Objective: Neuropathological abnormalities induced by HIV-1 are not al
ways predictable on the basis of the presence of HIV-related neurologi
cal symptoms. HIV-1 RNA load was measured in the cerebrospinal fluid (
CSF) of HIV-infected patients to verify whether it could be a marker o
f HIV-induced neuropathology. Design and methods: Histopathological an
d immunohistochemical examination of the brain for HIV-1 p24 antigen w
as performed in 50 HIV-infected patients with neurological symptoms; p
atients were defined as having HIV encephalitis in the presence of HIV
-related lesions or HIV-1 p24 antigen-positive cells. Quantitative pol
ymerase chain reaction for HIV-1 RNA was retrospectively applied to CS
F samples that had been drawn 1-60 days prior to death from these 50 p
atients; paired plasma samples of 28 patients were also analysed. Resu
lts: The CSF HIV-1 RNA copy numbers were significantly higher in 22 pa
tients with HIV encephalitis than in 28 patients without (median, 4.77
log(10) versus 3.45 log(10) copies/ml; P = 0.0003). No correlation wa
s found between CSF HIV-1 RNA load and the presence of opportunistic b
rain pathologies at post-mortem examination or between HIV-1 RNA loads
in paired CSF and plasma samples. Conclusions: High CSF HIV-1 RNA lev
els are associated with HIV encephalitis, regardless of the presence o
f opportunistic brain diseases or HIV-1 RNA levels in plasma. Quantita
tive CSF HIV-1 RNA may therefore be used as a specific marker of HIV-i
nduced neuropathology. (C) 1998 Rapid Science Ltd.