Heart Failure Risk and Events in People With HIV: The Randomized Trial to Prevent Vascular Events in HIV (REPRIEVE)

Abstract

BACKGROUND:

People with HIV (PWH) may have a higher risk of heart failure (HF) due to traditional and HIV-related factors. Incidence and risk prediction of HF in PWH are not well characterized. We aimed to quantify the risk of HF events in a global population of PWH with low-to-moderate estimated atherosclerotic cardiovascular disease risk.

METHODS:

HF incidence (events/1000 person years) was described overall and by demographic, HIV-specific, and HF factors, including estimated Predicting Risk of Cardiovascular Disease Events 10-year risk of HF. Confirmed HF events included adjudicated HF hospitalization and adverse events identified via a standardized Medical Dictionary for Regulatory Archives HF query.

RESULTS:

We analyzed 7769 REPRIEVE (Randomized Trial to Prevent Vascular Events in HIV) participants from 5 global regions (median, 50 years; 31% female). Over a median follow-up of 5.6 years (interquartile range, 4.3–5.9), HF incidence was higher in women, among Black participants in high-income regions, participants in sub-Saharan Africa, and among those with preexisting hypertension and obesity compared with the absence of these factors. Current and nadir CD4+T-cell count, and HIV-1 RNA level were not related to the incidence of HF events. Median (Q1–Q3) Predicting Risk of Cardiovascular Disease Events HF score was 1.66% (1.01–2.62). HF incidence was 1.65/1000 person-years (95% CI, 1.30–2.09). Expected number of HF events by Predicting Risk of Cardiovascular Disease Events HF (n=73) was consistent with observed (n=67).

CONCLUSIONS:

Select demographics, clinical factors, and global regions contribute to a higher incidence of HF events among PWH. In PWH, the observed overall number of HF events aligned with the estimated Predicting Risk of Cardiovascular Disease Events HF risk rates.

WHAT IS NEW?

• Heart failure was uncommon in the REPRIEVE (Randomized Trial to Prevent Vascular Events in HIV) trial at 5 years of follow-up.

• Hypertension and obesity were the most common modifiable risk factors for heart failure.

• Most of the hospitalizations for heart failure were recorded in high-income countries.

• The Predicting Risk of Cardiovascular Disease Events HF risk score accurately predicted the risk of heart failure in this low-to-moderate risk group of people with HIV.

WHAT ARE THE CLINICAL IMPLICATIONS?

• The Predicting Risk of Cardiovascular Disease Events HF risk calculator can likely be used to predict HF events in a low-to-moderate cardiovascular risk group of people with HIV.

• Strategies to control hypertension and obesity in people with HIV will likely reduce the risk of heart failure.

See Editorial by Shakil and Hsue

In the era of widespread access to antiretroviral therapy (ART), the risk of heart failure (HF) among people living with HIV (PWH) is 2-fold higher compared with the general population, with higher risk among younger adults, women, and those with higher viral load and lower CD4+T-cell counts.^1–3 Hospitalizations for HF are more common among PWH, and the mortality rate among PWH with HF is alarmingly high at 30% at 1 year after diagnosis, which is, 1.5× higher than those without HIV.^4–6 Although advances have been made to optimize and calibrate atherosclerotic cardiovascular disease (ASCVD) risk calculation for PWH,^7,8 relatively scant attention has been devoted to understanding the risk of HF in this population. Identifying those at greatest risk for HF and adverse HF events is critical to stratifying risk and tailoring timely interventions.

Prediction of HF events is challenging in PWH due to the lack of studies testing the performance of HF risk algorithms in this group. The American Heart Association’s Predicting Risk of Cardiovascular Disease Events (PREVENT) series of risk score equations, developed with US populations, updates prior ASCVD risk equations to include 10- (and 30-) year risk estimates for several specific CVD outcomes, including HF.⁹ The performance of the PREVENT HF risk algorithm in PWH has not been reported in prospective studies. Moreover, numerous studies of HF in PWH have highlighted the inattention to global disparities in HF with scant data from low- and middle-income countries (LMICs), yielding an incomplete understanding of HF risk in PWH in countries with high rates of HIV.^10,11

The REPRIEVE (Randomized Trial to Prevent Vascular Events in HIV) trial is a phase 3, global, multicenter, randomized clinical trial assessing a statin strategy for primary CVD prevention among ART-treated PWH with low-to-moderate ASCVD risk, inclusive of participants from 5 Global Burden of Disease (GBD) regions of the world.¹² REPRIEVE demonstrated efficacy to reduce cardiovascular events (a composite of cardiovascular death, myocardial infarction, hospitalization for unstable angina, stroke, transient ischemic attack, peripheral arterial ischemia, revascularization of a coronary, carotid, or peripheral artery, or death from an undetermined cause) by 36% over a median of 5.6 years.^13,14 HF hospitalization was a prespecified exploratory outcome. REPRIEVE offers an unprecedented opportunity to fill the critical gaps in understanding HF risk and incidence in a prospective study of PWH on stable ART at low-to-moderate ASCVD risk with adjudicated HF events.¹² The objectives of this analysis are to (1) describe the incidence of HF events in PWH and by risk factors for HF assessed at baseline, (2) to characterize predicted HF risk in this population via the PREVENT HF risk score, and (3) evaluate how predicted risk reflects observed events in REPRIEVE by the individual components of the PREVENT HF risk score.

Methods

Study Design and Participants

The data that support the findings of this study are available from the corresponding author on reasonable request. The REPRIEVE trial (URL: https://www.clinicaltrials.gov; Unique identifier: NCT02344290) was a randomized, double-blind, placebo-controlled study to determine whether daily 4 mg pitavastatin calcium could prevent major adverse cardiovascular events. REPRIEVE enrolled 7769 HIV-positive participants aged 40 to 75 years on stable ART with CD4+T-cell count >100 cells/mm³ and low-to-moderate ASCVD risk by the Pooled Cohort Equations. All participants were included in this analysis. Trial exclusions were any known clinical ASCVD, active cancer, kidney impairment, decompensated cirrhosis, certain viral infections without significant fibrosis, recent systemic infections, or diabetes with high low-density lipoprotein cholesterol; HF history was not a specific exclusion criterion (Table S1). The trial was stopped for demonstrated efficacy on March 30, 2023, and additional data were accrued through completion of final study visits between April and August 2023. REPRIEVE participants were enrolled from sites in countries across GBD regions, as previously described.^13,14 Each clinical research site obtained institutional review board/ethics committee approval and any other applicable regulatory entity approvals. Participants were provided with study information, including a discussion of risks and benefits, and signed the approved declaration of informed consent.

Outcomes

For this analysis, HF events included both hospitalized and nonhospitalized first HF events. HF hospitalization was a prespecified outcome in REPRIEVE, adjudicated per trial protocol by members of the Thrombolysis in Myocardial Infarction team blinded to group assignment. The standard definition for HF hospitalization included a hospitalization that met all these criteria: a primary diagnosis of HF, ≥24-hour length of stay, symptoms of HF, a combination of physical and laboratory criteria, and treatment specifically for HF. HF events without hospitalization were identified for review via a narrow prospectively determined standardized Medical Dictionary for Regulatory Archives query. Additional signs and symptoms of HF were identified via the wide cardiac failure standardized Medical Dictionary for Regulatory Archives query and confirmed by investigator review (G.S.B. and P.S.D.). In this analysis, HF is denoted as "HF hospitalization" and "confirmed HF" to denote those HF events adjudicated by trial protocol to include hospitalization, and HF events confirmed after review of adverse event triggers identified by a narrow standardized Medical Dictionary for Regulatory Archives query, respectively.

HF Risk Estimation

We used the American Heart Association PREVENT risk score to estimate HF risk in REPRIEVE. The American Heart Association PREVENT score estimates predicted 10-year HF risk in those without known HF at baseline using the following components: sex (female, male), age (years), systolic blood pressure, body mass index (BMI), estimated glomerular filtration rate, diabetes (yes, no), current smoking (yes, no), and antihypertensive medication use (yes, no).⁹ An online calculator for the PREVENT risk score is available at https://professional.heart.org/en/guidelines-and-statements/prevent-calculator, and model derivation has been described.⁹ Measurements below or above the allowable range (defined by the American Heart Association) for systolic blood pressure (90–200 mm Hg), BMI (18.5–39.9 kg/m²), and estimated glomerular filtration rate (15–140 mL/min per 1.73 m²) were assigned the lowest or highest limit of the range, respectively. No adjustments were made for measurements within the allowable range, and scores were not computed for participants missing components.

Statistical Analysis

Baseline participant characteristics were summarized in all participants and by confirmed HF (including HF with hospitalization) and no confirmed HF (including participants without HF events and those with signs or symptoms of HF). Incidence rates of HF hospitalization and confirmed HF events overall and by risk factor were estimated as the number of events divided by total person-years (PY) of time on study to the first HF event and presented per 1000 PY. For select risk factors, cumulative incidence over time was estimated using the Aalen estimator with death of all causes as a competing risk. The following risk factors of interest were a priori identified: age, natal sex, race, and ethnicity (within high income country [HIC] region), GBD region, hypertension status (no hypertension, undiagnosed HTN [BP ≥140/90 mm Hg], diagnosed HTN on medication, and diagnosed HTN not on medication), PREVENT risk score and its individual components, nadir and entry CD4+T-cell counts (cells/mm³), and entry HIV-1 RNA (copies/mL). The PREVENT HF risk was calculated as a 10-year risk per 100 participants. The expected numbers of HF events (based on our confirmed HF definition without hospitalization) were calculated over all participants and by risk factors of HF as the average PREVENT HF risk score multiplied by the total PY of study follow-up divided by 10; 95% CIs were obtained via 5000 bootstrapped samples. The ratio of the number of observed to expected (O/E ratio) events was used to assess the predictability of the PREVENT HF risk score in the analysis population. The O/E ratio was computed overall and within each characteristic, standardized for REPRIEVE follow-up, and with 95% CIs obtained via the same bootstrap sampling method as above. Participants missing 1 or more PREVENT HF risk score components (n=16) were excluded from analyses, including the risk score.

Results

Baseline Characteristics

The median age of all participants was 50 years (Q1–Q3, 45–55). There were 67 confirmed HF events and 28 HF hospitalizations over a median follow-up of 5.6 years. Participants with confirmed HF events were slightly older at 53.0 years (Q1–Q3, 47–57) than those with no HF (Table). There was a higher percentage of female participants with HF events compared with those with no HF (40% versus 31%). Among those with HF events, ≈50% were Black compared with 21% without HF. More than 50% of individuals with HF events had hypertension, with most taking antihypertensive medication. Around 60% with HF were current/former smokers compared with 50% without HF. BMI (29.4 versus 25.7 kg/m²) and ASCVD risk scores (6.6% versus 4.5%) were slightly higher, and estimated glomerular filtration rate (82.3 versus 96.2 mL/min per 1.73 m²) was slightly lower among participants with a HF event compared with no HF event. All other characteristics had similar distributions between participants with HF events and those with no HF (Table).

Table.

Baseline Characteristics by HF Outcomes

Most HF hospitalizations occurred in the HIC GBD region (n=24, 86%), while none of the HF events in sub-Saharan Africa (SSA) were adjudicated as a hospitalization. When evaluating confirmed HF events in the HIC region alone, there was an even higher percentage of participants identifying as Black with HF compared with those without HF (65% versus 41%; Table S2).

Confirmed HF and HF Hospitalization Event Incidence

Overall confirmed HF event incidence was 1.65 per 1000 PY (95% CI, 1.30–2.09) with similar rates by randomized (pitavastatin versus placebo) arm (Figure 1). Point estimates indicate that HF event incidence rates were higher among older participants, females, and those in HIC regions. Precision around these estimates was low, given relatively fewer events in some of these subgroups. Confirmed HF event incidence rates were highest in the subgroups of participants with hypertension, with the highest rates being seen among those on medication (3.20 per 1000 PY [95% CI, 2.18–4.69]; n=26 events). Black or African American race (3.68 per 1000 PY [95% CI, 2.60–5.20]; n=32 events), current smoker (2.35 per 1000 PY [95% CI, 1.55–3.56]; n=22 events), and BMI ≥30 kg/m² (3.31 [95% CI, 2.30–4.76]; n=29 events) subgroups showed some of the highest incidence rates. HF event incidence rates were high, but absolute HF event counts were low, in those with HIV-1 RNA level ≥400 copies/mL (n=3 events) and diabetes (n=1 event), precluding analysis.

Figure 1.

Incidence of first confirmed heart failure event by heart failure risk factors. Incidence rates were estimated using the Poisson distribution. x-axis truncated at 5, arrows indicate continuation of CI.

Incident HF hospitalization rate was 0.69 per 1000 PY (95% CI, 0.47–0.99). Similar to confirmed HF events, HF hospitalization rates were highest among Black race, controlled hypertension, and BMI ≥30 kg/m² categories (Figure S1).

Cumulative incidence of HF events according to GBD region, hypertension status, and BMI is shown in Figure 2. HF hospitalizations accumulated almost exclusively in the HIC GBD region in the first few years of follow-up; however, confirmed HF events appeared early in both HIC and SSA GBD regions. The probability of a confirmed HF event and of HF hospitalization was nearly 4-fold higher at 12 months with hypertension compared with not having hypertension. BMI ≥30 kg/m² was associated with early and consistently elevated risk of HF compared with lower BMI categories. Black and current/former smokers also showed early and consistent risk of HF hospitalizations and confirmed HF events during follow-up (Figures S2 and S3). Of note, we found no significant relationships between HF events and age, sex, or CD4 T+ cell count (nadir and at levels at entry).

Figure 2.

Cumulative incidence of heart failure events by Global Burden of Disease (GBD) region and hypertensive control. Cumulative incidence was calculated using the Aalen estimator for the probability of the subdistribution of failure of interest. Cummulative incidence is shown according to GBD region for first HF with hospitalization (A), GBD region for first Confirmed HF (B), hypertensive control for first HF with hospitalization (C), hypertensive control for first confirmed HF (D), body mass index for first HF with hospitalization (E), and body mass index for first confirmed HF (F). (Continued )

PREVENT-HF Risk Score

Sixteen participants were excluded from the following analyses due to missing at least 1 component of the PREVENT HF risk score at study entry. Among the participants included, 496 had BMI out of the allowable range, 12 had systolic blood pressure out of range, and 2 had estimated glomerular filtration rate out of range. The PREVENT-HF risk score was nearly 2-fold higher among those with confirmed HF events, including events with hospitalization (2.22%), compared with those with no HF (1.38%).

Observed and Expected HF Events

We calculated the number of expected HF events using the PREVENT-HF risk score and compared this to the observed events according to predefined subgroups (Figure 3). The expected number of HF events by PREVENT-HF (n=73) was similar to that of the observed (n=67). Although the observed confirmed HF events were generally consistent with the expected events across participant characteristics, there were some groups in which the PREVENT risk score predicted events less well. PREVENT under-predicted numbers of events in Black or African American participants (O/E ratio, 1.68 [95% CI, 1.11–2.29) and participants with BMI ≥30 kg/m² (O/E ratio, 1.55 [95% CI, 1.01–2.15]). The risk score over-predicted actual events in non-HIC GBD regions (O/E ratio, 0.63 [95% CI, 0.35–0.94]). When restricted to the subset with HF hospitalization events, the observed number of events was generally lower than the expected number of events by PREVENT-HF, indicating overprediction (Figure S4).

Figure 3.

Ratio of observed to expected number of confirmed HF events. Expected number of events is calculated as (average Predicting Risk of Cardiovascular Disease Events [PREVENT] HF risk score×total person years)/(100×10). PREVENT HF is a 10-year risk score that has been calculated per 100 participants; thus, the expected number of events is divided by 10×100 to determine the expected number of events for 1 participant over 1 year. The expected number of events is calculated using the arithmetic mean. The 95% CIs for expected events and the ratio of the number of observed to expected (O/E) are calculated using 5000 bootstrap samples. Due to low numbers of observed events in non-High Income Global Burden of Disease (GBD) regions, the expected number of events has only been calculated for all non-High Income regions instead of individual regions.

Discussion

To address the knowledge gaps related to incident HF events and the prediction of HF events in PWH, we analyzed, adjudicated, and prospectively observed HF events in the global REPRIEVE trial. Our main findings are as follows. The observed number of HF events over 5 years was relatively low in this low-to-moderate ASCVD risk population. Demographic risk factors for HF events included older age, female sex, and Black or African American race, while the strongest modifiable factors were hypertension and obesity. Most HF hospitalizations were observed in HIC regions, while other confirmed HF events were more common in SSA. The PREVENT-HF risk score was low at trial entry and accurately predicted HF events in the overall trial population. These data shed novel insights into the contemporary trends in HF among PWH.

PWH continue to face significant health threats from HF despite appropriate ART. Compared with individuals without HIV, PWH have a 70% to 80% higher relative risk of HF, which is not fully explained by differences in traditional cardiovascular risk factors.^1,15 Data from the Veterans Administration showed a higher incidence than the current study (7.21 per 1000 PY), while an analysis that employed physician-adjudicated HF found a prevalence more similar to what we report in REPRIEVE (3.7 per 1000 PY).² The risk of HF in PWH seems consistently greater in women when including studies that represent the general population, such as nongovernment electronic medical records. For example, the HIV HEART study (12% women) found a 2.5-fold greater adjusted hazard ratio for HF in women with HIV compared with men with HIV across 3 Kaiser Permanente national regions, and the US Partners Healthcare Database reported a 4-fold increased incidence of physician-adjudicated HF among women with HIV compared with women without HIV.^5,15 The constellation of demographic and clinical findings associated with HF in the present study tends to be more common in HF with preserved ejection fraction, which has become the most common HF presentation in PWH.¹⁰ Chronologic age, female sex, obesity, and hypertension are well-known mediators of the risk factors for HF with preserved ejection fraction in the general population, as well as in PWH in global settings.^6,16–18 Building on prior examples,¹⁹ prospective studies that incorporate cardiac imaging alongside deep phenotyping for the low-to-moderate predicted risk PWH are critical to allow for earlier detection and treatment.

Hypertension and obesity in PWH are of critical concern for their relationships to HF in this analysis, as well as the concomitant risk for diabetes, ASCVD, and immune dysfunction.²⁰ Hypertension (versus no hypertension) increases the risk for HF in the general population by 70%, and 46% to 74% of individuals in global HF registries have a history of HF.^21,22 Data on the HF risk from hypertension in PWH have thus far been limited to retrospective studies or single institution studies, with some data from the US indicating no excess risk for HF related to hypertension in PWH.²³ We posit that prospective ascertainment and adjudication of HF events in the current study are likely related to the strong signal linking hypertension and HF. The association between greater HF risk and HTN on medication status is likely attributable to higher underlying risk at trial entry. Further, regional variation in etiologies of HF (more nonatherosclerotic mechanisms in LMICs) suggests that the contribution of hypertension to HF in PWH may be related to regional factors.²⁴ Factors associated with increased weight gain among PWH include certain ART regimens (such as starting dolutegravir and tenofovir alafenamide or stopping weight-suppressing agents tenofovir disoproxil and efavirenz), women with HIV, and specific virological factors like lower baseline CD4+T-cell count and higher HIV viral load.²⁵ Globally, second-generation integrase strand transfer inhibitors have been associated with incident weight gain and obesity²⁶ which, if left unchecked, could portend future HF risk. Notably, high BMI is strongly related to increased myocardial steatosis in PWH, which may represent a mechanistic link between obesity and HF in this population.²⁷ With the advent of novel therapies for obesity (eg, glucagon-like peptide-1 receptor agonists) and recent data suggesting their benefits,²⁸ our findings support greater attention to these therapies to study their potential effects on HF risk in PWH, in addition to changing modifiable risk factors.

REPRIEVE enrolled a global cohort of PWH that represented many countries outside of HIC regions (≈47% of trial participants).¹³ We found that HF hospitalization rates were higher among participants in HICs compared with LMICs. Conversely, other HF confirmed events were the most common type of events in LMICs, particularly in the SSA region. The rates of HF stand in contrast to the lower rate of MACE events in the SSA region in REPRIEVE (1.1 per 1000 PY in the pitavastatin group and 1.8 per 1000 PY in the placebo group), as previously reported.¹³ The burden of HIV and disability-adjusted life years of HIV-associated CVD is greatest in LMICs.²⁹ Moreover, the mechanisms leading to HF in PWH in LMICs include both atherosclerotic and nonatherosclerotic cardiovascular mechanisms, as well as those related to the degree of HIV disease control.²⁴ Further indicating the elevated risk for HF in LMICs, data from South African PWH without CVD found that 31% had elevated circulating levels of NT-proBNP (N-terminal pro-B-type natriuretic peptide), a risk factor for incident HF.³⁰ REPRIEVE participants in low socio-demographic index regions have worse blood pressure control and more hypertension than those in high or middle socio-demographic index regions also signaling elevated risk of HF.³¹ Thus, it is likely that lower observed rates of HF hospitalization in LMICs are due to a shorter period of observation due to later enrollment in the trial, insufficient attention or technology focusing on these questions, or low access to medical care, as has been suggested previously, as well as the lower predicted cardiovascular risk at entry.^11,32,33 When we included both hospitalization for HF—which is a stringent diagnostic threshold requiring attuned health systems—and other confirmed HF events, we saw more nonhospitalized HF events in SSA, further supporting the idea that heterogeneity in HF diagnostic capabilities warrants attention in future work.

We found no treatment effect of pitavastatin on HF outcomes. The relationship between statin therapy and HF risk remains unclear. The Japanese Heart Failure Syndrome with Preserved Ejection Fraction registry reported a beneficial effect of statin therapy on 3-year mortality in individuals with HF with preserved ejection fraction without coronary artery disease, which was similar to the beneficial effect reported in a post hoc analysis from the Treatment of Preserved Cardiac Function Heart Failure with an Aldosterone Antagonist trial.^34,35 Some studies from cohorts in Asia also suggest improvement in intermediate HF outcomes and hospitalization with statins, including pitavastatin.³⁶ We are aware of no other trial that reports on the relationship between statin randomization in PWH and HF outcomes. Noting the low incidence of HF as a secondary outcome, our finding of a lack of an association over 5 years of follow-up is a novel contribution to the field that generates hypotheses regarding HF pathogenesis via non-statin-responsive pathways.

Risk prediction of HF has advanced recently with the publication of the PREVENT-HF risk calculator in 2023.^9,37 An advantage of the PREVENT score is the inclusion of numerous CVD subtypes, like HF, that were not captured using the Pooled Cohort Equations. We found a good overall level of agreement between the expected number of HF events and those observed in REPRIEVE over a median of 5.6 years using PREVENT-HF. Two distinctive aspects of the present study warrant consideration in this regard. First, of the 46 data sets representing >6 million adults used to generate the PREVENT-HF equation, most studies defined HF according to a hospitalization, physician medical examination, or a death event. For this analysis, we also included confirmed HF events that may not have met this level of acuity. Thus, even though our range of HF events includes both hospitalized and nonhospitalized events, the performance of PREVENT-HF in this cohort was accurate. Second, PREVENT-HF was not derived or validated outside the US or in PWH.³⁸ We are not aware of any longitudinal studies with adjudicated HF events in PWH that have been used to predict PWH-specific HF risk, or address the performance of the PREVENT-HF score in PWH at low-to-moderate predicted risk. Similarly, confirmatory studies in the general population are also currently lacking. We also observed under-estimation of HF risk in Black participants, which has also been observed for other cardiovascular disease risk equations in PWH.³⁹ The PREVENT scores omit race-based and ethnicity-based data elements in calculating risk for ASCVD, HF, or both, yet contextual determinants, or unmeasured variables (eg, diastolic dysfunction), may be contributing to the unmeasured excess risk in this group. Although the PREVENT-HF equation is a new, accurate risk estimation tool and represents a step forward in improving risk stratification and primary prevention of HF, several limitations and unknowns have been articulated, including underestimation in some key subgroups.⁴⁰ In the coming years, as additional data on risk prediction for HF are amassed globally, more precise insights into HF prevention and treatment will emerge that integrate regionally relevant risk factors as well as more conventional ones among PWH.⁴¹

In this analysis, we included a global cohort of PWH, inclusive of individuals from regions with the greatest burden of HIV. We incorporated a systematic approach to adjudication of HF hospitalization and further identified HF events by review of adverse event records. There are also limitations to consider. The number of HF events was lower than reported in some studies and was not evenly distributed by GBD region. As a result, an analysis of region-specific factors related to HF was not possible. Although HF hospitalization was a prespecified outcome in REPRIEVE and adjudicated per trial protocol, we did not routinely collect information on ejection fraction; therefore, this limited our ability to address HF subtypes (eg, according to ejection fraction). We may have also missed individuals with incident left ventricular dysfunction without symptoms, leading to underestimation of the early stages of HF. The REPRIEVE population included individuals without known CVD and with low-to-moderate predicted ASCVD risk; however, a history of HF was not an exclusion criterion. Any clinically actionable findings at the time of enrollment were managed at the discretion of the site principal investigator with assistance from local cardiologists or REPRIEVE trial medical officers as needed. For these reasons, we believe there is a low likelihood that participants with clinical HF were included in enrollment. Though our findings may lack generalizability to other groups, they provide critical data for the large number of PWH in the low-moderate risk groups, without known heart disease outside of a clinical trial setting. Future work in this area will benefit from focusing on populations across the spectrum of ASCVD risk to understand risk in the general population, as well as more detailed discrimination of HF risk prediction.

In conclusion, among a large, global cohort of PWH with low to moderate ASCVD risk followed prospectively, the incidence of HF hospitalization or confirmed HF events was modest and was accurately predicted by the PREVENT-HF score. The risks for HF were greatest for women, Black participants and those with hypertension or obesity. HF hospitalizations were more commonly seen in HIC regions. This is the first analysis of any global randomized clinical trial of PWH with adjudicated HF outcomes to suggest that the PREVENT-HF score is reliable in PWH, and to highlight potential global disparities in HF incidence. With increasing attention to CVD risk assessment in PWH, more detailed investigations and confirmation are needed in cohorts across the CVD risk spectrum, as it is vital that we integrate HF risk prediction in the care of this population.

ARTICLE INFORMATION

Acknowledgments

The study investigators thank the study participants, site staff, and study-associated personnel for their ongoing participation in the trial. In addition, they thank the following: the ACTG for clinical site support; ACTG Clinical Trials Specialists (Laura Moran, MPH and Jhoanna Roa, MD) for protocol development and implementation support; the data management center, Frontier Science Foundation, for data support; the Center for Biostatistics in AIDS Research for statistical support; and the Community Advisory Board for input for the community.

Sources of Funding

This study is supported through National Institutes of Health grants U01HL123336 and 1UG3HL164285, to the Clinical Coordinating Center, and U01HL123339 and 1U24HL164284, to the Data Coordinating Center, as well as funding from Kowa Pharmaceuticals America Inc, Gilead Sciences, and ViiV Healthcare. The National Institute of Allergy and Infectious Diseases supported this study through grants UM1 AI068636, which supports the ACTG Leadership and Operations Center; UM1 AI106701, which supports the ACTG Laboratory Center. This work was also supported by the Nutrition Obesity Research Center at Harvard (P30DK040561 to Dr Grinspoon).

Disclosures

The views expressed in this article are those of the authors and do not necessarily represent the views of the National Heart, Lung, and Blood Institute (NHLBI) or the National Institute of Allergy and Infectious Diseases (NIAID); the National Institutes of Health (NIH); or the US Department of Health and Human Services. Dr Aberg reports grants from Massachusetts General Hospital during the conduct of the study; institutional research support for clinical trials from Gilead Sciences, Glaxo Smith Kline (GSK), Janssen, Macrogenics, Merck, Pfizer, Regeneron, and ViiV Healthcare, and personal fees for advisory boards from GSK/ViiV, Invivyd, Merck, and Regeneron; and participation on Data and Safety Monitoring Board (DSMB) for Kintor Pharmaceuticals, all outside the submitted work. Dr Campbell reports research grant support through his institution from Gilead Sciences, ViiV Healthcare, all outside the submitted work. Dr Cespedes reports scientific advisory board consulting fees from Gilead Sciences and ViiV Healthcare, all outside the submitted work. Dr Currier reports consulting fees from Merck and Company and ResViroLogix, outside the submitted work. Dr Sponseller reports being employed by Kowa Pharmaceuticals America Inc. Dr Fichtenbaum reports research grant support through his institution from Gilead Sciences, ViiV Healthcare, GSK, and Merck, all outside the submitted work. Dr Lu reports grant support through his institution from the NIH/NHLBI and Kowa Pharmaceuticals America for the conduct of the study. He also reports research support to his institution from the American Heart Association, AstraZeneca, Ionis, Johnson & Johnson Innovation, MedImmune, the National Academy of Medicine, the NIH/NHLBI, and the Risk Management Foundation of the Harvard Medical Institutions Incorporated outside of the submitted work. Dr Malvestutto reports institutional research support by Lilly and honoraria from ViiV Healthcare, Gilead Sciences, and Pfizer for advisory board membership, all outside the submitted work. Dr Pierone reports research grant support from ViiV Healthcare, GSK and AbbVie, all outside the submitted work. Dr Zanni reports grant support through her institution from NIH/NIAID and Gilead Sciences Inc, relevant to the conduct of the study, as well as grants from NIH/NIAID and NIH/NHLBI; support for attending the Conference on Retroviruses and Opportunistic Infectionsand International Workshop for HIV and Women from conference organizing committee when abstract reviewer and speaker; and participation in DSMB for NIH funded studies, outside the submitted work. Dr Grinspoon reports grant support through his institution from NIH, Kowa Pharmaceuticals America Inc, Gilead Sciences Inc, and ViiV Healthcare for the conduct of the study; personal fees from Theratechnologies and ViiV; and service on the Scientific Advisory Board of Marathon Asset Management, all outside the submitted work. Dr Ribaudo reports grants from Kowa Pharmaceuticals during the conduct of the study, as well as grants from NIH/NIAID, NIH/NHLBI, NIH/National Institute of Diabetes and Digestive and Kidney Diseases, and NIH/NIA, outside of the submitted work. The other authors report no conflicts.

Supplemental Material

Tables S1 and S2

Figures S1–S4

Supplementary Material

hhf-19-e013382-s001.pdf ^{(2.4MB, pdf)}