OBJECTIVES: To assess the cost-effectiveness of esomeprazole (Nexium) compared to lansoprazole (Zoton FasTab) in the initial treatment of reflux oesophagitis over 12 weeks from the perspective of the UK NHS. METHODS: A probabilistic decision analysis model was constructed using Treeage DATA(tm) 4.0 to depict the sequential management of patients with unhealed reflux oesophagitis. Treatment pathways were based on a published 8-week UK healing model, however the model time horizon was extended by an additional 4 weeks to ensure the costs incurred by patients who remained unhealed after 8 weeks were also included. Beta distributions for the 4 and 8 week healing rates were calculated from a meta-analysis of the two available head-to-head studies comparing esomeprazole 40mg (4 weeks - r = 2087, n = 2765; 8 weeks r = 342, n =678) and lansoprazole 30mg (4 weeks - r = 1984, n = 2760; 8 weeks r = 347, n = 776) in the healing of reflux oesophagitis. These studies used the capsules formulation of lansoprazole. This data have been used, as the capsule and oro-dispersible formulation are bioequivalent. Triangular distributions were fitted to utility values reported by a study using the rating scale method in patients with gastro-oesophageal reflux disease. Estimates of resource utilisation were obtained from a survey of UK-based clinicians, and were multiplied by national published resource unit costs at 2003/04 prices. RESULTS: The mean cost per QALY gained with esomeprazole and lansoprazole were £1482 and £1633 respectively. Esomeprazole dominated lansoprazole (i.e. was more effective and less expensive) in 86.8% of the 10,000 Monte Carlo simulation patient iterations. Applying a willingness to pay threshold of £20,000 per QALY similar to that used by NICE indicates that esomeprazole is cost-effective in 98.9% of the 10,000 patient iterations. CONCLUSIONS: Esomeprazole is more cost-effective than the oro-dispersible formulation of lansoprazole in the initial treatment of reflux oesophagitis.