Variation in Heart Failure Risk by HIV Severity and Sex in People with HIV Infection

INTRODUCTION

As people with HIV (PWH) live longer lifespans on antiretroviral therapy (ART)¹, heart failure (HF) has become a major chronic health issue^2–4. PWH have an estimated 1.5- to 3-fold higher risk of developing HF than HIV-uninfected individuals, and the prevalence of comorbid HF is rising^4–6. This increased risk does not appear to be fully explained by the higher prevalence of traditional HF risk factors (e.g. smoking) among PWH, disparities in prevention or treatment of cardiovascular disease, interim acute coronary syndrome events, or adverse effects of current generation ART for HIV treatment^{4, 6}. Rather, accumulating evidence indicate that the severity of HIV infection and degree of HIV control may be key influential factors in HF risk.

Several epidemiologic studies have reported that HF risk is greatest among PWH who have HIV-related immunosuppression or high HIV viremia^7–11. In addition, the systemic immune activation/inflammation observed in HIV infection are hypothesized to affect men and women differently, potentially resulting in more pronounced HF risk among women with HIV^12–15, as has been observed with other cardiovascular events^16–19. However, sex differences in HF risk among PWH have received little attention in the literature since prior studies have been conducted either exclusively or predominantly among men¹⁹. Studies of HF risk among women with HIV have confirmed increased HF risk compared with HIV-uninfected women¹⁵, but most prior studies have lacked male comparators and/or analyses of PWH stratified by HIV severity^20–23. To our knowledge, no prior large-scale epidemiologic studies have evaluated how HIV severity may differentially affect risk of HF in men and women.

Prior studies of HIV severity and HF have also been limited by cross-sectional study design, lack of longitudinal measurements of HIV biomarkers, lack of a comparable HIV-uninfected reference group, and/or insufficient data on confounding factors for HF such as substance use. While the pathophysiological link between HIV-mediated immune dysfunction and HF remains unclear, characterizing subgroups of PWH at greatest risk for HF could inform strategies for targeted HF risk assessment or early intervention. The objectives of this study were to describe differences in incident HF risk by HIV severity and assess sex-specific risk of HF by HIV severity among members of a U.S. healthcare system. The study described here builds on prior work in this study population which found 1.73-fold elevated risk of incident HF among PWH compared to their uninfected counterparts but did not examine HF risk by HIV severity and sex⁶.

METHODS

RESULTS

DISCUSSION

In this study of PWH and matched PWoH, PWH with lower recent CD4 count or higher recent HIV viremia were at greatest risk of developing HF even after accounting for sociodemographics and clinical cardiovascular risk factors. Recent CD4 count was more strongly associated with HF than nadir CD4 count. Notably, preserved immunity (recent CD4 count ≥500 cells/μl and nadir CD4 count ≥200 cells/μl) was not associated with increased HF risk. The association of elevated HF risk with low CD4 or high HIV RNA was stronger among women than men. Our data also suggest that women with HIV and normal CD4 counts may remain at elevated risk for HF. Given the independent association of HIV severity with HF risk, effective treatment and management of HIV – including timely HIV diagnosis, prompt ART initiation, and promotion of ART adherence – may be important for primary prevention of HF. Increased attention to sex-specific risk factors for HF among PWH, and an improved understanding of underlying pathological mechanisms, will also be critical to HF prevention efforts given that women with HIV were at higher risk of HF than men with HIV.

Consistent with prior studies, we found that HIV-related immunodeficiency and detectable HIV viremia were associated with higher HF risk^{7–11, 23}. Moreover, we demonstrated elevated HF risk in a longitudinal cohort across multiple HIV measures, including recent and nadir CD4 count as well as HIV RNA level. Given that our study was conducted among PWH and demographically-matched PWoH, with analytic adjustment for behavioral factors and comorbidities, our findings support the hypothesis that HF risk among PWH is more reflective of HIV-related immunodeficiency and immune activation than the effects of substance use, aging, or accumulating age-associated conditions. Non-modifiable HIV-related factors may also contribute, including chronic HIV-mediated inflammation, which has been hypothesized to increase the likelihood of myocardial dysfunction⁴. Elevated HF risk among PWH in some other settings has been attributed, in part, to suboptimal treatment of cardiovascular disease^{2, 8, 19}. While we did not specifically examine cardiovascular disease management in this analysis, we previously showed that the excess HF risk in PWH was not explained by differential use of cardiopreventive therapies in our study setting¹⁵, and PWH had similar or superior cardiovascular disease risk profiles, including comparable uptake of lipid-lowering and hypertension therapy, compared with matched PWoH²⁵.

Our study population was primarily men, but its large size enabled analyses of sex differences in the association of HIV severity with HF, building on prior studies which have been underpowered for male-to-female comparisons^7–11. Although we interpret our sex-stratified results with caution given the available sample size, our analyses were adequately powered and mirrored the pattern of increasing HF risk with increasing HF severity from overall analyses but with distinctly stronger associations among women. Poor HIV control was more strongly associated with increased risk of HF among women than men, concordant with the hypothesis that the impacts of HIV-related immunodeficiency are differential by sex and can lead to enhanced HF risk among women with HIV^12–14. Observed differences in HF risk by sex may also be partially explained by sex-specific variation in cardiovascular aging, cardiometabolic risk factors, and pathophysiology of cardiovascular disease^33–37. Some conditions unique to women – for example, pre-eclampsia, gestational diabetes, polycystic ovary syndrome, and early menopause – can cause distinct patterns of metabolic dysregulation which increase cardiovascular disease risk^{34, 37–39}.

Our results have clinical implications in the current era of immediate ART initiation and improved ART regimens^{40, 41}. Nadir CD4 count was not as strongly associated with HF as recent CD4 count, highlighting ART initiation and adherence as central modifiable factors in HF risk, regardless of prior immunodeficiency or suboptimal HIV treatment. As ART use becomes more widespread, the epidemiology of HF is anticipated to change; specifically, HF with preserved ejection fraction will likely become the predominant HF type in virally-suppressed HIV populations^{3, 4, 15}. Since this type of HF has fewer evidence-based pharmacologic treatment options and lower treatment success^{36, 42}, effective interventions for primary prevention of HF will become increasingly important. Notably, women are approximately twice as likely as men to develop HF with preserved ejection fraction^{9, 22, 42}. With regard to sex-specific prevention of HF among PWH, women with treated HIV are also more likely than HIV-uninfected women to have pre-HF phenotypes characterized by diffuse myocardial fibrosis and reduced diastolic function, along with higher levels of circulating inflammatory markers that may promote myocardial damage⁴³. These potentially reversible HF precursors represent opportunities for early intervention among women with HIV, who experience higher rates of HF hospitalization, longer HF hospital stays, and higher mortality than HIV-uninfected women^{15, 22, 43}.

Overall, by identifying PWH most likely to develop HF, we contribute to current discussions on approaches to modify HF risk and optimize cardiac care in aging HIV populations^{2, 4}. There is accumulating evidence that HIV severity influences HF risk, but still-limited knowledge on its physiological underpinnings and longer-term effects. Potential HF prevention methods include early identification of subclinical myocardial damage using biomarkers associated with HF among PWH, pharmacotherapeutic options to minimize HIV-related inflammation, and EHR-based HF risk prediction to help clinicians better identify HF risk in selected populations^{2, 4}. The link between HIV severity and HF also justifies increased investment in strengthening the HIV care continuum, which focuses on increasing HIV diagnosis and engagement/retention in care, ART initiation, and viral suppression^{44, 45}. In addition to improving HIV-related outcomes, leveraging HIV care infrastructure to support expansion of cardiovascular disease prevention efforts for PWH could have positive impacts on cardiac health in this population^{2, 46}.

This study had several limitations. First, data were collected in routine clinical care and may have been affected by residual confounding due to factors not fully captured in the EHR such as diet quality and physical activity patterns. Second, while analyses adjusted for diagnoses of prior cardiovascular disease, diabetes mellitus, and dyslipidemia, we did not assess treatment for these conditions. The prevalence of these conditions was similar or lower among PWH compared with PWoH and thus is unlikely to have changed the direction of associations. Third, without data on the type or duration of ART use among PWH, we could not evaluate differences in HF risk among PWH who took ART regimens associated with lipodystrophy or cardiac side effects. However, the main goal of this study was to evaluate HF by HIV status across levels of HIV severity, and ART side effects would likely have been modest compared to the effects of diabetes and other comorbidities which were already included in adjusted models. Fourth, due to the overall low number of PWH who developed HF (n=230), we were unable to evaluate the association of HF with HIV severity by ART treatment era. However, in exploratory analyses, we found only 34 cases of HF among PWH who started study follow-up on January 1, 2010 or later, suggesting that ART improvements and better HIV control are linked to reduced HF risk in more recent years. Lastly, we did not stratify analyses by HF etiology (i.e., ischemic vs. non-ischemic) or degree of systolic dysfunction (i.e., reduced vs. preserved) but recognize that HF is a heterogeneous syndrome and that the epidemiology of HF may be rapidly evolving in more contemporary cohorts. Importantly, future larger studies powered to perform analyses of HIV severity and HF risk stratified by both sex and HF type would be important given known sex differences across HF subtypes.

This study builds methodologically on prior studies^7–11 in several ways, including: 1) use of time-updated and recent (i.e., less than 6 months old) CD4 and HIV RNA measures, which ensured evaluation of HIV severity within clinically-appropriate timeframes, 2) lagged analyses to ensure HIV measures preceded HF events by at least 6 months, which limited false associations due to reverse causality (i.e. HF influencing the HIV measure), 3) adjustment for time-updated comorbidity and substance use-related covariates, which accounted for changing epidemiology of cardiovascular risk factors over time, and 4) evaluation of HF risk across multiple measures of HIV severity, stratified according to clinically-relevant levels of CD4 and HIV RNA. Other strengths of the study included the large population of PWH and well-matched PWoH within the same setting, an HIV population that reflected the racial/ethnic diversity of the HIV population in the underlying community^47–51, use of a previously validated HF definition, and relatively large number of HF cases among women which enabled analyses of HF risk stratified by sex.

In summary, we found that HIV immunodeficiency and viremia were independent risk factors for incident HF. PWH who had lower CD4 count or higher HIV viremia were at greater risk of HF than PWoH, with stronger associations observed among women with HIV than men with HIV. In addition to addressing traditional cardiovascular risk factors, optimizing HIV management could be important for effective HF reduction among PWH.