq/L per day was expressedas an hourly rate. In fact, there is no evi-dence that a rapid hourly rate of correctionis harmful as long as the total increase inserum sodium concentration over a 24-hour period is not excessive.

References1. Sterns R, Nigwekar S, Hix J: The treatment of hyponatremia. Semin

Nephrol 29: 282–299, 2009

2. Santos-Soares PC, Bacellar A, Povoas HP, Oliveira-Filho J, Filguei-ras NM, Brito AF: Excessive water ingestion and repeated seizures:the domino effect. Arq Neuropsiquiatr 66: 552–553, 2008

3. Yalcin-Cakmakli G, Karli Oguz K, Shorbagi A, unda Bas D, Ergan-Arsava B, Kunt M, Akif Topcuoglu M: Hyponatremic encephalopa-thy after excessive water ingestion prior to pelvic ultrasound: Neu-roimaging findings. Intern Med 49: 1807–1811, 2010

4. Dvir D, Beigel R, Hoffmann C, Tsarfati G, Farfel Z, Pauzner R:Hyponatremic brain edema: Correlation with serial computed tomog-raphy scans. Isr Med Assoc J 11: 442–443, 2009

5. Gonzalez I, Perez N, Penas-LLedo EM, LLerena A, Gonzalez I, de laRubia A: High risk of polydipsia and water intoxication in schizo-phrenia patients. Schizophr Res 99: 377–378, 2008

6. Goldman MB: The assessment and treatment of water imbalance inpatients with psychosis. Clin Schizophr Relat Psychoses 4: 115–123,2010

7. Goldman MB: The mechanism of life-threatening water imbalance inschizophrenia and its relationship to the underlying psychiatric ill-ness. Brain Res Rev 61: 210–220, 2009

8. Torres IJ, Keedy S, Marlow-O’Connor M, Beenken B, Goldman MB:Neuropsychological impairment in patients with schizophrenia andevidence of hyponatremia and polydipsia. Neuropsychology 23: 307–314, 2009

9. Meulendijks D, Mannesse CK, Jansen PA, van Marum RJ, EgbertsTC: Antipsychotic-induced hyponatraemia: A systematic review ofthe published evidence. Drug Saf 33: 101–114, 2010

10. Rosner MH: Exercise-associated hyponatremia. Semin Nephrol 29:271–281, 2009

11. Siegel AJ, d’Hemecourt P, Adner MM, Shirey T, Brown JL, Lewand-rowski KB: Exertional dysnatremia in collapsed marathon runners: Acritical role for point-of-care testing to guide appropriate therapy.Am J Clin Pathol 132: 336–340, 2009

12. Chapman TH, Hamilton M: Water intoxication presenting as mater-nal and neonatal seizures: A case report. J Med Case Reports 2: 366,2008

13. Bergum D, Lonnee H, Hakli TF: Oxytocin infusion: Acute hypona-traemia, seizures and coma. Acta Anaesthesiol Scand 53: 826–827,2009


15. Koczmara C, Wade AW, Skippen P, Campigotto MJ, StreitenbergerK, Carr R, Wong E, Robertson K: Hospital-acquired acute hypona-tremia and reports of pediatric deaths. Dynamics 21: 21–26, 2010

16. Holliday MA, Segar WE. The maintenance need for water in paren-teral fluid therapy. Pediatrics 19: 823–832, 1957

17. Neville KA, Sandeman DJ, Rubinstein A, Henry GM, McGlynn M,Walker JL: Prevention of hyponatremia during maintenance intrave-nous fluid administration: A prospective randomized study of fluidtype versus fluid rate. J Pediatr 156: 313-9.e1–313-9.e2, 2010

18. Au A, Bell MJ: Prevention of hospital-acquired hyponatremia inchildren: Are hypotonic solutions safe? Pediatr Crit Care Med 11:528–529, 2010

19. Au AK, Ray PE, McBryde KD, Newman KD, Weinstein SL, BellMJ: Incidence of postoperative hyponatremia and complications incritically-ill children treated with hypotonic and normotonic solu-tions. J Pediatr 152: 33–38, 2008

20. Rahman M, Friedman WA: Hyponatremia in neurosurgical patients:Clinical guidelines development. Neurosurgery 65: 925–935, discus-sion 935–936, 2009

21. Sterns RH, Hix JK, Silver S: Treatment of hyponatremia. Curr OpinNephrol Hypertens 19: 493–498, 2010

22. Sterns RH, Nigwekar SU, Hix JK: The treatment of hyponatremia.Semin Nephrol 29: 282–299, 2009

23. Woo CH, Rao VA, Sheridan W, Flint AC: Performance characteris-tics of a sliding-scale hypertonic saline infusion protocol for the


treatment of acute neurologic hyponatremia. Neurocrit Care 11:228–234, 2009

24. Froelich M, Ni Q, Wess C, Ougorets I, Hartl R: Continuous hyper-tonic saline therapy and the occurrence of complications in neuro-critically ill patients. Crit Care Med 37: 1433–1441, 2009

25. Moritz ML, Ayus JC: 100 cc 3% sodium chloride bolus: A noveltreatment for hyponatremic encephalopathy. Metab Brain Dis 25:91–96, 2010

26. Oddo M, Levine JM, Frangos S, Carrera E, Maloney-Wilensky E,Pascual JL, Kofke WA, Mayer SA, LeRoux PD: Effect of mannitoland hypertonic saline on cerebral oxygenation in patients with severetraumatic brain injury and refractory intracranial hypertension. J Neu-rol Neurosurg Psychiatry 80: 916–920, 2009

27. Francony G, Fauvage B, Falcon D, Canet C, Dilou H, Lavagne P,Jacquot C, Payen JF: Equimolar doses of mannitol and hypertonicsaline in the treatment of increased intracranial pressure. Crit CareMed 36: 795–800, 2008

Chronic Hypotonic Hyponatremia

Differential DiagnosisChronic hyponatremia is usually caused by

vasopressin (AVP)-mediated water retention. Be-cause AVP levels should normally be undetectablewhen the plasma sodium concentration falls below135 mmol/L, secretion of AVP in the presence ofhyponatremia represents a nonosmotic stimulus forAVP release or ectopic secretion of the hormone.The most common cause for nonosmotic release ofAVP is an inadequate circulating blood volumecaused by true hypovolemia, heart failure, or he-patic cirrhosis and sensed by pressure and volumereceptors that relay signals to AVP-secreting cells inthe hypothalamus. These conditions also activatethe sympathetic nervous system and result in in-creased levels of renin, angiotensin, and aldoste-rone, factors that cause sodium retention and arereflected in a low urine sodium concentration.

Hyponatremia caused by heart failure or hepaticcirrhosis is clinically obvious because it occurs inpatients with advanced disease. Hypovolemic hypona-tremia may be equally obvious, with orthostatic hypo-tension, tachycardia, and prerenal azotemia, but, inmany cases, these features are absent and the clinicianmust turn to laboratory clues to distinguish betweenhypovolemic hyponatremia and syndrome of inappro-priate antidiuretic hormone secretion (SIADH). Theuse of fractional excretion of sodium and fractionalexcretion of urea in the differential diagnosis ofchronic hyponatremia were discussed in the last fluidand electrolyte issue of NephSAP and in a recentlypublished review (1). However, these diagnostictools are useful only in identifying patients who

have developed hyponatremia from extrarenal fluidlosses. Patients with renal salt-wasting syndromespose a greater diagnostic challenge because urineosmolality and urine sodium both are high andcannot be distinguished from urine chemistries inpatients with SIADH.

Causes of Renal Salt WastingAddison Disease. In patients with unexplained hypo-natremia, adrenal insufficiency must be excluded beforemaking a diagnosis of SIADH (2). Addison disease oftenpresents insidiously, and symptoms of postural hypoten-sion are late manifestations. In developed countries, Ad-dison disease is usually caused by autoimmune adrenali-tis, which can occur in isolation or as a part ofautoimmune polyendocrinopathy syndromes: (1) Auto-immune polyendocrinopathy syndrome type 1, includingautoimmune hypoparathyroidism, chronic mucocutane-ous candidiasis, and other autoimmune disorders (e.g.,type 1 diabetes, chronic active hepatitis, primary gonadalfailure, autoimmune thyroid disease) and (2) autoim-mune polyendocrinopathy syndrome type 2, primarilyincluding type 1 diabetes or autoimmune thyroid diseaseand sometimes primary gonadal failure, pernicious ane-mia, and vitiligo. Tuberculosis remains the most com-mon cause of Addison disease worldwide, and fungalinfections and cytomegalovirus adrenalitis may be com-plications of immunodeficiency. Adrenoleukodystrophy,caused by accumulation of long-chain fatty acids in theadrenal gland and central and peripheral nervous system,is an important cause of Addison disease in men. Amorning serum cortisol level �18 �g/dl (500 nmol/L)generally excludes Addison disease, whereas a level �6�g/dl is suggestive of the disease. The response toinjection of a synthetic analog of ACTH is the definitivediagnostic test.

Addison disease can often go unrecognized formonths or years. A German study of 270 patients whohad adrenal insufficiency (99 with Addison diseaseand 117 with hypopituitarism) and agreed to completea questionnaire found that �30% of women and 50%of men received a diagnosis within the first 6 monthsafter the onset of symptoms; in 20% of patients,diagnosis was delayed for �5 years (3). Fatigue andlack of energy were complaints in 74% of women and88% of men. Symptoms of nausea and vomiting oc-curred in 50% of women and 25% of men withAddison disease; weight loss, anorexia, abdominalpain, and diarrhea were also common. These symp-


toms often led to erroneous diagnoses of psychiatric orgastrointestinal disorders.Congenital Adrenal Hyperplasia. Congenital ad-renal hyperplasia (CAH) from 21-hydroxylase defi-ciency is an autosomal recessive disorder that causescortisol deficiency and androgen excess, with or with-out aldosterone deficiency. Severe life-threatening,salt-wasting hyponatremia, often with serum sodiumvalues �125 mmol/L, and hyperkalemia are commonfeatures that can result in death in early infancy; theseabnormalities respond to cortisol and mineralocorti-coid replacement with fludrocortisone (4,5). The dis-ease is most obvious in girls because of ambiguousgenitalia, but boys appear normal. Advances in therecognition and care of the disease have enabled manychildren to reach adulthood. Because of the unfamil-iarity of internists with the disease, young adults withCAH are often lost to follow-up and may stop takingcorticosteroid replacement (6,7). Salt-wasting crisesappear less often in adults but can emerge duringconcurrent illnesses.Cisplatin. Cisplatin has been used effectively in thetreatment of a variety of solid tumors. However, with-out adequate hydration, 50% of patients develop ausually reversible decline in GFR and, less commonly,a renal salt-wasting syndrome as a result of proximaltubular injury that may result in hyponatremia; acomprehensive review of this condition was publishedrecently (8). Renal salt wasting has been detected from12 hours up to 1 month after cisplatin administrationafter doses ranging from �200 to 600 mg/m2. Signs ofvolume depletion including orthostatic hypotensionand weight loss (as much as 12 lb in 3 days in onereported case) have been present in most reports ofcisplatin-induced hyponatremia. Hyponatremia causedby cisplatin is often treated as SIADH because bothconditions are characterized by hypotonic hyponatre-mia and increased urinary sodium concentrations(8,9). Clinical signs of hypovolemia, including weightloss, symptomatic orthostatic hypotension, and tachy-cardia, are responsive to volume replacement anddistinguish cisplatin-induced renal salt wasting fromSIADH. Orthostatic hypotension also occurs in pa-tients with SIADH caused by small cell lung cancer,but, in these cases, it is caused by autonomic insuffi-ciency and not volume depletion, and it does notrespond to intravenous fluids. It is important not totreat cisplatin-induced salt wasting with water restric-tion, diuretics, or AVP antagonists because this may

lead to more volume depletion and hemodynamiccollapse. Treatment should be aimed at restoring in-travascular volume with oral fluids and salt tablets orintravenous isotonic saline.Cerebral Salt Wasting. Hyponatremia is a com-mon complication of central nervous system disease.High urine sodium concentrations despite a low serumsodium concentration led to the term “cerebral saltwasting” in the early 1950s, before the pathophysiol-ogy of SIADH was understood. The term has reap-peared in recent years and is now considered by manyneurointensivists to be a common cause of hyponatre-mia in patients with intracranial pathology (10–13).As discussed in the last fluid and electrolyte issue ofNephSAP, there is controversy in the literature as towhether hyponatremia in patients with intracranialdisease should be ascribed to SIADH or to “cerebralsalt wasting.” Some have questioned whether cere-bral salt wasting actually exists and, if it does,whether it is possible to diagnose the conditionreliably. A valid diagnosis of cerebral salt wastingrequires the demonstration that urinary salt lossespersist despite hypovolemia. In the neurosurgicalliterature describing purported cases of cerebral saltwasting, hypovolemia is defined clinically as acentral venous pressure (CVP) �5 mmHg. How-ever, CVP is not a reliable measure of volumedepletion. Marik et al. (14) reviewed the literatureto identify studies that compared CVP measure-ments with other objective measures of intravacularvolume. The only study they could find demonstrat-ing the utility of CVP in predicting volume statuswas performed in seven standing, awake maresundergoing controlled hemorrhage. Of the 24 stud-ies of humans included in their analysis, five com-pared CVP with the measured circulating bloodvolume, whereas 19 studies determined the relation-ship between CVP and change in cardiac perfor-mance after a fluid challenge (generally defined as a�10 to 15% increase in stroke index/cardiac index).In all, 830 patients across a spectrum of medical andsurgical disciplines were studied; the pooled corre-lation coefficients between the CVP or �CVP andmeasured blood volume or change in stroke index/cardiac index were �0.18 for all of these compari-sons. The baseline CVP (reported in 11 studies) was8.7 � 2.3 mmHg in patients who responded to fluidchallenge, as compared with 9.7 � 2.2 mmHg in nonre-sponders (NS; P � 0.3). The receiver operating charac-


teristic curve for the CVP, a statistical tool that helpsassess the likelihood of a result being a true positiveversus a false positive, was 0.56; thus, at any CVP value,the likelihood that CVP can accurately predict fluidresponsiveness was only 56%.

Maesaka et al. (15) have advocated for the use ofthe fractional excretion of urate (FEurate) and frac-tional excretion of phosphate (FEphosphate) to distin-guish between SIADH and hyponatremia caused bysalt wasting. Data supporting this distinction are sum-marized in a recent review and a small series by theseauthors (16). SIADH is associated with hypouricemia,largely as a result of increased FEurate, and the FEur-ate typically improves with correction of hyponatre-mia. On the basis of a number of individual cases theyhave studied, these investigators argued that a subsetof patients with hyponatremia, some with cerebraldisease and some without, have “appropriate” ADHsecretion in response to hypovolemia caused by vol-ume depletion. These patients, unlike those withSIADH, had a persistently elevated FEurate and hy-pouricemia after correction of hyponatremia, and thishas been associated with increased levels of renin andaldosterone, increased FEphosphate (paired in somecases with a reduction in measured blood volume), anda response to saline. One of these patients had a urinesodium concentration of 6 mmol/L after a hip fracture,which many would attribute to hypovolemia as a resultof extrarenal fluid losses. Despite a low urine sodiumconcentration, the investigators attributed hypovole-mia to previous salt wasting by the proximal tubulebecause the persistently high FEuric acid and FEphos-phate were thought to reflect a proximal tubular ab-normality. This argument would be more convincing ifrecurrent hyponatremia associated with negative so-dium balance, weight loss, and clinical signs of hypo-volemia had been demonstrated after discontinuationof saline therapy.

Causes of Euvolemic HyponatremiaTumor-Associated SIADH. The first cases of theSIADH were reported in patients with small cell lungcancer (SCLC). This disease continues to be one of themost common causes of chronic SIADH. Among 455patients with a diagnosis of SCLC in a single Univer-sity hospital in Denmark, 44% had hyponatremia; theplasma sodium concentration was �125 mmol/L in11% of patients and 126 to 135 mmol/L in 33% (17).By comparison, only 1.6% of patients with non-SCLC

had hyponatremia at the time of diagnosis. Only 25%of patients who had serum sodium concentrations�130 mmol/L and were treated with water restrictionhad fully normalized their serum sodium to 135mmol/L by the time of the second cycle of chemother-apy, 3 to 4 weeks after the first cycle. Patients who didnot fully normalize their serum sodium had a signifi-cantly worse prognosis than patients who did. Atrialnatriuretic peptide (ANP) is produced by most SCLCcell lines, and ectopic secretion of ANP as well asSIADH may play a role in the pathogenesis of hypo-natremia. Patients with ectopic secretion of ANP havebeen reported to respond less positively to water re-striction, and lack of response to water restriction hasbeen suggested as a screening test for this abnormality.Hyponatremic Hypertensive Syndrome. Somepatients with unilateral renal artery stenosis develophypertension, hyponatremia, polyuria, hypokalemia,and marked polydipsia. Most patients are elderly,asthenic women with renal artery stenosis as a result ofatherosclerosis; the disease has also been describedoccasionally in children with fibromuscular dysplasia(18). Renin secretion induced by renal ischemia resultsin high circulating levels of angiotensin II, a potentdipsogen; high doses of angiotensin-converting en-zyme inhibitors to block the conversion from angio-tensin I to angiotensin II have been suggested. Hyper-tension in patients with Wilms’ tumor is relativelycommon and usually due to intrarenal ischemia, sec-ondary to compression of the normal renal vascula-ture, which increases renin production. A recent reportdescribed two children who had Wilms’ tumor andfirst presented with severe hypertension, hyponatre-mia, hypokalemia, polyuria, and polydipsia; preoper-ative treatment with an angiotensin-converting en-zyme inhibitor (captopril) caused a remission of theclinical syndrome, thereby allowing a safer curativenephrectomy (19).

Drug-Induced Euvolemic HyponatremiaCyclophosphamide. Cyclophosphamide is an alky-lating agent used extensively in the treatment of ma-lignant and rheumatologic diseases. Because one of itsadverse effects is hemorrhagic cystis, it is routinelyadministered with forced hydration. Although hypo-natremia was once a common complication when highdosages (�50 mg/kg) of cyclophosphamide were usedto treat neoplastic diseases, there have been only a fewcase reports of hyponatremia complicating more cur-


rent low-dosage (�20 mg/kg) regimens. A retrospec-tive analysis of 84 patients who were given intrave-nous pulse cyclophosphamide (500 to 750 mg/m2)with half-isotonic saline for prophylactic hydrationidentified hyponatremia (�135 mmol/L) in 15 treat-ment episodes (13.4%) in 12 patients (14.3%) (20).Hyponatremia was generally mild and largely asymp-tomatic, with serum sodium concentrations �130mmol/L (123 to 129 mmol/L) in only five patients.Urinary data from eight patients with hyponatremiawere consistent with SIADH with urine osmolality327 � 195 mOsm/kg H2O (range 182 to 790mOsm/kg H2O) and urine sodium 68 � 47 mmol/L(range 31 to 176 mmol/L). However, as shown in tworecent case reports, hyponatremia after low-dosagecyclophosphamide can be more serious (21,22). Forexample, a 49-year-old woman with diffuse systemicsclerosis presented with a serum sodium of 106mmol/L and urine osmolality of 620 mOsm/kg 1 dayafter receiving a 500-mg intravenous pulse of cyclo-phosphamide (21). She developed seizures 12 hoursafter presentation, prompting treatment with 3% sa-line, and subsequently (with an unreported increase inserum sodium concentration) a biphasic neurologiccourse associated with magnetic resonance imaging–documented central pontine myelinolysis. The causeof cyclophosphamide-induced hyponatremia is un-known, but because it was reported to cause an antidi-uresis in a patient with diabetes insipidus (DI), there isspeculation that the drug may enhance AVP’s effecton the collecting duct (20,23). The antidiuresis causedby cyclophosphamide is temporally related to the ap-pearance of its metabolites in the urine. Cyclophosph-amide metabolites (mafosfamide and 4-hydroperoxy-cyclophosphamide) decrease the production of IL-1and TNF-� in a dosage-dependent manner (23), andthese cytokines have been shown to downregulateexpression of the vasopressin 2 receptor (V2R) andaquaporin 2 (AQP2) (24); cyclophosphamide couldpotentially cause hyponatremia by upregulating ex-pression of V2R and AQP2 through suppression ofIL-1 and TNF-�, which are effector molecules in thedownregulation of VR2 (23).Carbamazepine. Carbamazepine is an anticonvul-sant and psychotropic medication commonly used inthe treatment of patients with epilepsy or intellectualdisability. The drug has an antidiuretic effect thatcommonly results in hyponatremia with clinical fea-tures of SIADH; this side effect has been exploited to

decrease the urinary volume in DI. Oxcarbazepine,which differs from carbamazepine only by the additionof one oxygen molecule, is also a recognized cause ofhyponatremia. The antidiuretic effect of carbamaz-epine is not well characterized; direct stimulation ofAVP release from the pituitary gland has been sug-gested, but there has also been evidence of a possibleeffect directly on the renal tubule. Braganca et al. (25)investigated the mechanism using in vitro and in vivoexperiments that support the latter hypothesis. Mi-croperfusion studies showed that in the absence ofADH, carbamazepine was able to increase water ab-sorption fourfold in the isolated rat inner medullarycollecting duct (IMCD); administration of specificinhibitors suggested that carbamazepine acts on theV2R–protein G complex and that its effect on waterflow is cAMP dependent. Administration of carbam-azepine in vivo to rats with nephrogenic DI signifi-cantly blunted the water diuresis induced by lithium,decreasing urine volume and increasing urine osmola-lity; consistent with this effect, AQP2 expression inIMCD obtained from rats given lithium and carbam-azepine was significantly increased in comparisonwith that in rats given lithium alone.Selective Serotonin Reuptake Inhibitors. Selec-tive serotonin reuptake inhibitors (SSRIs), the mostpopular antidepressants in use, have become one of themost common causes of hyponatremia, particularly inthe elderly during the first 2 weeks of treatment.Patients with SSRI-induced hyponatremia present witha clinical SIADH, but increased plasma levels of AVPhave not been demonstrated. Moyses et al. (26) treatedrats with fluoxetine, one of the most popular SSRIs,and found that the plasma sodium decreased from139.3 � 0.78 to 134.9 � 0.5 mmol/L (P � 0.01),whereas plasma AVP levels remained unchanged.IMCDs from fluoxetine-treated rats showed a 40%increase in AQP2 protein abundance, and incubationof IMCDs with fluoxetine resulted in a similar in-crease in AQP2. In addition, microperfusion studies ofIMCDs showed that in the absence of ADH, carbam-azepine doubled water absorption. These results areconsistent with the conclusion that SSRIs stimulatewater absorption by a non-ADH modulation ofAQP2 in the kidney. In addition to carbamazepineand fluoexetine, hydrochlorothiazide, chlorprop-amide, angiotensin I, and rosiglitazone all seem tohave AVP-independent effects on water reabsorp-tion in the collecting duct (26).


Other Drugs. Hyperkalemia is a common compli-cation of trimethoprim, which inhibits sodium influxvia the epithelial sodium channel (ENaC) in the col-leting duct. Much more rarely, the drug has beenassociated with hyponatremia, presumed to result fromrenal sodium wasting and hypovolemia-induced AVPrelease (27). Hyponatremia has also been occasionallyassociated with ciprofloxacin neurotoxicity (27).

Nephrogenic Syndrome of InappropriateAntidiuresis

A gain-of-function mutation of the V2R gene,named nephrogenic syndrome of inappropriate antidi-uresis (NSIAD), was first described in 2005 (28–35).Manifestations are opposite those of nephrogenic DI,which is caused by loss of function of the AVPreceptor. In the original description of the disease, twounrelated male infants were found to have the clinicalpicture of SIADH with undetectable plasma AVPlevels. In one of the original patients, the arginine oncodon 137 was substituted with cysteine (R137C), andhis mother was heterozygous for the R137C mutation,suggesting an X-linked inheritance. In the other pa-tient, arginine was substituted with leucine (R137L),with no mutation found in his mother, likely becauseof a spontaneous mutation. Functional analysis con-firmed constitutive activation of this receptor.

More than 200 different mutations on the V2Rgene are known to cause nephrogenic diabetes insipi-dus, so genetic heterogeneity might be expected forNSIAD. However, since the initial description of thedisease, R137C has been the mutation found in mostsubsequently described cases. One report identifiedthe activating missense mutation on R137C in threehemizygous males and four heterozygous females of alarge family pedigree (36). All except one female, whowas subsequently found to have skewed X inactiva-tion, had either spontaneous hyponatremia or an ab-normal water-loading test. The mother of another caseof NSIAD caused by R137C has been reported to havepersistent hyponatremia (128 mmol/L) (32); measure-ments of AVP levels found detectable AVP at plasmaosmolality levels below the normal osmolar thresholdfor AVP release, and the mother experienced thirst ata low plasma osmolality. Thus, in this patient, boththirst and renal water handling were abnormal. Thirsthas not been studied in other patients with activatingV2R mutations.

Mutations in the V2R gene may not explain all

cases of NSIAD. In 2001, a three-generation familywas found to have an “SIADH-like condition” withundetectable ADH levels. Studies of three affectedmembers of the family (all females) showed normallevels of urine AQP2, a marker of AVP-dependentAQP2 levels. A “gain of function” mutation involvingthe AVP receptor or AQP channel could have ex-plained these findings, but DNA sequencing for bothAVPR2 and AQP2 were normal.

Infants with NSIAD are at high risk for hypona-tremia because their diet consists predominantly ofliquid with little solute. As awareness of thirst evolvesover time, the risk for severe hyponatremia fromexcessive free water decreases. Because affected indi-viduals can often self-regulate fluid intake, even pa-tients who present in infancy reach adult life withminimal symptomatic episodes. The identification ofadult relatives of affected infants with NSIAD sug-gests that the disease could be present in geneticallyaffected individuals who remain asymptomatic or havenever had full evaluation. NSIAD should be suspectedin young patients with a clinical diagnosis of idio-pathic SIADH, especially if there is a good familyhistory; a good screening test for the condition may beunresponsiveness to V2R antagonists.

The identification of adult relatives of af-fected infants with NSIAD suggests thatthe disease could be present in geneticallyaffected individuals who remain asymp-tomatic or have never had full evaluation.

Pharmaceutical-grade urea in a dosage of 2 g/kgper d in four divided doses has been used to treatNSIAD in children, and it has been found to be welltolerated, despite the unpleasant taste as observed bythe mother and others (30). The need for urea de-creases with the development of a reliable thirst mech-anism.

References1. Decaux G: The syndrome of inappropriate secretion of antidiuretic

hormone (SIADH). Semin Nephrol 29: 239–256, 20092. Chakera AJ, Vaidya B: Addison disease in adults: Diagnosis and

management. Am J Med 123: 409–413, 20103. Bleicken B, Hahner S, Ventz M, Quinkler M: Delayed diagnosis of

adrenal insufficiency is common: A cross-sectional study in 216patients. Am J Med Sci 339: 525–531, 2010

4. Lee CT, Tung YC, Hsiao PH, Lee JS, Tsai WY: Clinical character-istics of Taiwanese children with congenital adrenal hyperplasia


caused by 21-hydroxylase deficiency in the pre-screening era. JFormos Med Assoc 109: 148–155, 2010

5. White PC: Neonatal screening for congenital adrenal hyperplasia. NatRev Endocrinol 5: 490–498, 2009

6. Auchus RJ: Management of the adult with congenital adrenal hyper-plasia. Int J Pediatr Endocrinol 2010: 614107, 2010

7. Auchus RJ: Congenital adrenal hyperplasia in adults. Curr OpinEndocrinol Diabetes Obes 17: 210–216, 2010

8. Hamdi T, Latta S, Jallad B, Kheir F, Alhosaini MN, Patel A:Cisplatin-induced renal salt wasting syndrome. South Med J 103:793–799, 2010

9. Brown KR, Leitao MM Jr: Cisplatin-induced syndrome of inappro-priate antidiuretic hormone (SIADH) in a patient with neuroendo-crine tumor of the cervix: A case report and review of the literature.Eur J Gynaecol Oncol 31: 107–108, 2010

10. Palmer BF: Hyponatremia in the intensive care unit. Semin Nephrol29: 257–270, 2009

11. Nakagawa I, Kurokawa S, Nakase H: Hyponatremia is predictable inpatients with aneurysmal subarachnoid hemorrhage” clinical signifi-cance of serum atrial natriuretic peptide. Acta Neurochir (Wien) 152:2147–2152, 2010

12. Costa KN, Nakamura HM, Cruz LR, Miranda LS, Santos-Neto RC,Cosme Sde L, Casulari LA: Hyponatremia and brain injury: Absenceof alterations of serum brain natriuretic peptide and vasopressin. ArqNeuropsiquiatr 67: 1037–1044, 2009

13. Audibert G, Steinmann G, de Talance N, Laurens MH, Dao P,Baumann A, Longrois D, Mertes PM: Endocrine response aftersevere subarachnoid hemorrhage related to sodium and blood volumeregulation. Anesth Analg 108: 1922–1928, 2009

14. Marik PE, Baram M, Vahid B: Does central venous pressure predictfluid responsiveness? A systematic review of the literature and thetale of seven mares. Chest 134: 172–178, 2008

15. Maesaka JK, Imbriano LJ, Ali NM, Ilamathi E: Is it cerebral or renalsalt wasting? Kidney Int 76: 934–938, 2009

16. Bitew S, Imbriano L, Miyawaki N, Fishbane S, Maesaka JK: More onrenal salt wasting without cerebral disease: Response to saline infu-sion. Clin J Am Soc Nephrol 4: 309–315, 2009

17. Hansen O, Sorensen P, Hansen KH: The occurrence of hyponatremiain SCLC and the influence on prognosis: A retrospective study of 453patients treated in a single institution in a 10-year period. LungCancer 68: 111–114, 2010

18. Neeli S: Renal artery stenosis with hyponatremic hypertensive syn-drome in a 7-year-old child. J Pediatr Urol 4: 407–408, 2008

19. D’Angelo P, Catania S, Zirilli G, Collini P, Tropia S, Perotti D,Terenziani M, Spreafico F: Severe polyuria and polydipsia in hy-ponatremic-hypertensive syndrome associated with Wilms tumor.Pediatr Blood Cancer 55: 566–569, 2010

20. Lee YC, Park JS, Lee CH, Bae SC, Kim IS, Kang CM, Kim GH:Hyponatraemia induced by low-dose intravenous pulse cyclophosph-amide. Nephrol Dial Transplant 25: 1520–1524, 2010

21. Jayachandran NV, Chandrasekhara PK, Thomas J, Agrawal S, Nar-simulu G: Cyclophosphamide-associated complications: We need tobe aware of SIADH and central pontine myelinolysis. Rheumatology(Oxford) 48: 89–90, 2009

22. Berger AK, Bellos F, Siegmund A, Eisenbach C, Lordick F: Symp-tomatic hyponatraemia caused by cylophosphamide. Onkologie 32:280–282, 2009

23. Park SJ, Kim JH, Shin JI: Insight on mechanism of hyponatraemiainduced by low-dose intravenous pulse cyclophosphamide. NephrolDial Transplant 25: 3453, author reply 3453–3454, 2010

24. Hocherl K, Schmidt C, Kurt B, Bucher M: Inhibition of NF-kappaBameliorates sepsis-induced downregulation of aquaporin-2/V2 recep-tor expression and acute renal failure in vivo. Am J Physiol RenalPhysiol 298: F196–F204, 2010

25. de Braganca AC, Moyses ZP, Magaldi AJ: Carbamazepine can

induce kidney water absorption by increasing aquaporin 2 expression.Nephrol Dial Transplant 25: 3840–3845, 2010

26. Moyses ZP, Nakandakari FK, Magaldi AJ: Fluoxetine effect onkidney water reabsorption. Nephrol Dial Transplant 23: 1173–1178,2008


28. Levtchenko EN, Monnens LA: Nephrogenic syndrome of inappro-priate antidiuresis. Nephrol Dial Transplant 25: 2839–2843, 2010

29. Rochdi MD, Vargas GA, Carpentier E, Oligny-Longpre G, Chen S,Kovoor A, Gitelman SE, Rosenthal SM, von Zastrow M, Bouvier M:Functional characterization of vasopressin type 2 receptor substitu-tions (R137H/C/L) leading to nephrogenic diabetes insipidus andnephrogenic syndrome of inappropriate antidiuresis: Implications fortreatments. Mol Pharmacol 77: 836–845, 2010

30. Cho YH, Gitelman S, Rosenthal S, Ambler G: Long-term outcomesin a family with nephrogenic syndrome of inappropriate antidiuresis.Int J Pediatr Endocrinol 2009: 431527, 2009

31. Marcialis MA, Dessi A, Contu S, Fanos V: Nephrogenic syndrome ofinappropriate antidiuresis: A novel cause of euvolemic hypotonichyponatremia in newborns. Diagnosis and practical management. JMatern Fetal Neonatal Med 22[Suppl 3]: 67–71, 2009

32. Gupta S, Cheetham TD, Lambert HJ, Roberts C, Bourn D, CoulthardMG, Ball SG: Thirst perception and arginine vasopressin productionin a kindred with an activating mutation of the type 2 vasopressinreceptor: The pathophysiology of nephrogenic syndrome of inappro-priate antidiuresis. Eur J Endocrinol 161: 503–508, 2009

33. Kocan M, See HB, Sampaio NG, Eidne KA, Feldman BJ, PflegerKD: Agonist-independent interactions between beta-arrestins andmutant vasopressin type II receptors associated with nephrogenicsyndrome of inappropriate antidiuresis. Mol Endocrinol 23: 559–571, 2009

34. Soule S, Florkowski C, Potter H, Pattison D, Swan M, Hunt P,George P: Intermittent severe, symptomatic hyponatraemia due to thenephrogenic syndrome of inappropriate antidiuresis. Ann ClinBiochem 45: 520–523, 2008

35. Marcialis MA, Faa V, Fanos V, Puddu M, Pintus MC, Cao A,Rosatelli MC: Neonatal onset of nephrogenic syndrome of inappro-priate antidiuresis. Pediatr Nephrol 23: 2267–2271, 2008

36. Ranchin B, Boury-Jamot M, Blanchard G, Dubourg L, Hadj-Aïssa A,Morin D, Durroux T, Cochat P, Bricca G, Verbavatz JM, Geelen G:Familial nephrogenic syndrome of inappropriate antidiuresis: Disso-ciation between aquaporin-2 and vasopressin excretion. J Clin Endo-crinol Metab 95: E37–E43, 2010

Clinical Outcomes of Chronic HyponatremiaHyponatremia is a common problem in hospitals

and in nursing homes. A number of clinical conditions,including heart failure, cirrhosis, pneumonia, HIV in-fection, chronic obstructive pulmonary disease, myo-cardial infarction, malignant diseases, and several neu-rologic conditions, are associated with hyponatremia,and the electrolyte disturbance is associated with aworse prognosis (1).

MortalityHyponatremia is known to increase the risk for

mortality in patients with acutely decompensated heartfailure. A single-center study of 628 patients whopresented to the emergency department with acutely


decompensated heart failure found that 24% ofpatients had serum sodium concentrations �135mmol/L. Patients with hyponatremia were less likelyto be male or to have hypertension or coronary arterydisease but significantly more likely to have severesymptoms, anemia, and higher amino-terminal pro-B-type natriuretic peptide (NT-proBNP) concentrations.In a multivariate Cox proportional hazards analysis,hyponatremia was an independent predictor of 1-yearmortality (hazard ratio 1.72; 95% confidence interval[CI] 1.22 to 2.37; P � 0.001) as was an NT-proBNPconcentration above the median value of 4690 pg/ml(hazard ratio 1.49; 95% CI 1.10 to 2.00; P � 0.009).Patients with hyponatremia and patients with a higherNT-proBNP were more likely to develop worseningrenal function during their hospitalization; they alsohad the highest 1-year mortality rates. Hyponatremiawas predictive of 1-year mortality only in patients withan elevated NT-proBNP (2).

Wald et al. (3) conducted a 7-year retrospectivestudy of all discharges of adult patients from a 400-bedacute care tertiary university-affiliated teaching hospitalin Boston and found that hospital-associated hyponatre-mia is independently associated with in-hospital mortal-ity, prolongation of length of stay (LOS), and dischargeto a facility rather than to home, regardless of whether theelectrolyte disturbance was present on admission, wasexacerbated after admission, or developed during hospi-talization. Similar to what has been reported for patientswith heart disease and liver disease, in this unselectedhospital population, even serum sodium values thatwould conventionally be classified as normal or slightlybelow normal (e.g., 133 to 137 mmol/L) were found tobe independently associated with mortality, prolongedLOS, and discharge to a facility. Because of the findingthat a serum Na �138 (corrected for the effect of hyper-glycemia) is associated with significantly higher mortal-ity, the authors suggested that our current definition ofwhat constitutes normonatremia should be reconsidered.Even in this teaching hospital, postadmission aggravationof hyponatremia was observed in 6% of patients who hadhyponatremia on admission, and the serum sodium fellbelow 138 mmol/L in 38% of normonatremic admis-sions. In patients with hospital-acquired hyponatremia, a15-fold increase in the risk for death was observed whenthe lowest serum sodium fell to �127 mmol/L. In com-parison, the odds ratios for mortality were more modestin patients with the most severe forms of community-acquired hyponatremia. A sophisticated statistical analy-

sis (restrictive cubic spline) demonstrated a U-shapedrelationship between the serum sodium concentrationand mortality with a value of 140 mmol/L associatedwith the lowest risk for mortality and a progressiveincrease in mortality risk as admission serum sodiumlevel declined (Figure 11); however, because very fewpatients had an admission serum sodium value �123mmol/L, the effect of severe hyponatremia (�120mmol/L) on mortality cannot be accurately estimatedfrom this relationship.

Waikar et al. (4) studied short- and long-termmortality of �95,000 patients who were admitted withand without hyponatremia (serum sodium �135mmol/L) to two large university teaching hospitals inBoston. Hyponatremia was present on admission in13% of those hospitalized for at least 2 days. Womenwith and without hyponatremia accounted for approx-imately half of all admissions overall, and two thirdsof all patients had severe hyponatremia. Patients withhyponatremia on admission were more likely to havecongestive heart failure, sepsis, pneumonia, metastaticdisease, and volume depletion than patients with nor-monatremia, and patients who were admitted withhyponatremia had more comorbidity (1.9 versus 1.4;P � 0.001) as measured by the Deyo modification ofthe Charlson index (D-CI), the sum of the weighted

0.20

0.10

0.05

0.15

110 115 120 125 130 135 140 145Admission Serum Sodium Concentration, mEq/L

Pred

icte

d Pr

obab

ility

of I

n-Ho

spita

l Mor

talit

yFigure 11. A statistical analysis (restricted cubic spline)relating mortality to admission serum sodium concentrationin a University hospital. Reprinted with permission fromreference 3 (Wald R, Jaber BL, Price LL, Upadhyay A,Madias NE: Impact of hospital-associated hyponatremia onselected outcomes. Arch Intern Med 170: 294–302, 2010).


number of comorbid conditions on the basis of 17diagnostic categories identified from InternationalClassification of Diseases, Ninth Revision, ClinicalModification diagnosis codes. Even mild hyponatre-mia (serum sodium concentration 130 to 134 mmol/L)carried a significantly increased risk for death in hos-pital, 1 year after discharge, and 5 years after dis-charge. Although there was a small difference in theabsolute prevalence of hyponatremia (12%) when cor-rected for hyperglycemia, there was no effect on theassociation between hyponatremia and mortality. Dif-ferences in mortality persisted after multivariable ad-justment in all categories of hyponatremia except forserum sodium concentration of �120 mmol/L. Com-pared with patients with normonatremia, the multivari-able-adjusted risk for death after 5 years after admis-sion was increased by 24% in patients with admissionserum sodium concentration of 130 to 134 mmol/L(P � 0.001), 33% in those with serum sodium con-centration of 125 to 129 mmol/L (P � 0.001), 29% inthose with serum sodium concentration of 120 to 124mmol/L (P � 0.003), and 9% in those with admissionserum sodium concentration of �120 mmol/L (P �0.52). Hyponatremia resolved in 3794 (7.2%), per-sisted in 4524 (8.6%), and was acquired during hos-pitalization in 1974 (3.8%). Mortality (in hospital, 1year after discharge, and 5 year after discharge) washighest in those with persistent or acquired hyponatre-mia, lower in those with hyponatremia that resolved,and lowest in those with normonatremia at both firstand last serum sodium measurements. The multivari-able-adjusted odds of in-hospital mortality associatedwith hyponatremia differed across clinical subgroups.For example, hyponatremia was not associated with anincreased risk for death compared with normonatremiain patients who were admitted with pneumonia, sepsis,liver disease, or medical admissions related to therespiratory system. The risk for death associated withhyponatremia seemed to be particularly strong in pa-tients with cardiovascular disease or cancer and thosewho underwent orthopedic procedures (P � 0.001).

None of these studies can determine whether acausal link exists between hyponatremia and mortal-ity. Severe, acute hyponatremia can lead to life-threat-ening cerebral edema, but the reasons for increasedmortality in less severe hyponatremia are less clear.Hyponatremia can be a marker of severity of illness,and although it may predict mortality, it may not bethe cause of death. Important physiologic derange-

ments of severe cardiovascular disease (e.g., low ef-fective circulating volume, decreased GFR, activationof neurohormonal responses) may be responsible forboth hyponatremia and an adverse outcome. Alterna-tively, the association between mild hyponatremia andmortality may reflect as-yet-unidentified adverse ef-fects on organ systems other than the central nervoussystem. Because the reasons for excess mortality as-sociated with hyponatremia are not well understood, itis premature to conclude that therapeutic interventions(e.g., water restriction, loop diuretics, the newly intro-duced and extremely expensive vasopressin 2 receptorantagonists) will alter outcomes. Because hyponatre-mia may not be associated with mortality in sepsis andpneumonia, the benefit of correcting the serum sodiumconcentration in those populations would be expectedto be limited. In contrast, the strong association withmortality in patients with heart disease or cancer andin patients who underwent orthopedic procedures sug-gests that these groups could be targeted to examinethe potential benefit of strategies to normalize serumsodium concentrations.

Hospital CostsA retrospective cohort study in a large academic

teaching hospital in New York City compared patientswith moderate to severe hyponatremia (serum sodiumlevel of �129 mmol/L at admission, n � 547) andpatients with mild to moderate hyponatremia (serumsodium level of 130 to 134 mmol/L, n � 1500) with7573 patients who had the same principal admittingdiagnoses and a serum sodium level of 135 to 145mmol/L (5). Patients who were admitted with hypo-natremia had significantly longer hospital LOS thanthose who were admitted without hyponatremia(median LOS: moderate to severe hyponatremia 8days, mild to moderate hyponatremia 8 days, normal6 days; P � 0.001). Patients with more severehyponatremia were also more likely to be admittedto the intensive care unit during the hospital stay(moderate to severe hyponatremia 32%, mild tomoderate hyponatremia 26%, normal 22%; P �0.001). These trends were also reflected in the totalcosts per admission, with median costs of $16,606for moderate to severe hyponatremia cases, $14,266for mild to moderate hyponatremia cases, and$13,066 for normal admissions (P � 0.001). Theauthors speculated that interventions or pharmaco-therapies for the prompt treatment of hyponatremia


could potentially reduce morbidity and LOS,thereby reducing the use of health care resources.

Falls and FracturesMild chronic hyponatremia is associated with

gait and attention deficits and with an increased riskfor falls (6–11). Asymptomatic hyponatremia seemedto cause more attention deficits than a serum alcohollevel of 0.6 g/L in an age- and gender-matched controlgroup; in a preliminary study, the threshold for gaitand attention deficits induced by hyponatremia were134 and 132 mmol/L, respectively.

A case-control study of 513 ambulatory patientswho were aged �65 years and presented to a generaluniversity hospital with bone fracture after fallingfound a 13.06% prevalence of hyponatremia (serumsodium �135 mmol/L) as compared with a 3.90%prevalence of hyponatremia among ambulatory age-and gender-matched control subjects who were ran-domly selected from ambulatory patients without ahistory of bone fracture (unadjusted odds ratio [OR]3.47 [95% CI 2.09 to 5.79]; adjusted OR 4.16 [95% CI2.24 to 7.71]) (10). Hyponatremia was mild andasymptomatic in all patients (mean serum sodium 131mmol/L) and was either drug induced (36% diuretics,17% selective serotonin reuptake inhibitors [SSRIs])or due to idiopathic syndrome of inappropriate antidi-uretic hormone secretion (SIADH; 37%). These find-ings were confirmed in a study of 364 cases of patientsaged 65 yr or older presenting to an urban emergencyroom with fractures of the hip, pelvis or femur; offracture patients with hyponatremia, 24.2% were tak-ing antidepressants, mostly selective serotonin recep-tor inhibitors (SSRIs), whereas none were taking thesemedications in the group without fracture (11). Simi-larly, a retrospective study of self-reported fracturesamong 1408 consecutive female patients who hadavailable laboratory data and underwent bone min-eral density (BMD) measurement found that hypo-natremia (serum sodium concentration �135mmol/L) was present in 8.7% of patients with aconfirmed fracture, whereas the prevalence of hy-ponatremia was 3.2% among patients without afracture (P � 0.001) (7). Multivariate logistic re-gression analysis controlling for age, T score,chronic kidney disease stage, and osteoporotic riskfactors and treatments found that a serum sodiumconcentration �135 mmol/L remained significantly

and independently associated with fracture occur-rence (P � 0.01).

Thiazide use has been associated with an increasedrisk for fractures in elderly patients living in nursinghomes, a paradoxic effect considering the beneficialeffect of thiazides on mineral balance; however, it is notknown whether the fracture risk associated with thiazidescan be attributed to thiazide-induced hyponatremia. Sim-ilarly, SSRIs in the treatment of depression are associatedwith falls and fractures with peak fracture risk occurringin the first 2 weeks after initiation of therapy, similar tothe time of onset of SSRI-induced hyponatremia; how-ever, it is not yet known whether the fall and fracture riskassociated with SSRIs can be attributed to SSRI-inducedhyponatremia.

Compounding the risk of falling attributable tohyponatremia, there is also evidence that hyponatre-mia is associated with osteoporosis (8). Approxi-mately one third of total body sodium resides in thebone, with 40% of bone sodium being exchangeablewith the extracellular sodium pool; therefore, long-term sodium depletion could theoretically lead to so-dium loss from the bone with consequent bone demin-eralization. Rats made hyponatremic (serum sodium110 � 2 mmol/L) for 3 months have a 30% reductionin BMD, as measured by dual-energy x-ray absorpti-ometry, as compared with controls. Reductions in bothtrabecular and cortical bone contents and an increasein the number of osteoclasts per bone area were found,as was a decreased serum concentration of osteocalcin,indicative of increased bone resorption and decreasedbone formation. The 30% reduction in BMD found inhyponatremic rats is approximately twice that reportedin various well-established rat osteoporosis modelsover similar periods of time using similar densitometrymethods. Data obtained from humans who were aged�50 years and had much milder hyponatremia areconsistent with the experimental data in severely hy-ponatremic rodents. The Third National Health andNutritional Examination Survey (NHANES III) pro-vides information on sodium concentrations and BMDof the hip in a nationally representative sample of USadults. An analysis of data from NHANES III foundthe adjusted OR for developing osteoporosis to benearly three times higher among adults with mildhyponatremia (SNa 133.0 � 0.2 mmol/L) comparedwith those without hyponatremia (OR 2.85; 95% CI1.03 to 7.86; P � 0.01). There was also a positivelinear association between SNa and femoral neck


BMD in patients with hyponatremia. Diuretics wereused by 11.1 and 6.8% of patients with hyponatre-mia and normonatremia, respectively. Of these, thi-azide diuretics were used by 10.5% of patients withhyponatremia and 4.7% of patients with normona-tremia.

A 30% reduction in BMD found in hypona-tremic rats and data obtained from hu-mans aged >50 years revealed the ad-justed OR for developing osteoporosis tobe nearly three times higher among adultswith mild hyponatremia (SNa 133.0 � 0.2mmol/L) compared with those without hy-ponatremia.

RhabdomyolysisSelf-induced water intoxication presents with

brain edema, coma, convulsions, and death as a resultof severe hyponatremia. Non-neurologic symptomssuch as rhabdomyolysis have been reported in singlecase reviews. A retrospective review of 22 patientswho were treated in a single medical referral center inJapan identified six patients with rhabdomyolysis(Creatine phosphokinase level 12,138 to 319,400) and16 patients without rhabdomyolysis (12); two patientswith rhabdomyolysis developed renal failure (creati-nine 2.7 and 6.9 mg/dl), and one patient developedextrapontine myelinolysis. The severity of hyponatre-mia in the two groups was not different, and there wasno association with convulsions, drugs, or alcohol, butthe rate of correction (attributable primarily to theexcretion of large quantities of dilute urine in theinitial few hours after presentation) was significantlymore rapid in the group with rhabdomyolysis (2.0 �1.3 versus 0.9 � 0.7 mmol/L per h [P � 0.017];21.3 � 6.0 versus 10.0 � 4.6 mmol/L per 24 h [P �0.001]).

Osmotic Demyelination SyndromeAn adaptive loss of organic osmolytes permits

brain cells to maintain osmotic equality with plasmawithout a large increase in brain water. Although thisadaptation permits survival with extremely low serumsodium concentrations, it also makes the brain vulnerableto injury from rapid correction of hyponatremia. Rapidcorrection of chronic hyponatremia is analogous to anacute hyperosmolar insult, and because the adaptation to

chronic hyponatremia restores brain volume to near nor-mal, rapidly increasing the serum sodium concentrationto normal after this adaptation occurs dehydrates thebrain just as acute hypernatremia dehydrates the brain inan individuals with normonatremia. Indeed, severe hy-pernatremia can cause osmotic demyelination in patientswho were never known to have hyponatremia (13).Excessive correction of chronic hyponatremia triggers acascade of injury in the brain beginning with breakdownof the blood-brain barrier and culminating in the pro-grammed death of oligodendrocytes, the cells that makemyelin in the central nervous system (14,15).

Studies in experimental animals have conclusivelydemonstrated that rapid correction of hyponatremia andnot hyponatremia itself is the cause of osmotic demyeli-nation (14,15). Consistent with this conclusion, it hasbeen shown that re-lowering the serum sodium concen-tration after excessive correction prevents demyelinatingbrain lesions and death in the rat (16). Re-induction ofhyponatremia also prevents the opening of the blood-brain barrier that occurs when hyponatremia is rapidlycorrected; but although treatment with dexamethasonealso prevents opening of the barrier, only re-induction ofhyponatremia resulted in a significant decrease in mor-tality (16). Similar to findings in the rat, in three singlecase reports, re-lowering of the serum sodium concen-tration was shown to reverse early findings of osmoticdemyelination after overcorrection of hyponatremia(17,18).

Slow recovery of organic osmolytes lost duringthe adaptation to a low plasma osmolality seems toplay a pivotal role in the pathogenesis of osmoticdemyelination. Brain regions that are slowest to re-cover organic osmolytes are the most severely injuredafter rapid correction of hyponatremia; uremia, whichreduces the incidence and severity of demyelinatingbrain lesions after rapid correction of hyponatremia, isassociated with more rapid reuptake of organic os-molytes, particularly myoinositol, and exogenous ad-ministration of myoinositol decreases the severity ofinjury caused by rapid correction (14). Myoinositolprotects glial cells from cytotoxicity induced by os-motic shrinkage (19). In cultured astrocytes grown inhypertonic medium, cell survival is reduced and canbe partially restored by adding myoinositol, but notother organic osmolytes, to the medium.

Central pontine myelinolysis (CPM), the mostclassic manifestation of brain injury caused by rapidcorrection of hyponatremia, was first described in


alcoholics in 1959, before measurements of the serumsodium concentration became routine. Shortly afterthe description of CPM, a symmetrical demyelinatinginjury in the center of the pons that spares neurons,similar lesions outside the pons (extrapontine my-elinolysis [EPM]) were described. Norenberg (15)recently published a compelling account of how CPMand EPM (also known as osmotic demyelination)came to be associated with excessive treatment ofhyponatremia. Often called a “rare” disorder, the num-ber of reported cases calls this characterization intoquestion. After the first reports in the 1980s, severalhundred cases in which CPM and EPM complicatedthe treatment of severe hyponatremia have been re-ported, and several new cases appear each year: Be-tween January 1, 2008, and August 31, 2010, morethan a dozen single case reports, a case series involv-ing adults (20–35), and three case reports involvingchildren (36–38) were published. A 12-year study at asingle university medical center in Canada identified12 patients with CPM and/or EPM documented bymagnetic resonance imaging (MRI) or autopsy (23);six of the 12 patients were known to have developedosmotic demyelination after rapid correction of hypo-natremia (including two who had diabetes insipidusand developed hyponatremia on desmopressin andexperienced a large water diuresis when desmopressinwas discontinued), and one case occurred in a patientwho had normonatremia and developed hypernatremia(to 176 mmol/L over 3 hours) because of a hemodial-ysis error. The five patients without a known rapidincrease in serum sodium concentration included threewho developed CPM/EPM after liver transplantationand two alcoholics for whom data were unavailable.Most patients with hyponatremia presented with theclassic biphasic course characterized by nonspecificsymptoms including confusion associated with severehyponatremia at presentation, followed by a relative im-provement lasting 2 to 3 days after correction of theelectrolyte disturbance, and then a second progressiveneurologic deterioration that led to a diagnosis ofCPM/EPM after a mean delay of 8 days. The neuro-logic symptoms that developed after overly rapid cor-rection were varied and included rigidity; cardiorespi-ratory autonomic dysfunction; seizures (generalizedtonic-clonic and partial complex); neuropsychiatricissues (catatonia, emotional lability); altered con-sciousness; and pyramidal, brainstem, and cerebellarsigns. As has been noted in previous reports, neuro-

imaging was initially normal with lesions appearingwith a delay of up to 14 to 21 days and with initialsparing of the pons.

Some of the extrapontine lesions seen after rapidcorrection of hyponatremia, such as cortical laminarnecrosis, are also seen in patients with hypoxic braindamage (35). This has led some investigators to attri-bute neurologic sequelae in chronic hyponatremia tohypoxia. The experimental models used to support thishypothesis have been criticized on methodologicgrounds, and they are difficult to reconcile with theobservation that therapeutic re-lowering of the serumsodium concentration prevents brain lesions (18). Fur-thermore, “hypoxic” brain lesions, such as corticallaminar necrosis, have been reported in patients with-out a hypoxic insult, having undergone rapid correc-tion of stable chronic hyponatremia (from 105 to 140mmol/L in 24 hours in one case) (35). Recent casereports of CPM/EPM after rapid treatment of hypona-tremia have included data on oxygen saturation onroom air (99% on room air in two cases [39,40] and aPO2 of 87 mmHg on room air in a third [40]) that makea diagnosis of hypoxic brain damage untenable.

There have been several reports of CPM/EPMafter orthotopic liver transplantation. A review of 1247patients who received a transplant in Korea identified11 (0.88%) patients with CPM/EPM (41). Risk factorswere compared with 44 control subjects without CPM/EPM, matched by age, gender, and date of operation.Preoperative serum sodium was significantly lower inthe CPM/EPM group than in control subjects (126 �7 versus 134 � 7 mmol/L; P � 0.001), and patientswho developed the disease had a significantly largerperioperative change in serum sodium concentration(15 � 6 versus 7 � 4 mmol/L; P � 0.001). Signifi-cantly larger amounts of blood products and crystalloidswere given in the CPM/EPM group, and a significantcorrelation was found between the volume administeredand the increase in serum sodium during surgery. Con-sistent with these observations, three patients with CPMidentified after liver transplantation over 3 years in asingle center in China each had preoperative hyponatre-mia (119 to 124 mmol/L), and each underwent correctionby �20 mmol/L during surgery (42).

In most patients with severe hyponatremia, neuro-logic findings suggestive of the osmotic demyelinationsyndrome emerge after correction by �10 mmol/L per24 h and/or �18 mmol/L per 48 h. However, there areoccasional reports of osmotic demyelination complicat-


ing severe hyponatremia after correction rates belowthese limits (43). There are several reports of CPMalcoholics with mild hyponatremia and occasionally nor-monatremia and with liver disease after correction by�10 mmol/L per d (44). In another report, MRI findingsconsistent with a diagnosis of EPM were identified in aseverely ill patient with Addison disease and miliarytuberculosis associated with a serum sodium of 118mmol/L that was corrected very gradually (45). How-ever, the title of some case reports is misleading: In areport entitled “Central Pontine Myelinolysis DespiteSlow Sodium Rise,” an otherwise healthy woman devel-oped MRI-documented CPM and EPM with classic butreversible symptoms after treatment of thiazide-inducedhyponatremia (serum sodium 96, serum potassium 2.5mmol/L) (39). A review of the data showed that theserum sodium increased by 18 mmol/L in �24 hours and24 mmol/L within 48 hours (39).

References1. Upadhyay A, Jaber BL, Madias NE: Epidemiology of hyponatremia.

Semin Nephrol 29: 227–238, 20092. Mohammed AA, van Kimmenade RR, Richards M, Bayes-Genis A,

Pinto Y, Moore SA, Januzzi JL Jr: Hyponatremia, natriuretic pep-tides, and outcomes in acutely decompensated heart failure: Resultsfrom the International Collaborative of NT-proBNP Study. CircHeart Fail 3: 354–361, 2010

3. Wald R, Jaber BL, Price LL, Upadhyay A, Madias NE: Impact ofhospital-associated hyponatremia on selected outcomes. Arch InternMed 170: 294–302, 2010

4. Waikar SS, Mount DB, Curhan GC: Mortality after hospitalizationwith mild, moderate, and severe hyponatremia. Am J Med 122:857–865, 2009

5. Callahan MA, Do HT, Caplan DW, Yoon-Flannery K: Economicimpact of hyponatremia in hospitalized patients: a retrospectivecohort study. Postgrad Med 121: 186–191, 2009

6. Ayus JC, Moritz ML: Bone disease as a new complication ofhyponatremia: Moving beyond brain injury. Clin J Am Soc Nephrol5: 167–168, 2010

7. Kinsella S, Moran S, Sullivan MO, Molloy MGM, Eustace JA:Hyponatremia independent of osteoporosis is associated with fractureoccurrence. Clin J Am Soc Nephrol 5: 275–280, 2010

8. Verbalis JG, Barsony J, Sugimura Y, Tian Y, Adams DJ, Carter EA,Resnick HE: Hyponatremia-induced osteoporosis. J Bone Miner Res25: 554–563, 2010

9. Sterns RH, Hix JK, Silver S: Treatment of hyponatremia. Curr OpinNephrol Hypertens 19: 493–498, 2010

10. Gankam Kengne F, Andres C, Sattar L, Melot C, Decaux G: Mildhyponatremia and risk of fracture in the ambulatory elderly. QJM101: 583–588, 2008

11. Sandhu HS, Gilles E, DeVita MV, Panagopoulos G, Michelis MF:Hyponatremia associated with large-bone fracture in elderly patients.Int Urol Nephrol 41: 733–737, 2009

12. Morita S, Inokuchi S, Yamamoto R, Inoue S, Tamura K, Ohama S,Nakagawa Y, Yamamoto I: Risk factors for rhabdomyolysis inself-induced water intoxication (SIWI) patients. J Emerg Med 38:293–296, 2010

13. Naik KR, Saroja AO: Seasonal postpartum hypernatremic encepha-

lopathy with osmotic extrapontine myelinolysis and rhabdomyolysis.J Neurol Sci 291: 5–11, 2010


15. Norenberg MD: Central pontine myelinolysis: Historical and mech-anistic considerations. Metab Brain Dis 25: 97–106, 2010

16. Gankam Kengne F, Soupart A, Pochet R, Brion JP, Decaux G:Re-induction of hyponatremia after rapid overcorrection of hypo-natremia reduces mortality in rats. Kidney Int 76: 614 – 621, 2009


18. Sterns RH, Hix JK: Overcorrection of hyponatremia is a medicalemergency. Kidney Int 76: 587–589, 2009

19. Hijab S, Havalad S, Snyder AK: The role of organic osmolytes in theresponse of cultured astrocytes to hyperosmolarity. Am J Ther Jan-uary 21, 2010 [epub ahead of print]

20. Lin CM, Po HL: Extrapontine myelinolysis after correction of hypo-natremia presenting as generalized tonic seizures. Am J Emerg Med26: 632 e5–e6, 2008

21. Wu YC, Peng GS, Cheng CA, Lin CC, Huang WS, Hsueh CJ, Lee JT:(99m)Tc-TRODAT-1 and (123)I-IBZM SPECT studies in a patientwith extrapontine myelinolysis with parkinsonian features. Ann NuclMed 23: 409–412, 2009

22. McNeill G, Halpenny D, Snow A, Geoghegan A, Torreggiani WC:Extrapontine myelinolysis after correction of hyponatraemia pre-senting as generalised tonic seizures. Am J Emerg Med 27: 243,2009

23. Odier C, Nguyen DK, Panisset M: Central pontine and extrapontinemyelinolysis: From epileptic and other manifestations to cognitiveprognosis. J Neurol 257: 1176–1180, 2010

24. Hawthorne KM, Compton CJ, Vaphiades MS, Roberson GH, KlineLB: Ocular motor and imaging abnormalities of midbrain dysfunc-tion in osmotic demyelination syndrome. J Neuroophthalmol 29:296–299, 2009

25. Biotti D, Durupt D: A trident in the brain, central pontine myelinol-ysis. Pract Neurol 9: 231–232, 2009

26. Shin HW, Song D, Sohn YH: Normal diffusion-weighted MR imag-ing predicts a good prognosis in extrapontine myelinolysis-inducedparkinsonism. Mov Disord 24: 1701–1703, 2009

27. Srimanee D, Bhidayasiri R, Phanthumchinda K: Extrapontine my-elinolysis in preoperative sellar region tumor: Report of two cases.J Med Assoc Thai 92: 1548–1553, 2009

28. Jayachandran NV, Chandrasekhara PK, Thomas J, Agrawal S, Nar-simulu G: Cyclophosphamide-associated complications: We need tobe aware of SIADH and central pontine myelinolysis. Rheumatology(Oxford) 48: 89–90, 2009

29. Patel AS, Matthews L, Bruce-Jones W: Central pontine myelinolysisas a complication of refeeding syndrome in a patient with anorexianervosa. J Neuropsychiatry Clin Neurosci 20: 371–373, 2008

30. Sharma VK, Rathakrishnan R: Delayed and extra-pontine myelinolysisafter osmotic dysregulation. Ann Acad Med Singapore 37: 623, 2008

31. Tuleja E, Chermann JF, Sereni C, Hart G, Sereni D: Kluver Bucysyndrome, unusual consequence of excessively rapid correction ofsevere hyponatremia. Presse Med 37: 975–977, 2008

32. Serrano-Castro PJ, Alonso-Verdegay G, Lopez-Martínez G, Arjona-Padillo A, Callejon JR, Olmedo VM, Guardado-Santervas P, Huete-Hurtado A, Olivares-Romero J, Naranjo Fernandez C: Possible caseof peripheral osmotic demyelination syndrome. J Neurol NeurosurgPsychiatry 79: 331–332, 2008

33. Funayama M, Hisamatsu T, Koreki A: Central pontine demyelinoly-sis following water intoxication in schizophrenia. Schizophr ResAugust 10, 2010 [epub ahead of print]

34. Howard SA, Barletta JA, Klufas RA, Saad A, De Girolami U: Bestcases from the AFIP: Osmotic demyelination syndrome. Radiograph-ics 29: 933–938, 2009

35. Roh JH, Kim JH, Oh K, Kim SG, Park KW, Kim BJ: Cortical laminar


necrosis caused by rapidly corrected hyponatremia. J Neuroimaging19: 185–187, 2009

36. Singh DK, Rastogi M, Husain M: Central pontine myelinolysis in apediatric head injury patient. Pediatr Neurosurg 46: 51–53, 2010

37. Tullu MS, Deshmukh I, Muranjan MN, Kher AS, Lahiri KR: Extra-pontine myelinolysis in a child with nephrotic syndrome. PediatrNeurol 43: 139–141, 2010

38. Tsutsumi S, Yasumoto Y, Ito M: Central pontine and extrapontinemyelinolysis in an infant associated with the treatment of craniophar-yngioma: Case report. Neurol Med Chir (Tokyo) 48: 351–354, 2008

39. Schuster M, Diekmann S, Klingebiel R, Volk T: Central pontinemyelinolysis despite slow sodium rise in a case of severe community-acquired hyponatraemia. Anaesth Intensive Care 37: 117–120, 2009


41. Lee EM, Kang JK, Yun SC, Kim KH, Kim SJ, Hwang KS, Lee SG: Riskfactors for central pontine and extrapontine myelinolysis followingorthotopic liver transplantation. Eur Neurol 62: 362–368, 2009

42. Zhang ZW, Kang Y, Deng LJ, Luo CX, Zhou Y, Xue XS, Wang D,Yin WH: Therapy of central pontine myelinolysis following livingdonor liver transplantation: Report of three cases. World J Gastro-enterol 15: 3960–3963, 2009

43. Pietrini V, Mozzani F, Crafa P, Sivelli R, Cademartiri F, Crisi G:Central pontine and extrapontine myelinolysis despite careful correc-tion of hyponatremia: Clinical and neuropathological findings of acase. Neurol Sci 31: 227–230, 2010

44. Orakzai RH, Orakzai SH, Hasley PB: Treating hyponatremia: Howslow is safe? Central pontine myelinolysis despite appropriate cor-rection of hyponatremia. Eur J Intern Med 19: e29–e31, 2008

45. Gujjar A, Al-Mamari A, Jacob PC, Jain R, Balkhair A, Al-Asmi A:Extrapontine myelinolysis as presenting manifestation of adrenalfailure: A case report. J Neurol Sci 290: 169–171, 2010

Treatment of Chronic HyponatremiaOsmotic demyelination syndrome can usually be

prevented by avoiding increases in the serum sodiumconcentration of �10 mmol/L in a single day and/or 18mmol/L over any 48-hour period; rather than therapeuticgoals, these are limits not to be exceeded. If a 4- to6-mmol/L increase in serum sodium concentration issignificant enough to improve the most severe symptomsin patients with acute hyponatremia, a therapeutic goal of6 mmol/L per d is reasonable in chronic hyponatremia,even when the serum sodium concentration falls to ex-tremely low levels (1). This can be translated to an easilyremembered “rule of sixes”: “Correction by 6 mmol/Lper d makes sense for safety, so correct by 6 mmol/L forsevere sx’s and stop.” In other words, for all patients withchronic hyponatremia, the correction goal is 6 mmol/Lduring the initial 24 hours, and for those with severesymptoms (seizure, severe delirium, and unresponsive-ness), the goal is preloaded in the first 6 hours, postpon-ing subsequent efforts to increase the serum sodium leveluntil the next day. As discussed previously, although theinitial rate of correction with this strategy is 1 mEq/L perh (6 mEq/L within 6 hours), the average rate over the

first 24 hours is 0.25 mEq/L per h (6 mEq/L per d), wellbelow the traditional limit of 12 mEq/L per d; unfortu-nately, the daily limit of 12 mEq/L is sometimes ex-pressed as 0.5 mEq/L per h—a formulation that manyclinicians find confusing because it can be misinterpretedas meaning that aggressive therapy for severe symptomsshould not be given.

Rule of sixes: “Correction by 6 mmol/L perd makes sense for safety, so correct by 6mmol/L for severe sx’s and stop.” In otherwords, for all patients with chronic hypo-natremia, the correction goal is 6 mmol/Lduring the initial 24 hours; for those withsevere symptoms (seizure, severe delir-ium, and unresponsiveness), the goal ispreloaded in the first 6 hours, postponingsubsequent efforts to increase the serumsodium level until the next day.

It is a good idea to set correction goals wellbelow rates that are known to result in iatrogenicinjury because it is quite easy to “overshoot the mark.”In many patients, the cause of water retention isreversible (e.g., vasopressin release secondary to vol-ume depletion), and when the cause is corrected (e.g.,volume repletion), vasopressin levels fall and the ensu-ing water diuresis increases the serum sodium concen-tration by as much as 2 mmol/L per h (2). Overcorrec-tion can be prevented or reversed by administeringdesmopressin to terminate a water diuresis; this strat-egy has been used successfully in a small series ofpatients (3). Alternatively, desmopressin can be givenbefore a water diuresis even begins with a concurrentslow infusion of hypertonic saline that is titrated tomaintain a correction rate of 6 mmol/L per d (1). Des-mopressin is administered every 6 to 8 hours to keep theurine concentrated, eliminating one of the variables thatincrease the serum sodium concentration. Desmopressinis given less frequently in patients with diabetes insipidus(DI) because the goals are different: In DI, dosing inter-vals of �12 hours allow some escape from water reten-tion to avoid hyponatremia, whereas in patients who havehyponatremia and no longer have a reason to retainwater, the goal is to prevent completely the excretion offree water until the serum sodium concentration has beenraised with hypertonic saline to a level closer to normal.


Alternatively, to eliminate the need for continuous infu-sion of 3% saline, desmopressin doses can be periodi-cally withheld each day to allow a brief period of waterexcretion to raise the serum sodium by the desired 6mmol/L; however, because the timing of water diuresiswhen desmopressin is discontinued is unpredictable, thisrequires much more rigorous hourly monitoring of thepatient to avoid unintentional overcorrection. A descrip-tion of this strategy was recently published in a casereport of a patient who survived without neurologicsequelae after presenting with a serum sodium of 96mmol/L (Figure 12) (1).

Overcorrection of hyponatremia can beprevented or reversed by administeringdesmopressin to terminate a water diure-sis. Alternatively, desmopressin can begiven before a water diuresis even beginswith a concurrent slow infusion of hyper-tonic saline that is titrated to maintain acorrection rate of 6 mmol/L per d. Desmo-pressin is administered every 6 to 8 hoursto keep the urine concentrated, eliminat-ing one of the variables that increase theserum sodium concentration.

Potassium (K) depletion makes patients withchronically hyponatremia especially vulnerable toovercorrection. The serum sodium concentration is afunction of the ratio of exchangeable body sodiumplus K divided by total body water. Therefore, admin-istration of K (increasing the numerator of the equa-tion) increases the serum sodium concentration. Arecent report described a mildly symptomatic patient(muscle weakness and slurred speech) with a serumsodium of 96 mmol/L and a serum K of 1.6 mmol/L;the patient developed osmotic demyelination afterovercorrection that could be attributed primarily toreplacement of her large K deficit (4). K replacementwas also responsible for a small but potentially dan-gerous increase in serum sodium concentration abovethe goal for the aforementioned patient treated withboth hypertonic saline and desmopressin (1). K chlo-ride can be given in a hypertonic solution containing20 mmol of K in 100 ml of water (400 mmol/L); thesolution is nearly as concentrated as 3% saline (513mmol/L); therefore, the rate of 3% saline administra-tion should be slowed when K is being given to avoidunintentional overcorrection.

Renal Replacement TherapyPatients with acute kidney injury may have con-

comitant severe hyponatremia. To avoid rapid correc-tion during renal replacement therapy, continuousvenovenous hemofiltration can be used with adjust-ment of the replacement fluid sodium concentration byadding sterile water (5). Stepwise correction of thepatient’s serum sodium concentration can be plannedusing replacement fluid made up to raise the serumsodium concentration successively. The concentrationof bicarbonate and K in the final solution will also bereduced, and additional sodium and/or bicarbonatesupplementation may be needed.

Vasopressin Antagonists (Vaptans)Once vasopressin is released into the circulation,

its biologic effects are mediated by three receptorsubtypes: V1A, V1B, and V2. V1A receptors arelocated on vascular smooth muscle, platelets, and theliver. By increasing intracellular calcium, activation ofthe V1A receptor results in vasoconstriction, plateletaggregation, and gluconeogenesis. V1B receptors arelocated in the anterior pituitary, and their activationresults in stimulation of ACTH release. V2 receptors(V2Rs) are located in the principal cells of the renal

Figure 12. Patient with profound hyponatremia treatedsuccessfully with concurrent administration of 3% NaCl anddesmopressin. Targeted (ideal) correction shown in redtriangles. Actual serum Na values shown in blue circles.Reprinted from reference 3 (Sterns RH, Hix JK, Silver S:Treating profound hyponatremia: A strategy for controlledcorrection. Am J Kidney Dis 56: 774–779, 2010), withpermission of W.B./Saunders Co.


collecting duct; when vasopressin binds to the V2Rs,they activate a cAMP-mediated signal transductionpathway that triggers an increased rate of insertion ofaquaporin channels in the apical membrane of therenal collecting duct, thereby increasing reabsorptionof free water along osmotic gradients, increasing urineosmolality, and producing an antidiuresis. Conversely,when vasopressin secretion is suppressed and the V2Ris unoccupied, dilute urine is excreted. Antagonists tothe vasopressin V2R, called “vaptans” because of thesuffix applied to their generic names, block the bind-ing of vasopressin to renal V2Rs, preventing vasopres-sin-mediated water reabsorption, decreasing urine os-molality, and increasing urine output. Unlike diureticagents, which increase the excretion of sodium andwater, vaptans enhance free water excretion withoutincreasing the excretion of sodium or K; therefore, theincrease in urine output caused by vaptans has beencalled an “aquaresis” and the vaptans have been called“aquaretic” agents (6).

Elegant experiments performed with V2R ago-nists and antagonists in awake, unrestrained ratsshowed that V2R-mediated effects of vasopressin arenot only antidiuretic but also antinatriuretic (7). Thesefindings are consistent with observations made inisolated collecting duct or in various mammalian oramphibian cell culture models. In these tissues, appli-cation of vasopressin to the basolateral side of themembrane increases amiloride-sensitive sodium trans-port, an effect mediated by the epithelial sodiumchannel. In healthy humans and individuals with neph-rogenic DI caused by mutations of aquaporin 2, admin-istration of the V2R agonist desmopressin (dDAVP)induces a twofold reduction in the rate of sodium excre-tion; no change in sodium excretion occurred whendDAVP was given to subjects with mutations of theV2R. Most studies describing the effects of selectivenonpeptide V2R antagonists in experimental animalsor humans concluded that these drugs behave as pure“aquaretics” with no effect on electrolyte excretion.The natriuretic effect of aquaretics is not readily ap-parent because compensatory sodium retention occursafter the initial loss. In rats given a V2R antagonist,24-hour sodium excretion is unchanged, but threefoldincreases in sodium excretion rate occur in the first 4to 6 hours. In humans, V2R antagonists have beenshown to increase sodium excretion significantly inpatients with hyponatremia and with cirrhosis andascites (an effect that can also be shown in rats with

experimental cirrhosis) but not in patients with thesyndrome of inappropriate secretion of antidiuretichormone (SIADH). Although vaptans increase sodiumexcretion to some extent, their natriuretic effect re-mains far smaller than their aquaretic effect; in the rat,the natriuresis produced by furosemide is sevenfoldgreater than that produced by a V2R antagonist atcomparable rates of urine flow.

In the rat, both V2R agonists and antagonistsincrease K excretion rate. Stimulation of V2R isknown to increase K secretion by the collecting duct,and this in vitro effect is mirrored by a dosage-dependent increase in transtubular K gradient (TTKG)seen in rats receiving dDAVP. In addition, large in-creases in urine flow rate increase K excretion by aflow-dependent mechanism; therefore, administrationof a V2R antagonist increased urinary excretion with-out changing the TTKG, an estimate of K secretion.

High dosages of vasopressin have been shown tobe natriuretic in rats, dogs, sheep, and humans. In the rat,this dosage-dependent natriuretic effect is largely abol-ished by the co-administration of a selective V1aR an-tagonist. Natriuresis associated with vasopressin contrib-utes to hyponatremia in SIADH, and it has beenattributed to a physiologic response to volume expansioncaused by water retention. In the rat, at least, this V1a-mediated effect is independent of volume expansion.Possible mechanisms include a pressure natriuresis re-sulting from the vasopressor effects of the hormone,either within the whole circulation or selectively withinthe kidney vasculature, and stimulation of prostaglan-dins, which increase sodium excretion by reducing so-dium transport in the collecting duct and by increasingmedullary blood flow. Vasopressin stimulates prosta-glandin production in interstitial medullary cells, whichexpress V1a receptors, and in cortical collecting ductcells, which express V1a receptors on their apical mem-branes; these apical membranes may respond to vaso-pressin in the collecting duct lumen, where levels of thehormone are higher than in peripheral blood.

The complex interplay of V2- and V1a-mediatedeffects on sodium excretion shown in the rat areillustrated in Figure 13 (7). The findings from theseexperiments have important implications. In patientswith hypervolemic hyponatremia as a result of heartfailure or cirrhosis, the additional effect of V2 antag-onism on sodium excretion may be beneficial and maydecrease diuretic use. V2R antagonists may also bebeneficial in some forms of salt-sensitive hyperten-


sion; however, more precise information is requiredregarding V1a and V2 effects in humans to knowwhether selective rather than mixed antagonists aremore appropriate in each of these disorders.Conivaptan. Intravenous conivaptan is the firstvasopressin receptor antagonist to be approved by theUS Food and Drug Administration (FDA) for treatingeuvolemic hyponatremia in hospitalized patients. Thedrug was initially approved for treating euvolemichyponatremia caused by SIADH, hypothyroidism, ad-renal insufficiency, or pulmonary disorders, and sub-sequently the FDA approved its use for treating hy-pervolemic hyponatremia in patients with heartfailure. Conivaptan has high affinity for the humanvasopressin subtypes V1a and V2. The peak diureticeffect of convipatan occurs 2 to 4 hours after itsinfusion, and the aquaretic effect persists for approx-imately 12 hours. Conivaptan is metabolized by theliver by CYP3A4, and it can interact with other drugsthat are metabolized by this enzyme; for this reason,concurrent administration of conivaptan with potentCYP3A4 inhibitors is contraindicated (see later).

Conivaptan’s efficacy for the treatment of hyponatre-mia has been assessed in three randomized, double-blind, placebo-controlled clinical trials and one open-label, nonrandomized trial; however, only one of therandomized trials studied the intravenous formulationof the drug that was approved for use by the FDA(6,8). Infusion site reactions were the most commonreason for discontinuation during clinical trials; a newformulation of the drug that does not include propyl-ene glycol was approved in October 2008 and wasexpected to decrease the irritation of veins, althoughno actual studies were done to show that infusion-sitereactions were reduced by the formulation.

In addition to these formal clinical trials, a sin-gle-center, observational study of 18 patients whowere treated with intravenous conivaptan for SIADHprovides a “real world” experience with the drug (9).The study found that only 12 (66.7%) patients met thecriterion for successful response, defined as an abso-lute increase in serum sodium 4 mmol/L over baseline.A lower pretreatment serum sodium concentration wasassociated with significantly larger increases in serumsodium concentration; all six patients with serum so-dium concentrations �120 mmol/L were corrected by�10 mmol/L per 24 hours, which most investigatorsnow consider excessive.

As noted, conivaptan antagonizes V1a receptorsas well as V2 receptors. Because V1a agonists havebeen effective in treating hepatorenal syndrome, ad-ministration of conivaptan to patients with hyponatre-mia caused by hepatic cirrhosis has not been recom-mended and is not approved by the FDA. However, arecent study reported a favorable experience in 24patients who had end-stage liver disease and hypona-tremia (serum sodium �130 mmol/L) and were treatedwith conivaptan (10). Systolic BP and serum creati-nine did not increase significantly, and variceal hem-orrhage or worsening of ascites was not observed. Inall patients, correction was �6 mmol/L, and in twopatients, the increase was 1 mmol/L per 24 h; noneurologic complications were noted in careful fol-low-up.

An attempt to enroll patients who were treated inneurointensive care units in a controlled trial of con-ventional therapy versus conventional therapy plusconivaptan was terminated because of difficulty re-cruiting patients after only six patients had entered thetrial (11). However, a single-center retrospective studyreported on 22 patients who were treated for hypona-

Figure 13. Dosage-dependent effects of AVP on sodiumexcretion rate and dissociation of these effects into V2R-and V1aR-mediated responses. The abscissa represents in-creasing levels of plasma AVP from left (undetectable) toright. A, the lowest values of AVP that reduce urine flowrate but not sodium excretion rate. V2R antinatriuretic andV1aR natriuretic effects are depicted as sigmoid curves withdifferent thresholds (B for V2R and C for V1aR effects).B and C, maximum effects depending on each receptor type,respectively. M, value of plasma AVP for which the anti-natriuretic and the natriuretic effects compensate each other.This value probably fluctuates according to a number offactors influencing the intensity of the responses mediatedby each of the two receptor types. Reprinted from reference7, with permission.


tremia in a neurointensive care unit with conventionaldosing of conivaptan (12). Half of the patients hadhyponatremia after failure of conventional therapies.The serum sodium concentration increased by �6mmol/L in 19 (86%) of 22 patients, with an averagetime to goal of 13.1 hours; no patient was rapidlycorrected. However, in a single-patient case report, a24-year-old woman with hyponatremia (121 mmol/L)associated with a large pituitary macroadenoma expe-rienced an excessively rapid correction of serum so-dium after receiving approved dosages of intravenousconivaptan. After approximately 25 mg of the drug,her sodium increased by 16 mmol/L over 8.5 hours;fortunately, the patient experienced no adverse effects.

The approved dosing for conivaptan is a 20-mgbolus followed by a 20-mg/d continuous infusion over1 to 4 days, a regimen that may increase the risk forphlebitis and that necessitates administration in largeveins and changing the infusion site every 24 hours(6,12). Administration of a conivaptan bolus to healthyadults without a subsequent infusion increased urineoutput and decreased urine osmolality, an effect thatpeaked in 2 hours and persisted for 6 hours. Anuncontrolled study of bolus dosing of conivaptan withno subsequent infusion in 19 patients who had acuteeuvolemic hyponatremia and were treated in a neuro-intensive care unit showed this to be an effectivedosing strategy; the serum sodium increased by at least4 mmol/L within 12 hours in 59% of patients (13). Thedrug consistently led to increased urine output and asignificant drop in urine specific gravity, and no casesof phlebitis were observed despite administration ofconivaptan through peripheral intravenous lines. In asingle case report, a single 20-mg bolus of conivaptan(without a sustaining infusion) was given to a patientwith traumatic brain injury complicated by increasedintracranial pressure when his serum sodium concen-tration fell to 128 mmol/L (14); an aquaresis devel-oped 3 to 5 hours after the dose, with urine outputpeaking at 1 L/h, and by 8 hours, the serum sodiumhad risen to 146 mmol/L (an increase of 18 mmol/Lwithin 8 hours). Intracranial pressure fell from 11 to15 mmHg before conivaptan to 2 mmHg after 4 hoursand remained reduced after 8 hours.Tolvaptan. Tolvaptan is the first oral nonpeptideselective vasopressin V2R antagonist approved for usein the United States by the FDA. Tolvaptan has anaffinity for the V2R that is 29 times greater than thatfor the V1A receptor, and it has no appreciable affinity

for the V1B receptor. The Study of Ascending Levelsof Tolvaptan in Hyponatremia (SALT-1 and SALT-2)showed that tolvaptan increases the serum sodiumconcentration over the short term (�30 days) in pa-tients with euvolemic and hypervolemic hyponatre-mia, and a multicenter, open-label extension of thesetrials, called SALTWATER, documented the long-term effectiveness and safety of tolvaptan in 111patients who had hyponatremia and were treated for amean follow-up of 701 days (15). Responses werecomparable between patients with euvolemia andthose with heart failure but more modest in patientswith cirrhosis.

A recent review of the literature identified ninepublished trials that investigated the use of the vaso-pressin V2R antagonist tolvaptan as of February 2010(16), and meta-analysis publications through 2009 (8)analyzed outcomes. When used to treat hyponatremia,tolvaptan was associated with significantly greaterincrease in serum sodium concentrations comparedwith placebo on treatment days 4 (3.62 � 2.68 versus0.25 � 2.08 mmol/L; P � 0.001) and 30 (6.22 � 4.10versus 1.66 � 3.59 mmol/L; P � 0.001). When usedto treat patients with heart failure, tolvaptan at dosagesof 30, 60, and 90 mg/d was associated with meanweight changes of �1.80, �2.10, and �2.05 kg,respectively, versus �0.60 kg with placebo (P �0.002, P � 0.002, and P � 0.009), but it does not alterdisease progression or mortality. The most commonlyreported adverse events associated with tolvaptan inclinical trials were dry mouth (4.2 to 23.0%), thirst(7.7 to 40.3%), and polyuria (0.6 to 31.7%), all con-sistent with the mechanism of action of the drug. Theacademic pharmacists who authored this review alsoprovide an extensive discussion of the pharmacologyof tolvaptan. Tolvaptan is metabolized exclusively inthe liver with an estimated half-life of 12 hours.Aquaretic and sodium-increasing effects of the drughave been reported to begin within 2 to 4 hours (ascompared with the diuretic effect of conventional loopdiuretics, which act within 30 to 60 minutes), andapproximately 60% of the effect occurs within the first12 hours. Because tolvaptan is metabolized primarilyby the CYP3A isoenzyme, numerous drug–drug inter-actions are possible. Concurrent administration of ke-toconazole, a strong inhibitor of CYP3A, increasestolvaptan exposure by 82%. Higher dosages of keto-conazole and other strong inhibitors of CYP3A (e.g.,itraconazole, clarithromycin, saquinavir, ritonavir, ne-


fazodone) administered at maximum dosages wouldbe expected to result in even greater increases intolvaptan exposure. Therefore, although no adverseeffects have been reported from these interactions,concurrent administration with strong inhibitors ofCYP3A is best avoided. There are no data on theeffects of moderate inhibitors of CYP3A (e.g., vera-pamil, diltiazem, fluconazole, erythromycin), but onthe basis of known metabolism and pharmacokinetics,a considerable increase in tolvaptan exposure wouldbe expected. Conversely, rifampin, which induces theenzyme, decreases tolvaptan levels by more than sev-enfold and would be expected to diminish the effect oftolvaptan. Similar interactions would be expected withother CYP3A inducers (e.g., rifabutin, barbiturates,phenytoin, carbamazepine, St. John’s wort).

Tolvaptan was shown to be effective in correct-ing hyponatremia in a 30-day placebo-controlled trialof 19 patients with schizophrenia and stable asymp-tomatic idiopathic hyponatremia (mean serum sodium130.3 mmol/L) (17); seven patients received tolvaptan,and 12 received placebo. All patients had evidence ofimpaired water excretion with persistent hyponatremiadespite water restriction or laboratory evidence ofSIADH, with urine osmolality �100 mOsm/kg despitea serum sodium �130 mmol/L. During the initial 4days of therapy, patients who were given the studydrug received increasing dosages of tolvaptan escalat-ing from 15 to 60 mg/d as needed to correct hypona-tremia slowly. Fluids were unrestricted unless morn-ing weights were 7.5% higher than their regular dryweight (and this occurred in only one placebo patientand one control subject). Beginning at approximately8 hours and continuing until the drug was stopped,mean serum sodium concentrations during treatmentwith tolvaptan were significantly higher than placebo.One patient underwent a 12-mmol/L increase in serumsodium concentration (from 127 to 139 mmol/L) in thefirst day of therapy associated with hypotensive symp-toms, ataxia, and slurred speech but ultimately recov-ered; no data on urine osmolality were reported for anyof the patients. The results of this small study aresimilar to other studies of tolvaptan and other vaso-pressin antagonists in patients with SIADH or hypo-natremia caused by heart failure or hepatic cirrhosis.Unfortunately, the study leaves unanswered the impor-tant question as to whether vasopressin antagonistswould be useful in preventing episodes of symptom-atic hyponatremia in patients with schizophrenia and

polydipsia and with self-induced water intoxication.Epidemiologic studies show that 10 to 20% of patientswith schizophrenia have polydipsia (water intake �3L/d), and a small percentage of this population withpolydipsia develop acute water intoxication. There isconsiderable evidence that most patients with schizo-phrenia and polydipsia develop hyponatremia duringperiods when their water intake exceeds normallyfunctioning water excretion. Even with maximallydilute urine (approximately 50 mOsm/L), maximumwater output cannot exceed approximately 1 L/h whileexcreting a normal dietary solute load (approximately900 mOsm/d).

References1. Sterns RH, Hix JK, Silver S. Treating profound hyponatremia: A

strategy for controlled correction. Am J Kidney Dis 56: 774–779,2010


3. Perianayagam A, Sterns RH, Silver SM, Grieff M, Mayo R, Hix J,Kouides R: DDAVP is effective in preventing and reversing inad-vertent overcorrection of hyponatremia. Clin J Am Soc Nephrol 3:331–336, 2008


5. Ostermann M, Dickie H, Tovey L, Treacher D: Management ofsodium disorders during continuous haemofiltration. Crit Care 14:418, 2010

6. Li-Ng M, Verbalis JG: Conivaptan: Evidence supporting its thera-peutic use in hyponatremia. Core Evid 4: 83–92, 2010

7. Perucca J, Bichet DG, Bardoux P, Bouby N, Bankir L: Sodiumexcretion in response to vasopressin and selective vasopressin recep-tor antagonists. J Am Soc Nephrol 19: 1721–1731, 2008

8. Rozen-Zvi B, Yahav D, Gheorghiade M, Korzets A, Leibovici L,Gafter U: Vasopressin receptor antagonists for the treatment ofhyponatremia: Systematic review and meta-analysis. Am J KidneyDis 56: 325–337, 2010

9. Velez JC, Dopson SJ, Sanders DS, Delay TA, Arthur JM: Intravenousconivaptan for the treatment of hyponatraemia caused by the syn-drome of inappropriate secretion of antidiuretic hormone in hospital-ized patients: A single-centre experience. Nephrol Dial Transplant25: 1524–1531, 2010

10. O’Leary JG, Davis GL: Conivaptan increases serum sodium inhyponatremic patients with end-stage liver disease. Liver Transpl 15:1325–1329, 2009

11. Naidech AM, Paparello J, Leibling SM, Bassin SL, Levasseur K,Alberts MJ, Bernstein RA, Muro K: Use of conivaptan (Vaprisol) forhyponatremic neuro-ICU patients. Neurocrit Care 13: 57–61, 2010

12. Wright WL, Asbury WH, Gilmore JL, Samuels OB: Conivaptan forhyponatremia in the neurocritical care unit. Neurocrit Care 11: 6–13,2009

13. Murphy T, Dhar R, Diringer M: Conivaptan bolus dosing for thecorrection of hyponatremia in the neurointensive care unit. NeurocritCare 11: 14–19, 2009

14. Dhar R, Murphy-Human T: A bolus of conivaptan lowers intracranialpressure in a patient with hyponatremia after traumatic brain injury.Neurocrit Care May 4, 2010 [epub ahead of print]

15. Berl T, Quittnat-Pelletier F, Verbalis JG, Schrier RW, Bichet DG,Ouyang J, Czerwiec FS, SALTWATER Investigators: Oral tolvaptan


is safe and effective in chronic hyponatremia. J Am Soc Nephrol 21:705–712, 2010

16. Nemerovski C, Hutchinson DJ: Treatment of hypervolemic or euv-olemic hyponatremia associated with heart failure, cirrhosis, or thesyndrome of inappropriate antidiuretic hormone with tolvaptan: Aclinical review. Clin Ther 32: 1015–1032, 2010

17. Josiassen RC, Curtis J, Filmyer DM, Audino B, Skuban N, Shaugh-nessy RA: Tolvaptan: A new tool for the effective treatment ofhyponatremia in psychotic disorders. Expert Opin Pharmacother 11:637–648, 2010

Hypernatremia

PhysiologyMaintenance of a normal serum sodium concen-

tration depends on the ability to balance water intakeand water excretion. Thirst is the primary defenseagainst hypernatremia because water intake shouldnormally compensate for water losses if water is avail-able; a very readable and comprehensive review of thesubject was recently published (1).

The ability to concentrate urine is essential whenwater is scarce. The renal medulla produces a concen-trated urine by generating an osmotic gradient extend-ing from the corticomedullary boundary to the tip ofthe inner medulla. In the presence of the antidiuretichormone arginine vasopressin (AVP), water is reab-sorbed across the collecting ducts so that the luminalfluid achieves osmotic equilibrium with the surround-ing interstitium; the reabsorbed water is returned to thecirculation in the ascending vasa recta.

The gradient is classically attributed to counter-current multiplication of small transepithelial concen-tration differences, the so-called “single effect.” Ourunderstanding of the complex interactions among thenephron segments of the renal medulla and its vascu-lature that generate and maintain the concentrationgradient is still incomplete and was the subject ofrecent reviews (2–5). In the outer medulla, the gradientcan be attributed to active sodium chloride reabsorp-tion in the thick ascending limb. Because the thinascending limb is incapable of active transport, pas-sive mechanisms for the single effect were proposedfor the inner medulla in 1972, based on what wasunderstood about differences in water, urea, and so-dium permeabilities of the thin descending (waterpermeable but impermeable to sodium and urea) andthin ascending limbs (impermeable to water but highlypermeable to sodium, less so to urea). However, math-ematical models using original assumptions regardingpermeability differences failed to account for the gra-dient. Other models based on hydrostatic pressure

generated by peristaltic contractions of the pelvic wallhave been proposed. More recently, new findings haveemerged about the complex functional architecture ofthe renal medulla and the diverse permeability char-acteristics of the loop of Henle: Tubules are organizedaround tightly packed vascular bundles, and differ-ences in their anatomic arrangements in the outer andinner medulla may have functional significance(Figures 14 and 15); approximately 85% of descend-ing thin limbs lack aquaporin 1 (AQP1) and aretherefore water impermeable, whereas AQP1 is pres-ent in those nephrons with longer loops of Henle;loops of Henle in the inner medulla lack urea trans-porters and are therefore urea impermeable. Thesenew findings have prompted new computer simula-tions that are consistent with a passive model.

Urea plays an important role in the urine-concen-trating mechanism because protein deprivation impairsand urea administration restores urine-concentratingability. AVP phosphorylates the urea transporterUT-A1 via a cAMP-mediated process that, togetherwith an AVP-independent effect of hyperosmolality,results in insertion of UT-A1 in the luminal membraneof the papillary collecting duct, increasing its perme-ability to urea. During an antidiuresis, water is osmot-ically reabsorbed from the urea-impermeable parts ofthe outer medullary collecting ducts through AQP2water channels inserted by AVP. With the extractionof water, urea concentration increases in the collectingduct lumen, and urea is absorbed once it reaches thepapillary tip, which is urea permeable when AVP isacting; this absorption achieves high concentrations inthe medullary interstitium and contributes to the med-ullary concentration gradient. Gene knockout micehave been created for the terminal collecting duct ureatransporter UT-A1, whose activity on the luminalmembrane is acutely regulated by AVP; for UT-A3,located on the basolateral membrane of the terminalcollecting duct; for UT-A2, located on thin descendinglimbs of the loop of Henle; and for the descendingvasa recta isoform, UT-B (3). Studies of these knock-outs have enhanced our understanding of the role ofurea transport in water conservation. Mice lackingboth terminal collecting duct urea transporters UT-A1and UT-A3 have a normal ability to conserve waterwhen fed a low-protein diet; on a high-protein diet,which generates more urea, the inability of the A1/A3knockout to reabsorb urea results in polyuria becauseof a urea diuresis. However, the A1/A3 knockout is


able to generate a corticomedullary sodium chlorideconcentration gradient that is indistinguishable fromthat of wild-type mice. Although this seems to argueagainst the passive model for sodium reabsorption inthe thin ascending limb, which relies on accumulationof the urea in the interstitium of the inner medulla (3),recent mathematical modeling of these data concludedthat the findings in the A1/A3 knockout are consistentwith predictions of the passive mechanism (6). UT-A2knockout mice, in which urea secretion into the thindescending limb is impaired, do not have a concen-trating defect, but UT-B mice, which lack a ureatransporter on vasa recta, have a high plasma ureaconcentration and a “urea-selective” urine-concentrat-ing defect with a lower inner medullary urea concen-tration. Humans with genetic loss of the UT-B trans-porter (present on vasa recta) are unable to concentratetheir urine above 800 mOsm/kg.

Halperin et al. (4) proposed a novel interpreta-tion of the concentrating mechanism, based on thepremise that its function is not only to conserve waterbut also to avoid an excessively concentrated urinethat would result in stone formation; we only brieflysummarize this detailed quantitative analysis. The au-thors suggested that the amount of dilution of themedullary interstitium caused by AVP-stimulated re-

absorption of water in the medullary collecting ductcontrols the amount of sodium chloride absorbed inthe thick ascending limb (which then restores thesodium chloride concentration of the interstitium) sothat it transports the “right” amount of salt into theinterstitium, too much of which would result in anexcessively concentrated urine. The proposed regula-tor of this process is the calcium receptor on thebasolateral membrane of the ascending limb, which,when occupied, inhibits ROMK function and de-creases sodium chloride reabsorption. The entry ofwater from the collecting duct into the interstitiumlowers the calcium concentration in the interstitialfluid, releasing the ROMK from inhibition from anoccupied calcium receptor. In the inner medulla, dilu-tion of the interstitial sodium chloride concentration byreabsorption of an iso-osmolal urea solution in responseto AVP (urea acting as an ineffective osmole whenthe papillary urea transporters allow it to diffuse out ofthe lumen with water) creates the driving force forpassive salt reabsorption without water from the water-impermeable thin ascending limb; with the addition ofsodium and chloride derived from the thin ascendinglimb, the sodium concentration of the medullary intersti-tium returns nearly to the concentration that preceded itsdilution with urea and water from the collecting duct.

Figure 14. Schematic diagrams of tubular organization in the rat renal medulla. (A) Cross-section through the inner stripe ofouter medulla, where tubules seem to be organized around a vascular bundle. (B) Cross-section through the upper innermedulla, where tubules and vessels are organized around a collecting duct cluster. (Inset) Configuration of a collecting duct,ascending vasa recta (AVR), an ascending thin limb, and a nodal space. Copyright 2011 by American Physiological Society.Reproduced from reference 5 (Layton AT, Layton HE, Dantzler WH, Pannabecker TL: The mammalian urine concentratingmechanism: Hypotheses and uncertainties. Physiology (Bethesda) 24: 250–256, 2009), with permission of AmericanPhysiological Society.


Whereas countercurrent multiplication createsthe medullary concentration gradient and countercur-rent exchange in the vasa recta sustains it, AVP isultimately responsible for the excretion of a concen-trated urine. In response to very minor dehydration,enough to increase the plasma osmolality by as little as2 mOsm/kg, AVP is released from the posterior pitu-itary. AVP binds to vasopressin 2 receptors (V2Rs) inthe basolateral membrane of collecting duct principalcells, stimulating production of cAMP, which acti-vates protein kinase A and phosphorylates AQP2,resulting in the exocytic insertion of AQP2 waterchannels stored in intracellular vesicles into the lumi-nal membrane of the collecting duct. In response toAVP, water is reabsorbed across the collecting duct

and achieves osmotic equilibrium with the surround-ing hyperosmotic medullary interstitium.

Recently, AVP-independent mechanisms control-ling water homeostasis have been identified. The bio-logic effects of secretin and oxytocin in this process haverecently been reviewed (7). Oxytocin, which plays animportant role in reproduction, and AVP are closelyrelated peptide hormones; both are secreted from theposterior pituitary, and seven of their nine amino acidsare identical. Oxytocin exerts its antidiuretic effect bybinding to the AVP V2R but with lower affinity. Itsantidiuretic effect occurs primarily at supraphysiologicconcentrations that can be achieved when oxytocin isgiven at high dosages to induce labor. Secretin is agastrointestinal hormone that stimulates water and elec-trolyte secretion in the intestine, liver, and pancreas.Secretin receptors and secretin-sensitive adenylcyclasehave been identified in renal epithelia. Secretin increasesintracellular cAMP levels via binding to SCTR on thebasolateral membrane which is coupled to the adenylcy-clase VI through the G-protein GS. Similar to the actionof AVP, increased cAMP induced by secretin leads to theexocytic insertion of AQP2-bearing vesicles into theapical plasma membrane and, through unknown tran-scription factors, increases the protein and transcriptlevels of AQP2. Several lines of evidence support aphysiologic role for secretin in water homeostasis:Plasma secretin levels increase in water-deprived mice,secretin has an antidiuretic effect in the AVP-deficientBrattleboro rat, and mice genetically deficient in thesecretin receptor have a defect in water conservation. Theauthors speculated that dysfunction of secretin and itsreceptor are potential causes of nephrogenic diabetesinsipidus (DI) in humans and that secretin may be acandidate for treating X-linked nephrogenic DI. Studiesin cultured cell models have identified a host of addi-tional factors that influence the abundance of AQP2 inthe renal collecting duct, including insulin, calcium, pro-inflammatory factors, and aldosterone (8).

Essential HypernatremiaAppropriate responses to an increase in serum

sodium concentration are greater water intake drivenby thirst and reduced urinary water output caused bythe secretion of AVP. “Essential hypernatremia” ischaracterized by upward resetting of the osmotic setpoint for both thirst and AVP release, resulting inpersistent euvolemic hypernatremia. In most cases,structural abnormalities as a result of trauma, tumor, or

Figure 15. The inner medullary concentrating processesmay be accomplished in stages that can be associated withfour subsections, or zones: (1) An outermost zone (OZ1)just below the outer medulla, where loops expressing neg-ligible or no AQP1 have their bends; (2) a larger outer zone(OZ2), which contains well-organized collecting duct clus-ters where tubules and vessels are tightly packed and whereloops bend within the central portions of the clusters; (3) anouter inner zone (IZ1), where the organization of the col-lecting duct clusters is diminishing and all vasa recta arefenestrated; and (4) an innermost zone (IZ2), where collect-ing duct clusters can no longer be distinguished, where thecollecting ducts seem to dominate all other structure. Col-lecting duct clusters, which coalesce into single collectingducts, are shown in blue. AQP1-positive and AQP1-nega-tive descending thin limbs are shown in red and yellow,respectively. Prebend segments and ascending thin limbsare shown in green. Copyright 2011 by American Physiolog-ical Society. Reproduced from reference 5 (Layton AT, LaytonHE, Dantzler WH, Pannabecker TL: The mammalian urineconcentrating mechanism: Hypotheses and uncertainties. Phys-iology (Bethesda) 24: 250–256, 2009), with permission ofAmerican Physiological Society.


inflammation are detected in the hypothalamic-pitu-itary area. However, several cases of idiopathic essen-tial hypernatremia have been reported, all of them inpatients who were younger than 13 years.

Studies in mice have demonstrated that the Nax

channel, expressed in the posterior pituitary and threecircumventricular organs in the brain (the subfornicalorgan [SFO], the organum vasculosum of the laminaterminalis [OVLT], and the median eminence) is theNa-level sensor of body fluids in the brain. Nax-knockout mice do not stop ingesting salt when dehy-drated, whereas wild-type mice avoid salt; unlikewild-type mice, salt-aversive behavior does not occurin knockout mice.

The investigators who first reported these find-ings recently identified a patient who had essentialhypernatremia and developed autoantibodies to Nax

(9). The patient, a 6.5-year-old Asian girl, had aganglioneuroma adjacent to the adrenal gland com-posed primarily of Nax-positive Schwann-like cells,and the neoplasm most likely evoked an antitumorimmune response (a paraneoplastic disease). She pre-sented with serum sodium levels as high as 199mmol/L with a normal BP and pulse and little changein consciousness. Serum AVP levels when her serumsodium was extremely high were detectable (2.1 pg/ml) but approximately half the levels that would beexpected in a patient with even mild dehydration.Resection of the tumor did not cure the hypernatremia,and she continued to exhibit abnormal osmoregulation4 years after surgery. Plasma AVP levels remainednearly constant with plasma osmolalities ranging from280 to 390 mOsm/kg. Administration of desmopressincombined with fluid deprivation resulted in a urineosmolality of 1009 mOsm/kg, indicating that the pa-tient could concentrate her urine normally, but she hadno complaint of thirst during the fluid deprivation test.She exhibited hyperphagia and hyperhydrosis (in-creased sweating), which contributed to water loss.Four years after surgical removal of the Nax-positivetumor, the patient’s serum continued to contain auto-antibodies to Nax.

Passive transfer of the patient’s IgG to wild-typemice reduced their water intake and inhibited AVP re-lease as a result of complement-mediated cell death in thecircumventricular organs where Nax is expressed. Thepatient’s IgG fraction depleted of autoantibodies to Nax

did not induce symptoms. The mice that were adminis-tered an injection of the patient’s IgG differed from

Nax-knockout mice in that the knockout mice had normalAVP levels. Nax is expressed in cells in the SFO andOVLT, which have projections to the supraoptic andparaventricular nuclei, which are responsible for theregulation of AVP secretion. Osmoreceptors, includingTRPV1 and TRPV4, are thought to be involved in theregulation of the activity of these neurons. After injectionof mice with the patient’s antibodies, evidence of celldegeneration in the SFO, OVLT, and posterior pituitarywere found.

Neurogenic DIFamilial Neurogenic DI. Familial neurohypophy-seal DI is an autosomal dominant disorder caused bymutations in the gene that encodes neurophysin II, theAVP carrier protein. Affected patients have a normalability to conserve water at birth, and progressivepolyuria develops later in childhood, usually withinthe first 6 years of life, with progressively worseningpolyuria and compensatory polydipsia. Because thesepatients have progressive loss of AVP, they mayinitially respond normally to water-deprivation testingand have normal pituitary findings on brain magneticresonance imaging (MRI). These normal findings havesometimes led to an erroneous diagnosis of psycho-genic polydipsia; genetic testing may be helpful inmaking the correct diagnosis and avoids the need forfrequent surveillance of family members (10). Studiesin knock-in mice expressing a mutant neurophysin IIgene that causes familial DI in humans showed thatpolyuria progresses without loss of AVP neurons (11–13). Inclusion bodies were found in the AVP neuronsin the supraoptic nucleus, and their size and numbergradually increased in parallel with the increases inurine volume. Electron microscopy showed that ag-gregates formed in the endoplasmic reticulum of AVPneurons; these aggregates, rather than cell death, areassociated with progressive polyuria. Interestingly, ad-ministration of desmopressin to knock-in mice withthe human disease phenotype decreased AVP mRNAexpression and diminished the formation of inclusionbodies in the AVP neurons (14). Animals treated withdesmopressin remained significantly less polyuric thancontrols for 14 days even after desmopressin admin-istration ended. Increasing expression of AVP with theadministration of hypertonic saline enhanced the for-mation of inclusion bodies in AVP cells. These datashow that activation of AVP neurons accelerates the


formation of aggregates and exacerbates the progres-sion of polyuria.Lymphocytic Hypophysitis. Lymphocytic hypo-physitis is an uncommon autoimmune disease charac-terized by inflammatory enlargement of the pituitarygland with eventual destruction of pituitary tissue andreplacement by fibrosis (15). The disease occurs morecommonly in women and often presents late in preg-nancy. Visual field impairment as a result of extracel-lular pituitary enlargement may develop, and patientspresent with varying degrees of dysfunction of theanterior (lymphocytic adenohypophysitis) and poste-rior pituitary (lymphocytic infundibuloneurohypophy-sitis).IgG4 Disease. IgG4-related disease, also known asIgG4-positive multiorgan lymphoproliferative syn-drome, is characterized by high serum levels of IgG4and dense infiltration of IgG4 plasma cells into mul-tiple organs and tissues, causing pancreatitis; Sjogrensyndrome; interstitial lung disease; and involvement ofthe liver, bile duct, gall bladder, kidney, and retroperi-toneum. A few cases of IgG4-related hypophysitiswith neurogenic DI have been reported (16).Neurosurgical DI. DI is a common complicationafter resections of pituitary adenoma and after trau-matic brain injury (17,18). In a study of 57 patientswho underwent resection of a pituitary adenoma at aGerman academic center, 38.5% developed isolatedDI and 15.7% developed DI followed by hyponatre-mia (17). The onset was typically on the second day ofsurgery, and in 8.7% of patients, desmopressin wasrequired for more than 3 months.

Congenital Nephrogenic DIDefects in the V2R or AQP2 lead to nephrogenic

DI (NDI), a disorder in which patients are unable toconcentrate their urine normally, resulting in polyuria,dehydration, and hypernatremia. The V2R gene, AVPR2,is located on the X chromosome, and AVPR2-linked NDIfollows an X-linked recessive inheritance pattern. TheAQP2 gene is located on chromosome 12. In most cases,AQP2-linked NDI follows an autosomal recessive inher-itance pattern.

Approximately 90% of congenital NDI cases arecaused by AVPR2 mutations encoding the V2R; be-cause this gene is located on the X chromosome, thedisease is generally a male disease with X-linkedrecessive inheritance. Female carriers of AVPR2 mu-tations are usually asymptomatic, but some females

have symptoms even though they are genetically char-acterized as carriers of the mutation. During X-chro-mosome inactivation, a natural process that occurs inevery female during embryonic development, one ofthe two X chromosomes is inactivated. The process israndom, but in individual cells or tissues, a dispropor-tionate number of chromosomes containing the unaf-fected gene can be inactivated. This can lead to mildsymptoms of the disease. In one reported family, themother of a boy and a girl with symptomatic NDI wasasymptomatic with a random pattern of X-chromo-some inactivation; her affected daughter expressed askewed inactivation pattern of 7:93 in favor of thevariant X chromosome (19). Patients with partial dis-ease are able to reach a urine osmolality �300mOsm/kg and �750 mOsm/kg after fluid deprivation,and they may respond to high dosages of AVP ordesmopressin (dDAVP). Combinations of indometha-cin and thiazide diuretics, however, are the most ef-fective therapy.

The V2R, expressed in vascular endothelial cellsas well as in kidney principal collecting duct epithelialcells, regulates secretion of the von Willebrand factor(vWF), the carrier protein for coagulation factor VIII(FVIII). Stimulation of the V2R with AVP leads toexocytosis of the vWF storage organelle Weibel Pal-ade bodies (WPb). In patients with NDI, the release ofWPb in response to the synthetic AVP analog dDAVPincreases plasma levels of vWF, FVIII, and tissue-typeplasminogen activator (t-PA; co-stored with vWF inWPb), indicating that the V2R is functioning normallyand that the concentrating defect is caused by a defectin the AQP2 gene. In patients with AVPR2-linkedNDI, there is no response to dDAVP.

Carriers of the AQP2 gene mutation have anincreased risk for thromboembolism (20). Elevatedplasma levels of FVIII are a risk factor of venous throm-bosis, and levels of FVIII depend on levels of its carrierprotein, vWF. Hypothetically, NDI-linked AQP2 muta-tions would be expected to cause upregulation of AVPrelease and V2R expression, resulting in increasedvWF secretion from the endothelium. Consistent withthis hypothesis, a study of 14 NDI carriers (nine withan AVPR2 mutation and five with an AQP2 mutation)found that in comparison with unaffected family mem-bers, AVP levels were increased only in AQP2 carriers(21). This suggests that AVP release is indeed upregu-lated in carriers of AQP2-linked NDI to compensateeffectively for increased fluid loss. The highest levels


of vWF propeptide, a measure of the vWF secretionrate and the highest levels of vWF antigen and factorFVIII activity, were also observed in carriers of AQP2mutations (21).Secondary NDI. Secondary NDI may also occur asa complication of other inherited tubulopathies, suchas cystinosis, nephronophthisis, Bartter syndrome andapparent mineralocorticoid excess, and distal tubularrenal tubular acidosis (22). Children with these disor-ders may present with recurrent episodes of hyperna-tremia with hypotonic urine unresponsive to dDAVP,leading to a misdiagnosis of congenital NDI, delayingrecognition of the underlying disorder. In cystinosis,trapping of cystine in lysosomes primarily effects theproximal tubule, with resulting Fanconi syndrome, butit can theoretically impair function of any cell; theincidence of concentrating defects in this disease hasnot been studied systematically, but one patient withhypernatremia (165 mmol/L) and a concurrent urineosmolality of 114 mosm/kg has been reported (22).Nephronophthisis is classically associated with polyuriaand polydipsia; because the disease causes generalizeddisruption of tubular function, reported patients typicallyexhibit isosthenuria (the inability to concentrate or dilutethe urine) (22). Bartter syndrome is caused by an inabilityto reabsorb salt in the thick ascending limb; although thiswould also be expected to result in isosthenuria, a mi-nority of patients exhibit hyposthenuria and are prone tohypernatremia when given salt supplements (22). Aurine-concentrating defect is a common feature in pa-tients with apparent mineralocorticoid excess, a diseasecharacterized by hypokalemic metabolic alkalosis, hy-percalciuria, and nephrocalcinosis. One patient with amaximum urine osmolality of 129 mosm/kg recoveredthe ability to concentrate her urine normally after treat-ment with spironolactone and amiloride (22). The pre-sumed causes of the concentrating defect in apparentmineralocorticoid excess and in distal renal tubular aci-dosis are hypokalemia and hypercalciuria.

Acquired NDIAcquired NDI results from hypokalemia, hyper-

calcemia, or medications. Lithium, used to treat bipo-lar disorder, is the most common cause of drug-induced NDI. Approximately 40% of patients who aretreated with lithium present with acquired NDI (23). Amulticenter medical chart review of 116 patients usinglithium found that 12 (26%) of the 46 patients withpolyuria used serotonergic antidepressants compared

with 10 (14%) of the 70 patients without polyuria. Therisk for polyuria in lithium users who concurrently useserotonergic antidepressants was significantly in-creased (odds ratio 2.86; 95% confidence interval 1.00to 8.21) (24).

Bedford et al. (25) studied the impact of psycho-tropic medications on urine-concentrating ability andurinary AQP2 excretion in patients who were takinglithium (n � 45), compared with those who weretaking alternative psychotropic medications (n � 42).Patients who were not taking lithium but were takingother medications could concentrate their urine nor-mally after overnight water deprivation and adminis-tration of desmopressin (958 � 51 mOsm/kg) andincreased urinary excretion of AQP2 and cAMP. Pa-tients who were taking lithium had varying degrees ofimpairment of urine-concentrating ability that wereassociated with decreased urinary AQP2 and cAMPand were correlated with the duration of lithium ther-apy; none of the patients harbored mutations in genescoding for the V2R or AQP2.

Nephrotoxic effects of lithium may be detected 8weeks after the start of treatment. Patients who are onlithium therapy for �10 years may develop chronickidney disease or hypercalcemia, prompting discon-tinuation of the drug. NDI symptoms may disappear inas little as 3 weeks after lithium is stopped; however,urine-concentrating defects usually persist for yearsafter ending treatment. Acute administration of lithiuminhibits the formation of cAMP, thereby preventingactivation of phosphokinase A (PKA) (26). Phosphor-ylation of AQP2 and UT-A1 by PKA is required fortranslocation and insertion of these key transportersinto the apical plasma membrane of the inner medul-lary collecting duct. This initial dysregulation of AVP-regulated water reabsorption contributes to the urine-concentrating defect observed in lithium-treated rats(26). Long-term lithium administration reduces theamount of AQP2 and UT-A1 in rat inner medulla.Although the urea transporters recovered to basal lev-els 14 days after discontinuation of lithium adminis-tration, AQP2 expression did not (27).

Lithium is a potent inhibitor of glycogen syn-thase kinase 3� (GSK3�), a serine/threonine proteinkinase. Knockout mice with a deletion GSK3� haveimpaired urine-concentrating ability in response towater deprivation or treatment with an AVP analogassociated with reduced levels of mRNA, protein, andmembrane localization of the AVP-responsive water


channel AQP2. The knockout also expressed lowerlevels of pS256-AQP2, a phosphorylated form of thewater channel crucial for membrane trafficking (28).Levels of cAMP, a major regulator of AQP2 expres-sion and trafficking, were also lower in the knockoutmice. Both GSK3� gene deletion and pharmacologicinhibition of GSK3� reduced adenylate cyclase activ-ity. Thus, GSK3� inactivation by lithium is a plausiblemechanism for its inhibitory effect on AVP action inthe renal collecting duct.

Amiloride inhibits the uptake of lithium in thecollecting duct by binding to the epithelial sodiumchannel and was used clinically to ameliorate polyuriain two open-label studies. Eleven patients who wereon long-term lithium therapy (patients aged 8 to 34)were enrolled in a randomized, placebo-controlled,crossover trial investigating the actions of amiloride(10 mg/d for 6 weeks) on dDAVP-stimulated urine-concentrating ability and AQP2 excretion (25). After 6weeks of amiloride but not after placebo, urine osmo-lality increased to 164.5 � 8.0% of baseline after theadministration of dDAVP (P � 0.05) with an associ-ated increase in urinary AQP2; because the increaseover baseline during the placebo portion of the trialwas approximately 134%, the affect of amiloride onabsolute urine osmolality seems to be modest.

NDI is a common complication of amphotericinB, and use of the less toxic liposomal preparation doesnot prevent NDI (29). Hypokalemia associated withthe drug may be an exacerbating factor, but renalconcentrating defects have been demonstrated in pa-tients who were treated with amphotericin and whoseplasma potassium concentrations were normal. Invitro, amphotericin B partially inhibits AVP-stimu-lated water permeability and urea transport in the ratinner medullary collecting duct, and it decreases theabundance of AQP2 water channels because of inhi-bition of adenylcyclase and/or G proteins. NDI is awidely known complication of another antibiotic, de-meclocycline, and this frequent adverse effect hasbeen exploited in the treatment of syndrome of inap-propriate antidiuretic hormone secretion. The druginhibits AVP-induced water flow in the toad bladder,but the mechanism of interference with the AVP-AQP2 cascade is unknown.

Gestational DIThe metabolic clearance of AVP increases by

four- to sixfold between week 8 and the middle of

pregnancy. The syncytiotrophoblast of the human pla-centa produces vasopressinase, which quickly de-grades AVP and oxytocin. Activity of the enzymeincreases gradually during pregnancy and reaches itspeak in the third trimester; activity remains high dur-ing labor and delivery and then decreases by 25% perday, becoming undetectable by 2 to 4 weeks postpar-tum (30). Normally, plasma concentrations of AVPduring pregnancy are maintained at normal levelsbecause of a compensatory increase in AVP secretion.Patients with subclinical neurogenic DI are unable toincrease AVP secretion during pregnancy and developpolyuria during their third trimester of pregnancy; inthis case, polyuria responds to the administration ofAVP or to desmopressin. Polyuria associated withsubclinical DI often recurs with each pregnancy, and itoccurs more often with multiple pregnancies; a largerplacental volume may correlate with increased secre-tion of vasopressinase. NDI can also develop at termin patients with an abnormal increase in vasopressi-nase activity but without any preexisting defect inwater metabolism. The increase in vasopressinase isoften associated with preeclampsia, the HELLP syn-drome, or acute fatty liver; hepatic dysfunction isthought to diminish the degradation of vasopressinase(30). Desmopressin, a synthetic analog of AVP, is notinactivated by vasopressinase and is the preferredtreatment; dosages should be equal to or slightlyhigher than those used in central DI in the absence ofpregnancy.

Consequences of HypernatremiaMortality. In general medical-surgical units thattreat non–critically ill patients, the prevalence of hy-pernatremia is approximately 1%; among critically illpatients who are treated in intensive care units (ICUs),the incidence ranges between 10 and 26%, and, inmost cases, hypernatremia develops during the ICUstay. Several studies suggested an association betweenhypernatremia and hospital mortality. However, mostof these studies were retrospective, single-center stud-ies with small numbers of patients and with inadequateadjustment for confounders or occurrence of adverseevents during the ICU stay. A retrospective, observa-tional study on a prospectively collected multicenterdatabase fed by 12 French ICUs was conducted todetermine the effect of mild (�145 but �150 mmol/L)and severe (�150 mmol/L) hypernatremia on mortal-ity after adjustment for confounders (31). The inves-


tigators compared 6895 patients who did not haveICU-acquired hypernatremia (hospital mortality15.2%) with 901 patients who had mild ICU-acquiredhypernatremia (hospital mortality 29.4%) and 344 pa-tients who had moderate to severe ICU-acquired hy-pernatremia (hospital mortality 46.2%). Most (69.8%)of the patients were admitted to the ICU for a medicalcondition, the most common of which were acuterespiratory failure, coma, and septic shock. The over-all prevalence of hypernatremia was 15.3%. Factorsindependently associated with ICU-acquired hyperna-tremia were male gender; greater disease severity onICU admission; and septic shock, acute respiratoryfailure, or coma at ICU admission. Other factors as-sociated with hypernatremia were the need for a blad-der catheter, central venous catheter, or vasoactiveagents and the use of steroids or antibiotics. No datawere available to evaluate fluid balance or use ofdiuretics; in addition, the investigators were unable tocorrect the serum sodium for the effect of hypergly-cemia, and hyperglycemia was not included as one ofthe confounding factors. Both mild and severe hyper-natremia remained highly significant independent riskfactors for mortality after stratification by center andadjustment for time-dependent and non–time-depen-dent confounders. However, the study could not de-termine why ICU-acquired hypernatremia was associ-ated with increased risk for death, and it remainsuncertain from this and other studies whether theassociation reflects a direct effect of hypernatremia ora marker for suboptimal quality of care.

Although ICU-acquired sodium disturbances arecommon in critically ill patients, few studies haveexamined sodium disturbances in patients after cardiacsurgery. A single-center study from a university hos-pital in Austria examined the incidence of hyperna-tremia in patients who were treated in a cardiothoracicsurgery ICU, including 2314 patients who underwentcoronary artery bypass grafting, open heart surgery,aortic surgery, heart or lung transplant, or thromboen-darterectomy of the pulmonary arteries (32). Duringtheir stay in the ICU, 221 (10%) patients developedhypernatremia (serum sodium �145 mmol/L), andpatients with ICU-acquired hypernatremia had higherICU mortality (19 versus 8%; P � 0.01) and a longerICU stay (17 versus 3 days; P � 0.01) compared withpatients without ICU-acquired hypernatremia. ICU-acquired hypernatremia was associated with an in-creased probability of mortality after adjustment for

the SAPS II score, length of surgery, surgery type, andmaximum lactate level immediately after surgery.However, when adjusted for other confounders, theinvestigators could not demonstrate an increased riskfor mortality dependent on the severity of ICU-ac-quired hypernatremia. However, the authors were ableto exclude factors that have been proposed as beingresponsible for the excess mortality with ICU-acquiredhypernatremia, showing that even after adjustment foracute renal failure, administration of furosemide,plasma lactate levels, and acid-base status, hyperna-tremia remained an independent predictor of mortality.A similar single-center study from a regional cardio-vascular ICU in Canada identified 6727 patients whounderwent cardiac surgery (74% for elective coronaryartery bypass grafting) and whose serum sodium levelswere normal on entry, excluding patients with preex-isting dialysis dependence and patients who receivedrenal replacement therapy on the first day in the cardiovas-cular ICU (33). Hypernatremia (�145 mmol/L) developedin 4% of patients, more commonly among patientswho had higher APACHE II scores, were on mechan-ical ventilation, had greater length of hospital staybefore ICU admission, had greater length of stay in theICU, and had the presence of hyperglycemia or abnor-mal serum potassium. Compared with patients withnormal serum sodium levels, hospital mortality wasincreased in patients with ICU-acquired hypernatremia(1.6 versus 14.0%; P � 0.001) and similar to that ofpatients with ICU-acquired hyponatremia (10%). Incontrast to the risk for hyponatremia, which was great-est in the first few days of ICU stay, the risk fordeveloping ICU-acquired hypernatremia steadily in-creased over time. Twenty-six (10.7%) patients withICU-acquired hypernatremia experienced a change inserum sodium �12 mmol/L per d; compared withpatients with lower rates of change, these patients hadhigher ICU (28.2 versus 1.3%; P � 0.001) and hos-pital mortality (31.8 versus 2.7%; P � 0.001). Theinvestigators were unable to determine whether ICU-acquired hypernatremia was a marker of illness sever-ity, and one cannot infer from the results of the studythat correction of hypernatremia would improve out-comes.

In patients with severe burns, microvascular in-tegrity is lost and a plasma-like fluid leaks into theinterstitial place. For ensuring oxygen delivery, fluidresuscitation is necessary until cellular integrity isrestored. Eventually, fluid volume needs to be normal-


ized, but hypernatremia should be avoided becausethere is evidence that it may induce apoptosis anddeepen the depth of the burn wound (34). A single-center study of 40 patients with severe burns (a totallyburned surface area �10%) compared 15 patients whohad hypernatremia (�145 mmol/L) with 25 patientswho did not have hypernatremia (34). All patientswere treated with albumin in Ringer’s lactate, and nopatient was treated with hypertonic fluids. However,there was a significant association between the use ofmechanical hyperventilation and the development ofhypernatremia. Hypernatremia occurred 5.0 � 1.4days after admission to the burn unit and persistedfor 4.6 � 2.7 days. Although there were no differ-ences in age, gender, or severity of the burns, all ofthe patients with normonatremia survived and three(20%) of the patients with hypernatremic died (P �0.045).

Brain Injury from HypernatremiaAs discussed in the section on hyponatremia,

severe hypernatremia can cause central pontine andextrapontine myelinolysis similar to the lesions thathave been described after rapid correction of hypona-tremia, except in certain cases with a predilection forthe limbic system. A 58-year-old woman who devel-oped extreme hypernatremia (serum sodium 200mmol/L), status epilepticus, and coma after a laparo-scopic excision of a hydatid cyst was reported. A 30%saline solution is used to eradicate these cysts, andwhen the procedure was complicated by intraperito-neal rupture of the cyst, the peritoneal cavity waslavaged with the 30% saline solution. Emergencyhemodialysis was started immediately to correct hy-pernatremia, but MRI of the brain in the ensuing 2weeks demonstrated cerebral lesions along the limbicsystem network, in a similar distribution that has beenseen after administration of glutamate (an excitotoxicorganic osmolyte). In a perplexing report from India,11 postpartum young women presented after delivery(during the second and third weeks in two and duringthe sixth week in one) with fever and neurologicsymptoms associated with very severe hypernatremia(�190 mmol/L in six of the patients) and azotemia(35). MRI of the brain before treatment of hyperna-tremia showed hyperintensity in the corpus callosumin all patients and various combinations of hyperin-tensities, consistent with demyelination, in the internalcapsule, corona radiata, cerebellar peduncles, and hip-

pocampus. Seven of the patients eventually improvedneurologically, and four died. Although it is plausiblethat the brain lesions were caused by hypernatremia,there was really no satisfactory explanation for theextreme electrolyte abnormalities in these youngwomen. The authors mentioned that it is customary insome parts of Southern India to restrict water and fluidintake during the puerperal state and speculated thatthis practice may have compounded fluid losses fromfever and the heat of the summer; none of the patientshad a history of polyuria, and those who recoveredneurologic function had no apparent abnormalities ofwater balance.

Treatment of HypernatremiaHypernatremia is common in patients who are

cared for in ICUs because critical illness is oftenassociated with impaired fluid regulation and withimpaired water intake caused by impaired conscious-ness and inability to drink. The physician must com-pensate for the patient’s inability to replace free waterlosses by prescribing appropriate fluid therapy. Fordetermination of the causes of hypernatremia treatedin a university hospital ICU in Austria, data wererigorously analyzed for 69 patients (6% of the patientsadmitted to the ICU during the study period) whoseserum sodium levels rose to �149 mmol/L (maximum150 to 164 mmol/L) after entering the ICU with aserum sodium of �146 mmol/L (36). Of the 69 pa-tients, 24 were excluded because of incomplete data.Positive sodium and potassium balance were found in38% of patients, and 44% had a negative fluid balance;the remaining patients had a combination of positivecation balance and negative fluid balance. Polyuriafrom diuretics (50%), osmotic diuresis from glucose orurea (30%), and DI (4%) were present in 38% ofpatients; excess extrarenal fluid losses from feverand/or diarrhea were found in 63%. In addition to theuse of bicarbonate (13% of cases), the addition ofpotassium chloride to isotonic saline created a hyper-tonic solution, leading to a positive sodium/potassiumbalance in 27% of patients. During the development ofhypernatremia, the serum sodium concentration in-creased by 5.3 � 3.3 mmol/L per d, and almost allpatients required approximately 2 days to develophypernatremia. Therefore, with frequent monitoring ofserum electrolytes (every 8 hours), it should have beenpossible to recognize the trend toward hypernatremiaand correct it with prescription of hypotonic fluids.


The investigators believed that prescribers avoidedhypotonic fluids because of the fear of causing hypo-natremia in patients with conditions expected to resultin nonosmotic release of AVP.

The same group analyzed balance data in theirICU population comparing the ability of three pub-lished formulas (Adrogue-Madias, Barsoum-Levine,and Kurtz-Nguyen) to predict measured changes inserum sodium using daily measurements of sodium/potassium and fluid/electrolyte balances (37). Al-though all of the formulas correlated significantly(P � 0.05), differences between predicted and mea-sured values as high as 15 mmol/L were observed, andwere too large for these formulas to be useful inclinical practice.

Rapid correction of hypernatremia in children isknown to cause cerebral edema, leading to seizures.Although it is generally accepted that a slow reductionof the serum sodium concentration by �10 mmol/Lper d or 0.5 mmol/L per h is desirable, there is littledocumented evidence as to what constitutes a safe rateof rehydration. A single-center study in a tertiarychildren’s hospital in China compared rehydrationregimens in 49 patients (aged 11.9 � 8.4 months) whodeveloped cerebral edema during treatment for hyper-natremia with 48 patients (aged 14.2 � 10.9 months)whose recovery was uneventful (38). The mean serumsodium of the 96 patients on presentation was 164.5mmol/L (range 151.0 to 184.0 mmol/L), and associ-ated symptoms included irritability in 72, irritabilityalternating with lethargy in 18, and lethargy alone inseven; all patients were febrile and had oliguria witheither thirst or dry mucous membranes, and only 11patients had hypotension with poor pulses. Cerebraledema occurred within the first 24 hours in all cases(range 4 to 23 hours). All patients were rehydratedwith hypotonic fluids ranging from 0.20 to 0.67%saline in dextrose after an initial fluid bolus of isotonicsaline in half of the patients. Risk factors for cerebraledema were an initial fluid bolus (29/49 versus 15/48;P � 0.006), severity of hypernatremia (167.7 � 7.8versus 161.3 � 7.9 mmol/L; P � 0.001), and theoverall rehydration rate (8.2 � 1.1 versus 6.4 � 0.6ml/kg per h; P � 0.001). On logistic regression, theodds ratio for developing cerebral edema was unaf-fected by the tonicity of the solution used for rehydra-tion, but it was significantly higher in patients with ahigher initial serum sodium concentration and forbolus therapy to expand the plasma volume; a rapid

rate of rehydration was the most significant contribu-tor to cerebral edema. The rate of reduction of serumsodium in the cerebral edema group was 1.0 � 0.3mmol/L per h and in the group without cerebral edemawas 0.5 � 0.1 mmol/L per h (P � 0.001). The authorsproposed that the overall rate of rehydration in the first24 hours of therapy for hypernatremia not exceed 6.8ml/kg per h, with preference for a slow and uniformrate. Initial normal saline fluid boluses should berestricted to patients with signs of circulatory collapseor shock. The greater the severity of hypernatremia,the slower the rate of rehydration and the higher thesodium concentration of the rehydration fluid shouldbe, with a goal of correction of not more than 0.5mmol/L per h (38).

Pediatric or adult patients with acute kidneyinjury may have concomitant severe hypernatremia.For avoiding rapid correction during renal replacementtherapy, continuous venovenous hemofiltration hasbeen used with adjustment of the replacement fluidsodium concentration using 30% saline additives (39).Each 5 ml of 30% saline added to a 5-L bag ofreplacement fluid containing a sodium concentrationof 140 mmol/L raises the fluid sodium concentrationby 5 mmol/L. Thus, for example, for a patient with aserum sodium concentration of 165 mmol/L, additionof 20 ml of 30% saline to the replacement solution willproduce a solution with a sodium concentration of 160mmol/L. In the United States, 23% saline is availablerather than the 30% solution used in this study; 6 ml of23% saline is equivalent to 5 ml of 30% saline, and, ifadded to a 5-L bag of replacement solution, it willraise the fluid sodium concentration by 5 mmol/L.Stepwise correction of the patient’s serum sodiumconcentration can be planned using replacement fluidmade up to lower the serum sodium concentrationsuccessively. If the serum sodium decreases by �2mmol/L in 6 hours, either the rate of filtration shouldbe decreased or the fluid replacement bag should bechanged to bags with a higher sodium concentration.Because the volumes of 30% saline additives aresmall, they will not substantially alter the concentra-tion of other electrolytes in the replacement solution.

References1. Thornton SN: Thirst and hydration: Physiology and consequences of

dysfunction. Physiol Behav 100: 15–21, 20102. Sands JM, Layton HE: The physiology of urinary concentration: An

update. Semin Nephrol 29: 178–195, 2009


3. Fenton RA: Urea transporters and renal function: Lessons fromknockout mice. Curr Opin Nephrol Hypertens 17: 513–518, 2008

4. Halperin ML, Kamel KS, Oh MS: Mechanisms to concentrate theurine: An opinion. Curr Opin Nephrol Hypertens 17: 416–422, 2008

5. Layton AT, Layton HE, Dantzler WH, Pannabecker TL: The mam-malian urine concentrating mechanism: Hypotheses and uncertain-ties. Physiology (Bethesda) 24: 250–256, 2009

6. Pannabecker TL, Dantzler WH, Layton HE, Layton AT: Role ofthree-dimensional architecture in the urine concentrating mechanismof the rat renal inner medulla. Am J Physiol Renal Physiol 295:F1271–F1285, 2008

7. Cheng CY, Chu JY, Chow BK: Vasopressin-independent mecha-nisms in controlling water homeostasis. J Mol Endocrinol 43: 81–92,2009

8. Hasler U, Leroy V, Martin PY, Feraille E: Aquaporin-2 abundance inthe renal collecting duct: New insights from cultured cell models.Am J Physiol Renal Physiol 297: F10–F18, 2009

9. Hiyama TY, Matsuda S, Fujikawa A, Matsumoto M, Watanabe E,Kajiwara H, Niimura F, Noda M: Autoimmunity to the sodium-levelsensor in the brain causes essential hypernatremia. Neuron 66:508–522, 2010

10. Stephen MD, Fenwick RG, Brosnan PG: Polyuria and polydipsia ina young child: diagnostic considerations and identification of novelmutation causing familial neurohypophyseal diabetes insipidus. Pi-tuitary April 18, 2010 [epub ahead of print]

11. Arima H, Oiso Y: Mechanisms underlying progressive polyuria infamilial neurohypophyseal diabetes insipidus. J NeuroendocrinolJuly 7, 2010 [epub ahead of print]

12. Arima H, Oiso Y: Mechanisms underlying progressive polyuria infamilial neurohypophyseal diabetes insipidus. J Neuroendocrinol 22:754–757, 2010

13. Hayashi M, Arima H, Ozaki N, Morishita Y, Hiroi M, Ozaki N,Nagasaki H, Kinoshita N, Ueda M, Shiota A, Oiso Y: Progressivepolyuria without vasopressin neuron loss in a mouse model forfamilial neurohypophyseal diabetes insipidus. Am J Physiol RegulIntegr Comp Physiol 296: R1641–F1649, 2009

14. Hiroi M, Morishita Y, Hayashi M, Ozaki N, Sugimura Y, NagasakiH, Shiota A, Oiso Y, Arima H: Activation of vasopressin neuronsleads to phenotype progression in a mouse model for familial neu-rohypophyseal diabetes insipidus. Am J Physiol Regul Integr CompPhysiol 298: R486–R493, 2010

15. Akahori H, Sugimoto T: Lymphocytic hypophysitis with a long latentperiod from onset of central diabetes insipidus to development ofpituitary enlargement. Intern Med 49: 1565–1571, 2010

16. Hori M, Makita N, Andoh T, Takiyama H, Yajima Y, Sakatani T,Fukumoto S, Iiri T, Fujita T: Long-term clinical course of IgG4-related systemic disease accompanied by hypophysitis. Endocr J 57:485–492, 2010

17. Kristof RA, Rother M, Neuloh G, Klingmuller D: Incidence,clinical manifestations, and course of water and electrolyte me-tabolism disturbances following transsphenoidal pituitary ade-noma surgery: A prospective observational study. J Neurosurg111: 555–562, 2009

18. Krahulik D, Zapletalova J, Frysak Z, Vaverka M: Dysfunction ofhypothalamic-hypophyseal axis after traumatic brain injury in adults.J Neurosurg 113: 581–584, 2010

19. Faerch M, Corydon TJ, Rittig S, Christensen JH, Hertz JM, Jendle J:Skewed X-chromosome inactivation causing diagnostic misinterpre-tation in congenital nephrogenic diabetes insipidus. Scand J UrolNephrol 44: 324–330, 2010

20. Nossent AY, Vos HL, Rosendaal FR, Bertina RM, Eikenboom JC:Aquaporin 2 gene variations, risk of venous thrombosis and plasmalevels of von Willebrand factor and factor VIII. Haematologica 93:959–960, 2008

21. Nossent AY, Ellenbroek JH, Frolich M, Bertina RM, Knoers NV,Eikenboom JC: Plasma levels of von Willebrand factor, von

Willebrand factor propeptide and factor VIII in carriers andpatients with nephrogenic diabetes insipidus. Thromb Res 125:554 –556, 2010

22. Bockenhauer D, Van’t Hoff W, Dattani M, Lehnhardt A, SubtireluM, Hildebrandt F, Bichet DG: Secondary nephrogenic diabetes in-sipidus as a complication of inherited renal diseases. Nephron Physiol116: p23–p29, 2010

23. Grunfeld JP, Rossier BC: Lithium nephrotoxicity revisited. Nat RevNephrol 5: 270–276, 2009

24. Wilting I, Egberts AC, Movig KL, Laarhoven JH, Heerdink ER,Nolen WA: The association between concomitant use of serotonergicantidepressants and lithium-induced polyuria: A multicenter medicalchart review study. Pharmacopsychiatry 41: 129–133, 2008

25. Bedford JJ, Weggery S, Ellis G, McDonald FJ, Joyce PR, Leader JP,Walker RJ: Lithium-induced nephrogenic diabetes insipidus: renaleffects of amiloride. Clin J Am Soc Nephrol 3: 1324–1331, 2008

26. Nielsen J, Kwon TH, Christensen BM, Frokiaer J, Nielsen S: Dys-regulation of renal aquaporins and epithelial sodium channel inlithium-induced nephrogenic diabetes insipidus. Semin Nephrol 28:227–244, 2008

27. Blount MA, Sim JH, Zhou R, Martin CF, Lu W, Sands JM, Klein JD:Expression of transporters involved in urine concentration recoversdifferently after cessation of lithium treatment. Am J Physiol RenalPhysiol 298: F601–F608, 2010

28. Rao R, Patel S, Hao C, Woodgett J, Harris R: GSK3beta mediatesrenal response to vasopressin by modulating adenylate cyclase activ-ity. J Am Soc Nephrol 21: 428–437, 2010


30. Aleksandrov N, Audibert F, Bedard MJ, Mahone M, Goffinet F,Kadoch IJ: Gestational diabetes insipidus: A review of an underdi-agnosed condition. J Obstet Gynaecol Can 32: 225–231, 2010

31. Darmon M, Timsit JF, Francais A, Nguile-Makao M, Adrie C, CohenY, Garrouste-Orgeas M, Goldgran-Toledano D, Dumenil AS, JamaliS, Cheval C, Allaouchiche B, Souweine B, Azoulay E: Associationbetween hypernatraemia acquired in the ICU and mortality: A cohortstudy. Nephrol Dial Transplant 25: 2510–2515, 2010

32. Lindner G, Funk GC, Lassnigg A, Mouhieddine M, Ahmad SA,Schwarz C, Hiesmayr M: Intensive care-acquired hypernatremia aftermajor cardiothoracic surgery is associated with increased mortality.Intensive Care Med 36: 1718–1723, 2010

33. Stelfox HT, Ahmed SB, Zygun D, Khandwala F, Laupland K:Characterization of intensive care unit acquired hyponatremia andhypernatremia following cardiac surgery. Can J Anaesth 57: 650–658, 2010

34. Namdar T, Siemers F, Stollwerck PL, Stang FH, Mailander P, LangeT: Increased mortality in hypernatremic burned patients. Ger Med Sci8: Doc11, 2010

35. Naik KR, Saroja AO: Seasonal postpartum hypernatremic encepha-lopathy with osmotic extrapontine myelinolysis and rhabdomyolysis.J Neurol Sci 291: 5–11, 2010

36. Lindner G, Kneidinger N, Holzinger U, Druml W, Schwarz C:Tonicity balance in patients with hypernatremia acquired in theintensive care unit. Am J Kidney Dis 54: 674–679, 2009

37. Lindner G, Schwarz C, Kneidinger N, Kramer L, Oberbauer R,Druml W: Can we really predict the change in serum sodium levels?An analysis of currently proposed formulae in hypernatraemic pa-tients. Nephrol Dial Transplant 23: 3501–3508, 2008

38. Fang C, Mao J, Dai Y, Xia Y, Fu H, Chen Y, Wang Y, Liu A: Fluidmanagement of hypernatraemic dehydration to prevent cerebral oe-dema: A retrospective case control study of 97 children in China. JPaediatr Child Health 46: 301–303, 2010

39. Ostermann M, Dickie H, Tovey L, Treacher D: Management ofsodium disorders during continuous haemofiltration. Crit Care 14:418, 2010


Edema and Diuretics

Cirrhosis and AscitesThe formation of ascites depends on the balance

between increased local filtration of fluid in the he-patic sinusoids and intestinal capillaries and aug-mented lymph drainage (1); accumulation of ascitesalso requires sodium and water retention. Although itwould be logical to attribute renal sodium retention incirrhosis to an adaptive response to underfilling of thecirculation caused by fluid transudation into the ab-dominal cavity, measurements of plasma volume donot support this explanation. An alternative theory, thevasodilation hypothesis, holds that systemic andsplanchnic vasodilation associated with increased cir-culating nitric oxide lowers arterial pressure despite anexpanded plasma volume (i.e., the capacitance of thecirculation is increased), triggering the usual neurohu-moral responses to arterial underfilling (renin-angio-tensin-aldosterone, the sympathetic nervous system,and nonosmotic release of vasopressin), which thenpromote renal sodium and water retention. However, arecent review that detailed study of the evolution ofsystemic hemodynamics and sodium balance in anexperimental model of cirrhosis in the dog found that

volume expansion preceded systemic vasodilation andthat natriuresis develops in patients with cirrhosis afterplacement of a porto-systemic vascular shunt despiteincreased peripheral vasodilation; these observationssuggest that systemic vasodilation is not the primaryevent in the pathogenesis of salt retention but rather ahomeostatic response to extracellular fluid volumeexpansion (2). The review proposed that renal saltretention is a primary event caused by pathologicactivation of a hepatic vascular sensor involved involume control. The presence of an afferent sensor inthe hepatic circulation is supported by several ob-servations: (1) The liver has sensors that regulaterenal function (e.g., electrical stimulation ofperivascular portal nerves increases glomerular fil-tration and renal sympathetic nervous system; he-patic denervation diminishes natriuresis; renal nerveactivity increases after an oral salt load or portalvein infusion of hypertonic saline); (2) hepatic veinthrombosis or occlusion stimulates renal salt andwater retention that is normalized by side-to-sideportocaval shunts (which lower both hepatic veinand portal vein pressure and maintain mixing ofportal venous and hepatic arterial blood supplying

PV HA

IVC

PressurePressureNormal pressure

Constriction

Cirrhosis orHV constriction

Cirrhosis andside-to-side shunt

Cirrhosis andend-to-side shunt

Na+ retention Na+ aNecnalab + retention

Figure 16. Hepatic vascular hemodynamics and sodium balance. (A) Cirrhosis, or restriction of hepatic vein flow, increasesintrahepatic vascular resistance and sinusoidal pressure, markedly decreasing portal vein (PV) flow and increasing hepaticartery (HA) flow. Changes in the physical forces or in the composition of the hepatic blood trigger sodium retention and edemaformation. (B) Insertion of a side-to-side porto-caval shunt decreases sinusoidal pressure and maintains mixing of portalvenous and hepatic arterial bloods, irrigating the liver. Under these conditions and despite cirrhosis, there is no sodiumretention. (C) Insertion of an end-to-side porto-caval shunt only partially decreases the elevated sinusoidal pressure andprevents mixing of the venous and arterial hepatic blood supplies as the portal vein blood is diverted to the inferior vena cava.Under these conditions and despite normalization of portal vein pressure, sodium retention continues unabated. HV, hepaticvein; IVC, inferior venacava. Reprinted from reference 2 (Oliver JA, Verna EC: Afferent mechanisms of sodium retention incirrhosis and hepatorenal syndrome. Kidney Int 77: 669–680, 2010), with permission of Blackwell Publishing, Inc.


the liver) but not by end-to-side shunts (which loweronly portal vein pressures and prevent mixing ofportal and hepatic arterial blood) (Figure 16); and(3) the hepatic artery has significant autoregulatorycapacity, suggesting the presence of an afferentsensor analogous to other circulations with thiscapacity—the kidney’s baroreceptor and the brain’scarotid sinus receptor.

The American Association for the Study of LiverDiseases (AASLD) has published evidence-basedguidelines for the management of ascites caused bycirrhosis (3). Since 1994, the guidelines have advo-cated combinations of furosemide and spironolactonefor the management of ascites. However, until re-cently, this recommendation was not backed by evi-dence from controlled trials (4). Angeli et al. (5)recently found that beginning with combined treat-ment was preferable because it achieved more rapidresolution of ascites (15 versus 21 days) with a lowerincidence of hyperkalemia than a strategy of beginningwith spironolactone alone and adding furosemide ifdiuretic resistance were encountered. Previous studiesshowing better results with monotherapy enrolled ahigh percentage of patients with new-onset ascites,whereas most of the patients in the more recent studyhad recurrent ascites.

A group in Italy reported favorable outcomes intreating patients who had diuretic-resistant heart fail-ure with extremely high dosages of intravenous furo-semide (250 to 1000 mg twice daily) combined withhypertonic saline (150 ml of 1.4 to 4.6% saline, de-pending on the serum sodium concentration) and moreliberal dietary sodium intake. The investigators re-ported a pilot study of the effectiveness of this novelapproach for patients with hepatic cirrhosis and ascitesresistant to �160 mg of furosemide and 400 mgspironolactone daily. High-dosage furosemide and hy-pertonic saline (60 patients) was compared with astandard diuretic schedule and two to three dailyparacenteses (24 patients) (6). The hypertonic saline/furosemide regimen was well tolerated and resulted inhigher urine output, high plasma sodium concentra-tions, lower body weight, less leg edema, and lesspleural fluid than patients who were treated withparacenteses and lower dosages of diuretic. Severalpossible explanations have been offered for why thiscounterintuitive approach to edema might work (7).Studies in isolated perfused hearts showed that hyper-tonicity may have a direct effect on myocardial per-

formance, possibly by increasing intracellular calciumand possibly by reducing cardiomyocyte edema. Otherproposed mechanisms include markedly reduced renalvascular resistance and reduced levels of inflammatorycytokines associated with hypertonic saline.

Transjugular intrahepatic portosystemic shunt(TIPS) has been proposed as an alternative to paracen-tesis in patients who have ascites and do not respondto diuretics, and a comprehensive review of the pro-cedure has been published (8). TIPS reduces the fil-tration pressure favoring ascites formation and in-creases urinary sodium excretion and urine volumeand reduces serum creatinine concentration; thesefindings are associated with decreased plasma renin,aldosterone, and norepinephrine and are consistentwith improvement in arterial underfilling. Meta-anal-yses of five randomized trials suggested an improvedsurvival after TIPS in comparison with that in patientswho required repeated and frequent paracenteses, butthe number of patients studied only totals 305. Severaluncontrolled studies recommended TIPS for the treat-ment of hepatic hydrothorax.

The vasopressin receptor antagonist tolvaptanhas been shown to increase urine output and decreasebody weight in patients with heart failure and to raisethe serum sodium concentration in patients with hy-ponatremia. In an open-label dose-ranging study, with-out a placebo control, the vasopressin antagonisttolvaptan was used as add-on therapy to 17 patientswith normonatremia and decompensated liver cirrho-sis and with ascites and edema resistant to furosemideat a dosage of �40 mg/d (9). Patients were allowed toincrease their intake of water if they became thirsty.Tolvaptan was initiated at 15 mg/d. If ascites andedema persisted after 3 days, then the dosage wastitrated to 30 mg/d for 3 days and then to 60 mg foranother 3 days. At a dosage of 15 mg, tolvaptanimproved clinical signs of ascites in edema as mea-sured by abdominal circumference, body weight, andlower limb pitting in 65% of patients; similar resultswere shown in 80% of patients at a dosage of 30 mgand 91% of patients at a dosage of 60 mg. After 9 daysof therapy, with escalating dosages of tolvaptan every3 days, the overall decrease in body weight averaged3.4 � 2.1 kg. Administration of tolvaptan significantlyincreased urine volume from baseline (1445 ml/d) to3240 ml/d on 15 mg, 3943 ml on 30 mg, and 4537 mlon 60 mg; compared with baseline, the drug decreasedurine osmolality (by approximately 50%), increased


serum sodium (by approximately 3 mmol/L), andincreased plasma vasopressin levels (by approximately2 to 3 pg/ml).

A similar randomized, placebo-controlled trial of148 patients with cirrhosis and ascites without hypo-natremia (all serum sodium values �130 mmol/L,mean serum sodium 136 to 137 at baseline) exploredthe effectiveness of the investigational vasopressin 2receptor antagonist satavaptan in combination withfixed dosages of spironolactone (100 mg/d) and furo-semide 20 to 25 mg/d (10). Administration of satavap-tan for 14 days was associated with a significantreduction in weight at all dosages (5.0, 12.5, and 25.0mg/d) averaging 2.08 to 2.46 kg compared with 0.36kg on placebo. The drug increased urine volume to ashigh as 5 L/d in some patients and, as would beexpected, decreased urine osmolality, although therewas no measured increase in sodium excretion. Fluidintake was not restricted, so the mean serum sodiumconcentration increased by only 0.8 to 2.5 mmol/L, ina dosage-dependent manner. Four of 38 patients whowere treated with the highest dosage experienced anincrease in serum sodium of �8 mmol/L per d (as highas 14 mmol/L) without associated neurologic compli-cations, and 28% of patients noted increased thirst atthe highest dosage. Renal function was not adverselyaffected.

Heart FailureAlthough high dosages of spironolactone (up to

400 mg/d) are widely accepted for the treatment of fluidretention in hepatic cirrhosis, in congestive heart failure,spironolactone has been administered in dosages that donot significantly increase sodium excretion (approxi-mately 25 mg/d) with the goal of preventing cardiac andvascular fibrosis, rather than conquering diuretic resis-tance. In patients who have advanced heart failure andare resistant to loop diuretics, diuretic dosages of spi-ronolactone have been avoided because concurrent treat-ment with angiotensin-converting enzyme inhibitors, an-giotensin receptor antagonists, and � blockers increasethe risk for hyperkalemia. In small trials of diuretic-resistant patients with heart failure, high dosages ofspironolactone (100 to 200 mg/d) significantly increasedsodium excretion without hyperkalemia. Recent editorialreviews argued that combinations of natriuretic dosagesof mineralocorticoid antagonists need not have a prohib-itive incidence of hyperkalemia when these are combined

with high dosages of loop diuretics and that this strategywarrants a carefully designed randomized trial (11,12).

Plasma levels of B-type natriuretic peptides(BNPs) have been widely used to guide treatment ofpatients with chronic heart failure. However, the ben-efits of this approach have been uncertain. A meta-analysis of eight randomized, controlled trials with atotal of 1726 patients found a significantly lower riskfor all-cause mortality (relative risk 0.76; 95% confi-dence interval 0.63 to 0.91; P � 0.003) in the BNP-guided therapy group compared with the control groupbut not in patients who were aged �75 years (13). Ahigher percentage of patients achieved target dosagesof angiotensin-converting enzyme inhibitors and �blockers during the course of these trials in the BNPgroup than in control subjects (21 to 22% in the BNPgroup versus 11.7 and 12.5% in control subjects,respectively).

Use of loop diuretics in the treatment of heartfailure has been implicated in the pathogenesis ofworsening renal function, and they may worsen sec-ondary hyperaldosteronism. Concern about the safetyof loop diuretics has spawned clinical trials to identifyalternative therapies for reducing congestion in acutedecompensated heart failure (ultrafiltration, natriureticpeptides, vasopressin receptor antagonists, and aden-osine type 1 receptor antagonists). Loop diuretics havebeen viewed as a “necessary evil” that may actually bedoing harm. Recent observations suggested that con-cern about diuretic-induced azotemia may be mis-placed. A retrospective analysis of the EvaluationStudy of Congestive Heart Failure and PulmonaryArtery Catheterization Effectiveness (ESCAPE) trialdatabase found that despite receiving higher dosagesof loop diuretics and a greater risk for worsening renalfunction, patients with the greatest degree of hemo-concentration (as evidenced by an increase in serumlevels of albumin and total protein and/or increase inhematocrit in response to treatment during hospitaliza-tion) had markedly reduced posthospitalization mor-tality when compared with patients who did not showevidence of hemoconcentration (14).

A desire to achieve adequate decongestion ofheart failure while avoiding the potentially adverseaffects of diuretic therapy has led to interest in the useof ultrafiltration rather than diuretics to treat heartfailure in patients with preserved renal function. In anunblinded trial funded by the manufacturer of anultrafiltration device, patients were randomly assigned


to ultrafiltration (n � 100) or to standard intravenousdiuretic therapy; patients in the diuretic arm receivedeither bolus injection (n � 68) or continuous infusion(n � 32) at the discretion of the treating physician(15). At 90 days, the ultrafiltration group had signifi-cantly fewer rehospitalizations and unscheduled visits(0.65 � 1.36) than the bolus (1.31 � 1.87; P � 0.05)or continuous infusion (2.29 � 3.23; P � 0.016)diuretic group. However, although the net fluid lossachieved by ultrafiltration (4.6 � 2.6 L) was notstatistically different from that in patients who weregiven bolus diuretics (2.9 � 3.5 L) or continuousdiuretics (3.6 � 3.5 L), the trend favored ultrafiltra-tion. Furthermore, weight loss at 48 hours in theultrafiltration group (5.0 � 3.1 kg) was significantlyhigher than in the bolus diuretic group (3.1 � 2.6 kg)or continuous infusion diuretic group (2.29 � 3.23kg). Therefore, it remains unclear whether the amountof fluid removed or the composition of fluid removedaffects outcomes. There was no difference in mortalityor serum creatinine between diuretic and ultrafiltrationtherapy. A larger blinded trial designed to achieveequal fluid removal is needed before this invasive andexpensive therapy can be recommended.

Nephrotic SyndromeThe pathogenesis of edema formation in the

nephrotic syndrome remains undefined. Loss of on-cotic pressure because of proteinuria and hypoalbu-minemia is an inadequate explanation for edema be-cause many patients have evidence of volumeexpansion with suppression of components of therenin-angiotensin-aldosterone system during times ofavid sodium retention, because natriuretic responses tosteroids often precede recovery of hypoalbuminemia,and because sodium retention occurs only on theaffected side of a unilateral nephrotic rat model. The“overfill hypothesis” suggests that volume overload inat least some case of the nephrotic syndrome is gen-erated by a primary defect in renal sodium handling.

Regulated by aldosterone and vasopressin, theepithelial sodium channel (ENaC) plays a key role inthe pathogenesis of edema formation in heart failure,hepatic cirrhosis, and the nephrotic syndrome. How-ever, adrenalectomized animals and the vasopressin-deficient Brattleboro rat can develop nephrosis, andnot all patients with the nephrotic syndrome haveelevated levels of aldosterone and vasopressin. Recentevidence suggested that a leaky glomerular filtration

barrier allows filtration of proteases or precursors ofproteases with the ability to activate ENaC (16). Plas-min, which has the ability to activate ENaC by cleav-ing its �-subunit, seems to be the predominant aprotinin-sensitive serine protease identified in nephrotic urinefrom rats and humans. Plasminogen filtered by the dam-aged glomerulus is cleaved to its active form, plasmin, byurokinase and other proteases in the urine (17).

The overfill hypothesis does not always fit pa-tient characteristics in clinical practice, particularlychildren with idiopathic nephrotic syndrome. Pediatri-cians are reluctant to treat patients with diureticswithout concurrent infusion of albumin because ofconcerns about hypovolemia and increased risk forthromboembolic complications. A recent report ana-lyzed 10 consecutive children with idiopathic ne-phrotic syndrome and found that a fractional excretionof sodium (FENa) �0.2% was predictive of highrenin, aldosterone, and vasopressin levels and a highblood urea nitrogen–creatinine ratio supporting a di-agnosis of volume contraction; a higher FENa seemedto reflect volume expansion. The investigators appliedthis standard to a second set of 20 children withnephrosis and classified nine as volume contracted and11 as volume expanded (18). The volume expandedgroup (FENa �0.2%) was treated without albumin(intravenous furosemide 1 mg/kg to a maximum of 40mg twice daily) and oral spironolactone (2.5 mg/kgper d to a maximum 100 mg twice daily). Volume-contracted patients (defined by a FENa �0.2%) weretreated with 25% albumin at 0.5 g/kg twice daily over2 to 3 hours, followed by intravenous furosemide at 1mg/kg per dose (maximum 40 mg) at the end ofalbumin infusion for severe edema. All patients re-ceived steroids and dietary sodium and fluid restric-tion. Treatment with diuretics without concurrent al-bumin infusion was well tolerated, and there was nodifference in hospital stay or weight loss between thevolume-contracted and volume-expanded groups aftertreatment. The authors suggested that FENa can suc-cessfully identify patients who are safe to treat withdiuretics without albumin.

References1. Moller S, Henriksen JH, Bendtsen F: Ascites: Pathogenesis and

therapeutic principles. Scand J Gastroenterol 44: 902–911, 20092. Oliver JA, Verna EC: Afferent mechanisms of sodium retention in

cirrhosis and hepatorenal syndrome. Kidney Int 77: 669–680, 20103. Runyon BA: Management of adult patients with ascites due to

cirrhosis: An update. Hepatology 49: 2087–2107, 2009


4. Bernardi M: Optimum use of diuretics in managing ascites in patientswith cirrhosis. Gut 59: 10–11, 2010

5. Angeli P, Fasolato S, Mazza E, Okolicsanyi L, Maresio G, Velo E,Galioto A, Salinas F, D’Aquino M, Sticca A, Gatta A: Combinedversus sequential diuretic treatment of ascites in non-azotaemicpatients with cirrhosis: Results of an open randomised clinical trial.Gut 59: 98–104, 2010

6. Licata G, Tuttolomondo A, Licata A, Parrinello G, Di Raimondo D,Di Sciacca R, Camma C, Craxì A, Paterna S, Pinto A: Clinical trial:High-dose furosemide plus small-volume hypertonic saline solutionsvs. repeated paracentesis as treatment of refractory ascites. AlimentPharmacol Ther 30: 227–235, 2009

7. Liszkowski M, Nohria A: Rubbing salt into wounds: Hypertonicsaline to assist with volume removal in heart failure. Curr Heart FailRep 7: 134–139, 2010

8. Rossle M, Gerbes AL: TIPS for the treatment of refractory ascites,hepatorenal syndrome and hepatic hydrothorax: a critical update. Gut59: 988–1000, 2010

9. Okita K, Sakaida I, Okada M, Kaneko A, Chayama K, Kato M, SataM, Yoshihara H, Ono N, Murawaki Y: A multicenter, open-label,dose-ranging study to exploratively evaluate the efficacy, safety, anddose-response of tolvaptan in patients with decompensated livercirrhosis. J Gastroenterol 45: 979–987, 2010

10. Gines P, Wong F, Watson H, Terg R, Bruha R, Zarski JP, Dudley F,NormoCAT Study Investigators: Clinical trial: Short-term effects ofcombination of satavaptan, a selective vasopressin V2 receptor an-tagonist, and diuretics on ascites in patients with cirrhosis withouthyponatraemia—A randomized, double-blind, placebo-controlledstudy. Aliment Pharmacol Ther 31: 834–845, 2010

11. Bansal S, Lindenfeld J, Schrier RW: Sodium retention in heart failureand cirrhosis: Potential role of natriuretic doses of mineralocorticoidantagonist? Circ Heart Fail 2: 370–376, 2009

12. Schrier RW, Masoumi A, Elhassan E: Aldosterone: Role in edema-tous disorders, hypertension, chronic renal failure, and metabolicsyndrome. Clin J Am Soc Nephrol 5: 1132–1140, 2010

13. Porapakkham P, Zimmet H, Billah B, Krum H: B-type natriureticpeptide-guided heart failure therapy: A meta-analysis. Arch InternMed 170: 507–514, 2010

14. Testani JM, Chen J, McCauley BD, Kimmel SE, Shannon RP:Potential effects of aggressive decongestion during the treatment ofdecompensated heart failure on renal function and survival. Circula-tion 122: 265–272, 2010

15. Costanzo MR, Saltzberg MT, Jessup M, Teerlink JR, Sobotka PA:Ultrafiltration is associated with fewer rehospitalizations than con-tinuous diuretic infusion in patients with decompensated heart failure:results from UNLOAD. J Card Fail 16: 277–284, 2010

16. Svenningsen P, Bistrup C, Friis UG, Bertog M, Haerteis S, KruegerB, Stubbe J, Jensen ON, Thiesson HC, Uhrenholt TR, Jespersen B,Jensen BL, Korbmacher C, Skøtt O: Plasmin in nephrotic urineactivates the epithelial sodium channel. J Am Soc Nephrol 20:299–310, 2009

17. Kleyman TR, Hughey RP: Plasmin and sodium retention in nephroticsyndrome. J Am Soc Nephrol 20: 233–234, 2009

18. Kapur G, Valentini RP, Imam AA, Mattoo TK: Treatment ofsevere edema in children with nephrotic syndrome with diureticsalone: A prospective study. Clin J Am Soc Nephrol 4: 907–913,2009


Nephrology Self-Assessment Program

Examination QuestionsInstructions to obtain 8 AMA PRA Category 1 CreditsTM

Date of Original Release: March 2011Examination Available Online: on or before Monday, March 7, 2011Audio Files Available: On or before Tuesday, March 15, 2011. A notice will be posted on the ASN website when the audiofiles become available.

CME Credit Eligible Through: February 29, 2012

Answers: Correct answers with explanations will be posted on the ASN website in March 2012 when the issue is archived.UpToDate Links Active: March and April 2011

Core Nephrology question links active: March, April, and May 2011

Target Audience: Nephrology Board and recertification candidates, practicing nephrologists, and internists.

Method of Participation:● Read the syllabus that is supplemented by original articles in the reference lists, and complete the online self-assessment

examination.● Examinations are available online only after the first week of the publication month. There is no fee. Each participant is

allowed two attempts to pass the examination (�75% correct) for CME credit.● Upon completion, review your score and incorrect answers.● Your CME certificate can be printed immediately after completion.● Answers and explanations are provided with a passing score and/or after the second attempt.● CME Credit will be posted to your transcript within 48 hours after checking the attestation box.

Instructions to Access the Online Examination and Evaluation:● Go to the ASN website: www.asn-online.org.● Click the CME tab at the top of the homepage.● Click the ASN CME Center button on the left side of the page.● Click on to the ASN CME Center icon.● Login to the ASN website.● Select Claim Credits for the NephSAP topic-activity you would like to complete.● Complete the NephSAP examination.● Complete the evaluation.● Enter the number of CME credits commensurate with your participation in the activity.● Check the box attesting that you have completed this activity.● You can print your CME certificate immediately.● CME credit will be posted to your transcript within 48 hours.● View or print your full transcript anytime at “My CME Center.”

Instructions to Obtain American Board of Internal Medicine (ABIM) Maintenance of Certification(MOC) Points:

Each issue of NephSAP provides 10 MOC points. Respondents must meet the following criteria:● Be certified by ABIM in internal medicine and/or nephrology and must be enrolled in the ABIM–MOC program

via the ABIM website (www.abim.org).● Take the self-assessment examination within the timeframe specified in this issue of NephSAP.● Designate the issue for MOC points by clicking on the MOC link on the CME certificate page after passing the examination.

You will be leaving the ASN site and transferring the information directly to the ABIM in real-time.● Provide your ABIM Certificate ID number and your date of birth.● You will receive a confirmation message from the ABIM indicating the receipt of your information.

MOC points will be applied to only those ABIM candidates who have enrolled in the program. It is your responsibility to completethe ABIM MOC enrollment process.


182

Volume 10, Number 2, March 2011—Fluid, Electrolyte, and Acid-Base Disturbances

1. A 50-year-old woman with a history of hyperten-sion and coronary artery disease is treated foresophageal squamous carcinoma with 4 days of5-fluorouracil followed by a single dose of cispla-tin. Five days later, she complains of dizziness.Supine heart rate (HR) is 120 bpm, and systolic BPis 100 mmHg. She refuses to have her orthostaticvital signs recorded because of severe light head-edness.Laboratory values are as follows: Serum Na 122mmol/L, serum osmolality 264 mOsm/kg, bloodurea nitrogen (BUN) 42 mg/dl, creatinine 1.0 mg/dl, glucose 90 mg/dl, serum cortisol 26.6 �g/L,urine Na 160 mmol/L, and urine osmolality 504mOsm/kg. The patient is given 4 L of isotonicsaline intravenously.

Which ONE of the following is the expectedresponse to this treatment?

A. It will correct the orthostatic symptoms butnot affect the hyponatremia.

B. It will correct the orthostatic symptoms butexacerbate the hyponatremia.

C. It will not correct the orthostatic symptomsor the hyponatremia.

D. It will not correct the orthostatic symp-toms, but it will correct the hyponatremia.

E. It will correct the orthostatic symptomsand the hyponatremia.

2. A 24-year-old with insulin-dependent diabetesand a history of cocaine and alcohol abuse,a dilated cardiomyopathy (Ejection Fraction20%), and ESRD treated with hemodialysis isadmitted on a Sunday night with complaints ofshortness of breath, abdominal pain, and vom-iting after skipping a dialysis treatment andstopping her insulin. She is obtunded, does notrespond to questions, and exhibits Kussmaul’srespirations. BP is 200/100 mmHg, HR is 100,respirations are 30, and temperature is 36°C.Oxygen saturation on 4 L/min nasal prong O2

is 90%. Jugular veins are distended to the ear-lobes, there are bilateral crackles at the lungbases and a prominent S3 gallop, the liver isdistended and tender, the abdomen is slightlydistended with hypoactive bowel sounds, andthere is 1� pretibial edema. There is a blizzardand there is going to be a substantial delay inassembling the dialysis team.Laboratory data are as follows: Serum Na 108mmol/L, serum K 7.2 mmol/L, serum Cl 75mmol/L, serum HCO3 6 mmol/L, BUN 112 mg/dl,creatinine 17 mg/dl, Ca 7 mg/dl, PO4 9 mg/dl,glucose 1800 mg/dl, and plasma osmolality 386mOsm/kg. Blood pH is 7.10, and PCO2 is 20mmHg. Electrocardiogram showed peaked Twaves with QRS 0.10.

In addition to insulin, which ONE of thefollowing is the BEST initial therapy?

A. 50 ml of 1 M NaHCO3 over 15 minutes,times three; 1 ampule of calcium glu-conate; and 30 g of sodium polystyrenesulfonate

B. 1 L of 0.9% saline over 1 hour, 1 ampuleof calcium gluconate, and 30 g of sodiumpolystyrene sulfonate

C. No intravenous fluids, 1 ampule of cal-cium gluconate, and 30 g of sodium poly-styrene sulfonate

D. 1 L of 0.45% over 4 hours

E. No additional therapy other than insulin

� 3. A 30-year-old man, despondent over a recentlydiagnosed brain tumor, goes on a drinking bingeand consumes several quarts of beer in a shortperiod. He is complaining of a headache andvomiting and is mildly confused.On physical examination, weight is 100 kg, BPis 120/80 mmHg, HR is 100. There is no jugularvenous distention, pupils are equal and reactive,ocular fundi are normal, chest is clear, heart hasno gallop, there is no edema, and there are no


183

focal neurologic findings. Laboratory data are asfollows: Serum Na 122 mmol/L, serum K 3.2mmol/L, serum Cl 88 mmol/L, serum HCO3 29mmol/L, BUN 4 mg/dl, creatinine 0.6 mg/dl,glucose 90 mg/dl, osmolality 270 mOsm/kg,aspartate aminotransferase 220, alanine amino-transferase 80, bilirubin 2.2, alkaline phospha-tase 120, urine Na 150 mmol/L, and urine os-molality 420 mOsm/kg.

Which ONE of the following is the BESTinitial therapy?

A. 100 ml of 3% saline over 15 minutes,times three

B. 3% saline at 20 ml/h for 24 hours

C. Fluid restriction

D. Isotonic saline 1 L over 1 hour

E. Conivaptan 20-mg bolus

4. A 63-year-old woman with a history of hyper-tension treated with losartan/hydrochlorothia-zide, 50.0/12.5 mg/d, presents with 5 days ofcough, diarrhea, anorexia, and progressiveweakness. She denies use of laxatives or herbalsupplements.

On physical examination, she is lethargicbut awake and oriented with slurred speech.Weight is 60 kg, BP is 138/70 mmHg supine and120/90 mmHg standing, HR is 88/min supineand 129/min standing, and temperature is 37°C.Oxygen saturation is 99% on room air. Jugularveins are just visible at 30°, lungs are clear, hearthas no murmur or gallop, and there is no edema.Neurologic exam shows 3/5 proximal muscleweakness with decreased deep tendon reflexes.Laboratory values are as follows: Serum Na 100mmol/L, serum K 1.7 mmol/L, serum Cl �60mmol/L, serum HCO3 37 mmol/L, BUN 10 mg/dl,creatinine 0.6 mg/dl, glucose 125 mg/dl, urine Na22 mmol/L, urine K 50 mmol/L, and urine osmo-lality 550 mOsm/kg.She is treated with 2 L of 0.9% saline and potas-sium replacement. On the second hospital day, herserum sodium concentration is 118 mmol/L, herserum potassium is 3.0 mmol/L, muscle weaknesshas improved, and she has no complaints. On thethird hospital day, her serum sodium is 123mmol/L and her serum potassium is 3.3 mmol/L,

but she is no longer responding to questions. Phys-ical examination reveals an oxygen saturation of90%, increased muscle tone in the lower extremi-ties, and bilateral Babinski reflexes. Magnetic res-onance imaging of the brain is normal.

In addition to improving oxygenation,which ONE of the following is the BESTtreatment for these findings?

A. KCl to increase the serum potassium to�4 mmol/L

B. 3% saline to increase the serum sodium to128 mmol/L

C. KCl to increase the serum potassium to�4 mmol/L and 3% saline to increase theserum sodium to 128 mmol/L

D. 1 g of solumedrol and 3% saline to in-crease the serum sodium to 128 mmol/L

E. Desmopressin and 5% dextrose in water tolower her serum sodium to 118 mmol/L

5. A 30-year-old who has schizophrenia and hasbeen treated with lithium in the past developstrigeminal neuralgia and is started on carbamaz-epine. Two days later, she presents with a majormotor seizure. On admission, she is comatosewith a BP of 200/100 mmHg. Weight is 50 kg.Neurologic examination reveals bilateral Babin-ski reflexes but no other localizing findings.Laboratory values are as follows: Serum Na 116mmol/l, serum K 3.4 mmol/L, serum Cl 80mmol/L, serum HCO3 10 mmol/L, BUN 5 mg/dl, creatinine 0.5 mg/dl, glucose 90 mg/dl, urineosmolality 340 mOsm/kg, urine Na 110 mmol/L, andurine K 20 mmol/L.Computed tomography scan of the head shows dif-fuse cerebral edema. In the emergency department(ED), she is given 1 L of 0.9% saline and then 300 mlof 3% NaCl over 3 hours. Her serum sodium con-centration increases to 122 mmol/L, but over the next6 hours she puts out 5 L of urine. Repeat urinechemistries obtained during her diuresis show osmo-lality 360 mOsm/kg, Na 120 mmol/L, and K 10mmol/L. After desmopressin, urine osmolality in-creases to 486 mOsm/kg.

Which ONE of the following is the BESTexplanation for her large urine output?


A. A physiologic water diuresis caused by herhyponatremia

B. Cerebral salt wasting

C. Diabetes insipidus (DI) as a result of braindeath

D. Nephrogenic diabetes as a result of previ-ous use of lithium

E. A physiologic solute diuresis caused byvolume expansion

6. A 5-year-old boy develops severe thirst, drink-ing up to twenty 8-oz glasses of water daily, andexcessive urination, going to the bathroom everyhour and wetting the bed at night. His father,paternal grandfather, and paternal aunt have DIthat responds to desmopressin. After a 7-hourdehydration test, the patient’s plasma osmolalityis 285 mOsm/kg and urine osmolality is 477mOsm/kg.

On the basis of these findings, which ONEof the following is the BEST course ofaction?

A. Tell the parents that the dehydration test isdiagnostic of psychogenic polydipsia andrecommend a psychiatric evaluation.

B. Recommend magnetic resonance imagingand if it shows that a pituitary bright spotis present, excluding neurogenic DI, thenrefer the child for a psychiatric evaluation.

C. Tell the parents that the child may havefamilial DI but that it would be best topostpone therapy because early treatmentwith desmopressin has been shown to ac-celerate the loss of vasopressin-secretingneurons in animal models.

D. Begin therapy with desmopressin becausethe dehydration test is diagnostic of DI.

E. Recommend a DNA sequence analysis ofthe AVP gene and treat with desmopressinif it confirms the presence of a mutation inthe gene encoding neurophysin II.

7. A 75-year-old man with a history of chroniclymphocytic leukemia and stage 3 renal diseasetreated with an angiotensin-converting enzymeinhibitor is seen in his doctor’s office complain-ing of fatigue. Outpatient laboratory data show a

hematocrit of 26, white blood cell count of250,000, predominantly mature lymphocytes,potassium 7.2 mmol/L with no other electrolyteabnormalities, and creatinine 2.0 mg/dl. Thepatient is sent to the ED for further treatment.The electrocardiogram on arrival in the EDshows no abnormalities.

Which ONE of the following should be car-ried out next?

A. Filter the blood sample in the laboratory toseparate the lymphocytes from the plasma.

B. Compare the plasma potassium with theserum potassium.

C. Start emergency treatment for hyperkale-mia.

D. Send the specimen to the laboratorythrough the tube system for a crisis mea-surement of whole-blood potassium.

E. Repeat the test using a blood sample handcarried to the laboratory and allowed toclot.

� 8. A mother brings in her 10-year-old daughter forevaluation of polydipsia and polyuria. Her father(the girl’s maternal grandfather) and two neph-ews have been diagnosed as having DI that doesnot respond to desmopressin. The girl’s motherand father are completely healthy and have noproblems with polyuria or polydipsia. The girl’surine tests negative for glucose, and a randomspecific gravity of her urine is 1.010.

Which ONE of the following is the BESTinterpretation of these findings?

A. DI in other family members is likely arecessively inherited form of nephrogenicDI (NDI).

B. DI in other family members is likely anX-linked form of NDI.

C. DI in other family members is likely anautosomal dominant form of NDI.

D. DI in other family members is likely anX-linked form of NDI, but the girl cannotbe affected because her mother is normal.

E. The girl may be a compound heterozygotefor X-linked NDI and the recessively in-herited form of NDI.


9. A 6-month-old infant is admitted in a lethargiccondition after several days of severe diarrhea.BP is 90/50 mmHg, and HR is 110 bpm. Mu-cous membranes are dry, and skin turgor is poor.The abdomen is soft with hyperactive bowelsounds. Laboratory data show a serum sodiumof 174 mmol/L.

Which ONE of the following treatmentsshould be used in an attempt to avoid neu-rologic complications?

A. An initial bolus of isotonic saline and cor-rection to a serum sodium of 150 mmol/Lin the first 12 hours

B. An initial bolus of isotonic saline followedby gradual correction to a serum sodiumof 162 mmol/L in the next 24 hours

C. No initial bolus of isotonic saline andgradual correction to a serum sodium of150 mmol/L in the first 24 hours

D. No initial bolus of isotonic saline andgradual correction to a serum sodium of162 mmol/L in the next 24 hours

10. A 50-year-old woman visiting from Sri Lanka isbrought in by her daughter with severe nausea,vomiting, and abdominal pain. The daughterfound a suicide note and a partially eaten pack-age containing what she believes to be Oleanderseeds, a poison commonly used in suicide at-tempts in Southern Asia. BP is 100/50 mmHg,and HR is 40 bpm. Electrocardiogram showscomplete atrioventricular block with a QRS of0.10; serum potassium is 7 mmol/L, and serumcreatinine is 0.6 mg/dl.

Which ONE of the following is the BESTtherapeutic approach to this problem?

A. Emergency hemodialysis

B. Calcium gluconate, insulin, and kayexalate

C. Calcium gluconate, insulin, and furosemide

D. Insulin and digoxin-specific Fab antibody

E. Insulin, saline, and fludrocortisone

11. A 30-year-old man with hypertension developssevere hypokalemia during treatment with a thi-azide diuretic. An evaluation reveals a lowrenin-aldosterone ratio and an adrenal mass. Onthe second postoperative day, after resection of

the adrenal tumor, the serum potassium is foundto be 6 mmol/L despite a normal serum creati-nine.

Which ONE of the following is MOSTlikely to explain the high potassium?

A. The original diagnosis was incorrect andthe patient most likely has Gordon syn-drome

B. The patient had one functioning adrenalgland and now has surgically induced Ad-dison disease

C. Hyperkalemia was caused by excessivepotassium replacement and should resolvein a few hours

D. Hyperkalemia was caused by the use ofdepolarizing muscle relaxants during sur-gery and should resolve in a few hours

E. Hyperkalemia is caused by suppression ofaldosterone secretion in the normal adrenalby the adenoma and should resolve in afew weeks

12. A 50-year-old man sustains multiple injuriesfrom a motor vehicle accident. He developsacute renal failure from rhabdomyolysis andhypernatremia from infusion of 3% saline usedto treat increased intracranial pressure. Fivedays after admission, his intracranial pressure isnormal but azotemia has progressed.Current laboratory data reveal the following:Serum Na 165 mmol/L, serum K 5.9 mmol/L,serum Cl 129 mmol/L, serum HCO3 18 mmol/L,BUN 95 mg/dl, creatinine 7 mg/dl, glucose 140mg/dl.

Which ONE of the following is the BESTstrategy?

A. Conventional hemodialysis

B. Conventional hemodialysis supplementedwith 50 ml of 3% saline every hour duringdialysis

C. Venovenous hemofiltration with standardreplacement fluid

D. Venovenous hemofiltration with standardreplacement fluid supplemented with 50ml of 3% saline after 6 hours

E. Venovenous hemofiltration with 25 ml of


30% saline or 30 ml of 23% saline addedto each 5-L bag of replacement solution

13. An African American man with stage 3 chronickidney disease (CKD) and recalcitrant hyperten-sion (BP 160/95 mmHg) despite treatment withan angiotensin-converting enzyme inhibitor anda diuretic sees his internist, who recommendsadditional treatment with spironolactone. Hisserum potassium is 4.4 mmol/L, and estimatedGFR (eGFR) is 47 ml/min.

Which ONE of the following is the MOSTlikely clinical outcome?

A. BP 125/80 mmHg, K 4.7 mmol/L, eGFR25 ml/min

B. BP 165/100 mmHg, K 4.7 mmol/L, eGFR45 ml/min

C. BP 125/80 mmHg, K 4.7 mmol/L, eGFR45 ml/min

D. BP 125/80 mmHg, K 6.5 mmol/L, eGFR45 ml/min

14. A 36-year-old woman develops amenorrhea,followed a few months later by progressiveanorexia, abdominal pains, and postural dizzi-ness. Physical examination reveals increasedskin pigmentation and BP of 100/50 mmHgsupine and 80/40 standing.Laboratory data reveal the following: Serum Na110 mmol/L, serum K 5.9 mmol/L, serum Cl 75mmol/L, serum HCO3 19 mmol/L, BUN 39mg/dl, creatinine 1.8 mg/dl, blood glucose 90mg/dl, urine osmolality 650 mOsm/kg, urinesodium 60 mmol/L, and urine potassium 24mmol/L.She is treated with hydrocortisone and 2 L ofisotonic saline. Eight hours later, her urine out-put increases to 1 L/h.

Which ONE of the following treatmentsshould be instituted at this time?

A. Match urine output with 0.9% NaCl

B. Give fludrocortisone 0.1 mg orally

C. Start 5% dextrose in water at 250 ml/h

D. Increase the dosage of hydrocortisone

E. Give desmopressin

15. A 20-year-old woman has had NDI diagnosed in

childhood. She has a history of deep venous throm-boembolic and pulmonary embolus. One of her foursisters but no other family members has the disease.

Which ONE of the following is the MOSTlikely explanation for her increased episodesof venous thrombosis?

A. The patient has an activating mutation ofthe vasopressin 2 receptor (V2R), whichresults in increased release of von Wille-brand factor from endothelial cells.

B. The patient has an inactivating mutation ofthe V2R, which results in failure to inhibitrelease of von Willebrand factor from en-dothelial cells.

C. The patient has a mutation of aquaporin 2,resulting in its failure to traffic to the en-dothelial cell membrane, where it normallyinhibits release of von Willebrand factor.

D. The patient has a normal V2R and persis-tently elevated levels of vasopressin,which results in increased release of vonWillebrand factor.

16. A 20-year-old woman with paroxysmal noctur-nal hemoglobinuria develops severe diuretic re-sistant ascites caused by hepatic vein thrombo-sis. A portosystemic shunt is proposed.

Which ONE of the following would beMOST successful in increasing sodiumexcretion in the urine?

A. A side-to-side anastomosis.

B. An end-to-side anastomosis.

C. The two shunts would be equally effective.

D. Neither form of shunt would improve so-dium excretion.

�17. A 50-year-old man with hepatic cirrhosis devel-ops progressive edema and ascites on spirono-lactone 100 mg/d. The serum potassium is 4.5mmol/L, and the serum creatinine is 0.6 mg/dl.A spot urine sample shows a urine Na concen-tration of 40 mmol/L and a urine creatinineconcentration of 100 mg/dl.

Which ONE of the following is the BESTapproach?

A. Continue the current dosage of spironolac-


tone and request a dietary consultation toinstruct the patient in a low-sodium diet.

B. Increase the dosage of spironolactone to200 mg/d

C. Order a large-volume paracentesis

D. Start furosemide 40 mg twice daily andcontinue spironolactone

E. Start furosemide 40 mg twice daily anddiscontinue spironolactone

�18. Milk-alkali syndrome was first described in theearly 1900s, when it developed as a result of thetreatment regimen used for peptic ulcer disease.Introduction of modern treatments for pepticulcers led to a marked reduction in the frequencyof this disorder. Now milk-alkali syndrome isagain being commonly diagnosed. There areseveral important differences between the syn-drome currently encountered and the disorderthat was seen in the early 1900s.

Which ONE of the following statementsabout the modern milk-alkali syndrome isCORRECT?

A. The calcium levels now are lower than inthe past.

B. The degree of kidney dysfunction now isworse than in the past.

C. The inorganic phosphorous levels now arelower than in the past.

D. The degree of alkalosis now is worse thanin the past.

19. Aminoglycosides can produce a clinical syn-drome that mimics which ONE of the follow-ing tubule disorders?

A. Gittelman syndrome

B. Liddle syndrome

C. Bartter syndrome

D. Pseudohypoaldosteronism type II

E. Gordon syndrome

20. A 58-year old man with CKD (eGFR 29 ml/min)as a result of hypertension is placed on bicar-bonate supplements to reduce his risk for pro-gression of his CKD.

Which ONE of the following effects of

bicarbonate administration is MOST likelyto explain the beneficial effect?

A. Reduced ammonia synthesis

B. Reduced renal tubular energy expenditure

C. Reduced intrarenal BP

D. Reduced angiotensin formation

21. A 60-year-old man presents to the ED withsevere cellulitis of the right arm. He remembersfalling down and lacerating his elbow severaldays before. He has no other history, and he ison no medications. Physical examination revealsmarked erythema of the lower half of his rightarm. He is febrile to 102°F. BP is 90/70 mmHg,and white blood cell count is 18,000/mm3 with aleft shift. A presumptive diagnosis of severesepsis is made. Chemistries reveal the follow-ing: Na 136 mEq/L, K 4.0 mEq/L, Cl 100mEq/L, HCO3 17 mEq/L, and lactate 6 mEq/L.Treatment with antibiotics and aggressive intra-venous fluid therapy is begun, and 2 hours later,a repeat lactate level is still 6 mEq/L.

Which ONE of the treatment options is theBEST choice?

A. Administer NaHCO3

B. Start continuous venovenous hemofiltra-tion

C. Give fomepazole

D. Give insulin and glucose

E. Maintain current therapy

�22. You are asked to evaluate a 40-year-old man whowas hospitalized after he sustained severe headtrauma in a motor vehicle accident. He was admit-ted to the intensive care unit, intubated, and se-dated. He was started on a morphine infusion forpain and a propofol infusion at 2 mg/kg per h forsedation and to help control intracranial hyperten-sion. The propofol dosage was titrated up to 5mg/kg per h to maintain cerebral perfusion pres-sure above 70 mmHg. On the fifth day of hospi-talization, several new biochemical abnormalitieswere noted. He had developed an anion gap met-abolic acidosis and high triglyceride levels, andcreatine phosphokinase was 15,000 IU/ml. At thistime, his vital signs were acceptable with a BP of150/90 mmHg and pulse of 110/min.


Which ONE of the following is the MOSTlikely acid that is accumulating?

A. L-lactic acid

B. Keto acids (acetoacetic and beta-hydroxylbutyric)

C. D-lactic Acid

D. Glyoxalic acid

E. Pyroglutamic acid (5-oxoproline)

23. A 24-year-old patient with a seizure disorder istreated with topiramate. After 2 weeks, a bio-chemical profile reveals the following results:Serum Na 141 mmol/L, serum K 3.4 mmol/L,serum Cl 110 mmol/L, and serum HCO3 16mmol/L.

Which ONE of the following is the MOSTlikely cause of this electrolyte pattern?

A. Distal renal tubular acidosis as a result ofa membrane backleak of hydrogen ion

B. A primary defect in renal ammoniagenesis

C. Inhibition of carbonic anhydrase

D. Stimulation of the respiratory center

E. Secondary hyperaldosteronism

24. A 45-year-old man presents with flank pain andhematuria and is found to have a right-sidedkidney stone. He has a history of severe mi-graine headaches managed with triptans. In ad-dition, his physician has prescribed a number ofmedications for him in the past, including cal-

cium channel blockers and �-adrenergic antag-onists, in an attempt to prevent the headachesfrom occurring so frequently. More recently,topiramate was found to be effective, and thepatient has been taking this medication for thepast year.

On the basis of this history, the compositionof the kidney stone is MOST likely to bewhich ONE of the following?

A. Calcium phosphate

B. Calcium carbonate

C. Uric acid

D. Calcium oxalate

E. Cysteine

25. A 45-year-old man is admitted to the hospitalwith an acute asthma attack. After treatmentwith glucocorticoids and � agonist inhalants, hiscondition improved so that mechanical ventila-tion was not necessary. His BP is 150/80 mmHg.

Which ONE of the following sets of arterialblood gases is Most likely to be obtained forthis patient?

Table 1.

pH PCO2 Torr HCO3 mEq/L

A 7.55 26 18B 7.40 20 13C 7.25 25 10D 7.25 55 28


Nephrology Self-Assesment Program

Core Knowledge QuestionsFluid, Electrolyte, and Acid-Base Disturbances

1. A 65-year-old man presents with the chief complaint of progressive weakness over the past several months.He is normotensive, and his physical examination is unremarkable. Laboratory studies reveal thefollowing: Na 135 mmol/L, Cl 105 mmol/L, K 3.0 mmol/L, HCO3 18 mEq/L, creatinine 1.8 mg/dl, BUN22 mg/dl, glucose 110 mg/dl, PCO2 28 Torr, pH 7.33, hematocrit 25%, white blood cell count 5600/mm3,and platelets 340,000/mm3; urinalysis shows trace protein, 1� glucose, normal sediment, and 24-h urineprotein of 4.8 g.

Which ONE of the following is a CHARACTERISTIC of the renal abnormality present in this pa-tient?

A. Evidence of nephrocalcinosis on kidney ureters-bladder x-ray of the abdomen.

B. The serum HCO3 concentration will increase after the administration of oral bicarbonate at 80 mEq/dbut then decrease to 18 mmol/L after the therapy is discontinued.

C. Bicarbonate therapy will cause the serum K to decline slightly as a result of a shift into cells.

D. The urine pH will be persistently alkaline.

E. The urine anion gap will be negative.

2. A 78-year-old Caucasian woman is brought to the emergency department secondary to abdominal pain.During the past 2 months, she has noticed periumbilical pain that is brought on by food ingestion. As a resultof worsening pain, the patient began taking acetaminophen 4 g/d for the past week. Medical history issignificant for stable two-block claudication and transient ischemic attack.Physical examination reveals the following: Temperature of 38.1°C, pulse 98 bpm, BP of 158/88 mmHg,left-sided carotid bruit, normoactive bowel sounds, abdominal bruit, and pain to deep palpation in themid-epigastrium without rebound.Admission laboratory studies reveal the following: Na 138 mmol/L, K 4.9 mmol/L, Cl 102 mmol/L, HCO3

7 mEq/L, creatinine 1.4 mg/dl, BUN 30 mg/dl, glucose 126 mg/dl, serum osmolality 295 mOsm/L. Arterialblood gas shows pH of 7.17, PCO2 of 18 Torr, and PO2 of 104 Torr; urinalysis shows pH of 5.5, trace ketones,and negative sediment.The anion gap acidosis in the presence of vascular disease and history consistent with intestinal angina ledto a diagnosis of ischemic bowel. The patient was taken to the operating room for an exploratory laparotomy.No evidence of ischemic bowel was found. A lactate level sent earlier came back at 2.8 mEq/L.

Which ONE of the following is the MOST likely cause of the metabolic acidosis in this patient?

A. Pyroglutamic acidosis resulting from the administration of acetaminophen.

B. D-Lactic acidosis as a result of bacterial overgrowth.

C. Malignant hyperthermia with secondary lactic acidosis.

D. Diabetic ketoacidosis.

E. Salicylate toxicity.

3. A 32-year-old man is brought to the emergency department and is described as combative and agitated. Afriend describes the patient as “into alcohol and all kinds of drugs and inhalants” but is unable to specify whatspecifically he may have taken. On physical examination, the patient is uncooperative and is slightlydisoriented. Laboratory studies reveal the following: Na 140 mmol/L, K 3.1 mmol/L, Cl 111 mmol/L, HCO3


190

21 mEq/L. Arterial blood shows pH of 7.5, PCO2 21 Torr, PO2 of 86 Torr, serum ketones present in 1:4dilution, urine ketones 4�, and urine glucose 1�.

Which ONE of the following is the MOST likely cause of the clinical syndrome?

A. Toluene inhalation.

B. Ethylene glycol intoxication.

C. Diabetic ketoacidosis.

D. Isopropyl alcohol ingestion.

E. Alcoholic ketoacidosis.

4. A 71-year-old woman who has had nocturia for several years is admitted to the hospital secondary toincreasing weakness and frequency of urination. She has been well until 2 days ago, when she felt weak andcould not climb the stairs to her apartment. She has a history of duodenal ulcer many years ago that respondedto intensive antacid therapy. She currently takes calcium carbonate for treatment of osteoarthritis, and shetakes bicarbonate of soda for heartburn. She has a 40 pack-year history of smoking.On physical examination, she is frail and oriented only to person. Pulse is 106/min, and BP is 110/80 supineand 90/70 mmHg sitting. The remainder of the examination is normal.Laboratory studies reveal the following: Hematocrit 41, Na 152 mmol/L, K 3.0 mmol/L, Cl 100 mmol/L,HCO3 39 mEq/L, BUN 98 mg/dl, creatinine 7.1 mg/dl, Ca 14.4 mg/dl, phosphate 6.3 mg/dl, serum1,25-dihydroxyvitamin D 30 pg/ml (35 to 85 pg/ml), parathyroid hormone 16 pg/ml (30 to 50 pg/ml).Urinalysis shows specific gravity of 1.007, trace protein, Na of 49 mmol/L, creatinine of 70 mg/dl, and urineosmolality of 260 mOsm/kgH2O. Renal ultrasound shows normal-sized kidneys and no hydronephrosis.

The clinical and laboratory findings are MOST consistent with which ONE of the following?

A. Vitamin D intoxication.

B. Chronic kidney disease as a result of longstanding hypertension.

C. Multiple myeloma.

D. Milk-alkali syndrome.

E. Primary hyperparathyroidism.

5. A 38-year-old woman with a strong family history of cardiovascular diseases and hypertension recentlyreceived a diagnosis of essential hypertension. Her BP, on three separate measurements, averages 154/94mmHg. Current medications include a daily multivitamin and birth control pills. The physical examinationand laboratory examination are normal. Because the patient is using birth control pills, her primary carephysician was comfortable prescribing lisinopril 10 mg/d. One month later, the patient returns for follow-up.BP is 142/88 mmHg. Laboratory examination shows the following: Na 140 mEq/L, K 5.5 mEq/L, Cl 100mEq/L, HCO3 22 mEq/L, creatinine 0.8 mg/dl, and BUN 10 mg/dl. The physician refers the patient to anephrologist for hyperkalemia evaluation.

Which ONE of the following is the MOST likely risk factor for hyperkalemia development afterprescribing an angiotensin-converting enzyme inhibitor for this patient?

A. High-grade bilateral renal artery stenosis.

B. Pseudohypoaldosteronism type II.

C. Acquired adrenal insufficiency.

D. Mineralocorticoid-blocking activity in birth control pill.

E. Daily ingestion of bananas.


Answers and Explanations

1. Answer: B. The serum HCO3 concentration will increase after the administration of oral bicarbon-ate at 80 mEq/d but then decrease to 18 mmol/L after the therapy is discontinued.The patient presents with hypokalemia and normal gap acidosis, in association with glycosuria in thesetting of a normal plasma glucose concentration. These findings suggest the Fanconi syndrome and atype II or proximal renal tubular acidosis (RTA). The large amount of proteinuria but only a tracepositive dipstick suggests the excretion of a cationic protein, suggesting that multiple myeloma as theunderlying cause. With a proximal RTA, administration of bicarbonate will only transiently increase theserum bicarbonate concentration, and, once discontinued, the plasma concentration will fall to a lowerTmax, usually approximately 16 to 18 mEq/L. Choice A is incorrect because distal and not proximal RTAis associated with nephrocalcinosis. An adverse effect of bicarbonate therapy is a worsening ofhypokalemia as a result of increased renal K wasting as distal Na delivery is increased. Thus, choice Cis incorrect. Choice D is incorrect because the urine will acidify in proximal RTA once the serumbicarbonate falls to the lower Tmax after discontinuation of oral bicarbonate therapy. The urine anion gapis not negative in proximal RTA because proximal tubular dysfunction also impairs renal ammoniagen-esis, making choice E incorrect.

● Hoorn EJ, Zietse R: Combined renal tubular acidosis and diabetes insipidus in hematological disease.Nat Clin Pract Nephrol 3: 171–175, 2007● Lacy MQ, Gertz MA: Acquired Fanconi’s syndrome associated with monoclonal gammopathies.Hematol Oncol Clin North Am 13: 1273–1280, 1999

2. Answer: A. Pyroglutamic acidosis resulting from the administration of acetaminophenThe patient presents with an anion gap metabolic acidosis in association with ingestion of large amounts ofacetaminophen. This drug leads to depletion of glutathione, thereby altering the gamma glutamyl cycle insuch a way that pyroglutamic acid is overproduced. Choice B is incorrect because the clinical setting doesnot suggest a bacterial overgrowth syndrome, making D-lactic acidosis unlikely. Choice C is incorrectbecause neither the clinical setting nor the signs and symptoms in this case suggest the presence of malignanthyperthermia. The absence of ketonuria and the normal serum glucose exclude diabetic ketoacidosis, sochoice D is incorrect. Salicylate toxicity often has features of both an anion gap acidosis and respiratoryalkalosis with the latter being more prominent in adults. As a result, choice E is incorrect.

● Fenves AZ, Kirkpatrick HM III, Patel VV, Sweetman L, Emmett M: Increased anion gap metabolicacidosis as a result of 5-oxoproline (pyroglutamic acid): A role for acetaminophen. Clin J Am Soc Nephrol1: 441–447, 2006Peter JV, Rogers N, Murty S, Gerace R, Mackay R, Peake SL: An unusual cause of severe metabolicacidosis. Med J Aust 185: 223–225, 2006● Green TJ, Bijlsma JJ, Sweet DD: Profound metabolic acidosis from pyroglutamic acidemia: An undera-ppreciated cause of high anion gap metabolic acidosis. CJEM 12: 449–452, 2010

3. Answer: D. Isopropyl alcohol ingestionA patient with a history of substance abuse presents with mild hypokalemia and respiratory alkalosis. Theworkup is noteworthy for positive serum and urine ketones. This clinical presentation best fits with isopropylalcohol ingestion, or choice D. Ingestion of this substance leads to acetone production, giving rise to positiveketones. There is no consumption of bicarbonate in the production of acetone, so metabolic acidosis is not afeature of this intoxication. The respiratory alkalosis is likely due to alcohol withdrawal or the underlyingagitation. The hypokalemia may be due to a shift of K into cells as a result of the alkalosis. Toluene inhalation,ethylene glycol intoxication, diabetic ketoacidosis, and alcoholic ketoacidosis all are conditions associatedwith metabolic acidosis, which is not present in this case.


● Zaman F, Pervez A, Abreo K: Isopropyl alcohol intoxication: A diagnostic challenge. Am J Kidney Dis 40:E12, 2002Abramson S, Singh AK: Treatment of the alcohol intoxications: Ethylene glycol, methanol and isopropanol.Curr Opin Nephrol Hypertens 9: 695–701, 2000Jammalamadaka D, Raissi S: Ethylene glycol, methanol and isopropyl alcohol intoxication. Am J Med Sci339: 276–281, 2010

4. Answer: D. Milk-alkali syndromeThe patient presents with metabolic alkalosis, acute renal failure, and hypercalcemia. The history describesa patient who likely is taking large quantities of antacids for treatment of dyspepsia. These features stronglysuggest a diagnosis of milk-alkali syndrome. The increased frequency of urination may be the result of a renalconcentrating defect caused by hypercalcemia. Vitamin D intoxication (choice A) can be a cause ofhypercalcemia but is excluded by the low levels of active vitamin D. A patient with chronic kidney diseaseas a result of hypertension (choice B) would be expected to present with small kidneys. A serum calciumconcentration of 14.4 would not be expected in such a patient. Multiple myeloma (choice C) is a considerationbut is not the best answer given the clinical features in this case. Anemia and a low anion gap are often presentin patients with myeloma. Primary hyperparathyroidism is excluded by the suppressed parathyroid hormonelevels, making choice E incorrect.

● Felsenfeld AJ, Levine BS: Milk alkali syndrome and the dynamics of calcium homeostasis. Clin J Am SocNephrol 1: 641–654, 2006● Picolos M, Orlander P: Calcium carbonate toxicity: The updated milk-alkali syndrome: Report of 3 casesand review of the literature. Endocr Pract 11: 272–280, 2005● Beall D, Scofield R: Milk-alkali syndrome associated with calcium carbonate consumption: Report of 7patients with parathyroid hormone levels and an estimate of prevalence among patients hospitalized withhypercalcemia. Medicine (Baltimore) 74: 89–96, 1995● Patel AM, Goldfarb S: Got calcium? Welcome to the calcium-alkali syndrome. J Am Soc Nephrol 21:1440–1443, 2010

5. Answer: D. Mineralocorticoid blocking activity in birth control pillThe oral contraceptive Yasmin-28 contains the non–testosterone-derived progestin drospirenone, whichpossesses mineralocorticoid-blocking effects similar to what is seen with spironolactone. The product labelingrecommends K� monitoring in the first month after prescribing the drug for patients who are receiving K�

supplements, renin-angiotensin blockers, or nonsteroidal anti-inflammatory drugs. Despite this recommen-dation, there are many instances in which monitoring does not occur or patients are prescribed thecontraceptive in the setting of other drugs that either provide a K� load or interfere in renal K� secretion.There is nothing in the clinical history to suggest the disease states provided in choices A, B, and C. The K�

load contained in bananas would be unlikely to cause hyperkalemia in the setting of normal renal function,making choice E incorrect.

● McAdams M, Staffa J, Pan G: The concomitant prescribing of ethinyl estradiol/drospirenone and potentiallyinteracting drugs. Contraception 76: 278–281, 2007● Eng P, Seeger J, Loughlin J, Oh K, Walker A: Serum potassium monitoring for users of ethinylestradiol/drospirenone taking medications predisposing to hyperkalemia: Physician compliance and survey ofknowledge and attitudes. Contraception 75: 101–107, 2007● Schutt B, Kunz M, Blode H: Coadministration of estradiol/drospirenone and indomethacin does not causehyperkalemia in healthy postmenopausal women: A randomized open-label crossover study. J Clin Pharma-col 47: 774–781, 2007


2011fluids nephsap

Documents