Association of High Ratio of CSF/Plasma HIV-1 RNA with Central Nervous System Co-Infection in HIV-1-Positive Treatment-Naive Patients

Abstract

Cerebrospinal fluid (CSF) human immunodeficiency virus-1 (HIV-1) ribonucleic acid (RNA) at higher levels than in plasma has been observed in HIV-1-positive patients and defined as CSF/plasma discordance or CSF escape. Discordance is particularly seen in untreated patients with antiretroviral agents. Quantitative data regarding its association with blood–brain barrier (BBB) damage and intracranial co-infection with other pathogens are limited. Therefore, we used the CSF to plasma HIV-1 RNA ratio (HRR) to determine its relation to central nervous system (CNS) co-infection in HIV-1-positive treatment-naïve individuals. We retrospectively recruited the subjects with HIV-1-positive and potential neurological deficits. A lumbar puncture was performed before the antiretroviral therapy. The paired CSF/plasma HIV-1 RNA samples were analyzed. Univariate and multivariate logistic regression models and multiple spine regression analyses were performed to assess the association between the HRR and CNS co-infection. A total of 195 patients with 78% males (median age: 49 years) were included in this study, of whom 98 (50.2%) had CNS co-infection with other pathogens. The receiver-operating characteristic curve analysis showed that the optimal cutoff value for the HRR to predict the CNS co-infection was 1.00. Higher HRR (≥1) was significantly associated with tuberculous meningitis (OR 6.50, 95% CI 2.08–20.25, p = 0.001), cryptococcus meningitis (OR 7.58, 95% CI 2.10–27.32, p = 0.001), and multiple co-infection (OR 4.04, 95% CI 1.02–16.04, p = 0.047). Higher HRR (≥1) (OR 3.01, 95% CI 1.09–8.73, p = 0.032) was independently associated with the CNS co-infection after adjusting for covariates. No significant nonlinear association was found between the HRR and CNS co-infection in the multivariate spline regression (p > 0.05) and a positive relationship was found between the HRR and CNS co-infection when the HRR was ≥0.78. Higher HRR was associated with an increased risk of CNS co-infection in HIV-1-positive patients. The relationship between the HRR and CNS co-infection may be related to the BBB disturbance and warrants further investigation with a large, longitudinal cohort.