OBJECTIVE To review current knowledge of anti-HIV therapy and the implications for patient management. BACKGROUND In three large clinical studies, combination therapy with zidovudine plus didanosine or zalcitabine improved survival and delayed clinical events in comparison with zidovudine alone, with greater benefits in zidovudine-naive than zidovudine-experienced patients. Initial studies suggest addition of an HIV protease inhibitor to combination therapy with two nucleoside analogues may result in greater reduction in viral loads, raising hopes about eradication therapy. Two clinical endpoint studies involving protease inhibitors are available. In one of these, saquinavir/zalcitabine combination therapy offered significant clinical benefits over monotherapy with either drug and in the other ritonavir improved survival in individuals with late HIV infection. It remains unclear at what stage therapy should be started. Evidence from clinical studies suggests initial therapy should be a combination regimen, probably two nucleoside inhibitors. Data from surrogate marker studies suggest that even greater reductions in viral load and increments in CD4+ count can be seen using triple therapy, either with two nucleosides and a non-nucleoside reverse transcriptase inhibitor, or two nucleoside analogues and a protease inhibitor. It remains unclear whether there is a qualitative difference between reducing viral load below detectable levels and more modest reductions in viral load, and whether immune destruction caused by HIV can be reversed and if sustained suppression of HIV replication can be achieved. CONCLUSIONS It is now possible to offer real hope of improvements in survival to HIV-infected patients, through the use of combinations of antiretroviral agents, but more evidence from ongoing and further studies is still needed.