Impact of Integrase Strand Transfer Inhibitors on Cognition in the HAILO Cohort

Abstract

Background: Integrase inhibitors (INSTIs) have been associated with poorer cognition in people with HIV (PWH). We examined the impact of switching to INSTIs on neuropsychological (NP) outcomes in PWH 40 years of age and older. Methods: From the AIDS Clinical Trials Group observational cohort study, HAILO, we identified PWH who switched to INSTIs, had ≥2 NP assessments before and at least 1 after switch, and maintained viral suppression while on INSTIs. NP performance was assessed with a composite score (NPZ4) including Hopkins Verbal Learning Test (HVLT-R), Digit Symbol test (DSY), Trail Making A, and Trail Making B, while adjusting for covariates and learning effects. Outcomes changes from preswitch and postswitch periods were estimated using piecewise linear mixed models. Results: Among 395 PWH (mean age 54 years, 81% male, 20% Hispanic, and 29% Black) NPZ4 increased preswitch and postswitch. There was no difference in slopes between periods for NPZ4 [preswitch 0.036/year (95% CI: 0.03 to 0.043); postswitch 0.022/year (95% CI: 0.006 to 0.005); P = 0.147]. All tests scores improved preswitch (P < 0.01). Postswitch, Trail Making A and DSY increased (all P < 0.01) without differences in rate of change (all P > 0.05). HVLT-R had a nonsignificant decrease postswitch (P = 0.22), resulting in a significant preswitch vs postswitch difference in slopes (P = 0.03). Conclusions: NP performance improved regardless of INSTI use. There was an attenuation of improvement in verbal memory in the postswitch vs preswitch period. The clinical significance of these changes is unclear but, overall, INSTIs did not have a consistent detrimental effect on NP outcomes.