Long-term efavirenz use is associated with worse neurocognitive functioning.

Type: Poster
Title: Long-term efavirenz use is associated with worse neurocognitive functioning.
Authors: Letendre S, Vaida F, Wong S, Croteau D, FitzSimons C, Grant I, Ellis R.
Year: 2013
Conference: Conference on Retroviruses and Opportunistic Infections

Background: Neurocognitive (NC) complications continue to afflict a substantial proportion of people living with HIV taking effective antiretroviral therapy (ART). One explanation for this is antiretroviral neurotoxicity. For instance, efavirenz (EFV) is associated with short-term central nervous system (CNS) side effects and an efavirenz metabolite can cause dendritic spine injury in vitro. Our objective was to compare long-term use of EFV to a comparator, lopinavir-ritonavir (LPV/r) in a cohort of well-characterized subjects.

Methods: 445 subjects were selected from the CHARTER cohort because they were using either EFV (n=272, mean duration 17.9 months) or LPV/r (n=173, mean duration 16.4 months) and did not have severe NC comorbidities. Standardized comprehensive NC testing was summarized as global and domain deficit scores as well as by impairment ratings. Univariable analyses were performed using parametric and non-parametric tests as appropriate and multivariable regression using Akaike Information Criterion selection.

Results: Compared with LPV/r users, EFV users were more likely to be taking their first ART regimen (36% vs. 14%, p<.001); less likely to have AIDS (64% vs. 82%, p<.001) or HCV (23% vs. 32%, p=.05); had higher CD4+ T-cell nadirs (187 vs. 129/mm3, p<.001) and lower peak (5.3 vs. 5.6 log10 c/mL, p<.001) and current (2.1 vs. 2.6 log10 c/mL, p<.001) plasma HIV RNA levels; and were less likely to have detectable HIV RNA in CSF (8% vs. 26%, p<.001). Overall, EFV users had worse verbal fluency (p=.03), speed of information processing (p=.04), and working memory (p=.03). In the subgroup with undetectable plasma viral loads (n=269), EFV users had worse speed of information processing (p=.02) and executive functioning (p=.03). An interaction with HCV serostatus was identified in multivariable analyses. Among HCV seronegative individuals (n=329), EFV users had worse speed of information processing and executive functioning. Among HCV seropositive individuals (n=116), LPV/r users had worse learning, recall, and motor functioning.

Conclusions: Substantial differences exist between EFV and LPV/r users in observational cohorts, possibly because of channeling by clinicians who appear to prescribe LPV/r to more severely ill patients or as second-line therapy. Despite these differences, EFV users had worse NC functioning in several cognitive abilities. A potentially important interaction was identified that could indicate that the neurotoxicity of specific antiretrovirals may differ based on HCV co-infection. The complexity of these data is substantial and differences would best be evaluated in a randomized clinical trial.