BACKGROUND: Little is known about adherence to inhaled medication in chronic obstructive pulmonary disease (COPD) and the impact on mortality and morbidity. METHODS: Data on drug adherence from a randomised double-blind trial comparing inhaled salmeterol 50 microg + fluticasone propionate 500 microg twice daily with placebo and each drug individually in 6112 patients with moderate to severe COPD over 3 years in the TORCH study were used. All-cause mortality and exacerbations leading to hospital admission were primary and secondary end points. The study of adherence was not specified a priori as an ancillary study. RESULTS: Of the 4880 patients (79.8%) with good adherence defined as >80% use of study medication, 11.3% died compared with 26.4% of the 1232 patients (20.2%) with poor adherence. The annual rates of hospital admission for exacerbations were 0.15 and 0.27, respectively. The association between adherence and mortality remained unchanged and statistically significant after adjusting for other factors related to prognosis (hazard ratio 0.40 (95% CI 0.35 to 0.46), p<0.001). The association was even stronger when analysing on-treatment deaths only. Similarly, the association between adherence and hospital admission remained unchanged and significant in a multivariate analysis (rate ratio 0.58 (95% CI 0.44 to 0.73, p<0.001). The association between increased adherence and improved mortality and reduction in hospital admission was independent of study treatment. The effect of treatment was more pronounced in patients with good adherence than in those with poor adherence. CONCLUSION: Adherence to inhaled medication is significantly associated with reduced risk of death and admission to hospital due to exacerbations in COPD. Further research is needed to understand these strong associations.