Logistic risk model predicting postoperative renal failure requiring dialysis in cardiac surgery patients

Renal failure requiring dialysis represents a serious complication following cardiac surgery. This study was designed to determine the incidence and predictors of renal failure requiring dialysis in a contemporary patient population. We also aimed to create a model based on these risk factors that could serve as a tool for the prediction of renal failure requiring dialysis.

Between October 2007 and June 2009, 2511 consecutive patients (mean age 69 ± 12 years, 68% male) underwent on-pump cardiac surgery at our institution. The main outcome investigated was postoperative renal failure requiring temporary or permanent dialysis. Other postoperative parameters included in the analysis were hospital mortality, major morbidity, length of hospital stay, and discharge condition. Predictors of renal failure requiring dialysis were determined using multivariate regression models. The discriminatory power was evaluated by calculating the area under the receiver-operating-characteristic (ROC) curves (c-statistic).

Renal failure requiring dialysis occurred in 3.9% (n = 98) of patients. Hospital mortality among patients with dialysis-dependent renal failure was 37.8% compared with a mortality rate of 1.3% in patients without this complication (p<0.001). Multivariate analysis revealed pulmonary hypertension (odds ratio (OR) = 8.1), preoperative renal dysfunction (creatinine >2.0 mg dl⁻¹) (OR=4.6), cardiopulmonary bypass (CPB) time >120 min (OR=3.9), peripheral vascular disease (OR = 3.1), previous myocardial infarction (OR=3.0), atrial fibrillation (OR = 2.8), age > 75 years (OR = 2.6), New York Heart Association (NYHA) class IV (OR = 2.5), and diabetes (OR = 2.0) as independent predictors for postoperative renal failure requiring dialysis. A logistic equation including the coefficients of the regression analysis accurately predicted individual patient's risk for the occurrence of renal failure requiring dialysis (area under the ROC curve: 0.829, 95% confidence interval 0.78-0.86).

Renal failure requiring dialysis remains a serious complication, particularly in patients with pulmonary hypertension and previous renal dysfunction. Our logistic risk model allows the prediction of renal failure requiring dialysis, based on the individual presentation of risk factors and, therefore, helps to determine the perioperative risk in cardiac surgery patients.