(1) The reference antiretroviral treatment for HIV-infected patients is a three-drug regimen combining drugs from different families, namely two nucleoside reverse transcriptase inhibitors + a protease inhibitor or a non nucleoside reverse transcriptase inhibitor. In our opinion, indinavir is the first-choice protease inhibitor. (2) The clinical file on amprenavir is limited. In particular, it is not known whether the recommended dose regimen (1 200 mg twice a day) is the best one. A 48-week trial involving 232 patients previously untreated with antiretrovirals showed that the combination of amprenavir (1 200 mg) + lamivudine (150 mg) + zidovudine (300 mg), administered twice a day, was more effective than lamivudine + zidovudine on viral load. (3) An unblinded trial involving 504 patients, who had already received a reverse transcriptase inhibitor but no protease inhibitor, compared three-drug regimens consisting of amprenavir (1 200 mg twice a day) and two reverse transcriptase inhibitors, or indinavir (800 mg three times a day) and two reverse transcriptase inhibitors, for 48 weeks. Changes in viral load and the CD4+ lymphocyte count did not favour the amprenavir-containing regimen. (4) Laboratory studies suggest that the risk of cross-resistance between amprenavir and other HIV protease inhibitors is lower than with other compounds of this class. However, it is unclear whether this means that amprenavir would be clinically effective after failure of one or several other protease inhibitors. (5) The main adverse effects of amprenavir are gastrointestinal disorders and paraesthesias. Overall, the safety profile resembles that of the other protease inhibitors. (6) Like other protease inhibitors, amprenavir can potentially interact with many other drugs, owing to its hepatic metabolism. (7) Amprenavir is taken in two daily doses irrespective of meal times, but this advantage over other protease inhibitors is cancelled out by the fact that patients have to swallow 8 large soft capsules at each intake. (8) In our opinion, pending further data, amprenavir should be used only in well-conducted clinical trials.