Predicting mortality in patients with malarial acute renal failure

Acute Physiology and Chronic Health Evaluation (APACHE) III scores, calcula ted within the first 24 h of admission, were analysed in 108 patients with acute renal failure due to falciparum malaria who were admitted to Bangkok Hospital for Tropical Diseases, Thailand. Twelve (11.1%) patients died. The mean APACHE III score was 82.0+/-25.5 (range, 45-171). There was a close r elation between the APACHE III score and the hospital mortality rate. The n on-survivors had significantly higher APACHE III scores than the survivors, 109.8+/-36.7 and 75.7+/-21.6, respectively (P<0.001). Patients with APACHE III score greater than or equal to 82 had a 4.2-fold higher risk of dying compared with patients with a lower score (95% CI 1.2-14.7; P=0.013). Haemo dialysis treatment was performed in 97 (89.8%) of the patients. The mean AP ACHE III score for patients who were not treated with haemodialysis (95.9+/ -38.0) was not significantly higher than those were received haemodialysis (80.4+/-23.5; P>0.05), but the former had a 4.4-times higher risk of dying compared with those dialysed (95% CI 1.6-12.3; P=0.019). Using the APACHE I II score and its ability to predict death, we calculated its sensitivity, s pecificity and accuracy to be 0.92, 0.31 and 0.41, respectively, at a cut-o ff score of 67 points. The area under the receiver operating characteristic (ROC) curve was 0.75. The APACHE III scoring system correlated well with t he outcome of critically ill malaria patients with acute renal failure, alt hough it was not possible to identify individual survivors or non-survivors . APACHE III should not be used for individual prognosis or treatment decis ions.