The TOwards a Revolution in COPD Health (TORCH) study was a 3-yr multicentre trial of 6,112 patients randomised to salmeterol (Salm), fluticasone propionate (FP), a Salm/FP combination (SFC) or placebo (P). Here the cost-effectiveness of treatments evaluated in the TORCH study is assessed. For four regions, 3-yr all-cause hospitalisation, medication and outpatient care costs were calculated. The sample was restricted to the 21 countries (n = 4,237) in which European quality of life five-dimension (EQ-5D) data were collected in order to estimate the number of quality-adjusted life years (QALYs). Regression models were fitted to survival, study medication cost, other medication cost and EQ-5D data in order to estimate total cost, number of QALYs and cost per QALY, adjusted for missing data and region. SFC had a trial-wide estimate of cost per QALY of 43,600 US dollars (USD) compared with P (95% confidence interval 21,400-123,500 USD). Estimates for Salm versus P (197,000 USD) and FP versus P (78,000 USD) were less favourable. The US estimates were greater than those from other regions; for SFC versus P, the cost per QALY was 77,100 (46,200-241,700) USD compared to 24,200 (15,200-56,100) USD in Western Europe. Compared with P, SFC has a lower incremental cost-effectiveness ratio than either FP or Salm used alone, and is, therefore, preferred to these monotherapies on the grounds of cost-effectiveness.