Impact of postoperative nadir hemoglobin and blood transfusion on outcomes after operations for atherosclerotic vascular disease.

OBJECTIVE: Controversy surrounds the topic of transfusion policy after noncardiac operations. This study assessed the combined impact of postoperative nadir hemoglobin (nHb) levels and blood transfusion on adverse events after open surgical intervention in patients who undergo operative intervention for atherosclerotic vascular disease.
METHODS: Consecutive patients who underwent peripheral arterial disease (PAD)-related operations were balanced on baseline characteristics by inverse weighting on propensity score calculated as their probability to have nHb greater than 10 gm/dL on the basis of operation type, demographics, and comorbidities, including the revised cardiac risk index. A multivariate generalized estimating equation analysis was performed to investigate associations between nHb, transfusion, and a composite outcome of perioperative death and myocardial infarction. Logistic and Cox proportional hazards regressions were used to assess the impact of nHb and transfusion on respiratory and wound complications; and a composite end point (CE) of death, myocardial infarction during a 2-year follow-up. Level of statistical significance was set at alpha of 0.0125 to adjust for the increased probability of type I error attributable to multiple comparisons.
RESULTS: The analysis cohort included 880 patients (1074 operations). After adjusting for nHb level, the number of units transfused was not associated with the perioperative occurrence of the CE (odds ratio [OR], 1.13; P = .025). Adjusted for the number of units transfused, nHb had no impact on the perioperative CE (OR, 0.62; P = .22). An interaction term between transfusion and nHb level remained nonsignificant (P = .312), indicating that the impact of blood transfusion was the same regardless of the nHb level. Perioperative respiratory complications were more likely in patients receiving transfusions (OR, 1.22; P = .009), and perioperative wound infections were less common in patients with nHb >10 gm/dL (OR, 0.65; P = .01). During an average follow-up of 24 months, transfused patients were more likely to develop the CE (hazard ratio [HR], 1.15, P = .009), whereas nHb level did not impact the long-term adverse event rate (HR, 0.78; P = .373). The above associations persisted even after adjusting the Cox regression model for the occurrence of perioperative cardiac events.
CONCLUSIONS: Although nHb less than 10 gm/dL is not associated with death or ACS after PAD-related operations, maintaining nHb greater than 10 gm/dL appears to decrease the risk of wound infection. Blood transfusion is associated with increased risk of perioperative respiratory complications. Until a randomized trial settles this issue definitively, a restrictive transfusion strategy is justified in patients undergoing operations for atherosclerotic vascular disease. Published by Mosby, Inc.