Prediction of five-year survival following presentation with scleroderma -Development of a simple model using three disease factors at first visit

Objective. To identify combinations of easily acquired clinical variables, at first presentation with scleroderma, that would predict subsequent morta lity. Methods. In this prospective study of all new patients at one major sclerod erma center, 280 patients with definite scleroderma (according to the Ameri can College of Rheumatology criteria) whose disease onset occurred from 198 2 to 1991 and who were followed up to the end of 1996 were identified. Stan dardized data collection was performed at entry to obtain data on major cli nical and laboratory variables. Vital status was determined by linkage to t he National Death Index. Results. At 5 years, 55 (26%) of 215 women and 21 (32%) of 65 men had died. Univariate analysis showed that older age, diffuse skin disease, higher sk in score, low carbon monoxide diffusing capacity, abnormal electrocardiogra m findings, proteinuria, hematuria, low hemoglobin level, elevated erythroc yte sedimentation rate (ESR), and presence of antitopoisomerase antibody we re all associated with increased mortality. A logistic regression model, va lidated by Monte Carlo simulation, identified 3 factors, proteinuria, eleva ted ESR, and low carbon monoxide diffusing capacity, that in combination, h ad an accuracy of >80% in predicting mortality. The absence of these 3 fact ors was associated with 93% survival. Conclusion, A simple model has been developed that appears to accurately pr edict mortality over 5 years in a cohort of patients newly presenting with scleroderma.