Disparities by Sex, Race, and Ethnicity in Use of Left Ventricular Assist Devices and Heart Transplants Among Patients With Heart Failure With Reduced Ejection Fraction

Abstract

Background

The extent to which sex, racial, and ethnic groups receive advanced heart therapies equitably is unclear. We estimated the population rate of left ventricular assist device (LVAD) and heart transplant (HT) use among (non‐Hispanic) White, Hispanic, and (non‐Hispanic) Black men and women who have heart failure with reduced ejection fraction (HFrEF).

Methods and Results

We used a retrospective cohort design combining counts of LVAD and HT procedures from 19 state inpatient discharge databases from 2010 to 2018 with counts of adults with HFrEF. Our primary outcome measures were the number of LVAD and HT procedures per 1000 adults with HFrEF. The main exposures were sex, race, ethnicity, and age. We used Poisson regression models to estimate procedure rates adjusted for differences in age, sex, race, and ethnicity. In 2018, the estimated population of adults aged 35 to 84 years with HFrEF was 69 736, of whom 44% were women. Among men, the LVAD rate was 45.6, and the HT rate was 26.9. Relative to men, LVAD and HT rates were 72% and 62% lower among women (P<0.001). Relative to White men, LVAD and HT rates were 25% and 46% lower (P<0.001) among Black men. Among Hispanic men and women and Black women, LVAD and HT rates were similar (P>0.05) or higher (P<0.01) than among their White counterparts.

Conclusions

Among adults with HFrEF, the use of LVAD and HT is lower among women and Black men. Health systems and policymakers should identify and ameliorate sources of sex and racial inequities.

Nonstandard Abbreviations and Acronyms

HFrEF

heart failure with reduced ejection fraction

HT

heart transplant

INTERMACS

Interagency Registry of Mechanically Assisted Circulatory Support

MESA

Multi‐Ethnic Study of Atherosclerosis

UNOS

United Network for Organ Sharing

Clinical Perspective.

What Is New?

• Among the population with heart failure with reduced ejection fraction, the rates of left ventricular assist device and heart transplant use increased steadily between 2010 and 2018 among Black, Hispanic, and White adults aged 35 to 84 years.

• Among women with heart failure with reduced ejection fraction, the rates of left ventricular assist device and heart transplant were about one‐third that of the rates for their male counterparts.

• Among men with heart failure with reduced ejection fraction, the rates of left ventricular assist device and heart transplant for Black patients were significantly lower than and for Hispanic patients similar to that among White patients.

What Are the Clinical Implications?

• Health systems and providers should examine the processes of care for patients with advanced heart failure, including the protocols of referral for mechanical support, to identify sources of sex and racial inequities.

For patients with heart failure with reduced ejection fraction (HFrEF) refractory to medical management, the use of durable left ventricular assist devices (LVADs) and heart transplants (HTs) has steadily increased over the past decade and improved patient life expectancy and quality of life. ¹ , ² Improved outcomes are due to increased accessibility and safety with LVAD as well as an increase in the donor pool for HT. ¹ , ³ , ⁴ Though access to advanced heart failure therapies has increased overall, the extent to which women and minority racial/ethnic groups have received such therapies equitably remains unclear.

Prior analyses of the distribution of advanced heart failure therapies by sex, race, and ethnicity have been limited by the paucity of data specific to the population with HFrEF. ⁵ , ⁶ , ⁷ , ⁸ Evidence on LVAD use by sex, race, and ethnicity using convenience samples from a few health systems has been mixed. ⁹ , ¹⁰ , ¹¹ Studies that compared population‐level rates of LVAD use by sex, race, and ethnicity have relied on the overall census population as the denominator. ⁵ , ⁶ However, this is a significant limitation due to well‐documented sex, racial, and ethnic differences in the prevalence of HFrEF. ³ , ⁶ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ Prior studies of the rates of procedure use based on the overall census population indicated higher rates among Black adults relative to White adults, but Black adults also have a higher underlying prevalence of HFrEF. ⁵ , ⁶ , ¹⁷ While many studies have identified lower rates of advanced heart failure therapies among women, measurement of the differences has been limited by the lack of data on the prevalence of HFrEF. ⁵ , ¹⁷ , ¹⁸ For demographic comparisons of the receipt of heart transplantations, prior studies have used the population on the organ transplant waitlist as the denominator. ⁷ , ⁸ However, patients eligible for a transplant may be excluded on the basis of socioeconomic criteria (insurance coverage, family and social supports) that are associated with sex, race, and ethnicity. ¹⁹ , ²⁰ , ²¹ Overall, the choice of denominators in the prior studies likely underestimated sex, racial, and ethnic disparities in LVAD and HT use.

Identifying adults with HFrEF as the study population, we obtained LVAD and HT rates. ¹⁷ , ²² We used data on LVAD and HT procedures performed from 2010 to 2018 in 19 states that comprise 85% and 70% of the national population of Hispanic and Black adults, respectively. To understand the relationship between denominator choice and procedure rate estimates by sex, race, and ethnicity, we improved prior estimates by using the population with HFrEF and those with heart failure as alternative denominators. We hypothesized that the procedure rates among the populations with heart failure and HFrEF would be lower for women and Hispanic and Black adults.

Methods

Data Sources

As some of the data used are covered by data use agreements, some of the data that support the findings of this study, along with all the programming code, are available from the corresponding author upon reasonable request (see Data S1 for details). We combined data from multiple sources. First, data on the number of LVAD and HT procedures were obtained using annual state inpatient discharge records from 2010 to 2018 from 19 states (Arkansas, Arizona, California, Colorado, Iowa, Illinois, Kentucky, New Jersey, New York, Oregon, Pennsylvania, Maryland, Florida, Georgia, North Carolina, Texas, Virginia, Wisconsin, South Carolina; Figure S1). ¹⁸ , ²³ , ²⁴ , ²⁵ Each state's database contains all inpatient discharge records for all payers from all short‐term acute care hospitals, except for federally owned hospitals (Data S1). Our selection of the 19 states was influenced by multiple factors. First, data from all the states are unavailable, and for some states, the data on key fields (particularly the identification of race and ethnicity) has relatively high rates of missingness. Second, for budgetary reasons, we selected states with larger Hispanic and Black populations. Our second data source was the annual census of adults residing in each state from the American Community Survey, which was used to estimate the population with heart failure and HFrEF. ²⁶ We used National Health and Nutrition Examination Survey data to obtain estimates of heart failure prevalence by age, sex, race, and ethnicity. (Data S2) ¹² , ²⁷ Finally, we obtained data on the incidence of HFrEF from MESA (Multi‐Ethnic Study of Atherosclerosis). For a diverse population of asymptomatic adults enrolled between 2000 and 2002 and followed through 2018, MESA data contain longitudinal information on the diagnosis of heart failure and ejection fraction over a follow‐up period of up to 18.5 years (median, 16.7 years; Data S2) ²⁸ , ²⁹ The institutional review board at the Wake Forest University School of Medicine approved this study and waived consent as the study used deidentified data.

Study Cohort

The selected states account for 85%, 70%, and 61% of the national population of Hispanic, non‐Hispanic Black, and non‐Hispanic White adults (Table S1). We first obtained estimates of the census population from each study state each year stratified by sex, age, and race and ethnicity (Hispanic, non‐Hispanic Black, non‐Hispanic White, and others [including Asian or Pacific Islander and Native American]). ²⁶ We limited our study cohort to adults aged 35 to 84 years; few patients beyond age 84 years received an LVAD or HT. Using National Health and Nutrition Examination Survey data, we estimated the prevalence of heart failure by age, sex, race, and ethnicity. ¹² To identify a subgroup with HFrEF, we estimated the cumulative incidence of reduced ejection fraction among patients with heart failure, defined as an ejection fraction ≤45%, using MESA data and applied the incidence rates by age, sex, race, and ethnicity to obtain estimated counts of the state‐level population of adults with HFrEF (Data S2 and Table S2). ¹⁷ , ²² We followed the Strengthening the Reporting of Observational Studies in Epidemiology reporting guidelines. ³⁰

Measures

Our primary outcome measures are the rates of LVAD and HT procedures among the population with HFrEF (number of procedures per 1000 adults with HFrEF). ⁵ , ⁶ As secondary measures, we obtained the procedure rates among 2 alternative denominators: (1) adults with heart failure (ie, not restricted to HFrEF) and (2) all adults. We identified LVAD and HT procedures using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD‐9‐CM) procedure codes for data from January 1, 2010 to September 30, 2015, and the Tenth Revision, Clinical Modification and Procedure Coding System (ICD‐10‐CM and ICD‐10‐PCS) for data from October 1, 2015 to December 31, 2018 (Data S3 and Tables S3 and S4). ²³ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ We identified potential LVAD device exchange on the basis of multiple hospitalizations for the same patient; this analysis was limited to California, as data for other states do not permit identification of multiple hospitalizations (Data S3). To obtain the HFrEF population and the procedure rates, by state, we excluded out‐of‐state patients. We used the race and ethnicity field in the data sources and defined 4 groups: Hispanic, non‐Hispanic Black (hereafter Black), non‐Hispanic White (hereafter White), and Others. Missing race and ethnicity were observed only in discharge data and among 1.5% of discharges. As discharge data are compiled at the hospital level, we used the Agency for Healthcare Research and Quality guidelines for identifying hospitals with suspect race and ethnicity information and excluded all discharges for all years from these hospitals (Data S4 and Table S5). ³⁶ We used the sex field in the data sources to identify male and female population cohorts.

Statistical Analysis

Our analytic data consisted of annual state‐level cohorts of the HFrEF population stratified by age, sex, race, and ethnicity. We used Poisson regression models with robust standard errors, treating the cohort‐specific procedure volume as the dependent variable and HFrEF population as the exposure measure, to obtain age‐adjusted rates of procedure rates by sex, race, and ethnicity. We also obtained the incidence rate ratio of differences in the procedure rates among Black and Hispanic patients, treating the White patients as the reference group along with the associated 95% CI. We used pooled data (2010–2018) to obtain procedure rates by year and state. As a secondary outcome, we obtained the incidence rate of potential LVAD device exchange by race and ethnicity in California. To assess the sensitivity of the estimates to potential overdispersion, we obtained corresponding estimates using negative binomial regression. Data processing and all statistical analyses were conducted using Stata version 16.1 (StataCorp, College Station, TX) from June 12, 2022 to August 31, 2023.

Results

Among the overall census population of 114.1 million adults aged 35 to 84 years from the 19 study states in 2018, an estimated 1.9 million had heart failure and 69 736 had HFrEF (Table 1). Women constituted 51.9% of the census population and 44.2% of the population with HFrEF. Black and Hispanic adults comprised 12.1% and 18.3% of the census population, and 29.5% and 9.5% of the HFrEF population, respectively. In 2018, 2008 LVADs and 1330 HTs were performed among the study population. Women received 19.1% of LVAD and 25% of HT procedures. Black and Hispanic patients accounted for 27.5% and 10.6% of LVAD procedures, and 22.3% and 14.7% of HT procedures, respectively.

Table 1.

Characteristics of the Study Population, 2018

Characteristics	Census population	Population with heart failure (NHANES)	Population with HFrEF	LVAD procedures	HTs
Total count, 2018	114 121 247	1 894 878	69 736	2008	1330
Age distribution, y, %
35–44	24.7	14.3	15.0	11.4	11.7
45–54	24.6	14.2	15.0	19.5	22.0
55–64	24.3	13.9	14.8	32.8	40.0
65–74	17.5	38.5	38.1	29.5	26.2
75–84	8.9	19.1	18.8	6.8	0.0
Female, %	51.9	42.5	44.2	19.1	25.0
Race and ethnicity, %
White, non‐Hispanic	60.9	59.8	55.1	54.9	53.2
Black, non‐Hispanic	12.1	20.1	29.5	27.5	22.3
Hispanic	18.3	12.7	9.5	10.6	14.7
Other*	8.7	7.5	7.6	5.2	8.0
Missing/unknown	0.0	0.0	0.0	1.8	1.7

The numbers reported refer to individuals aged 35 to 84 years residing in the 19 study states: Arkansas, Arizona, California, Colorado, Florida, Georgia, Iowa, Illinois, Kentucky, Maryland, New Jersey, New York, North Carolina, Oregon, Pennsylvania, South Carolina, Texas, Virginia, and Wisconsin. The census population counts are from the American Community Survey population estimates for 2018. Population with heart failure was obtained by applying the prevalence of heart failure estimates by age, sex, race and ethnicity obtained from NHANES data to the census population counts from the American Community Surveys. See Data S1 for details. Population with heart failure and low ejection fraction was obtained by applying the prevalence of reduced ejection fraction among heart failure patients in the MESA cohort study. See Data S1 for details. Data for the counts of LVADs and HTs are from the state inpatient discharge data from 2018. Characteristics of the recipients of all LVADs and HTs performed in 2018 in the study states were obtained for the counts of LVADs and from the state inpatient discharge data from 2018. For New York we used 2017 data since 2018 data were unavailable. HFrEF indicates heart failure with reduced ejection fraction; HT, heart transplant; LVAD, left ventricular assist device; MESA, Multi‐Ethnic Study of Atherosclerosis; and NHANES, National Health and Nutrition Examination Survey.

^*Other race groups include Asian or Pacific Islander and Native American.

The HFrEF prevalence rate among adults aged 35 to 64 years was 6% lower among women (8.6 per 1000) than among men (9.2 per 1000), and among adults aged 65 to 84 years, the rate was 51% lower among women (23.6 per 1000) than among men (48.1 per 1000) (Table S2). The prevalence of HFrEF was higher among Black men and women in both age groups, relative to their White counterparts. In particular, the rate among Black women in both age groups was >100% higher than among White women. The pattern in White–Hispanic differences in HFrEF prevalence was mixed, with a 9% higher rate among Hispanic men aged 35 to 64 years as well as among Hispanic women aged 65 to 84 years, relative to their White counterparts.

In 2018, there were 32.0 LVAD and 20.3 HT procedures per 1000 HFrEF population (Table 2). Adjusted for age, race, and ethnicity, the LVAD rate among women (14.6 [95% CI, 12.5–16.7]) was 72% lower than among men (51.4 [95% CI, 46.6–56.2]) and the HT rate among women (13.3 [95% CI, 11.6–15.0]) was 62% lower than among men (34.8 [95% CI, 31.3–38.3]). The combined rate of LVAD and HT among women was 68% lower compared with men. Estimation using negative binomial regression indicated little change in the estimates (Table S6).

Table 2.

Rate of Use of LVADs and HTs by Sex, 2018

Sex group	Procedure rate (no. of procedures/1000 adults with HFrEF)
	Crude rate	Age‐race adjusted rate (95% CI)	Incidence rate ratio (95% CI)
LVAD
All	32.04	32.04	NA
Men	45.63	51.37 (46.58–56.15)	Reference
Women	14.26	14.6 (12.47–16.73)	0.28 (0.24–0.33)
HT
All	20.25	20.25	NA
Men	26.85	34.78 (31.26–38.3)	Reference
Women	11.61	13.27 (11.55–14.99)	0.38 (0.33–0.44)
LVAD and HT
All	52.28	52.28	NA
Men	72.48	86.28 (80.31–92.25)	Reference
Women	25.88	27.7 (24.79–30.61)	0.32 (0.29–0.36)

All estimates are for 2018, other than New York, for which we used 2017 data since 2018 data were unavailable. Crude rate gives the unadjusted (observed) rate of procedure counts. Age‐adjusted rates were obtained from Poisson regression of procedure counts on age, race, and ethnicity categories. Separate regression was estimated for each procedure. Incidence rate ratio estimates were also obtained from the same Poisson regression. As a test of sensitivity of the estimates to potential overdispersion, we also obtained corresponding estimates using negative binomial regression in place of Poisson regression (Table S6). We found little difference in all the estimates. HFrEF indicates heart failure with reduced ejection fraction; HT, heart transplant; LVAD, left ventricular assist device; and NA, not applicable.

Among women, the age‐adjusted rate of LVAD was 70% and 30% higher among Black patients (15.7 [95% CI, 12.6–18.9]) and Hispanic (20.3 [95% CI, 10.9–29.7]) compared with White patients (12.2 [95% CI, 10.2–14.2]; Table 3). Also, the age‐adjusted HT rate was 51% higher among Hispanic women (16.5 [95% CI, 12.0–21.0]) and similar among Black women (10.7 [95% CI, 8.3–13.1]) compared with White women (11.1 [95% CI, 9.2–13.1]). Among men, relative to White patients, the rates of LVAD and HT were 25% and 46% lower among Black patients, and among Hispanic patients, the LVAD rate was similar and the HT rate was 27% higher. The combined LVAD and HT procedure rate among Black men was 33% lower and similar among Hispanic men, compared with White men.

Table 3.

Rate of Use of LVADs and HTs by Race, Ethnicity, and Sex, 2018

Race and ethnicity group	Procedure rate (no. of procedures/1000 adults with HFrEF)
	Crude rate	Age‐adjusted rate (95% CI)	Incidence rate ratio (95% CI)
Women
LVAD
All	14.26	14.26	NA
Hispanic	20.61	20.31 (10.92–29.7)	1.7 (1.03–2.78)
Black	16.23	15.74 (12.58–18.9)	1.3 (1–1.68)
White	11.92	12.16 (10.15–14.18)	Reference
HT
All	11.61	11.61	NA
Hispanic	17.18	16.51 (12.03–20.99)	1.51 (1.09–2.08)
Black	11.31	10.7 (8.32–13.09)	0.96 (0.72–1.28)
White	10.73	11.13 (9.17–13.08)	Reference
LVAD and HT
All	25.88	25.88	NA
Hispanic	37.79	36.82 (25.57–48.07)	1.61 (1.16–2.24)
Black	27.54	26.42 (22.36–30.47)	1.14 (0.93–1.38)
White	22.65	23.28 (20.44–26.12)	Reference
Men
LVAD
All	45.63	45.63	NA
Hispanic	48.96	50.35 (39.35–61.35)	0.92 (0.72–1.17)
Black	44.28	41.64 (35.68–47.59)	0.75 (0.63–0.9)
White	44.95	55.41 (49.75–61.06)	Reference
HT
All	26.85	26.85	NA
Hispanic	41.70	42.46 (33.03–51.89)	1.27 (0.98–1.64)
Black	20.32	18.38 (14.93–21.82)	0.54 (0.44–0.68)
White	25.40	33.89 (30.13–37.64)	Reference
LVAD and HT
All	72.48	72.48	NA
Hispanic	90.66	92.86 (75.58–110.14)	1.05 (0.86–1.29)
Black	64.60	59.83 (52.5–67.16)	0.67 (0.58–0.77)
White	70.35	89.31 (82.67–95.96)	Reference

All estimates are for 2018, other than New York, for which we used 2017 data since 2018 data were unavailable. Crude rate gives the unadjusted (observed) rate of procedure counts. Age‐adjusted rates were obtained from Poisson regression of procedure counts on age, race, and ethnicity. Separate regression was estimated for each procedure. Incidence rate ratio estimates were also obtained from the same Poisson regression. Age‐adjusted rate and ratios are applicable only for comparison across race and ethnic subgroups within each sex group. HFrEF indicates heart failure with reduced ejection fraction; HT, heart transplant; LVAD, left ventricular assist device; and NA, not applicable.

All 3 racial and ethnic groups experienced increases in the combined rate of LVAD and heart transplants from 2010 to 2018 (Table 4). The increase between 2010 and 2018 in the combined rate was 89.3% (95% CI, 57.0%–128.2%) among Black adults, 88.6% (95% CI, 47.6%–141.0%) among Hispanic adults and 49.7% (95% CI, 29.8%–72.7%) among White adults.

Table 4.

Change in LVAD and HT Rates Between 2010 and 2018 Among the Population With HFrEF

Race and ethnicity group	Age‐ and sex‐adjusted rate (95% CI)		Change between 2010 and 2018, % (95% CI)
	2010	2018
LVAD
Hispanic	16.54 (12.32–20.76)	34.56 (27.15–41.97)	108.9 (49.8–191.6)
Black	13.29 (10.55–16.04)	28.96 (25.17–32.74)	117.8 (70.6–178.1)
White	21.28 (18.18–24.39)	33.82 (30.53–37.10)	58.9 (33.4–89.2)
HT
Hispanic	18.49 (13.86–23.11)	31.59 (25.85–37.32)	70.9 (25.5–132.6)
Black	9.86 (8.05–11.65)	15.00 (12.70–17.31)	52.2 (19.9–93.2)
White	17.25 (14.66–19.84)	23.78 (21.50–26.05)	37.8 (15.4–64.6)
LVAD+HT
Hispanic	35.10 (28.75–41.45)	66.19 (55.20–77.19)	88.6 (47.6–141.0)
Black	23.17 (19.63–26.71)	43.85 (39.11–48.58)	89.3 (57.0–128.2)
White	38.48 (33.65–43.30)	57.61 (53.60–61.62)	49.7 (29.8–72.7)

We pooled annual data from 2010 to 2018 to examine the change in procedure rates by race and ethnicity cohort. Age‐ and sex‐adjusted rates for 2010 and 2018 and the estimated change between 2010 and 2018 were obtained from a single Poisson regression of procedure counts on age, sex, and interaction of race and ethnicity group with year indicators. Separate regression was estimated for each procedure. HFrEF indicates heart failure with reduced ejection fraction; HT, heart transplant; and LVAD, left ventricular assist device.

Pooling data from 2010 to 2018, the combined rate of LVAD and HT procedures ranged across the 19 study states among White adults from 21.1 (95% CI, 16.9–25.2) in South Carolina to 69.9 (95% CI, 64.6–75.2) in Texas per 1000 adults with HFrEF (Figure and Table S7). The combined rate was lower among Black adults relative to White adults in 9 of the 19 study states, and among Hispanic adults in 6 states (P<0.05).

Figure 1. Combined rate of use of left ventricular assist devices and heart transplants by race, ethnicity, and state, 2010–2018.

We aggregated data on procedure volumes from 2010 to 2018 to examine differences in procedure rates by state and race and ethnicity cohort. Age‐ and sex‐adjusted rates were estimated from a Poisson regression of procedure counts on age, sex, and state indicators. Separate regression was estimated for each race and ethnicity group. HFrEF indicates heart failure with reduced ejection fraction.

In secondary analysis of incidence of LVAD exchange in California, we found that 4.3% of the patients who received an LVAD during 2010 through 2018 had a device exchange (Table S8), and this rate ranged from 2.6% among Black adults to 5.9% among White adults.

The combined LVAD and HT procedure rates using 2 alternative denominator populations, the population with heart failure and the overall census population, were lower among women by at least 68% compared with that among men (Table S9). However, both alternative measures were higher among Black adults (≥22%) relative to White adults, while the rate using the heart failure population was similar and the rate using the overall census population was lower among Hispanic adults (Table S10).

Discussion

Based on data covering 67% of the national adult population, and 85% and 70% of the Hispanic and Black population, our study highlights 5 findings. First, among the subpopulation with HFrEF, the combined rate of LVAD and HT is 68% lower among women relative to men. Second, among women, the combined rate is 61% higher among Hispanic women and similar among Black women relative to that among White women. Third, among men, the combined rate is 33% lower among Black men and similar among Hispanic men when compared with that among White men. Fourth, between 2010 and 2018, LVAD and HT rates increased for each of the 3 racial and ethnic groups and for both men and women. Fifth, among the subpopulation with HFrEF, rates of LVAD and HT varied considerably by state, with a 3‐fold difference between states with the lowest and highest procedure rates.

While prior studies have noted the potential underuse of advanced heart therapies among women, when compared with men, to our knowledge, our study is the first to quantify the extent among a population at higher risk for the therapies. We recognize that not all those with heart failure with ejection fraction <45% will be candidates for advanced heart therapies, and that reduced ejection fraction may be one criterion for referral for advanced heart therapies. ³⁷ Even after adjusting for the lower prevalence of HRrEF among women, we found that the rate of use of LVAD and HT was 68% lower among women when compared with that for men. Reasons for this disparity remain unclear. Prior studies were based on examination of patients referred for evaluation for eligibility for advanced heart therapy. ⁹ , ³⁸ In the study by Herr et al ⁹ of patients referred for advanced heart therapy from 9 medical centers, only 27% of the 515 patients were women. Our study estimated that women constituted 44% of the population with HFrEF; therefore, a lower rate of referral among women, compared with men, may be one source of disparities in the procedure rate by sex. The prior studies point to explicit and implicit biases in the evaluation process for eligibility for advanced heart therapy, particularly in the subjective psychosocial criteria, including provider perceptions about availability of caregiver support, health literacy, and resources for care continuity. ¹⁷ , ³⁸ , ³⁹

To our knowledge, our study also provides the first estimates of population‐level rates of use of LVAD and HT among Black, Hispanic, and White adults with HFrEF. Our estimates of within‐sex comparisons indicate significantly lower use of LVAD and HT among Black men compared with White men, while among Hispanic men and women and Black women the procedure rates are either higher or similar to that of their White counterparts. Prior studies of comparisons are based on convenience samples of health system data and provide mixed evidence. In the study by Herr et al, ⁹ among the patients referred for evaluation, White patients were more likely to be offered LVAD and HT compared with non‐White patients. Among 702 patients with advanced heart failure in a single high‐volume center, Jones et al ¹⁰ found similar rates of LVAD receipt among White and Black patients. Prior studies of LVAD rates using large‐population–based data were based on the overall census population as the denominator. Combining counts of LVAD procedures from the Interagency Registry of Mechanically Assisted Circulatory Support (INTERMACS) registry database with census population counts, Breathett et al ⁵ found that the LVAD rate in 2012 among Black adults was 84% higher and among Hispanic adults 55% lower compared with that among White adults. Similarly, using INTERMACS, Medicare, and census population data, Bourque et al ⁶ found that, compared with White adults, those from minoritized groups had an 87% higher LVAD rate in 2014. Consistent with these findings, our study estimates of LVAD rates at the census population (2018) indicate a 166% higher rate among Black and 13% lower rate among Hispanic adults compared with White adults. However, after accounting for the prevalence of HFrEF, our study indicates that the LVAD rate among adults with HFrEF was 25% lower among Black men and similar among Hispanic men compared with White men. The change in pattern of White–Black differences is largely accounted for by the higher prevalence of (1) heart failure and (2) reduced ejection fraction among Black adults. This is an important methodological correction for the literature; for example, among adults aged 35 to 64 years, the prevalence of heart failure was 106% higher and HFrEF was 171% higher among Black men compared with White men.

Prior evidence of heart transplant rates by race and ethnicity have been estimated using the population in the transplant waitlist registry as the denominator. Chouairi et al ⁷ used data on patients listed for an HT from the United Network for Organ Sharing (INOS) registry database (2011–2020) and estimated that, compared with White patients, Black patients had a 13% lower likelihood of receiving an HT, while Hispanic patients had a similar likelihood. Using UNOS data, Trivedi et al ⁸ reported that in 2019 the proportion of Black patients listed for transplant was 25% and the proportion of Black patients out of all transplanted patients was 26%; that is, Black patients did not experience inequity in the receipt of HT. Assignment to be placed on transplant waitlists may not be equitable across the population of eligible patients, with some patients excluded on the basis of lack of insurance and potentially other socioeconomic criteria (access to providers, family and social supports). ¹⁷ , ¹⁹ , ²³ , ⁴⁰ Using a population of patients with HFrEF is an important methodological advance, as it avoids confounding race and ethnicity with socioeconomic criteria that influence placement on the transplant list. Indeed, our study found that the rate of HT was 46% lower among Black men and 27% higher among Hispanic men relative to White men.

INTERMACS data indicate an increase in the proportion of LVAD as destination therapy. ⁴¹ As such, the combined rate of LVAD and HT among the population with HFrEF enables comparison of overall access to mechanical support and transplants across subgroups. We found a significantly lower rate of combined procedures in 2018 among Black men (33% lower). compared with White men. Also, consistent with prior studies, we found that the use of LVAD and HT has trended to increase over time, not only overall but also among Black, Hispanic, and White adults. Further, our findings point toward a greater increase among Black and Hispanic adults than White adults. ⁵ , ⁷ , ⁸

Inequities in LVAD and HT use are consistent with the ubiquitous disparities in receipt of cardiovascular therapies, specifically in interventions for structural heart disease. ³ , ⁴² Recent studies using administrative and health systems data have documented lower rates of LVAD, HT, transcatheter mitral valve repair, left atrial appendage occlusion, and transcatheter aortic valve among Black and Hispanic patients. ¹¹ , ¹⁷ , ⁴³ , ⁴⁴ Treatment for structural heart disease requires patients to navigate referrals to multiple specialists involving complex diagnostic evaluations. Navigating systems of care may be more difficult for patients who are uninsured or underinsured. Prior studies have also identified lower adherence to treatment protocols, limited social supports, and patient preference as factors in lower treatment rates among Black men. ¹⁷ However, lower adherence may be driven by limited health literacy, poor patient–provider communication, and high financial burden. Physician characterization of poor adherence and preference may be influenced by subjective psychosocial criteria and perceived bias among providers. ⁹ , ¹⁷ In an interview‐based study of clinician decision making for advanced heart failure intervention based on hypothetical clinical vignettes, Breathett et al ⁴⁰ found that clinicians aged ≥40 years who were randomized to vignettes of Black patients were less likely to allocate heart transplantation. A recent study with explicit patient preference data found similar preference for advanced heart failure therapy among Black and White patients. ¹¹ Inequities in receipt of LVAD and HT also manifest the broader health care disparities arising from structural racism, through embedded biases in the laws and practices in various domains of everyday life, including government, economy, education, and housing. ¹⁷ , ³⁹ , ⁴⁵ , ⁴⁶ Our findings of LVAD and HT use by state identify several states with no significant difference in use by race and ethnicity. However, in states with larger minoritized populations (California, New York, and Texas), use of the procedures is systematically lower among Black and Hispanic adults.

Limitations and Strengths

This study has several limitations. First, our estimates of heart failure and HFrEF prevalence were based on self‐reported and not adjudicated heart failure. Using heart failure identified in Medicare claims data (among those aged ≥65 years) as the reference, Gure et al ⁴⁷ found that self‐report data underestimated heart failure prevalence, particularly among Black and Hispanic adults. However, given the limited sources of data on prevalence of heart failure, the National Health and Nutrition Examination Survey provides the most comprehensive objective estimates by race and ethnicity at the population level. Second, we used incidence of reduced ejection fraction (≤45%) among adults with heart failure in MESA to identify a subgroup of heart failure patients at higher risk of advanced heart failure therapies. In the absence of detailed phenotype information relating to heart failure symptoms and severity, our denominator counts of the population with HFrEF do not represent the clinically adjudicated population indicated for advanced heart therapy. ¹⁷ Third, our data on LVAD use do not account for device exchanges. ⁴⁸ However, data from one state that permits identification of device exchange (California) indicates that the incidence of device exchange among LVAD recipients is about 4%, and the differences by race and ethnicity are relatively small (with incidence ranging from 2.6% among Hispanic adults to 5.9% among White adults) and not likely to significantly affect our estimates of the relative differences in LVAD use by race and ethnicity. Finally, LVAD and HT use are influenced by patient preference and treatment adherence, which are unobserved in our data and may confound our estimates. ⁹ , ¹¹

Recognizing the limitations, the main strength of this study is that it operationalizes the measurement of differences in use of LVAD and HT among Black and Hispanic adults relative to White adults using available objective measures representative of the 3 racial and ethnic subpopulations at the national level. Our estimates of lower rates among Black and Hispanic adults are consistent with the growing evidence of inequities in use of advanced heart failure therapies on the basis of clinical experience. ⁹ , ¹⁰ , ¹¹ , ¹⁷ , ³⁹ , ⁴⁰ , ⁴⁶ These estimates are in contrast with higher rates among Black adults reported by prior studies based on census population as the denominator. Nevertheless, studies using more detailed heart failure phenotype data are needed to examine the precision and robustness of our estimates of LVAD and HT rates by race and ethnicity.

Conclusions

Adjusting for age and differences in the prevalence of heart failure and HFrEF, the use of LVAD and HT is lower among women compared with men and among Black men compared with White men. Among women, use of LVAD and HT among Black and Hispanic patients was no lower than among White patients. From 2010 to 2018 racial disparities may have narrowed, with a faster increase in LVAD and HT rates among Black adults compared with White adults; nevertheless, there are still significant inequities in the procedure rates.

Sources of Funding

This research was supported by a grant from the National Institute on Minority Health and Health Disparities (R01MD011594, A. Hanchate, PI).

Disclosures

None.

Supporting information

Data S1–S4

Tables S1–S10

Figure S1