Abstract Body

Background

Little is known about variations in cognitive performance over the long term among older adults with HIV (OAWH) and the risk of progressive cognitive decline may be overstated based on transient decline between two assessments. This analysis aimed to characterize visit-to-visit and overall evolution of cognition over up to 10 years in OAWH.

Methods

268 participants (87% men; mean duration of HIV 17.3 years, SD: 8.0; 95% undetectable) of the multisite Canadian + Brain Health Now study were followed for ≥ 5 years (mean 6 visits over 7.7 years). They were on average 53.7 years of age (range: 35.1 to 81.2). Cognition was assessed every 9-12 months using the B-CAM, a computerized battery assessing multiple domains. 1) to identify distinct groups of individuals with a similar evolution over time, we used Group-Based Trajectory Analysis of B-CAM scores, using absolute (sensitive to initial value) and centered (more sensitive to change) values 2) to assess if shorter term decline predicted longer term decline, for each pair of visits of a given participant, we assessed meaningful change (≥ 0.5 SD difference), and compared proportions of episodes of such change in the stable and declining trajectories, and 3) among those with at least one episode of decline between visits, Poisson regression was used to compare those on a declining versus stable trajectory on personal characteristics.

Results

Two trajectories of centered B-CAM were present: stable (68.7%) and declining (31.3%). Uncentered B-CAM had 5 patterns: low (intercept: 0.0, 4.9%), medium-low (0.7, 34.2%), intermediate (1.3, 41.9%), medium-high (1.9, 14.1%) and high (2.6, 4.9%). Trajectories of intermediate and higher uncentered B-CAM were less likely to follow a declining trajectory. Experiencing at least one episode of meaningful worsening between visits was common, 50% in the stable group and 100% in the overall declining group, and was associated with an increasing likelihood of a declining trajectory in older participants (RR 2.0 per decade).

Conclusions

Over 10 years, more than half of OAWH had at least one meaningful worsening in cognitive performance which did not reliably predict long-term decline as only a quarter had sustained cognitive decline. Older individuals and those with worse overall performance were at higher risk. These findings suggest that shorter-term worsening need not portend longer-term decline, although it may be more worrisome in older people.