- Email: [email protected]
Oxygen consumption in hemodialysis patients undergoing cardiopulmonary bypass
Oxygen Consumption in Hemodialysis Patients Undergoing Cardiopulmonary Bypass Waheedullah Karzai, MD and Hans-Joachim Priebe, MD Objectives: Cardiopulmonary bypass (CPB) can be successfully performed in patients on hemodialysis. However, ischemic complications occur more often in these patients. This could partly be because of shunting through the arteriovenous (AV) fistula during CPB, resulting in reduced peripheral flow and oxygen (Oz) delivery. Inadequate oxygen delivery during. CPB should be reflected in a lower oxygen consumption (VOz) compared with patients without an AV fistula. Design: To test the hypothesis, the authors analyzed VO2 in three groups of patients retrospectively. Group I included 14 patients with end-stage renal failure (creatinine level 9.1 _+ 0.3 mg/dL, urea level 126 -+ 8 mg/dL) requiring hemodialysis through an AV fistula. Group 2 included 13 patients with compensated renal insufficiency (creatinine level 3.1 -+ 0.4 mg/dL, urea level 106 +- 10 mg/dL) without an AV fistula. Group 3 included 14 patients with normal renal function (creatinine level 1.0 +-- 0.1 mg/dL, urea levet 44 -- 4 mg/dL). Setting: An operating room of a university hospital Participants: Patients undergoing cardiac surgery requiring CPB.
Measurements and Main Results: ~lOz was calculated from the recorded hemodynamic and blood gas data using standard formulae. Data were analyzed using a two-way analysis of variance with a repeated measurement on one factor. Before undergoing CPB, VO2 was similar in all three groups. VO2 decreased in all three groups during hypothermic CPB (standard flow rate 2.2 L/min/m z, standard temperature 29°C) and returned to prebypass levels during rewarming. There was no difference in VO2 among the three groups during hypothermic CPB or during rewarming. Only base excess decreased more in group 1 patients compared with the other groups (p < 0.001). Conclusion: During hypothermic CPB at a flow rate of 2.2 L/min/m 2, shunting through an AV fistula is unlikely to lead to decreased VO2 in dialysis patients.
MPROVED SURGICAL TECHNIQUES and perioperative care have made cardiopulmonary bypass (CPB) feasible in patients with chronic renal failure dependent on hemodialysis. 1,2 However, the incidence of splanchnic ischemia after cardiac surgery appears to be higher in this patient population, a Accelerated atherosclerosis of splanchnic vessels may be one of the causes of inadequate regional perfusion and ischemia in these patients. In addition, shunting through the arteriovenous (AV) hemodialysis fistula during CPB may also affect regional flow and oxygen (02) delivery. Normally, shunt flow through an AV fistula is compensated for by an increase in cardiac: output. However, pump flow rates during CPB are rarely increased to compensate for shunt flow. It was, therefore, hypothesized that standard pump flow rates during CPB may be inadequate in these patients. Inadequate flow or O2 delivery during CPB should be reflected in greater decreases in oxygen consamption (402). To test this hypothesis, the authors retrospectively analyzed 4 0 2 in patients undergoing CPB with and without AV fistula.
started and continued throughout the operation. Increments of flunitrazepam and pancuronium were added during surgery at the discretion of the anesthesiologist. In addition, 1 mg of flunitrazepam and 2 to 4 mg of pancuronium were administered 5 to 10 minutes before CPB. Valve replacement and aortocoronary bypass or the combination of both were the cardiac procedures requiring CPB (Table t). Extracorporeal circulation was maintained using nonpulsatile pump flow (Sttckert, Mtinchen, Germany). Oxygenation was provided by a Cobe CML membrane oxygenator (Laboratories GmbH, Heimstetten, Germany). An asanguineous prime containing 1,000 mL of lactated Ringer's solution, 500 mL of 6% hydroxyethylstarch, and 200 mL of 20% mannitol was used in control patients. In patients with renal disorders (groups 1 and 2), hydroxyethylstarch was replaced by an electrolyte solution or by 1 to 2 units of packed red cells (group 1, 10 patients; group 2, 5 patients). The standard CPB flow rate at this institution is 2.2 L/min/mz. However, the rounding of calculated flow rates for surface area of individual patients produced small deviations. During CPB, hemofiltration, initiated after stable CPB conditions had been attained, was performed in nine group 1 patients, in two group 2 patients, and in none of the group 3 patients. Although records of blood gas analysis (all alpha-stat) and hemodynamic parameters were available for several periods of the operation and CPB, three periods that were least confounded by such factors as surgical stress, temperature variations, and vasoactive drug management were chosen. The three periods were as follows: (1) after induction of anesthesia but before surgical incision, (2) during steady-state hypothermic CPB (blood temperature 29°C, rectal temperature within 0.5°C of blood temperature, standard flow rate), and (3) during rewarming (blood temperature 35 ° to 35.5°C, rectal temperature within
I
METHODS
Forty-one patients who underwent CPB at this institution between 1993 and 1995 were studied retrospectively. Group 1 consisted of 14 patients with end-stage renal failure requiring hemodialysis through an AV shunt. Group 2 consisted of 13 patients with chronic compensated renal insufficiency (creatinine level > 2.5 mg/dL) not requiring hemodialysis. Group 3 consisted of 14 patients with normal renal function who underwent surgery within 10 days of the respective group 1 patients, matched for age, sex, and surgical procedure. All group 1 patients were hemodialyzed 1 day before surgery (Table 1). All patients received 1 to 2 mg of flunitrazepam 1 to 2 hours before arriving in the operating room. An arterial catheter was placed using local anesthesia. Anesthesia was induced with fentanyl, 5 to 8 pg/kg; flunitrazepam, 10 to 15 lag/kg; and pancuronium, 0.1 to 0.15 mg/kg, and was maintained with 0.4% to 0.6% isoflurane in a 50% oxygen and nitrous oxide mixture. A continuous fentanyl infusion, 0.5 rag/h, was
Copyright© 1998by WB. Saunders Company KEY WORDS: cardiopulmonary bypass, hemodialysis, oxygen delivery, oxygen consumption, arteriovenous fistula
From the Department ofAnesthesiology, University Hospital Freiburg, Freiburg, Germany. Address reprint requests to Waheedullah Karzai, MD, Department of Anesthesiology, University Hospital Jena, Bachstr. 18, 07740 Jena, Germany. Copyright © 1998 by W.B. Saunders Company 1053-0770/98/1204-0009508.00/0
Journal of Cardiothoracic and Vascular Anesthesia, Vol 12, No 4 (August), 1998: pp 415-417
415
416
KARZAI AND PRIEBE
Table 1. Patient Characteristics Group 1 (n = 14) Age (yr) Weight (kg) Aortocoronary bypass (no.) Valve replacement (no.) Preoperative creatinine level (mg/dL) Preoperative urea level (mg/dL)
59 -+ 2 68 -+ 3 11 6
Group2 (n = 13)
DISCUSSION Group3 (n = 14)
65 -+ 3 76 + 3 11 5
59 -+ 2 72 -+ 2 11 5
9.1 -+ 0.3
3.1 -+ 0.4*
1.0 _+ OAt
126 -+ 8
106 + 10"
44 -+ 4 t
NOTE. Values expressed as mean -+ standard error of the mean. Group 1 included patients requiring hemodialysis, group 2 included patients with chronic renal failure not requiring hemodialysis, and group 3 included patients without renal failure. Because of combined surgical procedures, the sum of operations was greater than the total number of patients in each group. * p < 0,05 compared with groups 1 and 3. ?p < 0.05 compared with groups 1 and 2.
0.5°C of blood temperature, standard flow rate). ~'O2 was calculated using the AV O2 content difference and either cardiac index or CPB flow rate. Statistical analysis was performed using a one-way or a two-way analysis of variance (with a repeated measure on one factor) to test for differences among and within groups. A correction was made for multiple comparisons. In addition, the authors used the chi-square test when appropriate. The program SAS (SAS, Carey, NJ) was used for statistical analysis. RESULTS
Demographic and preoperative laboratory data are listed in Table 1. At the times of intraoperative measurements, no differences in the total amount of anesthetic drugs used (ie, fentanyl, flunitrazepam, pancuroniutu) existed among groups. Total amounts of anesthetics administered before CPB were as follows: fentanyl, 1.1 - 0.08, 1.2 +- 0.07, and 1.2 + 0.05 mg per patient in groups 1, 2, and 3, respectively; flunitrazepam, 3.5 -+ 0.2, 3.4 _+ 0.2, and 3.8 -+ 0.2 tug; and pancuroniutu 14 _+ 1, 14.5 -+ 1, and 16 _+ 1 rag, respectively. Before CPB, cardiac index and 02 delivery were greater in group 1 patients than in patients of groups 2 and 3. Hemoglobin levels were less in patients of groups 1 and 2 than in patients of group 3 (Table 2). During both hypothermic (29°C) CPB and rewarming (35°C) on CPB, there were no differences (p > 0.05) in body temperature (blood and rectal), duration of CPB, hemoglobin, 02 delivery, and mean arterial pressure among groups. "QO2 decreased (p < 0.001) during hypothermic CPB and returned to prebypass values during rewarming in all three groups. At all three time periods (ie, pre-CPB, during hypothermic CPB, during rewarming on CPB), YO2 was similar among groups. Base deficit increased only in patients with an AV fistula (group 1). It is interesting to note that it took longer (p < 0.05) for the rectal temperature to decrease to the predetermined level at the beginning of CPB in group 1 (27.4 + 2.8 minutes) than in groups 2 (18.2 _+ 2.5 minutes) and 3 (20.5 _+ 2.1 minutes). More patients in groups 1 (13 patients requiring zero to four units; median, two units) and 2 (7 patients requiring zero to four units; median, 1 unit) required blood transfusions than did patients of group 3 (one patient requiring two units).
It was hypothesized that vQO2 would decrease during CPB in patients with an AV fistula. In contrast to this hypothesis, during hypothermic CPB as well as during rewarming on CPB, 4 0 2 of hemodialysis patients (with AV shunts) was not less than that of patients with compensated renal insufficiency (but without AV shunts) or of patients without renal disease. In all three groups of patients, VO2 decreased similarly in response to hypothermic CPB, and it retumed to prebypass levels on rewarming. Before CPB, patients with an AV fistula had a higher cardiac output. This was most likely the physiologic response to the AV shunt. This agrees with experimental studies that have also shown increases in cardiac output in the presence of chronic AV fistulae) Anemia alone cannot entirely account for the elevated cardiac output, because patients with compensated renal disease (ie, group 2) also suffered from anemia but had no increase in cardiac output when compared with controls. Depending on size and location of the fistula and general circulatory status, shunt flow varies between 165 and 2,600 mL/min (mean, 806 mL/min).4 Occasionally, high flow rates through AV fistulae causes high-output cardiac failure) Furthermore, gastric mucosal acidosis occurrs during hemodialysis. 6
Table 2. Cardiorespiratory and Laboratory Parameters Before and During CPB Variable
Pre-CPB
CPB (29°C)
CPB (35°C)
~/O2 (m!-,/min/m 2) Group 1 80 -+ 6 62 -+ 4f 84 -+ 4 Group 2 80 + 8 62 + 2f 92 -+ 5 Group 3 80 _+ 6 61 _+ 2t 82 + 3 DO2 (mL/min/m 2) Group 1 406 _+ 45* 247 _+ 14 271 _+ 14 Group 2 323 -+ 27 262 - 8 299 _+ 16 Group3 352 -+ 16 268 -+ 8 287 _+ 9 Mixed-venous saturation (%) Group 1 85 -+ 1 81 _+ 1 74-+ 2f Group 2 79-+2 80-+1 73-+17 Group3 83+1 82+1 76_+21" CI/CPB flow rate (L/min/m 2) Group 1 3.0 -+ 0.2* 2.3 -+ 0.1 2.4 -+ 0.1 Group 2 2.4 _+ 0.1 2.3 _+ 0.1 2.4 _+ 0.1 Group 3 2.1-+0,1 2.2+0.1 2.2-+0,1 MAP (mmHg) Group 1 78 -+ 4 74 -+ 4 72 _+ 5 Group 2 82 -+ 3 72 -+ 2 73 + 4 Group 3 78 _+ 2 74 -+ 4 74 -+ 3 Hgb (g/dL) Group 1 9.6 -+ 1.9 7.4 -+ 0.5 7.9 _+ 0.4 Group 2 9.5 -+ 0.4 8.0 -+ 0.3 8.9 -+ 0.4 Group 3 11.9-+0.3" 8.8_+0.3 9.0_+0.3 BD (mmol/L) Group 1 - 1 . 0 + 0,6 - 4 . 0 + 0.5 - 5 . 4 + 0.4? Group 2 - 3 . 3 -+ 0.6* - 3 . 7 -+ 0.7 - 4 . 4 -+ 0.7 Group 3 - 0 . 3 -+ 0.4 - 1 . 6 -+ 0.5 - 1 . 9 -+ 0.5 NOTE. Values expressed as mean -+ SEM. Abbreviations: CPB, cardiopulmonary bypass; 402, 02 consumption; DO2, 02 delivery; CI, cardiac index; MAP, mean arterial pressure; Hgb, hemoglobin; BD, base deficit. * p < 0.05 between groups, l-p < 0.05 between time periods.
CPB AND HEMODIALYSIS FISTULA
4t7
These findings suggest impairment in regional perfusion during periods of high shunt rates relative to cardiac output. The decreases in '~O2 during hypothermic CPB in patients with AV fistulae were similar to those without AV shunts. This would suggest that the standard CPB flow rate provided adequate flow and oxygen to the peripheral tissues in spite of shunting through AV fistulae. It would thus appear that with pump flow rates between 2.1 and 2.5 L/min/m2, shunting through dialys:s shunts is unlikely to be primarily responsible for the more frequent ischemic complications observed in hemodialysis patients undergoing CPB. When using lower flow rates than those used in this study, adequacy of 02 delivery in patients with AV fistulae must be reassessed: Within reasonable flow rates, comparing the patient's actual "qO2 with a normal "QO2 value for a given temperature at a given institution may be one method of assessing adequacy of 02 delivery during CPB. Mixed-venous 02 saturation, which is frequently used to assess adequacy of flow, is an unreliable means of monitoring regional 02 delivery because it can be artificially elevated because of shunting of arterial blood. Another way of assessing peripheral 02 delivery may be by alternately opening and closing the AV fistula (by applying external pressure) durin.g CPB and simultaneously determining "QO2.An increase in VO2 on closure of the fistula would suggest that at a given CPB flow rate, fistula flow may indeed be compromising regional perfusion and oxygen delivery. However, the possibility (although small) remains of damaging the AV fistula during such a maneuver. Although global 02 delivery and 402 seemed to be preserved
in hemodialysis patients, two findings would suggest some impairment of regional perfusion. First, the hemodialysis patients were the only group of patients in whom base deficit increased with time, necessitating bicarbonate administration in some cases. Whether this greater tendency for accumulation of nonvolatile acid was primarily because of the underlying renal disease per se, or whether it was indirect evidence of impaired regional perfusion and 02 delivery, cannot be answered on the basis of this study. Second, hemodialysis patients (with AV shunts) required a longer time to reach hypothermic temperatures during CPB compared with the other two groups, also suggesting disturbed distribution of flow. Certain limitations of this study have to be acknowledged. These findings are based on retrospective analysis. Thus, the limitations inherent to all retrospective analyses also apply to the results of this study. In addition, global 402 does not necessarily reflect the state of regional perfusion and 02 supply. Regional low perfusion during CPB in patients with AV shunts might have occurred despite CPB flow rates similar to those in patients without shunts. Similar global flow might have been accompanied by regional flow redistribution resulting in (undetected) regional ischemia. In conclusion, ~rO2 in hemodialysis patients with AV fistulae during CPB was not less than in respective control patients, suggesting adequate CPB flow rate and 02 delivery. However, regional flow and 02 delivery may nevertheless be compromised. Prospective studies in patients with AV fistulae undergoing CPB are needed to address this issue.
REFERENCES
1. Ko W, Kreiger KH, Isom OW: Cardiopulmonarybypass procedures in dialysispatients.Ann Thorac Surg 55:677-684, 1993 2. Deutsch E, Bernstein RC, Addonizio R Kussmaul WG III: Coronary artery bypass surgery in patients on chronic hemodialysis.A case-control study.Ann InternMed 110:369-372, 1989 3. Huang M, Hester RL, GuytonAC: Hemodynamicchan~esin rats after openingof an arteriovenousfistula.Am J Physio1262:H846-H851, 1992
4. Rodriguez Moran M, Rodriguez Rodriguez JM, Ramos Boyero M, et al: Flow of dialysis fistulas. Noninvasivestudy performed with standard Doppler equipment.Nephron40:63-66, 1985 5. Engelberts I, Tordoir JHM, Boon ES, Schreij G: High-output cardiac failure due to excessive shunting in hemodialysis fistula:An easily overlookeddiagnosis.Am J Nephro115:323-326, 1995 6. DiebelL, KozolR, WilsonRE et al: Gastricintramucosalacidosis in patientswith chronickidneyfailure.Surgery 113:520-526, 1993
Measurements and Main Results: ~lOz was calculated from the recorded hemodynamic and blood gas data using standard formulae. Data were analyzed using a two-way analysis of variance with a repeated measurement on one factor. Before undergoing CPB, VO2 was similar in all three groups. VO2 decreased in all three groups during hypothermic CPB (standard flow rate 2.2 L/min/m z, standard temperature 29°C) and returned to prebypass levels during rewarming. There was no difference in VO2 among the three groups during hypothermic CPB or during rewarming. Only base excess decreased more in group 1 patients compared with the other groups (p < 0.001). Conclusion: During hypothermic CPB at a flow rate of 2.2 L/min/m 2, shunting through an AV fistula is unlikely to lead to decreased VO2 in dialysis patients.
MPROVED SURGICAL TECHNIQUES and perioperative care have made cardiopulmonary bypass (CPB) feasible in patients with chronic renal failure dependent on hemodialysis. 1,2 However, the incidence of splanchnic ischemia after cardiac surgery appears to be higher in this patient population, a Accelerated atherosclerosis of splanchnic vessels may be one of the causes of inadequate regional perfusion and ischemia in these patients. In addition, shunting through the arteriovenous (AV) hemodialysis fistula during CPB may also affect regional flow and oxygen (02) delivery. Normally, shunt flow through an AV fistula is compensated for by an increase in cardiac: output. However, pump flow rates during CPB are rarely increased to compensate for shunt flow. It was, therefore, hypothesized that standard pump flow rates during CPB may be inadequate in these patients. Inadequate flow or O2 delivery during CPB should be reflected in greater decreases in oxygen consamption (402). To test this hypothesis, the authors retrospectively analyzed 4 0 2 in patients undergoing CPB with and without AV fistula.
started and continued throughout the operation. Increments of flunitrazepam and pancuronium were added during surgery at the discretion of the anesthesiologist. In addition, 1 mg of flunitrazepam and 2 to 4 mg of pancuronium were administered 5 to 10 minutes before CPB. Valve replacement and aortocoronary bypass or the combination of both were the cardiac procedures requiring CPB (Table t). Extracorporeal circulation was maintained using nonpulsatile pump flow (Sttckert, Mtinchen, Germany). Oxygenation was provided by a Cobe CML membrane oxygenator (Laboratories GmbH, Heimstetten, Germany). An asanguineous prime containing 1,000 mL of lactated Ringer's solution, 500 mL of 6% hydroxyethylstarch, and 200 mL of 20% mannitol was used in control patients. In patients with renal disorders (groups 1 and 2), hydroxyethylstarch was replaced by an electrolyte solution or by 1 to 2 units of packed red cells (group 1, 10 patients; group 2, 5 patients). The standard CPB flow rate at this institution is 2.2 L/min/mz. However, the rounding of calculated flow rates for surface area of individual patients produced small deviations. During CPB, hemofiltration, initiated after stable CPB conditions had been attained, was performed in nine group 1 patients, in two group 2 patients, and in none of the group 3 patients. Although records of blood gas analysis (all alpha-stat) and hemodynamic parameters were available for several periods of the operation and CPB, three periods that were least confounded by such factors as surgical stress, temperature variations, and vasoactive drug management were chosen. The three periods were as follows: (1) after induction of anesthesia but before surgical incision, (2) during steady-state hypothermic CPB (blood temperature 29°C, rectal temperature within 0.5°C of blood temperature, standard flow rate), and (3) during rewarming (blood temperature 35 ° to 35.5°C, rectal temperature within
I
METHODS
Forty-one patients who underwent CPB at this institution between 1993 and 1995 were studied retrospectively. Group 1 consisted of 14 patients with end-stage renal failure requiring hemodialysis through an AV shunt. Group 2 consisted of 13 patients with chronic compensated renal insufficiency (creatinine level > 2.5 mg/dL) not requiring hemodialysis. Group 3 consisted of 14 patients with normal renal function who underwent surgery within 10 days of the respective group 1 patients, matched for age, sex, and surgical procedure. All group 1 patients were hemodialyzed 1 day before surgery (Table 1). All patients received 1 to 2 mg of flunitrazepam 1 to 2 hours before arriving in the operating room. An arterial catheter was placed using local anesthesia. Anesthesia was induced with fentanyl, 5 to 8 pg/kg; flunitrazepam, 10 to 15 lag/kg; and pancuronium, 0.1 to 0.15 mg/kg, and was maintained with 0.4% to 0.6% isoflurane in a 50% oxygen and nitrous oxide mixture. A continuous fentanyl infusion, 0.5 rag/h, was
Copyright© 1998by WB. Saunders Company KEY WORDS: cardiopulmonary bypass, hemodialysis, oxygen delivery, oxygen consumption, arteriovenous fistula
From the Department ofAnesthesiology, University Hospital Freiburg, Freiburg, Germany. Address reprint requests to Waheedullah Karzai, MD, Department of Anesthesiology, University Hospital Jena, Bachstr. 18, 07740 Jena, Germany. Copyright © 1998 by W.B. Saunders Company 1053-0770/98/1204-0009508.00/0
Journal of Cardiothoracic and Vascular Anesthesia, Vol 12, No 4 (August), 1998: pp 415-417
415
416
KARZAI AND PRIEBE
Table 1. Patient Characteristics Group 1 (n = 14) Age (yr) Weight (kg) Aortocoronary bypass (no.) Valve replacement (no.) Preoperative creatinine level (mg/dL) Preoperative urea level (mg/dL)
59 -+ 2 68 -+ 3 11 6
Group2 (n = 13)
DISCUSSION Group3 (n = 14)
65 -+ 3 76 + 3 11 5
59 -+ 2 72 -+ 2 11 5
9.1 -+ 0.3
3.1 -+ 0.4*
1.0 _+ OAt
126 -+ 8
106 + 10"
44 -+ 4 t
NOTE. Values expressed as mean -+ standard error of the mean. Group 1 included patients requiring hemodialysis, group 2 included patients with chronic renal failure not requiring hemodialysis, and group 3 included patients without renal failure. Because of combined surgical procedures, the sum of operations was greater than the total number of patients in each group. * p < 0,05 compared with groups 1 and 3. ?p < 0.05 compared with groups 1 and 2.
0.5°C of blood temperature, standard flow rate). ~'O2 was calculated using the AV O2 content difference and either cardiac index or CPB flow rate. Statistical analysis was performed using a one-way or a two-way analysis of variance (with a repeated measure on one factor) to test for differences among and within groups. A correction was made for multiple comparisons. In addition, the authors used the chi-square test when appropriate. The program SAS (SAS, Carey, NJ) was used for statistical analysis. RESULTS
Demographic and preoperative laboratory data are listed in Table 1. At the times of intraoperative measurements, no differences in the total amount of anesthetic drugs used (ie, fentanyl, flunitrazepam, pancuroniutu) existed among groups. Total amounts of anesthetics administered before CPB were as follows: fentanyl, 1.1 - 0.08, 1.2 +- 0.07, and 1.2 + 0.05 mg per patient in groups 1, 2, and 3, respectively; flunitrazepam, 3.5 -+ 0.2, 3.4 _+ 0.2, and 3.8 -+ 0.2 tug; and pancuroniutu 14 _+ 1, 14.5 -+ 1, and 16 _+ 1 rag, respectively. Before CPB, cardiac index and 02 delivery were greater in group 1 patients than in patients of groups 2 and 3. Hemoglobin levels were less in patients of groups 1 and 2 than in patients of group 3 (Table 2). During both hypothermic (29°C) CPB and rewarming (35°C) on CPB, there were no differences (p > 0.05) in body temperature (blood and rectal), duration of CPB, hemoglobin, 02 delivery, and mean arterial pressure among groups. "QO2 decreased (p < 0.001) during hypothermic CPB and returned to prebypass values during rewarming in all three groups. At all three time periods (ie, pre-CPB, during hypothermic CPB, during rewarming on CPB), YO2 was similar among groups. Base deficit increased only in patients with an AV fistula (group 1). It is interesting to note that it took longer (p < 0.05) for the rectal temperature to decrease to the predetermined level at the beginning of CPB in group 1 (27.4 + 2.8 minutes) than in groups 2 (18.2 _+ 2.5 minutes) and 3 (20.5 _+ 2.1 minutes). More patients in groups 1 (13 patients requiring zero to four units; median, two units) and 2 (7 patients requiring zero to four units; median, 1 unit) required blood transfusions than did patients of group 3 (one patient requiring two units).
It was hypothesized that vQO2 would decrease during CPB in patients with an AV fistula. In contrast to this hypothesis, during hypothermic CPB as well as during rewarming on CPB, 4 0 2 of hemodialysis patients (with AV shunts) was not less than that of patients with compensated renal insufficiency (but without AV shunts) or of patients without renal disease. In all three groups of patients, VO2 decreased similarly in response to hypothermic CPB, and it retumed to prebypass levels on rewarming. Before CPB, patients with an AV fistula had a higher cardiac output. This was most likely the physiologic response to the AV shunt. This agrees with experimental studies that have also shown increases in cardiac output in the presence of chronic AV fistulae) Anemia alone cannot entirely account for the elevated cardiac output, because patients with compensated renal disease (ie, group 2) also suffered from anemia but had no increase in cardiac output when compared with controls. Depending on size and location of the fistula and general circulatory status, shunt flow varies between 165 and 2,600 mL/min (mean, 806 mL/min).4 Occasionally, high flow rates through AV fistulae causes high-output cardiac failure) Furthermore, gastric mucosal acidosis occurrs during hemodialysis. 6
Table 2. Cardiorespiratory and Laboratory Parameters Before and During CPB Variable
Pre-CPB
CPB (29°C)
CPB (35°C)
~/O2 (m!-,/min/m 2) Group 1 80 -+ 6 62 -+ 4f 84 -+ 4 Group 2 80 + 8 62 + 2f 92 -+ 5 Group 3 80 _+ 6 61 _+ 2t 82 + 3 DO2 (mL/min/m 2) Group 1 406 _+ 45* 247 _+ 14 271 _+ 14 Group 2 323 -+ 27 262 - 8 299 _+ 16 Group3 352 -+ 16 268 -+ 8 287 _+ 9 Mixed-venous saturation (%) Group 1 85 -+ 1 81 _+ 1 74-+ 2f Group 2 79-+2 80-+1 73-+17 Group3 83+1 82+1 76_+21" CI/CPB flow rate (L/min/m 2) Group 1 3.0 -+ 0.2* 2.3 -+ 0.1 2.4 -+ 0.1 Group 2 2.4 _+ 0.1 2.3 _+ 0.1 2.4 _+ 0.1 Group 3 2.1-+0,1 2.2+0.1 2.2-+0,1 MAP (mmHg) Group 1 78 -+ 4 74 -+ 4 72 _+ 5 Group 2 82 -+ 3 72 -+ 2 73 + 4 Group 3 78 _+ 2 74 -+ 4 74 -+ 3 Hgb (g/dL) Group 1 9.6 -+ 1.9 7.4 -+ 0.5 7.9 _+ 0.4 Group 2 9.5 -+ 0.4 8.0 -+ 0.3 8.9 -+ 0.4 Group 3 11.9-+0.3" 8.8_+0.3 9.0_+0.3 BD (mmol/L) Group 1 - 1 . 0 + 0,6 - 4 . 0 + 0.5 - 5 . 4 + 0.4? Group 2 - 3 . 3 -+ 0.6* - 3 . 7 -+ 0.7 - 4 . 4 -+ 0.7 Group 3 - 0 . 3 -+ 0.4 - 1 . 6 -+ 0.5 - 1 . 9 -+ 0.5 NOTE. Values expressed as mean -+ SEM. Abbreviations: CPB, cardiopulmonary bypass; 402, 02 consumption; DO2, 02 delivery; CI, cardiac index; MAP, mean arterial pressure; Hgb, hemoglobin; BD, base deficit. * p < 0.05 between groups, l-p < 0.05 between time periods.
CPB AND HEMODIALYSIS FISTULA
4t7
These findings suggest impairment in regional perfusion during periods of high shunt rates relative to cardiac output. The decreases in '~O2 during hypothermic CPB in patients with AV fistulae were similar to those without AV shunts. This would suggest that the standard CPB flow rate provided adequate flow and oxygen to the peripheral tissues in spite of shunting through AV fistulae. It would thus appear that with pump flow rates between 2.1 and 2.5 L/min/m2, shunting through dialys:s shunts is unlikely to be primarily responsible for the more frequent ischemic complications observed in hemodialysis patients undergoing CPB. When using lower flow rates than those used in this study, adequacy of 02 delivery in patients with AV fistulae must be reassessed: Within reasonable flow rates, comparing the patient's actual "qO2 with a normal "QO2 value for a given temperature at a given institution may be one method of assessing adequacy of 02 delivery during CPB. Mixed-venous 02 saturation, which is frequently used to assess adequacy of flow, is an unreliable means of monitoring regional 02 delivery because it can be artificially elevated because of shunting of arterial blood. Another way of assessing peripheral 02 delivery may be by alternately opening and closing the AV fistula (by applying external pressure) durin.g CPB and simultaneously determining "QO2.An increase in VO2 on closure of the fistula would suggest that at a given CPB flow rate, fistula flow may indeed be compromising regional perfusion and oxygen delivery. However, the possibility (although small) remains of damaging the AV fistula during such a maneuver. Although global 02 delivery and 402 seemed to be preserved
in hemodialysis patients, two findings would suggest some impairment of regional perfusion. First, the hemodialysis patients were the only group of patients in whom base deficit increased with time, necessitating bicarbonate administration in some cases. Whether this greater tendency for accumulation of nonvolatile acid was primarily because of the underlying renal disease per se, or whether it was indirect evidence of impaired regional perfusion and 02 delivery, cannot be answered on the basis of this study. Second, hemodialysis patients (with AV shunts) required a longer time to reach hypothermic temperatures during CPB compared with the other two groups, also suggesting disturbed distribution of flow. Certain limitations of this study have to be acknowledged. These findings are based on retrospective analysis. Thus, the limitations inherent to all retrospective analyses also apply to the results of this study. In addition, global 402 does not necessarily reflect the state of regional perfusion and 02 supply. Regional low perfusion during CPB in patients with AV shunts might have occurred despite CPB flow rates similar to those in patients without shunts. Similar global flow might have been accompanied by regional flow redistribution resulting in (undetected) regional ischemia. In conclusion, ~rO2 in hemodialysis patients with AV fistulae during CPB was not less than in respective control patients, suggesting adequate CPB flow rate and 02 delivery. However, regional flow and 02 delivery may nevertheless be compromised. Prospective studies in patients with AV fistulae undergoing CPB are needed to address this issue.
REFERENCES
1. Ko W, Kreiger KH, Isom OW: Cardiopulmonarybypass procedures in dialysispatients.Ann Thorac Surg 55:677-684, 1993 2. Deutsch E, Bernstein RC, Addonizio R Kussmaul WG III: Coronary artery bypass surgery in patients on chronic hemodialysis.A case-control study.Ann InternMed 110:369-372, 1989 3. Huang M, Hester RL, GuytonAC: Hemodynamicchan~esin rats after openingof an arteriovenousfistula.Am J Physio1262:H846-H851, 1992
4. Rodriguez Moran M, Rodriguez Rodriguez JM, Ramos Boyero M, et al: Flow of dialysis fistulas. Noninvasivestudy performed with standard Doppler equipment.Nephron40:63-66, 1985 5. Engelberts I, Tordoir JHM, Boon ES, Schreij G: High-output cardiac failure due to excessive shunting in hemodialysis fistula:An easily overlookeddiagnosis.Am J Nephro115:323-326, 1995 6. DiebelL, KozolR, WilsonRE et al: Gastricintramucosalacidosis in patientswith chronickidneyfailure.Surgery 113:520-526, 1993