The use of combination antiretroviral therapy is supported by clinical evidence for its superiority over monotherapy. Lamivudine (3TC) has been studied in combination with zidovudine (ZDV) and is recommended for use specifically in combination therapy. With the associated increase in drug acquisition cost, the economics of combination therapy versus monotherapy warrant study. An economic evaluation was undertaken to compare 3TC/ZDV combination therapy with ZDV monotherapy, taking a UK public finance perspective. The cost effectiveness of each of the 2 treatments was estimated using a Markov model of progression through 3 HIV-positive disease states: (i) CD4 cells > 200 and < 500 cells/mm3; (ii) CD4 < 200 cells/mm3, non-AIDS; and (iii) AIDS to eventual death. Progression probabilities and life expectancy were derived from a cohort treated at Chelsea and Westminster Hospital in London, using data for 1987 to 1995, along with cost data for a ZDV intent-to-treat population for 1994 and 1995. The relative risk of progression for 3TC/ZDV compared with ZDV monotherapy was estimated from meta-analysis of 4 completed comparative trials. To predict the effect of 3TC/ZDV on life expectancy and lifetime costs, progression probabilities were adjusted by the relative risk statistic for the duration of treatment with 3TC/ZDV. On the basis of an estimated relative risk of progression of 0.509 (95% CI 0.365 to 0.710), treatment with 3TC/ZDV is predicted to yield an incremental cost-effectiveness ratio of Pounds 6276 (95% CI Pounds 5337 to Pounds 9075) per life year saved (discounted at 6% per year). Extensive sensitivity analyses were performed to test the effects of varying values of input parameters on the model results. Under reasonable assumptions, the predicted cost effectiveness of 3TC/ZDV combination therapy compares favourably with previously reported economic analyses of various HIV treatments.