There were no significant differences in the percent of predicted values for FEV1, FVC, total lung capacity (TLC), or diffusing capacity of the lung for carbon monoxide (Dlco) [Table 3]
. The FEV1/FVC ratio was higher in the deceased group but did not reach statistical significance (p = 0.06). The changes in values for exercise Pao2, resting Pco2, exercise Pco2, and V̇o2max, expressed as a volume per the weight of each patient, were not statistically significant. Fourteen of 48 patients had arterial blood gas levels measured during exercise, and 27 of 48 patients had results of cardiopulmonary exercise testing. All of the hemodynamic measurements were recorded during right heart catheterization. There was no significant difference in the systolic pulmonary artery pressure, diastolic pulmonary artery pressure, mean pulmonary artery pressure, or cardiac output between the two groups. The mean Pao2 in the alive group was 62.4 mm Hg, and in the deceased group it was 53.9 mm Hg (p = 0.04). The only statistically significant gas exchange abnormality between the two groups was resting Pao2. In order to determine the clinical relevance of the parameters that were statistically significant between the two groups, a survival analysis was performed. Because the FEV1/FVC ratio approached statistical significance, it was included in the survival analysis. A univariate Cox regression analysis demonstrated that the FEV1/FVC ratio was a significant predictor of survival (hazards ratio, 1.07; 95% confidence interval [CI], 1.00 to 1.14; p = 0.04). In stepwise multivariate analysis, Pao2 and FEV1/FVC ratio were significantly associated with survival (hazards ratio, 1.07; 95% CI, 1.00 to 1.14; p = 0.03).