Since 1996 and the introduction of highly active antiretroviral therapies, multiple drugs have been developed. The best choice of drugs to start with still remains debated, especially regarding the choice between protease inhibitor-based or nonnucleoside reverse transcriptase inhibitor-based regimens. Regarding the regimen's ability to control viral replication, both have been proven to be very efficient in large settings of patients. Regarding short-term tolerability, both classes are responsible for various side effects in 10-25% of patients. Regarding long-term tolerability, protease inhibitor-based regimens may be responsible for metabolism abnormalities (hypertriglyceridemia, diabetes), although the role of the associated nucleoside reverse transcriptase inhibitor is not always clear. Patient adherence to the regimen is the cornerstone of efficacy. Resistance acquisition in case of poor adherence leading to treatment failure is a key issue regarding the ability to build future efficient regimens. Failure while using protease inhibitor-based regimens seldom leads to acquired resistance. Failure while using nonnucleoside reverse transcriptase inhibitor-based regimens leads to resistance, but the new available nonnucleoside reverse transcriptase inhibitors do not share the same patterns of resistance, so that the well-known "class resistance" in case of nonnucleoside reverse transcriptase inhibitor failure is no longer valid. Drug-drug interactions are frequent with all antiretrovirals and require close monitoring. In conclusion, protease inhibitor-based and nonnucleoside reverse transcriptase inhibitor-based regimens each have many advantages. The best choice will essentially be made taking into account the patients' characteristics.