- Email: [email protected]
Gas exchange during dialysis
Gas Exchange during Dialysis Contrasting Mechanisms Contributing to Comparable Alterations with Acetate and Bicarbonate Buffers
JUDSON M. HUNT, M.D.’ TIMOTHY R. CHAPPELL, M.D. WILLIAM L. HENRICH, M.D. LEWIS J. RUBIN, M.D. Dallas,Texas
From the Department of Internal Medicine, Pulmonary and Nephrology Dlvisions, University of Texas Health Science Center, Southwestern Medical School, and Veterans Administration Medical Center, Dallas, Texas. This work was supported in part by Eli Lilly and Company, Erika Corporation, and the Medical Service of the Veterans Administration Medical Center, Dallas, Texas. Requests for reprints should be addressed to Dr. Lewis J. Rubin, Pulmonary 11 lF, Veterans Administration Medical Center, 4500 South Lancaster Road, Dallas, Texas 75216. Manuscript accepted February 9, 1984. Current address: University of Texas Health Science Center, Nephrology Section, 7703 Floyd Curl Drive, San Antonio, Texas 78229. l
Although arterial hypoxemia during hemodiaiysis is common and may contribute to dialysis morbidity, the mechanisms responsible remain uncertain. Additionally, controversy exists as to whether bicarbonate diaiysate produces less hypoxemia than acetate dialysate. The short- and long-term effects of acetate dialysate and bicarbonate dfaiysate on gas exchange were compared in eight stable patients undergoing dialysis using a closed, proportioning system and a double-blind, crossover study destgn. Diaiysate was sampied immediateiy proximal and distal to the diaiyzer to determine its contribution to total carbon dioxide elimination. Ventilatory parameters and blood gas values were measured before dialysis, at one hour, and after dialysis. Arterial oxygen tension fell signpicantiy and comparably at one hour with both diaiysates, whereas the aiveoiar-arterial oxygen gradient tncreased only silghtiy. Despite hypoxemia, minute ventilation decreased by 4 to 18 percent, and arterial carbon dioxide tension was unchanged. Aithough total carbon dioxide elimination was unchanged in all groups, there was a significant decrease in lung total carbon dioxide eiiminatlon with acetate diaiysate of 9.23 f 2.89 to 7.74 f 1.57 mmoi per minute on Day 1 (mean f SD, p <0.025) concomitant with a loss of total carbon dioxide into the bath of 2.04 f 0.20 mmoi per minute, resuiting in a significant reduction in respiratory quotient (0.92 f 0.07 to 0.75 f 0.05, p KO.01). in contrast, there was a gain of total carbon dioxide into the blood of 1.84 f 0.45 mmoi per minute with bicarbonate diaiysate, which resulted in an increased pH at one hour compared with acetate diaiysate (7.39 f 0.04 versus 7.35 f 0.03, p <0.05). Hypoxemia persisted after dialysis in ail groups and was associated with an increased alveolar-arterial oxygen gradient in three of the four groups. it is concluded that transitory hypoventiiation contributes to comparable hypoxemia with both acetate and bicarbonate diaiysates by different mechanisms. With acetate diaiysate, there is a decrease in carbon dioxlde load to the lungs, whereas with bicarbonate diaiysate, the mechanism responslbie appears to be a suppression of respiratory drive resuittng from a gain of bicarbonate from the diaiysate. Additionally, neither dialysate prevents post-dialysis hypoxemia, which is associated with an increased alveolar-arterial oxygen gradient resulting from a mechanism that remains to be elucidated.
August 1984
The American Journal of Meclklne
Volume 77
255
HYPOXEMIA DURING DIALYSIS-HUNT
TABLE I
ET AL
Dialysate Composition Acetate 136.8 f 108.8 f 2.08 f 2.2 f 36’ 7.0 f 1.7 f 204.0 f
Sodium (meqlliter) Chloride (meq/liter) Potassium (meqlliter) Bicarbonate (meq/liter) Acetate (meqlliter) Calcium (meq/liter) Magnesium (meq/liter) Dextrose (mg/dl)
Bicarbonate
0.6 0.3 0.05 0.2
135.0 106.4 2.12 33.2
0.04 0.1 1.0
6.26 f 0.09 1.6 f 0.0 203.0 f 2.4
Not measured; value given is from manufacturer’s tions. l
f 1.1 f 4.0 f 0.04 f 0.6
specifica-
Arterial hypoxemia can be a significant contributing factor to morbidity during acetate hemodialysis. Although the pathogenesis of dialysis-induced hypoxemia remains unclear [l-4], several possible explanations have been proposed, including ventilation-perfusion mismatching resulting from complement-mediated leukostatic plugging of the pulmonary vasculature [ 11, an increased affinity of hemoglobin for oxygen (Bohr effect) induced by increasing blood pli during dialysis [5], and alveolar hypoventilation. Alveolar hypoventilation seems particularly plausible since the alveolararterial oxygen gradient does not increase with hypoxemia that develops early in dialysis [6,7]. Hypoventilation has been suggested to be due to decreased
delivery of carbon dioxide to central chemoreceptors resulting from either a loss of carbon dioxide into the dialysate [8] or to a decreased rate of total body carbon dioxide production as the result of acetate metabolism [9], both resulting in a decreased respiratory quotient. Since the metabolic quotient may be different from the respiratory quotient if lung carbon dioxide elimination is not equal to total carbon dioxide production, measurement of carbon dioxide loss in dialysate is crucial to discerning the mechanism responsible. The use of a bicarbonate dialysate in lieu of acetate has been found to reduce the incidence of hypoxemia by some investigators [ 8, lo] but not by others [ 11,121.
TABLE II
Dialysate Gas Values before and after Dialyrer with Saline Recirculation* Oxygen (torr)
Carbon Dloxide (torr)
0.012 0.016
111.8f8.8 113.2 f 8.4
84.9 f 5.6 82.7 f 6.0
0.036 0.036
105.5 f 7.6 106.9 f 6.0
PH Bicarbonate dialysate Before 7.124 f After 7.127 f Acetate dialysate Before 6.939 f After 6.940 f l
Sixteen observations,
256
August 1984
each group.
7.9 f 7.9 f
1.2 1.2
The issue of which dialysate buffer produces the least hypoxemia is of clinical importance, particularly in patients with a marginal cardiopulmonary status. Accordingly, the present study was designed to evaluate the effects of both dialysates on gas exchange and to clarify the mechanisms responsible for the alterations that have been observed. PATIENTS AND METHODS Patient Population. Eight stable patients undergoing long-term hemodialysis were included in the study after giving informed consent. The mean age of the patients was 55.5 f 7.6 years and the mean duration of dialysis was 55.2 f 39.6 months. The cause of kidney failure was hypertensive nephrosclerosis in three, and diabetic nephropathy, chronic glomerulonephritis, interstitial nephritis, recurrent nephrolithiasis, and unknown in one patient each. The average interdialytic weight gain was 2.32 f 0.71 kg. Each patient participated in both dialysate protocols, which lasted two weeks each and consisted of five-hour periods of dialysis three times per week. The same dialyzer (Organon-Technica, 1 m*, 220 ml per minute average blood flow) and closed delivery system (Drake-Willock proportioning system), modified for either acetate or bicarbonate concentrates, was utilized in each patient. Dialysis Protocol. The dialysis protocol consisted of a two-week course (six treatments) employing either an acetate or bicarbonate dialysate. The dialysate was sampled immediately proximal to the dialyzer, and its composition was verified by laboratory analysis (Table I). Aside from the presence of either acetate or bicarbonate, no significant differences in dialysate composition were noted. Four patients initially began with acetate dialysis and were switched to bicarbonate dialysis at the end of the two-week period, whereas the other four patients began with bicarbonate and switched to acetate dialysis in a similar fashion. Neither the patients nor the dialysis medical staff were aware of the dialysate composition or crossover time. Predialysis weight and blood pressure were measured and arterial and venous lines were inserted in the usual fashion in preparation for
hemodialysis in each patient. Gas Exchange Studies. Short-term studies were performed on the first day of dialysis using either acetate or bicarbonate dialysate, and long-term studies were performed during the sixth dialysis treatment. Timed collections of exhaled gas were obtained before dialysis, after the first hour of dialysis, and immediately following dialysis after the patient’s blood had been returned. The oxygen and carbon dioxide contents of exhaled gas were measured using a mass spectrometer (Perkin Elmer MGA 1100, Pomona, California), and the volumes were measured in a Tissot gasometer (W. Collins, Braintree, Massachusetts). From these data, minute ventilation, tidal volume, ratio of dead space to tidal volume, total body oxygen consumption, carbon dioxide elimination, respiratory quotient, and alveolar-arterial oxygen gradient were calculated using the standard formulas [ 131. Blood was collected anaerobically for arterial blood gas measurements at times corresponding to ventilatory measurements. Diafysate gas tensions, pH, and bicarbonate concentration were
The American Journalof Medicine Volume 77
HYPDXEMIA DURING DIALYSIS-HUNT
ET AL
Irsa
Figure 1. The effects of short- and long-term acetate (open and closed circk?, respective&) and bizarbonate(open and closedsqvares) henwdialysis on gas exchange. Pa4 = arterial oxygen tension; A-a02D = alveolar-arterial oxygen gnradient; pace* = arterial carbondioxide tension: VE = minute ventilation. Measurements were madebefore (pre),at one hour, and a&r (post)dialysis ( l p I 0.025, 7 p 10.05 compared with predialysis values).
tII II
\iE 8L/min
76-
1111
54
1 ;IO”R
t
measured from samples anaerobically drawn proximal and distal to the dialyzer after the first hour of dialysis. Arterial blood gas and diilysate measurements were performad using an IL 813 blood gas analyzer (Instrumentation Laboratory, Lexington, MA). Prior to the initiation of dialysis, dialysate samples were collected proximal and distal to the dialyzer while saline se lution was circulating through the dialysis machine (Table II). From the dialysate data, obtained at one hour, the amount of total carbon dioxide lost or added was calculated using a modification of the Fick equation: carbon dioxide uptake or elimination by dialysis equals the dialysate flow rate multiplied by the carbon dioxide difference across the dialyzer (calculated by subtracting the total carbon dioxide content of dialysate proximal to the dialyzer from the carbon dioxide content of fluid distal to the dialyzer). Total carbon dioxide was calculated by multiplying the arterial carbon dioxide tension by 0.03 and then adding the bicarbonate value; it is expressed in mmol/llter. In order to ensure accurate dialysate flow rate measurements, timed collections of dialysate were performed at least twice during each five-hour dialysis period using a graduated cylinder. The data are presented as the
P&ST
mean f SD. Statistical analysis was performed using a two-way analysis of variance for sequential data. Linear regression analysis was performed using the least-squares method. RESULTS The arterial blood gas measurements for the short- and long-term studies with both dialysates are shown in
Figure 1. With short-term acetate dialysate, the PaOp fell from 92 f 8 to 79 f 7 mm Hg at one hour, and remained at 80 f 5 mm Hg after dialysis. A similar decline in arterial oxygen tension was observed with short-term bicarbonate dialysate, from 98 f 10 to 85 f 8 mm Hg at one hour and to 8 1 f 10 mm Hg immediately after dialysis. There were no differences in oxygen tension between the groups at each interval. The arterial oxygen tension was also significantly decreased both at one hour and after hemodialysis in the long-term acetate dialysate group. With long-term bicarbonate dialysate, in contrast, there was a gradual
August 1984
The American Journal of Medicine
Volume 77
257
HYPOXEMIA
DURING DIALYSIS-HUNT
TABLE III
ET AL
Arterfal Blood pH and Bicarbonate and Carbon Dioxide Lost from Blood to Dlalysate or Gained from
Dlalysate to Blood (at one hour) during Hemodlalysls
Before Acetate Day 1 Day6 Bicarbonate Day 1 Day 6 ’
pH OneHour
After
Before
Bicarbenate (meq/llter) One Hour
After
7.35 f 0.04 7.33 f 0.03
7.35 f 0.03 7.34 f 0.03
7.39 f 0.03’ 7.39 f 0.03+
19.0 f 2.5 16.4 f 1.8
18.0 f 2.4 16.5 f 1.4
19.7 f 19.8 f
1.6 1.9’
7.35 f 0.04 7.36 f 0.03
7.39 f 0.04 7.39 f 0.03’S
7.46 f 0.03++ 7.45 f 0.03++
17.1 f 3.8 18.8 f 2.1
21.2 f 3.6’+ 21.4 f 2.0’+
22.8 f 2.7++ 23.3 f 2.8++
CarbonDioxide Change (mmol/mlnutet
-2.04 -1.88
f 0.20 l 0.19
+1.64 f 0.45 +1.21 f 0.68
p < 0.05, + p < 0.01 compared with predialysis values; + p < 0.05 compared with both acetate dialysis studies.
decline in arterial oxygen tension, which was different from control only after dialysis. As in the long-term studies, there were no differences in arterial oxygen tension between the groups at each interval. Despite arterial hypoxemia. the alveolar-arterial oxygen gradient at one hour was not statistically different from the predialysis values in any of the groups, although there was a trend for the alveolar-arterial oxygen gradient to increase slightly in three of the four groups. In contrast, the post-dialysis alveolar-arterial oxygen gradient was significantly higher compared with predialysis values in three of the four groups. The ratio of dead space to tidal volume was normal in all groups (range 0.34 to 0.42) and was not significantly changed throughout the study. Arterial carbon dioxide tension increased slightly with short- and long-term bicarbonate dialysate (31.5 f 5.2 to 36.1 f 5.0 mm Hg, p = 0.06; and 34.7 f 2.8 to 36.8 f 2.9 mm Hg, p = O.l), but there were no differences in arterial carbon dioxide tension between the groups at each interval. Despite hypoxemia, minute ventilation fell slightly in all groups. The change in minute ventilation at one hour was significant, however, only on Day 6 with acetate dialysate. No significant differences in minute ventilation were observed between acetate and dialysate buffers at any time period. The effects of acetate and bicarbonate dialysates on blood pH and bicarbonate values are shown in Table III. Blood pH increased steadily with both buffers and was significantly higher after dialysis compared with predialysis values. Additionally, pH and bicarbonate were significantly higher both at one hour and after dialysis in the bicarbonate dialysate groups compared with the acetate dialysate groups. Oxygen consumption and total carbon dioxide elimination were comparable in all groups and unchanged by dialysis (Figure 2). Respiratory carbon dioxide elimination decreased at one hour with acetate dialysate, in both short- and long-term studies, concomitant with a loss of total carbon dioxide into the dialysate. As
258
August 1984
The American Journal of Medicine
a result, the respiratory quotient dropped significantly at one hour, whereas the metabolic quotient remained unchanged. COMMENTS Routine hemodialysis is associated with a variety of adverse effects including hypotension, hypoxemia, muscular cramps, nausea, and vomiting. Hypotension is particularly troublesome since it hinders the adequate removal of fluid and, if symptomatic, may limit tolerance of dialysis. The hypoxemia of dialysis generally occurs early [l] (usually within 30 to 60 minutes after dialysis commences) and may contribute to hypotension by impairing myocardial function or inducing arrhythmias, particularly in patients with underlying cardiopulmonary disease. Accordingly, any maneuver that would reduce the incidence or severity of hypoxemia and hypotension might be expected to improve patient tolerance of dialysis. The use of bicarbonate dialysate achieved an early measure of popularity because it was found to reduce the incidence and severity of hypoxemia and adverse symptoms compared with acetate buffer [3,8]. Subsequent reports, however, noted that hypoxemia also occurs with bicarbonate dialysate, leaving this issue unsettled [ 11,121. Additionally, the mechanisms responsible for the changes in ventilatory function have not been clearly elucidated. The present study demonstrates that significant and comparable degrees of hypoxemia occur at one hour of either acetate or bicarbonate dialysis and that hypoventilation, produced by different stimuli, contributes to the observed reductions in arterial oxygen tension. Although the alveolar-arterial gradient was increased slightly at one hour of dialysis, it is unlikely that abnormalities in gas exchange that increase the alveolararterial oxygen gradient are solely responsible for hypoxemia at one hour, since there was not a close correlation between the increase in the alveolar-arterial oxygen gradient and the decrease in arterial oxygen
Volume 77
HYFQXEMIA DURING DIALYSIS-HUNT
ET AL
1.2&
14
1.1.
131
LO0.9lo-
0.7-
9-
9'/co,
0.8-
L
8-
%,
mmol/min
mmol/min
8-
RQ
0.60.5-
7-
7-
6-
6-
0.4-
5-
5-
0.3-
4-
4-
0.2-
3-
3-
0.1-
/
e
e
ACETATE
BICARB
-
ACETATE
BICARB
ACETATE
BICARB
Figure 2. The effects of short-term acetate or bicarbonate dialysis on total carbon dioxide elimination (i’C02), oxygen consumption N02), and respiratory quotient (RQ). Measurements were made before dialysis, at one hour, and after dialysis. Solid area denotes carbon dioxide elimination by dialyzer ( l p I 0.05 compared with pre-dialysis value).
tension (r = -0.63). In contrast, a closer inverse correlation was observed between the alveolar-arterial oxygen gradient and the arterial oxygen tension after dialysis (r = -0.80) at a time when hypoxemia persisted and minute ventilation had returned to baseline levels, implying that other factors contribute to postdialysis hypoxemia. Our observations are most compatible with the following mechanisms of hypoxemia. During acetate hemodialysis, a loss of total carbon dioxide, associated with an unchanged total body carbon dioxide production, results in a decreased carbon dioxide load to the lungs, promoting relative alveolar hypoventilation [ 141 without changing arterial carbon dioxide tension. In contrast, a net gain of total carbon dioxide from the bath associated with significant increases in blood pH and bicarbonate was the stimulus for hypoventilation [ 151 during bicarbonate dialysis. It is of interest that the hypoxemia persisted after both acetate and bicarbonate dialysis treatments were discontinued and was associated with an increased alveolar-arterial oxygen gradient. Although the mechanism is unclear, it is possible that a diffusion defect related to intrapulmonary leukostasis develops during hemodialysis [ 11. Alternatively, arterial hypoxemia may have resulted from a reduction in cardiac output, which has been shown to occur with dialysis [ 16-181, thereby reducing the oxygen tension of mixed venous blood-an
important factor in the oxygenation of arterial blood in patients with significant ventilation/perfusion mismatch [ 19,201, or from microatelectasis resulting from hypoventilation during dialysis. However, it would appear that ventilation/perfusion mismatch, a diffusion defect, or intrapulmonary shunting are not likely to be the sole factors, since these abnormalities are generally associated with an increased physiologic deadspace. Further studies, particularly more sophisticated measurements of the relationship between regional ventilation and perfusion or diffusion will be necessary to clarify the mechanism responsible for the gradually increasing alveolar-arterial oxygen gradient we observed. In summary, both acetate and bicarbonate hemodialysis, in short- and long-term studies, produced significant and comparable hypoxemia. Although the mechanisms contributing to hypoxemia are different, bicarbonate dialysate does not offer more protection from dialysis-induced hypoxemia than acetate, in either the short- or long-term. ACKNOWLEDGMENT Technical support was provided by Myers Henry. We appreciate the cooperation of the Dialysis Unit staff of the Dallas Veterans Administration Medical Center and the secretarial assistance of Ms. Becky Rendon and Ms. Virginia Mitchell.
August 1984
The American Journal of Medicine
Volume 77
259
HYF’OXEMIA DURING DIALYSIS-HUNT
ET AL
REFERENCES 1.
2.
3.
4. 5.
6.
7.
6.
9.
10.
260
Craddock PR, Fehr J, Brigham KL, Kronenberg RS, Jacob HS: Complement and le;kocyte-mediated pulmonary dysfunction in hemodialvsis. N Enal J Med 1977: 296: 769774. Aurigemma NM, Feldman NT, Gottlieb M, Ingram RH, Lazarus JM, Lowrie EG: Arterial oxygenation during hemodialysis. N Engl J Med 1977; 297: 671-673. Tolchin N, Roberts JL, Lewis EJ: Respiratory gas exchange by high-efficiency hemodialyzers. Nephron 1976; 2 1: 137-145. Burns CB, Scheinhorn DJ: Hypoxemia during hemodialysis. Arch Intern Med 1962; 142: 1350-1353. Wathen RL, Ferris FZ, Nagar D, Keshaviak P: An alternative explanation for dialysis-induced arterial hypoxemia (abstr). Am Sot Nephrol 1978; 11: 55A. Sherlodc J, Ledwith J, Letteri J: Hypoventilation and hypoxemia during hemodialysis: reflex response to removal of CO* across the dialyzer. Trans Am Sot Artif Intern Organs 1977; 23: 406-410. Patterson RW, Nissenson AR, Miller J, Smith RT, Narins RG. Sullivan SF: Hypoxemia and pulmonary gas exchange during hemodialysis. J Appl Physiol 1961; 50: 259-264. Dolan MJ, Whipp BJ, Davidson WD, Weitzman RE, Wasserman K: Hypopnea associated with acetate hemodialysis: carbon dioxide-flow dependent-ventilation. N Engl J Med 1961; 305: 72-75. Oh MS; Uribarri JV, Del Monte ML, Friedman EA. Carroll HJ: Consumption of CO1 in metabolism of acetate as an explanation for hypoventilation and hypoxemia during hemodialysis. Proc Clin Dial Transplant Forum 1979; 9: 226-229. Nissenson AR: Prevention of dialysis-induced hypoxemia by bicarbonate dialysis. Trans Am Sot Artif Intern Organs
August 1984
The American
Journal of Medicine
1960; 26: 339-342. Eiser AR, Jayamanne D, Kokseng C, Che H, Slifkin RF, Neff MS: Contrasting alterations in pulmonary gas exchange during acetate and bicarbonate hemodialysis. Am J Nephrol 1962; 2: 123-127. 12. Henrich WL, Woodard TD. Meyer BD, Chappell TR, Rubin LJ: High sodium bicarbonate and acetate hemodialysis: a double-blind crossover comparison of hemodynamic and ventilatory effects. Kidney Int 1963; 24: 245-250. 13. Slonim NB, Bell BP, Christensen SE: Cardiopulmonary laboratory basic methods and calculations. Springfield, Illinois: Charles C Thomas, 1967; 155-157. 14. Ponte J, Purves MJ: Carbon dioxide and venous return andtheir interaction as stimuli to ventilation in the cat. J Physiol 1978; 274: 455-475. 15. Heinemann HO, Goldring RM: Bicarbonate and the regulation of ventilation. Am J Med 1974; $7: 361-370. 16. Aizawa Y, Ohmore T, lmae I, Nara Y. Matsuoka M, Hirasawa Y: Depressant action of acetate upon the human cardiovascular system. Clin Nephrol 1977; 6: 477-460. 17. Azancot I, Degoulet P, Juillet Y, Rottembourg J, Legrain M: Hemodynamic evaluation of hypotension during chronic hemodialysis. Clin Nephrol 1977; 6: 312-316. ia. Handt A, Farber MO, Szwed JJ: lntradialytic measurement of cardiac output by thermodilution and impedance cardiography. Clin Nephro! 1977; 7: 61-64. 19. Kelman GR, Nunn, JF, Prys-Roberts C, Greenbaum R: The influence of cardiac output on arterial oxygenation: a theoretical study. Br J Anaesth 1967; 39: 450-456. 20. Mithoefer JC, Ramierez C, Cook W: The effect of mixed venous oxygenation on arterial blood in chronic obstructive pulmonary disease. Am Rev Respir Dis 1976; 117: 259264. 11.
Volume 77
JUDSON M. HUNT, M.D.’ TIMOTHY R. CHAPPELL, M.D. WILLIAM L. HENRICH, M.D. LEWIS J. RUBIN, M.D. Dallas,Texas
From the Department of Internal Medicine, Pulmonary and Nephrology Dlvisions, University of Texas Health Science Center, Southwestern Medical School, and Veterans Administration Medical Center, Dallas, Texas. This work was supported in part by Eli Lilly and Company, Erika Corporation, and the Medical Service of the Veterans Administration Medical Center, Dallas, Texas. Requests for reprints should be addressed to Dr. Lewis J. Rubin, Pulmonary 11 lF, Veterans Administration Medical Center, 4500 South Lancaster Road, Dallas, Texas 75216. Manuscript accepted February 9, 1984. Current address: University of Texas Health Science Center, Nephrology Section, 7703 Floyd Curl Drive, San Antonio, Texas 78229. l
Although arterial hypoxemia during hemodiaiysis is common and may contribute to dialysis morbidity, the mechanisms responsible remain uncertain. Additionally, controversy exists as to whether bicarbonate diaiysate produces less hypoxemia than acetate dialysate. The short- and long-term effects of acetate dialysate and bicarbonate dfaiysate on gas exchange were compared in eight stable patients undergoing dialysis using a closed, proportioning system and a double-blind, crossover study destgn. Diaiysate was sampied immediateiy proximal and distal to the diaiyzer to determine its contribution to total carbon dioxide elimination. Ventilatory parameters and blood gas values were measured before dialysis, at one hour, and after dialysis. Arterial oxygen tension fell signpicantiy and comparably at one hour with both diaiysates, whereas the aiveoiar-arterial oxygen gradient tncreased only silghtiy. Despite hypoxemia, minute ventilation decreased by 4 to 18 percent, and arterial carbon dioxide tension was unchanged. Aithough total carbon dioxide elimination was unchanged in all groups, there was a significant decrease in lung total carbon dioxide eiiminatlon with acetate diaiysate of 9.23 f 2.89 to 7.74 f 1.57 mmoi per minute on Day 1 (mean f SD, p <0.025) concomitant with a loss of total carbon dioxide into the bath of 2.04 f 0.20 mmoi per minute, resuiting in a significant reduction in respiratory quotient (0.92 f 0.07 to 0.75 f 0.05, p KO.01). in contrast, there was a gain of total carbon dioxide into the blood of 1.84 f 0.45 mmoi per minute with bicarbonate diaiysate, which resulted in an increased pH at one hour compared with acetate diaiysate (7.39 f 0.04 versus 7.35 f 0.03, p <0.05). Hypoxemia persisted after dialysis in ail groups and was associated with an increased alveolar-arterial oxygen gradient in three of the four groups. it is concluded that transitory hypoventiiation contributes to comparable hypoxemia with both acetate and bicarbonate diaiysates by different mechanisms. With acetate diaiysate, there is a decrease in carbon dioxlde load to the lungs, whereas with bicarbonate diaiysate, the mechanism responslbie appears to be a suppression of respiratory drive resuittng from a gain of bicarbonate from the diaiysate. Additionally, neither dialysate prevents post-dialysis hypoxemia, which is associated with an increased alveolar-arterial oxygen gradient resulting from a mechanism that remains to be elucidated.
August 1984
The American Journal of Meclklne
Volume 77
255
HYPOXEMIA DURING DIALYSIS-HUNT
TABLE I
ET AL
Dialysate Composition Acetate 136.8 f 108.8 f 2.08 f 2.2 f 36’ 7.0 f 1.7 f 204.0 f
Sodium (meqlliter) Chloride (meq/liter) Potassium (meqlliter) Bicarbonate (meq/liter) Acetate (meqlliter) Calcium (meq/liter) Magnesium (meq/liter) Dextrose (mg/dl)
Bicarbonate
0.6 0.3 0.05 0.2
135.0 106.4 2.12 33.2
0.04 0.1 1.0
6.26 f 0.09 1.6 f 0.0 203.0 f 2.4
Not measured; value given is from manufacturer’s tions. l
f 1.1 f 4.0 f 0.04 f 0.6
specifica-
Arterial hypoxemia can be a significant contributing factor to morbidity during acetate hemodialysis. Although the pathogenesis of dialysis-induced hypoxemia remains unclear [l-4], several possible explanations have been proposed, including ventilation-perfusion mismatching resulting from complement-mediated leukostatic plugging of the pulmonary vasculature [ 11, an increased affinity of hemoglobin for oxygen (Bohr effect) induced by increasing blood pli during dialysis [5], and alveolar hypoventilation. Alveolar hypoventilation seems particularly plausible since the alveolararterial oxygen gradient does not increase with hypoxemia that develops early in dialysis [6,7]. Hypoventilation has been suggested to be due to decreased
delivery of carbon dioxide to central chemoreceptors resulting from either a loss of carbon dioxide into the dialysate [8] or to a decreased rate of total body carbon dioxide production as the result of acetate metabolism [9], both resulting in a decreased respiratory quotient. Since the metabolic quotient may be different from the respiratory quotient if lung carbon dioxide elimination is not equal to total carbon dioxide production, measurement of carbon dioxide loss in dialysate is crucial to discerning the mechanism responsible. The use of a bicarbonate dialysate in lieu of acetate has been found to reduce the incidence of hypoxemia by some investigators [ 8, lo] but not by others [ 11,121.
TABLE II
Dialysate Gas Values before and after Dialyrer with Saline Recirculation* Oxygen (torr)
Carbon Dloxide (torr)
0.012 0.016
111.8f8.8 113.2 f 8.4
84.9 f 5.6 82.7 f 6.0
0.036 0.036
105.5 f 7.6 106.9 f 6.0
PH Bicarbonate dialysate Before 7.124 f After 7.127 f Acetate dialysate Before 6.939 f After 6.940 f l
Sixteen observations,
256
August 1984
each group.
7.9 f 7.9 f
1.2 1.2
The issue of which dialysate buffer produces the least hypoxemia is of clinical importance, particularly in patients with a marginal cardiopulmonary status. Accordingly, the present study was designed to evaluate the effects of both dialysates on gas exchange and to clarify the mechanisms responsible for the alterations that have been observed. PATIENTS AND METHODS Patient Population. Eight stable patients undergoing long-term hemodialysis were included in the study after giving informed consent. The mean age of the patients was 55.5 f 7.6 years and the mean duration of dialysis was 55.2 f 39.6 months. The cause of kidney failure was hypertensive nephrosclerosis in three, and diabetic nephropathy, chronic glomerulonephritis, interstitial nephritis, recurrent nephrolithiasis, and unknown in one patient each. The average interdialytic weight gain was 2.32 f 0.71 kg. Each patient participated in both dialysate protocols, which lasted two weeks each and consisted of five-hour periods of dialysis three times per week. The same dialyzer (Organon-Technica, 1 m*, 220 ml per minute average blood flow) and closed delivery system (Drake-Willock proportioning system), modified for either acetate or bicarbonate concentrates, was utilized in each patient. Dialysis Protocol. The dialysis protocol consisted of a two-week course (six treatments) employing either an acetate or bicarbonate dialysate. The dialysate was sampled immediately proximal to the dialyzer, and its composition was verified by laboratory analysis (Table I). Aside from the presence of either acetate or bicarbonate, no significant differences in dialysate composition were noted. Four patients initially began with acetate dialysis and were switched to bicarbonate dialysis at the end of the two-week period, whereas the other four patients began with bicarbonate and switched to acetate dialysis in a similar fashion. Neither the patients nor the dialysis medical staff were aware of the dialysate composition or crossover time. Predialysis weight and blood pressure were measured and arterial and venous lines were inserted in the usual fashion in preparation for
hemodialysis in each patient. Gas Exchange Studies. Short-term studies were performed on the first day of dialysis using either acetate or bicarbonate dialysate, and long-term studies were performed during the sixth dialysis treatment. Timed collections of exhaled gas were obtained before dialysis, after the first hour of dialysis, and immediately following dialysis after the patient’s blood had been returned. The oxygen and carbon dioxide contents of exhaled gas were measured using a mass spectrometer (Perkin Elmer MGA 1100, Pomona, California), and the volumes were measured in a Tissot gasometer (W. Collins, Braintree, Massachusetts). From these data, minute ventilation, tidal volume, ratio of dead space to tidal volume, total body oxygen consumption, carbon dioxide elimination, respiratory quotient, and alveolar-arterial oxygen gradient were calculated using the standard formulas [ 131. Blood was collected anaerobically for arterial blood gas measurements at times corresponding to ventilatory measurements. Diafysate gas tensions, pH, and bicarbonate concentration were
The American Journalof Medicine Volume 77
HYPDXEMIA DURING DIALYSIS-HUNT
ET AL
Irsa
Figure 1. The effects of short- and long-term acetate (open and closed circk?, respective&) and bizarbonate(open and closedsqvares) henwdialysis on gas exchange. Pa4 = arterial oxygen tension; A-a02D = alveolar-arterial oxygen gnradient; pace* = arterial carbondioxide tension: VE = minute ventilation. Measurements were madebefore (pre),at one hour, and a&r (post)dialysis ( l p I 0.025, 7 p 10.05 compared with predialysis values).
tII II
\iE 8L/min
76-
1111
54
1 ;IO”R
t
measured from samples anaerobically drawn proximal and distal to the dialyzer after the first hour of dialysis. Arterial blood gas and diilysate measurements were performad using an IL 813 blood gas analyzer (Instrumentation Laboratory, Lexington, MA). Prior to the initiation of dialysis, dialysate samples were collected proximal and distal to the dialyzer while saline se lution was circulating through the dialysis machine (Table II). From the dialysate data, obtained at one hour, the amount of total carbon dioxide lost or added was calculated using a modification of the Fick equation: carbon dioxide uptake or elimination by dialysis equals the dialysate flow rate multiplied by the carbon dioxide difference across the dialyzer (calculated by subtracting the total carbon dioxide content of dialysate proximal to the dialyzer from the carbon dioxide content of fluid distal to the dialyzer). Total carbon dioxide was calculated by multiplying the arterial carbon dioxide tension by 0.03 and then adding the bicarbonate value; it is expressed in mmol/llter. In order to ensure accurate dialysate flow rate measurements, timed collections of dialysate were performed at least twice during each five-hour dialysis period using a graduated cylinder. The data are presented as the
P&ST
mean f SD. Statistical analysis was performed using a two-way analysis of variance for sequential data. Linear regression analysis was performed using the least-squares method. RESULTS The arterial blood gas measurements for the short- and long-term studies with both dialysates are shown in
Figure 1. With short-term acetate dialysate, the PaOp fell from 92 f 8 to 79 f 7 mm Hg at one hour, and remained at 80 f 5 mm Hg after dialysis. A similar decline in arterial oxygen tension was observed with short-term bicarbonate dialysate, from 98 f 10 to 85 f 8 mm Hg at one hour and to 8 1 f 10 mm Hg immediately after dialysis. There were no differences in oxygen tension between the groups at each interval. The arterial oxygen tension was also significantly decreased both at one hour and after hemodialysis in the long-term acetate dialysate group. With long-term bicarbonate dialysate, in contrast, there was a gradual
August 1984
The American Journal of Medicine
Volume 77
257
HYPOXEMIA
DURING DIALYSIS-HUNT
TABLE III
ET AL
Arterfal Blood pH and Bicarbonate and Carbon Dioxide Lost from Blood to Dlalysate or Gained from
Dlalysate to Blood (at one hour) during Hemodlalysls
Before Acetate Day 1 Day6 Bicarbonate Day 1 Day 6 ’
pH OneHour
After
Before
Bicarbenate (meq/llter) One Hour
After
7.35 f 0.04 7.33 f 0.03
7.35 f 0.03 7.34 f 0.03
7.39 f 0.03’ 7.39 f 0.03+
19.0 f 2.5 16.4 f 1.8
18.0 f 2.4 16.5 f 1.4
19.7 f 19.8 f
1.6 1.9’
7.35 f 0.04 7.36 f 0.03
7.39 f 0.04 7.39 f 0.03’S
7.46 f 0.03++ 7.45 f 0.03++
17.1 f 3.8 18.8 f 2.1
21.2 f 3.6’+ 21.4 f 2.0’+
22.8 f 2.7++ 23.3 f 2.8++
CarbonDioxide Change (mmol/mlnutet
-2.04 -1.88
f 0.20 l 0.19
+1.64 f 0.45 +1.21 f 0.68
p < 0.05, + p < 0.01 compared with predialysis values; + p < 0.05 compared with both acetate dialysis studies.
decline in arterial oxygen tension, which was different from control only after dialysis. As in the long-term studies, there were no differences in arterial oxygen tension between the groups at each interval. Despite arterial hypoxemia. the alveolar-arterial oxygen gradient at one hour was not statistically different from the predialysis values in any of the groups, although there was a trend for the alveolar-arterial oxygen gradient to increase slightly in three of the four groups. In contrast, the post-dialysis alveolar-arterial oxygen gradient was significantly higher compared with predialysis values in three of the four groups. The ratio of dead space to tidal volume was normal in all groups (range 0.34 to 0.42) and was not significantly changed throughout the study. Arterial carbon dioxide tension increased slightly with short- and long-term bicarbonate dialysate (31.5 f 5.2 to 36.1 f 5.0 mm Hg, p = 0.06; and 34.7 f 2.8 to 36.8 f 2.9 mm Hg, p = O.l), but there were no differences in arterial carbon dioxide tension between the groups at each interval. Despite hypoxemia, minute ventilation fell slightly in all groups. The change in minute ventilation at one hour was significant, however, only on Day 6 with acetate dialysate. No significant differences in minute ventilation were observed between acetate and dialysate buffers at any time period. The effects of acetate and bicarbonate dialysates on blood pH and bicarbonate values are shown in Table III. Blood pH increased steadily with both buffers and was significantly higher after dialysis compared with predialysis values. Additionally, pH and bicarbonate were significantly higher both at one hour and after dialysis in the bicarbonate dialysate groups compared with the acetate dialysate groups. Oxygen consumption and total carbon dioxide elimination were comparable in all groups and unchanged by dialysis (Figure 2). Respiratory carbon dioxide elimination decreased at one hour with acetate dialysate, in both short- and long-term studies, concomitant with a loss of total carbon dioxide into the dialysate. As
258
August 1984
The American Journal of Medicine
a result, the respiratory quotient dropped significantly at one hour, whereas the metabolic quotient remained unchanged. COMMENTS Routine hemodialysis is associated with a variety of adverse effects including hypotension, hypoxemia, muscular cramps, nausea, and vomiting. Hypotension is particularly troublesome since it hinders the adequate removal of fluid and, if symptomatic, may limit tolerance of dialysis. The hypoxemia of dialysis generally occurs early [l] (usually within 30 to 60 minutes after dialysis commences) and may contribute to hypotension by impairing myocardial function or inducing arrhythmias, particularly in patients with underlying cardiopulmonary disease. Accordingly, any maneuver that would reduce the incidence or severity of hypoxemia and hypotension might be expected to improve patient tolerance of dialysis. The use of bicarbonate dialysate achieved an early measure of popularity because it was found to reduce the incidence and severity of hypoxemia and adverse symptoms compared with acetate buffer [3,8]. Subsequent reports, however, noted that hypoxemia also occurs with bicarbonate dialysate, leaving this issue unsettled [ 11,121. Additionally, the mechanisms responsible for the changes in ventilatory function have not been clearly elucidated. The present study demonstrates that significant and comparable degrees of hypoxemia occur at one hour of either acetate or bicarbonate dialysis and that hypoventilation, produced by different stimuli, contributes to the observed reductions in arterial oxygen tension. Although the alveolar-arterial gradient was increased slightly at one hour of dialysis, it is unlikely that abnormalities in gas exchange that increase the alveolararterial oxygen gradient are solely responsible for hypoxemia at one hour, since there was not a close correlation between the increase in the alveolar-arterial oxygen gradient and the decrease in arterial oxygen
Volume 77
HYFQXEMIA DURING DIALYSIS-HUNT
ET AL
1.2&
14
1.1.
131
LO0.9lo-
0.7-
9-
9'/co,
0.8-
L
8-
%,
mmol/min
mmol/min
8-
RQ
0.60.5-
7-
7-
6-
6-
0.4-
5-
5-
0.3-
4-
4-
0.2-
3-
3-
0.1-
/
e
e
ACETATE
BICARB
-
ACETATE
BICARB
ACETATE
BICARB
Figure 2. The effects of short-term acetate or bicarbonate dialysis on total carbon dioxide elimination (i’C02), oxygen consumption N02), and respiratory quotient (RQ). Measurements were made before dialysis, at one hour, and after dialysis. Solid area denotes carbon dioxide elimination by dialyzer ( l p I 0.05 compared with pre-dialysis value).
tension (r = -0.63). In contrast, a closer inverse correlation was observed between the alveolar-arterial oxygen gradient and the arterial oxygen tension after dialysis (r = -0.80) at a time when hypoxemia persisted and minute ventilation had returned to baseline levels, implying that other factors contribute to postdialysis hypoxemia. Our observations are most compatible with the following mechanisms of hypoxemia. During acetate hemodialysis, a loss of total carbon dioxide, associated with an unchanged total body carbon dioxide production, results in a decreased carbon dioxide load to the lungs, promoting relative alveolar hypoventilation [ 141 without changing arterial carbon dioxide tension. In contrast, a net gain of total carbon dioxide from the bath associated with significant increases in blood pH and bicarbonate was the stimulus for hypoventilation [ 151 during bicarbonate dialysis. It is of interest that the hypoxemia persisted after both acetate and bicarbonate dialysis treatments were discontinued and was associated with an increased alveolar-arterial oxygen gradient. Although the mechanism is unclear, it is possible that a diffusion defect related to intrapulmonary leukostasis develops during hemodialysis [ 11. Alternatively, arterial hypoxemia may have resulted from a reduction in cardiac output, which has been shown to occur with dialysis [ 16-181, thereby reducing the oxygen tension of mixed venous blood-an
important factor in the oxygenation of arterial blood in patients with significant ventilation/perfusion mismatch [ 19,201, or from microatelectasis resulting from hypoventilation during dialysis. However, it would appear that ventilation/perfusion mismatch, a diffusion defect, or intrapulmonary shunting are not likely to be the sole factors, since these abnormalities are generally associated with an increased physiologic deadspace. Further studies, particularly more sophisticated measurements of the relationship between regional ventilation and perfusion or diffusion will be necessary to clarify the mechanism responsible for the gradually increasing alveolar-arterial oxygen gradient we observed. In summary, both acetate and bicarbonate hemodialysis, in short- and long-term studies, produced significant and comparable hypoxemia. Although the mechanisms contributing to hypoxemia are different, bicarbonate dialysate does not offer more protection from dialysis-induced hypoxemia than acetate, in either the short- or long-term. ACKNOWLEDGMENT Technical support was provided by Myers Henry. We appreciate the cooperation of the Dialysis Unit staff of the Dallas Veterans Administration Medical Center and the secretarial assistance of Ms. Becky Rendon and Ms. Virginia Mitchell.
August 1984
The American Journal of Medicine
Volume 77
259
HYF’OXEMIA DURING DIALYSIS-HUNT
ET AL
REFERENCES 1.
2.
3.
4. 5.
6.
7.
6.
9.
10.
260
Craddock PR, Fehr J, Brigham KL, Kronenberg RS, Jacob HS: Complement and le;kocyte-mediated pulmonary dysfunction in hemodialvsis. N Enal J Med 1977: 296: 769774. Aurigemma NM, Feldman NT, Gottlieb M, Ingram RH, Lazarus JM, Lowrie EG: Arterial oxygenation during hemodialysis. N Engl J Med 1977; 297: 671-673. Tolchin N, Roberts JL, Lewis EJ: Respiratory gas exchange by high-efficiency hemodialyzers. Nephron 1976; 2 1: 137-145. Burns CB, Scheinhorn DJ: Hypoxemia during hemodialysis. Arch Intern Med 1962; 142: 1350-1353. Wathen RL, Ferris FZ, Nagar D, Keshaviak P: An alternative explanation for dialysis-induced arterial hypoxemia (abstr). Am Sot Nephrol 1978; 11: 55A. Sherlodc J, Ledwith J, Letteri J: Hypoventilation and hypoxemia during hemodialysis: reflex response to removal of CO* across the dialyzer. Trans Am Sot Artif Intern Organs 1977; 23: 406-410. Patterson RW, Nissenson AR, Miller J, Smith RT, Narins RG. Sullivan SF: Hypoxemia and pulmonary gas exchange during hemodialysis. J Appl Physiol 1961; 50: 259-264. Dolan MJ, Whipp BJ, Davidson WD, Weitzman RE, Wasserman K: Hypopnea associated with acetate hemodialysis: carbon dioxide-flow dependent-ventilation. N Engl J Med 1961; 305: 72-75. Oh MS; Uribarri JV, Del Monte ML, Friedman EA. Carroll HJ: Consumption of CO1 in metabolism of acetate as an explanation for hypoventilation and hypoxemia during hemodialysis. Proc Clin Dial Transplant Forum 1979; 9: 226-229. Nissenson AR: Prevention of dialysis-induced hypoxemia by bicarbonate dialysis. Trans Am Sot Artif Intern Organs
August 1984
The American
Journal of Medicine
1960; 26: 339-342. Eiser AR, Jayamanne D, Kokseng C, Che H, Slifkin RF, Neff MS: Contrasting alterations in pulmonary gas exchange during acetate and bicarbonate hemodialysis. Am J Nephrol 1962; 2: 123-127. 12. Henrich WL, Woodard TD. Meyer BD, Chappell TR, Rubin LJ: High sodium bicarbonate and acetate hemodialysis: a double-blind crossover comparison of hemodynamic and ventilatory effects. Kidney Int 1963; 24: 245-250. 13. Slonim NB, Bell BP, Christensen SE: Cardiopulmonary laboratory basic methods and calculations. Springfield, Illinois: Charles C Thomas, 1967; 155-157. 14. Ponte J, Purves MJ: Carbon dioxide and venous return andtheir interaction as stimuli to ventilation in the cat. J Physiol 1978; 274: 455-475. 15. Heinemann HO, Goldring RM: Bicarbonate and the regulation of ventilation. Am J Med 1974; $7: 361-370. 16. Aizawa Y, Ohmore T, lmae I, Nara Y. Matsuoka M, Hirasawa Y: Depressant action of acetate upon the human cardiovascular system. Clin Nephrol 1977; 6: 477-460. 17. Azancot I, Degoulet P, Juillet Y, Rottembourg J, Legrain M: Hemodynamic evaluation of hypotension during chronic hemodialysis. Clin Nephrol 1977; 6: 312-316. ia. Handt A, Farber MO, Szwed JJ: lntradialytic measurement of cardiac output by thermodilution and impedance cardiography. Clin Nephro! 1977; 7: 61-64. 19. Kelman GR, Nunn, JF, Prys-Roberts C, Greenbaum R: The influence of cardiac output on arterial oxygenation: a theoretical study. Br J Anaesth 1967; 39: 450-456. 20. Mithoefer JC, Ramierez C, Cook W: The effect of mixed venous oxygenation on arterial blood in chronic obstructive pulmonary disease. Am Rev Respir Dis 1976; 117: 259264. 11.
Volume 77