Comparison of multiple organ dysfunction scores in the prediction of hospital mortality in the critically ill*

Publisher Website
Google Scholar
Abstract
To compare the scales and predictive power for hospital mortality of three recent multiple organ dysfunction scores. Prospective, observational, validation cohort study. A ten-bed medical-surgical intensive care unit in a Finnish tertiary care hospital. Among the 591 consecutive patients admitted, 520 patients who stayed >4 hrs were studied. Clinical and laboratory data were collected daily. Acute Physiology and Chronic Health Evaluation (APACHE) III, Multiple Organ Dysfunction Score, Logistic Organ Dysfunction score, and Sequential Organ Failure Assessment score all were calculated and compared for hospital mortality. The areas under receiver operating curves (se) for day-1 scores were 0.825 (0.02) for APACHE III, 0.805 (0.02) for Logistic Organ Dysfunction, 0.776 (0.02) for SOFA, and 0.695 (0.02) for Multiple Organ Dysfunction Score in prediction of hospital mortality. The highest discriminative power was revealed with total maximum scores. No statistical differences existed between the total maximum scores (p values, .06 to .97). Calibration was good for all scores of day-1 multiple organ dysfunction scales and APACHE III by chi-square test (values between 10.14 and 5.42). Discriminative power (ability to distinguish between patients who die and those who live) of day-1, of daily maximum, and especially of total maximum multiple organ dysfunction scores, were rather good, comparable with each other, and comparable with APACHE III in prediction of hospital mortality.