Table 2.
Results from efficacy analyses. Mixed effects model analyses adjusted for covariates and Wilcoxon Mann-Whitney test adjusted for centre
| Efficacy parameter | Placebo | Fluticasone propionate | Treatment difference between drug and placebo (95% CI) | P value |
|---|---|---|---|---|
| FEV1 after bronchodilator: | ||||
| No of patients | 325* | 339* | ||
| Mean change in FEV1 ml/year (SE) | −59 (4.4) | −50 (4.1) | 9 (6.0) (−3 to 20) | 0.161 |
| Predicted FEV1 at 3 months | 1.37 | 1.44 | 0.076 (0.056 to 0.097) | <0.001 |
| Predicted FEV1 at 3 years | 1.20 | 1.30 | 0.100 (0.064 to 0.135) | <0.001 |
| Health status: | ||||
| No of patients | 291* | 309* | ||
| Mean change in questionnaire score (SE) (units/year) | 3.17 (0.31) | 2.00 (0.29) | −1.17 (0.40) (−1.95 to −0.39) | 0.004 |
| Annual exacerbation rate: | ||||
| No of patients | 370 | 372 | ||
| Mean (SD) rates | 1.90 (2.63) | 1.43 (1.93) | ||
| Median (range) rates | 1.32 (0 to 30) | 0.99 (0 to 26) | −0.3 (−0.4 to 0.0)† | 0.026 |
FEV1=forced expiratory volume in one second in litres.
*
Numbers are smaller than randomised population for FEV1 and health status because of patient withdrawals, missing assessments, or respiratory infections or exacerbations (affects FEV1 only).
†
Zero values are possible in 95% confidence intervals with non-parametric analyses that show P values ⩽0.05 because method of calculation of confidence intervals differs from non-parametric test.