HIV infection and Heart Failure Outcomes in Women

There is a 2.5-fold increased risk of incident heart failure (HF) among women living with HIV (WLWHIV) (1). Whether HF outcomes differ by HIV status among women has not been established. Leveraging data from a large current and established U.S. healthcare system-based cohort (2), we explored differences in HF outcomes among women with HF with and without HIV. We used ICD-9 codes to identify HF and all diagnoses were also adjudicated. The cohorts were matched 1:3 by age/race/ethnicity. Our primary outcome was incident HF hospitalization. The follow-up period started from the diagnosis of HF. We also recorded the cumulative HF hospitalization rate (total hospitalizations/100 person-years). Our secondary outcomes included length of stay of first HF hospitalization, all-cause mortality, and cardiovascular mortality (any cardiac cause) determined through physician electronic health record review. Univariate and multivariate Cox proportional hazard regression analyses were performed to determine the association between baseline covariates and outcomes. Subgroup analyses on the use of guideline-recommended HF therapy (ACEI/ARB + beta-blocker) were conducted among those with a reduced left ventricular ejection fraction (LVEF) (HFrEF, LVEF cut-off of <50%). There was a broadly similar prevalence of traditional cardiovascular risk factors (smoking, hypertension, dyslipidemia, diabetes mellitus, valvular disease, presence of coronary artery disease, family history of coronary artery disease) between WLWHIV with HF and uninfected women with HF. There was a higher prevalence of sleep apnea and cocaine use and a lower BMI among WLWHIV. The LVEF was similar between the groups (54 ± 15 vs. 52 ± 14%, p = 0.48) and HFpEF was more common than HFrEF (HIV+ 71 vs. 29%, HIV- 63 vs. 37%, p = 0.44); WLWHIV with HF had a wider QRS (113 ± 19 vs. 97 ± 20 msec, p <0.0001) and higher PASP (49 ± 12 vs. 39 ± 10 mmHg, p <0.0001). Among WLWHIV, 92% were on ART with a mean duration of 13 years, 50% had achieved viral suppression (<400 copies/ml) with a mean CD4 count of 367 cells/mm³ and mean nadir CD4 count of 160 cells/mm³. The cohort had an average age of 59 ± 9 years and over a median follow-up of 9 years, 34/1388 WLWHIV developed HF (2.5% cumulative incident rate, 0.27% incident rate per year), 4 times that of uninfected control women (102 of 13,781, 0.74% cumulative incident rate, 0.07% incident rate per year). The median follow-up from the initial HF diagnosis for WLWHIV was 4 years (IQR 1–7) and for controls was 5 years (IQR 2–8, p = 0.362). As compared to controls, WLWHIV with HF had increased rates of cumulative HF hospitalization (42 vs. 9 per 100 person years, p <0.0001), length of stay for HF hospitalization (8 vs. 5 days, p <0.0001), higher all-cause mortality (53 vs. 21%, p <0.001) and CV mortality (83 vs. 33 %, p <0.006). The mortality rate was 18/310 = 0.058 person-years for WLWHIV and 21/788 = 0.027 person-years for the uninfected group. The incident HF hospitalization rate was 20 per 100 person-years for WLWHIV as compared with 8 per 100 person-years for the controls (p <0.0001). In multivariate model, adjusting for known predictors of HF hospitalization including age, hypertension, diabetes mellitus, obstructive sleep apnea, coronary artery disease, angiotensin converting enzyme inhibitor or angiotensin receptor blocker use and LVEF, the adjusted hazard for incident HF hospitalization was 2.58 (95% CI: 1.55–4.29 P <0.0001). We further analyzed the use of optimal HF pharmacological therapy among HFrEF and found WLWHIV with HFrEF (n = 10) had lower use of optimal HF pharmacological therapy (40 vs. 83%, p = 0.01) as compared to control women with HFrEF. To try and understand other potential contributors to differences, we also tried to compare socioeconomical status by analyzing health insurance coverage. The percentage of patients who had more than one health insurance was higher in WLWHIV as compared to the uninfected group, although the difference did not reach statistical significance (p=0.107). Medicare insurance was more prevalent in WLWHIV (55.9% WLWHIV vs. 33.7% controls, p=0.023), while Medicaid use did not differ substantially between the 2 groups significantly (50.0% WLWHIV vs. 33.0% controls, p=0.072). A limitation of this report includes the long timeframe over which data were captured as the treatment of HIV and heart failure has changed substantially.

In summary, leveraging data from a large US health care system, we found that HF outcomes differ markedly by HIV status among women. Specifically, WLWHIV with HF have higher rates of HF hospitalization, longer HF hospitalizations, and higher rates of all-cause and cardiovascular mortality. To our knowledge, this study is the first to report HF outcome differences among women with and without HIV. In this era of effective ART, understanding how HIV status influences the development of HF and HF outcomes in women is of critical public health importance to the 14 million women aging with HIV worldwide.

Figure 1.

A: Incident HF hospitalization rate. Kaplan-Meier survival curve for incident HF hospitalization showing that WLWHIW (red line) have decreased event free survival as compared to uninfected women (blue line). HF: heart failure, WLWHIV: women living with HIV. B: Cardiovascular Mortality rate. Kaplan-Meier survival curve for cardiovascular mortality showing that WLWHIV (red line) have increased cardiovascular mortality as compared to uninfected women (blue line).