ABSTRACT: The effect of hyperoncotic (20%-25%) human albumin on cardiac-surgery-associated acute kidney injury (CS-AKI) is uncertain. This study evaluated whether intraoperative or postoperative administration of 20%-25% albumin influences the risk of CS-AKI compared with crystalloids or iso-oncotic (4%-5%) albumin. Systematic review and meta-analysis of randomized controlled trials (RCTs) and risk-adjusted cohort studies were conducted. Risk ratios (RRs) were pooled using random-effects models. Literature search of PubMed, Embase (Ovid), and Cochrane CENTRAL was carried out from January 1, 1995, to July 17, 2025. Adults undergoing on-pump CS. Four eligible studies-two RCTs and two cohort studies, encompassing 6651 patients-were included. About 20%-25% of human albumin was administered either intraoperatively or within 24 h postoperatively. Comparators were crystalloids or 4%-5% albumin. The primary outcome was the incidence of any-stage AKI within 7 days of surgery. Pooled analysis showed 20%-25% albumin increased the risk of CS-AKI (RR 1.10, 95% confidence interval CI 1.05-1.16; I² =0%). Restriction to RCTs yielded a similar result (RR 1.12, 95% CI 1.04-1.20). The increased risk was consistent for both intraoperative (RR 1.09) and postoperative (RR 1.12) administration, but the interaction between infusion timing was non-significant (P = 0.59). Infusion of 20%-25% albumin is associated with a modest but consistent increase in postoperative AKI, independent of infusion timing from intraoperative to postoperative within 24 h. Until adequately powered trials resolve the remaining imprecision, routine perioperative administration of hyperoncotic albumin should be approached with caution.
Abo et al. (2026) studied this question.