The relationship of serum human immunodeficiency virus-1 (HIV-1) RNA levels to HIV-1 RNA levels in other compartments, such as the lungs, is not well characterized. The purpose of this study was to determine the viral burden of HIV-1 in the lungs by comparing HIV-1 RNA in cell-free bronchoalveolar lavage fluid (BALF) with that in serum. Specimens were examined from 77 HIV-seropositive adults (CD4(+) cell counts: 0 to 700 cells/mm(3); 48% receiving prescribed antiretroviral agents), comprising 43 asymptomatic individuals who were compared with 34 persons with active lung disease caused by Pneumocystis carinii (n = 26), bacteria (n = 3), Mycobacterium avium complex (n = 2), Nocardia sp. (n = 1), Aspergillus sp. (n = 1), or pulmonary Kaposi's sarcoma (n = 1). For serum HIV-1 RNA, the proportion of subjects with detectable levels and the mean values were similar for asymptomatic individuals and persons with active lung disease (85% versus 86%, respectively) (6.64 x 10(4) versus 1. 81 x 10(5) HIV-1 RNA copies/ml; p = 0.13). In contrast, HIV-1 RNA in BALF was more often detected (16% versus 62%; p = 0.001), and mean values were higher (1.04 x 10(5) versus 3.31 x 10(6) HIV-1 RNA copies/ml; p = 0.032), in subjects with active lung disease than in asymptomatic subjects, independent of early or advanced clinical stages of HIV-related disease. For both study groups, HIV-1 RNA levels in BALF exceeded those in serum in 56% of cases by up to 66-fold, and did not correlate with local levels of tumor necrosis factor-alpha, granulocyte-macrophage colony-stimulating factor, or interleukin-16. HIV-1 proviral DNA in cells from BALF was detected in up to 86% of subjects, more frequently in persons with advanced HIV disease (p = 0.0496), and often involved > 10% of BALF cells, but did not correlate with HIV-1 RNA detected in BALF. These data provide evidence for active HIV-1 replication in the lungs. HIV-1 replication is compartmentalized relative to serum, may be restricted, is independent of HIV-1 proviral DNA and clinical stage of HIV, and may be influenced by pulmonary disease such as P. carinii pneumonia or by other local or lung-specific factors. The lungs represent a large reservoir for HIV-1, and may present a source of persistent HIV-1 replication even during periods of apparent clinical latency of HIV-1 infection.