There is a lack of methodological conformity in cost-effectiveness analyses of hypertension treatment. They differ with respect to assumptions about the effectiveness of treatment, the outcome measure chosen, the cost-concept, the discounting of effects and the duration of therapy. The aim of this paper is to review these issues and estimate the importance of different assumptions for the cost per life-year gained. To analyse these assumptions a computer simulation model was constructed based on the Framingham logistic risk equations and Swedish cost data. It is shown that the cost per life-year gained is highly sensitive towards many of these assumptions. It is also shown that the average cost-effectiveness ratios calculated in previous studies and the relevant marginal cost-effectiveness ratios can differ by several hundred per cent. The results of cost-effectiveness analyses in the hypertension field have to be interpreted with caution. Due to the lack of standardized methodology, the comparability between studies is limited. There is also a need to complement cost-effectiveness analysis in this area with other approaches, for example based on WTP.