Acute kidney injury and prognosis after cardiopulmonary bypass: a meta-analysis of cohort studies

Background: Robust estimates and sources of variation in risks of clinical outcomes for cardiopulmonary bypass (CPB)-associated acute kidney injury (AKI) are needed to inform clinical practice and policy. We aimed to assess whether the methods for defining acute kidney disease modify the estimated association of AKI with CPB.

Study design: Systematic review and meta-analysis.

Setting & population: Adults undergoing CPB.

Selection criteria for studies: Cohort studies reporting adjusted associations between CPB-associated AKI and early mortality, later mortality, stroke, myocardial infarction, congestive heart failure, all-cause hospitalization, chronic kidney disease, end-stage kidney disease, bleeding complications, or perioperative infection.

Predictors: CPB-associated AKI and renal replacement therapy.

Outcomes: The primary outcome was early mortality (in-hospital or within 90 days of surgery) in studies reporting adjusted associations and secondary outcomes including total and cardiovascular mortality, major adverse cardiovascular events, rehospitalization, end-stage kidney disease, bleeding, and perioperative infection.

Results: 46 studies with 47 unique cohorts comprising 242,388 participants were included. The pooled rate of CPB-associated AKI was 18.2%, and of renal replacement therapy, 2.1%. CPB-associated AKI was associated with early mortality (risk ratio [RR], 4.0; 95% CI, 3.1-5.2; crude mortality with CPB-associated AKI, 4.6%; without CPB-AKI, 1.5%) with considerable heterogeneity between studies (I(2)=87%). The AKI definition did not modify prognostic estimates (P for subgroup analysis = 0.9). When heterogeneity was fully accounted for using credibility ceilings, risks of early mortality were attenuated (RR, 2.2; 95% CI, 1.8-2.8) but remained high. Renal replacement therapy also was associated with early mortality (RR, 5.3; 95% CI, 3.4-8.1). CPB-associated AKI also was associated with long-term mortality (RR, 2.0; 95% CI, 1.7-2.3) and stroke (RR, 2.2; 95% CI, 1.1-4.5). No other outcomes were reported in more than 3 studies.

Limitations: Unclear attrition from follow-up in most studies and variable adjustment for confounders across studies.

Conclusions: CPB-associated AKI is associated with a more than 2-fold increase in early mortality regardless of AKI definition.