Title

HIV infection independently increases the risk of developing heart failure: The HIV HEART study

Presenter

Alan S. Go

Authors

A.S. Go¹, M. Horberg², K. Reynolds³, W.J. Towner³, H. Avula¹, R.C. Hechter³, D. Klein¹, S. Vupputuri², K.K. Lee¹, T.K. Leong¹, W. Leyden¹, R. Neugebauer¹, S.H. Sung¹, M.J. Silverberg¹

Institutions

¹Kaiser Permanente Northern California, Division of Research, Oakland, United States, ²Kaiser Permanente Mid-Atlantic States, Mid-Atlantic Permanente Research Institute, Rockville, United States, ³Kaiser Permanente Southern California, Department of Research and Evaluation, Pasadena, United States

Background: HIV infection has been associated with excess atherosclerotic events, but limited data exist about its link to developing heart failure (HF) and possible contributing factors. In a large, multi-institutional, community-based population, we evaluated the independent association of HIV infection with incident HF.
Methods: Within 3 large U.S. integrated healthcare delivery systems, we identified all eligible HIV(+) adults (≥21 years) between 2000-2016 without prior HF and frequency-matched up to 10:1 to HIV(-) subjects without prior HF based on entry year, age (±1 year), gender, race, and primary treating facility. Through 2016, we identified cases of incident HF based on validated algorithms using electronic health records (EHR). Demographic features, cardiovascular risk factors, pertinent medical history and medication use were ascertained from EHR and other health system databases. We evaluated the independent association of HIV infection with incident HF through a series of multivariable Cox regression models that sequentially adjusted for: health system and calendar era, demographics, lifestyle factors, cardiovascular history, other comorbidities, and cardiopreventive and other medication use. In the final model, we adjusted for acute coronary syndrome events during follow-up as a potential explanatory variable.
Results:
We identified 38,868 HIV(+) and 386,586 matched HIV(-) adults during the study period. HIV(+) patients were more likely to have low neighborhood-level educational attainment and household income, prior cancer, dementia or depression but less likely to have prior cardiovascular conditions or cardiovascular risk factors. The rate (per 100 person-years) of incident HF was higher in HIV(+) (0.24, 95%CI:0.22-0.26) vs. matched HIV(-) (0.16, 95%CI:0.15-0.16) patients (P< 0.0001) (Figure). In multivariable analyses, HIV infection was associated with an increased rate of developing HF that strengthened after serial adjustment for demographic characteristics; cardiovascular and medical history; and cardiopreventive medication, antidiabetic therapy and NSAID use, with an 75% increased rate in the fully adjusted model (Table, Models 1-3). Further adjustment for acute coronary syndrome events during follow-up only modestly attenuated the association of HIV infection with incident HF (Model 4).

[Cumulative incidence of newly-diagnosed heart failure by HIV status]

	Model 1	Model 2	Model 3	Model 4
	Health System, Entry Year and Demographics	Model 1 + Cardiovascular and Non-Cardiovascular History	Model 2 + Cardiopreventive Medications, Antidiabetic therapy and NSAIDs	Model 3 + Acute Coronary Syndrome Events During Follow-Up
Adjusted Hazard Ratio (95% CI) for HIV(+) vs. HIV(-)	1.54 (1.40-1.69)	1.69 (1.54-1.86)	1.75 (1.59-1.93)	1.66 (1.50-1.83)

[Multivariable association of HIV infection with incident heart failure]

Conclusions: HIV infection independently increases the risk of developing HF and this excess risk does not appear to be mediated through atherosclerotic disease pathways or differential use of cardiopreventive therapies.

Go to Session