Assessing Mortality Models in Systemic Sclerosis-Related Interstitial Lung Disease

The gender, age, and lung physiology (GAP) model, interstitial lung diseases-GAP (ILD-GAP) model, and the smoking history, age, and diffusion capacity of the lung (SADL) model were compared using a systemic sclerosis-ILD (SSc-ILD) cohort to evaluate which best determined prognosis. The models were applied to a cohort of 179 patients with SSc seen at a tertiary care center within 1 year of ILD diagnosis. Demographics, clinical characteristics, and mortality were recorded. The performance of the models was assessed using standardized mortality ratios (SMR) of observed versus predicted outcomes for calibration and concordance (c)-statistics for discrimination. SSc-ILD patients with usual interstitial pneumonia (31, 17%) had a higher mortality than those with non-specific interstitial pneumonia (147, 83%) (hazard ratio 2.27; 95%CI 1.03-4.97). All 3 models had comparable discrimination (c = 0.72, 0.72, and 0.71, respectively). Regarding calibration, the ILD-GAP model underestimated mortality (SMR 1.50; 95%CI 1.05-2.14). Calibration was acceptable for SADL (SMR 1.00; 95%CI 0.70-1.44) and GAP (SMR 0.90; 95%CI 0.63-1.29). The SADL model underestimated mortality in Stage I ILD. The ILD-GAP model underestimated mortality, and the SADL model underestimated mortality in certain subgroups. However, the GAP model performed well in this cohort, providing the best prognostic information for SSc-ILD.