Table 4.
Univariate analyses utilized to predict clinical deterioration and mortality
| (A) Logistic regression models utilized to predict clinical deterioration within 6 months | |||
| Odds ratio | 95 % CI | p-value | |
| %FVC | 0.959 | 0.922–0.992 | 0.014 |
| %DLCO | 0.903 | 0.824–0.964 | <0.001 |
| Serum MMP-7 | 0.483 | ||
| Serum MMP-10 | 2.716 | 1.184–7.830 | 0.017 |
| BALF MMP-7 | NA | ||
| BALF MMP-10 | NA | ||
| (B) Cox hazard models utilized to predict mortality | |||
| Hazard ratio | 95 % CI | p-value | |
| %FVC | 0.952 | 0.917–0.986 | 0.005 |
| %DLCO | 0.892 | 0.828–0.950 | <0.001 |
| Serum MMP-7 | 0.595 | ||
| Serum MMP-10 | 1.773 | 1.032–2.887 | 0.039 |
| BALF MMP-7 | 0.992 | ||
| BALF MMP-10 | 0.377 | ||
The odds ratio and hazard ratio are shown when p < 0.10
CI confidence interval, %FVC percentage of predicted forced vital capacity, %DL CO percentage of predicted diffusing capacity of the lung for carbon monoxide, MMP matrix metalloproteinase, BALF bronchoalveolar lavage fluid, NA not available