Preoperative renal function, postoperative acute kidney injury, and 5-year mortality after coronary artery bypass grafting: a causal mediation analysis

Background Preoperative chronic kidney disease (CKD) is associated with adverse outcomes following coronary artery bypass grafting (CABG). However, whether postoperative acute kidney injury (AKI) mechanistically mediates the pathway from impaired renal function to long-term mortality, and whether this relationship differs according to cardiopulmonary bypass strategy, remains poorly understood. Purpose To evaluate the independent associations of preoperative estimated glomerular filtration rate (eGFR) with postoperative AKI and 5-year all-cause mortality in CABG patients, and to quantify the mediating role of AKI stratified by on-pump versus off-pump surgery. Methods Patients undergoing CABG were stratified by preoperative eGFR into three groups: ≥90, 60–89, and <60 mL/min/1.73m². Multivariable logistic regression and Cox proportional hazards models were used to assess associations between eGFR categories, postoperative AKI, and 5-year mortality. Models were progressively adjusted for demographics, comorbidities, surgical variables, and baseline biomarkers. Causal mediation analysis was performed to decompose the total effect of eGFR on mortality into direct and AKI-mediated indirect effects, both in the overall cohort and stratified by surgical strategy (on-pump vs. off-pump). Results Among 25,268 CABG patients, AKI incidence increased progressively across declining eGFR categories (≥90: 43.7%; 60–89: 61.6%; <60: 84.4%; p for trend <0.001). After full adjustment, eGFR <60 remained independently associated with both AKI (OR 5.19, 95%CI 4.49–6.01) and 5-year mortality (HR 1.50, 95%CI 1.19–1.88), compared with eGFR ≥90. Causal mediation analysis revealed that AKI mediated 41.4% (95%CI 15.8–70.0%, p=0.001) of the eGFR–mortality association in the overall cohort. Notably, in on-pump CABG, AKI accounted for 53.6% of the total effect (p=0.020), with a non-significant direct effect (P=0.270), suggesting AKI as the predominant mediating pathway. Conversely, in off-pump CABG, the mediated proportion was 36.2% and did not reach statistical significance (p=0.060). Conclusion Reduced preoperative eGFR independently increases risks of postoperative AKI and 5-year mortality after CABG. AKI substantially mediates the eGFR–mortality association, acting as the dominant pathway particularly in on-pump surgery, whereas direct mechanisms appear more prominent in off-pump procedures. These findings underscore the critical importance of AKI prevention strategies in CKD patients undergoing on-pump CABG.