0595 Inflammatory Plasma Biomarker Cluster Associations with Sleep in People with and without HIV

Introduction Sleep problems are commonly reported in people with HIV (PWH) and may be exacerbated by HIV-induced inflammation. We determined associations between systemic inflammation and objective/subjective sleep measures in PWH and demographically/lifestyle similar HIV-negative controls. Methods Objective sleep measures from 7-day actigraphy (e.g. mean/standard deviation (SD) of wake after sleep onset [WASO], sleep duration/efficiency), overnight oximetry (oxygen desaturation index [ODI]), and patient-reported measures (Insomnia Severity Index [ISI] and Patient-Reported Outcomes Measurement Information System [PROMIS] sleep questionnaires) were assessed in participants in the multicenter POPPY-Sleep Study in the UK and Ireland. Principal Component Analysis using 31 plasma inflammatory biomarkers followed by cluster analysis previously identified 3 distinct inflammatory clusters: 1 (low inflammation), 2 (immune activation) and 3 (systemic inflammation). Baseline characteristics and between-cluster differences in sleep outcomes were assessed using Kruskal-Wallis or logistic regression/Chi-squared tests. Results The 465 participants (74% PWH, median [interquartile range] age 54 [50-60] years) were mainly male (80%), men having sex with men (71%) and white (88%). Among PWH, most (98%) were on antiretroviral therapy, 92% had viral load ≤50 cps/mL and CD4 cell count was 610 [470-785] cells/mm3. Overall, 18% met ISI criteria for insomnia (ISI≥ 15), and other sleep measures suggested generally good sleep (e.g., ODI 3.1/hr [1.5-6.4]). Clusters 1 (n=209), 2 (n=47) and 3 (n=209) differed significantly for HIV status (73%, 60%, 78%, p=0.03); BMI (24.8, 25.9, 26.2 kg/m2, p=0.002); systolic blood pressure (126, 135, 126 mmHg, p=0.002); cardiovascular disease (39%, 28%, 53%, p=0.001) and arthritis (8%, 9%,16%, p=0.02) – all factors associated with sleep problems. There were no clinically relevant between-biomarker-cluster differences in the proportions with insomnia (17%, 18%, 20%) either before (p=0.76) or after (p=0.75) adjustment for potential confounders. Few associations were observed among other actigraphy, oximetry and PROMIS measures. Conclusion Despite observed differences in clinical factors associated with sleep problems, we found no consistent or strong associations between inflammatory biomarker clusters and a range of sleep outcomes. Although associations could exist with other sleep outcomes (e.g. sleep architecture) or biomarker types (e.g. cerebrospinal fluid) not assessed, our findings do not support a strong association between sleep and plasma inflammatory biomarkers in this population. Support (If Any) NIH R01HL131049