Quantifying HIV-1 RNA using the polymerase chain reaction on cerebrospinal fluid and serum of seropositive individuals with and without neurologic abnormalities.

We quantified HIV-1 RNA levels (copies per milliliter) in cerebrospinal fluid (CSF) and serum from subjects at various stages of HIV-1 disease and determined the relationship of RNA levels to clinical and neurologic disease status (HND) and to laboratory values. Ninety-seven HIV-1-seropositive men without CNS opportunistic infections, tumors, or neurosyphilis and 13 high-risk seronegative controls were included in the study. Each individual underwent a structured interview and physical and neurologic examinations, followed by standardized collection of blood and CSF. A custom-designed, fully automated polymerase chain reaction (PCR) system was used to perform a minimum of four separate amplifications per specimen, using two HIV-1 gag primer pairs. Southern blotting followed by hybridization with product-specific probes was used for post-PCR detection. The number of copies per milliliter was determined by relating unknowns to a built-in dilution-series standard curve using an image analysis system. HIV-1 RNA was detectable in 96% of the sera, 78% of the concentrated CSF samples, and 54% of the unconcentrated CSF samples. Serum RNA levels were significantly higher than in CSF. Serum RNA levels were significantly inversely correlated with CD4+ cell counts (p = -0.34; p = 0.03): i.e., higher RNA levels in seropositive subjects were associated with lower numbers of CD4+ cells. Serum RNA levels correlated positively with number of AIDS-related symptoms, dysfunction scores for total neurological examination, mental status score, cranial nerve score, and CNS motor signs score. Serum RNA levels did not correlate significantly with length of time on zidovudine therapy, intrathecal IgG synthesis rate, or albumin leakage. RNA levels in CSF significantly correlated only with intrathecal IgG synthesis rate and with serum RNA levels. These results confirm that serum levels of HIV-1 RNA correlate with HND and inversely correlate with CD4 counts, demonstrating that HND occurs predominantly in late stages of HIV-1 disease, although HIV-1 RNA can be detected in CSF from a majority of HIV-1-seropositive individuals at all stages of disease, which suggests that there can be early penetration of HIV into the CNS. However, HND can occur in the absence of high levels of CSF HIV-1 RNA. We also found that the concentration of HIV-1 in CSF is correlated with intrathecal IgG synthesis rate.