Background: Before the introduction of combination antiretroviral therapy (CART), neurological disease correlated with cerebrospinal fluid (CSF) levels of human immunodeficiency virus (HIV) RNA.

Objective: To investigate the relationships among HIV RNA levels, immune activation markers, and neurological status in patients receiving CART.

Design: Multicenter cohort study.

Setting: Academic neurology departments.

Patients: A total of 371 patients unselected for neurological complaints and with CD4 cell counts less than 200/microL or with cognitive symptoms and CD4 cell counts less than 300/microL were enrolled into the Northeastern AIDS Dementia cohort in 1998-2002. Diagnoses of HIV-associated dementia (HIV-D) and minor cognitive-motor disorder (MCMD) were obtained with a computerized algorithm. Plasma and CSF levels of HIV RNA, monocyte chemotactic protein 1, macrophage colony-stimulating factor, and tumor necrosis factor alpha were quantified.

Results: The mean +/- SD age was 41.5 +/- 7.2 years, and the mean +/- SD educational level was 12.3 +/- 2.2 years. Seventy percent of the cohort was black, and 30% were women. The mean +/- SD CD4 cell count was 136.8 +/- 87.9/microL, and CART was used in 71%. Twenty-nine percent of the patients were unimpaired (n = 106), 36% had MCMD (n = 133), and 35% had HIV-D (n = 128). Mean log(10) CSF HIV RNA copies per milliliter was 2.6 +/- 0.8, with no differences among the neurological groups, even after adjustments for baseline CD4 cell counts and antiretroviral therapy. Cerebrospinal fluid HIV RNA was undetectable in 47% of unimpaired, 46% of MCMD, and 43% of HIV-D patients (P = .91). Plasma levels of monocyte chemotactic protein type 1 and tumor necrosis factor alpha correlated weakly with HIV RNA levels but did not distinguish those with neurological deficits.

Conclusions: In contrast to observations in individuals not treated with CART, we found no relationship between CSF markers and neurological status in this CART-using cohort with advanced HIV/AIDS. This was not explicable by demographic differences or plasma virological control. CART may substantially attenuate the degree of central nervous system HIV infection and immune activation, and in CART users, CSF HIV RNA and immune activation markers may fail to discriminate milder degrees of HIV-D and MCMD.