Table 3—

Results of univariate and multivariate Cox proportional-hazards models relating parameters of the timed walk test and pulmonary function to mortality

VariablesUnivariate modelsMultivariate# models
Relative hazard (95% CI)p-valueRelative hazard (95% CI)p-value
Resting room air SP,O20.93 (0.75–1.14)0.4541.06 (0.83–1.37)0.637
End-exercise SP,O2 2% units0.73 (0.59–0.91)0.0040.74 (0.55–1.00)0.051
Change in SP,O2 with exercise0.80 (0.69–0.92)0.0020.80 (0.67–0.96)0.014
Walk distance 30-m units0.89 (0.81–0.97)0.0100.91 (0.81–1.02)0.098
Walk velocity 0.1-m·s−1 units0.77 (0.65–0.91)0.0020.76 (0.63–0.93)0.006
DL,CO % pred0.93 (0.88–0.97)0.0010.92 (0.87–0.98)0.005
FVC % pred+1.00 (0.97–1.02)0.7810.94 (0.97–1.02)0.646
Resting Pa,O2 mmHg0.97 (0.94–1.00)0.0760.96 (0.91–1.01)0.101
Resting PA–a,O2 difference mmHg1.01 (0.99–1.04)0.3281.01 (0.97–1.05)0.672
  • CI: confidence interval; SP,O2: arterial oxygen saturation measured by pulse oximetry; DL,CO: diffusing capacity for carbon monoxide; % pred: % predicted; FVC: forced vital capacity; Pa,O2: arterial oxygen tension; PA–a,O2: alveolar–arterial oxygen tension. #: the multivariate model included age, sex, FVC % pred, time from the onset of symptoms and supplemental oxygen administration during the test; : for consistency, change in saturation was entered as a negative change such that a lesser degree of desaturation would be associated with reduced mortality; +: the multivariate model excluded FVC as a covariate. 1 kPa = 0.133 mmHg.