Background: Triple combination antiretroviral therapy, recommended as standard of care, is unaffordable for much of the developing world.Objectives: To establish whether half doses of zidovudine (AZT) and zalcitabine(ddC) are as effective as standard doses in a Thai population with lower body weight than Western populations and predominantly infected with HIV-1 subtype E.Methods: A group of 116 antiretroviral naive patients, with CD4 cell counts 100±500 3 106 cells/l, were randomized to: AZT 200 mg three times daily plus ddC0.75 mg three times daily versus AZT 100 mg three times daily plus ddC 0.375 mg three times daily and followed-up regularly for 48 weeks.Results: The study enrolled 111 patients: 59 men and 52 women, body weight (mean±standard deviation) 56.4 ±12.3 kg, mean CD4 cell count 324 ´ 106 cells/l, mean HIV RNA 4.7 log10 copies/ml. There were no signi®cant differences between the two groups. Twelve patients discontinued, including two deaths that were unrelated to study medication. No signi®cant differences in adverse events were seen. Week 48 data for the standard dose and half dose arms, respectively, were mean CD4 cell count increases of 52 and 78 3 106 cells/l (P . 0.34), mean plasma HIV-1 RNA reduction of1.4 and 1.1 log10 copies/ml (P . 0.10), HIV RNA of , 400 copies/ml in 52 and 20%{(P . 0.001). Participants with higher than mean baseline CD8 cell counts (mean1062 3 106 cells/l) showed greater decline in CD8 cells on standard doses. Further analysis showed improved reduction in HIV RNA (P , 0.0001) and in the percentage with undetectable HIV RNA (P . 0.0137) in the standard dose arm, corrected for baseline HIV RNA, which if , 4.75 log10 copies/ml signi®cantly correlated with HIV RNA , 400 copies/ml at week 48.Conclusion: At week 48, the proportion with HIV RNA , 400 copies/ml was signi®-cantly higher in the standard dose arm; lower baseline HIV RNA correlated with better HIV RNA outcome at 48 weeks. The arms did not differ in CD4 cell response but standard doses correlated with greater CD8 cell decline.