Human immunodeficiency virus (HIV)-1 invades the central nervous system (CNS) soon after infection and is partially protected there from host immunity and antiretroviral drugs (ARVs). Sanctuary from highly active antiretroviral therapy (HAART) in the CNS could result in ongoing viral replication, promoting the development of drug resistance and neurological disease. Despite the importance of these risks, no previous study has directly assessed HAART's effects on brain tissue viral load (VL). The authors evaluated 61 HIV-infected individuals for whom both histories of HAART treatment and postmortem brain tissue VL measurements were available. Two groups were defined based on HAART use in the 3 months prior to death: HAART(+) subjects had received HAART, and HAART(-) subjects had not received HAART. HIV RNA was quantified in postmortem brain tissue (log10 copies/10 microg total tissue RNA) and antemortem plasma (log10 copies/ml) by reverse transcriptase-polymerase chain reaction (RT-PCR). Brain tissue VLs were significantly lower among HAART(+) subjects compared to HAART(-) subjects (median 2.6 versus 4.1; P= .0007). These findings suggest that despite the limited CNS penetration of many antiretroviral medications, HAART is at least partially effective in suppressing CNS viral replication. Because some HAART regimens may be better than others in this regard, regimen selection strategies could be used to impede CNS viral activity, limit neuronal dysfunction, and prevent or treat clinical neurocognitive disorders in HIV-infected patients. Furthermore, such strategies might help to prevent the development of ARV resistance.