CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

417 IMPACT OF WEB-BASED COGNITIVE TRAINING ON WORKING MEMORY IN COCAINE USERS WITH HIV Sheri L. Towe 1 , Yunan Xu 1 , Christina S. Meade 1 1 Duke University, Durham, NC, USA Background: Cocaine use is disproportionately prevalent among persons with HIV, and it is known to exacerbate HIV-associated neurocognitive impairments, specifically working memory, that contribute to impulsive decision making. This study tested the effectiveness of a web-based cognitive training intervention to improve working memory and reduce impulsivity in HIV-infected cocaine users. Methods: In this randomized controlled trial, participants were assigned to one of two conditions of 48 cognitive training sessions, each lasting 20-30 minutes, over 10 weeks. Games in the active condition (ACT) targeted working memory, while games in the control condition (CON) targeted other domains. Each session included a random sampling of 4 out of possible 8 games repeated once back-to-back. Participants completed clinical interviews and comprehensive neuropsychological testing at baseline and post-intervention, as well as a process measure to provide feedback on the intervention. Results: The sample of 58 participants was 48.6 years old on average, mostly male (71%) and African American (88%). Participants completed 37.3 of the 48 possible sessions on average, with no difference by condition, and 56 participants (97%) completed the post-intervention follow-up. We conducted repeated measures ANCOVAs on working memory (domain deficit score) and delay discounting (natural log k-value), controlling for age, education, and number of games improved (as proxy of intervention engagement). In the intent-to-treat sample, there was a significant group-by-time interaction for working memory with a medium effect size [F(1, 51) = 4.470, p = .039, eta squared = 0.081], such that ACT had greater improvements relative to CON. For delay discounting, there was a similar pattern, again with a medium effect, but the interaction effect was not significant [F(1, 48) = 3.546, p = .066, eta squared = 0.069]. Overall, participants rated the sessions as helpful (M=4.09, on 5 point scale), but those in ACT perceived greater improvement on the games over time [M=4.39 (0.69) vs. 3.89 (0.92); t(54)= 2.31, p= .025]. Conclusion: Our findings support the acceptability and potential effectiveness of cognitive training to improve working memory in HIV-infected cocaine users. A larger trial with a longer duration of training targeting more domains is needed to test the durability of effects and improvement in daily living. 418 RETINAL THINNING CORRELATES WITH BRAIN ATROPHY IN WELL- CONTROLLED HIV INFECTION Bryan Smith 1 , Katrina Geannopoulos 1 , Ramiro Maldonado 2 , Tianxia Wu 1 , Elizabeth F. Horne 1 , Lillian Ham 2 , Joseph Snow 2 , Govind Nair 1 , Daniel S. Reich 1 , Chuen-Yen Lau 3 , Emily Chew 2 , Avindra Nath 1 1 National Institute of Neurological Disorders and Stroke, Bethesda, MD, USA, 2 NIH, Bethesda, MD, USA, 3 NIAID, Bethesda, MD, USA Background: Retinal measurements correlate well with neurologic disease in multiple sclerosis, however whether such measurements correlate with neurologic disease in well-treated persons living with HIV (PLWH) is unknown. We evaluated differences in retinal measures by spectral domain optical coherence tomography (SD-OCT) between PLWH and uninfected controls and correlations with the retinal measures and brain volumes, neuropsychological (NP) function, and markers of neuronal injury and neuroinflammation. Methods: SD-OCT was performed on 69 PLWH on antiretroviral therapy (ART) and 28 uninfected controls. Participants also underwent brain MRI, neuropsychological testing, and an optional lumbar puncture. All procedures, including the SD-OCT, were completed for research only and there were no

416 MITOCHONDRIAL DNA, COGNITIVE FUNCTION, AND FRAILTY IN OLDER ADULTS WITH HIV Carrie Johnston 1 , Michelle Rice 1 , Heather Derry 1 , Chelsie Burchett 1 , Eugenia

Siegler 1 , Mary Choi 1 , Marshall Glesby 1 1 Weill Cornell Medicine, New York, NY, USA

Poster Abstracts

Background: Older PLWH experience more comorbidities and geriatric syndromes, including cognitive impairment and frailty. Mitochondrial DNA (mtDNA), released from dying cells, is a biomarker of inflammation, a mediator of immune activation and has been detected at elevated levels in the plasma of PLWH. We hypothesized that in older PLWH, plasma mtDNA would be associated with lower cognitive performance, frailty, and higher serum IL-6 level. Methods: We analyzed cross-sectional data from PLWH over age 55 at a single urban medical center. Participants completed a psychosocial questionnaire, biomedical visit, cognitive assessment (Montreal Cognitive Assessment, MoCA) and frailty testing by Fried criteria. Plasma and urine cell-free mtDNA were measured by qPCR detection of mitochondrial NADH dehydrogenase-1. Results: There were 164 participants; mean age 61 (SD: 6) range 54-87 years. One-third (55) were female. Half identified as Black, 29% as White, and 21% as other. Median time living with HIV was 25 years (IQR 22-29). The majority (93%) had HIV-1 viral load <200 copies/mL and median CD4 count was 582 cells/ uL (IQR 402-795). Geometric mean mtDNA level in plasma was 221 copies/µl (geometric SD: 2) and 2.4x108 copies/gram of urine creatinine (geometric SD: 4) in urine. Levels of plasma and urine mtDNA (Spearman correlation rho, ρ=0.05, p=0.54) were unrelated. Median MoCA score was 24 (IQR 21-26); 63 (39%) scored <23. 49 (30%) were non-frail, 95 (58%) pre-frail and 20 (12%) frail. Age was not related to MoCA score (ρ=-0.1, p=0.19), but was associated with frailty status by Jonckheere-Terpstra (JT) test (p=0.008). Plasma mtDNA level was higher in those with low MoCA score (p=0.028) by t-test [Figure 1]. Higher plasma mtDNA level was associated with slow walk (p=0.007) and exhaustion (p=0.04), but not weight loss (p=0.62), grip strength (p=0.06), low physical activity (p=0.71) or composite frailty score (p=0.98). Serum IL-6 levels were associated with frailty status (p=0.018) but not with low MoCA score (p=0.89 by JT). Neither plasma nor urine mtDNA levels were correlated with IL-6 level, ρ-0.05 (p=0.55) and 0.09 (p=0.29), respectively. Conclusion: In this study we show a relationship between elevated levels of plasma mtDNA and lower performance on the MoCA, greater exhaustion, and slower walk, suggesting mtDNA may have a role as a novel biomarker in assessing pathogenic inflammation associated with cognitive dysfunction and some components of frailty in PLWH.

CROI 2020 145