Table 4—

Hazard ratios for prediction of all-cause mortality using Cox’s regression analysis#

FEV1QFEV1·Ht−2FEV1·Ht−3FEV1PP
Chi-squared10408103761032010257
Sex
 Females1111
 Males1.82.01.81.5
Age decile
 1 (youngest)1111
 21.61.71.71.8
 32.22.32.32.7
 42.93.13.13.9
 53.53.83.95.0
 64.24.54.76.3
 75.35.86.18.6
 86.67.37.711.5
 98.59.410.015.8
 10 (oldest)13.014.615.627.3
Lung function decile
 1 (best)1111
 21.41.41.31.1
 31.61.61.61.2
 42.01.71.61.2
 52.02.01.91.4
 62.62.32.11.6
 73.02.72.51.8
 83.73.43.22.4
 94.84.44.22.9
 10 (worst)6.96.55.94.1
  • The youngest, best lung function and female sex groups were used as the reference groups. Chi-squared for each model are shown as a measure of fit, and, although the differences in Chi-squared appear small, the models in the leftmost columns are significantly better than those in the rightmost columns (p<0.05). FEV1Q: forced expiratory volume in 1 s (FEV1) quotient; FEV1·Ht−2: FEV1 divided by height squared; FEV1·Ht−3: FEV1 divided by height cubed; FEV1PP: percentage predicted FEV1. #: in 26,967 subjects, using sex, age and lung function indices as predictors, with the latter two split into deciles.