Objective: To evaluate the ability of the RIFLE classification to predict hospital mortality in adult patients who underwent cardiac surgery.
Methods: From October 1st 2006 to December 31st 2006, five hundred and nine adult patients who underwent coronary artery bypass grafting and/or valve operation were enrolled in this study. Renal function was assessed daily according to the RIFLE classification, meanwhile, APACHE II score and SOFA score were also evaluated, as well as the maximum scores were recorded.
Results: Mean duration of ventilation support was 18 (14-19) hours, the time of ICU stay was 1.4 +/- 1.0 days, and the time of postoperative hospital stay was 12.0 (10.0-15.0) days. 167 patients (32.8%) incurred postoperative ARF according to the RIFLE classification. The overall mortality was 4.3% (22/502). A significant increase (P < 0.01) was observed for mortality based on RIFLE classification. By applying the area under the receiver operating characteristic curve, the RIFLE classification had more powerful discrimination power [0.933, (95% CI 0.872-0.995), P < 0.001].
Conclusions: ARF is one of the major complications in postcardiotomy patients. Analytical data suggested the good discriminative power of the RIFLE classification for predicting inpatient mortality of adult postoperative patient with ARF, and the RIFLE classification is simple and practically performed. According to the RIFLE classification, patients with RIFLE class I or class F incur a significantly increased risk of in-hospital mortality compared with those who never develop ARF.