Hemodilution during cardiopulmonary bypass is an independent risk factor for acute renal failure in adult cardiac surgery

BackgroundThis observational study sought to determine whether the degree of hemodilution during cardiopulmonary bypass is independently related to perioperative acute renal failure necessitating dialysis support.MethodsData were prospectively collected on consecutive patients undergoing cardiac operations with cardiopulmonary bypass from 1999 to 2003 at a tertiary care hospital. The independent relationship was assessed between the degree of hemodilution during cardiopulmonary bypass, as measured by nadir hematocrit concentration, and acute renal failure necessitating dialysis support. Multivariate logistic regression was used to control for variables known to be associated with perioperative renal failure and anemia.ResultsOf the 9080 patients included in the analysis, 1.5% (n = 134) had acute renal failure necessitating dialysis support. There was an independent, nonlinear relationship between nadir hematocrit concentration during cardiopulmonary bypass and acute renal failure necessitating dialysis support. Moderate hemodilution (nadir hematocrit concentration, 21%-25%) was associated with the lowest risk of acute renal failure necessitating dialysis support; the risk increased as nadir hematocrit concentration deviated from this range in either direction (P = .005). Compared with moderate hemodilution, the adjusted odds ratio for acute renal failure necessitating dialysis support with severe hemodilution (nadir hematocrit concentration <21%) was 2.34 (95% confidence interval, 1.47-3.71), and for mild hemodilution (nadir hematocrit concentration >25%) it was 1.88 (95% confidence interval, 1.02-3.46).ConclusionsGiven that there is an independent association between the degree of hemodilution during cardiopulmonary bypass and perioperative acute renal failure necessitating dialysis support, patient outcomes may be improved if the nadir hematocrit concentration during cardiopulmonary bypass is kept within the identified optimal range. Randomized clinical trials, however, are needed to determine whether this is a cause-effect relationship or simply an association