RÉSUMÉ
OBJECTIVES: Over the last 30 years, despite the increasing sophistication in medical care, the mortality of acute renal failure(ARF) has remained virtually unchanged at 40-50%, but the reasons remain unknown. This study intend to identify prognostic risk factors influencing survival and predict the mortality in ARF patients. METHODS: We retrospectively analyzed 152 patients with ARF who required renal replacement therapy, or whose serum creatinine level above 5 mg/dl, from Jan. 1988 to May. 1995. Multiple factors which may influence mortality were evaluated by univariate and multivariate analysis. RESULTS: 1) Of the 152 patients, 97 were male and 55 were female. The mean age was 47 years and the overall mortality was 36.8%. 2) Based on the univariate analysis, age>60 years, cause of ARF, APACHE II score, number of failing organs, peak serum creatinine level, PaO2, coma, hypotension, ARDS, GI bleeding, ventilatory support, need for antiarrhythmics, DIC, cardiovascular failure, pulmonary failure, neurological failure, and gastrointestinal failure were all significant factors discriminating between survivors and nonsurvivors(p<0.05) 3) By multivariate analysis, hypotension, coma, ventilatory support, and age over 60 years were significant independent predictors influencing survival in ARF patients and logistic equation and logit score were as follows : z=-2.04+1.32(age over 60)+2.18(hypotension)+2.88 (ventilatory support) + 3.28(coma) P=ez/(1+ ez) 4) In ROC(receiver-operating characteristic)curve, when the cutoff point was 0.2, maximum sensitivity was 75% and maximum specificity was 82%. CONCLUSION: In ARF, prognostic risk factors for mortality were age over 60 years, hypotension, assisted ventilation and coma. The logit score by multiple analysis is a reliable predictor of mortality in ARF patients, however the further studies are required to confirm these results.