Depression and suicidal ideation among HIV-infected adults receiving efavirenz versus nevirapine in Uganda: A prospective cohort study

Background: Evidence regarding potential adverse neuropsychiatric effects of efavirenz is conflicting, and data from sub-Saharan Africa, where 70% of persons living with HIV (PLHIV) reside and efavirenz is used as first-line therapy, are limited. Objective: To estimate associations between efavirenz use and depression and suicidal ideation among PLHIV in Uganda. Design: Prospective observational cohort study. (ClinicalTrials .gov: NCT01596322) Setting: Mbarara, Uganda. Participants: Adult PLHIV enrolled at the start of antiretroviral therapy (ART) and observed every 3 to 4 months from 2005 to 2015. Measurements: The exposure of interest was time-varying efavirenz use, defined as use during the 7 days and in 60 or more of the 90 days before a study visit, compared with nevirapine use. Self-reported outcomes were depression, defined as a mean score greater than 1.75 on the Hopkins Symptom Checklist depression subscale, and suicidal ideation. Multivariable-adjusted generalized estimating equations (GEE) logistic regression, Cox proportional hazards regression, and marginal structural models were fit to estimate the association between efavirenz use and the risk for depression and suicidal ideation. Results: 694 participants (median age, 33 years; median pretreatment CD4 ⁺ count, 1.8 × 10 ⁹ cells/L) contributed 1200 person-years of observation (460 person-years receiving efavirenz). No baseline differences in depression or suicidal ideation were found between patients ever exposed to efavirenz and those never exposed to efavirenz and receiving nevirapine (P > 0.80 for both). Of 305 participants ever-exposed to efavirenz, 61 (20.0%) and 19 (6.2%) had depression and suicidal ideation, respectively, on at least 1 follow-up visit, compared with 125 (32.1%) and 47 (12.1%) of the 389 who received nevirapine. In adjusted GEE models, efavirenz use was associated with decreased odds of depression compared with nevirapine use (adjusted odds ratio, 0.62 [95% CI, 0.40 to 0.96]) and was not significantly associated with suicidal ideation (adjusted odds ratio, 0.61 [CI, 0.30 to 1.25]). Time-to-event and marginal structural models yielded similar estimates. Limitation: Nonrandom assignment to treatment and substantial differences between the efavirenz and nevirapine groups. Conclusion: No evidence was found that use of efavirenz in firstline ART increased the risk for depression or suicidal ideation compared with nevirapine use among PLHIV in Uganda.