metabolic and hemodynamic consequences of sodium bicarbonate administration in patients with heart...

8
- CLINICAL STUDIES Metabolic and Hemodynamic Consequences of Sodium Bicarbonate Administration in Patients with Heart Disease ROBERTM. BERSIN, M.D., KANU CHATTERJEE, M.B., ALLEN I. ARIEFF, M.D. SanFrancisco, Ca/ifornia PURPOSE: The use of sodium bicarbonate (NaHCOs) in cardiopulmonary arrest has been questioned, but the effects of NaHC03 in patients with heart diseaseare not known. We therefore prospectively evaluated the effects of NaHC03 in patients with congestive heart failure. PATIENTSANDMETHODS: Tenpatientsreceived NaHCOa and control infusions of equimolar sodium chloride (NaCl). Measurements were made of blood gases,2,3diphosphoglyceric acid (2,3-DPG), glucose, lactate, cardiac hemodynamics, and oxy- gen consumption. RESULTS: The arterial oxygen tension (~02) fell an average of 10 mm Hg after NaHCOa administra- tion in patients with congestive heart failure, whereas it rose with NaCl (p <0.005). Myocardial oxygen consumption decreased by 17% (p <0.002) without an accompanying change in oxygen de- mand. Systemic oxygen consumption fell by 21%. Red blood cell 2,3-DPG levels were elevated at baseline, but did not change with NaHCO3 admin- istration. The oxygen pressure at 50% hemoglobin saturation (PRO) was correspondingly elevated at baseline in these patients and decreased signifi- cantly with NaHCOs (Bohr effect) (p cO.003). The arterial and mixed venous carbon dioxide tensions increased with NaHCOs but decreased with NaCl administration (p cO.05). Blood glucose concentra- tions fell by 1.7 mmol/L with NaHC03 (p <0.003) and blood lactate concentrations increased uni- formly (p <O.OOl). Three patients developed net myocardial lactate generation during NaHC03 ad- ministration; two of these three developed symp- toms of angina. Coronary blood flow did not change with NaHC03 but increased with NaCl (p <0.04). Two patients developed transient pump fail- ure. CONCLUSION: These data demonstrate that NaHCOa impairs arterial oxygenation and reduces systemic and myocardial oxygen consumption. The decrease in oxygen utilization is associated with anaerobic metabolism, enhanced glycolysis, and el- evation of the blood lactate level, and may lead to transient myocardial ischemia in some patients. Thus, the use of NaHC03 in such patients warrants re-evaluation. From the Cardiology Division, University of California Medical Center, and Geriatrics Division, Veterans Administration Medical Center, San Francisco, California. This work was supported by grants from the National Heart, Lung and Blood Institute (no. HL01791). the National Institute of Arthritis, Metabolism, Digestive and Kidney Diseases (no. DK18350). the American Heart Association (no. 88-l 185) and the Research Service of the Veterans Administration. Dr. Bersin is a recipient of the Physician-Scientist Award. National Institutes of Health, and is a Winthrop Pharmaceuticals Grant-in- Aid Awardee of the American Heart Association. Dr. Chatterjee is the Lucie Stern Professor of Cardiology of the University of California. Requests for reprints should be addressed to Robert M. Bersin. M.D., Cardiology Division (BoxO124). UCSF Moffitt Hospital, 505 Parnassus Avenue, San Francisco. California 94143. Manuscript submitted June 29, 1988, and accepted in revised form May 11, 1989. I n cardiopulmonary arrest, hypoxia and circulatory insufficiency combine to reduce tissue oxygen avail- ability, resulting in anaerobic metabolism and lactic acidosis [I]. Sodium bicarbonate (NaHCOs) is the most frequently employed therapeutic agent for the management of lactic acidosis in this setting [2-41. However, controversy exists as to whether the admin- istration of NaHCOs is of benefit [5-121. NaHCOs raisesthe carbon dioxide tension (pCO3 of blood, and thus of the cells, as carbon dioxide diffuses readily across tissue membranes [12-161. An increase of intra- cellular carbon dioxide is believed to exacerbate the intracellular acidosisalready present in hypoxic states [16,17] and cardiopulmonary arrest [18,19]. Experi- mental studies suggest that this may be a clinically important phenomenon. In separate studies, either the administration of NaHCOs or COs results in de- pression of cardiac function, presumably asa result of lowering intracellular pH in both situations [2O-221. In dogs with hypoxia and lactic acidosis, we confirmed a rise of venous pCOs and a concomitant fall of tissue intracellular pH with NaHCOs administration [16]. Hypotension and cardiac failure resulted [23]. The acid-base status of these animals not only failed to improve, but actually worsened as bicarbonate was infused. Data in humans on the effects of NaHCOs adminis- tration during cardiopulmonary arrest (or other states characterized by hypoxia and circulatory insufficien- cy) are limited [11,12,18,19]. Reliable data on hemody- namic and metabolic function during cardiopulmo- nary arrest are difficult to obtain. We therefore elected to investigate the effects of NaHCOs administration in patients with heart failure. Patients with heart failure generally have impaired ventilatory capacities, pulmo- nary congestion, mild hypoxia, and circulatory insuffi- ciency [24]. As a result, tissue oxygen availability is frequently limited, but not to the extent observed in cardiopulmonary arrest. Their clinical circumstances are thus similar to those of patients suffering cardio- pulmonary arrest, only not assevere and more suitable for study. They also constitute a subpopulation of pa- tients with a high risk of cardiopulmonary arrest [25]. In this study, we describe the metabolic and hemody- namic consequences of NaHCOs administration in pa- tients with heart disease. PATIENTS AND METHODS Patients Studies were carried out in 10 patients with stable New York Heart Association (NYHA) class III or IV congestive heart failure [26]. This group included sev- en men and three women aged 45 to 85 years (median, 65 years), All 10 patients had depressedleft ventricu- lar function as assessed by a resting gated equilibrium radionuclide angiogram (range of ejection fractions, 10% to 34%; median, 20%). Nine patients had cardiac failure secondary to coronary artery diseaseand one had primary idiopathic cardiomyopathy. All studies July 1989 The American Journal of Medicine Volume 87 7

Upload: robert-m-bersin

Post on 31-Oct-2016

213 views

Category:

Documents


1 download

TRANSCRIPT

- CLINICAL STUDIES

Metabolic and Hemodynamic Consequences of Sodium Bicarbonate Administration in Patients with Heart Disease ROBERTM. BERSIN, M.D., KANU CHATTERJEE, M.B., ALLEN I. ARIEFF, M.D. SanFrancisco, Ca/ifornia

PURPOSE: The use of sodium bicarbonate (NaHCOs) in cardiopulmonary arrest has been questioned, but the effects of NaHC03 in patients with heart disease are not known. We therefore prospectively evaluated the effects of NaHC03 in patients with congestive heart failure.

PATIENTSANDMETHODS: Tenpatientsreceived NaHCOa and control infusions of equimolar sodium chloride (NaCl). Measurements were made of blood gases, 2,3diphosphoglyceric acid (2,3-DPG), glucose, lactate, cardiac hemodynamics, and oxy- gen consumption.

RESULTS: The arterial oxygen tension (~02) fell an average of 10 mm Hg after NaHCOa administra- tion in patients with congestive heart failure, whereas it rose with NaCl (p <0.005). Myocardial oxygen consumption decreased by 17% (p <0.002) without an accompanying change in oxygen de- mand. Systemic oxygen consumption fell by 21%. Red blood cell 2,3-DPG levels were elevated at baseline, but did not change with NaHCO3 admin- istration. The oxygen pressure at 50% hemoglobin saturation (PRO) was correspondingly elevated at baseline in these patients and decreased signifi- cantly with NaHCOs (Bohr effect) (p cO.003). The arterial and mixed venous carbon dioxide tensions increased with NaHCOs but decreased with NaCl administration (p cO.05). Blood glucose concentra- tions fell by 1.7 mmol/L with NaHC03 (p <0.003) and blood lactate concentrations increased uni- formly (p <O.OOl). Three patients developed net myocardial lactate generation during NaHC03 ad- ministration; two of these three developed symp- toms of angina. Coronary blood flow did not change with NaHC03 but increased with NaCl (p <0.04). Two patients developed transient pump fail- ure.

CONCLUSION: These data demonstrate that NaHCOa impairs arterial oxygenation and reduces systemic and myocardial oxygen consumption. The decrease in oxygen utilization is associated with anaerobic metabolism, enhanced glycolysis, and el- evation of the blood lactate level, and may lead to transient myocardial ischemia in some patients. Thus, the use of NaHC03 in such patients warrants re-evaluation.

From the Cardiology Division, University of California Medical Center, and Geriatrics Division, Veterans Administration Medical Center, San Francisco, California. This work was supported by grants from the National Heart, Lung and Blood Institute (no. HL01791). the National Institute of Arthritis, Metabolism, Digestive and Kidney Diseases (no. DK18350). the American Heart Association (no. 88-l 185) and the Research Service of the Veterans Administration. Dr. Bersin is a recipient of the Physician-Scientist Award. National Institutes of Health, and is a Winthrop Pharmaceuticals Grant-in- Aid Awardee of the American Heart Association. Dr. Chatterjee is the Lucie Stern Professor of Cardiology of the University of California. Requests for reprints should be addressed to Robert M. Bersin. M.D., Cardiology Division (BoxO124). UCSF Moffitt Hospital, 505 Parnassus Avenue, San Francisco. California 94143. Manuscript submitted June 29, 1988, and accepted in revised form May 11, 1989.

I n cardiopulmonary arrest, hypoxia and circulatory insufficiency combine to reduce tissue oxygen avail-

ability, resulting in anaerobic metabolism and lactic acidosis [I]. Sodium bicarbonate (NaHCOs) is the most frequently employed therapeutic agent for the management of lactic acidosis in this setting [2-41. However, controversy exists as to whether the admin- istration of NaHCOs is of benefit [5-121. NaHCOs raises the carbon dioxide tension (pCO3 of blood, and thus of the cells, as carbon dioxide diffuses readily across tissue membranes [12-161. An increase of intra- cellular carbon dioxide is believed to exacerbate the intracellular acidosis already present in hypoxic states [16,17] and cardiopulmonary arrest [18,19]. Experi- mental studies suggest that this may be a clinically important phenomenon. In separate studies, either the administration of NaHCOs or COs results in de- pression of cardiac function, presumably as a result of lowering intracellular pH in both situations [2O-221. In dogs with hypoxia and lactic acidosis, we confirmed a rise of venous pCOs and a concomitant fall of tissue intracellular pH with NaHCOs administration [16]. Hypotension and cardiac failure resulted [23]. The acid-base status of these animals not only failed to improve, but actually worsened as bicarbonate was infused.

Data in humans on the effects of NaHCOs adminis- tration during cardiopulmonary arrest (or other states characterized by hypoxia and circulatory insufficien- cy) are limited [11,12,18,19]. Reliable data on hemody- namic and metabolic function during cardiopulmo- nary arrest are difficult to obtain. We therefore elected to investigate the effects of NaHCOs administration in patients with heart failure. Patients with heart failure generally have impaired ventilatory capacities, pulmo- nary congestion, mild hypoxia, and circulatory insuffi- ciency [24]. As a result, tissue oxygen availability is frequently limited, but not to the extent observed in cardiopulmonary arrest. Their clinical circumstances are thus similar to those of patients suffering cardio- pulmonary arrest, only not as severe and more suitable for study. They also constitute a subpopulation of pa- tients with a high risk of cardiopulmonary arrest [25]. In this study, we describe the metabolic and hemody- namic consequences of NaHCOs administration in pa- tients with heart disease.

PATIENTS AND METHODS Patients

Studies were carried out in 10 patients with stable New York Heart Association (NYHA) class III or IV congestive heart failure [26]. This group included sev- en men and three women aged 45 to 85 years (median, 65 years), All 10 patients had depressed left ventricu- lar function as assessed by a resting gated equilibrium radionuclide angiogram (range of ejection fractions, 10% to 34%; median, 20%). Nine patients had cardiac failure secondary to coronary artery disease and one had primary idiopathic cardiomyopathy. All studies

July 1989 The American Journal of Medicine Volume 87 7

SODIUM BICARBONATE IN HEART DISEASE / BERSIN ET AL

PC.005

-20 L 0 30 60 90 120

TIME Figure 1. Net change in arterial oxygenation (A ~02 in mm Hg) with NaCl (+) and NaHCOs (a) infusions. The means f SEM are shown.

w -7 PC.004

0 30 60 90 120

TIME

Figure 2. Net change in myocardial oxygen consumption (A mL 02. minute) with NaCl (+) and NaHCOs (8) infusions. The means f SEM are shown.

were carried out electively in the Coronary Care Unit of the Moffitt Hospital, University of California, San Francisco. Arterial pressure was monitored via a 20- gauge catheter placed in a radial artery. Pulmonary pressures and cardiac output were measured with a balloon-tipped, flow-directed catheter placed in the pulmonary arterial bed. Coronary sinus blood flow was measured by placement of a Wilton Webster thermo- dilution catheter (Webster Laboratories, Altadena, California) in the coronary sinus. Correct position of the pulmonary arterial and coronary sinus catheters was confirmed by fluoroscopy and contrast adminis- tration. All blood sampling was performed via these catheters.

Protocol Stable baseline values were established with three

consecutive sets of cardiac hemodynamic measure- ments and coronary sinus blood flows within 10% of each other. Baseline blood sampling followed. All pa- tients then received 2.5 mEq/kg of body weight of NaHCOs over one hour by continuous infusion of 595 mmol/L NaHCOs. Each infusion was followed by a one-hour post-infusion recovery period. Cardiac he- modynamics and coronary sinus blood flows were measured at 15-minute intervals during the infusion and recovery periods (total of eight determinations per infusion), whereasblood samples were obtained at

30-minute intervals (total of four determinations). To control for the effect of the volume administered, sodi- um content, and the osmolality of the solution, a sec- ond control infusion of 2.5 mEq/kg of body weight 513 mmol/L sodium chloride (NaCl) was then adminis- tered to each subject following an identical protocol. The order in which each subject received the two infu- sions was randomized and all patients were unaware of the identity of the infusions administered.

This protocol was approved by the Institutional Re- view Board of the University of California, San Fran- cisco. Informed consent was obtained from each pa- tient. No patient underwent right heart catheterization solely for purposes of this study. In every case, this procedure was performed for other clinical purposes prior to entry into this study.

Cardiac Hemodynamic Measurements Measurements were made of the right atria1 mean

pressure (RA), pulmonary artery systolic, diastolic, and mean (PAM) pressures, pulmonary capillary wedge pressure (PCW), and systemic arterial systolic, diastolic, and mean (MAP) pressures (all in mm Hg). Cardiac output (CO) was determined in L/minute in triplicate by the thermodilution technique [27]. Sys- temic vascular resistance (SVR in dynes/second/ cm-s) was calculated by the formula: SVR = ([MAP - RA] SO)/CO. Similarly, pulmonary vascular resistance (PVR) was calculated by the formula: PVR = ([PAM - PCW] SO)/CO. Coronary sinus blood flow was mea- sured in ml/minute by continuous thermodilution P81. Blood Measurements

Arterial, mixed venous (pulmonary artery distal), and coronary sinus blood gases were measured at 37°C with a Corning model 168 or model 178 blood gas ana- lyzer (Corning Medical, Medfield, Massachusetts). Oxygen saturations were determined with a Radiome- ter model OSM-2 oximeter (Radiometer, Copenhagen, Denmark). The hemoglobin content of blood (g/dL of blood) was measured every 30 minutes using the Drob- kins cyanmethemoglobin method. Oxygen content (volume percent [vol%]) was calculated by the formu- la: oxygen content = 1.34 X hemoglobin X percent oxygen saturation). Systemic oxygen consumption was calculated by the formula: systemic oxygen con- sumption = arterial-venous oxygen content difference X cardiac output. Myocardial oxygen consumption was estimated by the formula: myocardial oxygen con- sumption = arterial-coronary sinus oxygen content difference X coronary sinus blood flow. The oxygen pressure at 50% hemoglobin saturation (Pss in mm Hg) was calculated on mixed venous blood corrected for temperature using directly measured oxygen satura- tions and published tables [29,30].

Measurements of 2,3-diphosphoglyceric acid (2,3- DPG) and lactate were made on arterial, mixed ve- nous, and coronary sinus blood deproteinized immedi- ately with 5.5 mmol perchloric acid and frozen at -5°C. Both were measured spectrophotometrically using NAD/NADH assays (Sigma Chemicals, St. Lou- is, Missouri) [31,32]. Plasma glucose determinations were made on arterial samples using the glucose oxi- dase method. Plasma electrolytes were measured on arterial samples by flame photometry.

Statistical comparisons were made using the area

8 July 1989 The American Journal of Medicine Volume 87

SODIUM BICARBONATE IN HEART DISEASE / BERSIN ET AL

-4’ / -20 -10 0 10 20 30 40 50 60

r = .43

I 0 131 -6'

I"

-30 -20 -10 0 10 20 30

CORONARY BLOOD FLOW

Figure 3. Relation of myocardial oxygen consumption (A mL Oz/ minute) to coronary sinus blood flow (A ml/minute) with NaCl (0) and NaHCOs (0) infusions.

under the difference curve method with a zero base- line. Significance was estimated using the repeated- measures analysis of variance technique. Data are re- ported as means f SEM. Differences between the two infusion curves were considered significant when p <0.05. Statistical analyses were accomplished with a Macintosh Plus (Apple Computer Corp., Cupertino, California) computer and a Statview 512f statistical software package (Calabasas, California).

RESULTS Arterial Oxygenation

A large and statistically significant decrease of arte- rial oxygen tension (~0s) occurred with the infusion of NaHCOs, whereas an increase occurred with NaCl (Figure 1). The mean decrease of the arterial pOz was 9 mm Hg in a group of patients whose average baseline room-air ~0s was only 73 mm Hg (median, 74 mm Hg). Arterial oxygen content values correspondingly fell an average of 1.64 ~01%. Myocardial oxygen delivery and consumption also decreased significantly in all pa- tients with the infusion of NaHCOs, whereas no change was observed with NaCl (Figure 2). A multiple regression analysis demonstrated that the normal mechanisms by which myocardial oxygen demands are met were altered by the administration of NaHCOs.

With the infusion of NaCl, the normal physiology of myocardial oxygenation was unaltered. The most im- portant means by which myocardial oxygen demands were met remained coronary blood flow (Figure 3). Oxygen consumption remained independent of myo- cardial oxygen extraction (Figure 4). However, with the infusion of NaHCOs, myocardial oxygen consump- tion failed to correlate with coronary blood flow and instead was affected most by a reduction of myocardi- al oxygen extraction (Figures 3 and 4).

A decrease of the arterial to venous (A-V) gradient of oxygen across a vascular bed implies that less oxy-

4l

r = .21

0

O 0

z? 2 0

F P 0 --- O n-

: O 0 0

0 2 0 0 0

z -2

0

0 0 0 "s 00 0

s -4 2 -4 -3 -2 -1 0 1

:: I r = .80 I

MYOCARDIAL OXYGEN EXTRACTION

Figure 4. Relation of myocardial oxygen consumpbon (A mL OZ/ minute) to myocardial oxygen extraction (A ~01%) with NaCl(0) and NaHCOs (0) infusions.

0 30 60 90 120

TIME

igure 5. Net change in systemic oxygen consumption (ra) (A mL 02 minute/m*) and blood glucose (+) (mmol/L) with NaHCOa infusion. Mean net changes from baseline are shown. Note that the blood glucose level decreases in parallel with reductions in systemic oxygen consumption.

gen is extracted by the tissues supplied by that vascu- lar bed unless blood flow is increased in compensation. In the present study, NaHCOs decreased arterial oxy- gen content (p <O.OOl) and increased coronary venous oxygen content (p <O.OOl) at a time when coronary blood flow did not change. The net result was a signifi- cant decrease of myocardial oxygen extraction, which was not seen when NaCl was administered to the same patients (Figure 2).

Systemic oxygen consumption was also substantial- ly reduced with NaHCOs administration, from 152 to 114 mL oxygen/minute/m2 of body surface area (BSA) (mean decrease -21%, p <0.004) (Figure 5). This was due to a significant reduction of systemic oxygen ex-

July 1989 The American Journal of Medicine Volume 87 9

SODIUM BICARBONATE IN HEART DISEASE / BERSIN ET AL

TABLE I

Complete List of Blood Gas Values Measured in This Study*

Time PH PC& (mm Hg) POZ (mm Hg) HCOs-(mEq/L) % 02 Saturation

Art PA CS Art PA CS Art PA CS Art PA CS Art PA CS

NaCl Infusions Baseline 7.43 7.39 7.39 36.9 42.7 48.0 78.4 34.8 20.6 24.0 25.6 27.7 95.1 61.9 29.5 30 minutes 7.41 7.38 7.38 38.6 42.9 47.7 87.5 34.6 21.0 24.1 24.9 27.5 94.3 60.4 28.1 60 minutes 7.42 7.38 7.38 37.6 41.7 48.7 79.1 34.7 21.3 23.8 24.7 26.8 92.9 60.7 29.3 90 minutes 7.43 7.38 7.38 35.6 43.6 47.2 91.1 33.6 19.8 23.1 25.7 25.8 95.0 29.4 120 minutes 7.43 7.40 7.40 36.6 42.1 48.7 86.9 34.1 20.4 23.9 25.1 28.1 94.3

E 29.4

NaHCO3 Infusions Baseline 7.44 7.39 7.36 34.7 42.8 49.4 80.2 32.9 19.0 30 minutes 7.48 7.43 7.41 37.6 43.3 50.2 70.6 32.5 20.2 60 minutes 7.51 7.45 7.43 38.0 44.5 50.7 76.5 31.9 18.3

90 minutes 7.50 7.45 7.42 36.7 45.0 50.5 77.1 33.0 18.7 120 minutes 7.50 7.45 7.42 35.6 44.2 49.0 81.7 29.6 18.7

Art = arterial; PA = pulmonary arterial; CS = coronary sinus; HCOs- = bicarbonate; 02 = oxygen. l The means are shown. None of the baseline values differed significantly from each other.

23.0 25.8 27.6 54.6 23.5 27.2 28.5 31.1 z: 59.4 25.7 29.4 30.5 33.2 94:4 60.7 25.6

28.3 30.5 32.1 94.7 27.2 30.0 31.1 94.0 E . :i: .

traction (from 5.70 to 4.50 vol%, p <O.OOl), which was not seen with NaCl (p <0.04).

Blood Gas Analysis The arterial pH rose 0.08 pH units with NaHC03

administration (from 7.43 to 7.51; p <O.OOl), whereas it did not change with NaCl administration (7.43 to 7.42, p = NS). The arterial pCOs rose with NaHCOs (from 34.7 to 38.0 mm Hg) as did the mixed venous pCOs (from 42.8 to 44.5 mm Hg), whereas neither changed with NaCl. Although the increase of both ar- terial and mixed venous pCOs was significant with NaHCOs administration (p <0.05), the magnitude of change observed was small relative to that of the bicar- bonate concentration, and thus the overriding effect of NaHCOs was to increase blood pH. The rise of the arterial pH with NaHCOs administration shifted the oxygen-hemoglobin binding curve to the left (Bohr effect). The Pss of mixed venous blood was elevated at baseline, 29.5 mm Hg (normal = 26.6), and decreased with NaHCOs administration (from 29.5 to 28.0, p cO.003). However, the fall of P50 was less than expect- ed for the magnitude of change of arterial pH observed (calculated Bohr effect = log Psc/pH = 0.36; normal = 0.48). Arterial 2,3-DPG levels were also ,elevated at

< -0.2

E E -0.6

9 -1 .o

0 30 60 90 120

TIME 1

Figure6. Net change in arterial lactate concentrations (A in mmol/L) with NaCl (+) and NaHC03 ( q ) infusions. The means f SEM are shown.

baseline, but did not change with either infusion (from 26.4 to 29.3 ctrnollg hemoglobin, p = NS). Thus, the reduction of arterial oxygenation and of myocardial and systemic oxygen extraction with NaHCOs admin- istration could not be explained entirely by changes of oxygen-hemoglobin binding or of 2,3-DPG levels. A complete list of blood gas measurements is provided in Table I.

Blood Lactate The arterial lactate concentration increased from

1.0 to 1.5 mmol/L with NaHCOs and decreased with NaCl administration in the same patients (Figure 6). Mixed venous and coronary sinus lactate levels fol- lowed the same pattern. Six patients developed marked reductions of myocardial lactate extraction (decreases of more than 50% below baseline values), and three patients had net myocardial lactate produc- tion as a result of NaHCOs administration. Two of these three patients developed symptoms of angina. This is in contrast to the infusion of NaCl, in which lactate concentrations uniformly fell and none of the same patients developed net myocardial lactate pro- duction or symptoms.

Hemodynamics The administration of NaCl increased coronary si-

nus blood flow (p <O.OOl), as expected with volume expansion. However, this increase failed to occur with NaHCOs, and coronary sinus blood flow did not change during the infusion period but actually de- creased in the recovery period (p = NS). Coronary vascular resistance decreased with NaCl and did not change with NaHCOs. Mean coronary perfusion pres- sure did not change appreciably with either infusion (p = NS). The right atrial mean pressure, pulmonary artery mean pressure, pulmonary capillary wedge pressure, and stroke volume increased to the same degree with both infusions (p = NS). Overall, cardiac outputs did not change appreciably from baseline val- ues with either infusion (p = NS). However, four pa- tients developed transient reductions of cardiac out- put of more than 10% of baseline values for at least two consecutive time points, and two patients developed transient pump failure during NaHC03 administra- tion, with the cardiac index decreasing 25% to 50% below baseline values to outputs below 1.80 L/minute/

10 July 1989 The American Journal of Medicine Volume 87

SODIUM BICARBONATE IN HEART DISEASE / BERSIN ET AL

TABLE II

Complete List of Hemodynamic Values Measured In This Study

N&I InfusIons Heart rate MAP HRXBP Cl Strokevolume cs flow RA PAM PCW PVR SW

0 15 30 45 Time (minutes)

60 75 90 105 120

ii a2 70 89 8”; i:, tz iti fi: li;

“$Z 6,764 2.68 7.189 2.59 “iz 6,764 2.76 6,660 2.76 6.497 2.66 6,4:; 2.62 6,424 2.52 58.3 60.8 59.9 60.5 65.3 64.4 65.2 61.0 59.2

96 I07 ii4 112 107 105 106 101 ,109

2: 288 2; 2; 2: 2: 286 3; 3: 2:; 1;: 2:: 2;: 2:; 2:: 193 15 2:: 305 13

1,325 1,333 1,405 1.415 1,339 1,320 1,227 1,346 1,418

R = heart rate; BP = blood pressure; Cl = cardiac indp: CS = coronary sinus. Other abbreviations as in text. The means are shown. Time 0 reoresents the baseline values before the infusions were administered. None of the baseline values differed significantly from each

other for the two infusions.

HRXBP 7,811 7.754 7.835 7.704 7,546 7,269 7.555 7,603 756 Cl 2.72 2.66 2.77 2.92 2.71 2.63 2.85 2.60 2.87 Sirokevolurne 60.1 59.9 62.2 66.1 63.8 63.2 66.9 60.5 66.5 cs flow 123 125 135 123 122 119 122 111 113 RA 7 7 a 8 7 8 7 7 8 PAM 25 27 27 27 27 25 27 28 27 PCW 14 17 18 17 17 15 16 17 17 PVR 198 180 159 Ii1 163 191 175 209 lj3 SVR 1.488 1,480 1.416 1,302 1,487 1.468 1,405 1,520 1,410

m2 BSA. The heart rate and systemic vascular resis- tance did not change appreciably with either infusion (p = NS); however, pulmonary vascular resistance fell significantly with NaHCOa (p <O.O03). Mean arterial pressure also decreased slightly but nonetheless sig- nificantly with NaHCOa as compared with NaCl ad- ministration (p <0.05). Although the mean arterial pressure did decrease with NaHCOa, the overall (heart rate) X (blood pressure) product fell to the same de- gree with the two infusions (-226 with NaCl versus -226 with NaHCOs, p = NS). Thus, although signifi- cant hemodynamic changes were observed transiently in four of the 10 patients with NaHCOs administra- tion, these changes were generally not of sufficient magnitude to result in hemodynamic instability. All but two patients tolerated the procedures without ap- parent discomfort. A complete list of cardiac hemody- namic measurements is provided in Table II.

Plasma Electrolytes and Glucose Following NaHCOs infusion, the plasma sodium

changed from 135 to 138 mM, potassium from 3.9 to 3.6 mM, chloride from 102 to 98 mM, bicarbonate from 23.2 to 29.5 mM, and osmolality from 288 to 296 mOsm/kg HsO. With NaCl, the plasma sodium changed from 132 to 140 mM, potassium from 4.2 to 4.0 mM, chloride from 96 to 103 mM, bicarbonate from 24.0 to 23.1 mM, and osmolality from 277 to 287 mOsm/kg HsO. Only the changes in bicarbonate and chloride concentrations achieved significance when the two infusions were compared (p G.001).

The blood glucose concentration fell significantly with NaHCOs administration, but increased with NaCl (-1.7 versus +I..4 mmol/L, p <0.003) (Figure 5). Three patients developed frank hypoglycemia during the NaHCOs infusion period, with blood glucose val-

ues ranging from 1.2 to 2.9 mmol/L. This decrease occurred despite the administration of approximately 400 mL of 285 mmol/L glucose in water per infusion (OF 16 g glucose per hour) for the determination of cardiac outputs and coronary sinus blood flows.

COMMENTS These data demonstrate that the infusion of

NaHCOs reduces both the extraction and the con- sumption of oxygen by the heart in patients with con- gestive heart failure. In addition, systemic oxygen con- sumption and the arterial ~0, are reduced. As a consequence, blood lactate concentrations increase and myocardial ischemia may develop. Hemodynamic function may also be affected transiently in some pa- tients. These consequences of NaHCOs administra- tion are not due to alterations of extracellular volume, serum sodium level, or osmolality, as the NaCl infu- sions provided a paired control for each of these vari- ables. Such consequences were not seen with NaCl administration in the same patients.

The most striking finding in this study was the ob- servation that NaHCOs resulted in a significant de- cline of myocardial oxygen consumption. The mecha- nism by which NaHC($ decreased myocardial oxygen consumption was a primary reduction of myocardial oxygen extraction without a compensatory increase of coronary blood flow. Coronary sinus blood flow was measured in this study with the thermodilution tech- nique [z]. Although this technique does not provide absolute quantitation of coronary blood flow, it does furnish a reproducible estimate with which changes in coronary blood flow can be evaluated. Coronary sinus blood flow did not change. Normally, changes in myo- cardial oxygen demands are met by changes in coro- nary blood flow [33-353. Oxygen extraction is ordinari-

July 1989 The American Journal of Medicine Volume 87 11

SODIUM BICARBONATE IN HEART DISEASE / BERSIN ET AL

ly near maximal and normally varies little over a wide range of coronary blood flows [33-351. However, with the administration of NaHCOs, myocardial oxygen ex- traction was reduced primarily, and since coronary sinus blood flow did not increase in compensation, myocardial oxygen consumption correlated best with oxygen extraction rather than with blood flow (Figures 3 and 4).

Myocardial oxygen extraction may have been par- ticularly affected in our patient population for several reasons. First, all of our patients had symptomatic heart failure (NYHA class III or IV). Second, patients with symptomatic heart failure have, in general, high- er than normal resting myocardial oxygen require- ments, due in a large measure to ventricular enlarge- ment and increased wall stress [36,37]. Third, a shift of oxygen-hemoglobin binding affects oxygen extraction in the coronary circulation more than in other vascular beds. Oxygen extraction rates are highest in this vas- cular bed [33] and the oxygen-hemoglobin shift is uti- lized in the coronary circulation more than in any oth- er in order to ensure adequate oxygen delivery.

Last, red blood cell 2,3-DPG levels were elevated. The compound 2,3-DPG plays an important role in regulating oxygen-hemoglobin binding [38]. An in- crease in 2,3-DPG reduces oxygen-hemoglobin bind- ing and facilitates the delivery of oxygen to tissues. It is of interest that our patients had elevated 2,3-DPG levels (mean, 26.4 pmmol/g hemoglobin; normal, 13.1 f 2.4 pmmol/g hemoglobin, p <O.OOl). Indeed, this increase in 2,3-DPG reduced oxygen-hemoglobin binding as measured by the Pss. The Psa is a direct measure of oxygen-hemoglobin binding and repre- sents the PO:! at which blood is 50% saturated with oxygen [29,30]. The higher the value of Psc, the less saturated the blood is with oxygen at a given ~02, and thus the less avidly hemoglobin binds oxygen. The baseline Psc was significantly elevated in our patients. It would thus appear that 2,3-DPG is increased in congestive heart failure as a compensatory mechanism to increase tissue oxygen delivery to meet normal rest- ing oxygen demands. The same phenomenon has been described in children with congenital heart disease and impaired tissue oxygenation [39]. In our patients with heart failure, the administration of NaHCOs re- duced Pss and shifted the oxygen-hemoglobin disso- ciation curve toward normal, removing this compensa- tory mechanism and reducing tissue oxygen availability to the point at which even resting oxygen demands could no longer be met. As a consequence, transient anaerobic metabolism occurred.

A reduction of myocardial oxygen consumption would have been appropriate if the workload being performed by the heart, and thus the demands for oxygen, decreased commensurately [40,41]. We mea- sured two of the major determinants of myocardial oxygen demand: the heart rate and the arterial pres- sure [40-42]. The heart rate-blood pressure product did not change with the two infusions. Of the determi- nants not measured, left ventricular volume would have been expected to increase with both infusions as a result of the volume load and sodium administered. An increase of left ventricular volume would have in- creased rather than decreased oxygen demands. If the decrease in myocardial oxygen consumption matched a reduction in oxygen demand, then the other major determinant not measured, left ventricular contractil-

12 July 1989 The American Journal of Medicine Volume 87

ity (dP/dt), must have decreased. Thus, either myo- cardial oxygen supply and consumption decreased with NaHCOs administration inappropriately at a time when oxygen demands most probably remained constant, or oxygen demands actually decreased be- cause of a reduction of myocardial contractility.

In any case, anaerobic metabolism became neces- sary in order to maintain resting basal hemodynamic function. The fact that three of 10 patients receiving NaHCOs had net myocardial lactate production and two of these three developed angina supports the no- tion that transient anaerobic metabolism occurred. The decrease in blood glucose concentrations with NaHCOs administration also supports this observa- tion, as the blood glucose should have increased if oxygen consumption fell as glucose was administered. The fact that glucose concentrations decreased sug- gests that either anaerobic glycolysis was stimulated [43], gluconeogenesis was impaired, or insulin levels were increased (or anti-insulin levels decreased) as a result of the alkalemia induced.

Systemic oxygen consumption was also significantly reduced with NaHCOs administration by the same mechanism. The cardiac index was generally un- changed (i-70 ml/minute/m*, p = NS), and thus flow was not affected substantially in most instances. The principal effect of NaHCOs was on systemic oxygen extraction. Mixed Venus oxygen contents increased at a time when blood flow did not change. Evidence for anaerobic metabolism was also present systemically, as arterial lactate concentrations rose uniformly in all patients (Figure 6).

It could be argued that the decrease in oxygen ex- traction and consumption was due entirely to a shift of oxygen-hemoglobin binding caused by the alkalosis per se, in so far as the arterial pH increased out of the normal range with NaHCOs administration (from 7.43 to 7.51). Indeed, the Pss fell in every patient. However, the shift was less than anticipated and individual changes of the Pr,s did not correlate directly with indi- vidual changes of either myocardial or systemic oxy- gen extraction (r = 0.27 and r = 0.01, respectively, p = NS). This effect does not appear to be related to any alteration of 2,3-DPG levels either, as these levels did not change significantly with either infusion.

Circulatory shunting or an impairment of oxygen use at a cellular level may have also occurred; however, the relative contributions of each of these to the im- pairment of oxygen use are not known. The intracellu- lar pH decreases with NaHCOs infusion in experimen- tal hypoxic lactic acidosis as a result of both increased carbon dioxide generation and intracellular lactate ac- cumulation [15,16]. Such may also be the case in our patients with heart failure, as arterial and mixed ve- nous pCO2 levels increased with NaHCOs administra- tion. Nonetheless, a direct link between carbon diox- ide generation, intracellular acidosis, and reduced oxidative phosphorylation cannot be confirmed with the available data.

An increase of oxygen-hemoglobin binding with NaHCOs also cannot explain the fall of the arterial ~0s. Increased oxygen-hemoglobin binding, to the de- gree seen in this study with NaHCOs administration, should either increase or not change the arterial pop [38]. Since the pulmonary capillary wedge pressure increased to the same degree with both infusions and the cardiac index did not change, neither alterations of

SODIUM BICi\RBONATE IN HEAtiT DISEASE / BERSIN ET AL

pulmonary blood flow nor alterations of pulmonary venous pressure explain this phenomenon. Most Iike- ly, the explanation lies with the development of intra- pulmonary circulatory shunts. The arterial pCOs rose as expected to a level appropriate for the amount of NaHCOs administered [13,44], suggesting that minute ventilation probably remained constant. Since hypox- ia: developed at a time when ventilation was presum- ably constant and pulmonary blood flow was not changing, intrapulmonary shunting is suggested [45].

The lack of an increase in coronary sinus blood flow with NaHCOs administration is of interest. The ef- fects of alkalinization on coronary blood flow by means of increasing the bicarbonate concentration in animal experiments in oiuo have generally showed in- creases in coronary blood flow [20,46-531. However, increases in coronary blood flow with NaHCOs are not always seen [52], and Scheuer [53] demonstrated that coronary blood flow increases only if the blood pCOs is raised in the process: if the increase of the blood pCOr with NaHCOs administration is counteracted by si- multaneous hyperventilation to keep the pCOz con- stant, then coronary blood flow does not change de- spite an even more marked increase of the blood pH. Although the arterial and mixed venous pCOz rose with NaHCOs administration in the present study, the absolute magnitude of the increase was small (less than 4 mm Hg) relative to the magnitude of the blood pH change (+0.08 pH units). Thus, our current obser- vation that coronary sinus blood flow in humans did not change corroborates the finding in animal studies that the blood pCOs, and not the pH, regulates coro- nary blood flow during alkalinization with NaHC03.

It has been postulated that coronary blood flow de- creases reflexively in response to primary reductions in myocardial oxygen demands 1331. However, in the present study, oxygen demands did not appear to change. Rather, oxygen supply and extraction were affected primarily. This effect also appears to not be related to the serum potassium, as serum potassium levels were reduced to the same degree with both infu- sions.

The elevation of the blood lactate concentration ap- pears to be secondary to impaired tissue oxygenation and resultant anaerobic metabolism rather than to be an effect of alkalosis per se. There are several mecha- nisms by which the blood lactate concentration may increase during alkalosis, including volume contrac- tion, hyperventilation, liver disease, and increased ac- tivity of the glycolytic enzyme phosphofructokinase (PFK) [54,55]. Our patients had volume expansion, exhibited normal ventilation, and did not show evi- dence of liver disease. An increase in PFK activity would have occurred if intracellular pH increased [55]. However, we do not know whether intracellular pH changed in our patients with heart failure in this study. To the contrary, experimental studies have demonstrated a significant decrease of intracellular pH when NaHCOs is administered to dogs with hyp- oxic lactic acidosis [ 161, or to lambs with cerebral isch- emia [56]. Nonetheless, any effect on PFK activity would appear to have been overshadowed by the sub- stantial changes observed in oxygen consumption, The development of an oxygen debt is also supported by the occurrence of net myocardial lactate production in 30% of our patients [57].

Curiously, such problems are generally not encoun-

tered when NaHCOs is administered to patients un- dergoing hemodiaiysis, or to patients with renal tubu- lar acidosis or diarrhea1 conditions. In these instances, hypoxemia and circulatory insufficiency are generally not present and ventilatory capacities are normal. These patients are thus able to compensate for any effect NaHC03 may have on oxygen delivery to tis- sues. However, the occasional patient who develops hypoxemia during bicarbonate hemodialysis may in fact have heart disease with limited cardiac reserve ]531.

We therefore conclude that the administration of NaHCOs to patients with heart failure results in pro- found alterations of tissue oxygen extraction, which in turn decrease oxygen consumption of both the coro- nary and systemic vascular beds below basal levels of oxygen demand. These alterations result in anaerobic metabolism both in the coronary and in the systemic circulations. The mechanisms by which oxygen extrac- tion is altered are complex and vary in different re- gional vascular beds. In the coronary circulation, myo- cardial oxygen extraction is affected primarily as a result of increased oxygen-hemoglobin binding and perhaps circulatory shunting or an impairment of oxy- gen use at the cellular level. The systemic circulation is affected similarly. However, in the pulmonary circula- tion, circulatory shunting appears to be the primary mechanism. Regardless of the cause, the coronary cir- culation appears to be the most profoundly affected and the consequence may frequently be myocardial ischemia. Thus, NaHCOs has potentially deleterious consequences when administered to patients with heart failure. These findings also suggest that other patients critically dependent on tissue oxygen trans- port to meet oxygen demands, such as children with congenital heart disease 139) and adults with acute myocardial infarction or cardiac arrest [l,ll,lZ], are likely to manifest similar responses to NaHCOs ad- ministration.

ACKNOWLEDGMENT We thank Kevrn Tanaka. B.S.. Anne Horton. B.S.. and Phitip Samacki. B.S.. for thetr technical assistance. Janet Collins, B.S., for the performance of blood gas analy- ses, and all of the nurstng staff of the Moffdt Hospital Coronary Care Unrt for their clinical assistance wrth the performance Of thus study.

REFERENCES 1. Filmore SJ, Shaprro M, Killip T: Serial blood gas studres during cardropulmonary resuscitation, Ann Intern Med 1970; 72: 465469. 2. Park R. Arieff Al: Lactic acidosis: current concepts. Ckn Endocrinol Metab 1983; 12: 339-358. 3. Cohen RD. Woods HF: Clinical and biochemical aspects of lactic acidosis. Ox- ford: Blackwell Scientific Publications, 1976: 10-76. 4. Cohen RD. Woods HF: Lactic acidosis revistted. Diabetes 1983: 32: 181-191. 5. National Conference on Cardiopulmonary Resuscitation: Standards and guide- lknes for cardiopulmonary resusitation (CPR) and emergency cardiac care (ECC). Par! fff: Adult advanced cardiac life support. JAMA 1986; 2.55: 2933-2954. 6. Stacpoole PW: Lacbc acidosis: the case against bicarbonate therapy. Ann Intern Med 1986; 105: 276279. I.Graf H. Arieff Al: The use of sodium bicarbonate in the therapy of organic acidosis. Intensive Care Med 1986: 12: 1-4. 8. Madias NE: Lactic acidosis, Kidney Int 1986; 29: 752-774. 9. Narins R. Cohen JJ Bicarbonate therapy for organic acidosis: the case for its continued use. Ann intern Med 1987: 106: 615-618. 10. Fraley OS, Adler S. Bruns FJ. Zett B: Stimulation of lactate productton by admmrslration of brcarbonate in a patient wdh a solid neoplasm and lactic acidosts. N Engl J Med 1980; 303: 1100-1102. 11. Mattar JA. Weil MH. Shubin H. Stein L: Cardtac arrest in the crdically ill, II. Hyperosmolar states following cardiac arrest. Am J Med 1974: 56: 162-168. 12. Bishop RL. Weisfeldt ML: Sodium bicarbonate administration during cardiac

July 1989 The American Journal of Medicine Volume 87 13

SODIUM BICARBONATE IN HEART DISEASE / BERSIN ET AL

arrest. Effect on arterial pH, pCOa, and osmolality. JAMA 1976; 235: 506-509. 13. Ostrea EM, Odell GB: The influence bf bicarbonate administration on blood pH in a “closed system”: clinical implications. J Pediatr 1972; 80: 671680. 14. Singer RB. Deering RC, Clark JK: The acute effects in man of a rapid intrave- nous infusion of hypertonic sodium bicarbonate solution. II. Changes in respiration and output of carbon dioxide. J Clin Invest 1956; 35: 245-253. 15. Arieff Al, Park R, Leach W. Lazdrowitz VC: Systemic effects of NaHCOs in experimental lactic acidosis in dogs. Am J Physiol 1982; 242: F586-F591. 16. Graf H. Leach W, Arieff Al: Metabolic effects of sodium bicarbonate in hypoxic lactic acidosis in dogs. Am J Physiol 1985; 249: F630-F635. 17.Arieff Al, Graf H: Pathogenesis of hypoxic lactic acidosis in the dog. Am J Physiol 1987; 253: E271-E276. 18. Weil MH. Rackow EC, Trevino R, et at Difference in acid-base state between venous and arterial blood during cardiopulmonary resuscitation. N Engl J Med 1986: 315: 153-156.

19. Androgue HJ, Rashad MN, Gorin AB. eta/: Assessing acid-base status in circula- tory failure. Differences between arterial and central venous blood. N Engl J Med 1989: 320: 1312-1316. 20. Ng ML, Levy MN, Zieske HA: Effects of changes in pH and of carbon dioxide tension on left ventricular performance. Am J Physiol 1967; 213: 116-120. 21. Clancy RL. Cingolani HE, Taylor RR, Graham TP, Gilmore JP: Influence of sodium bicarbonate on myocardial performance. Am J Physiol 1967; 212: 917- 923. 22. Weisfeldt ML, Bishop RL. Greene HL: Effects of pH and pCOa on performance of ischemic myocardium. In: Roy PE, Harris P, eds. Recent advances in studies on cardiac structure and metabolism. Baltimore: Uhiversity Park Press, 1975; 355- 364. 23. Graf H, Leach W. Arieff Al: Evidence for a detrimental effect of bicarbonate therapy in hypoxic lactic acidosis. Science 1985; 227: 754-756. 24. Tattersfield AE, McNicol MW, Sillett RW: Relationship between hemodynamic and respiratory function in patients with myocardial infarction and left ventricular failure. Clin Sci (Oxford) 1972; 42: 751-768. 25. Bigger JT: Why patients with congestive heart failure die: arrhythmias and sudden cardiac death. Circulation 1987: 75(suppl IV): IV-28-IV-30. 26. Criteria Committee, New York Heart Association, Inc: Diseases of the heart and blood vessels. Nomenclature and criteria for diagnosis, 6th ed. Boston: Little, Brown, 1964; 114. 27. Ganz W, Donose R, Marcus HS, et at A new technique for measurement of cardiac output by thermodilution in man. Am J Cardiol 1971; 27: 392-396. 28. Ganz W. Tamura K, Marcus HS. Donose R, Yoshida S, Swan HJC: Measurement of coronary sinus blood flow by continuous thermodilution in man. Circulation 1971; 44: 181-195.

29. Severinghaus JW: Blood gas calculator. J Appl Physiol 1966; 21: 1108-1116. 30. AJtman PL. Dittmer DS: Blood oxygen dissociation curves: mammals. In: Biolo- gy data book, 2nd ed. vol Ill. Bethesda: Federation of American Societies for Experimental Biology, 1974; 1863-1871. 31. Rose ZB, Leibowitz J: Direct determination of 2.3-diphosphoglycerate. Anal Biochem 1970; 35: 177-180. 32. Sigma Technical Bulletin, 726/826-UV-10-68, p. I-10. 33. Marcus ML: The coronary circulation in health and disease. New York: McGraw-Hill, 1983; 65-92. 34. Alella A, Williams FL, Williams CB, Katz LN: Interrelation between cardiac oxy- gen consumption and coronary blood flow. Am J Physiol 1955; 183: 570-582. 35. Lombard0 TA, Rose L, Taeschler M, Tukiy S, Bing RJ: The effect of exercise on coronary blood flow, myocardial oxygen consumption and cardiac efficiency in

man. Circulation 1953; 7: 71-78. 36. Strauer BE: Myocardial oxygen consumption in chronic heart disease: role of wall stress, hypertrophy and coronary reserve. Am J Cardiol 1979: 44: 730-740. 37. DeMarco T. Chatterjee K, Rouleau J-L, Parmley Ww: Abnormal coronary he- modynamics and energetics in patients with chronic heart failure caused by isch- emit heart disease and dilated cardiomyopathy. Am Heart J 1988; 115: 809-815. 38. Bellingham Al, Detter JC, Lenfant C: Regulatory mechanisms of hemoglobin- oxygen affinity in acidosis and alkalosis. J Clin Invest 1971; 50: 700-706. 39. Berman W, Wood SC, Yabek SM, Dillon T. Fripp RR, Burstein R: Systemic $gc3;$snsport in patients with congenital heart disease. Circulation 1987; 75:

40. Hoffman JIE, Buckberg GD: The myocardial supply:demand ratio-a critical review. Am J Cardiol 1977: 41: 327-332. 41. Weber KT. Janicki JS: The metabolic demand and oxygen supply of the heart: physiologic and clinical considerations. Am J Cardiol 1979; 44: 722-729. 42. Braunwald E: Control of myocardial oxygen consumption. Physiologic and clinlcal considerations. Am J Cardiol 1971; 27: 416-432. 43. Coleman JE: Anaerobic glycolysis: In: Bondy PK, Rosenberg LE, eds. Metabolic control and disease. Philadelphia: WB Saunders, 1980; 163-176. 44. Forster RE, Crandall ED: Time course of exchanges between red cells and extracellular fluid during COa uptake. J Appl Physiol 1975; 38: 710-718. 45. Klocke RA: Arterial blood gases. In: Baum GL, Wolinsky E, eds. Textbook of pulmonary diseases. Boston: Little Brown, 1985; 117-143. 46. Mehmul HC, Duvelleroy MA, Laver MB: Response of coronary blood flow to pH- induced changes in hemoglobin-Da affinity. J Appl Physiol 1973: 35: 485-489. 47. Elek SR. Katz LN: Further observations on the action of drugs on coronary vessel caliber. J Pharmacbl.Exp Ther 1942; 75: 178-182. 48. Feinberg H. Gerola A, Katz LN: Effect of changes in blood COa on coronary flow and myocardial 02 consumption. Am J Physiol 1960; 199: 349-354. 49. Goodyer AVN. Eckhardt WF, Ostberg RH, Goodkind MJ: Effects of metabolic acidosis and alkalosis on coronary blood flow and myocardial metabolism in the intact dog. Am J Physiol 1961; 200: 628-632. 50. Rowe GG, Castillo CA, Crumpton CW: Effects of hyperventilation on systemic and coronary hemodynamics. Am Heart J 1962; 63: 67-77. 51. Rowe GG, McKenna DH, Jaramillo CV, Crumptori CW: The systemic and coro- nary hemodynamic effect of sodium bicarbonate. Am J Med Sci 1964; 248: 426 428. 52. Wang H, Katz RL: Effects of changes in corona’ry blood pH on the heart. Circ Res 1965; 17: 1,14-122. 53. Scheuer J: The effectsof respiratory and metabolic alkalosison coronary flow, hemodynamics and myocardial carbohydrate metabolism. Cardiologia 1968; 52: 275-286. 54. Bersin RM, Arieff Al: Primary lactic alkalosis. Am J Med 1988; 85: 867-871. 55. Relman AS: Metabolic consequences of acid-base disorders. Kidney Int 1972; 1: 347-359. 56. Hope PL, Cady EB. Delpy DT, Ives NK, Gardiner RM, Reynolds EOR: Brain metabolism and intracellular pH during ischemia: effects of systemic glucose and bicarbonate administration studied by 31P and 1H nuclear magnetic resonance spectroscopy in the lamb. J Neurochem 1988; 50: 1394-1402. 57. Gertz EW. Wisneski JA, Neese R. Houser A, Korte R, Bristow JD: Myocardial lactate extraction: multidetermined metabolic function. Circulation 1980: 61: 256-261. 58. Wathen RL. Ward RA: The dialysis patient: acid-base disorders and hypoxemia. In: Massry SG, Glassock RI, eds. Textbookof nephrology, vol II. Baltimore: Williams and Wilkins, 1989; 1273-1277.

14 July 1989 The American Journal of Medicine Volume 87