Trends and demographic disparities in heart failure mortality rates in non-alcoholic fatty liver disease: a population-based retrospective study in the United States from 1999 to 2020

Abstract

Objectives: This study aimed to analyze two decades heart failure (HF) mortality data in Non-Alcoholic Fatty Liver Disease (NAFLD), now known as Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD), in the United States (US), identifying patterns and disparities in mortality rates. Methods: A retrospective analysis was conducted using mortality data from the CDC WONDER database spanning 1999-2020. Age-adjusted mortality rates (AAMRs) per 1,000,000 persons were calculated, and trends were assessed using Average Annual Percentage Change (AAPC) and Annual Percent Change (APC) using Joinpoint 5.0.2. Data were stratified by year, sex, race/ethnicity, urbanization, and census regions. Results: From 1999-2020, 68,436 HF-related deaths occurred among US adults with NAFLD. The overall AAMR increased from 12.49 in 1999 to 24.30 in 2020, with an AAPC of 3.05 (95% CI: 2.80 to 3.31, P < 0.001), with a steep rise in AAMR from 2017-2020 and an APC of 12.35 (95% CI: 9.71 to 15.99). American Indian or Alaskan natives had the highest AAMRs (28.63), followed by Hispanics (20.05), and African Americans or Blacks (14.51). The highest mortality regionally was in the Southern region (AAMR: 16.05) and nonmetropolitan areas had higher AAMRs than metropolitan areas (16.63 vs. 13.76). Conclusions: This analysis demonstrated increasing mortality rates from HF in NAFLD, with a sharper increase in recent years. This also showed nonmetropolitan areas, the Southern region of the US, and minority populations had higher mortality rates, which highlights at-risk populations and opportunities for important public health interventions.

Introduction

Non-Alcoholic Fatty Liver Disease (NAFLD), now known as Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD), is on the rise. NAFLD has become the most common cause of chronic liver disease worldwide and has been estimated to affect 25% of the global population and continues to increase [1,2]. NAFLD refers to the presence of steatosis in greater than 5% of hepatocytes without significant alcohol exposure and is preferably detected via elastography or ultrasound [3]. When steatosis exists with hepatic inflammation the condition is termed Non-Alcoholic Steatohepatitis (NASH), which can present with elevated liver enzymes and lead to fibrosis, and is often asymptomatic until complications of cirrhosis develop if fibrosis progresses [3]. In 2024, resmetirom became an approved drug for various stages of fibrosis from NAFLD and can reverse fibrosis in ~25% of cases, though a total body weight loss of 10% has also shown to lead to a resolution in NASH in up to 90% of cases [4,5]. The United States (US), which has the highest prevalence of overweight and obese individuals worldwide, is particularly impacted by NAFLD and is estimated to bear an annual cost of $103 billion from this condition alone [6,7]. Projections estimate 100 million Americans will suffer from NAFLD by 2030 with no signs of improvement on the horizon given the epidemic of childhood obesity and diabetes [8].

While noting the increasing prevalence of NAFLD, parallel focus must be given to the heart, as adverse cardiovascular events are the leading cause of death in patients with liver disease [9]. HF has been repeatedly linked to NAFLD, as those with metabolic liver disease are 20% more likely to suffer from HF than those without [10]. NAFLD has been found to be an independent risk factor for diastolic dysfunction as well as a contributor to coronary disease which can further impact systolic function [11]. The pathophysiologic explanations of this linkage have largely been attributed to effects of metabolic disease, including the toxicity of lipid metabolism, inflammatory cascades from fatty deposition, and neurohormonal activation [11,12].

Evidence suggests the burden of HF and NAFLD are not felt equally throughout the US. Lower household income, minority status, and South and West US regions have been shown to suffer from increased liver related mortality [7,13]. In heart failure, minorities also suffer disproportionately, and these inequities are worsening rather than improving [14,15]. However, data demonstrating population details of the specific impact of HF in NAFLD is limited and there remains a critical knowledge gap regarding the influence of demographic disparities on these mortality trends in the US. The aim of this study was to investigate the trends and disparities by sex, race, geography, and overall HF-related death in patients with NAFLD from 1999 to 2020.

Methods

Population and study setting

This comprehensive investigation used the Centers for Disease Control and Prevention’s Wide-Ranging Online Data for Epidemiologic Research (CDC WONDER) database to collect information from death certificates. The study focused on instances of persons with heart failure related mortality who held a concurrent diagnosis of NAFLD/NASH between the years 1999 to 2020. Inclusion criteria consisted of decedents aged 25 to 85 years whose death certificates included ICD-10 codes corresponding to both heart failure (I50.x) and NAFLD/NASH (K74.0, K74.6, K75.8, K76.0, K76.9) from the 10th version of the International Classification of Diseases and Related Health Problems (ICD-10) [16-18]. Exclusion criteria included individuals outside this age range, and unrelated causes of mortality. The dataset comprises information regarding the causes of death from death certificates from all 50 states and the District of Columbia. At-death records from the Multiple Causes of Death Public Use registry were examined to find cases of NAFLD and HF. The research did not necessitate any approval from a regional institutional review board because it relied on deidentified public use data issued by the government. The STROBE standards for reporting observational research were followed in this study.

Data abstraction

NAFLD and HF-related deaths, population size, and location of death (including medical facilities [outpatient, emergency room, inpatient, death on arrival or status unknown], home, hospice, and nursing home/long-term care facility) were extracted. Demographics (sex, race and ethnicity, and age) and regional information (urban-rural and state) were extracted from 1999 to 2020. Races were defined as Non-Hispanic (NH) White, NH Black, Hispanic, Native American or Alaska Native, and NH Asian or Pacific Islander patients. These racial and ethnic categories have previously been used within analyses from the CDC WONDER database and rely on reported data on death certificates. Under the National Center for Health Statistics Urban-Rural Classification Scheme, the population was categorized into rural versus urban areas. Rural regions were defined as those with fewer than 50,000 people based on the 2013 U.S. Census. Urban areas were divided into medium/small metropolitan areas with a population ranging from 50,000 to 999,999, and large metropolitan areas claimed a population of 1 million or more. These areas were further divided into four main geographical categories: Northeast, Midwest, South, and West, using the United States Census Bureau.

Statistical analysis

Mortality rates per 1,000,000 individuals were computed for both crude and age-adjusted data across the period from 1999 to 2020 as the main indicators of trend. These rates were divided into categories by year, gender, race/ethnicity, state, and urban/rural status, with 95% confidence intervals (CIs). The crude mortality rates were determined by dividing the total number of HF-related deaths in patients with NAFLD by the corresponding U.S. population each year. Age-adjusted mortality rates (AAMRs) were calculated by standardizing HF-related deaths in NAFLD patients to the United States population in 2000. The annual percent change (APC) and its associated 95% CI in AAMR were calculated using the Join-Point Regression Program (Version 5.0.2, National Cancer Institute). This method identifies significant changes in AAMR over time by fitting a series of joined straight lines on a log scale where temporal variation occurred. The Join-point regression statistical software was set to determine a maximum of 3 Joinpoints (21 years of analysis were included) where significant temporal variation existed in the trend. APCs with 95% CI for the AAMR were calculated at the identified line segments linking Joinpoints using the Monte Carlo permutation test. The weighted average of the APCs was calculated and reported as average APCs (AAPCs) and corresponding 95% CIs as a summary of the reported mortality trend for the entire study period. APC was classified as increasing or decreasing based on whether the slope reflected changes in mortality significantly differed from zero when assessed via two-tailed t-testing at a significance level of P < 0.05.

Results

Overall and annual trends for HF-related AAMR in NAFLD

From 1999 to 2020, HF was reported as a contributing cause for 68,436 deaths among adults with NAFLD aged 25 and older in the United States (Supplementary Data). The overall AAMR for HF-related deaths in NAFLD ranged between 12.49 (95% CI 11.97 to 13.01) per 1 million individuals in 1999 to 24.30 (95% CI 23.71 to 24.9) per 1 million individuals in 2020 (Figure 1A), with an Average Annual Percentage Change (AAPC) of 3.05 (95% CI 2.80-3.31) (p value < 0.001). AAMR experienced an initial decline from 1999 to 2010 (APC: -1.49, 95% CI -2.23 to -0.89) (p value = 0.003), followed by a notable upward slope from 2010 to 2017 (APC: 6.58, 95% CI 3.92 to 7.86) (p value = 0.01) and a drastic rise from 2017 to 2020 (APC 12.35, 95% CI 9.71 to 15.99) (p value < 0.001).

Figure 1.

Heart failure-related mortality trends in NAFLD in the United States, 1999 to 2020 with (A) overall HF-related AAMR in NAFLD per million, (B) sex-stratified HF-related AAMR in NAFLD per million, (C) race/ethnicity Stratified HF-related AAMR in NAFLD per million, (D) regionally stratified HF-related AAMR in NAFLD per million, (E) urbanization stratified HF-related AAMR in NAFLD per million.

HF-related AAMR in NAFLD stratified by sex

HF in NAFLD resulted in 40,777 deaths in the male population in the last two decades while 27,659 deaths were reported in the female population. Adult males exhibited higher overall AAMRs consistently over the study period compared to adult females (overall AAMR for males: 19.33, 95% CI: 19.14 to 19.52; for females: 10.26, 95% CI: 10.14 to 10.38). The AAMRs of both male and female cohorts increased in the two decades with a more prominent rise in females [Males: AAPC: 2.72, 95% CI: 2.46 to 2.99 (p value < 0.001); Females: AAPC: 3.27, 95% CI: 2.90 to 3.67 (p value < 0.001)] (Figure 1B).

The AAMR for adult males decreased steadily from 17.24 (95% CI: 16.28 to 18.19) in 1999 to 15.30 (95% CI: 14.50 to 16.11) in 2010 (APC: -1.80; 95% CI: -2.59 to -1.11) (p value = 0.02), succeeded by a substantial increase to 23.09 by 2017 (APC: 6.66; 95% CI: -0.3 to 7.68) (p value = 0.055); and a sharp incline to 31.29 in 2020 (APC: 10.93; 95% CI: 8.24 to 14.34) (p value < 0.001). Similarly, the AAMR for adult females, decreased from 8.87 in 1999 to 7.56 in 2009 (APC: -1.60; 95% CI: -3.31 to -0.59) (p value < 0.006), followed by an increase to 11.49 by 2016 (APC: 5.18; 95% CI: 1.73 to 7.48) (p value < 0.01) and a steep ascent to 18.54 by 2020 (APC: 12.85; 95% CI: 10.09 to 18.57) (p value < 0.001).

HF-related AAMR in NAFLD stratified by race/ethnicity

Pronounced differences were found when examining deaths among different racial/ethnic groups. The highest number of deaths occurring in Non-Hispanic (NH) White individuals (51,581; 75.4%), followed by Hispanic individuals (7,468; 10.9%), NH Black individuals (6,956; 10.2%), NH Asian or Pacific Islanders (1,455; 2.1%), and the lowest number was in NH American Indian or Alaska Native individuals (807; 1.2%). AAMRs were highest among NH American Indian or Alaska Natives, followed by Hispanic or Latinos, NH Black or African Americans, NH Whites, and NH Asian or Pacific Islander populations (overall AAMR: NH American Indian or Alaska Native: 28.63, 95% CI: 26.56 to 30.7; Hispanic or Latino: 20.05, 95% CI: 19.57 to 20.52; NH Black or African American: 14.51, 95% CI: 14.16-14.86; NH White: 13.71, 95% CI: 13.59 to 13.83; NH Asian or Pacific Islander: 8.02, 95% CI: 7.60 to 8.44).

The AAMR of all the races increased to variable degrees from 1999 to 2020, with the incline most pronounced in NH White cohort [NH White: AAPC: 3.16, 95% CI: 2.90 to 3.46 (p value < 0.001); NH Black: AAPC: 2.26, 95% CI: 1.66 to 2.94 (p value < 0.001); Hispanic: AAPC: 1.95 95% CI: 1.65 to 2.51 (p value < 0.001); NH Asian or Pacific Islander: AAPC: 0.87, 95% CI: -0.36 to 2.32) (p value = 0.13)] (Figure 1C). Trend analysis of NH American Indian or Alaskan Native was not possible due to unreliable data.

HF-related AAMR in NAFLD stratified by geography

Disparities in AAMRs were observed among different states, with AAMRs ranging from as low as 9.23 (95% CI: 8.89 to 9.57) in New York to 25.08 (95% CI: 23.39 to 26.77) in West Virginia. States falling within the top 90th percentile included West Virginia, Texas (23.04), Mississippi (22.46), Kentucky (21.62), and Oklahoma (21.19), while states in the lower 10th percentile included New York, Nevada (9.34), Arizona (9.48), Florida (10.11) and District of Columbia (10.14) (Figure 2).

Figure 2.

State-Stratified HF-related AAMRs in NAFLD per 1,000,000 in the United States from 1999 to 2020.

For the census regions, highest mortality rates were observed in the South (16.05, 95% CI: 15.86 to 16.24), followed by the West (15.07, 95% CI: 14.83 to 15.30), Midwest (12.73, 95% CI: 12.52 to 12.94), and Northeast (11.53, 95% CI: 11.31 to 11.75). The AAMRs of all regions increased over the course of the study with greatest rise in the South [South: AAPC: 3.46, 95% CI: 3.11 to 3.95 (p value < 0.001); Midwest: AAPC: 3.21, 95% CI: 2.84 to 3.60 (p value < 0.001); West: AAPC: 2.25, 95% CI: 1.81 to 2.78 (p value < 0.001); Northeast: AAPC: 2.23, 95% CI: 1.52 to 2.90 (p value < 0.001)] (Figure 1D).

Regarding urbanization, nonmetropolitan areas exhibited higher AAMRs than metropolitan areas throughout the study period, with overall AAMRs of 16.63 (95% CI: 16.35 to 16.91) and 13.76 (95% CI: 13.64 to 13.88), respectively. The AAMR of both, Metropolitan and Nonmetropolitan areas increased from 1999 to 2020, with a greater increase in Nonmetropolitan areas [Nonmetropolitan: AAPC: 4.00, 95% CI: 3.62 to 4.39 (p value < 0.001); Metropolitan: AAPC: 2.83, 95% CI: 2.57 to 3.11 (p value < 0.001)] (Figure 1E).

Discussion

This analysis showed a concerning increase in HF-related mortality in NAFLD patients across recent years. Trends demonstrated significant gender, racial (Figure 3), and geographic disparities in HF mortality. Higher mortality rates were seen in males, though with a steeper increase in females, in non-metropolitan areas, the southern US, and in minorities, specifically with Non-Hispanic Alaska Natives and Hispanics experiencing the highest mortality rates.

Figure 3.

CENTRAL ILLUSTRATION: Demographic profiles of heart failure-related mortality among adults with NAFLD Aged 25 to 85+ in the United States, 1999 to 2020.

Overall trends

In this study, the overall AAMR for HF deaths in those with NAFLD decreased from 1999 to 2010 but showed an increase from 2010 to 2017 and a sharp rise from 2017 to 2020. This initial decline in mortality could be attributed to several factors, including improvement in healthcare resources, improved detection of clinically significant liver disease, more effective antihypertensive and lipid lowering medications, and public health efforts such as limiting tobacco use. Similar studies on HF mortality trends in the US have suggested the latter exacerbation of mortality, however, can likely be attributed to impacts of the worsening obesity epidemic and metabolic dysfunction [14]. These metabolic risk factors, including insulin resistance and impaired glucose and lipid metabolism, are primary contributors to cardiometabolic disease, and result in decreased myocardial energy metabolism leading to cardiac dysfunction [19-21]. Particular attention should be given to 2020, the peak of mortality in our study, in which the impact of COVID-19 must be considered as the virus’ severity and outcomes was worse in obese individuals [7,11].

Disparities by sex

Our study showed a higher AAMR for HF-related deaths in those with NAFLD in males compared to females, though with a steeper increase in the rates of female deaths. This is concordant with prior studies which found lifetime risk of HF and HF mortality to be higher in men than women. Cardiometabolic risk factors are a necessary link between these two diseases and can be further affected hormonally, as, for example, estrogen production can impact lipid metabolism and possibly lead to reduction of cardiac fibrosis in females [22]. The increasing rates of female mortality are notable, as evidence has found diabetes to be more strongly associated with HF in women than in men, planting a logical hypothesis that the obesity epidemic could disproportionately afflict women in relation to HF [23].

Disparities by race and ethnicity

The results of our study show a racial gap in HF-NAFLD mortality, with Non-Hispanic (NH) American Indian or Alaska Native individuals suffering from the highest mortality rates followed by other minorities. This correlates with prior demographic data in both hepatic and cardiac literature. Higher rates of liver-related mortality are reported in minorities, with the highest prevalence of NAFLD in Hispanic populations, and HF prevalence and mortality more heavily impact those who identify as American Indian, African American, or Hispanic [7,24,25]. An American Heart Association 2020 report echoed similar findings, showing the highest HF-related mortality rates in African American men [26]. A likely culprit for these racial disparities could be social determinants of health, as low socioeconomic status has been found to be independently associated with inpatient HF-related mortality [25,27]. These factors include socioeconomic position, social support, reduced access to healthcare, poor residential environments, and food deserts - all of which have a negative association with HF morbidity and mortality [25].

Geographic disparities

Our study found that mortality events were highest in the Southern and non-metropolitan areas. In terms of geographical presentation, the South exhibited the highest mortality followed by the West, a trend also elicited when studying regional tendencies in liver disease mortality [13]. Furthermore, states across the US have also shown variation in NAFLD and HF mortality. The states in the top 90^th percentile were West Virginia, Texas, Mississippi, Kentucky, and Oklahoma; while the states in the bottom 10^th percentile were New York, Nevada, Arizona, Florida, and District of Columbia. Prior data has also demonstrated that from 1999-2017 the South and Midwest regions had higher HF AAMRs than other geographic regions, most notably Oklahoma and Mississippi [22]. This geographic disparity has been so historically prevalent, the southern states have at times been referred to as “the Heart Failure Belt” though this has been contested, with recent trends showing mortality may be more closely linked to urbanization than region [28]. When studying population density in relation to HF, rural counties have consistently been found to have higher mortality rates compared to urban counties [29]. Ultimately, the variation in public health investment and infrastructure, lack of access to both primary and specialty care in rural areas, and sociodemographic factors, and sociocultural attitudes all influence these disparities observed among regions and states [13,29].

Limitations

Several limitations should be acknowledged in this study. Firstly, reliance on the CDC WONDER database introduces potential errors or omissions in the data. Moreover, the study’s focus on individuals aged 25 and older may overlook variations in younger age groups, potentially biasing trend interpretations. The lack of detailed clinical data, such as biomarkers, treatment modalities, and interventions inhibits comprehensive understanding of mortality drivers. The absence of granular analyses for distinct cases of NAFLD and HF further complicates causal inference, hindering targeted investigations. Furthermore, socioeconomic factors, which can profoundly influence health outcomes, were not accounted for in this study.

Conclusion

Main conclusion

HF-related mortality in patients with NAFLD, now known as MASLD, has increased from 1999 to 2020. Most notably, our study underscores a significant recent mortality increase from 2017 to 2020. Particularly high mortality rates were observed among NH American Indians or Alaska Natives, residents in the Southern US regions, and in non-metropolitan areas. HF in NAFLD significantly impacts cost and mortality for Americans and is a serious healthcare problem with a worsening trajectory. This data shows the disparities felt by Americans with these conditions and understanding these disparities is essential for guiding healthcare policies and developing targeted public health interventions to better serve at-risk populations.

Future studies

Future work on this subject should continue to focus on connections and interventions involving the interrelatedness of NAFLD and HF. As treatment options increase for metabolic disease, such as the use of glucagon-like peptide-1 receptor agonists, these mortality trends should continue to be studied to monitor for population level impacts.

Disclosure of conflict of interest

Chien-Jung Lin is supported by the American Heart Association Career Development Award 24CDA1277194.

Abbreviations

HF

Heart Failure

NAFLD

Non-Alcoholic Fatty Liver Disease

NASH

Non-Alcoholic Steatohepatitis

MASLD

Metabolic Dysfunction-Associated Steatotic Liver Disease

AAMRs

Age-adjusted mortality rates

AAPC

Average Annual Percentage Change

APC

Annual Percent Change

CDC WONDER

the Centers for Disease Control and Prevention’s Wide-Ranging Online Data for Epidemiologic Research

NH

Non-Hispanic