Predictors of Mortality in Pulmonary Sarcoidosis

Abstract

OBJECTIVE: The goal of this study was to assess the prognostic strength of factors in predicting respiratory death in a large cohort of patients with sarcoidosis with at least 8 years' follow-up. METHODS: Data were collected on age, sex, self-declared race, time of death, spirometry findings, diffusing capacity for carbon monoxide, chest radiograph stage, extent of fibrosis on high-resolution CT (HRCT) scanning, and presence of precapillary pulmonary hypertension (PH). The Gender-Age-Physiology (GAP) index and Walsh model criteria for high vs low risk were calculated. RESULTS: A total of 452 patients were identified who had complete pulmonary function testing and chest imaging. The median age at time of entry into the study was 50 years (range, 25-78 years). PH was confirmed by right heart catheterization in 29 (6.4%). Of 452 patients, 42 died during the time of the study, including 38 (8.4%) who died of sarcoidosis-associated respiratory failure and 4 who died of non-sarcoidosis causes. The overall mortality from sarcoidosis was 3.9% and 9.0% at 5 and 10 years, respectively. Alive patients were significantly younger than patients who died from sarcoidosis. Increased mortality was seen for black patients, stage 4 chest radiographs, > 20% fibrosis on HRCT scanning, or PH. The two composite scores (GAP and the Walsh model) were predictive of increased mortality according to univariate analysis. Using the Cox proportional hazards regression model, only age, extent of fibrosis on HRCT scanning, and PH were independent predictors of mortality. CONCLUSIONS: Although most patients with sarcoidosis do well, increased mortality was seen in those patients who were older, had extensive fibrosis on HRCT scanning, or who had PH.