Objectives: Associations between inflammatory markers and prevalent or incident frailty, cognitive impairment, clinical events and mortality in older people with HIV are poorly understood.

Design: An observational cohort study.

Methods: Participants ≥50 years from the ACTG HAILO cohort study were included. Participants completed annual evaluations for cognitive impairment and frailty. Clinical events included non-AIDS-defining cancers, diabetes and cardiovascular, liver and kidney diseases. Associations between inflammatory markers (hsCRP, IL-6, TNFR1, CXCL-9 and inflammatory index score [IIS]) at baseline and prevalence and incidence of frailty, cognitive impairment, any clinical event and non-accidental mortality were examined. We used 10-fold cross validation to examine whether the combination of inflammatory markers and frailty improved the ability to predict incident outcomes.

Results: Among 484 participants (17% assigned female at birth, 25% Black and 20% Hispanic), median age was 56 years. Median BMI was 27 kg/m 2 , median CD4 count was 627 cells/mm 3 , and 95% had HIV-1 RNA <200 copies/mL. HsCRP, IL-6, TNFR1, CXCL-9 and IIS were associated with increased risk of prevalent frailty and clinical events, but not cognitive impairment. CXCL-9 (Q4 vs. Q1) and TNFR1 were associated with an increased incidence of both frailty and clinical events; Q4 vs Q1 of the IIS was associated with clinical events; increased inflammatory markers (except CXCL-9) were associated with an increased risk of mortality. TNFR1 combined with frailty modestly improved the predictability of incident clinical events and mortality over frailty alone.

Conclusions: Several inflammatory markers were associated with increased risk of frailty, clinical events, and mortality, but not cognitive impairment.

Keywords: CXCL-9; clinical events; frailty; inflammation; inflammatory index score.