Table 1—

Treatment effect(salmeterol relative to salmeterol/fluticasone) according to pre-trial inhaled corticosteroid (ICS) use

Post hoc analysis groupsTime to first exacerbation HR (95% CI)Count of exacerbations RR (95% CI)
No ICS pre-study 1.37 (1.02–1.85)1.31 (1.05–1.64)
ICS pre-study 1.50 (1.18–1.91)1.63 (1.35–1.96)
  • HR: hazard ratio; CI: confidence interval; RR: ratio rate. Values more than one indicate a higher rate for salmeterol relative to salmeterol/fluticasone treatment. Time to first exacerbation was analysed by Cox's proportional hazard model with the following predictive variables: treatment, sex, age, smoking status, forced expiratory volume in one second (% predicted at start of randomised treatment period) and long-acting β-adrenergic use pre-study. Count of exacerbation was analysed by use of a Poisson regression model with the same covariables as for time to first exacerbation.