2360. Serum Cholesterol Variability and Mortality in PLWH On Lipid-Lowering Therapy

Abstract

Abstract Background In people living with HIV (PLWH), consistent use of lipid lowering therapy (LLT) is associated with lower all-cause mortality and use of tenofovir disoproxil fumarate (TDF) with reduced serum cholesterol (C). In the general population, high cholesterol variability (CV) is independently associated with all-cause mortality. We explored the relationship between consistency of LLT/TDF exposure, cholesterol variability, and mortality in PLWH. Methods In a cohort of 15,860 PLWH who achieved undetectable viremia on ART between 2001 and 2011 we used multivariable Cox modeling, stratified by prevalent cardiovascular disease (CVD) to identify predictors of mortality and built multi-level marginal structural models for different LLT/TDF exposure levels based on percentage of days covered (PDC) over the past year, categorizing exposures into low (PDC < 50%), high (50–93%), and consistent (≥94%). The latter two were further subdivided into single and dual exposures for a composite LLT model where we explored effect sizes before and after inclusion of time-updated absolute C values and its 1-year coefficient of variation (CoV). We also examined the relationship between consistency of LLT/TDF use and C variability in a mixed linear model. Results During a median follow-up (f/u) of 4.3 years, there were 2,269 deaths. For 43% of f/u time patients were exposed to TDF, for 13% to non-statin LLT, and for 17% to statins. The hazard ratio (HR) of death for consistent use was 0.38, 95% confidence interval (CI) 0.32–0.45 for TDF, 0.40 (CI:0.28–0.58) for non-statin LLT, and 0.50 (CI:0.40–0.63) for statins (Fig 1). In the composite model, increasing LLT intensity was inversely associated with mortality risk but low C and high CV remained strong independent predictors of death. High HDL C values were protective (Table). The effect sizes for high LLT intensity and high HDL-CV moderately decreased when they were modeled jointly. Compared to unexposed patients, HDL CoV was significantly lower during consistent LLT exposures but higher for patients with low LLT exposure (Fig 2). Conclusion In PLWHA, intensity of LLT exposure is inversely related with all-cause mortality. The highest LLT intensity levels were associated with the lowest HDL-CV. This could serve as a partial explanation for the lower mortality. Disclosures Roger Bedimo, MD, Gilead Sciences: Advisor/Consultant|Janssen: Advisor/Consultant|Merck & Co.: Advisor/Consultant|Merck & Co.: Grant/Research Support|ViiV Healthcare: Advisor/Consultant.