Incidence, risk factors and neurocognitive impact of CSF viral escape: Longitudinal analysis of the CHARTER & HNRP Cohorts.

Type: Poster
Title: Incidence, risk factors and neurocognitive impact of CSF viral escape: Longitudinal analysis of the CHARTER & HNRP Cohorts.
Authors: Perez-Valero I, Letendre S, Deutsch R, Heaton R, Clifford DB, McArthur J, Morgello S, Gelman B, Collier A, Grant I, the CHARTER Group, and the HNRP Group.
Year: 2013
Conference: Conference on Retroviruses and Opportunistic Infections

Background: No large longitudinal studies have evaluated risk factors incidence, evolution and consequences of CSF viral escape (CVE).

Methods: Subjects on antiretroviral therapy (ART) with plasma viral suppression (<50 cop/mL) at baseline were selected. Demographics, neurocognitive performance (NP), clinical data and blood/CSF tests results were recorded. 3 types of CVE were defined: CSF blip [single occurrence of CVE while suppressed in plasma], persistent CVE (P-CVE) [≥2 consecutive CVE while suppressed in plasma] and CVE next to a period of loss of HIV-suppression in plasma (CVE-LS). Rates of CVE were reported as incidence rates. To identify risk factors of CVE, variables were identified univariably (p<0.10) and multivariable models were developed (minimal AIC) using mixed effects logistic regression for all visits where HIV-RNA was <50cop/mL in plasma. To analyze the impact of CVE in the evolution of NP, a multiple linear regression model including demographics & GDS change was performed for subjects that were always undetectable in plasma and have never performed a previous neurocognitive assessment.

Results: 849 HIV+ subjects (mean education 12.8 years, mean age 45 years, Caucasian 49.8% & male 80.9%) were included. Their median CD4 nadir was 120 cells/mm3, 24.3% had hepatitis C and 70% AIDS. At baseline, subjects had been on ART for median of 4.6 years and their median CD4 was 456 cells/mm3. 49.6% were treated with a PI-based and 35.5% with NNRTI-based ART. 37.3% were cognitively impaired. Subjects were followed for a median of 30.3 months & 3304 visits were analyzed. CVE was detected in 88 visits (2.7%) in 60 subjects (7.1%): CSF blip (45 visits – 41 subjects), persistent CVE (20 - 9) and CVE-LS (23 - 17). The incidence rate of CVE was 37.4 cases per 1000 person-years (CSF blip 19.1, P-CVE 8.5 & CVE-LS 9.8). CVE was associated with detection of HIV-RNA in plasma below 50 cop/mL [p=0.03; OR: 1.6 (1.1-2.4) per +10cop/mL] and WBC in CSF [p<0.01; OR: 3.1 (1.9-4.9) per +5 cells]. NP evolution was analyzed in 236 subjects with (n=13) & without (n=223) CVE. Mean GDS changes were -0.1±0.34 (CVE) & -0.02±0.38 (No CVE). CVE development was not associated (p=0.84) with differences NP evolution.

Conclusions: CVE is uncommon and generally transitory. CVE is associated with HIV low-level viremia in plasma and with the number of WBC in CSF. We have not found an association between CVE and NP evolution.