Serum creatinine measurement immediately after cardiac surgery and prediction of acute kidney injury

Background: After heart surgery, acute kidney injury (AKI) confers substantial long-term risk of death and chronic kidney disease. We hypothesized that small changes in serum creatinine (SCr) levels measured within a few hours of exit from the operating room could help discriminate those at low versus high risk of AKI.

Study design: Prospective cohort of 350 elective cardiac surgery patients (valve or coronary artery bypass grafting) recruited in Winnipeg, Canada. Baseline SCr level was obtained at the preoperative visit 2 weeks before surgery. The postoperative SCr level was drawn within 6 hours of completion of surgery and then daily while the patient was in the hospital.

Predictor: Immediate (ie, <6 hours) postoperative SCr level change (ΔSCr), categorized as within 10% (reference), decrease >10%, or increase >10% relative to baseline.

Outcome: AKI, defined according to the new KDIGO (Kidney Disease: Improving Global Outcomes) consensus definition as an increase in SCr level >0.3 mg/dL within 48 hours or >1.5 times baseline within 1 week.

Measurements: We compared the C statistic of logistic models with and without inclusion of immediate postoperative ΔSCr.

Results: After surgery, 176 patients (52%) experienced a decrease >10% in SCr level, 26 (7.4%) experienced an increase >10%, and 143 had ΔSCr within ±10% of baseline. During hospitalization, 53 (14%) developed AKI. Bypass pump time, baseline estimated glomerular filtration rate, and European System for Cardiac Operative Risk Evaluation (euroSCORE) were associated with AKI in a parsimonious base logistic model. Added to the base model, immediate postoperative ΔSCr was associated strongly with subsequent AKI and significantly improved model discrimination over the base model (C statistic, 0.78 [95% CI, 0.71-0.85] vs 0.69 [95% CI, 0.62-0.77]; P < 0.001). A ≥10% SCr level decrease predicted significantly lower AKI risk (OR, 0.37; 95% CI, 0.18-0.76), whereas a ≥10% SCr level increase predicted significantly higher (OR, 6.38; 95% CI, 2.37-17.2) AKI risk compared with the reference category.

Limitations: We used a surrogate marker of AKI. External validation of our results is warranted.

Conclusion: In elective cardiac surgery patients, measurement of immediate postoperative ΔSCr improves prediction of AKI.