npt el final

Editorials

Echoes of the past?*

T he hemodynamic abnormali-ties of sepsis, and their appro-priate treatment, are familiarto caregivers in the intensive

care unit (ICU) and the emergency de-partment. The key is early recognition,resuscitation with fluids and vasopressorsif required, and treatment of the under-lying cause. This approach is the basis ofthe hemodynamic management guide-lines of the Surviving Sepsis Campaign(1), and of the campaign’s 6-hr resuscita-tion care bundle. Yet, this consensushides fundamental difficulties that be-devil our ability to improve care, espe-cially for patients who do not stabilize orreverse their shock after initial resuscita-tion and require more advanced and pro-longed support. Once into this secondphase, agreement disappears as to themost appropriate hemodynamic monitor-ing techniques, the correct targets andend points of treatment, and even whichfluids and vasoactive drugs to give. Prac-tice varies internationally, with a prefer-ence for crystalloid resuscitation in theUnited States, and greater use of artificialcolloids in Europe, although the resultsof recent large studies (2, 3) mean thatthese patterns are in a state of flux. Thereare also differences in the hemodynamicmonitoring used to guide therapy, andtherefore the paradigms of care. In gen-eral terms, there is a greater reliance onfilling pressures in the United States, anda more “mixed economy” in Europe, withvarious flow and volume measurementtechniques (transpulmonary thermodilu-tion, arterial pulse wave analysis, esoph-ageal Doppler) more widely employed.There is general agreement about the im-portance of using volume challenges anddemonstrating volume responsiveness,but less clarity about the best way of

doing this. This variation in practice pos-sibly is due to the gaps that still exist inour fundamental understanding of thepathophysiology of sepsis, because accu-rate description of complex physiology inthe ICU remains a challenge.

It is against this challenging back-ground that Dr. Viellard-Baron and col-leagues (4) publish their intriguing studyof the incidence of left ventricular hypo-kinesia in septic shock in this issue ofCritical Care Medicine. They assessed 67patients with septic shock on a daily basisfor the first 3 days of ICU admission usingechocardiography and again after wean-ing from vasoactive support in those whorecovered. Patients were treated withstandard supportive therapies, with fluidresuscitation being guided echocardio-graphically using the pattern of respira-tory cycle flow variation in the superiorvena cava during mechanical ventilation,an approach that the authors have previ-ously validated, and initial vasopressorsupport with norepinephrine. They foundthat primary global hypokinesia (definedas a left ventricular ejection fraction de-rived from the long axis view of �45%)was present in 26 patients (39%), andthat secondary global hypokinesia (de-fined similarly, but occurring in patientswhose initial study was within the normalrange) developed in a further 14 patients(21%). In these patients, dobutamine wasadded and the norepinephrine dosage re-duced, with epinephrine also used if hy-potension persisted. Nonsurvivors had asignificantly higher cardiac index on day1 than survivors (3.8 � 1.3 vs. 3.1 � 0.9,p � .017), although this difference disap-peared subsequently, but the authorsfound no convincing evidence of differ-ences between ejection fraction, left ven-tricular (LV) end-diastolic volume (EDV),or ventricular elastance (estimated by therelationship between systolic arterialpressure and left ventricular end-systolicvolume). The authors therefore concludethat left ventricular global hypokinesia iscommon in severe sepsis, that its pres-ence does not discriminate between sur-vivors and nonsurvivors, that its develop-ment may be influenced by the use ofnorepinephrine via an increase in left

ventricular afterload, and that it respondsto and is a possible indication for an ino-trope such as dobutamine.

Why are these findings important?Certainly, this is not the first paper todescribe the high incidence of global LVdysfunction in severe sepsis; indeed, thesame authors do exactly this in theirlarger previous series (5), reportingbroadly similar hemodynamic patterns tothose that they document on this occa-sion. What is different is the lower mor-tality in their current series, which maybe expected given the general improve-ments in supportive care that have oc-curred over the last decade, and the factthat they no longer observe the pattern ofhigher LVEDV and lower LV ejection frac-tion in nonsurvivors as compared withsurvivors that they saw in their firststudy. The association between severesepsis and LV dysfunction characterizedby ventricular dilation has been knownsince the work of two pioneering criticalcare research groups in the 1980s. Dr.Calvin and colleagues (6) demonstrateddecreased ejection fraction and increasedLVEDV using electrocardiogram-gatedradionuclide cineangiography in 1982.Dr. Parker and colleagues (7) expandedthis work in 1984 to demonstrate that thepattern of LV dilation and decreased ejec-tion fraction was characteristic of survi-vors recovering after several days and in-troduced the concept of adaptive LVdilation as a mechanism by which a heartwith a reduced ejection fraction could stillmaintain an adequate stroke volume. It isthis later concept that Dr. Viellard-Baronand colleagues are now challenging.

Were the early radionuclide studieswrong? Certainly, technology has movedon considerably since that time, and highquality bedside echocardiography is nowavailable in most ICUs, although usuallyfor traditional diagnostic assessmentrather than as a bedside monitor to guidetreatment. This latter approach is becom-ing more common as more ICUs acquirethe technology, and as training schemesfor ICU physicians develop, and the au-thors’ unit is a pioneer of this approach.There are also doubts as to the accuracyof the method used by Dr. Parker and

*See also p. 1701.Key Words: septic shock; echocardiography; car-

diac function; myocardial depression; hemodynamicsDr. Beale has received departmental support

from Philips Healthcare and research support fromLiDCO Ltd.

Copyright © 2008 by the Society of Critical CareMedicine and Lippincott Williams & Wilkins

DOI: 10.1097/CCM.0b013e3181761147

1950 Crit Care Med 2008 Vol. 36, No. 6

colleagues for generating absolute valuesfor LVEDV, because they derived strokevolume from thermodilution cardiac out-put studies using a pulmonary arterycatheter and ejection fraction from theradionuclide studies, then combined thetwo to back-calculate end-diastolic andend-systolic volumes. However, even withhigh-quality modern echocardiographicequipment, Doppler flow estimates fromthe LV outflow tract and the difference inventricular volumes in systole and dias-tole give different values for stroke vol-ume, and these inconsistencies raisedoubts about the conclusions of echo-based studies such as this in the minds ofnon–echocardiography intensivists moreused to conceptualizing hemodynamicsusing standard ICU monitoring technol-ogies. In any event, unlike the currentstudy, the authors’ prior study did find ahigher LVEDV in nonsurvivors. Just toconfound the situation further, a recentstudy of patients with severe sepsis pre-senting to the emergency room (8) foundthat a hyperdynamic left ventricle, as ev-idenced by an increased ejection fraction,was highly specific for severe sepsis, al-though the degree of echocardiographicdetail provided is limited. So what is thehemodynamic truth in severe sepsis—hyperkinetic (usually associated with va-sodilatation) or hypokinetic? Indeed, isthere a single characteristic pattern, or isthe situation more complex?

There is another fundamental prob-lem with studies that characterize sepsisin terms of a specific hemodynamic pro-file. This is the idea that there are dichot-omous categories that patients with se-vere sepsis can be classified into; in thisinstance, either ventricular hypokinesiaor not. In fact, the examinations withinthe authors’ current paper are point as-sessments in time along a physiologiccontinuum, when “time zero” is oftenunclear. The more extreme disease man-ifestations, perhaps including adaptiveventricular dilation, probably only occurin some patients some of the time, soapplying group analysis to these physio-logic phenomena risks missing individualevents, and also losing the opportunity tounderstand them. Furthermore, in severesepsis hemodynamics also are influencedby fluid loading, vasopressor and inotropeuse, and host response. Weaknesses in allour current monitoring technologiesmake it impossible to be certain that pa-tients have been treated precisely thesame, however hard the authors try tostandardize these factors. So whether

adaptive LV dilation really does occur,how often and in whom, and what itmeans is less important than recognizingthat ventricular hypokinesia is common-place and knowing when it is present.This means using a bedside monitoringtechnology that allows its detection; i.e.,making echocardiography much morereadily available in the ICU.

The other issue is the potential thera-peutic consequence of this study’s find-ings. The Surviving Sepsis Campaignguidelines recommend the use of dobut-amine as an inotrope in severe sepsis(grade 1C) in the presence of myocardialdysfunction as suggested by elevated car-diac filling pressures and low cardiac out-put (1). Unfortunately, in their currentstudy, the authors do not provide anyfilling pressure data, and it is not clear ifthey used the central venous pressure atall in the management of this patientgroup, although in their previous studythat found differential ventricular kinet-ics between survivors and nonsurvivorsthey fluid-loaded patients to a central ve-nous pressure of 12 mm Hg. There hasbeen debate as to the value of filling pres-sures as a reliable guide to fluid therapyin sepsis, and this area remains highlycontroversial. All patients in the currentstudy also showed relative preservation ofcardiac index, with values of at leastabout 3 L/min/m2, so would not neces-sarily have qualified for dobutamine ther-apy according to Surviving Sepsis Cam-paign criteria. When they were givendobutamine at 5.4 � 1.7 �g/kg/min, pa-tients demonstrated significant increasesin ejection fraction, but not in cardiacindex or changes in LVEDV. The situationwas made more complex by a concomi-tant reduction in norepinephrine dosefrom 0.8 � 0.7 �g/kg/min to 0.5 � 0.5�g/kg/min. These are high doses of nor-epinephrine, and a relatively low dose ofdobutamine, and it is impossible to knowwhether the hemodynamic changes aredue to dobutamine, or to norepinephrinereduction, or their influence on outcome,which the authors acknowledge. Even ifthese data are interpreted as suggestingthat inotropic therapy should be usedmore widely in severe sepsis, other datasuggest that dobutamine may not neces-sarily be the optimum inotrope (9).

What does this study mean for routinepractice? Ventricular hypokinesia is a fre-quent phenomenon in severe sepsis, butrequires bedside echocardiography in theICU to be readily detected and quantified.Filling pressures or volumes and cardiac

output measurement alone are not ade-quate to describe fully the hemodynamicdefects of sepsis. Nor is echocardiogra-phy, in spite of the authors’ enthusiasm,because it is not a continuous monitoringtechnique but an intermittent measure-ment technique with a degree of subjec-tivity. It does, however, provide directvisualization of the response to therapythat no other technology allows. Perhapsit is time for us to learn from neurosur-gical critical care, and make a virtue ofthe individual shortcomings of specificmonitoring approaches by actively favor-ing a multimodal approach that com-bines the different technologies to allowus to describe and ultimately treat thesecomplex patients more effectively thanwe manage to do at present.

Richard Beale, MB, BSGuy’s and St Thomas’

Hospital NHS FoundationTrust

London, UK

REFERENCES

1. Dellinger RP, Levy MM, Carlet JM, et al: Surviv-ing Sepsis Campaign: International guidelinesfor the management of severe sepsis and septicshock: 2008. Crit Care Med 2008; 36:296–327

2. The SAFE Study Investigators. A Comparisonof albumin and saline for fluid resuscitation inthe intensive care unit. N Engl J Med 2004;350:2247–2256

3. Brunkhorst FM, Engel C, Bloos F, et al: In-tensive insulin therapy and pentastarch resus-citation in severe sepsis. N Engl J Med 2008;358:125–139

4. Viellard-Baron A, Caille V, Charron C, et al:Actual incidence of global left ventricular hy-pokinesia in adult septic shock. Crit Care Med2008; 36:1701–1706

5. Jardin F, Fourme T, Page B, et al: Persistentpreload defect in severe sepsis despite fluidloading. A longitudinal echocardiographicstudy in patients with septic shock. Chest1999; 116:1354–1359

6. Calvin JE, Driedger AA, Sibbald WJ: An assess-ment of myocardial function in human sepsisutilizing ECG-gated cardiac scintigraphy.Chest 1981; 80:579–586

7. Parker MP, Shelhamer JH, Bacharach SL,et al: Profound but reversible myocardial de-pression in patients with septic shock. AnnIntern Med 1984; 100:483–490

8. Jones AE, Craddock PA, Tiyal VS, et al: Diag-nostic accuracy of left ventricular function foridentifying sepsis among emergency depart-ment patients with nontraumatic symptom-atic undifferentiated hypotension. Shock2005; 24:513–517

9. Morelli A, De Casto S, Teboul JL, et al: Effectsof levosimendan on systemic and regionalhaemodynamics in septic myocardial depres-sion. Intensive Care Med 2005; 31:639–644

1951Crit Care Med 2008 Vol. 36, No. 6

Growth hormone to catabolic patients revisited*

T he muscle protein depletionassociated with critical illnessand in particular multiple or-gan failure is one of the major

challenges in intensive care medicine.Adequate nutrition attenuates the deple-tion, but nutrition is not sufficient tocounteract the catabolic processes. De-spite nutritional support, the protein lossfrom leg muscle amounts to 10% perweek (1). In parallel to the protein losses,there is a free glutamine depletion (2, 3),although the efflux of free glutamine fromskeletal muscle is moderately increased (4).The low plasma concentration of glutamineis an independent predictor for mortality inseptic shock patients (5), and an extra sup-ply of glutamine has been demonstratedto decrease mortality in multiple organ fail-ure patients requiring parenteral nutrition(6–8).

The well-known anabolic effect ofgrowth hormone was recognized previ-ously (9) as a potential agent to counter-act muscle protein depletion in criticalillness. Shortage of growth hormone lim-ited the exploration of this possibility, butwhen recombinant growth hormone be-came available, a number of encouragingcase reports were published (10, 11). Inmechanistic studies, a stimulation ofskeletal muscle protein synthesis rate andan attenuation of muscle free glutaminedepletion were reported (12–14). A con-siderable optimism arose, although con-cerns were raised about the potentialdanger of suppressing endogenous glu-tamine production (15) as well as theinsufficient knowledge of the effects ofthe pharmacologic doses of growth hor-mone that were used (16). A phase IIImulticenter study treating mechanicallyventilated intensive care unit (ICU) pa-tients with pharmacologic doses ofgrowth hormone was launched (17). At

an interim safety control, an elevatedmortality rate in the treated patients wasfound. Overnight the enthusiasm wasgone and ongoing studies were stopped.

A number of possible explanations forthe disastrous result of the phase III trialwere suggested, such as elevation of theconcentration of free fatty acids in plasmaresulting in cardiac toxicity and arrhyth-mias, shortage of glutamine as indicatedpreviously, and immunodepressive ef-fects. Thorough scrutiny of the clinicalreport forms from the phase III studydoes not provide any clear-cut clues tothe mechanism behind the excess mor-tality rate. The mortality was evenly dis-tributed over time related to the initia-tion of therapy as well as in relation tothe total length of ICU stay. Cause ofdeath was not different between the con-trols and the treatment group.

Following the publication of the phaseIII study, there has virtually been a totalabsence of clinical studies involvinggrowth hormone in catabolic patients.Therefore, it is extremely welcome thatinvestigators have come back to the ideaof growth hormone substitution as a pos-sible mechanism to overcome the deple-tion of muscle proteins in critical illness.A working group from the Czech Repub-lic has chosen an approach with whichthey mimic the physiologic endogenoussecretion pattern for growth hormone us-ing intravenous pulses (18). Further-more, in the experimental setup the in-vestigators supplement their patientswith exogenous glutamine intravenously,to compensate for any decrease in endog-enous glutamine production. As a result,Dr. Duška and colleagues (19) were ableto demonstrate an improved nitrogen re-tention in a model study involving me-chanically ventilated trauma patients inthe ICU, published in this issue of CriticalCare Medicine.

This study represents an important re-vival of the idea that hormonal stimula-tion of muscle protein metabolism maybe a possibility in the future when therisks involved with growth hormonetherapy are identified and controlled. Theeffects of growth hormone on glucosehemostasis necessitate a close surveil-

lance of blood glucose. Furthermore, theeffect of growth hormone on endogenousglutamine production should be sepa-rately elucidated, and it is recommendedthat in clinical studies plasma glutamineconcentration be normalized (to �0.5mmol/L) by exogenous supplementation.In the present study the exogenous sup-plementation was only sufficient to pre-vent a further decrease in plasma glu-tamine concentration. It is even possiblethat a normalization of plasma glutamineconcentration will facilitate blood glu-cose control, as it is reported that exog-enous glutamine supplementation de-creases insulin resistance (20).

The study by Dr. Duška and colleagues(19) is an important step forward afterthe total silence in the field of clinicalresearch over the effects of growth hor-mone as a possible therapeutic agent incatabolic states.

Jan Wernerman, MD, PhDIntensive Care MedicineKarolinska University Hospital

HuddingeStockholm, Sweden

REFERENCES

1. Gamrin L, Essen P, Forsberg AM, et al: Adescriptive study of skeletal muscle metabo-lism in critically ill patients: Free amino ac-ids, energy-rich phosphates, protein, nucleicacids, fat, water, and electrolytes. Crit CareMed 1996; 24:575–583

2. Roth E, Funovics J, Muhlbacher F, et al:Metabolic disorders in severe abdominal sep-sis: Glutamine deficiency in skeletal muscle.Clin Nutr 1982; 1:25–41

3. Tjader I, Rooyackers O, Forsberg AM, et al:Effects on skeletal muscle of intravenousglutamine supplementation to ICU patients.Intensive Care Med 2004; 30:266–275

4. Vesali RF, Klaude M, Rooyackers OE, et al:Longitudinal pattern of glutamine/glutamatebalance across the leg in long-stay intensivecare unit patients. Clin Nutr 2002; 21:505–514

5. Oudemans-van Straaten HM, Bosman RJ,Treskes M, et al: Plasma glutamine depletionand patient outcome in acute ICU admis-sions. Intensive Care Med 2001; 27:84–90

6. Griffiths RD, Jones C, Palmer TE: Six-monthoutcome of critically ill patients given glu-tamine-supplemented parenteral nutrition.Nutrition 1997; 13:295–302

*See also p. 1707.Key Words: growth hormone; critical care; glu-

tamine; catabolism; protein metabolismThe author has not disclosed any potential con-

flicts of interest.Copyright © 2008 by the Society of Critical Care

Medicine and Lippincott Williams & Wilkins

DOI: 10.1097/CCM.0b013e318176aa4a


7. Goeters C, Wenn A, Mertes N, et al: Paren-teral L-alanyl-L-glutamine improves6-month outcome in critically ill patients.Crit Care Med 2002; 30:2032–2037

8. Novak F, Heyland DK, Avenell A, et al: Glu-tamine supplementation in serious illness: Asystematic review of the evidence. Crit CareMed 2002; 30:2022–2029

9. Dahn MS, Lange P: Hormonal changes andtheir influence on metabolism and nutritionin the critically ill. Intensive Care Med 1982;8:209–213

10. Ziegler TR, Young LS, Ferrari-Baliviera E, et al:Use of human growth hormone combined withnutritional support in a critical care unit.JPEN J Parenter Enteral Nutr 1990; 14:574–581

11. Herndon DN, Barrow RE, Kunkel KR, et al:Effects of recombinant human growth hor-mone on donor-site healing in severely burnedchildren. Ann Surg 1990; 212:424–429

12. Gamrin L, Essén P, Hultman E, et al: Pro-tein-sparing effect in skeletal muscle of

growth hormone treatment in critically illpatients. Ann Surg 2000; 231:577–586

13. Biolo G, Iscra F, Bosutti A, et al: Growthhormone decreases muscle glutamine pro-duction and stimulates protein synthesis inhypercatabolic patients. Am J Physiol Metab2000; 279:E323–E332

14. Carroll PV, Jackson NC, Russell-Jones DL,et al: Combined growth hormone/insulin-like growth factor I in addition to glutamine-supplemented TPN results in net proteinanabolism in critical illness. Am J PhysiolEndocrinol Metab 2004; 286:E151–E157

15. Wernerman J: Protein wasting in severe ill-ness: Pathogenesis and therapy. DiabetesNutr Metab 2000; 13:21–24

16. Van den Berghe G: Increased mortality asso-ciated with growth hormone treatment incritically ill adults. N Engl J Med 2000; 342:134–135

17. Takala J, Ruokonen E, Webster NR, et al:Increased mortality associated with growth

hormone treatment in critically ill adults.N Engl J Med 1999; 341:785–792

18. Duška F, Fric M, Pažout J, et al: Frequentintravenous pulses of growth hormone to-gether with alanylglutamine supplementa-tion in prolonged critical illness after multi-ple trauma: Effects on glucose control,plasma IGF-I and glutamine. Growth HormIGF Res 2008; 18:82–87

19. Duška F, Fric M, Waldauf P, et al: Frequentintravenous pulses of growth hormone to-gether with glutamine supplementation inprolonged critical illness after multiple trau-ma: Effects on nitrogen balance, insulin re-sistance, and substrate oxidation. Crit CareMed 2008; 36:1707–1713

20. Dechelotte P, Hasselmann M, Cynober L, et al:L-alanyl-L-glutamine dipeptide-supplementedtotal parenteral nutrition reduces infectiouscomplications and glucose intolerance in crit-ically ill patients: The French controlled, ran-domized, double-blind, multicenter study. CritCare Med 2006; 34:598–604

Insights from PAC-Man on flow-monitoring devices: Curbyour enthusiasm*

I n the past few years, a number oflarge, well-conducted, clinical tri-als have cast doubt on or dis-proved the effectiveness of com-

monly employed interventions forcritically ill patients, each of which wassupported by strong physiologic rationaleand investigation. Examples include theuse of albumin for resuscitation (1), cor-ticosteroids for traumatic brain injury (2)or septic shock (3), and nitric oxide forhypoxemic respiratory failure (4). Theprecise benefit of intensive insulin con-trol in a heterogeneous population ofcritically ill patients is uncertain (5), andthe use of activated protein C in severesepsis is being reevaluated in anotherlarge multicenter trial (6).

Perhaps the mother of all controversiesin critical care over the past 2 decades hasrevolved around goal-directed therapy and

the role of the pulmonary artery catheter(PAC). A number of randomized controlledtrials and observational studies have notfound a benefit in clinical outcomes whenpulmonary catheters have been used to di-rect intravenous fluid and hemodynamicsupport among high-risk surgical patientsor those with acute lung injury or sepsis(7–11). Accordingly, we have seen a dra-matic decrease in pulmonary artery cathe-ter use in the United States over the past15 yrs (12). Adding insult to injury is thepotential for increasing risk of misuse,side effects, and harm if the skill of pul-monary artery catheter insertion and datainterpretation is lost through waning ex-perience (13).

The PAC-Man study was a large prag-matic trial of pulmonary artery catheter-ization or not for critically ill patientsthat took place in 65 intensive careunits (ICUs) in the United Kingdom.Unlike the ARDS Network Fluid andCatheter Treatment Trial (FACTT), PAC-Man did not use prescribed managementalgorithms in the treatment or controlarm of the trial. The primary objectivewas to compare hospital mortality andpulmonary artery catheter use. However,nearly 80% of ICUs enrolling patients de-cided, a priori, to use an alternative car-

diac output measurement device in thecontrol arm, such as Doppler ultrasonog-raphy, or indicator dilution techniques ofPiCCO and LiDCO. Hence, PAC-Man didnot compare pulmonary artery catheter-derived cardiac output and vascular pres-sure measurement with no cardiac out-put measurement but rather comparedpulmonary artery catheter use vs. a fre-quently less invasive flow measurementstrategy (11).

In this issue of Critical Care Medicine,Dr. Harvey and colleagues (14) present apost hoc analysis of the PAC-Man trial,providing descriptive information on pa-tients and exploring important subgroupsthat were considered for managementwith a pulmonary artery catheter or al-ternative flow-measuring device. The au-thors specified these subgroups and theanalytic method after initial trial resultswere known. Cox proportional hazardsmodeling was used for time-to-eventanalyses and multilevel logistic regres-sion was used for other evaluations tomake comparisons between groups re-ceiving different flow-monitoring strate-gies or not and to adjust for differences inmeasured potential confounders.

After stratifying for time from ICU ad-mission to randomization, age, severity of

*See also p. 1714.Key Words: heart catheterization; critical care;

clinical trial; mortality; cardiac outputDr. Ferri received an unrestricted research grant

from Edwards Lifesciences Canada in 2007. Dr. Fowlerhas not disclosed any potential conflicts of interest.


DOI: 10.1097/CCM.0b013e318176a7d7


illness, surgical status, the use of otherflow-monitoring devices, and ICU andhospital characteristics, the authorsfound no difference in hospital survivalamong patients managed with a pulmo-nary artery catheter compared with an-other cardiac output monitoring device.The only exception was among scheduledpostoperative patients admitted to ICUwithin whom pulmonary artery catheter-ization had a significant hospital survivalbenefit. However, we should be cautiousin the interpretation of results from onesmall subgroup among a large number ofcomparisons, on which no adjustmentsin the required p value to denote signifi-cance have been made. It is difficult toimagine the mechanism that would un-derlie this benefit.

The main strength of the currentstudy is that it addresses a very importantcomparison not extensively examined inthe main PAC-Man publication. Becauseuse of an alternative flow-measuring de-vice may have been a postrandomizationdecision for patients enrolled in PAC-Man, the benefits of randomization inallocating nonintervention characteris-tics equally among the two groups is lost,and the authors must therefore rely onmultivariable modeling to adjust for therecognized and measured potential con-founding variables. Although the authorsdid their best to remove the influence ofsuch variables from their analysis, it ispossible that unrecognized or unmea-sured confounders still conspire to makethe two patient groups different and thusthe comparisons inconclusive. The otherchief limitation of this study is its posthoc nature and the risk that in asking anumber of questions, associations will befound, or not, by chance alone.

Should we be surprised that the PAC-Man investigators did not find benefit inuse of either pulmonary artery catheteror other flow-measuring devices? In the

absence of an incrementally effectivetreatment strategy that is aided by flow-measuring device information, it is diffi-cult to believe that such devices shouldaffect outcome. For example, if we did nothave effective medications, such as�-blockers, angiotensin-converting en-zyme inhibitors, and receptor blockers, totreat congestive heart failure, would theuse of echocardiography alone to diag-nose left ventricular function be associ-ated with improved patient outcomes? Itis unlikely that other devices, providingsimilar (or less) physiologic informationabout cardiac output and vascular pres-sures than do pulmonary artery cathe-ters, will prove more effective. We shouldbe prudent in employing expensive diag-nostic-therapeutic strategies withoutdocumented clinical benefit. We shouldalso likely temper our enthusiasm for re-peating such difficult, expensive, and im-portant trials, using end points of mor-tality, with each next generation ofsimilar technology in the absence of ad-vances in therapy.

Mauricio Ferri, MDRobert A. Fowler, MD, MS(Epi)

Sunnybrook Health SciencesCentre

Toronto, OntarioCanada

REFERENCES

1. The SAFE Study Investigators: A comparisonof albumin and saline for fluid resuscitationin the intensive care unit. N Engl J Med2004; 350:2247–2256

2. Edwards P, Arango M, Balica L, et al, and theCRASH trial collaborators: Final results ofMRC CRASH, a randomised placebo-con-trolled trial of intravenous corticosteroid inadults with head injury—Outcomes at 6months. Lancet 2005; 365:1957–1959

3. Sprung CL, Annane D, Keh D, et al, and theCORTICUS Study Group: Hydrocortisonetherapy for patients with septic shock.N Engl J Med 2008; 358:111–124

4. Adhikari NK, Burns KE, Friedrich JO, et al:Effect of nitric oxide on oxygenation andmortality in acute lung injury: Systematicreview and meta-analysis. BMJ 2007; 334:779

5. Van den Berghe G, Wilmer A, Hermans G,et al: Intensive insulin therapy in the medicalICU. N Engl J Med 2006; 354:449–461

6. Efficacy and Safety of Drotrecogin Alfa (Ac-tivated) in Adult Patients With Septic Shock.NCT00604214. http:// www.clinicaltrials.gov.Accessed February 22, 2008

7. Connors AF Jr, Speroff T, Dawson NV, et al,SUPPORT Investigators: The effectiveness ofright heart catheterization in the initial careof critically ill patients. JAMA 1996; 276:889–897

8. Sandham JD, Hull RD, Brant RF, et al, forthe Canadian Critical Care Clinical TrialsGroup: A randomized, controlled trial of theuse of pulmonary-artery catheters in high-risk surgical patients. N Engl J Med 2003;348:5–14

9. Richard C, Warszawski J, Anguel N, et al,French Pulmonary Artery Catheter StudyGroup: Early use of the pulmonary arterycatheter and outcomes in patients withshock and acute respiratory distress syn-drome: A randomized controlled trial. JAMA2003; 290:2713–2720

10. The National Heart, Lung, and Blood Insti-tute Acute Respiratory Distress Syndrome(ARDS) Clinical Trials Network: Comparisonof two fluid-management strategies in acutelung injury. N Engl J Med 2006; 354:2564–2575

11. Harvey S, Harrison DA, Singer M, et al, PAC-Man study collaboration: Assessment of theclinical effectiveness of pulmonary arterycatheters in management of patients in in-tensive care (PAC-Man): A randomised con-trolled trial. Lancet 2005; 366:472–477

12. Wiener RS, Welch HG: Trends in the use ofthe pulmonary artery catheter in the UnitedStates, 1993–2004. JAMA 2007; 298:423–429

13. Rubenfeld GD, McNamara-Aslin E, RubinsonL: The pulmonary artery catheter, 1967–2007: Rest in peace? JAMA 2007; 298:458–461

14. Harvey SE, Welch CA, Harrison DA, et al:Post hoc insights from PAC-Man—The U.K.pulmonary artery catheter trial. Crit CareMed 2008; 36:1714–1721


Burden of psychological symptoms and illness in family ofcritically ill patients: What is the relevance for criticalcare clinicians?*

I n this issue of Critical Care Med-icine, Dr. Seigel and colleagues (1)describe a high prevalence of psy-chological illness among family

members whose loved one died in theintensive care unit (ICU). This study isunique, as the first of its kind from NorthAmerica. The burden of psychologicalsymptoms has previously been describedamong family members of ICU patients inFrance. Pochard et al. (2) found that 73%of family members had a significant levelof symptoms of anxiety and 35% had sig-nificant levels of symptoms of depression3–5 days after a loved one was admittedto the ICU. Another French study, byAzoulay et al. (3), found the prevalence ofa significant burden of anxiety symptomsamong 49% and of depression symptomsamong 20% of family members of pa-tients in the ICU when measuring symp-toms 3 months after ICU stay. A recentintervention study, also from France, ex-amined psychological symptoms amongfamily members whose loved one haddied in the ICU 3 months prior. In theircontrol group, the prevalence of a signif-icant level of depression symptoms was67% and the prevalence of a significantlevel of anxiety symptoms was 56% (4).For each of these French studies, a “sig-nificant level” of symptoms was the levelpreviously defined in studies of survey-based screening measures that were as-sociated with a high risk for a clinicaldiagnosis of anxiety or depression, al-though these measures have not beenvalidated specifically among family mem-bers of critically ill patients.

There have been few studies in NorthAmerica that similarly document theprevalence and risk factors of psycholog-ical symptoms among family members ofcritically ill patients. The North Americanexperience may be different from that re-ported by French investigators becausefamilies’ preferences for care and clini-cians’ delivery of care are affected by re-gional, racial, religious, and cultural in-fluences (5). This study by Dr. Seigel andcolleagues (1) finds the overall prevalenceof one of four psychiatric illnesses (majordepressive disorder, generalized anxietydisorder, panic disorder or complicatedgrief disorder) to be 34%, with the mostcommon single disorder being major de-pressive disorder, with a prevalence of27%. Symptoms were assessed 3 to 12months after the patient’s death. Giventhe differences in study design, it is diffi-cult to directly compare the currentstudy with the French studies, but thisnew study shows that this problem is notconfined to France.

The study by Dr. Seigel and colleagues(1) is unique in the authors’ use of Struc-tured Clinical Interviews for DSM-IV(SCID) rather than screening surveymeasures to assess symptoms of psychi-atric illness. The SCID allows not onlythe assessment of symptoms of psychiat-ric illness but also a sensitive and specificdiagnosis of psychiatric illness by DSM-IVcriteria. This is clearly a strength of thisstudy; by contrast, the studies fromFrance all used the Hospital Anxiety andDepression Scale (HADS) as a screeningtool. A formal structured clinical inter-view, while it provides more reliable di-agnostic information than screeningtools, is more time and resource inten-sive. Therefore, screening tools like theHospital Anxiety and Depression Scaleand other psychological symptom mea-sures are more feasible for large researchstudies, and many of these survey toolshave been shown to provide reliable and

valid assessment of the burden of psycho-logical symptoms (6, 7).

Particularly interesting findings ofthis new study are the risk factors thatwere associated with psychiatric illness,including being the patient’s spouse, ex-periencing an additional major stressorafter the patient’s death, and the patient’sillness lasting �5 yrs (1). These risk fac-tors are important to elucidate to betterunderstand the process behind develop-ment of psychiatric illness in this groupand to identify a subgroup of familymembers who may be at increased riskfor psychiatric illness. Additionally, fam-ily members who did not consider thecritical care physician a source of comfortwere more likely to have psychiatric ill-ness. In this retrospective study, it is notpossible to determine whether physicianbehavior or communication caused fam-ily distress, but in the context of otherstudies this seems to be a likely explana-tion and it seems likely that improvingphysician communication will reduce theburden on the family (4, 8).

The study by Dr. Seigel and colleagues(1) included only family members of pa-tients who died in the ICU. There is evi-dence that these family members are atincreased risk for psychological symp-toms compared with the families of pa-tients who survive (3). However, eventhough patient death is a risk factor forpsychological symptoms among familymembers, families of patients who sur-vive are also at increased risk of thesesymptoms compared with the generalpopulation (3). There are two other im-portant reasons for focusing on the fam-ilies of all critically ill patients, not onlythe families of those patients whom weexpect to die. First, it is generally notclear whether critically ill patients willsurvive at the time when most clinician-family communication is occurring. Sec-ond, there is evidence that family mem-bers of patients who survive are actually

*See also p. 1722.Key Words: depression; anxiety; critical care; in-

tensive care; palliative care; end of lifeThe authors have not disclosed any potential con-



DOI: 10.1097/CCM.0b013e31817616c0


less satisfied with communication fromICU clinicians than are family membersof patients who die (9). If we are to betruly effective in minimizing the long-term effects of critical illness and criticalcare on family members, we must focuson our communication with the familiesof all critically ill patients.

As Dr. Seigel and colleagues (1) dis-cuss, the major limitations of their studyare the small sample size and a single-center study design; nonetheless, thisstudy highlights an important phenome-non. It is important for critical care cli-nicians to understand and address psy-chological morbidity among familymembers for several reasons. First, careprovided by critical care clinicians hasbeen associated with these symptoms (4,8). Second, because families often serveas surrogate decision makers for criticallyill patients (10), communication with andcare for family members of critically illpatients are important components of thequality of care delivered to critically illpatients. Third, the family’s experience inthe ICU is a high priority for most criti-cally ill patients. Fourth, psychologicalmorbidity among family members of crit-ically ill patients is likely to have impor-tant long-term consequences in many do-

mains, including future health care,employment, and quality of life. Studieslike this one will be critical in both de-fining the scope of the problem and un-derstanding risk factors and mechanismsfor development of these symptoms. Thisinformation will be necessary to developinterventions that assist family throughthe difficult experience of having a criti-cally ill loved one and also enhance thequality of care we provide to critically illpatients.

Erin K. Kross, MDJ. Randall Curtis, MD, MPH

Division of Pulmonary andCritical Care Medicine

Department of MedicineSchool of MedicineUniversity of WashingtonSeattle, WA

REFERENCES

1. Seigel MD, Hayes E, Vanderwerker LC, et al:Psychiatric illness in the next of kin of pa-tients who die in the intensive care unit. CritCare Med 2008; 36:1722–1728

2. Pochard F, Azoulay E, Chevret S, et al: Symp-toms of anxiety and depression in familymembers of intensive care unit patients:Ethical hypothesis regarding decision-making capacity. Crit Care Med 2001; 29:1893–1897

3. Azoulay E, Pochard F, Kentish-Barnes N, etal: Risk of post-traumatic stress symptoms infamily members of intensive care unit pa-tients. Am J Respir Crit Care Med 2005; 171:987–994

4. Lautrette A, Darmon M, Megarbane B, et al: Acommunication strategy and brochure forrelatives of patients dying in the ICU. N EnglJ Med 2007; 356:469–478

5. Vincent JL: Cultural differences in end-of-lifecare. Crit Care Med 2001; 29(2 Suppl):N52–N55

6. Lowe B, Spitzer RL, Grafe K, et al: Compar-ative validity of three screening question-naires for DSM-IV depressive disorders andphysicians’ diagnoses. J Affect Disord 2004;78:131–140

7. Kroenke K, Spitzer RL, Williams JB, et al:Anxiety disorders in primary care: preva-lence, impairment, comorbidity, and detec-tion. Ann Intern Med 2007; 146:317–325

8. Tilden VP, Tolle SW, Garland MJ, et al:Decisions about life-sustaining treatment:Impact of physicians’ behaviors on thefamily. Arch Intern Med 1995; 155:633– 638

9. Wall RJ, Curtis JR, Cooke CR, et al: Familysatisfaction in the ICU: Differences betweenfamilies of survivors and non-survivors.Chest 2007; 132:1425–1433

10. Prendergast TJ, Luce JM: Increasing inci-dence of withholding and withdrawal of lifesupport from the critically ill. Am J RespirCrit Care Med 1997; 155:15–20

Pediatric organ donation after cardiac death—Still in its infancy*

Organ donation after cardiacdeath (DCD) has been identi-fied as an important tool inthe struggle to close the gap

between the number of organ donors andthe nearly 100,000 individuals in theUnited States who currently are waitingfor lifesaving organs. An important goalof the National Organ Donation Break-through Collaborative, launched in 2003,has been to increase the percentage ofDCD donors from approximately 5% to

�10% of all donors. Significant progresshas been made nationally, with the totalnumber of DCD donors steadily increas-ing each year and 26 of 59 organ procure-ment organizations (OPOs) reportingthat �10% of their total donors wereDCD donors in 2007 (1).

Despite this overall progress, the storyis quite different when the focus is shiftedto pediatric DCD. While DCD donors �18yrs old accounted for 10.2% of total adultdeceased donors nationally in 2007, only7.3% of deceased donors under �18 yrsold were DCD donors. Since 1995, threeOPOs have accounted for 27% of the na-tional pediatric DCD donor activity. Dur-ing the same time period, 34 OPOs re-ported fewer than five pediatric DCDdonor recoveries, and ten OPOs reportedno pediatric DCD donor activity whatso-ever (2). A qualitative study by Curley etal. (3) may offer some clues as to why

DCD in pediatrics is lagging behind. Inthis study of hospital staff involved inorgan donation at a large children’s hos-pital, respondents reported concerns re-garding DCD that may be particularly sa-lient with pediatric patients. Theseconcerns included: 1) questions regard-ing the legitimacy of parental consent forDCD, since DCD is not strictly in thedying child’s best interests; 2) worriesthat DCD, even if done well, might neg-atively affect the dying experience for pa-tients and families; and 3) concern thatpublic trust in the hospital might erode ifa perception grew that dying childrenwere being “used” for their organs (3).

Another likely barrier to the develop-ment of pediatric DCD programs is thelack of published clinical research as wellas the striking lack of formal guidancefrom professional organizations specifi-cally with regard to pediatric DCD. The

*See also p. 1729.Key Words: organ donation; donation after cardiac

death; non-heart-beating organ donation; cardiorespi-ratory death; pediatrics; ethics

The author has not disclosed any potential con-flicts of interest.


DOI: 10.1097/CCM.0b013e318176131f


Institute of Medicine’s reports on DCD donot address the issue of DCD in children(4). The Society of Critical Care Medi-cine’s position paper on DCD briefly ad-dresses pediatric DCD and concludes thatit is ethically reasonable to perform DCDon children, although the report givesscant attention to the specifics and sub-tleties of developing pediatric DCD pro-grams (5).

This is the milieu in which Dr. Naimand colleagues (6), in this issue of Criti-cal Care Medicine, present their 11-yrexperience with pediatric DCD. They re-port the DCD experience at Children’sHospital of Philadelphia over an 11-yrperiod from 1995 to 2005. While therewere 14 pediatric DCD cases during thisperiod, charts were only available for 12.The donors in all 12 cases were �18 yrsof age. All donors had suffered severeneurologic injury but did not meet braindeath criteria. Death was declared after 2mins of acirculation, and organ procure-ment began after an additional 5-minwaiting period. Overall graft survival at 1yr was excellent, and all 12 donors suc-cessfully donated at least one lifesavingorgan. Despite this relatively small seriesof cases, this experience actually ac-counts for 6% of the total national pedi-atric DCD donor activity during this timeperiod (2). Hospitals across the countrythat are interested in considering, devel-oping, or improving a pediatric DCD pro-gram will be able to use this relativelyrich descriptive study to move towardtheir goals.

While the reader is likely to see in thisstudy a variety of findings that advancehis or her knowledge of the subject, sev-eral observations merit discussion here.In 1998, authors from the same institu-tion published a study intended to ex-plore the potential impact of pediatricDCD on overall deceased organ donationand determined that 6.9 patients per yearwould qualify for DCD and four patientsper year would donate through DCD (7).This estimate contrasts sharply with theactual rate of pediatric DCD of 1.3 DCDdonors per year observed by Dr. Naim andcolleagues (6). The findings of the cur-rent study are more in line with otherstudies from large intensive care unitsthat have attempted to estimate the im-

pact of pediatric and adult DCD on pro-curement rates (8, 9). While a number ofpotential explanations for these discrep-ancies have been offered, one factor thatmay have significantly reduced the ob-served rate of DCD donation from thatpredicted is the dominant practice duringthe study period of family-initiated dona-tion. If all potential DCD donor familieshad been approached during the period ofobservation, it is possible that DCD pro-curement rates might have been higher.This experience, while primarily reflec-tive of the preceding decade, should givepause to those who would argue that pe-diatric DCD has the potential to dramat-ically increase pediatric organ donationrates.

Another notable aspect of this caseseries relates to the waiting period be-tween the development of acirculationand the start of organ procurement. Anadequate time interval is essential to en-sure that death has occurred before organremoval begins, but as the time intervalincreases, the viability of the donated or-gans diminishes. Addressing this criticalelement, the Institute of Medicine recom-mended that “an interval of at least 5mins elapse after complete cessation ofcirculatory function . . . before death ispronounced and organ perfusion or do-nation begins” (4). While this recommen-dation has been widely adopted in theUnited States, some American hospitalscontinue to use a 2-min interval and oth-ers (particularly in Europe) use a 10-mintime interval (10). Lingering questionsremain about the effect of the duration ofthe time interval on graft function, andfew data are available to guide the waythat children’s hospitals approach thiscritical component of the DCD protocol.While the experience offered by Dr. Naimand colleagues (6) is small, the authorsshow that excellent graft function canresult from a total interval of 7 mins in apediatric population.

This article brings into focus, but doesnot answer, many of the most difficultquestions regarding the development ofpediatric DCD programs. As a rich de-scriptive study of a single center’s suc-cessful pediatric DCD program, the arti-cle will illuminate the challenging andcontroversial aspects of pediatric DCD

and provide guidance to the national de-bate. Future research should address thebiological, psychological, emotional, eth-ical, and social particularities of pediatricorgan donation after cardiac death. Theconcerns voiced by Curley et al. (3) willbe addressed and overcome only afterconsiderable dialogue, reflection, and de-liberation on the part of hospital commu-nities across the country, informed by thepediatric community at large and by worklike that of Dr. Naim and colleagues (6).

George Hardart, MD, MPHColumbia University College

of Physicians andSurgeons

New York, NY

REFERENCES

1. Organ Procurement and TransplantationNetwork. www.optn.org. Accessed February11, 2008

2. United Network for Organ Sharing. www.unos.org. Accessed February 11, 2008

3. Curley MA, Harrison CH, Craig N, et al: Pe-diatric staff perspectives on organ donationafter cardiac death in children. Pediatr CritCare Med 2007; 8:212–219

4. Institute of Medicine: Non-Heart-Beating Or-gan Transplantation: Medical and Ethical Is-sues in Procurement. Washington, DC, Na-tional Academy Press, 1997

5. Recommendations for nonheartbeating or-gan donation: A position paper by the Eth-ics Committee, American College of Criti-cal Care Medicine, Society of Critical CareMedicine. Crit Care Med 2001; 29:1826 –1831

6. Naim MY, Hoehn KS, Wertlieb S, et al: TheChildren’s Hospital of Philadelphia’s experi-ence with donation after cardiac death. CritCare Med 2008; 36:1729–1733

7. Koogler T, Costarino AT: The potential ben-efits of the pediatric nonheartbeating donor.Pediatrics 1998; 101:1049–1052

8. Lacroix JD, Mahoney JE, Knoll GA: Renaltransplantation using non-heart-beating do-nors: A potential solution to the organ donorshortage in Canada. Can J Surg 2004; 47:10–14

9. Durall AL, Laussen PC, Randolph AG: Poten-tial for donation after cardiac death in achildren’s hospital. Pediatrics 2007; 119:e219–e224

10. Shemie SD, Baker AJ, Knoll G, et al: Nationalrecommendations for donation after cardio-circulatory death in Canada: Donation aftercardiocirculatory death in Canada. Can MedAssoc J 2006; 175:S1


The curse of the drinking class*

Work is the curse of the drinking class—Oscar Wilde

T he adage that the definition ofan alcoholic is someone whodrinks more than his doctorturns on the truth that one

often has a blind spot to one’s own weak-nesses. In this issue of Critical Care Med-icine, Dr. Gacouin and colleagues (1)have unveiled another blind spot relatedto prior alcohol consumption in intensivecare unit (ICU) patients, putting them atincreased risk of developing bacterial in-fections and ventilator-associated pneu-monia (VAP) during their ICU stays.Heavy drinkers also had almost twice therate of death compared with those whoabstained from alcohol or those who weremoderate drinkers.

The study is a prospective, observa-tional study of 358 patients admitted for�3 days to a 21-bed ICU at a Frenchuniversity medical center over a 1-yr pe-riod. The preadmission alcohol drinkinghabits of patients were evaluated usingtwo tools: the National Institute of Alco-hol Abuse and Alcoholism criteria for at-risk drinkers and the Simplified MichiganAlcohol Short Test (SMAST). To improvethe accuracy of the diagnosis, the deter-mination of being at-risk was made,whenever possible, by interviewing bothpatients and family. The at-risk groupwas further subdivided into a high-dosegroup (five or more drinks per day) and alow-dose group (fewer than five drinksper day). The at-risk group was comparedwith a larger group (n � 247) who eitherwere abstinent (74%) or were moderatedrinkers (37%).

These results confirm those seen pre-viously in trauma and postoperative pa-tients where alcohol consumption was

significantly associated with excess mor-bidity and mortality during their ICUstays (2–4). At-risk drinkers had higherrates of acquired bacterial infection (36%vs. 19%), VAP (27% vs. 16%), and septicshock (18% vs. 8%). The effect of drink-ing also appears dose-related, as thegroup of drinkers consuming five ormore drinks per day had higher acquiredbacterial and VAP infection rates thandrinkers consuming fewer than fivedrinks per day. The length of time ofheavy drinking was also important: Ac-quired bacterial infection occurred inonly 5% of at-risk drinkers for �5 yrs, roseto 17% for those drinking 5–10 yrs, andreached 78% for those drinking �10 yrs.

The risk of developing VAP in at-riskdrinkers was increased by 70% in thisstudy. The baseline rate of VAP in thisunit is within the typical range reportedfor ICUs in the United States and Europe.What could cause such an increase? Bar-ring some difference in how at-risk pa-tients were provided with measures forpreventing VAP, these data imply intrin-sic factors at work. Could high daily al-cohol consumption cause a delay in seek-ing medical attention, so that patientswere sicker and more likely infected be-fore admission? A higher but not signifi-cantly different neutrophil count sug-gests this. It is well established that theprogression of head and neck cancers ismore advanced in patients with high al-cohol consumption, related to delay inseeking attention for an incipient prob-lem (5). Furthermore, it has been sug-gested that excessive alcohol consump-tion leads to a relative immunodeficientstate. Tonnesen et al. (2) found that de-layed hypersensitivity reactions were de-pressed in postoperative patients withhigh alcohol consumption (�60 g of eth-anol alcohol per day � five drinks) com-pared with a matched group of low alco-hol consumption.

One question about the study con-cerns the accuracy of the ethanol con-sumption history. How reliable is self-

reporting of alcohol consumption? Studiesfrom the United States, the United King-dom, and Denmark have addressed thisquestion and surprisingly have found thatunderreporting or forgetting is not usu-ally present (6–8). Others have foundthat the interviewing technique is impor-tant for avoiding underreporting and thatthe quantity-frequency approach, as usedin this study, is better than a graduatedfrequency approach (9). In addition, theauthors sought to improve accuracy byinterviewing family to confirm the tworeports. So although the authors did notprovide evidence of the accuracy of theirhistory taking, we can be fairly confidentthat the data on alcohol consumption wereaccurate in this study.

There are two other aspects of thestudy that deserve consideration. First, towhat extent was treatment for ethanolwithdrawal applied in the at-risk patientgroup? These patients were consumingenough ethanol regularly to manifestsigns and symptoms of ethanol with-drawal and thus were probably treatedwith benzodiazepines or other agents toameliorate the withdrawal syndrome. Al-though the proportions of patients whorequired intubation and ventilation werenot different between the two groups, thelength of time patients required mechan-ical ventilation is not reported. It isknown that excessive alcohol use in-creases the need for and duration of me-chanical ventilation in medical patientsand may have contributed to the in-creased VAP rates seen in this study (10).In addition, many physicians treat pa-tients at risk for alcohol withdrawal withprophylactic doses of benzodiazepines. Ifthe at-risk group received additional sed-atives, whether or not they were mechan-ically ventilated, they would be expectedto have a higher rate of complications.

Another aspect that is not well char-acterized is the severity of liver impair-ment in the at-risk group. The authorsstate that only eight of 111 patients hadcirrhosis, based on a diagnosis made be-

*See also p. 1735.Key Words: alcohol; nosocomial infection; sepsis;

infection; liver; cirrhosisThe authors have not disclosed any potential con-



DOI: 10.1097/CCM.0b013e318176112b


fore entry into the study. However, morethan three fourths of the at-risk grouphad been drinking �10 yrs. With thisamount of alcohol consumption in theat-risk group, there may be more patientswith end-stage liver disease. Why is thisimportant? Chronic liver injury second-ary to chronic ethanol ingestion causesimpaired clearance of intestinal bacteriathat enter the portal circulation. Im-paired hepatic filtering function of gutflora could be a mechanism for the in-creased infection rate in at-risk drinkersand may explain higher rates of septicshock (11).

So what is the take-home messagefrom the findings of Dr. Gacouin andcolleagues (1)? It is essential to take adetailed alcohol ingestion history in pa-tients admitted to the ICU. This history isimportant not only for the prediction ofwell-recognized complications, such asliver dysfunction and alcohol withdrawalsyndrome, but also to recognize that

these patients are at increased risk forsevere infection.

C. Spencer Yost, MDMichael A. Gropper, MD, PhD

UCSF–Anesthesia andPerioperative Care

San Francisco, CA

REFERENCES

1. Gacouin A, Legay F, Camus C, et al: At-riskdrinkers are at higher risk to acquire a bac-terial infection during an intensive care unitstay than abstinent or moderate drinkers.Crit Care Med 2008; 36:1735–1741

2. Tonnesen H, Petersen KR, Hojgaard L, et al:Postoperative morbidity among symptom-freealcohol misusers. Lancet 1992; 340:334–337

3. Jurkovich GJ, Rivara FP, Gurney JG, et al:The effect of acute alcohol intoxication andchronic alcohol abuse on outcome fromtrauma. JAMA 1993; 270:51–56

4. Spies CD, Neuner B, Neumann T, et al: In-tercurrent complications in chronic alco-holic men admitted to the intensive care unitfollowing trauma. Intensive Care Med 1996;22:286–293

5. Tromp DM, Brouha XD, De Leeuw JR, et al:Psychological factors and patient delay inpatients with head and neck cancer. EurJ Cancer 2004; 40:1509–1516

6. Crawford A: Bias in a survey of drinkinghabits. Alcohol 1987; 22:167–179

7. Grant KA, Arciniega LT, Tonigan JS, et al:Are reconstructed self-reports of drinking re-liable? Addiction 1997; 92:601–606

8. Stockwell T, Donath S, Cooper-Stanbury M, etal: Under-reporting of alcohol consumption inhousehold surveys: A comparison of quantity-frequency, graduated-frequency and recent re-call. Addiction 2004; 99:1024–1033

9. Gronbaek M, Heitmann BL: Validity of self-reported intakes of wine, beer and spirits inpopulation studies. Eur J Clin Nutr 1996;50:487–490

10. de Wit M, Best AM, Gennings C, et al: Alcoholuse disorders increase the risk for mechani-cal ventilation in medical patients. AlcoholClin Exp Res 2007; 31:1224–1230

11. Balzan S, de Almeidaq Uadros C, de Cleva R,et al: Bacterial translocation: Overview ofmechanisms and clinical impact. J Gastroen-terol Hepatol 2007; 22:464–471

Tracheostomy protocol compliance: Herding cats?*

T racheostomies are commonprocedures in the intensivecare unit (ICU), occurring inabout 10% of critically ill pa-

tients who require mechanical ventila-tion (1). A tracheostomy is reported toallow a more secure airway, earlier andsafer enteral feeding, easier oral care, andenhanced patient comfort while reducingsedation needs. However, the benefits oftracheostomy have to be weighed againstpotential complications, which includeinfections, hemorrhage, pneumothorax,pneumomediastium, and tracheal steno-sis (2).

A review published in 1998 concludedthat there was insufficient evidence tosupport the view that timing of a trache-ostomy can alter the duration of mechan-ical ventilation or prevent airway injury

in the critically ill patient (3). Since then,many other studies have reported de-creased days of mechanical ventilation(4), decreased duration of ICU and hospi-tal length of stay (5), and less damage tothe upper airway (4) with early tracheos-tomies. A meta-analysis attempted to an-swer this question definitively in 2005.Griffiths et al. (6) reported that early tra-cheostomy (�7 days) may lead to reducedduration of mechanical ventilation andshorter ICU stays, but the timing of tra-cheostomies did not alter mortality orincrease the risk of developing hospital-acquired pneumonia. Unfortunately, thismeta-analysis only included a limitednumber of studies (five trials) and a smallnumber of patients (406 patients). Heter-ogeneity was also high in the meta-analysis due to variability of inclusionand exclusion criteria, definitions of earlyand late tracheostomy, characteristics ofenrolled patients, and diagnostic criteriafor hospital-acquired pneumonia.

Part of the difficulty in designing stud-ies to address the timing of tracheostomylies in identifying patients who may havea simple subsequent weaning process af-ter failing the first attempt at extubation.

Tracheostomies should be performed inthose patients who will encounter a dif-ficult or prolonged wean. Unfortunately,no validated specific and sensitive test orscoring system is available that predictsthe need for prolonged ventilation, so theselection of patients for tracheostomy re-mains a subjective, arbitrary decision dueto the lack of solid data.

In this issue of Critical Care Medicine,Dr. Freeman and colleagues (7) under-take the task of developing a standardizedapproach to tracheostomy selection andtiming. The authors felt that one way todecrease arbitrary choices would be tointegrate the decision-making processinto a standard ventilator weaning proto-col. The primary purpose of the study wasto determine the feasibility of developingsuch a protocol. In their pilot implemen-tation group of 125 patients, patients,who met criteria for tracheostomy, failedthe preliminary weaning assessment(PWA) or spontaneous breathing trial(SBT) for two successive days after either5 days of mechanical ventilation if theyhad not been reintubated or 3 days ofmechanical ventilation if they had beenreintubated.

*See also p. 1742.Key Words: tracheostomy; mechanical ventilation;

protocol; quality improvement; critical care; sedationThe authors have not disclosed any potential con-



DOI: 10.1097/CCM.0b013e3181761228


The article by Dr. Freeman and col-leagues (7) has certain limitations thatthe authors readily acknowledge. Despitefocused educational sessions for physi-cians, nurses, and ancillary staff, visibleprotocols posted at each bedside, and ex-tensive discussions on ICU rounds, only45% of the patients underwent tracheos-tomy consistent with the protocol. Thir-ty-three percent of the patients under-went tracheostomy earlier than dictatedby protocol, and 22% of the patients, de-spite passing one or more SBTs, contin-ued to tracheostomy.

The fact that the majority of physi-cians did not follow the study protocolreflects the experience with other clinicalprotocols (e.g., daily sedation cessationprotocol) (8). As anyone who has tried toimplement a protocol into widespreadclinical practice knows, this can be a dif-ficult feat. Implementation of clinicalpractice guidelines is often slow to non-existent. Lack of physician buy-in to pro-tocols is attributed to physicians’ under-lying knowledge and attitudes, theculture in which they practice, minimalincentive to change, and lack of time orstaff support (9).

The study was preliminary and there-fore could not be extrapolated to all crit-ical care patients, such as those on non-invasive ventilation. The study could notevaluate other questions related to tra-cheostomy, such as technique, patientcomfort, and reintubation rates. Anotherstudy designed to look at tracheostomyprotocol efficacy compared with standard

practice will have to be completed, but inthe meantime, this introduction of astandardized approach to tracheostomybased on weaning performance may bejust the breakthrough we need to usetracheostomy in a more consistent andeffective fashion. This protocol may notbe perfect, but one logical approach mustbe chosen to promote consistent clinicaldecisions. The consistency then allowsfor the rigorous evaluation of the inter-vention, which in this instance is thetiming of tracheostomy placement (10).

Although it is too early to recommendwidespread implementation of tracheos-tomy protocols, perhaps someday a pa-tient admitted to the ICU with severesepsis will be enrolled in an early resus-citation protocol, a low tidal volume ven-tilation protocol, a daily sedation cessa-tion protocol, a spontaneous breathingtrial protocol, a glucose control protocol,and an assessment for tracheostomy pro-tocol. The ultimate success of this proto-col-based care may well hinge on ourability to “herd the cats” (ourselves) intofollowing our own recommendations.

Linda Liu, MDMichael A. Gropper, MD, PhD

UCSF–Anesthesia andPerioperative Care

San Francisco, CA

REFERENCES

1. Esteban A, Anzueto A, Frutos F, et al: Char-acteristics and outcomes in adult patientsreceiving mechanical ventilation: A 28-day

international study. JAMA 2002; 287:345–355

2. Stauffer JL, Olson DE, Petty TL: Complica-tions and consequences of endotracheal in-tubation and tracheotomy: A prospectivestudy of 150 critically ill adult patients. Am JMed 1981; 70:65–76

3. Maziak DE, Meade MO, Todd TR: The timingof tracheotomy: A systematic review. Chest1998; 114:605–609

4. Rumbak MJ, Newton M, Truncale T, et al: Aprospective, randomized, study comparingearly percutaneous dilational tracheotomy toprolonged translaryngeal intubation (delayedtracheotomy) in critically ill medical pa-tients. Crit Care Med 2004; 32:1689–1694

5. Rodriguez JL, Steinberg SM, Luchetti FA,et al: Early tracheostomy for primary airwaymanagement in the surgical critical care set-ting. Surgery 1990; 108:655–659

6. Griffiths J, Barber VS, Morgan L, et al: Sys-tematic review and meta-analysis of studiesof the timing of tracheostomy in adult pa-tients undergoing artificial ventilation. BMJ2005; 330:1243

7. Freeman BD, Kennedy C, Robertson TE,et al: Tracheostomy protocol: Experiencewith development and potential utility. CritCare Med 2008; 36:1742–1748

8. Mehta S, Burry L, Fischer S, et al: Canadiansurvey of the use of sedatives, analgesics, andneuromuscular blocking agents in criticallyill patients. Crit Care Med 2006; 34:374–380

9. Cabana MD, Rand CS, Powe NR, et al: Whydon’t physicians follow clinical practiceguidelines? A framework for improvement.JAMA 1999; 282:1458–1465

10. Morris AH: Developing and implementingcomputerized protocols for standardizationof clinical decisions. Ann Intern Med 2000;132:373–383

Sleep and breathing: The impact of mechanical ventilation on thequality of sleep*

T he field of sleep medicine hasrapidly expanded in the last 10yrs. Although sleep distur-bances have received much at-

tention in the medical community and

the outpatient setting, sleep disturbanceshave not been a focus in the hospitalsetting, particularly the intensive careunit (ICU). Patients admitted to intensivecare are critically ill with multiple organsystem dysfunctions. Studies have consis-tently demonstrated that sleep distur-bances are a common and significantconcern for these patients. Patients in theICU have poor sleep efficiency with frag-mented sleep, increased arousals result-ing in awakenings, and a low percentageof time spent in rapid eye movement(REM) sleep (1–4). Few studies have un-

dertaken the task of evaluating the phys-iologic, neurocognitive, and psychologi-cal consequences of poor sleep in thispatient population. Also, how sleep con-tributes to the recovery of illness has notbeen evaluated.

Sleep disturbances in the ICU are mul-tifactorial. Some of these factors includethe ICU environment, medications, noise,pain, illness, frequent care-related inter-ruptions by nurses and physicians, andmechanical ventilation (5, 6). Cooperet al. (2) evaluated sleep architecture in20 critically ill patients requiring me-

*See also p. 1749.Key Words: sleep; respiration; artificial respiration;

critical illnessThe authors have not disclosed any potential con-



DOI: 10.1097/CCM.0b013e3181761260


chanical ventilation and found that allpatients had abnormal sleep with severelyreduced sleep efficiency and REM sleep.Severe sleep fragmentation was reflectedby a high frequency of arousals and awak-enings. The mode of mechanical ventila-tion used may contribute to sleep distur-bance. The study by Meza et al. (7)showed that pressure support ventilation(PSV) caused arousals and awakeningsdue to central apneas in healthy subjects.Parthasarathy and Tobin (4) evaluatedthe relationship between the mode of me-chanical ventilation and sleep disruptionby comparing assist control ventilation(ACV) with pressure support in 11 me-chanically ventilated patients. Five pa-tients (with congestive heart failure andejection fraction �50%) developed cen-tral apneas on PSV leading to arousalsand increased sleep fragmentation com-pared with subjects on ACV.

Problems associated with mechanicalventilation, such as dyssynchrony ofbreathing, endotracheal tube discomfort,and increased patient effort, can contrib-ute to frequent arousals and fragmentedsleep. Bosma et al. (3) looked at the rela-tionship between patient-ventilator inter-action and sleep quality in 13 mechani-cally ventilated patients by comparingpressure support to proportional assistventilation. The overall sleep quality wasbetter on proportional assist ventilationdue to fewer arousals, fewer awakenings,a greater proportion of REM and slow-wave sleep, and less patient-ventilatorasynchrony. Toublanc et al. (1) con-ducted a recent study comparing assist-control ventilation to low levels of PSV(at 6 cm H2O) on sleep quality in intu-bated patients. Consistent with previousfindings, regardless of the mode of me-chanical ventilation used, sleep architec-ture was abnormal, with up to 40% of thenight spent in wakefulness. There was apredominance of stages 1 and 2, whileslow-wave sleep and REM periods weresignificantly reduced. However, ACVcompared with PSV reduced wakefulnessduring the first part of the night andincreased slow-wave sleep during the sec-ond part of the night. These studiesraised the possibility that there could bean optimal mode of mechanical ventila-tion that provides good-quality sleep inpatients requiring support from a me-chanical ventilator.

In this issue of Critical Care Medicine,the interesting although small study byDr. Cabello and colleagues (8) addressedsleep quality on three different modes of

mechanical ventilation. Fifteen patientsrequiring ventilatory support were placedon assist ACV, clinically adjusted PSV(cPSV), and automatically adjusted PSV(aPSV). Consistent with previous studies,the authors found reduced REM sleep anda high frequency of arousals and awaken-ings. However, unlike previous studies byParthasarathy and Tobin (4) and Tou-blanc et al. (1), where ACV seemed to bethe superior mode, Dr. Cabello and col-leagues found that none of the three ven-tilatory modes were superior to the oth-ers. Sleep architecture and quality wereequally affected in all three modes of me-chanical ventilation. The reduced amountof REM sleep, increased arousals andawakenings, and high sleep fragmenta-tion were similar in all three modes. Cen-tral apneas were present during pressuresupport ventilation but occurred muchless frequently compared with the studyby Parthasarathy and Tobin, accountingfor �10% of sleep fragmentation. Thismay be explained in part by a loweramount of pressure delivered during thisstudy. Using more sensitive indicators forrespiratory effort would be helpful, suchas the esophageal pressure monitor. Inthis study, respiratory efforts were mea-sured by chest and abdominal inductanceplethysmography. Esophageal pressuremonitoring is invasive, and in the case ofpatients requiring ventilatory support, itmeasures diaphragmatic excursions.Stoohs et al. (9) evaluated noninvasive es-timation of esophageal pressure based onintercostal electromyography in patientswith sleep-disordered breathing and foundthat this method can be used to estimateesophageal pressure with sufficient accu-racy. Increased respiratory effort causingarousals may be better quantified usingmore sensitive monitoring, such as thetechnique described by Stoohs and col-leagues.

Based on studies by Toublanc et al. (1)and Parthasarathy and Tobin (4), pres-sure support ventilation is an inferiormode of maintaining good-quality sleep.Fanfulla et al. (10) reported that the typeof ventilator mode might have a majorinfluence on sleep quality and neuromus-cular function in patients with knownneuromuscular disease. Chen and Tang(11) showed that sleep loss might impairinspiratory muscle endurance in intu-bated patients. Muscle strength is a sig-nificant determinant of successful wean-ing, and therefore the quality of sleepshould be an important focus for physi-cians. It is a common practice in the

intensive care setting to maintain or in-crease ventilatory support for intubatedpatients during nocturnal hours. Tou-blanc et al. suggested that PSV may notbe the optimal mode to ensure adequateventilatory support while patients arerested during nocturnal hours of sleep.The study by Dr. Cabello and colleagues(8) highlights the importance of an ap-propriate amount of pressure delivered toprevent the occurrence of central apneasleading to sleep fragmentation. If pres-sure support ventilation is to be used,then perhaps it is important to define thecorrect pressure to avoid central apneasleading to arousals/awakenings and poorsleep efficiency.

Circadian rhythm disturbances cansignificantly affect sleep. Dr. Cabello andcolleagues (8) determined sleep start at2 p.m. with three 6-hr sessions ending at8 a.m. the following day. Sleep start wasdetermined at 2 p.m. because most rou-tine ICU care was performed between8 a.m. and 2 p.m. Patients in the ICU maynot demonstrate normal circadianrhythm fluctuations as defined by day-light and night. Hardin et al. (12) dem-onstrated that 18 ICU patients monitoredwith 24-hr polysomnography displayedabnormal sleep/wake cycles with erraticprogression through sleep stages. Pa-tients were not quantitatively sleep de-prived, but �50% of sleep occurred indaytime. Starting sleep time at 2 p.m.may contribute to the already present cir-cadian disturbances in this populationgroup. Rapid eye movement sleep is mostdifficult to achieve and is at its lowest inpercentage of total sleep time duringmid-afternoon. In Dr. Cabello and col-leagues’ study, REM sleep was nonexist-ent in the daytime period and the propor-tion of REM sleep progressively increasedin the second nocturnal period. The au-thors should take into account the alter-ation of normal circadian rhythm in theintensive care setting and how it contrib-utes to poor quality sleep.

Medications routinely used in the ICUcommonly cause sleep disturbances, inparticular benzodiazepines and opioids.In this study by Dr. Cabello and col-leagues (8), the average duration of ven-tilatory support for a patient was 22 � 17days. These patients required prolongedventilatory support and therefore likelyrequired prolonged sedation and paincontrol. Although the patients had beenoff sedatives and opioids for �24 hrs, theeffects of these medications may still be


clinically significant. The dose-dependentrelationships of these medications arebased on renal and hepatic metabolism,as well as the volume of distribution, allof which are abnormal in the critically illpatient. Benzodiazepines in particular areknown to decrease REM sleep, and opioidshave been shown to cause respiratory de-pression and induce central apneas. Allow-ing for a �24-hr benzodiazepine and opi-oid-free duration before the start of sleeptime recording would be prudent.

Dr. Cabello and colleagues (8) addressa difficult but important issue in the ICU.Sleep disturbances in the intensive caresetting can bring about acute and long-term physiologic, psychological, and be-havioral consequences for the patient.Physicians should pay closer attention tosleep-related issues in this patient popu-lation and attempt to find resolution.Many questions remain unanswered inregard to the topic of sleep in the inten-sive care environment. Much more is ex-pected from the field of sleep researchfocusing on this patient population, andwe hope that larger well-designed studieswill answer our questions.

Michelle Cao, DOPriscilla S. A. Sarinas, MD

Stanford University SleepDisorders Clinic

Stanford University School ofMedicine

Division of Pulmonary,Critical Care, and SleepMedicine

VA Palo Alto Health CareSystem

Palo Alto, CA

REFERENCES

1. Toublanc B, Dominique R, Jean-Charles G,et al: Assist-control ventilation vs. low levelsof pressure support ventilation on sleep qual-ity in intubated ICU patients. Intensive CareMed 2007; 33:1148–1154

2. Cooper A, Thornley K, Young B, et al: Sleepin critically ill patients requiring mechanicalventilation. Chest 2000; 117:809–818

3. Bosma K, Ferreyra G, Ambrogio C, et al:Patient-ventilator interaction and sleep inmechanically ventilated patients: Pressuresupport versus proportional assist ventila-tion. Crit Care Med 2007; 35:1048–1054

4. Parthasarathy S, Tobin M: Effect of ventilatormode on sleep quality in critically ill pa-tients. Am J Respir Crit Care Med 2002; 166:1423–1429

5. Gabor J, Cooper A, Crombach S, et al: Con-tribution of the intensive care unit environ-ment to sleep disruption in mechanically

ventilated patients and healthy subjects.Am J Respir Crit Care Med 2003; 167:708–715

6. Freedman N, Gazendam J, Levan L, et al:Abnormal sleep/wake cycles and the effect ofenvironmental noise on sleep disruption inthe intensive care unit. Am J Respir Crit CareMed 2001; 163:451–457

7. Meza S, Mendez M, Ostrowski M, et al: Sus-ceptibility to periodic breathing with assistedventilation during sleep in normal subjects.J Appl Physiol 1998; 85:1929–1940

8. Cabello B, Thille AW, Drouot X, et al: Sleepquality in mechanically ventilated patients:Comparison of three ventilatory modes. CritCare Med 2008; 36:1749–1755

9. Stoohs RA, Blum HC, Knaack L, et al: Non-invasive estimation of esophageal pressurebased on intercostal EMG monitoring. In:Proceedings of the 26th Annual InternationalConference of the IEEE EMBS, September2004, San Francisco, CA

10. Fanfulla F, Delmastro M, Berardinelli A, et al:Effects of different ventilator settings onsleep and inspiratory effort in neuromuscu-lar patients. Am J Respir Crit Care Med 2005;172:619–624

11. Chen HI, Tang YR: Sleep loss impairs in-spiratory muscle endurance. Am Rev RespirDis 1989; 140:907–909

12. Hardin K, Seyal M, Stewart T, et al: Sleep incritically ill chemically paralyzed patients re-quiring mechanical ventilation. Chest 2006;129:1468–1477

Acute compartment syndrome: Are we close to making an earlydiagnosis?*

Compartment syndrome, an or-thopedic emergency, is thecondition of increased pres-sure within a limited anatom-

ical space that compromises tissue perfu-sion and, ultimately, causes tissuenecrosis, rhabdomyolysis, renal failure,and even death (1). Compartment syn-drome can develop anywhere skeletalmuscle is surrounded by substantial fas-cia and most commonly is caused by frac-tures, soft tissue trauma, arterial injury,

burns, and strenuous physical exertion(2–5).

Timely diagnosis is important to avoidirreversible consequences of compart-ment syndrome. Although time of injuryor ischemia to tissue necrosis can vary,currently the upper limit of accepted tis-sue viability is 6 hrs (6). Diagnosis can bechallenging even for those who are awareof the possibility of compartment syn-drome in certain clinical settings due tolack of a definitive diagnostic measure.Moreover, frequent evaluations of pa-tients may be needed if clinical findingsare not clear. Relevant history and classicclinical symptoms of pain, paresthesia,palor, and pulselessness may be enoughfor diagnosis, whereas early cases or un-conscious trauma patients may not havetypical presentation. In those instances,laboratory tests (serum creatine phos-

phokinase, serum and urine myoglobin),radiography, and ultrasound/Doppler ofthe extremity may give additional infor-mation. A number of commercially avail-able tonometers (i.e., Stryker, ACE) canbe used to measure the compartmentpressures. While pressures of 0–15 mmHg are considered normal, pressure �30mm Hg in the presence of appropriateclinical signs is diagnostic for compart-ment syndrome (7). While investigatorshave studied methods of diagnosingchronic exertional compartment syn-drome with newer noninvasive tools,such as using magnetic resonance imag-ing at 0.1 T (8) or near-infrared spectros-copy to measure hemoglobin saturationof deep tissues as an indicator of bloodflow (9), the diagnosis of acute compart-ment syndrome relies on conventionalmethods.

*See also p. 1756.Key Words: acute compartment syndrome; ex-

tremity; leg; diagnosis; infraredThe author has not disclosed any potential con-



DOI: 10.1097/CCM.0b013e318176129c


In this issue of Critical Care Medicine,Dr. Katz and colleagues (10) describeearly detection of acute compartmentsyndrome in trauma patients by using anovel technique of infrared imaging. Thistechnique is proposed to provide an earlydiagnosis of acute compartment syn-drome by using thermographic images ofextremities to detect lower surface tem-peratures, which can be seen in compart-ment syndrome due to decreased bloodflow in the affected extremities. In thisobservational study, the authors mea-sured the average temperature of the an-terior surface of the proximal and distalregion of each leg, termed the thigh-footindex, with an uncooled microbolometerinfrared camera in 164 patients who pre-sented to a level I trauma center withtrauma. Research assistants obtained thethermographic images using the sameangle and distance (90 cm) in patientswho presented within 4 hrs of trauma.Eleven of 164 analyzed patients developedcompartment syndrome. Those with acuteleg compartment syndrome (unilateral orbilateral) had significantly lower anteriorleg surface temperatures than did the pa-tients with contralateral leg syndrome(unilateral compartment syndrome) andthe control group (bilateral compartmentsyndrome). Similarly, thigh-foot index wasgreater in the legs with compartment syn-drome. Analysis of covariance revealed nosignificance in the surface temperaturemeasurements for the variables of bloodpressure, presence of shock, body temper-ature, ambient temperature, and lower ex-tremity fractures.

Although this is a unique method andholds promise as a noninvasive diagnostictool for acute compartment syndrome,future studies are needed to eliminate thelimitations of the current study and val-idate its clinical utility. As noted by the

authors, the study has numerous limita-tions. First, this was an observationalstudy based on retrospective chart reviewfor final diagnosis of compartment syn-drome. Therefore, clinical utility, espe-cially the decision for fasciotomy, cannotbe drawn from the current data. Second,no correlation with conventional diag-nostic methods (clinical symptoms, phys-ical examination, measurements of com-partment pressures) was performedprospectively. Therefore, we have no dataregarding whether this new techniquewould provide an early and accurate di-agnosis of acute compartment syndrome.Moreover, no data were provided regard-ing whether a diagnosis of compartmentsyndrome was missed based on chart re-view and thermographic measurements.Third, most patients diagnosed withacute compartment syndrome had sus-tained blunt extremity trauma and frac-ture. We have no data as to whether in-frared imaging would help determinecompartment syndrome from othercauses. Fourth, the study had technicallimitations: Technical skills are requiredto properly calibrate the camera and useit to collect data. As mentioned in thearticle, images from seven patients couldnot be used due to poor quality. Fifth, thecost of equipment and having readilyavailable personnel who can operate thesystem can be problematic in the emer-gency department.

Once these limitations are addressed infuture studies, infrared imaging may poten-tially provide a supportive measure to diag-nose acute compartment syndrome. Untilthen, clinical judgment and conventionaldiagnostic measures must be used to diag-nose acute compartment syndromes.

H. Erhan Dincer, MD, FCCP,FAASM

Division of Pulmonary,Critical Care & SleepMedicine

VA Southern Nevada HealthCare System

Las Vegas, NV

REFERENCES

1. Matsen FA: Compartmental Syndrome. NewYork, Grune and Stratton, 1980

2. Tornetta P, Templeman D: Compartmentsyndrome associated with tibial fracture. In-str Course Lect 1997; 46:303–308

3. Roberge RJ, McLane M: Compartment syn-drome after simple venipuncture in an anti-coagulated patient. J Emerg Med 1999; 17:647–649

4. Mubarak SJ, Hargens AR: Acute compart-ment syndromes. Surg Clin North Am 1983;63:539–565

5. Berlemann U, Al-Momani Z, Hertel R: Exer-cise-induced compartment syndrome in theflexor-pronator muscle group. Am J SportsMed 1998; 26439–441

6. Matsen FA, Winquist RA, Krugmire RB: Di-agnosis and management of compartmentalsyndromes. J Bone Joint Surg Am 1980; 62:200–203

7. Heppenstall RB, Sepega AA, Scott R, et al:The compartment syndrome: An experimen-tal and clinical study of muscular energymetabolism using phosphorous nuclearmagnetic resonance spectroscopy. Clin Or-thop 1988; 226:138

8. Eskelin MK, Lotjonen JM, Mantysaari MJ:Chronic exertional compartment syndrome:MR imaging at 0.1 T compared with tissuepressure measurement. Radiology 1998; 206:333–337

9. van Den Brand JG, Verleisdank EJ: Near in-frared spectroscopy in the diagnosis ofchronic exertional compartment syndrome.Am J Sports Med 2004; 32:452–457

10. Katz LM, Nauriyal V, Nagaraj S, et al: Infra-red imaging of trauma patients for detectionof acute compartment syndrome of the leg.Crit Care Med 2008; 36:1756–1761


Death by TPN . . . the final chapter?*

I t seems obvious that the humanbody was designed to receivenourishment and hydration viathe gastrointestinal route; this

model of nourishment is found through-out the animal kingdom. It is now widelyaccepted that the gastrointestinal tractis the preferred route of delivering nutri-tional support even in hospitalizedpatients (1). Furthermore, consensusguidelines strongly recommend enteralover parenteral nutrition in critically illpatients (2, 3). Indeed, for patients inwhom enteral nutrition is possible (i.e.,who have a functional gastrointestinaltract), there is no disease state that isknown to benefit from parenteral nutri-tion (1). The institution of early enteralnutrition in critically ill medical andpostoperative patients has been demon-strated to improve outcome (4, 5). Yetparenteral nutrition continues to bewidely used in patients who can be fedenterally. The results from the GermanCompetence Network Sepsis (SepNet)study by Dr. Schmitz and colleagues (6)in this issue of Critical Care Medicineconfirm this observation; only 20.1% ofintensive care unit (ICU) patients withsevere sepsis/septic shock received exclu-sively enteral feeding during their ICUstay. In addition, these authors confirmwhat should be no surprise; parenteralnutrition was an independent predictor ofdeath (odds ratio of 2.09). The adversesequelae associated with parenteral nutri-tion result from 1) not directly feedingthe bowel; 2) the metabolic, immuno-logic, endocrine, and infective complica-tions associated with parenteral nutri-tion; and 3) the fact that parenteralnutrition is infused into the patient’s sys-temic venous system, bypassing the liver.This latter concept is supported by the

recent study by Grau et al. (7), who in alarge cohort of critically ill patients dem-onstrated that parenteral nutrition wasstrongly associated with the developmentof liver dysfunction, while early enteralnutrition was protective. The hepatotox-icity of parenteral nutrition may partlyexplain the perniciousness of this admix-ture, considering the central role the liverplays in metabolism and homeostasis.

The findings of the SepNet study areall the more important as the study tookplace in the era of “tight glycemic con-trol” and the mean blood glucose concen-tration was similar between enterally andparenterally fed patients. It has previ-ously been argued that the increased riskof infections and death associated withparenteral nutrition is due to poor glyce-mic control and that with adequate glu-cose control parenteral nutrition is safe(8, 9). Clearly, this is not the case! Further-more, it is likely that the parenteral nutri-tion admixtures in the United States aremore deadly than those available in Europe,because intravenous glutamine solutionsand non-soy-based lipid emulations, whichhave been demonstrated to reduce the tox-icity of parenteral nutrition, are not avail-able in the United States (10).

Why would clinicians purposefullyharm their patients with a toxic solutionthat would certainly not meet currentsafety standards? Again, the article by Dr.Schmitz and colleagues (6) may shedsome light on this question. Mechanicalventilation, intra-abdominal disease, andthe presence of shock were independentlyassociated with the failure to feed enter-ally. These exclusions are clearly contraryto current evidence, which suggests thatearly enteral nutrition improves out-comes in these patient groups and thatenteral nutrition may protect the gastro-intestinal tract from vasopressor-inducedischemia (4, 11, 12). In addition, manyclinicians are still under the misguidedimpression that enteral nutrition andparenteral nutrition are equivalent. Thismyth (9, 13) is unfortunately propagatedby experts who contend that “appropri-ately administered parenteral nutritionmay provide similar or more benefit thanenteral (nutrition) and clearly needs

more widespread acceptance” (9). Enteralnutritional support may not be possiblein patients with intestinal failure, such asshort-gut syndromes, and in those withhigh-output small-bowel fistulas. How-ever, these patients are uncommon inclinical practice. Almost all patients whoreceive parenteral nutrition in the ICUhave a functional gastrointestinal tractand can be fed enterally.

The risk of pulmonary aspiration andventilator-associated pneumonia (VAP) isoften cited as a reason to delay enteralfeeding. Feeding mechanically ventilatedpatients with a nasogastric tube in thesupine position is clearly a risk factor forVAP (14). However, it is unclear whetherearly enteral nutrition increases the riskof VAP when measures to prevent VAP aresystematically applied and patients arefed slowly (15). Data from the ProjectImpact Data System suggested that early(as opposed to delayed) enteral nutritionwas associated with a small but nonsig-nificant increase in the risk of VAP. How-ever, overall mortality was lower in the“early enteral group,” with both groupshaving an equivalent number of ventila-tor-free days (4). This suggests that de-spite a possible increase in the risk ofVAP, early enteral feeding has an overallbenefit. In addition, it is unclear whethersmall-bowel feeding decreases the risk ofVAP (16, 17). Therefore, I recommendthat an orogastric tube be placed in allventilated patients on admission to theICU and tube feeding be started within 12hrs (except when extubation within 24hrs is planned). Tube feeding should beinitiated slowly (20 mL/hr) with permis-sive underfeeding for the first 3–5 days(15–20 kcal/kg/day) and then advanced toa goal of 20–25 kcal/kg/day as the pa-tient’s condition improves. In patientswith gastric intolerance, promotility agents(such as erythromycin, 100 mg intrave-nously every 8 hrs) and small-bowel feedingshould be considered. With this approach,almost 100% of ICU patients can be nour-ished in the way that nature intended.

Paul E. Marik, MD, FCCMDivision of Pulmonary and

Critical Care Medicine

*See also p. 1762.Key Words: parenteral nutrition; enteral nutrition;

sepsis; shock; critically ill; nutrition; mechanical ven-tilation

The author has not disclosed any financial interestin any of the products mentioned in this article.


DOI: 10.1097/CCM.0b013e31817612d6


Thomas Jefferson UniversityPhiladelphia, PA

REFERENCES

1. Zaloga GP: Parenteral nutrition in adult in-patients with functioning gastrointestinaltracts: Assessment of outcomes. Lancet2006; 367:1101–1111

2. Heyland DK, Dhaliwal R, Drover JW, et al:Canadian clinical practice guidelines for nu-trition support in mechanically ventilated,critically ill adult patients. JPEN J ParenterEnteral Nutr 2003; 27:355–373

3. Kreymann KG, Berger MM, Deutz NE, et al:ESPEN guidelines on enteral nutrition: In-tensive care. Clin Nutr 2006; 25:210–223

4. Artinian V, Krayem H, DiGiovine B: Effects ofearly enteral feeding on the outcome of crit-ically ill mechanically ventilated medical pa-tients. Chest 2006; 129:960–967

5. Marik PE, Zaloga GP: Early enteral nutritionin acutely ill patients: A systematic review.Crit Care Med 2001; 29:2264–2270

6. Schmitz G, Schädler D, Engel C, et al: Cur-

rent practice in nutritional support and itsassociation with mortality in septic pa-tients—Results from a national, prospective,multicenter study. Crit Care Med 2008; 36:1762–1767

7. Grau T, Bonet A, Rubio M, et al: Liver dys-function associated with artificial nutritionin critically ill patients. Crit Care 2007; 11:R10

8. Perioperative total parenteral nutrition insurgical patients. The Veterans Affairs TotalParenteral Nutrition Cooperative StudyGroup. N Engl J Med 1991; 325:525–532

9. Bistrian BR, McCowen KC: Nutritional andmetabolic support in the adult intensive careunit: Key controversies. Crit Care Med 2006;34:1525–1531

10. Marik PE: Maximizing efficacy from paren-teral nutrition in critical care: Appropriatepatient populations, supplemental parenteralnutrition, glucose control, parenteral glu-tamine, and alternative fat sources. CurrGastroenterol Rep 2007; 9:345–353

11. Marik PE, Zaloga GP: A meta-analysis of par-

enteral nutrition vs enteral nutrition inacute pancreatitis. BMJ 2004; 328:1407–1409

12. Zaloga GP, Roberts PR, Marik PE: Feedingthe hemodynamically unstable patient: Acritical evaluation of the evidence. Nutr ClinPract 2003; 18:285–293

13. Jeejeebhoy KN: Total parenteral nutrition:potion or poison? Am J Clin Nutr 2001; 74:160–163

14. Drakulovic MB, Torres A, Bauer TT, et al:Supine body position as a risk factor fornosocomial pneumonia in mechanically ven-tilated patients: A randomised trial. Lancet1999; 354:1851–1858

15. Collard HR, Saint S, Matthay MA: Preventionof ventilator-associated pneumonia: An evi-dence-based systematic review. Ann InternMed 2003; 138:494–501

16. Goncalves RR, Leite HP, Nogueira PC: Doessmall bowel feeding decrease the risk of ven-tilator-associated pneumonia? JPEN J Par-enter Enteral Nutr 2004; 28:60

17. Marik PE, Zaloga G: Gastric vs post-pyloricfeeding? A systematic review. Crit Care 2003;7:R46–R51

Glutamine and tight glycemic control: Chicken or egg?*

F or a clinician striving to de-liver the optimal nutrient reg-imen to the critically ill, it is areassurance to learn that two

of the most significant clinical develop-ments in nutrition that impact on out-come are themselves intimately related.The glutamine story (1) is one of systemfailures arising as a consequence of aconditional deficiency due to a limitedendogenous production in skeletal mus-cle failing to meet increased demands.The activation of muscle proteolysis withrelative inhibition of protein synthesisleads to muscle wasting and release ofamino acids, many of which are metabo-lized to glutamine. However, it is alsorecognized that part of the carbon usedin glutamine production also is linkedto the glucose delivered to muscle. Thisis optimal in hyperinsulinemic euglyce-mia and is impaired in hypoglycemia or

when insulin resistant (2). Perhaps it maybe possible to get an optimal glutaminesupply without scavenging the body pro-tein reserve by simply combining ade-quate nutrient provision with sufficientexogenous insulin to give tight glycemiccontrol?

The elegant clinical research study byDr. Gianni Biolo and colleagues (3) inthis issue of Critical Care Medicine startsto suggest that this actually may be pos-sible. They have studied a select group ofcancer patients undergoing radical ab-dominal surgery. Using carefully con-trolled metabolic studies involving stableisotope tracers, they show that by main-taining strict euglycemia with insulin incompany with good nutrient delivery,they were able to increase glucose uptakeand increase protein synthesis with amore neutral protein balance indicativeof less catabolism of muscle. This re-sulted in a higher glutamine synthesisand plasma glutamine concentration.This contrasts with their previous workon growth hormone (4), which showedthat, although stimulating protein syn-thesis, it was associated with a reductionin glutamine availability.

Each patient provided his/her owncontrol in a 24-hr crossover study in ran-

domized order, once with moderate hy-perglycemia (mean, 9.6 mmol) and theother with insulin-mediated tight eugly-cemia (mean blood glucose, 5.8 mmol)soon after surgery. They received contin-uous intravenous nutrition of glucose,lipids, and amino acids to a total energyof 28 kcal/kg/day for the 48 hrs of study.An interesting feature was that the in-creased endogenous insulin productionduring the hyperglycemic periods couldbe reduced (shown by lower C-peptide)during the tight euglycemia period andthis was associated with improved insulinsensitivity. Are there limitations? Yes,these are stressed patients but not septicor critically ill with multiple organ fail-ure. They only studied eight female pa-tients for 48 hrs, so we do not know if theeffect is sustained. However, their find-ings are consistent and the changes ob-served are of a magnitude to suggest clin-ical significance.

Does this seems too good to be true?Will using a tight glycemic protocol alonesolve the glutamine issue as well? Wemust recognize that intensive insulintherapy in some hands appears problem-atic (5) with increased hypoglycemia, butthen these workers did not follow a de-tailed nutrition protocol that prescribed a

*See also p. 1768.Key Words: glutamine; insulin; tight glycemic con-

trol; euglycemia; nutrition; glucoseDr. Griffiths has received grant support from

Fresenius Kabi.Copyright © 2008 by the Society of Critical Care


DOI: 10.1097/CCM.0b013e3181761672


reasonable glucose load to avoid this risk.From the paper, it appears that in theirmoderately obese population (BMI of 27)over the first few days, they were deliver-ing on average �1000 kcals per day intotal. This is a period of great instabilitywith considerable variability and incon-sistency of nutrient delivery. This con-trasts with the careful attention to con-sistent nutrient delivery shown in a paperfrom the Leuven group (6). Studies inseptic critically ill patients show that glu-cose oxidation is not impaired; it is thenonoxidative disposal (e.g., into skeletalmuscle) that accounts for reduced totalutilization and appearance of hyperglyce-mia. Therefore, with physiologic levels ofinsulin, an optimal glucose utilizationrate in sepsis (and hence safe delivery rateif not exceeded) is 4 mg/kg fat free mass/min (7). Contrary to what some appear tobelieve (5), it may be important to deliverglucose within this range of between 8 to12 g/hr, which is precisely what was donein the original studies (6, 8).

But perhaps the glutamine deficiencythat occurs plays a role in worsening in-sulin sensitivity? Glutamine, as a paren-teral infusion, can favorably influence in-sulin-mediated glucose utilization (9)and in healthy adults can beneficially in-fluence postprandial insulin action, glu-cose disposal, and fat oxidation (10). Clin-ical studies of glutamine (11), as well asshowing a substantially reduced infectionand pneumonia rate, demonstrate signif-icant reduction of hyperglycemia and asignificant reduction in the number ofpatients requiring insulin. As has beenelegantly demonstrated, multiple-traumapatients (12) receiving parenteral glu-tamine show improved insulin sensitiv-

ity. This suggests that we cannot ignoreexogenous glutamine provision, and thatperhaps glutamine and tight glycemiccontrol are complementary with each fa-cilitating the other.

There is no doubt that tight glycemiccontrol is not a simple therapy, butrather a complex integrated process ofcare that requires considerable attentionto all details to be performed safely. Sim-ply controlling glucose levels is notenough, because it masks the signs andtherefore risks of overfeeding (13), whichsadly appears endemic and cavalier withparenteral nutrition use (14). For tightglycemic control to work safely and effec-tively, making sure there is consistentnutrient delivery, neither overfeeding norunderfeeding, with attention to glucoseand glutamine in the initial period maybe the key.

Richard D. Griffiths, BSc, MBBS,MD, FRCP

University of LiverpoolLiverpool, UK

REFERENCES

1. Bongers T, Griffiths RD, McArdle A: Exoge-nous glutamine: The clinical evidence. CritCare Med 2007; 35(Suppl 9):S545–S552

2. Meyer C, Woerle HJ, Gerich J: Paradoxicalchanges of glutamine release during hyperin-sulinemia euglycemia and hypoglycaemia inhumans: Further evidence for glucose–glutamine cycle. Metabolism 2004; 53:1208–1214

3. Biolo G, De Cicco M, Lorenzon S, et al:Treating hyperglycaemic improves skeletalmuscle protein metabolism in cancer pa-tients after major surgery. Crit Care Med2008; 36:1768–1775

4. Biolo G, Iscra F, Bosutti A, et al: Growthhormone decreases muscle glutamine pro-duction and stimulates protein synthesis in

hypercatabolic patients. Am J Physiol Endo-crinol Metab 2000; 279:E323–E332

5. Brunkhost FM, Engel C, Bloos F, et al: In-tensive insulin therapy and pentastarch re-suscitation in severe sepsis. N Engl J Med2008; 358:125–139

6. Van den Berghe G, Wilmer A, Milants I, et al:Intensive insulin therapy in mixed medical/surgical intensive care unit: Benefits versusharm. Diabetes 2006; 55:3151–3159

7. Carlson G: Hunterian Lecture: Insulin resis-tance in human sepsis: Implications for thenutritional and metabolic care of the criti-cally ill surgical patient. Ann R Coll SurgEngl 2004; 86:75–81

8. Van den Berghe G, Wilmer A, Hermans G, etal: Intensive insulin therapy in the medicalICU. N Engl J Med 2006; 354:449–461

9. Borel MJ, Williams PE, Jabbour K, et al:Parenteral glutamine infusion alters insulin-mediated glucose metabolism. JPEN J Par-enter Enteral Nutr 1998; 22:280–285

10. Iwashita S, Mikus C, Baier S, et al: Glutaminesupplementation increases postprandial en-ergy expenditure and fat oxidation in hu-mans. JPEN J Parenter Enteral Nutr 2006;30:76–80

11. Déchelotte P, Hasselmann M, Cynober L, et al:L-alanyl-L-glutamine dipeptide–supplementedtotal parenteral nutrition reduces infectiouscomplications and glucose intolerance in crit-ically ill patients: The French controlled, ran-domized, double-blind, multicenter study. CritCare Med 2006; 34:598–604

12. Bakalar B, Duska F, Pachl J, et al: Parenter-ally administered dipeptide alanyl-glutamineprevents worsening of insulin sensitivity inmultiple-trauma patients. Crit Care Med2006; 34:381–386

13. Griffiths RD: Too much of a good thing: Thecurse of overfeeding. Crit Care 2007; 11:176

14. Dissanaike S, Shelton M, Warner K, et al: Therisk for bloodstream infections is associatedwith increased parenteral caloric intake inpatients receiving parenteral nutrition. CritCare 2007; 11:R114


Hypothermia and coronary intervention after cardiac arrest:Thawing a cool relationship?*

T herapeutic hypothermia (TH)represents one of the mostpromising treatments for car-diac arrest victims since the

introduction of defibrillation and cardio-pulmonary resuscitation �50 yrs ago (1).Supported by two landmark randomizedcontrolled trials (2, 3) as well as a num-ber of smaller clinical and laboratorystudies, current resuscitation guidelinesrecommend that comatose survivors ofout-of-hospital ventricular fibrillation(VF) cardiac arrest be cooled to 32°C to34°C for �12 hrs after return of sponta-neous circulation (ROSC). This therapyhas been shown to improve survival as wellas neurologic outcomes among survivors.It is generally believed that the cooling pro-cess should be initiated as soon as possibleafter ROSC, based on animal investigationsof postarrest TH (4, 5).

A large number of questions remain tobe answered regarding the application ofhypothermia, and such questions havecontributed to the slow adoption of thiscrucial therapy (6). A number of hospitalselect to treat patients after cardiac arrestinvolving non-VF rhythms, such as asys-tole or pulseless electrical activity, or pa-tients suffering in-hospital cardiac arrest(7); it remains to be definitively shownwhether such patients benefit from cool-ing. In addition, the time-sensitivity ofcooling initiation has not been clearlydefined. This latter question is of greatpractical importance, as a number ofpost-VF survivors are found to have ST-segment elevation myocardial infarction(STEMI) requiring immediate percutane-

ous coronary intervention (PCI). Shouldcooling be started before PCI or can itwait until after this process, introducinga likely delay of several hours? Wouldhypothermia induction delay the time-sensitive intervention of PCI? TH carriesa potential risk of increased bleeding; is itsafe to perform PCI, with concomitantuse of anticoagulants, such as heparin orglycoprotein IIb/IIIa inhibitors, while hy-pothermia is being induced? These con-cerns have created tension between twocompeting modalities that are essentialtreatments for survivors of cardiac arrestdue to STEMI, which represents a sizablefraction of arrest patients.

The report by Dr. Wolfrum and col-leagues (8) in this issue of Critical CareMedicine is a useful addition to the sparseliterature addressing such practical ques-tions of TH implementation. In the au-thors’ observational study from one hos-pital, VF cardiac arrest patients achievingROSC and found to have electrocardio-graphic evidence of STEMI were treatedimmediately with TH. They subsequentlyreceived PCI during hypothermia treat-ment. These patients were compared withhistorical controls, namely resuscitatedSTEMI patients who received PCI withoutcooling. The authors found that the pro-cess of TH implementation did notlengthen door-to-balloon times for PCI.They also found that while complications,such as bleeding or infection, appeared tobe more common in the TH/PCI group,survival was not worse with the combina-tion therapy; indeed, there was a trendtoward improved survival. Finally, theyshowed that cooling can be quickly initi-ated with intravenous infusions of chilledsaline and application of external icepacks, a cost-effective method that ispractical for most settings.

While the investigation suffers fromsmall sample size (n � 33 patients) andthe likely confounders of patient selec-tion bias and secular trend (door-to-balloon times may have improved overtime independent of TH, especially givenincreased scrutiny of these patients by

investigators), the finding that coolingand PCI are not incompatible lays impor-tant groundwork for collaboration be-tween the emergency department and thecatheterization laboratory. Emergencyphysicians and intensivists need notworry about delaying PCI to establish hy-pothermia induction if systems are ap-propriately established for rapid cooling.Although further study is required, car-diologists may be reassured that coolingcan be initiated before PCI without sig-nificantly increased risks to the patientduring catheterization. This work adds toseveral other publications that haveshown the benefit of combined TH andPCI in postarrest patients (9–11); in oneof these other studies, Sunde et al. (11)astutely suggest that cooling and PCI arenecessary components of the larger suiteof postresuscitation critical care, includ-ing careful glycemic control, neurologicmonitoring, and appropriate hemody-namic support.

While a number of established proto-cols for hypothermia involve the use ofchilled saline infusions or external icepacks, such as those described by Dr.Wolfrum and colleagues (8) (a compila-tion of hospital protocols can be found athttp://www.med.upenn.edu/resuscitation/hypothermia/protocols.shtml), a note ofcaution should be given: These coolingmethods depend on careful monitoringby operators, as there are no inherentthermostatic controls to prevent over-cooling. Chilled saline and ice packs havebeen shown to be useful during coolinginduction, but these are blunt and un-wieldy tools that can overcool (12), withconcomitant increased risk of bleeding andarrhythmia, and that fail to adequatelymaintain cooling over 12–24 hrs (13).Newer thermostatically controlled devices,while expensive, appear to induce andmaintain cooling in a consistent fashion(14). Further investigations will be re-quired to determine the utility of combinedapproaches, such as using chilled saline asa cooling “accelerant” in concert with athermostatically controlled device.

*See also p. 1780.Key Words: cardiac arrest; cardiopulmonary resus-

citation; hypothermia; myocardial infarction; percuta-neous coronary intervention

Dr. Abella has received speaking honoraria fromAlsius Corporation, Medivance Corporation, GaymarIndustries, and Laerdal Medical Corporation. He hasalso received grant support from Philips Medical Sys-tems and Cardiac Science Corporation.


DOI: 10.1097/CCM.0b013e318176a874


Other important questions regardingPCI and cardiac arrest remain to be ad-dressed. While PCI is clearly indicated forarrest survivors with STEMI, what aboutsurvivors of VF without STEMI? It islikely that a large fraction of VF patientshave significant coronary disease or evenacute coronary syndrome without elec-trocardiographic evidence of STEMI.Should the majority of post-VF arrest pa-tients receive prompt coronary evalua-tion and possible intervention? In a pro-vocative report, Grogaard et al. (15)describe successful PCI of two VF patientsbefore ROSC, during active cardiopulmo-nary resuscitation. Should PCI be consid-ered part of the Advanced CardiovascularLife Support algorithm for VF patients? Itis hoped that future investigations usingcardiac arrest registry data and activeclinical trial networks will elucidate fur-ther the boundaries and practical issuessurrounding TH and postarrest care.These pragmatic advances in optimizingpostarrest care will allow cardiac arrestpatients to receive the best chance ofreturning home from the hospital aliveand neurologically intact.

Benjamin S. Abella, MD, MPhilDepartment of Emergency

Medicine and the Centerfor Resuscitation Science

University of PennsylvaniaSchool of Medicine

Philadelphia, PA

REFERENCES

1. Safar PJ, Kochanek PM: Therapeutic hypo-thermia after cardiac arrest. N Engl J Med2002; 346:612–613

2. Bernard SA, Gray TW, Buist MD, et al: Treat-ment of comatose survivors of out-of-hospital cardiac arrest with induced hypo-thermia. N Engl J Med 2002; 346:557–563

3. The Hypothermia After Cardiac Arrest StudyGroup: Mild therapeutic hypothermia to im-prove the neurologic outcome after cardiacarrest. N Engl J Med 2002; 346:549–556

4. Kuboyama K, Safar P, Radovsky A, et al:Delay in cooling negates the beneficial effectof mild resuscitative cerebral hypothermiaafter cardiac arrest in dogs: A prospective,randomized study. Crit Care Med 1993; 21:1348–1358

5. Abella BS, Zhao D, Alvarado J, et al: Intra-arrest cooling improves outcomes in a mu-rine cardiac arrest model. Circulation 2004;109:2786–2791

6. Merchant RM, Soar J, Skrifvars MB, et al:Therapeutic hypothermia utilization amongphysicians after resuscitation from cardiacarrest. Crit Care Med 2006; 34:1935–1940

7. Arrich J, European Resuscitation CouncilHypothermia After Cardiac Arrest RegistryStudy Group: Clinical application of mildtherapeutic hypothermia after cardiac arrest.Crit Care Med 2007; 35:1041–1047

8. Wolfrum S, Pierau C, Radke PW, et al: Mildtherapeutic hypothermia in patients afterout-of-hospital cardiac arrest due to acuteST-segment elevation myocardial infarc-tion undergoing immediate percutaneouscoronary intervention. Crit Care Med 2008;36:1780 –1786

9. Hovdenes J, Laake JH, Aaberge L, et al: Ther-apeutic hypothermia after out-of-hospitalcardiac arrest: Experiences with patientstreated with percutaneous coronary inter-vention and cardiogenic shock. Acta Anaes-thesiol Scand 2007; 51:137–142

10. Knafelj R, Radsel P, Ploj T, et al: Primarypercutaneous coronary intervention andmild induced hypothermia in comatose sur-vivors of ventricular fibrillation with ST-elevation acute myocardial infarction. Resus-citation 2007; 74:227–234

11. Sunde K, Pytte M, Jacobsen D, et al: Imple-mentation of a standardised treatment pro-tocol for post resuscitation care after out-of-hospital cardiac arrest. Resuscitation 2007;73:29–39

12. Merchant RM, Abella BS, Peberdy MA, et al:Therapeutic hypothermia after cardiac ar-rest: Unintentional overcooling is commonusing ice packs and conventional coolingblankets. Crit Care Med 2006; 34:S490–S494

13. Kliegel A, Janata A, Wandaller C, et al: Coldinfusions alone are effective for induction oftherapeutic hypothermia but do not keep pa-tients cool after cardiac arrest. Resuscitation2007; 73:46–53

14. Al-Senani FM, Graffagnino C, Grotta JC,et al: A prospective, multicenter pilot studyto evaluate the feasibility and safety of usingthe CoolGard System and Icy catheter follow-ing cardiac arrest. Resuscitation 2004; 62:143–150

15. Grogaard HK, Wik L, Eriksen M, et al: Con-tinuous mechanical chest compressions dur-ing cardiac arrest to facilitate restoration ofcoronary circulation with percutaneous cor-onary intervention. J Am Coll Cardiol 2007;50:1093–1094

Acute lung injury after blood product transfusion: Are the timeschanging?*

I n this issue of Critical Care Med-icine, Dr. Wright and colleagues(1) present a provocative study onthe United Kingdom’s experience

with the frequency of acute lung injury(ALI) after ruptured abdominal aortic an-

eurysm repair. This study raises a num-ber of interesting issues, including thegender selection of specific blood prod-ucts and the pathogenesis of ALI in pa-tients experiencing large amounts ofblood loss.

Acute lung injury from blood transfu-sions has most commonly been referredto as transfusion-related acute lung in-jury (TRALI). TRALI is an underreportedand underrecognized serious complica-tion of transfusion therapy. TRALI is de-fined as acute lung injury that developswithin 6 hrs (often within 30–60 mins) ofa blood product transfusion (2). However,there is also a more general association

with blood product transfusions and thedevelopment of ALI and an associationwith worse outcomes from ALI (3, 4). Theexact mechanisms of these latter associ-ations are unclear, but possibilities in-clude cases of overt and subclinicalTRALI, transfusion-related immuno-modulation predisposing to infectiouscauses of ALI (5), misclassification oftransfusion-associated circulatory over-load cases (6), or even an immune primingeffect of blood transfusions, predisposingthe patient to a more robust reaction to asecond, more common cause of ALI (i.e.,neutrophil priming followed by pulmonaryaspiration, pneumonia, sepsis).

*See also p. 1796.Key Words: acute lung injury; transfusion-related

acute lung injury; fresh frozen plasma; aortic aneu-rysm; blood transfusion



DOI: 10.1097/CCM.0b013e318176a8b2


There has been a general associationof TRALI with the transfusion of donorhuman leukocyte antigen (HLA) or neu-trophil antibodies that recognize the cog-nate antigen in the recipient. Many of thedonors implicated in TRALI cases are al-loimmunized women who harbor HLAantibodies. The U.K. National Blood Ser-vice, though monitoring from its SeriousHazards of Transfusion (SHOT) database,and the American Red Cross have identi-fied that donors from reported TRALIcases are commonly female and that theblood products implicated in these TRALIcases often contained HLA antibodies(7, 8). There appears to be a dose-dependent effect of the number of preg-nancies on HLA alloimmunization (9),with �20% of multiparous women har-boring HLA antibodies. Plasma frommultiparous donors was even shown pre-viously in a randomized, controlled trialto produce more hypoxemia in criticallyill patients (10). Recently, it has beenreported that intensive care patients whoreceived high plasma volume blood prod-ucts (fresh frozen plasma [FFP] or plate-lets) from female donors vs. male donorsdeveloped worsened oxygenation (11).Given the association of female donorswith TRALI cases, the U.K. NationalBlood Service initiated in 2003 the diver-sion of female blood donations away fromFFP and platelet concentrates. Is this reg-ulation of blood products misogyny orprescient policy?

The current study by Wright and col-leagues may yield valuable insight intotransfusion policy and important clinicaloutcomes (1). The investigators capital-ized on the U.K. policy change to ask thefollowing important question: In a criti-cally ill patient population requiring mul-tiple blood product transfusions, willmale-only FFP lead to a decreased fre-quency of postoperative ALI? The investi-gation was a retrospective, before/afterobservational trial at a single center thatcares for a large number of patients withruptured abdominal aortic aneurysms.Despite the overall limitations of thestudy design, interesting conclusions canbe made due to the large sample size (211patients over 7.5 yrs), reasonable match-ing of the two groups, and the before/after FFP policy. The authors should alsobe commended for using a sham studyprotocol to blind the data extractors andexpert radiologists to the working hy-pothesis. The primary outcome was the

frequency of ALI in the first 6 postoper-ative hours, a time course similar to clin-ical TRALI. Importantly, the patients re-ceiving male-only FFP for resuscitationhad a lower occurrence rate of postoper-ative ALI (21% vs. 36%). This was upheldby a multivariable logistic regressionanalysis controlling for potential con-founding variables. The patients receiv-ing male-only FFP also had a lower fre-quency of postoperative hypoxemia (PaO2/FIO2 �300 cutoff) and a trend for a highermean PaO2/FIO2 ratio. Given the trial de-sign, confounders that could bias theconclusions (and would not be controlledby the multivariate logistic regressionanalysis) include potential unmeasuredimbalances in the two patient groups andthe general improvement in intensivecare over time. In addition, the contribu-tion of male vs. female platelet transfu-sions was not specifically addressed inthis study, although platelet transfusions(or the plasma fraction of the plateletpool) are directed to be male-only in theUnited Kingdom. Also, even though therewas a lower frequency of ALI in the male-only FFP group, it is curious that theoverall mortality in the female vs. maleFFP groups was unchanged.

Nevertheless, these results are provoc-ative for a protective effect of male-onlyFFP in the resuscitation of these surgicalpatients. So, what exactly was the causeof ALI in this patient population? As ac-knowledged by the authors, the patho-genesis of ALI in patients with largeamounts of blood loss is probably multi-factorial, with contributions from shock,ischemia-reperfusion of the lungs, andpossibly the blood products themselves(TRALI). Importantly, only one case ofpostoperative ALI in this study was re-ported to the blood bank as possibly beingtransfusion-related, which highlights theunderrecognition and underreporting ofpotential TRALI cases.

Data are still being collected in theUnited Kingdom on how the change indonor allocation of blood products hasaffected the frequency of TRALI, althoughthere may be a trend for fewer referredcases (12). There are several other blood-banking policy initiatives that could po-tentially decrease the risk of TRALI (13).The American Association of Blood Bankshas made similar recommendations to di-vert potentially alloimmunized donorsfrom FFP and eventually from plateletdonations (14). Time will tell if the oc-

currence rate of TRALI, or perhaps evenALI, will change with these new recom-mendations. Prospective clinical trials ofALI and TRALI that do not rely on passivereporting may be helpful in detectingchanges in lung injury after blood prod-uct transfusion. For now, gender-specificblood product selection seems like a rea-sonable approach to improve the safety ofallogeic blood transfusions.

Mark R. Looney, MDDepartment of MedicineDivision of Pulmonary and

Critical Care MedicineUniversity of California,

San Francisco

REFERENCES

1. Wright SE, Snowden CP, Athey SC, et al:Acute lung injury after ruptured abdominalaortic aneurysm repair: The effect of exclud-ing donations from females from the produc-tion of fresh frozen plasma. Crit Care Med2008; 36:1796–1802

2. Toy P, Popovsky MA, Abraham E, et al:Transfusion-related acute lung injury: Def-inition and review. Crit Care Med 2005;33:721–726

3. Netzer G, Shah CV, Iwashyna TJ, et al: Asso-ciation of RBC transfusion with mortality inpatients with acute lung injury. Chest 2007;132:1116–1123

4. Gong MN, Thompson BT, Williams P, et al:Clinical predictors of and mortality in acuterespiratory distress syndrome: Potential roleof red cell transfusion. Crit Care Med 2005;33:1191–1198

5. Vamvakas EC: Pneumonia as a complica-tion of blood product transfusion in thecritically ill: Transfusion-related immuno-modulation (TRIM). Crit Care Med 2006;34:S151–S159

6. Skeate RC, Eastlund T: Distinguishing be-tween transfusion related acute lung injuryand transfusion associated circulatoryoverload. Curr Opin Hematol 2007; 14:682– 687

7. Stainsby D, Jones H, Asher D, et al: Serioushazards of transfusion: A decade of hemovigi-lance in the UK. Transfus Med Rev 2006;20:273–282

8. Eder AF, Herron R, Strupp A, et al: Transfu-sion-related acute lung injury surveillance(2003–2005) and the potential impact of theselective use of plasma from male donors inthe American Red Cross. Transfusion 2007;47:599–607

9. Densmore TL, Goodnough LT, Ali S, et al:Prevalence of HLA sensitization in femaleapheresis donors. Transfusion 1999; 39:103–106

10. Palfi M, Berg S, Ernerudh J, et al: A ran-domized controlled trial of transfusion-related acute lung injury: Is plasma frommultiparous blood donors dangerous?Transfusion 2001; 41:317–322


11. Gajic O, Yilmaz M, Iscimen R, et al: Trans-fusion from male-only vs. female donors incritically ill recipients of high plasma volumecomponents. Crit Care Med 2007; 35:1645–1648

12. Serious Hazards of Transfusion Annual Re-

port 2006. Available at: http://www.shotuk.org/SHOT%20report%202005.pdf. Accessed March25, 2008

13. Mair DC, Hirschler N, Eastlund T: Blood do-nor and component management strategiesto prevent transfusion-related acute lung in-

jury (TRALI). Crit Care Med 2006; 34:S137–S143

14. Strong DM, Lipton KS: American Associationof Blood Banks Association Bulletin 06–07.http://www.nibb.org/pdf/ab06 – 07.pdf. Ac-cessed February 25, 2008

Is extravascular lung water measurement in acute respiratorydistress syndrome worth the effort?*

Since it was first described byLyman Brewer et al. (1) inWorld War II as “wet lung” andnamed by Ashbaugh et al. in

1967 (2), acute respiratory distress inadults or adult respiratory distress syn-drome (ARDS) has remained a “syn-drome” and not a specific disease that canbe easily characterized or diagnosed. In1994, the American-European ConsensusConference determined specific criteriabased on chest radiograph (a gross exam),PaO2/FIO2 ratio (a gross physiologic vari-able), and absence of left atrial hyperten-sion (to rule out cardiogenic edema) (3).The search for a more reliable, reproduc-ible, and measurable variable continues.

In this issue of Critical Care Medicine,Dr. Berkowitz and colleagues (4) reintro-duce the use of extravascular lung water(EVLW) to help in the diagnosis andtreatment of the “noncardiogenic pulmo-nary edema” subset of acute lung injury.This is a technique first described usingtwo indicators (indocyanine green dyeand chilled saline) in 1951 by Newman etal (5). In the late 1970s and early 1980s,Elings et al. (6) perfected the techniqueand were able to use a single-indicatortechnique to estimate EVLW.

The twist that Dr. Berkowitz and col-leagues (4) add to this technique is theuse of predicted body weight (PBW) andadjusted body weight (AdjBW) based onheight, instead of actual body weight oran index based on height and weight(body surface area).

The authors propose that indexing toPBW or AdjBW increases the proportionof ARDS patients with elevated EVLW(defined as �10 mL/kg) and may helpdifferentiate them from patients withoutARDS. By adjusting for PBW, the authorscreated a situation in which all but one ofthe ARDS patients had at least one ele-vated measurement of EVLW. The num-ber of septic patients without clinicalARDS and an elevated EVLW decreased to50%.

This is a novel approach, because allprevious data, including animal data, havebeen expressed in mL/kg units. The origi-nal studies correlated EVLW via dye dilu-tion technique to wet weight/dry weightlung changes in animals (6). The formulaswere not ideal-weight adjusted. The for-mula for EVLW uses cardiac output, whichcan be indexed by body surface area as ameans of eliminating differences in size. Allthe authors’ hemodynamic measures (car-diac index, as shown in their Table 2) wereindexed. Using a new measurement (AdjBWor PBW) that is not used to normalize othervariables will need validation in futurestudies.

The use of EVLW as a tool to diagnoseand treat ARDS has drawbacks: It re-quires a pulmonary artery catheter, a5-Fr femoral artery catheter, and multi-ple measurements. EVLW should not beused at a single point in time but ratheras a method to follow the physiologicchanges of the capillary leak in ARDS andidentify the point at which the leak im-proves. This is labor intensive at present.Other physiologic, less invasive tech-niques, such as pulmonary leak scans (7,8), may be more useful to diagnose ARDS.

On the positive side, EVLW might be abetter determinant of the onset and res-olution of the acute phase of ARDS thanthe presently used criteria. Perhaps thepatients who had sepsis and elevated

EVLW but did not meet strict ARDS clinicalcriteria actually had subclinical ARDS. Ifthis is the case, changing variables, such asadjusting weight, would not be necessary tomake the patient fit the technology. In-stead, we believe that the technology im-proves the diagnosis of the patient’s disease.We need a better way to diagnose ARDS,and the article by Dr. Berkowitz and col-leagues (4) reinforces our desire to con-tinue to search for an objective measure toidentify this entity.

H. David Reines, MD, FACS, FCCMInova Fairfax HospitalDepartment of SurgeryFalls Church, VA

REFERENCES

1. Brewer LA, Burbank BB, Samson PC, et al:The wet lung in war casualties. Ann Surg1946; 123:343–362

2. Ashbaugh DG, Bigelow DB, Petty TL, et al:Acute respiratory distress in adults. Lancet1967; 2:319–323

3. Bernard GR, Artigas A, Brigham KL, et al: TheAmerican-European Consensus Conferenceon ARDS. Am J Respir Crit Care Med 1994;149:818–824

4. Berkowitz DM, Danai PA, Eaton S, et al: Ac-curate characterization of extravascular lungwater in acute respiratory distress syndrome.Crit Care Med 2008; 36:1803–1809

5. Newman EV, Merrell M, Genecin A, et al: Thedye dilution method for describing the centralcirculation: An analysis of factors shaping thetime-concentration curves. Circulation 1951;4:735–742

6. Elings VB, Lewis FR: A single indicator tech-nique to estimate extravascular lung water.J Surg Res 1982; 33:375–385

7. Spicer KM, Reines HD, Frey GD: Diagnosis ofadult respiratory distress syndrome with Tc-99m human serum albumin and portableprobe. Crit Care Med 1986; 14:669–676

8. Groenevald AB, Verheij: Extravascular lungwater to blood volume ratios as measures ofpermeability in sepsis induced ALI/ARDS. In-tensive Care Med 2006; 32:1315–1321

*See also p. 1803.Key Words: acute respiratory distress syndrome;

extravascular lung water; noncardiogenic edema; sepsisThe author has not disclosed any potential con-



DOI: 10.1097/CCM.0b013e318176a92c


Interleukin-6 plasma level in organ donors influences outcome forrecipients of organ transplantation: Is it clinically relevant?*

Organ shortage is a well-known and well-substanti-ated fact. In the UnitedStates in 2007, �26,000 pa-

tients received organ transplantationsfrom �13,000 donors. However, thenumber of patients waiting for an organis still �98,000 (1). Based on the contin-uous increased gap between available or-gans for transplantation and patientswaiting on the list for organ transplanta-tion, in �70,000 patients (1), differentstrategies have been developed to in-crease organ transplantation. The appli-cation of “expanded donor” criteria, theuse of organs that in the past were notconsidered for transplantation, has beensuccessful for this purpose. Expanded do-nor criteria include use of donors of olderage, donors who were exposed to hepatitisB and C viruses, grafts infiltrated with fat,non-heart-beating donors, donors with hy-pernatremia, or donors with multiple co-morbidities (2). Another approach used toaddress the increased demand in organtransplantation has been the introductionof live donor transplantation, which in2001 exceeded transplantation from cadav-eric donors (1). However, live donor trans-plantation is not applicable to every solidorgan, and not every program has the re-quired expertise and infrastructure to runthis type of program (3). Living donortransplant poses complex ethical chal-lenges in part due to widespread altru-istic donation in the United States anddue to the monetary exchange known totake place in countries different fromthe United States (4).

In this issue of Critical Care Medicine,Dr. Murugan and colleagues (5) describethe correlation between cadaveric donorplasma level of interleukin (IL)-6 and the

recipient’s outcome after transplantation.It is known that brain death is accompa-nied by a “cytokine storm,” which inturns creates donor instability and organdamage (6–8); however, this is the firststudy that correlates the plasma level of acytokine in the donor to the recipient’s6-month hospital-free survival. The authorsdemonstrated with an elegant method thathigher levels of IL-6 in the donor are asso-ciated with a worse outcome in the recipi-ent. This article introduces a novel andsomewhat substantiated approach to pre-dict the recipient’s outcome; nevertheless,the study raises many questions. First, it isdifficult to introduce the measurement ofcytokine plasma level on a routine basis asone of the screening tests to be performedon every cadaveric donor. Time and costare two important factors that drive deci-sions of organ procurement organizationsaround the country, and a future feasibilitystudy might help in determining if the ben-efit of this data can overcome the burden oftime and cost involved with the acquisi-tion of this information. Second, there isnot an accepted standardized protocol forthe maintenance of cadaveric donors inthe United States. Therefore, what theauthors observed in their cohort of do-nors and recipients might not reflect theimplications of IL-6 plasma levels in therest of the country. Third, there is con-flicting evidence in the medical literatureabout the effect of IL-6 on different typesof cells. As an example, Bin Gao (9) de-scribed a protective effect of this cytokineon hepatocytes. Finally, based on the ev-idence that corticosteroids decreaseplasma levels of IL-6 (10), it might havebeen interesting to know the outcomes ofthe recipients of the eight donors whoreceived corticosteroids.

Recently only a few articles about trans-plantation have caught my interest asmuch as this one. The results of this re-search, although difficult to apply clini-cally, open up a new way of assessing out-comes in organ transplantation, and whilewe are not at the point where organs will bedeclined from donors solely based on ele-

vated IL-6 level, this may become an im-portant factor in evaluating donors, espe-cially in a time when expanded donors areincreasingly evaluated and used in trans-plantation.

Cataldo Doria, MD, PhD, FACSJefferson Medical CollegeThomas Jefferson University

HospitalDepartment of SurgeryDivision of TransplantationPhiladelphia, PA

REFERENCES

1. United Network for Organ Sharing Web site.Available at: http://www.unos.org. AccessedDecember 31, 2007

2. Cameron A, Busuttil RW: AASLD/ILTS trans-plant course: Is there an extended donorsuitable for everyone? Liver Transpl 2005;11(Suppl 1):S2–S5

3. Berg CL, Gillespie BW, Merion RM, et al:Improvement in survival associated withadult-to-adult living donor liver transplanta-tion. Gastroenterology 2007; 133:1806–1813

4. Aulisio MP, DeVita M, Luebke D: Taking valuesseriously: Ethical challenges in organ donationand transplantation for critical care professionals.Crit Care Med 2007; 35:S95–S101

5. Murugan R, Venkataraman R, Wahed AS,et al: Increased plasma interleukin-6 in do-nors is associated with lower recipient hos-pital-free survival after cadaveric organtransplantation. Crit Care Med 2008; 36:1810–1816

6. Nijboer WN, Schuurs TA, van der HoevenJAB, et al: Effect of brain death on stress andinflammatory response in the human donorkidney. Transplant Proc 2005; 37:367–369

7. Avlonitis VS, Wigfield CH, Golledge HDR,et al: Early hemodynamic injury during do-nor brain death determines the severity ofprimary dysfunction after lung transplanta-tion. Am J Transplant 2006; 7:83–90

8. Weiss S, Kotsch K, Francuski M, et al: Braindeath activates donor organs and is associatedwith a worse I/R injury after liver transplantation.Am J Transplant 2007; 7:1584–1593

9. Gao B: Therapeutic potential of interleukin-6 inpreventing obesity- and alcohol-associated fattyliver transplant failure. Alcohol 2004; 34:59–65

10. Kuecuek O, Mantouvalou L, Klemz R, et al:Significant reduction of proinflammatory cy-tokines by treatment of the brain-dead do-nor. Transplant Proc 2005; 37:387–388

*See also p. 1810.Key Words: cytokine; solid organ transplantation;

outcome; cadaveric organ donation; interleukin-6The author has not disclosed any potential con-



DOI: 10.1097/CCM.0b013e318176a974


Epidemiology on intra-abdominal hypertension: An urgent call formulticenter trials*

I n this issue of Critical Care Med-icine, Dr. Vidal and colleagues (1)report the results of a study enti-tled “Incidence and Clinical Ef-

fects of Intra-abdominal Hypertension inCritically Ill Patients.” On first sight, itlooks as if this study just repeats thepreviously performed multicenter studieson prevalence and incidence of intra-abdominal hypertension (IAH) and ab-dominal compartment syndrome (ACS)(2, 3). Is this really the case? Dr. Vidal andcolleagues (1) collected prospective dataon 83 critically ill patients in a sample of153. They found that IAH (sustained in-crease in intra-abdominal pressure [IAP]of �12 mm Hg) was present at admissionin 23–31% of cases (depending onwhether mean or maximal IAP valueswere used), whereas 12% developed ACS.Primary IAH was seen in 32% of patientswith IAH. Patients with IAH were sicker,had more organ failure, and a higher hos-pital mortality (53% vs. 27%) than pa-tients with a normal IAP. Risk factors forIAH were fluid resuscitation, acidosis, hy-potension, ileus, mechanical ventilation,and acute respiratory distress syndrome.The most important message is that incritically ill patients with capillary leak orsepsis, futile resuscitation may lead topositive daily and cumulative fluid bal-ance, IAH, and subsequent morbidity (or-gan failure) and mortality. Therefore, in-tensive care unit (ICU) physicians nolonger have an excuse not to obtain abaseline IAP value in patients with two ormore risk factors for IAH (4, 5).

This epidemiologic study has a lot ofsimilarities with previous multicenterstudies, in which multiple logistic regres-sion analysis showed that the occurrenceof IAH during the first week of ICU stay isan independent outcome predictor,whereas the mean IAP at admission isnot. In the present study, Dr. Vidal andcolleagues (1) showed that maximal IAPwas also identified as an independent pre-dictor of mortality. This is a new finding,as is the association of IAH with the Se-quential Organ Failure Assessment score,and more particular, renal and respira-tory subscores, abdominal perfusion pres-sure (APP � mean arterial pressure �IAP), and the filtration gradient (FG �mean arterial pressure � 2 � IAP). Theauthors are the first to have successfullyincorporated the new consensus defini-tions recently published by the World So-ciety of the Abdominal CompartmentSyndrome (WSACS, www.wsacs.org),confirming the results from previousstudies (4, 6, 7). They also looked at dif-ferences in the prognostic ability of IAP-mean vs. IAPmax and performed theproper analysis identifying thresholds forIAP and APP with receiver operatingcharacteristic curves. The authors foundno significant difference between meanor maximal IAP at admission (days 1 and2) between survivors and nonsurvivors,whereas APP was significantly lower innonsurvivors already from day 2 onwards,suggesting that APP could be a betteroutcome predictor or resuscitation targetthan IAP. This is an interesting point thatcould have been stressed more because ithas only been studied retrospectively sofar (8). Another good point is that prog-nostic markers were not based on singleIAP measurements but on the mean oftwo IAP measurements with 3-min inter-vals, so as to rule out possible measure-ment outliers (7). The study also tacklesthe issue of the ideal frequency of IAPmonitoring. The article stresses again theimportance of measuring IAP in the ICUon a daily basis and the implications IAHand ACS can have on mortality.

To play the devil’s advocate, one couldargue that the questions that Dr. Vidaland colleagues (1) tried to answer havealready been addressed. From the intro-duction, it is unclear why yet anothersimilar study is needed. Patients with IAHare sicker than those without, and this isassociated with morbidity and mortality.Although this is certainly true, it is noth-ing new: reading the conclusions aboutthe relationship between IAH/ACS andmorbidity and mortality brings to mind anumber of similar articles that havereached the same conclusions again andagain. Open questions related to riskstratification, early diagnosis, and inno-vative preventive or treatment methodswere not addressed in this article.

Does this study really add to the grow-ing body of evidence showing that in-creased IAP has a negative effect on organfunction and patient outcome? This ques-tion is difficult to answer because data ononly 83 patients can hardly be extrapo-lated to the general ICU population, es-pecially because 60 patients (or 39.2%) ofthe initial sample were excluded from thefinal analysis. The authors conducted aprospective 9-month study but only in-cluded 83 patients? How many beds hasthe ICU under study? What was the bodymass index in the study population? Wereobese patients excluded? Was this really acohort study or a random sample? Epide-miologic conclusions can only be drawnif the collected data are complete and ofgood quality; was this the case? Further-more, the results reported are difficult tointerpret because Table 1 contains datacollected at admission and during theICU stay (64% of patients developed IAHbut only 31% presented with IAH at ad-mission). The authors could also havedone a better job on other points, as wewill demonstrate below.

Dr. Vidal and colleagues (1) used theold-fashioned Kron technique for IAPmeasurement, which is probably inferiorto other techniques. Recently, standard-ized IAP measurement kits to avoid inter-observer and intra-observer variability

*See also p. 1823.Key Words: abdominal pressure; abdominal com-

partment; epidemiology; risk factorsDr. Malbrain has received honoraria from Glaxo-

SmithKline royalties from Holtech Medical, and the 2003ESICM Chris StoutenBeek grant; he owns stock in PulsionMedical Systems, and a patent for Cimon (Pulsion Med-ical System). Dr. De Laet has not disclosed any potentialconflicts of interest.


DOI: 10.1097/CCM.0b013e318176ab25


have become available, such as the Foley-Manometer (www.holtech-medical.com),the AbViser valve (www.wolfetory.com),or even a fully automated continuoustechnique by means of a balloon-tippednasogastric tube (www.spiegelberg.de orwww.pulsion.com) (9–14). It is impor-tant that a good IAP measurement tech-nique is used for study purposes becausedefinitions of IAH and ACS (and thus theepidemiologic prevalence and incidencedata) stand or fall with the accuracy andreproducibility of the method used.

Another major drawback is that a vol-ume of 50 mL was used to prime thebladder before IAP measurement, whereasthe World Society of the Abdominal Com-partment Syndrome advocates the instil-lation of a maximal amount of 25 mL (6,7). Recent studies have shown that in-stilling even a relatively small volumeinto the bladder may increase intrinsicintravesical pressure 1 to 2 mm Hg. Witha mean IAP of around 10 mm Hg, thiswould produce a 10–20% error (15–17).The authors also used the wrong zeroreference point, namely, the symphysispubis. The consensus definitions advo-cate the use of the midaxillary line at thelevel of the iliac crest (6). This has re-cently been studied in a multicenter fash-ion (18, 19). The authors state the num-ber of patients who were mechanicallyventilated but failed to mention if theywere under sedation or curarization,which can also affect baseline IAP values.

Finally, the authors looked at the ef-fect of IAH on organ function and foundan inverse relation with renal function(as expressed by the Sequential OrganFailure Assessment renal subscore andfiltration gradient). However, it is unfor-tunate that they did not look at the acutekidney injury and RIFLE criteria fortheir study population, which wouldhave given extra and new value to theresults (20).

The data presented in this issue ofCritical Care Medicine may raise evenmore questions than they answer, andthis may also hold true for the majority ofthe available literature data available be-cause of the previous lack of standardizedIAP pressure measurement methods andgood consensus definitions.

Currently, no multicenter, random-ized, interventional, controlled clinicaltrial is available to tackle the question ofwhether an increase in IAP is a phenom-enon or an epiphenomenon and whetherany intervention to normalize IAP or APPwill eventually affect patient outcome.

Until that study exists, there will alwaysbe believers and nonbelievers (21).

The development of a management al-gorithm for IAH/ACS compares with themultifaceted approach emerging for earlygoal-directed therapy in sepsis (22). In-deed, these two syndromes overlap, andthe 2004 Surviving Sepsis Campaignguidelines call for careful fluid manage-ment to prevent IAH (23). To survive sep-sis, consensus campaigns have called forvigilance in diagnosis, careful monitoringduring goal-directed resuscitation, andmultifaceted early interventions, includ-ing antibiotics, vasopressors, inotropes,steroids, activated protein C, and tightglucose control, along with other generalICU standards of care (24–26). With thesebundles of evidence-based recommenda-tions, small supportive studies have beencompleted and large multicenter trialsare underway (27). Clinicians and re-searchers interested in IAH/ACS musttread the same path. In many ways, wehave already started down this path (21).In general, we understand who is at risk,how to monitor them, and we have nu-merous preventive and therapeutic inter-ventions for IAH/ACS (4, 7). Similar toearly goal-directed therapy in sepsis, fewof these interventions are based on level 1evidence (4). However, at least some evi-dence exists for many of these therapies,thus providing us with a start for an IAH/ACS therapeutic bundle. The urgency forprogress is real. The prevalence and inci-dence data for IAH/ACS demonstrate thatIAH and ACS are real and do exist, notonly in surgical, trauma, or burn pa-tients, but also in patients admitted to amedical ICU. The prevalence and inci-dence data show that at a minimum, IAHand sepsis go hand in hand (2, 3). It isalso not unrealistic to suggest that IAH/ACS may affect an even greater numberof ICU patients than those affected bysepsis (2).

Considerable progress has been madeover the past decade, but significant workstill needs to be done. We must study andlearn from the past and, at the same time,proactively “invent” the future. As aptlydescribed by Ivatury and Sugerman (28),IAH/ACS is “. . . a clinical entity that hadbeen ignored for far too long . . . . . . themystery of IAH and ACS continues tounfold, transgressing the boundaries ofacute and chronic illness and medical andsurgical specialities.” Recently, Sugrue (29)asked for practice guidelines, and Mal-brain (30) questioned whether it wouldbe wise not to think about IAP. The fu-

ture of patients with IAH and ACS is inour hands, and the results of the presentstudy and other recent multicenter stud-ies confirm the importance of IAH andACS on patient outcome (1–3, 31). Forthose who carry the mandate to futureIAH/ACS research, the path ahead isclear. Using available evidence, we mustdevelop an IAH/ACS therapeutic bundleand apply it in a multicenter, prospective,outcome trial.

We invite others to join the WorldSociety of the Abdominal CompartmentSyndrome, to adhere to the consensusdefinitions posted at the Web site, and tosubmit some prospective data for the nextworld congress (www.wcacs.org) to beheld in Dublin, Ireland, June 24 –27,2009.

Manu L. N. G. Malbrain, MD,PhD

Inneke De Laet, MDDepartment of Intensive CareZiekenhuisNetwerk

AntwerpenZNA StuivenbergAntwerpenBelgium

REFERENCES

1. Vidal M, Ruiz Weisser J, Gonzalez F, et al:Incidence and clinical effects of intra-abdominal hypertension in critically ill pa-tients. Crit Care Med 2008; 36:1823–1831

2. Malbrain ML, Chiumello D, Pelosi P, et al:Prevalence of intra-abdominal hypertensionin critically ill patients: A multicentre epide-miological study. Intensive Care Med 2004;30:822–829

3. Malbrain ML, Chiumello D, Pelosi P, et al:Incidence and prognosis of intra-abdominalhypertension in a mixed population of criti-cally ill patients: A multiple-center epidemi-ological study. Crit Care Med 2005; 33:315–322

4. Cheatham ML, Malbrain ML, Kirkpatrick A,et al: Results from the International Confer-ence of Experts on Intra-abdominal Hyper-tension and Abdominal Compartment Syn-drome: II. Recommendations. Intensive CareMed 2007; 33:951–962

5. Malbrain ML: You don’t have any excuse, juststart measuring abdominal pressure and actupon it! Minerva Anestesiol 2008; 74:1–2

6. Malbrain ML, Cheatham ML, Kirkpatrick A,et al: Results from the International Confer-ence of Experts on Intra-abdominal Hyper-tension and Abdominal Compartment Syn-drome: I. Definitions. Intensive Care Med2006; 32:1722–1732

7. Malbrain ML, De laet I, Cheatham M: Con-sensus conference definitions and recom-mendations on intra-abdominal hyperten-sion (IAH) and the abdominal compartment


syndrome (ACS): The long road to the finalpublications, how did we get there? Acta ClinBelg Suppl 2007; 62:44–59

8. Cheatham ML, White MW, Sagraves SG, et al:Abdominal perfusion pressure: A superior pa-rameter in the assessment of intra-abdomi-nal hypertension. J Trauma 2000; 49:621–626

9. De Potter TJ, Dits H, Malbrain ML: Intra- andinterobserver variability during in vitro vali-dation of two novel methods for intra-abdominal pressure monitoring. IntensiveCare Med 2005; 31:747–751

10. Balogh Z, De Waele JJ, Malbrain ML: Contin-uous intra-abdominal pressure monitoring.Acta Clin Belg Suppl 2007; 62:26–32

11. De Waele JJ, De laet I, Malbrain ML: Rationalintra-abdominal pressure monitoring: Howto do it? Acta Clin Belg Suppl 2007; 62:16–25

12. Malbrain ML, De laet I, De Waele JJ: Contin-uous intra-abdominal pressure monitoring:This is the way to go! Int J Clin Pract 2008;62:359–362

13. Malbrain ML, De Laet I, Viaene D, et al: Invitro validation of a novel method for con-tinuous intra-abdominal pressure monitor-ing. Intensive Care Med 2008; 34:740–745

14. Malbrain ML: Different techniques to mea-sure intra-abdominal pressure (IAP): Timefor a critical re-appraisal. Intensive Care Med2004; 30:357–371

15. De Waele J, Pletinckx P, Blot S, et al: Salinevolume in transvesical intra-abdominal pres-sure measurement: Enough is enough. In-tensive Care Med 2006; 32:455–459

16. Malbrain ML, Deeren DH: Effect of bladdervolume on measured intravesical pressure: Aprospective cohort study. Crit Care 2006; 10:R98

17. Chiumello D, Tallarini F, Chierichetti M,et al: The effect of different volumes andtemperatures of saline on the bladder pres-sure measurement in critically ill patients.Crit Care 2007; 11:R82

18. Cheatham ML, De Waele J, De Keulenaer B,et al: The effect of body position on intra-abdominal pressure measurement: A multi-center analysis. Acta Clin Belg Suppl 2007;62:246

19. De Waele J, Cheatham ML, De Keulenaer B,et al: The optimal zero reference transducerposition for intra-abdominal pressure mea-surement: A multicenter analysis. Acta ClinBelg Suppl 2007; 62:247

20. Bellomo R, Kellum JA, Ronco C: Definingand classifying acute renal failure: From ad-vocacy to consensus and validation of theRIFLE criteria. Intensive Care Med 2007; 33:409–413

21. Kimball EJ, Kim W, Cheatham ML, et al:Clinical awareness of intra-abdominal hyper-tension and abdominal compartment syn-drome in 2007. Acta Clin Belg Suppl 2007;62:66–73

22. Rivers E, Nguyen B, Havstad S, et al: Earlygoal-directed therapy in the treatment of se-vere sepsis and septic shock. N Engl J Med2001; 345:1368–1377

23. Dellinger RP, Carlet JM, Masur H, et al: Sur-viving Sepsis Campaign guidelines for man-

agement of severe sepsis and septic shock.Intensive Care Med 2004; 30:536–555

24. Annane D, Sebille V, Charpentier C, et al:Effect of treatment with low doses of hydro-cortisone and fludrocortisone on mortality inpatients with septic shock. JAMA 2002; 288:862–871

25. van den Berghe G, Wouters P, Weekers F,et al: Intensive insulin therapy in the criti-cally ill patients. N Engl J Med 2001; 345:1359–1367

26. Bernard GR, Vincent JL, Laterre PF, et al:Efficacy and safety of recombinant humanactivated protein C for severe sepsis. N EnglJ Med 2001; 344:699–709

27. Otero RM, Nguyen HB, Huang DT, et al:Early goal-directed therapy in severe sepsisand septic shock revisited: Concepts, contro-versies, and contemporary findings. Chest2006; 130:1579–1595

28. Ivatury RR, Sugerman HJ: Abdominal com-partment syndrome: A century later, isn’t ittime to pay attention? Crit Care Med 2000;28:2137–2138

29. Sugrue M: Intra-abdominal pressure: Timefor clinical practice guidelines? IntensiveCare Med 2002; 28:389–391

30. Malbrain ML: Is it wise not to think aboutintra-abdominal hypertension in the ICU?Curr Opin Crit Care 2004; 10:132–145

31. Malbrain ML: Incidence of intra-abdominalhypertension in the intensive care unit: Forthe Critically Ill and Abdominal Hyperten-sion (CIAH) Study Group. Crit Care Med2005; 33:2150–2153

The ups and downs of a good idea: Phased chest and abdominalcompression–decompression cardiopulmonary resuscitation incardiac arrest*

Approximately 166,000 out-of-hospital cardiac arrests occurannually in North America,with a median reported sur-

vival to discharge of 6.4% (1, 2). Despiteadvances in the delivery of care and in-novations in cardiopulmonary resuscita-tion (CPR), prognosis remains poor. Im-

proving outcome from sudden cardiacarrest is a healthcare urgency, but clini-cal human research is very difficult sec-ondary to ethical concerns over informedconsent and poor funding (3). Innova-tions to improve vital organ blood flowwere introduced into the 2005 AmericanHeart Association CPR guidelines, whichattempted to maximize coronary and ce-rebral perfusion pressures and minimizetheir loss from excessive ventilation dur-ing the hemodynamic phase of cardiacarrest (4, 5). Nevertheless, disappoint-ment with outcomes from conventionalCPR has prompted researchers to inves-tigate alternative methods to improve vi-tal organ blood flow and postresuscita-tion survival.

Active phased compression– decom-pression resuscitation (APCDR) is an al-ternative to standard CPR that traces itsorigins from computer-based analogmodeling designed to optimize coronaryand cerebral perfusion (6). It is per-formed by a hand-held seesaw-like devicethat delivers coincident positive intratho-racic pressure and abdominal decompres-sion, followed by negative intrathoracicpressure and abdominal compression.During the chest compression phase, thepositive intrathoracic pressure is thoughtto augment cerebral perfusion pressurewith coincident decompression of the ab-domen, which reduces ventricular after-load. Active chest decompression with ab-dominal compression increases venous

*See also p. 1832.Key Words: cardiopulmonary resuscitation; cardiac

arrest; active compression–decompression resuscita-tion; death; sudden



DOI: 10.1097/CCM.0b013e318176ad02


return and augments coronary perfusionpressure (7). Adequate coronary and ce-rebral perfusion pressures after pro-longed cardiac arrest but before defibril-lation are the main determinants ofsurvivability and neurologic outcome (8).

Early animal studies with APCDR sug-gested improved survival with increasedcoronary and cerebral perfusion pres-sures when compared with standard CPR.Tang et al. (9) induced ventricular fibril-lation in a porcine model of cardiac ar-rest. Animals were randomized to eitherLifestick (Datascope Corporation, Fair-field, NJ) APCDR or standard CPR. Sig-nificant improvements in coronary perfu-sion pressure, aortic pressure, end-tidalCO2, 48-hr survival, and neurologic re-covery over standard CPR were demon-strated. Wenzel et al. (10) performed aconfirmatory study of the hemodynamiceffects of Lifestick APCDR relative tostandard CPR, also in the porcine model.Lifestick APCDR animals again demon-strated improved coronary and cerebralperfusion, mean arterial pressure, andend-tidal CO2 compared with standardCPR. More significantly, Wenzel et al.(10) were able show that the survival andneurologic improvements from earlieranimal studies were associated with im-proved coronary and cerebral perfusionpressures compared with standard CPR.

Arntz et al. (11) performed a prospec-tive, blinded feasibility and safety study of50 patients randomized to LifestickAPCDR (n � 24) or conventional CPR(n � 26) in patients with nontraumaticout-of-hospital cardiac arrest. Althoughno improvement in outcome was demon-strated, Lifestick APCDR demonstratedno significant chest or abdominal com-plications. Rescue personnel consideredapplication of the Lifestick device practi-cal and relatively easy to fixate to thechest. Notably absent was any objectivehemodynamic assessment of Lifestick orconventional CPR. Overall, Lifestick ap-plication seemed feasible and safe, butthe small sample size and variability inpresenting rhythms between the groupslimited any assessment of hemodynamicefficacy.

In the current issue of Critical CareMedicine, Dr. Havel and colleagues (12)examine the safety, feasibility, and hemo-dynamic consequences of LifestickAPCDR compared with mechanical chestcompression (Thumper device, Michigan

Instruments, Grand Rapids, MI) in a pro-spective, case-control, clinical study ofprolonged cardiac arrest.

Caveats that are appropriately notedby the authors at the outset are that theexperimental and control groups werenot randomized, physicians were notblinded to the interventions, and thesmall sample sizes (Lifestick, 20;Thumper, 11) preclude any definitivecommentary on efficacy or safety. Never-theless, the article by Dr. Havel and col-leagues (12) provides interesting andvaluable commentary on the hemody-namic and echocardiographic changes as-sociated with APCDR in humans.

In five patients, direct measurementsrevealed increased coronary perfusionpressures and, in 15 patients, increases inend-tidal CO2 during Lifestick APCDRrelative to mechanical precordial com-pression, supportive of preserved andpossibly improved coronary perfusionpressures and cardiac output.

Transesophageal echocardiography wasperformed in a limited number of Lifestickand Thumper patients, revealing that likelyboth thoracic and cardiac pump mecha-nisms are mutually operative to maintainblood flow during CPR (13).

At this time, Lifestick resuscitationcannot be recommended in cardiac arrestbecause of the lack of any objective ben-efit and limited human studies. Neverthe-less, Dr. Havel and colleagues (12) shouldbe congratulated on moving forward anadjunctive resuscitation technique thatremains promising. Future studiesshould be directed to further clarify thehemodynamic consequences of Lifestickresuscitation earlier in cardiac arrest,with higher compression and lower ven-tilation ratios to be consistent with cur-rent American Heart Association CPRguidelines.

Edward M. Omron, MD, MPH,FCCP

Pulmonary and Critical CareService

Ingham Regional MedicalCenter

Lansing, MI

REFERENCES

1. Rea TD, Eisenberg MS, Sinibaldi G, et al:Incidence of EMS-treated out of hospital car-diac arrest in the United States. Resuscita-tion 2004; 63:17–24

2. Nichol G, Stiell IG, Laupacis A, et al: A

cumulative meta-analysis of the effective-ness of defibrillator-capable emergencymedical services for victims of out-of-hospital cardiac arrest. Ann Emerg Med1999; 34:517–525

3. Halperin H, Paradis N, Mosesso V, et al: Rec-ommendations for implementation of com-munity consultation and public disclosureunder the Food and Drug Administration’s“Exception From Informed Consent Require-ments for Emergency Research”: A specialreport from the American Heart AssociationEmergency Cardiovascular Committee andCouncil on Cardiopulmonary, Perioperativeand Critical Care. Endorsed by the AmericanCollege of Emergency Physicians and the So-ciety for Academic Emergency Medicine. Cir-culation 2007; 116:1855–1863

4. 2005 American Heart Association guidelinesfor cardiopulmonary resuscitation and emer-gency cardiovascular care. Circulation 2005;112(Suppl 1)IV-1–IV-211

5. Ewy GA: Cardiocerebral resuscitation: Thenew cardiopulmonary resuscitation. Circula-tion 2005; 111:2134–2142

6. Babbs CF, Weaver C, Ralston S, et al: Cardiac,thoracic, and abdominal pump mechanismsin cardiopulmonary resuscitation: Studies inan electrical model of the circulation. Am JEmerg Med 1984; 2:299–308

7. Babbs CF: CPR techniques that combine chestand abdominal compression and decompres-sion: Hemodynamic insights from a spread-sheet model. Circulation 1999; 100:2146–2152

8. Ewy GA, Kern KB, Sanders AB, et al: Cardio-cerebral resuscitation for cardiac arrest.Am J Med 2006; 119:6–9

9. Tang W, Weil MH, Schock RB, et al: Phasedchest and abdominal compression-decom-pression: A new option for cardiopulmonaryresuscitation. Circulation 1997; 95:1335–1340

10. Wenzel V, Lindner KH, Prengel AW, et al:Effect of phased chest and abdominal com-pression-decompression cardiopulmonaryresuscitation on myocardial and cerebralblood flow in pigs. Crit Care Med 2000; 28:1107–1112

11. Arntz HR, Agrawal R, Richter H, et al: Phasedchest and abdominal compression-decom-pression versus conventional cardiopulmo-nary resuscitation in out-of-hospital cardiacarrest. Circulation 2001; 104:768–772

12. Havel C, Berzlanovich A, Sterz F, et al:Safety, feasibility, and hemodynamic andblood flow effects of active compression–decompression of thorax and abdomen inpatients with cardiac arrest. Crit Care Med2008; 36:1832–1837

13. Andreka P, Frenneaux MP: Haemodynamicsof cardiac arrest and resuscitation. Curr OpinCrit Care 2006; 12:198–203


Gender differences after hemorrhagic shock from blunt trauma:What is helping the women?*

A great debate has been surfac-ing in intensive care units inrecent years. At stake in thisdebate is the validity of the no-

tion that gender-differentiated outcomesare unequal for similar disease entities.The majority of investigators interestedin this topic seem to think that womenappear more physiologically positioned towithstand a septic challenge. This stanceis more clearly defensible for female ratsthan female humans, as human studieshave yielded conflicting gender-stratifiedresults, depending on the population anddisease studied (1).

A less-debated aspect of the argumentis the mechanism by which women maytheoretically gain their advantage overtheir male counterparts in the aforesaidseptic challenge. The exact mechanismfor the protection is unclear, but thestance by the pro-female advocates has al-ways been initiated by the most fundamen-tal of the difference between the sexes: thefemale sex hormones. They reason that be-cause female sex hormones in the studiedpopulations are the only discernible differ-ences among the groups, these entitiesmust contain the elixir that staves off thewoes of sepsis. Which sex hormones is lessclear, but what else could women have insufficient quantities over men to create aphysiologic difference in a given clinicalsetting?

The recent study by Dr. Sperry andcolleagues (2) in this issue of CriticalCare Medicine gives some much neededsupport for the first debate of whetherany gender-differentiated outcomes be-tween men and women exist (they do).However, in attempting to authenticatethe purported superiority of their femalepatients, the authors have also thrownthe proverbial monkey wrench into thebasis for this advantage. These investiga-tors found that adult female patients aftersevere blunt trauma complicated by hem-orrhagic shock fared better in regard tomultiple organ failure and nosocomial in-fection rates. However, these advantagesoccurred in both premenopausal andpostmenopausal women, indicating thatthe effects of sex hormones were unlikelyto be a valid mechanism.

On further analysis of these intriguingresults, several observations require ru-mination. First, this retrospective study,gleaned from the data of The Inflamma-tion and Host Response to Injury grant,was not designed to solve the gender-protection question. Next, the reader seesthat questions arise concerning the base-line equality between the male and fe-male groups. The Injury Severity Scoresfor men and women are identical, but theinitial resuscitation and initial AcutePhysiology and Chronic Health Evalua-tion II scores differ significantly betweenthe groups. The transfusion require-ments for the initial 12 hrs also differbetween the two groups, but not signifi-cantly, and the authors did not reporttheir 24- or 48-hr transfusion require-ments. In addition, the men in this studydid have a higher rate of alcoholism com-

pared with their female counterparts.Most importantly, however, the need forinitial laparotomy and thoracotomy diddiffer between the two groups, implyingthat the men were more severely in-jured than their female counterparts,even if this feature was not demon-strated by the Injury Severity Scores.

Whether these aforementioned ob-servations indeed explain the subse-quent differences in nosocomial infec-tion or organ dysfunction rates in menover women is a matter for conjecture.Plausibility and common sense wouldclaim that the summation of thesebaseline differences, rather than sexhormones or unidentified elusive female-specific factors, would bear the burdenfor creating the adverse secondary out-comes for the male patients reportedherein. Nonetheless, in this retrospec-tive blunt trauma study evaluating thehost response to hemorrhagic shock,Dr. Sperry and colleagues (2) have con-tributed more valuable data for all sidesof the ignited outcome debates regard-ing gender and sepsis.

Soumitra R. Eachempati, MDDepartment of SurgeryNew York Presbyterian HospitalNew York, NY

REFERENCES

1. Eachempati SR, Hydo L, Barie P: Gender-based differences in outcome for septic pa-tients admitted to the surgical intensive careunit. Arch Surg 1999; 134:1342–1347

2. Sperry JL, Nathens AB, Frankel HL, et al:Characterization of the gender dimorphismafter injury and hemorrhagic shock: Are hor-monal differences responsible? Crit Care Med2008; 36:1838–1845

*See also p. 1838.The author has not disclosed any potential con-



DOI: 10.1097/CCM.0b013e318176ae01


Critical illness myopathy: Deeper insights*

Acutely acquired neuromuscu-lar dysfunction occurs in�50% of intensive care unitpatients with sepsis for a dura-

tion of �1 wk (1). Often, this involvesneuropathy and myopathy together, butthese entities can occur independently inmany patients (2, 3). The pathogenesis ofcritical illness neuropathy and myopathyis not known, but recent research is pro-viding promising insights.

In this issue of Critical Care Medicine,Dr. Rossignol and colleagues (4) reporttheir study of skeletal muscle taken fromseptic rats (cecal perforation model) inwhich they found reduced force of con-traction but faster contraction and relax-ation times compared with controls. Fur-thermore, they found that the musclefatigued more quickly, despite reachingmaximal tetanic force more quickly. In ex-amining the molecular mechanisms under-lying these phenomena, they discovered anincrease in the ryanodine receptor expres-sion, which related to increased messengerRNA activity responsible for its synthesis.As a separate issue, they observed increasedsensitivity to the neuromuscular blockingagent atracurium. What does it all mean?

The increased sensitivity to neuro-muscular blockade suggests a problemwith neuromuscular transmission (in thesense of reduced acetylcholine fromnerve terminals or a decrease in receptorsat the endplate), but this has never beendemonstrated in critical illness neuromy-

opathy. Indeed, the problem seems to bedissociable from neuropathy. It couldwell relate to the diminished activity ofsodium channels, including the endplateregion, which Dr. Rossignol and col-leagues have also described (5).

The most intriguing aspect of this ar-ticle is the potential insight into alteredcytoplasmic ionized calcium homeostasisin skeletal muscle, which may play a rolein some of the myopathies of critical ill-ness (6, 7). To activate muscle contrac-tion, ionized calcium must be releasedfrom the sarcoplasmic reticulum withinthe muscle cells. Sarcoplasmic reticulumcalcium release (ion) channels are gov-erned by the large macromolecular com-plex known as the ryanodine receptor(RyR1). The activity of the receptor, andtherefore, the release of ionized calciuminto the cytoplasm, is affected by numer-ous factors that phosphorylate its regula-tory subunits or nitrosylate-free sulfhy-dryl groups on its cysteine residues (8).These include calmodulin, calstabin-1,and others that can, in turn, be influ-enced by various “stressors,” includingcatecholamines, ischemia, acidosis, andpossibly, inflammatory mediators. Exces-sive cytoplasmic calcium can cause muchmischief within the muscle cell, includ-ing oxidation of myofibrillar proteins (9).This might explain the observations ofDr. Rossignol and colleagues (4). Thelarge muscle slow-twitch fibers arehigher in oxidative reserves than the type2, rapid-twitch fibers; selective dysfunc-tion or loss of the larger, slow-twitchfibers might then allow for the morerapid onset, but less tetanic strength, ofcontraction that they observed.

Further research is clearly needed tosubstantiate these findings and to explorethe mechanisms in more detail. Dr.Rossignol and colleagues (4) are to be

commended for their continued pursuitof muscular weakness in critically ill/sepsis models.

G. Bryan Young, MD, FRCPCDepartment of Clinical

Neurological SciencesUniversity HospitalThe University of Western

OntarioLondon, OntarioCanada

REFERENCES

1. Khan J, Harrison TB, Rich MM, et al: Earlydevelopment of critical illness myopathy andneuropathy in patients with severe sepsis.Neurology 2006; 67:1421–1425

2. Visser LH: Critical illness polyneuropathy andmyopathy: Clinical features, risk factors andprognosis. Eur J Neurol 2006; 13:1203–1212

3. Bolton CF: Neuromuscular manifestations ofcritical illness. Muscle Nerve 2005; 32:140–163

4. Rossignol B, Gueret G, Pennec JP, et al: Ef-fects of chronic sepsis on contractile proper-ties of fast-twitch muscle in an experimentalmodel of critical illness neuromyopathy in therat. Crit Care Med 2008; 36:1855–1863

5. Rossignol B, Gueret G, Pennec JP, et al: Ef-fects of chronic sepsis on the voltage-gatedsodium channel in isolated rat muscle fibers.Crit Care Med 2007; 35:351–357

6. Larsson L: Experimental animal models ofmuscle wasting in intensive care unit patients.Crit Care Med 2007; 35(9 Suppl):S484–S487

7. Schweickert WD, Hall J: ICU-acquired weak-ness. Chest 2007; 131:1273–1274

8. Bellinger AM, Mongillo M, Marks AR: Stressedout: The skeletal muscle ryanodine receptor asa target of stress. J Clin Invest 2008; 118:445–453

9. Dalla Libera L, Ravara B, Gobbo V, et al: Skel-etal muscle myofibrillar protein oxidation inheart failure and the protective effect of Carve-dilol. J Mol Cell Cardiol 2005; 38:803–807

*See also p. 1855.Key Words: sepsis; myopathy; critical illness;

pathogenesisThe author has not disclosed any potential con-



DOI: 10.1097/CCM.0b013e318176aa87


A treatment worth its salt?*

A new era in the treatment ofshock began with the 1980 re-port that hypertonic (7.5%) sa-line (HS) could resuscitate

dogs in hemorrhagic shock (1). Thisstudy sparked numerous investigationsinto the benefits of HS either alone orwith dextran or hetastarch to resuscitateexperimental animals from hemorrhagichypotension (cited in Ref. 2) and led toseveral clinical trials in patients withtraumatic hypotension (3–5). Other stud-ies investigated the efficacy of these fluidsin traumatic brain injury or intracranialhemorrhage (5, 6). These studies pro-vided solid evidence for the physiologiceffects of hypertonic fluids on expandingplasma volume and improving cardiacoutput, regional blood flows, and the mi-crocirculation.

Sprinkled among the earlier studieswere a few from the 1980s and early1990s that investigated whether the he-modynamic effects of HS would be usefulas supportive therapy in the treatment ofsepsis in experimental animals (7–9).However, in the past decade recognitionof possible immunomodulating effects ofHS has opened new possibilities that thebenefits of HS could extend beyond itshemodynamic effects. Hypertonicity wasshown to inhibit leukocyte adherence andactivation. Subsequent to improved mi-crocirculatory flow, use of hypertonic flu-ids was associated with a reduction ofneutrophils rolling and sticking to theendothelial cells of blood vessels (10–12).In addition, in isolated neutrophils fromhealthy human volunteers, it was ob-served that HS alone or with dextran in-hibited their respiratory burst and de-creased �2-integrin expression, superoxideproduction, and elastase release, but onlyif HS was added before neutrophils be-came primed or activated (10–14). How-

ever, if HS was added after neutrophilpriming or activation, superoxide pro-duction and elastase release were actuallyenhanced (13). Immunomodulatory ef-fects of HS have since been investigatedin animal models of hemorrhage. Pascualet al. (12) observed that infusion of HSreduced neutrophil adherence to pulmo-nary endothelium and reduced lung my-eloperoxidase activity compared with in-fusion of lactated Ringer’s solution.These authors also reported that the sus-ceptibility to sepsis after hemorrhage wasdiminished, possibly due to inhibition ofsepsis-induced P-selectin expression or tobacterial challenge. Taken together, thesedata suggested that modulation of im-mune function by HS could possibly re-duce secondary complications of infec-tion and that initial or early hypertonicsaline resuscitation would be of greaterbenefit than infusion after administrationof conventional fluids. A similar conclu-sion that hypertonic fluids should be theinitial resuscitation fluid was also drawnfrom the hemorrhage studies in experi-mental animals (2).

In this issue of Critical Care Medicine,Dr. Shih and colleagues (15) present awell-designed, comprehensive investiga-tion of the use of HS vs. normal saline ina peritonitis-induced (cecal ligation andpuncture [CLP]) septic shock model inrats. HS was infused 3 hrs after CLP. Theinvestigators monitored standard hemo-dynamics and the systemic pressor re-sponse to norepinephrine over an 18-hrperiod. In addition, they measured bloodglucose, indices of hepatic and renalfunction, lactate dehydrogenase as a gen-eral index of cellular injury, and plasmainterleukin-1� and nitric oxide levels.Upon euthanasia, thoracic aorta, lung,liver, and kidney were assayed for super-oxide levels and lung and liver for expres-sion of inducible nitric oxide synthase.Histology was performed on lung, liver,kidney, and aorta for evidence of neutro-phil infiltration, and survival rates at 9and 18 hrs were quantified. The investi-gators observed that HS infusion main-tained mean arterial pressure in CLP ratswith no effect on heart rate. In addition,HS infusion in these rats improved sys-

temic vascular reactivity and indices oforgan and general cellular injury. Mech-anistically, the investigators reportedthat HS attenuated the elevated levels ofseveral mediators in both plasma (inter-leukin-1� and nitric oxide) and tissues(superoxide). This translated into re-duced neutrophil infiltration in lung andliver and 27% and 47% higher survivalrates at 9 hrs and 18 hr, respectively, inCLP rats that received HS compared withnormal saline. This study shows that inan animal model that may be the closestmimic available to human septic shock, 4mL/kg HS was able to prevent hypoten-sion, reduce organ dysfunction, and im-prove survival, possibly through its actionson hemodynamics and the inflammatoryresponse.

Although this study has limitations re-garding the clinical significance of thelevel of improvement seen with HS overnormal saline, as well as the relevance ofthe results to human sepsis and septicshock, it does expand the therapeutic po-tential of HS in this field. Currently, amulticenter trauma trial sponsored bythe U.S. National Institutes of Health andthe U.S. Army is underway evaluating HSalone or with dextran compared with nor-mal saline. Some of the enrolling centersare investigating the potential benefit ofthese hypertonic fluids on immune func-tion and their ability to reduce multior-gan failure. The results of these studiesare highly anticipated. Perhaps this studyby Dr. Shih and colleagues (15) will en-courage the planning of similar clinicalstudies to evaluate HS as the initial re-suscitation fluid in the therapy of sepsisand septic shock, beyond its typical phys-iologic effects.

Michael A. Dubick, PhDU.S. Army Institute of

Surgical ResearchSan Antonio, TX

REFERENCES

1. Velasco IT, Pontieri V, Rocha e Silva M Jr,et al: Hyperosmotic NaCl and severe hemor-rhagic shock. Am J Physiol 1980; 239:H664–H673

2. Dubick MA, Bruttig SP, Wade CE: Issues ofconcern regarding the use of hypertonic/

*See also p. 1864.Key Words: hypertonic saline; sepsis; hemorrhage;

rats; inflammatory responseThe author has not disclosed any potential con-



DOI: 10.1097/CCM.0b013e318176aabe


hyperoncotic fluid resuscitation of hemor-rhagic hypotension. Shock 2006; 25:321–328

3. Wade CE, Kramer GC, Grady JJ, et al: Effi-cacy of hypertonic 7.5% saline and 6% dex-tran-70 in treating trauma: A meta-analysisof controlled clinical studies. Surgery 1997;122:609–616

4. Olsson JSC: The use of hypertonic solutions inprehospital care in Scandinavia. ScandJ Trauma Resusc Emerg Med 2004; 12:78–85

5. Wade CE, Grady JJ, Kramer GC, et al: Indi-vidual patient cohort analysis of the efficacyof hypertonic saline/dextran in patients withtraumatic brain injury and hypotension.J Trauma 1997; 42(5 Suppl):S61–S65

6. Himmelseher S: Hypertonic saline solutionsfor treatment of intracranial hypertension.Curr Opin Anaesthesiol 2007; 20:414–426

7. Luypaert P, Vincent JL, Domb M, et al: Fluid

resuscitation with hypertonic saline in endo-toxic shock. Circ Shock 1986; 20:311–320

8. Kreimeier U, Frey L, Dentz J, et al: Hyper-tonic saline dextran resuscitation during theinitial phase of acute endotoxemia: Effect onregional blood flow. Crit Care Med 1991;19:801–809

9. Horton JW, Walker PB: Small-volume hyper-tonic saline dextran resuscitation from canineendotoxin shock. Ann Surg 1991; 214:64–73

10. Angle N, Hoyt DB, Coimbra R, et al: Hypertonicsaline resuscitation diminishes lung injury bysuppressing neutrophil activation after hemor-rhagic shock. Shock 1998; 9:164–170

11. Pascual JL, Ferri LE, Seely AJ, et al: Hypertonicsaline resuscitation of hemorrhagic shock di-minishes neutrophil rolling and adherence toendothelium and reduces in vivo vascular leak-age. Ann Surg 2002; 236:634–642

12. Pascual JL, Khwaja KA, Ferri LE, et al: Hy-pertonic saline resuscitation attenuates neu-trophil lung sequestration and transmigra-tion by diminishing leukocyte-endothelialinteractions in a two-hit model of hemor-rhagic shock and infection. J Trauma 2003;54:121–130

13. Ciesla DJ, Moore EE, Zallen G, et al: Hyper-tonic saline attenuation of polymorphonu-clear neutrophil cytotoxicity: Timing is ev-erything. J Trauma 2000; 48:388–395

14. Junger WG, Liu FC, Loomis WH, et al: Hy-pertonic saline enhances cellular immunefunction. Circ Shock 1994; 42:190–196

15. Shih CC, Chen SJ, Chen A, et al: Therapeuticeffects of hypertonic saline on peritonitis-induced septic shock with multiple organdysfunction syndrome in rats. Crit Care Med2008; 36:1864–1872

Levo is in the air: Take a deep breath!*

Mrs. Fictive-Fitness is a 40-yr-old “healthy” womanwho underwent a laparo-scopic cholecystectomy

that was complicated by iatrogenic intes-tinal perforation, necessitating a 10-cmbowel resection with end-to-end anasto-mosis. On the second postoperative day,she developed septic shock and requiredhigh doses of norepinephrine to maintaina sufficient mean arterial blood pressure.Despite a hemoglobin value of 10 g/dLand appropriate intravascular volume re-placement, her central venous oxygensaturation was unacceptably low, so in-cremental doses of dobutamine were con-tinuously infused with the aim to achievethreshold values of 70%. Although thesurgical focus had been eliminated andantibiotic therapy was judged as appro-priate, septic shock progressed. Probablydue to adrenergic receptor down-regula-tion, excessive catecholamine doses wereneeded to maintain at least a minimalcirculation. In the meantime, Mrs. Fic-tive-Fitness had undergone metamorpho-sis and converted into Mrs. Ultra-Weak-

ness. As she suffered from multiple organfailure despite (or because of?) havingreceived 1 �g/kg/min norepinephrine and25 �g/kg/min dobutamine, the intensivecare physicians in charge considered ad-ministering levosimendan as drug of lastresort, hoping that a somewhat “magicbullet” could revive her. Because she hadsevere myocardial insufficiency with acardiac index of 1.5 L/min/m2 and a cen-tral venous oxygen saturation of only45%, the team decided to give a bolusdose of 24 �g/kg body weight. Some min-utes later, her mean arterial blood pres-sure dropped dramatically, and a coupleof hours thereafter she was dead. Whatwent wrong with this case?

In the last years, levosimendan hasemerged as a promising agent in themanagement of reversible cardiovasculardysfunction. Levosimendan is a calciumsensitizer that reliably increases cardiacoutput at low energetic costs. Due to si-multaneous activation of adenosinetriphosphate-sensitive potassium chan-nels, levosimendan contributes to globalvasodilatation within the systemic andpulmonary circulation (1). Whereas thedecrease in afterload may be desirable,the subsequent decrease in systemic vas-cular resistance may threaten organ per-fusion. Therefore, it is important to guar-antee both appropriate vascular fillingand tone before levosimendan infusion.Since the vasodilatory properties are dosedependent, vasodilation is most pro-nounced if a bolus is given (without pre-

ceding volume loading and coadministra-tion of a vasoconstrictor agent). But whatdo we do if the patient suffers from sys-temic inflammation, associated with pro-nounced vasodilatation in conjunctionwith pulmonary hypertension and rightheart failure? In this life-threatening con-dition, it would be smart to administer ananti-inflammatory agent that amelioratescardiopulmonary dysfunction withoutfurther deteriorating systemic hemody-namics and tissue perfusion.

In this issue of Critical Care Medicine,Dr. Boost and colleagues (2) report theresults of a timely and carefully con-ducted study investigating the role ofaerosolized levosimendan in the experi-mental setting of ventilator-induced lunginjury (VILI) in rats. The authors demon-strated that nebulization of levosimen-dan, when applied preinjury, reduced therelease of inflammatory mediators andimproved survival. The data support theconcept that prophylactic inhalation oflevosimendan may be a useful option toprevent pulmonary derangement in ven-tilated subjects.

Previous studies provided evidencethat among the factors involved in thepathogenesis of VILI, overdistension andcollapse of distal bronchioli and alveoliduring mechanical ventilation, as well asactivation and release of proinflammatorymediators, play a pivotal role. In addition,increased expression of intercellular ad-hesion molecule-1, cytokine-inducedneutrophil chemoattractant-1, and

*See also p. 1873.Key Words: levosimendan; acute lung injury;

ventilator-induced lung injury; sepsis; septic shock;hemodynamics

The authors have not disclosed any potential con-flicts of interest.


DOI: 10.1097/CCM.0b013e318176aae9


monocyte chemoattractant protein-1 isimplicated in the inflammatory orches-tration (3– 6).

Levosimendan has proven to inhibitthe expression of intercellular adhesionmolecule-1 and vascular cell adhesionmolecule-1 on the surface of endothelialcells (7) and to reduce the release of en-dothelin-1 (8). Additional anti-inflamma-tory properties of levosimendan havebeen reported during ischemia and reper-fusion, acute heart failure, and experi-mental sepsis (9–11). Pharmacologic pre-conditioning with levosimendan iscurrently known to represent a powerfulcell-protective mechanism conferringrelative resistance against cell death. Fi-nally, an improvement in myocardial per-formance associated with a decrease inpulmonary vascular resistance (1) mightalso contribute to an increase in tissueperfusion and oxygenation.

In the study by Dr. Boost and col-leagues (2), levosimendan exerted anti-inflammatory properties, as indicated bya decrease in proinflammatory cytokines(i.e., macrophage inhibitory protein-2and interleukin 1�) as well as a reductionin nitric oxide release and matrix metal-loproteinase-9 expression. The presentstudy confirms the notion that calciumsensitizing may be a promising therapeu-tic option in a condition in which theright ventricle is acutely overloaded (1)and suggests that levosimendan attenu-ates the degree of inflammation in thesetting of mechanical ventilation (2).

Another important aspect of the cur-rent study is that levosimendan has beenadministered preventively. During thelast years, it has become more and moreobvious that the efficacy of goal-directedtherapies—whether the approach focuseson balanced oxygen demand-supply rela-tionship (12), hemodynamic stabilization(13), or appropriate antimicrobial ther-apy (14)—critically depends on early in-stitution of therapy. In this regard, it isespecially noteworthy that even a power-ful therapeutic agent may result in un-changed or even increased mortality ifadministration is delayed (14). In clinicalpractice, however, levosimendan is com-monly used as last resort, when conven-tional (less expensive) inotropes havefailed to improve hemodynamics. Thismay result in clinicians’ biased impres-sion that many patients die after havingbeen treated with levosimendan. The ma-jor question that needs to be addressed iswhy we are not using levosimendan earlyin the stage of the disease. In this regard,

it may be of relevance that physiciansmore and more are forced to work aseconomically as possible. Therefore, itsometimes happens that potentiallypromising therapeutic options are with-held because of their considerable pricecompared with standard therapy, whichmay be less effective. Of course, it makesno sense to break a fly on the wheel.However, it is important to figure outwhether, when, and how potentially effec-tive compounds may be used under spe-cial circumstances. Although the resultsof the current study strengthen the hy-pothesis that levosimendan confers organprotection and reduces mortality in theexperimental setting, it needs to be takeninto consideration that levosimendan wasgiven preinjury and that no randomizedclinical trials have shown that calciumsensitizing reduces mortality in patientssuffering from VILI and right ventricularoverload.

Nevertheless, the work by Dr. Boostand colleagues (2) is of special impor-tance, since it demonstrated for the firsttime that levosimendan can be effectivelyaerosolized. Unfortunately, the study didnot determine the effects of inhaled levo-simendan on cardiac output and vascularresistance within the systemic and pul-monary circulation. Although it appearsthat this route of administration may bepreferable compared with classic intrave-nous infusion, it needs to be taken intoaccount that inhalation of 240 �g of le-vosimendan in 500-g rats corresponds to30–40 mg (approximately three vials) inadult humans. In this context, it remainsinconclusive a) which amount of the drugactually reached the alveolar and pulmo-nary microvascular compartment; b)whether innovative nebulization tech-niques may enable a more economic wayof drug distribution within the airways; c)to what extent inhaled levosimendan ex-erts (possibly adverse) systemic hemody-namic effects; and d) whether the (poten-tially increased) costs associated with thisinnovative approach contribute to a bet-ter outcome compared with conventionalinotropes or intravenous levosimendaninfusion, respectively.

When we view the published studieson this topic together with the currentwork of Dr. Boost et al. (2), it appearsreasonable to investigate the role of levo-simendan in the early stage of the disease(without a loading bolus dose, if possi-ble). Since levosimendan administrationmay contribute to arterial hypotension,invasive monitoring is desirable to detect

and treat a relevant decrease in meanarterial blood pressure. Future clinicalstudies are needed to elucidate the safety,efficacy, and costs of inhaled vs. intrave-nous levosimendan. If the experimentalresults by the current authors (2) arereplicable in humans, inhaled levosimen-dan may become an attractive approach,especially in hemodynamically unstableconditions, such as acute heart failure orsepsis-related myocardial depression (1).We hope that the near future will revealwhether our patients should take a deepbreath, when there is levo in the air.

Christian Ertmer, MDMartin Westphal, MD, PhD

Department of Anesthesiologyand Intensive Care

University Hospital ofMuenster

Muenster, GermanyAndrea Morelli, MD

Department of Anesthesiologyand Intensive Care

University of Rome“La Sapienza”Rome, Italy

REFERENCES

1. Morelli A, Teboul JL, Maggiore SM, et al:Effects of levosimendan on right ventricularafterload in patients with acute respiratorydistress syndrome: A pilot study. Crit CareMed 2006; 34:2287–2293

2. Boost KA, Hoegl S, Hofstetter C, et al: In-haled levosimendan reduces mortality andrelease of proinflammatory mediators in a ratmodel of experimental ventilator-inducedlung injury. Crit Care Med 2008; 36:1873–1879

3. Parker JC, Hernandez LA, Peevy KJ: Mecha-nisms of ventilator-induced lung injury. CritCare Med 1993; 21:131–143

4. Li LF, Yu L, Quinn DA: Ventilation-inducedneutrophil infiltration depends on c-Jun N-terminal kinase. Am J Respir Crit Care Med2004; 169:518–524

5. Schmeck J, Janzen R, Munter K, et al: Endo-thelin-1 and thromboxane A2 increase pul-monary vascular resistance in granulocyte-mediated lung injury. Crit Care Med 1998;26:1868–1874

6. Mendez MP, Morris SB, Wilcoxen S, et al:Shedding of soluble ICAM-1 into the alveolarspace in murine models of acute lung injury.Am J Physiol Lung Cell Mol Physiol 2006;290:L962–L970

7. Parissis JT, Karavidas A, Bistola V, et al: Ef-fects of levosimendan on flow-mediated va-sodilation and soluble adhesion molecules inpatients with advanced chronic heart failure.Atherosclerosis 2008;197:278–282

8. Gruhn N, Nielsen-Kudsk JE, Theilgaard S, etal: Coronary vasorelaxant effect of levosi-mendan, a new inodilator with calcium-


sensitizing properties. J Cardiovasc Pharma-col 1998; 31:741–749

9. Zager RA, Johnson AC, Lund S, et al: Levo-simendan protects against experimental en-dotoxemic acute renal failure. Am J PhysiolRenal Physiol 2006; 290:F1453–F1462

10. Lepran I, Pollesello P, Vajda S, et al: Precon-ditioning effects of levosimendan in a rabbitcardiac ischemia-reperfusion model. J Car-diovasc Pharmacol 2006; 48:148–152

11. Adamopoulos S, Parissis JT, IliodromitisEK, et al: Effects of levosimendan versusdobutamine on inflammatory and apopto-tic pathways in acutely decompensatedchronic heart failure. Am J Cardiol 2006;98:102–106

12. Rivers E, Nguyen B, Havstad S, et al: Earlygoal-directed therapy in the treatment of se-vere sepsis and septic shock. N Engl J Med2001; 345:1368–1377

13. Morelli A, Ertmer C, Lange M, et al: Contin-uous terlipressin infusion in patients withseptic shock: Less may be best, and the ear-lier the better? Intensive Care Med 2007;33:1669–1670

14. Kumar A, Roberts D, Wood KE, et al: Dura-tion of hypotension before initiation of effec-tive antimicrobial therapy is the critical de-terminant of survival in human septic shock.Crit Care Med 2006; 34:1589–1596

Calcium sensitizing in sepsis: Is levosimendan on the right path?*

Levosimendan is a triple-actioncompound exerting positiveinotropic, vasodilatory, and an-ti-ischemic effects at the same

time (1). Whereas the increase in myo-cardial contractility is primarily mediatedby calcium sensitization, global vasodila-tion and anti-ischemic properties are as-cribed to activation of adenosine triphos-phate–sensitive potassium channels(KATP) in vascular smooth muscle cellsand the mitochondria, respectively. By itsunique mechanism, levosimendan maythus increase cardiac output and improveregional perfusion simultaneously. Inthis context, previous experimental andclinical studies demonstrated that levosi-mendan attenuates right and left ventric-ular dysfunction and increases both re-gional blood flow and global oxygentransport (2–4).

In addition to stimulation of KATP

channels, levosimendan may contributeto vasodilation by blocking the release ofendothelin-1 (5), a potent vasoconstric-tive peptide involved in the pathophysiol-ogy of macro- and microvascular maldis-tribution seen in septic shock. Finally,levosimendan may beneficially affect en-dothelial dysfunction by inhibiting theexpression of soluble adhesion molecules,such as intercellular adhesion molecule-1and vascular cell adhesion molecule-1(6), thereby potentially attenuating sep-sis-related microvascular dysfunction.

Dr. Schwarte and colleagues (7) also per-formed experiments in this area and re-ported that levosimendan increases gas-trointestinal mucosal oxygenation inhealthy dogs, as determined by reflec-tance spectrophotometry. Interestingly,the latter authors demonstrated that theincrease in gastrointestinal perfusion re-sulting from levosimendan infusion wasnot completely dependent on an increasein cardiac output, but was rather causedby redistribution of blood flow toward thegastrointestinal tract due to opening ofvascular smooth muscle KATP channels(7). The fact that healthy animals werestudied, however, raises the crucial ques-tion whether or not the beneficial micro-circulatory effects caused by levosimen-dan may also be replicated in thepresence of septic shock, a condition thatis characterized by maldistribution of re-gional and microvascular blood flow.

In this issue of Critical Care Medicine,Dr. Fries and colleagues (8) report theresults of a carefully conducted study de-signed to compare the effects of norepi-nephrine and levosimendan on microvas-cular perfusion and oxygenation in a ratmodel of septic shock. Using sidestreamdarkfield imaging and oxygen-dependentquenching of phosphorescence, the au-thors showed that both norepinephrine andlevosimendan had comparable effects onthe restoration of sepsis-induced alter-ations in systemic hemodynamics (i.e., car-diac output) with no significant impact onmicrovascular perfusion. Whereas levosi-mendan significantly improved microvas-cular oxygenation, norepinephrine con-tributed to a further impairment. Despitethe restoration of cardiac output withboth vasoactive agents, neither norepi-nephrine nor levosimendan improved mi-crocirculatory blood flow. Unfortunately,

it cannot be excluded that the uncouplingbetween unaffected microvascular oxygendelivery (microvascular blood flow) andincreased microvascular oxygenation hasto be attributed to the techniquesadopted in this study. However, it appearsthat �PO2 measurements are more sen-sitive than the semiquantitative measure-ments derived from the sidestream dark-field imaging. Whether a larger samplesize would have yielded significant differ-ences in terms of microvascular bloodflow remains inconclusive.

It also may be critically discussedwhether the animals have suffered fromhypovolemia at the time of levosimendanadministration (3, 4), as suggested bymarked hemoconcentration. The findingthat norepinephrine, an alpha-adrenergicvasoconstrictor with scarce inotropic ef-fects, resulted in a similar increase incardiac output when compared with levo-simendan, a strong inodilator, is rathersurprising and in contrast to clinical ob-servations. A potential explanation maybe that hypovolemia limited the increasein cardiac output in the levosimendangroup, whereas the relatively high nor-epinephrine dose exerted positive inotro-pic effects by stimulation of beta-1 recep-tors. Unfortunately, the authors did notstudy an additional group treated withcombined levosimendan and norepineph-rine. Such an approach would have helpedstabilize mean arterial pressure and in-crease cardiac output simultaneously, andpotentially could have resulted in differentmicrocirculatory effects.

Opening of KATP channels in responseto levosimendan administration typicallyis associated with significant vasodila-tion. In the presence of an adequate vol-ume status, the drop in blood pressuremay be limited by a simultaneous in-

*See also p. 1886.Key Words: levosimendan; septic shock; norepi-

nephrineThe authors have not disclosed any potential con-



DOI: 10.1097/CCM.0b013e318176a807


crease in cardiac output. In this context,it is especially important to note that thevasodilatory effects of levosimendan aredose-dependent and most pronounced if aloading bolus is administered. In thepresence of systemic inflammationand/or vasodilatory shock states, levosi-mendan should therefore only be givenwhen adequate fluid resuscitation can beassured. In a condition in which a changein myocardial contractility is not re-quired immediately (e.g., in sepsis), itmay be reasonable to continuously infuselevosimendan without a preceding load-ing bolus dose (3, 4). Finally, we urgentlyrecommend combining levosimendanwith a vasopressor agent to circumvent a(further) decrease in perfusion pressure.

To explain the levosimendan-associ-ated increase in �PO2, Dr. Fries and col-leagues (8) focused their attention on thepossible effects on mitochondrial KATP

channels and cellular mechanisms inde-pendent of microvascular perfusion. Un-der hemodynamic stress conditions asso-ciated with microcirculatory alterations,the opening of mitochondrial KATP chan-nels may exert a positive effect on energyconsumption, as recently demonstratedin the failing heart (9, 10). In this con-text, it may be important that in the pres-ence of reduced oxygen delivery and cel-lular hypoxia, cells can adapt to maintainviability by down-regulating oxygen con-sumption, energy requirements, and ATPdemand (11), an adaptive phenomenonoften referred to as hibernation-type re-sponse (11). Whether this protectivemechanism plays a similar role in sepsis-related dysfunction in organs other thanthe heart remains to be determined infuture studies. However, because mito-chondrial dysfunction is critically in-volved in the pathogenesis of multipleorgan failure (11), a compound reducingoxygen consumption and preserving mi-tochondrial function represents an inter-esting therapeutic strategy in a state ofreduced energy availability, such as sep-

sis. By opening mitochondrial KATP chan-nels, levosimendan may contribute tomaintenance of mitochondrial volume, toamelioration of Ca2 overload duringischemia, and to preservation of mito-chondrial function. Because these mech-anisms are implicated in the preventionof apoptosis (12), it is conceivable thatlevosimendan likewise offers some degreeof protection against programmed celldeath.

The current literature on this topic,together with the work of Dr. Fries andcolleagues (8), makes it appear very likelythat besides its positive effect on myocar-dial contractility, levosimendan plays apivotal role in cellular protection. How-ever, although it appears that levosimen-dan is on the right path, future large-scale, randomized clinical trials are nowneeded to explicitly clarify whether or notcalcium sensitizing represents a rationalconcept in patients suffering from septicshock. If it can be proven that the bene-ficial effects of levosimendan translateinto a reduced length of hospital stay anda reduction in mortality, the discussionabout initially increased costs (as com-pared with current standard therapy)would certainly change significantly.

Christian Ertmer, MDDepartment of Anesthesiology

and Intensive CareUniversity Hospital of

MuensterMuenster, Germany

Andrea Morelli, MDDepartment of Anesthesiology

and Intensive CareUniversity of Rome,

“La Sapienza”Rome, Italy

Martin Westphal, MD, PhDDepartment of Anesthesiology

and Intensive CareUniversity Hospital of

MuensterMuenster, Germany

REFERENCES1. Westphal M, Morelli A, Van Aken H: Dear

Levosimendan, the right ventricle will thankyou! Crit Care Med 2007; 35:952–953

2. Dubin A, Murias G, Sottile JP, et al: Effects oflevosimendan and dobutamine in experimen-tal acute endotoxemia: A preliminary con-trolled study. Intensive Care Med 2007; 33:485–494

3. Morelli A, De Castro S, Teboul JL, et al:Effects of levosimendan on systemic and re-gional hemodynamics in septic myocardialdepression. Intensive Care Med 2005; 31:638–644

4. Morelli A, Teboul JL, Maggiore SM, et al:Effects of levosimendan on right ventricularafterload in patients with acute respiratorydistress syndrome: A pilot study. Crit CareMed 2006; 34:2287–2293

5. Gruhn N, Nielsen-Kudsk JE, Theilgaard S,et al: Coronary vasorelaxant effect of levosi-mendan, a new inodilator with calcium-sensitizing properties. J Cardiovasc Pharma-col 1998; 31:741–749

6. Parissis JT, Karavidas A, Bistola V, et al: Ef-fects of levosimendan on flow-mediated va-sodilation and soluble adhesion molecules inpatient with advanced chronic heart failure.Atherosclerosis 2007; 197:278–282

7. Schwarte LA, Picker O, Bornstein SR, et al:Levosimendan is superior to milrinone anddobutamine in selectively increasing micro-vascular gastric mucosal oxygenation indogs. Crit Care Med 2005; 33:135–142

8. Fries M, Ince C, Rossant R, et al: Levosi-mendan but not norepinephrine improvesmicrovascular oxygenation during experi-mental septic shock. Crit Care Med 2008;36:1886 –1891

9. Kevelaitis E, Peynet J, Mouas C, et al: Open-ing of potassium channels: The common car-dioprotective link between preconditioningand natural hibernation? Circulation 1999;99:3079–3085

10. De Luca L, Colucci WS, Nieminem MS, et al:Evidence-based use of levosimendan in dif-ferent clinical settings. Eur Heart J 2006;27:1908–1920

11. Singer M: Mitochondrial function in sepsis:Acute phase versus multiple organ failure.Crit Care Med 2007; 35(Suppl 9):S441–S448

12. Buckley JF, Singer M, Clapp LH: Role ofKATP channels in sepsis. Cardiovasc Res2006; 72:220–230


Monitoring the brain: Lack of tools or lack of will?*

I n this issue of Critical Care Med-icine, Dr. Jia and colleagues (1)address the role of an electrophys-iological marker to predict the

neurologic functional outcome associatedwith temperature manipulations followingcardiac arrest in rats. While hypothermiapreviously has been demonstrated to pro-vide neuroprotection in rats, the importantmessage of this article is that in usinga quantified (q) electroencephalogram(EEG), the authors were able to predict theeffect of various treatments on neurologicoutcome. This marker was predictive at atime when the animal was in a coma andthus the clinical neurologic exam unreveal-ing.

Currently, imaging is the predomi-nant mode of investigating the brain.However, it is not designed for continu-ous monitoring, only for point investiga-tion, even with the advance of scannertechnology allowing bedside imaging.Only a few techniques, such as positronemission tomography or single photonemission computed tomography, give in-formation on metabolic state, and thesetypes of investigations are not readilyavailable (2). Jugular bulb oximetry, aglobal reflection of the brain metabolicstatus, cannot be maintained in situ forprolonged periods (3). More recently,brain cortical microdialysis techniqueshave become available and bring a vastarray of information (4). However, thistype of monitoring also is highly invasive,provides focal rather than global moni-toring, and has yet to accumulate evi-dence that it improves outcome. Thus, wehave no easily applied tools available forcontinuous cerebral monitoring.

The EEG always has been a staple ofbrain evaluation. However, it is not rou-tinely utilized for brain-injured patients,

owing to what is felt to be a necessity forskilled interpretation. Therefore, EEGdata rarely are used in the acute settingfor clinical decision-making. It is remark-able that intensive care personnel com-monly employ continuous monitoring fororgans such as the heart and lungs, butnot the brain. A simple 5-lead EEG cannow be displayed readily on most inten-sive care and operating room monitorswith small computerized modules. Fur-thermore, qEEG facilitates the process ofdata interpretation. Commercially avail-able qEEG monitors are used in the op-erating room setting, developed in re-sponse to demand by clinicians toquantify the effects of anesthetic drugs onthe brain and to improve patient out-comes (5–6). These monitors, such as thebispectral index type, while far from per-fect, have given clinicians a new tool tomonitor their patients.

Information quantity (IQ) is derivedthrough processing of hemispheric EEGrecordings and represents the variabilitypresent in the tracings (7). Short win-dows at various time points after resusci-tation from cardiac arrest are analyzed bycomputer algorithm. Discrete wavelettransform is first applied to remove re-dundancy after which Shannon’s entropyis calculated (8). IQ is presented as a ratiocomparing this measurement to the pre-arrest baseline. The qEEG is able to iden-tify patterns associated with significantinjury; burst suppression, generalizedsuppression, and seizure activity are allperiodic patterns and score low by IQ.The authors demonstrate that early re-turn of variability, as measured by higherIQ, is predictive of neurologic recovery.This correlation is shown to be signifi-cant as early as 30 mins postresuscita-tion.

IQ as described by Dr. Jia and col-leagues (1), despite limitations, is a tech-nology that could be adapted rapidly tobring EEG to the bedside. Further inves-tigation must be made to translate IQinto clinical practice. IQ is expressed as aratio, relative to a baseline measurementthat one would not expect to be availablefor most patients. The authors suggestthat IQ could be computed based on a

standardized baseline. Clinicians couldthen follow a positive or negative trend,and possibly use this information toguide therapy. The absolute value may bepredictive as well. IQ could be subject topitfalls secondary to the effects of seda-tives and anesthetics on human patients.In high doses, many of these lead to slow-ing and burst suppression, which wouldbe expected to lower IQ (9). In this study,halothane did not appear to interfere withIQ. Furthermore, it is the unresponsivepatient, in whom sedation is rarely re-quired, where IQ would be most useful.Nevertheless, more work needs to bedone to define the influence of anesthet-ics on qEEG. Technology to suppress ar-tifact and improve signal–noise ratio alsomay be required to realize the full poten-tial of quantitative EEG at the bedside.

IQ may be valuable clinically as anearly predictive tool following neurologicinjury, particularly in the setting of thecomatose patient. As Dr. Jia and col-leagues suggest, IQ could help predictoutcome in patients resuscitated aftercardiac arrest (10). Conceivably, IQ alsocould provide useful information in othertypes of brain injury, such as intracranialhemorrhage and traumatic brain injury.In addition to providing prognostic infor-mation, this tool possibly could monitorresponse to therapeutic interventions. Ina more immediate translation to the clin-ical arena, the return of variability alsocould be monitored following major neu-rosurgical procedures when secondarybrain injury may be expected and directlycompared with a preanesthetic baseline.These data would help us to better followour patients’ evolution in case of pooroutcome, in conjunction with imagingdata and other tools.

In conclusion, Dr. Jia and colleaguesdemonstrate the validity of a novelmethod for continuous EEG monitoringto evaluate response to a guided therapy.At present, continuous cerebral monitor-ing in humans remains imprecise. Thisshould not deter us from doing so. Asevidenced in this work, new noninvasivemonitoring is on the horizon. The brainshould be monitored, especially in acutesituations. Only in doing so routinely will

*See also p. 1909.Key Words: brain monitoring; electroencephalog-

raphy; quantified EEG, information quantity; entropy;cardiac arrest

The authors have not disclosed any potential con-flicts of interest.


DOI: 10.1097/CCM.0b013e318176ae41


we learn from experience and stimulatenew ideas and technology. In this sense,the work of Dr. Jia and colleagues pointsthe way forward.

René Tempelhoff, MDJeffrey Yoder, MD

Washington UniversitySt. Louis, MO

REFERENCES

1. Jia X, Koenig MA, Nickl R, et al: Early elec-trophysiologic markers predict functionaloutcome associated with temperature ma-nipulation after cardiac arrest in rats. CritCare Med 2008; 36:1909–1916

2. Wartenberg KE, Schmidt JM, Mayer SA: Mul-

timodality monitoring in neurocritical care.Crit Care Clin 2007; 23:507–538

3. Feldman Z, Robertson CS: Monitoring of ce-rebral hemodynamics with jugular bulbcatheters. Crit Care Clin 1997; 13:51–77

4. Tisdall MM, Smith M: Cerebral microdialysis:Research technique or clinical tool. Br J An-aesth 2006; 97:18–25

5. Tempelhoff R, Modica PA, Rich KM, et al: Useof computerized electroencephalographicmonitoring during aneurysm surgery. J Neu-rosurg 1989; 71:24–31

6. Modica PA, Tempelhoff R: A comparison ofcomputerized EEG with internal carotid ar-tery stump pressure for detection of ischemiaduring carotid endarterectomy. J NeurosurgAnesthesiol 1989; 1:211–218

7. Jia X, Koenig MA, Shin HC, et al: Quantita-

tive EEG and neurological recovery withtherapeutic hypothermia after asphyxial car-diac arrest in rats. Brain Res 2006; 1111:166–175

8. Shin HC, Tong S, Yamashita S, et al: Quanti-tative EEG assessment of brain injury and hy-pothermic neuroprotection after cardiac arrest.Conf Proc IEEE Eng Med Biol Soc 2006;1:6229–6232

9. Sleigh J, Voss L, Bernard J: What are elec-troencephalogram entropies really measur-ing? Int Congr Ser 2005; 1283:231–234

10. Bernard SA, Gray TW, Buist MD, et al:Treatment of comatose survivors of out-of-hospital cardiac arrest with induced hypo-thermia. N Engl J Med 2002; 346:557–563

Brain tissue oxygen monitors: More than an ischemia monitor*

Brain cells need a constant sup-ply of oxygen and glucose fornormal energy metabolism.Therefore, continuous cere-

bral blood flow (CBF) and cerebral oxy-gen tension and delivery are essential tobrain function. In the absence of effectiveneuroprotective treatments, clinicalmanagement of patients with severe trau-matic brain injury (TBI) tries to maintainadequate cerebral perfusion pressure(CPP) and control intracranial pressure(ICP). Recent studies, however, havefailed to demonstrate that adequate CPPimproves patient outcome (1). This raisesthe possibility that there may be addi-tional mechanisms of tissue hypoxia, andso the interstitial partial pressure of oxy-gen in brain tissue (PbtO2) is emerging asan additional complementary therapeutictarget in TBI. This issue is importantsince brain tissue hypoxia is associatedwith worse outcome after severe TBI (2).

Direct PbtO2 monitors have been avail-able for clinical use in the United Statesfor several years. Despite this, there re-

main several unanswered questions: Inparticular, what are the determinants ofPbtO2? In this issue of Critical Care Med-icine, Dr. Rosenthal and colleagues (3)try to answer this question. The authorsexamined 14 patients with severe TBI andchallenged them with an increase in FIO2

to 1.0 (oxygen reactivity), an increase inmean arterial blood pressure by 10 mmHg (cerebral autoregulation), and a de-crease in PaCO2 of 10 mm Hg (CO2 cere-bral vascular reactivity). The study hasseveral limitations; for example, somephysiologic responses were measured us-ing local monitors (Licox and Hemedex),the challenges were administered in aconsecutive manner, hyperoxygenationwas maintained during the CO2 and meanarterial blood pressure challenges, andinformation from the challenges wasused to guide therapy. Nevertheless, thedata are robust (119 data points) and sug-gest that PbtO2 reflects the product ofcerebral blood flow and the arteriovenousdifference in oxygen tension, that is, PbtO2 �CBF � AVTO2. This finding is consistentwith experimental studies which suggestthat PbtO2 does not simply mirror CBF(4), positron emission tomography stud-ies in humans after TBI which show thatdiffusion rather than perfusion-limitedischemia may be responsible for cerebralhypoxia (5), and microdialysis studieswhich indicate that increases in the lac-tate-pyruvate ratio, a marker of anaerobicmetabolism, can occur independentlyfrom CPP (6). The findings of Dr.Rosenthal and colleagues (3) may apply

only to normal-appearing tissue and topatients with preserved cerebral autoreg-ulation after TBI. Additional studies areneeded to further explore the determi-nants of PbtO2 in pericontusional braintissue, where cerebral metabolism andtissue oxygenation may behave differ-ently, in TBI patients with impaired au-toregulation, or in conditions such assubarachnoid hemorrhage.

While earlier studies using xenoncomputed tomography suggested thatPbtO2 reflects regional CBF, the findingthat PbtO2 � CBF � AVTO2 implies arelationship between the amount of dis-solved plasma oxygen passing through agiven volume of brain per unit time andthe steady-state oxygen concentration inbrain tissue. This is important in TBIsince magnetic resonance imaging stud-ies demonstrate that cytotoxic edema isthe predominant edema after TBI and somay limit diffusion of needed metabolitesand oxygen to and from the cell (7). Inthese situations, the brain might requirehigher tissue oxygen tensions to main-tain sufficient tissue oxygenation. Fur-thermore, the results imply that a PbtO2

monitor is not simply an ischemia mon-itor or a measure of CBF. While reducedCBF (ischemia) is a cause of cellular hyp-oxia, there are other potential causes ofreduced oxygen. This is relevant sincepositron emission tomography studies innormal volunteers during visual activa-tion and hypoxia suggest that adequatelevels of tissue oxygenation can be main-tained without an increase in CBF (8). In

*See also p. 1917.Key Words: brain oxygen; cerebral blood flow;

cerebral metabolism; traumatic brain injuryDr. Le Roux has received honoraria from Integra

and grants from the Integra Foundation, Neurologica,and the Groff Foundation. Dr. Oddo has not disclosedany potential conflicts of interest.


DOI: 10.1097/CCM.0b013e318176a9f6


animal and clinical studies of TBI, vari-ables such as hemoglobin and lung func-tion influence PbtO2 independent of CBF(9–11), and cellular distress may be ob-served in the absence of ischemia (12).Certainly CBF is important, and how ox-ygenation of the brain responds to vari-ous therapies depends in part on CBF (13,14). However, knowing what is in theblood vessels of the brain (CPP and CBF)is only part of the information needed toprevent cellular dysfunction. While PbtO2

it is not a measure of supply and metab-olism, the findings of Dr. Rosenthal andcolleagues (3) suggest that at the veryleast, since PbtO2 reflects diffusion of dis-solved plasma oxygen, we now know whatis getting out of the blood vessels andinto the brain tissue.

There are other important observa-tions from the study. First, hyperoxygen-ation decreased CBF. In some circum-stances this may be dangerous, but as Dr.Rosenthal and colleagues (3) found, CPPincreased in part because ICP is reduced,likely through hyperoxic induced vaso-constriction. Similarly, hyperoxia can im-prove cellular metabolism because ofgreater oxygen availability (15), suggest-ing that the reduced CBF may not alwaysbe deleterious. However, Dr. Rosenthaland colleagues did not address the safety,feasibility, or efficacy of normobaric hy-peroxygenation. This topic remains con-troversial, since some studies suggest abenefit while others do not (14, 16). How-ever, hyperoxygenation (an FIO2 of 1.0)may miss the point, since Dr. Rosenthaland colleagues also observed that themean ratio of tissue to arterial or venousoxygen concentration is very small. Thisis consistent with Kety Schmidt’s originalhypothesis that the concentration of ox-ygen in brain tissue is very small relativeto the oxygen content of arterial and ve-nous blood. While the minimum tissuePO2 required to provide sufficient intra-cellular oxygen is unknown, neuronalmitochondria require an intracellular PO2

of 1.5 mm Hg to maintain aerobic me-tabolism. Together, these data suggestthat small changes in brain oxygen ten-sion to correct compromised brain oxy-gen may be all that is necessary to main-tain cellular health. A PbtO2 monitortherefore may help identify the cause forcellular distress and so tailor therapy tothe individual patient. In this way, a PbtO2

monitor is more than an ischemia mon-itor. Whether this knowledge can be usedto improve patient outcome after TBI willneed to be studied (17). Knowing whatPbtO2 represents is the first step in thatdirection.

Mauro Oddo, MDPeter Le Roux, MD

Department of NeurosurgeryUniversity of PennsylvaniaPhiladelphia, PA

REFERENCES

1. Robertson CS, Valadka AB, Hannay HJ, et al:Prevention of secondary ischemic insults af-ter severe head injury. Crit Care Med 1999;27:2086–2095

2. van den Brink WA, van Santbrink H, Steyer-berg EW, et al: Brain oxygen tension in se-vere head injury. Neurosurgery 2000; 46:868–876

3. Rosenthal G, Hemphill JC III, Sorani M, et al:Brain tissue oxygen tension is more indica-tive of oxygen diffusion than oxygen deliveryand metabolism in patients with traumatic braininjury. Crit Care Med 2008; 36:1917–1924

4. Scheufler K-M, Lehnert A, Rohrborn H-J,et al: Individual values of brain tissue oxygenpressure, microvascular oxygen saturation,cytochrome redox level and energy metabo-lites in detecting critically reduced cerebralenergy state during acute changes in globalcerebral perfusion. J Neurosurg Anesthesiol2004; 16:210–219

5. Menon DK, Coles JP, Gupta AK, et al: Diffu-sion limited oxygen delivery following headinjury. Crit Care Med 2004; 32:1384–1390

6. Vespa PM, O’Phelan K, McArthur D, et al:Pericontusional brain tissue exhibits persis-tent elevation of lactate/pyruvate ratio inde-pendent of cerebral perfusion pressure. CritCare Med 2007; 35:1153–1160

7. Marmarou A, Signoretti S, Fatouros PP, et al:Predominance of cellular edema in traumaticbrain swelling in patients with severe headinjuries. J Neurosurg 2006; 104:720–730

8. Mintun MA, Lundstrom BN, Snyder AZ, et al:Blood flow and oxygen delivery to humanbrain during functional activity: Theoreticalmodeling and experimental data. Proc NatlAcad Sci U S A 2001; 98:6859–6864

9. Rosenthal G, Hemphill JC, Sorani M, et al:The role of lung function in brain tissueoxygenation following traumatic brain in-jury. J Neurosurg 2008; 108:59–65

10. Smith MJ, Stiefel MF, Magge S, et al: Packedred blood cell transfusion increases local ce-rebral oxygenation. Crit Care Med 2005; 33:1104–1108

11. Hare GM, Mazer CD, Hutchison JS, et al:Severe hemodilutional anemia increases ce-rebral tissue injury following acute neuro-trauma. J Appl Physiol 2007; 103:1021–1029

12. Vespa P, Bergsneider M, Hattori N, et al:Metabolic crisis without brain ischemia iscommon after traumatic brain injury: Acombined microdialysis and positron emis-sion tomography study. J Cereb Blood FlowMetab 2005; 25:763–774

13. Hlatky R, Valadka AB, Gopinath SP, et al:Brain tissue oxygen tension response to in-duced hyperoxia reduced in hypoperfusedbrain. J Neurosurg 2008; 108:53–58

14. Nortje J, Coles JP, Timofeev I, et al: Effect ofhyperoxia on regional oxygenation and me-tabolism after severe traumatic brain injury:Preliminary findings. Crit Care Med 2008;36:273–281

15. Tolias CM, Reinert M, Seiler R, et al: Normo-baric hyperoxia-induced improvement in ce-rebral metabolism and reduction in intracra-nial pressure in patients with severe headinjury: A prospective historical cohort-matched study. J Neurosurg 2004; 101:435–444

16. Diringer MN, Aiyagari V, Zazulia AR, et al:Effect of hyperoxia on cerebral metabolic ratefor oxygen measured using positron emissiontomography in patients with acute severe headinjury. J Neurosurg 2007; 106:526–529

17. Stiefel MF, Spiotta A, Gracias VH, et al: Re-duced mortality rate in patients with severetraumatic brain injury treated with brain tis-sue oxygen monitoring. J Neurosurg 2005;103:805–811


Pediatric delirium: A new diagnostic challenge of which to beaware*

Probably one of the most fre-quent behavioral symptoms ofacute brain dysfunction amongpediatric patients is delirium,

which is a complex neuropsychiatric syn-drome that complicates physical illness,negatively impacting prognosis (1).

Delirium is a particularly derelict area,although it seems associated with longerhospital stay and higher mortality amonghospitalized pediatric patients (2).

The delirium evaluation is a multipro-fessional issue, and all professionals in-volved in the treatment of a hospitalizedchild—pediatricians, nurses, respiratorytherapists, etc.—should be aware of thepotential for this diagnosis and should bequalified to assess it. Though these pro-fessionals in pediatric intensive careunits are used to monitoring end organdysfunction, they often do not seem torecognize delirium (3).

The 2002 guidelines of the Society ofCritical Care Medicine for sedation andanalgesia published the parameters forthe use of sedatives and analgesics in thecritically ill adult, which also includedthe evaluation of delirium, clearly recom-mending the routine assessment of delir-ium (grade B recommendation) and thetreatment of delirium with haloperidol(grade C recommendation) (4). As long asthere is no similar guideline for pediatricpatients, these adult recommendationscan be followed.

There are two basic neurologic assess-ment steps to evaluate sedation and de-lirium. These involve evaluating the pa-tient’s level of consciousness/sedationand the brain’s function directed to thelevel of arousal. This is where the bigchallenge begins, because in pediatric pa-tients there is absence of a clear defini-

tion of delirium, the lack of a reliable andvalidated assessment tool, and great diffi-culty in distinguishing pain from delirium.

Delirium physiopathology remains amatter of intense research, but an impor-tant role seems to be played by the alter-ation of neuronal activity due to localbrain production of cytokines, cell infil-tration and tissue injury in response tosystemic infections, and inflammatory in-juries of the central nervous system. Theinflammatory injuries that affect thebrain include those induced by endo-toxin, cytokines, and hypoxemia (5).

There also is evidence that three neu-rotransmitter systems are involved in thedevelopment of delirium: Dopamine in-creases neuronal excitability, while ace-tylcholine and gamma-aminobutyric aciddecrease the excitability of neurons. Animbalance of one or more of these neu-rotransmitter systems in their synthesis,release, and inactivation results in neu-ronal instability and erratic neurotrans-mission, impairing an adequate controlof cognitive function, behavior, andmood. Serotonin imbalance, increasedcentral noradrenergic activity, and en-dorphin hyperfunction also may be in-volved in the development of delirium(6). Inadequate cerebral blood flowcaused by shock and coagulopathies andmetabolic disturbances also can be thesource of delirium.

Among all the causes that may influencedelirium, medical therapies, such as me-chanical ventilation and sedative and anal-gesic drug use, should also be mentioned.

These lead us to the risk factors, whichcan be split in five categories (7, 8):

● Host factors: Age, baseline comorbidi-ties, surgical procedure, pain;

● Acute illness itself: Sepsis, hypoxemia,disease severity score, stroke;

● Psychological: Difficult temperament,separation anxiety, premorbid psychi-atric condition, pain;

● Social: Anxious caregiver, caregiverpresence/absence, caregiver pain per-ceptions;

● Environmental or iatrogenic: Pediatric

intensive care unit admission, meta-bolic derangements, anticholinergicmedications, use of sedative and anal-gesic medications, noise, cool temper-ature, light, high number of hospitalprocedures, and pain.

Therefore, in the clinical setting, severalaspects of delirium—such as the use ofsedatives and analgesics, sepsis, and sur-gery and postoperative cognitive dysfunc-tion—should be taken into account, be-cause they are potentially modifiablefactors. Given this approach, pain, anxiety,agitation, and of course delirium cannot bemanaged through trial and error drug ad-ministration, as has been done in the past.

We also can look at the risk factors fordelirium from another point of view, thatof predisposing and precipitating factors(9). For instance, age would be a predis-posing factor and a metabolic disturbancea precipitating factor. It is interesting tonote that using disease severity scoresmay lead one to look mainly at the pre-cipitating factors and overlook the predis-posing ones.

For these reasons, we should aim to-ward the prevention of delirium, giventhe inherent diagnostic and treatmentcomplexity, the uncertainty about theclinical significance of delirium in chil-dren in the short and long term, the factthat it is biologically plausible that delir-ium is not only a marker of end organdamage but also a promoter of other or-gan system dysfunction (7), and the trendof intensivists to diagnose delirium onlyupon the presence or absence of an obvi-ous medical etiology (10).

In this issue of Critical Care Medicine,Dr. Schieveld and colleagues (11) presenttwo pediatric illness severity measures—Pediatric Index of Mortality (PIM) andPediatric Risk of Mortality (PRISM)II—as possible means to assess the risk ofpediatric delirium in pediatric intensivecare units. The authors acknowledge thedifficulty of diagnosing pediatric delir-ium, so they took extra care in presentingtheir data and in concluding that it is“worthwhile to employ PIM and PRISM II

*See also p. 1933.Key Words: delirium; critical care; pediatrics; risk

assessment mechanical ventilationThe authors have not disclosed any potential con-



DOI: 10.1097/CCM.0b013e318176aeba


in the risk assessment of pediatric delir-ium.” We could not agree more, mainlywhen we take into account the difficultyin correctly and quickly diagnosing andtreating delirium in the pediatric patient.Let’s prevent it rather than manage it.

Werther Brunow de Carvalho,MD, PhD

Marcelo Cunio MachadoFonseca, MD

Federal University of SaoPaolo

Sao Paolo, Brazil

REFERENCES

1. Leentjens AF, Schieveld JN, Leonard M, et al:A comparison of the phenomenology of pe-

diatric, adult, and geriatric delirium. J Psy-chosom Res 2008; 64:219–223

2. Turkel SB, Tavaré CJ: Delirium in childrenand adolescents. J Neuropsychiatry ClinNeurosci 2003; 15:431–435

3. Pandharipande P, Jackson J, Ely EW: Delir-ium: Acute cognitive dysfunction in the crit-ically ill. Curr Opin Crit Care 2005; 11:360–368

4. Jacobi J, Fraser GL, Coursin DB, et al:Clinical practice guidelines for the sus-tained use of sedatives and analgesics inthe critically ill adult. Crit Care Med 2002;30:119 –141

5. Gunther ML, Morandi A, Ely EW: Pathophys-iology of delirium in the intensive care unit.Crit Care Clin 2008; 24:45–65

6. Webb JM, Carlton EF, Geehan DM: Delir-ium in the intensive care unit: Are we

helping the patient? Crit Care Nurs Q2000; 22:47– 60

7. Scott GM, Gold JI: Emergence delirium: Are-emerging interest. Semin Anesth PerioperMed Pain 2006; 25:100–104

8. Devlin JW, Fong JJ, Fraser GL, et al: Delir-ium assessment in the critically ill. IntensiveCare Med 2007; 33:929–940

9. Burns A, Gallagley A, Byrne J: Delirium.J Neurol Neurosurg Psychiatry 2004; 75:362–367

10. Cheung CZ, Alibhai SM, Robinson M, et al:Recognition and labeling of delirium symp-toms by intensivists: Does it matter? Inten-sive Care Med 2008; 34:437–446

11. Schieveld J, Lousberg R, Berghmans E, et al:Pediatric illness severity measures predictdelirium in a pediatric intensive care unit.Crit Care Med 2008; 36:1933–1936

Steroid therapy of septic shock: The decision is in the eye of thebeholder

I n 1911 Rupert Waterhouse re-ported a case of cardiovascularcollapse in a septic infant with bi-lateral adrenal hemorrhage noted

at autopsy, calling the condition “supra-renal apoplexy.” This likely representedthe peak of clarity to date on adrenalinsufficiency in sepsis (Fig. 1; 1–15).

In this issue of Critical Care Medi-cine, Marik and colleagues, represent-ing an international task force, publishclinical practice guidelines (SCPG) forthe diagnosis and management of cor-ticosteroid insufficiency in critical ill-ness (16). The authors are to be con-gratulated for this very timely andclinically relevant publication. The rec-ommendations are evidence based andin general on target and useful for thebedside clinician. They coin a new term,critical illness related corticosteroid in-sufficiency (CIRCI), defined as inade-quate cellular corticosteroid (steroid)activity for the severity of the patient’sillness. CIRCI occurs as a result of ei-ther a decrease in adrenal steroid pro-

duction (adrenal insufficiency) or tissueresistance to glucocorticoids (with orwithout adrenal insufficiency). Al-though recommending that adrenal in-sufficiency in critical illness is best di-agnosed by a delta cortisol following250 mg cosyntropin of �9 �g/dL orrandom total cortisol of �10 �g/dL, theauthors recommend using clinical as-sessment and not the ACTH (cortico-tropin) stimulation test to identify pa-tients with septic shock or acuterespiratory distress syndrome (ARDS)who should receive steroid therapy forCIRCI. The document does point to thepotential for free cortisol measurement,not currently available at most centers,as having promise for decisions on ste-roid therapy in the future. The SCPGdocument recommendation for septicshock states “hydrocortisone should beconsidered in the management strategyof patients with septic shock, particu-larly those patients who have respondedpoorly to fluid resuscitation and vaso-pressor agents.” This is a 2B (weakstrength with moderate level of evi-dence support in a three-tiered qualityof evidence grading system) recommen-dation. This recommendation is similarto that of the recently published Sur-viving Sepsis Campaign (SSC) recom-mendation that states “we suggest thatintravenous hydrocortisone be givenonly to adult septic shock patients after

it has been confirmed that their bloodpressure is poorly responsive to fluidresuscitation and vasopressor therapy”(17). The SSC recommendation isgraded 2C (weak strength with a four-tiered quality of evidence grading sys-tem). Discussion in both sets of guide-lines point to the difference in patientpopulations enrolled in the French trial(which showed benefit of stress-dosesteroids in septic shock) and the COR-TICUS trial (which showed no survivalbenefit with a similar treatment strat-egy). The septic shock patients enrolledin the French trial were (1) more crit-ically ill by severity scores, (2) moreunstable from a blood pressure stand-point, and (3) enrolled earlier in septicshock. Any or all of these characteris-tics could lead to the difference in theoutcome as far as survival advantagebetween the two studies. As the authorspoint out, in addition, steroids wereavailable and widely used throughoutthe countries participating in theCORTICUS trial. Although there are nodata to allow ascertainment for oragainst selection bias, this remains apotential concern. It should also be rec-ognized that all four of the randomizedtrials of stress-dose steroid therapy ofseptic shock (including two single cen-ter studies) support the capability ofsteroids to reverse septic shock (11, 12,14, 15). In the CORTICUS trial the dif-

*See also p. 1937.Key Words: corticosteroid; sepsis; intensive care;

critical care; acute respiratory distress syndromeThe author has not disclosed any potential con-



DOI: 10.1097/CCM.0b013e31817d7ee4


ference in number of patients who hadshock reversal did not reach statisticalsignificance but a trend was present asit was in the single-center studies. Morerapid reversal of shock in those patientswho had shock reversal was demon-strated in the CORTICUS trial. A similarfinding was present in the ARDSnettrial studying stress-dose steroids inpersistent ARDS, where a decreased in-cidence of development of septic shockoccurred in the steroid-treated group(18). So in the end as to steroid therapyof septic shock, it appears to be allabout patient selection with a lingeringquestion of survival advantage. Unfor-tunately, but out of necessity, both theSCPG recommendation and the SSCrecommendation for septic shock arebased on bedside clinician judgment of“poor” response of blood pressure tofluid resuscitation and vasopressoragents without any specific definedthresholds for this situation. There isno literature evidence that would allowspecific threshold selection, i.e., “vaso-

1900 2000 20101970 1980 19901940 1950 19601910 1920 1930

Use of adrenal cortical extractsto treat infection (3)

Positive effect of steroids on rheumatoid arthritis

inflammation (4)

Recognition of link betweenexcessive inflammation and

lethal sepsis (5)

Prophylactic high dosesteroids protect animals

from sepsishemodynamic changes

(6)

High single dose steroidsincrease survival in septic shock

in single-center study (7)

Link between steroids,decrease in NF kappa B and amelioration of

pro-inflammatoryresponse in severe

sepsis (10)

Single center studies show↓ ICU stay and reversal of

septic shock with stressdose sustained steroid

therapy (11,12)

Multicenter, single country clinical trial

shows mortalityreduction with

sustained stressdose steroid therapyof septic shock (14)

European multi-centertrial attempting to

duplicate results of2002 positive trial of

sustained stress dosesteroids in septicshock fails (15)

WWI autopsies of sepsisdeaths show adrenal cortex

edema/necrosis (2)

No benefit of shorthigh dose steroids

in multicenterrandomized

controlled trials in septic shock (8,9)

Suprarenalapoplexy isdescribed(1)

Single center studydemonstrating increased

survival in persistentARDS with sustained

stress dose steroids (13)

Figure 1. A history of the literature that has influenced our thought process concerning utility of steroid therapy of septic shock and acute respiratorydistress syndrome.

Table 1. Steroid therapy of acute respiratory distress syndrome (ARDS)

Early/Severe ARDS Late/Persistent ARDS

Short duration high-dosesteroids

No Not studied. Would not beexpected to work

Low-dose (stress-dose) sustainedsteroids

Potential. Needs furtherstudy

Improves oxygenation andpatients off mechanicalventilation quicker, but nodifference in outcome

Select patients might benefit.Controversial

Table 2. Steroid therapy of septic shock

Septic Shock with BloodPressure Poorly Responsive to Fluids

and Vasopressors (highdose/multiple agents)

Septic Shock RequiringVasopressors but Single Pressor/

Lower Dose Range

Short duration high-dosesteroids

No No

Low-dose (stress-dose)sustained steroids

Yes for shock reversal; morecontroversial for survival butprobably should be given withlower level of evidence thanshock reversal.

No


pressors greater than ? with adequatefluid resuscitation judged by ?” My ownpersonal bias is that following adequate(optimal?) fluid resuscitation, patientsreceiving two vasopressors or high-dosenorepinephrine (or equivalent vaso-pressor) or increasing moderate dosesof norepinephrine (or equivalent)would qualify for steroid treatment.Earlier is likely better. Comparing theSCPG and SSC document concerningstress-dose steroid therapy of septicshock, the recommendations for use ofACTH stimulation test, tapering of ste-roids with resolution of shock, and pref-erence for hydrocortisone are essen-tially identical.

Concerning ARDS, it is recom-mended in the SCPG document thatsteroids be considered in patients withearly severe ARDS (PaO2:FIO2 �200) andin addition before day 14 in patientswith unresolving ARDS. Again thestrength of recommendation is 2B. Thedata for decision making in acute earlysevere ARDS are less robust than thedata for decision making in unresolvingARDS (18 –20). The current data on ste-roids for acute severe ARDS, althoughencouraging, needs further validationin larger randomized trials before thistherapy should be considered for gen-eral use in clinical practice. As to use ofsteroids in persistent (unresolving)ARDS, the authors of the SCPG docu-ment draw different conclusions thanthose of the ARDSnet investigators whocarried out the trial of steroids in un-resolving ARDS, despite that study con-tributing the majority of evidence forthe SCPG guidelines recommendation(16, 18). The ARDSnet investigatorsconcluded that there was no clinicalbenefit from use of stress-dose steroidsin unresolving ARDS. Decisions for ste-roid use in the circumstance of unre-

solving ARDS need to be individualizedand steroids should not routinely beused in this circumstance as no survivaladvantage was demonstrated in theARDSnet trial (18). Tables 1 and 2 pro-vide my current opinions on use of ste-roid therapy in ARDS and septic shock,respectively.

R. Phillip Dellinger, MD, FCCMCooper Health

Systems - Critical Care SectionCamden, NJ

REFERENCES

1. Waterhouse R: A case of suprarenal apoplexy.Lancet 1911; I:577–578

2. Meduri GU: An historical review of glucocor-ticoid treatment in sepsis. Disease patho-physiology and the design of treatment in-vestigation. Sepsis 1999; 3:21–38

3. Perla D, Marmorston J: Suprarenal corticalhormone and salt in the treatment of pneu-monia and other severe infections. Endocri-nology 1940; 27:367–374

4. Kinsell LW, Jahn JP: The use of corticoids inassociation with antibiotics in the manage-ment of unusually severe infections. Ann NYAcad Sci 1955; 61:397–407

5. Breen GE, Talukdar PK: Corticosteroids inacute infections. Lancet 1965; 1:158 –160

6. Motsay GJ, Alho A, Jaeger T, et al: Effects ofcorticosteroids on the circulation in shock:experimental and clinical results. Fed Proc1970; 29:1861–1873

7. Schumer W: Steroids in the treatment of clin-ical septic shock. Ann Surg 1976; 184:333–339

8. Bone RC, Fisher CJ, Clemmer TP, et al: Acontrolled clinical trial of high dose methylprednisolone in the treatment of severe sep-sis and septic shock. N Engl J Med 1987;317:653–658

9. The Veterans Administration Systemic SepsisCooperative Study Group: Effect of high-doseglucocorticoid therapy on mortality in pa-tients with clinical signs of systemic sepsis.N Engl J Med 1987; 317:659–665

10. Scheinman RI, Cogswell PC, Lofquist AK, et al:Role of transcriptional activation of IkBa in

medication of immunosuppression by glu-cocorticoids. Science 1995; 270:283–290

11. Bollaert PE, Charpentier C, Levy B, et al:Reversal of late septic shock with supraphysi-ologic doses of hydrocortisone. Crit CareMed 1998; 26:645–650

12. Briegel J, Forst H, Haller M, et al: Stressdoses of hydrocortisone reverse hyperdy-namic septic shock: A perspective, random-ized, double-blind, single-center study. CritCare Med 1999; 27:723–732

13. Meduri GU, Headley S, Golden E, et al: Effectof prolonged methylprednisolone therapy inunresolving acute respiratory distress syn-drome. A randomized controlled trial. JAMA1998; 280:159–165

14. Annane D, Sebille V, Charpentier C, et al:Effect of treatment with low doses of hydro-cortisone and fludrocortisone on mortality inpatients with septic shock. JAMA 2002; 288:862–871

15. Sprung CL, Annane D, Briegel J, et al: Corti-costeroid therapy of septic shock (CORTICUS).Abstr. Am Rev Respir Crit Care Med 2007;175:A507

16. Marik PE, Pastores SM, Annane D, et al:Recommendations for the diagnosis andmanagement of corticosteroid insufficiencyin critically ill adult patients: Consensusstatements from an international task forceby the American College of Critical CareMedicine. Crit Care Med 2008; 36:1937–1949

17. Dellinger RP, Levy MM, Carlet JM, et al:Surviving Sepsis Campaign: Internationalguidelines for management of severe sepsisand septic shock. Crit Care Med 2008; 36:296–327

18. The Acute Respiratory Distress SyndromeNetwork: Efficacy and safety of corticoste-roids for persistent acute respiratory distresssyndrome. N Engl J Med 2006; 354:1671–1684

19. Confalonieri M, Urbino R, Potena A, et al:Hydrocortisone infusion for severe commu-nity-acquired pneumonia: a preliminary ran-domized study. Am J Respir Crit Care Med2005; 171:242–248

20. Meduri GU, Golden E, Freire AX, et al: Meth-ylprednisolone infusion in patients with earlysevere ARDS: results of randomized trial.Chest 2007; 131:954–963


npt el final

Documents

initial resuscitation

septic shock

paradigms of care

hr resuscitation care

initial vasopressorsupport

vasoactive support

issue ofcritical care

intensivecare unit icu