TY - JOUR T1 - Predictors of mortality and risk prediction among patients with scleroderma related interstitial lung disease JF - European Respiratory Journal JO - Eur Respir J DO - 10.1183/13993003.congress-2015.OA4962 VL - 46 IS - suppl 59 SP - OA4962 AU - Julie Morisset AU - Brett M. Elicker AU - Xiaowen Hu AU - Stephanie Le AU - Jay H. Ryu AU - Kirk D. Jones AU - Brett Ley AU - Paul J. Wolters AU - Talmadge E. King AU - Harold R. Collard AU - Joyce S. Lee Y1 - 2015/09/01 UR - http://erj.ersjournals.com/content/46/suppl_59/OA4962.abstract N2 - Interstitial lung disease (ILD) is an important cause of mortality and morbidity in scleroderma (Ssc). Risk prediction and prognostication in Ssc-ILD patients is challenging. This study aims to determine factors predicting mortality and to develop a clinical risk prediction model for Ssc-ILD.Patients with Ssc-ILD were identified from two ongoing longitudinal cohorts: 135 patients at the University of California, San Francisco (UCSF, derivation cohort) and 90 patients at the Mayo Clinic (validation cohort). Cox regression analysis was used to evaluate the relationship between predictor variables and time to death, censoring lung transplantation. Model discrimination was assessed using the C-statistic.At baseline, the UCSF patients had a mean age of 54 years, were predominantly female (69%), and 39% were ever smokers. The mean forced vital capacity (FVC) % predicted was 72 and the mean diffusing capacity for carbon monoxide (DLCO) % predicted was 51. These baseline characteristics were similar in the Mayo cohort. On bivariate analysis in the UCSF cohort, age (HR 1.07, P=0.001), gender (HR 0.38, P=0.016), FVC-% predicted (HR 0.97, P=0.020), DLCO-% predicted (HR 0.94, P<0.001) and long-term oxygen therapy (HR 5.25, P=0.033) were significant predictors of mortality. These variables were also significant predictors in the Mayo cohort. A combination of 3 variables was included in the final risk prediction model: age, ever smoking history, and DLCO-% predicted. This model had similar performance in the derivation cohort (C-statistic = 0.852) and in the validation cohort (C-statistic = 0.839).A model including age, smoking history and DLCO can predict mortality in Ssc-ILD. ER -