To identify the determinants of survival in patients with idiopathic pulmonary fibrosis (IPF), we performed a survival analysis on 74 subjects with IPF. The study subjects were on average 64 yr of age (range, 25 to 83 yr), 62% were male, and 29% were never smokers. A tissue diagnosis was made in 67 (91%) of our study subjects. These subjects were followed for a mean period of 4 yr (range, 1.4 to 118.8 months) after the onset of pulmonary symptoms. During the period of observation, 41 subjects died (median survival = 28.2 months) and 33 continue to survive (median follow-up period = 60.9 months). A univariate analysis demonstrated that diminished survival was significantly associated with male gender (hazard ratio = 1.98; 95% confidence interval [CI] = 1.01-3.85), a higher FEV1/FVC ratio (hazard ratio = 1.82 [per 10% increase in the FEV1/FVC ratio]; 95% CI = 1.21-2.73), a lower percent predicted FVC (hazard ratio = 0.74; 95% CI = 0.60-0.91), a lower percent predicted total lung capacity (TLC) (hazard ratio = 0.75; 95% CI = 0.60-0.94), a lower percent predicted diffusing capacity of carbon monoxide (DLCO) (hazard ratio = 0.69; 95% CI = 0.53-0.89), a higher ILO profusion category on chest radiograph (hazard ratio = 3.52; 95% CI = 1.58-7.87), and an enhanced release of prostaglandin E2 (PGE2) by cultured alveolar macrophages (hazard ratio = 1.32 [per 10 pm/ml of PGE2]; 95% CI = 1.07-1.62). In fact, after controlling for age, Cox's regression analyses demonstrated that diminished survival was independently associated with male gender (hazard ratio = 9.05;95% CI = 1.84-44.5), a higher FEV,/FVC ratio (hazard ratio = 3.91 [per 10% increase in the FEV,/FVC ratio]; 95% CI = 1.68-8.07), and an enhanced release of PGE2 by cultured alveolar macrophages (hazard ratio = 1.45 [per 10 pm/ml of PGE2] ; 95% CI = 1.14-1.83). Survival was not found to be independently associated with the concentration of cells in the lavage fluid or other cytokines (tumor necrosis factor(alpha) [TNFa] and interleukin-1beta [IL-1beta]) released by alveolar macrophages. Limiting our analysis to subjects closely followed in our SCOR Program (n = 41), Cox's regression analysis (adjusted for age) demonstrated that diminished survival was independently associated with male gender (hazard ratio = 8.21; 95% CI = 1.30-50.6), a higher ILO profusion category (hazard ratio = 4.00; 95% CI = 1.43-11.42), and a lower concentration of lymphocytes in bronchoalveolar lavage fluid (hazard ratio = 0.95; 95% CI = 0.91-0.99). In aggregate, our findings suggest that adverse prognosticators in IPF include male gender, advanced disease (restrictive lung function, abnormal gas exchange, and increased interstitial abnormalities on the chest radiograph), and, possibly, increased release of PGE2 from cultured alveolar macrophages.
D.A. Schwartz, Department of Internal Medicine, Department of Veterans Administration Medical Center, Iowa City, IA 52242