National, Regional, and Global Cardiomyopathy Burden from 1990 – 2019

Objective: To gauge the incidence of both alcoholic cardiomyopathy (AC) and other cardiomyopathy (OC) in 204 nations and regions over the 1990 – 2019 period. Methods: The Global Burden of Disease Study 2019 was leveraged to gather comprehensive information pertaining to both AC and OC. The calculated prevalence rates, mortality rates, and disability-adjusted life years (DALYs) related to these forms of cardiomyopathy were presented as counts and age-standardized rates (ASRs) per 100,000 persons with percentage change and 95% uncertainty intervals (UIs). The overall burden of cardiomyopathy was established by combining data pertaining to both AC and OC, with calculations and corresponding 95% condence intervals (CIs) being determined based upon corresponding standard error values determined by dividing the width of 95% UI by 1.96×2. Results: At the national level, ASR rates for cardiomyopathy varied by 144.6-fold among surveyed countries in 2019, while the rates ASR rates for AC and OC respectively varied by 408.0-fold and 380.6-fold. Age-standardized cardiomyopathy-associated mortality rates also varied by 58.4-fold among surveyed nations. The highest such AC-related death rates were observed in Latvia, with an age-standardized mortality rate of 16.3 (95%UI 11.0 to 21.1) per 100,000 persons, whereas some nations reported 0 AC-related deaths in 2019. National age-standardized OC-related mortality rates varied by 105-fold among surveyed nations. Despite overall reductions in cardiomyopathy ASRs from 1990 – 2019, the overall numbers of cases, deaths, years lost in life (YLLs), and years lost due to disability (YLDs) have all increased over this period for both AC and OC. Conclusion: These data suggest that cardiomyopathy, including both AC and OC, remains a persistent threat to global public health as evidenced by increased numbers of cases, deaths, and DALYs over the past three decades, with clear geographic differences in overall cardiomyopathy burden. Additional study will be essential to more fully understand the risk factors associated with the development of this disease in an effort to guide its prevention, early diagnosis, and treatment. DALY-related the burden of AC and as of 2019 at the national, regional, and global levels. these analyses, there were approximately 0.7 million and 3.7 million worldwide AC and cases, respectively, in 2019. there were 2.4 million and 5.7 million DALYs AC and respectively, 2019 at in a manner that evaluated differences in global AC and OC prevalence as a function of SDI and region. At all time points, AC prevalence were higher in regions with high SDI values irrespective of sex, while OC prevalence rates were higher in areas with low SDI values. The greatest decreases in overall cardiomyopathy prevalence were observed in high-middle SDI regions, while the largest increase in AD prevalence was observed in low-middle SDI regions, and no increasing trends were observed for OC. The primary components of SDI include gross fertility rates, per capita income levels, and years of education. As such, social can inuence cardiomyopathy prevalence rates and associated changes. One possible explanation for this may be that healthcare surveillance systems in high SDI regions are more comprehensive, while there is signicant room for improving healthcare systems in low, low-middle, and middle SDI regions in order to improve the prevention and diagnosis of both AC and OC, particularly among individuals with asymptomatic disease. 2019 varied among geographic regions, with highest age-standardized rates overall Eastern


Introduction
Cardiovascular disease (CVD) is the most prominent driver of global morbidity and mortality, resulting in high rates of disability or lost productivity that impose a marked burden on affected patients and society as a whole (1) . Cardiomyopathy is an uncommon and thus often-overlooked yet severe form of CVD that is characterized by clinical ndings including abnormal cardiac structural characteristics, arrhythmia, and/or heart failure (2) . Cardiomyopathy can result in the sudden death of affected individuals during childhood or adolescence and may require patients to undergo cardiac transplantation in some cases (3) .
Among the all types of cardiomyopathy, alcoholic cardiomyopathy (AC) is a special one. It is caused by heavy alcohol use over a prolonged period. Alcohol and its metabolites are cardiotoxic. Myocardial inhibition secondary to alcohol is initially reversible, but prolonged alcohol consumption can lead to irreversible dysfunction (4) . However, there is neither a certain amount of alcohol known to be toxic to myocardial cells nor is there a special period of exposure time to cause it to happen (5) . Furthermore, not all chronic alcohol abusers develop AC nally and prevalence of AC still cannot be accurately assessed due to incomplete data and its uncertainty (6) .
A comprehensive overview of the epidemiological characteristics of cardiomyopathy in particular populations and countries would be invaluable as a tool for guiding the development of public health policy and informing decision-making aimed at preventing, diagnosing, and treating this debilitating disease.
However, robust datasets pertaining to cardiomyopathy morbidity and mortality are primarily available for highly developed nations that had standardized the diagnosis and evaluation of affected patients (7) . The Global Burden of Disease (GBD) study, which is an ongoing analysis of the incidence of over 350 diseases in around 195 nations and regions throughout the world, offers a unique opportunity to gain a more robust overview of the true global burden of cardiomyopathy. Since 2016, the GBD study has subdivided cardiomyopathy into the alcoholic cardiomyopathy (AC) and other cardiomyopathy (OC) subtypes (8) . There is thus a clear need to more fully explore the national, regional, and global AC and OC burden and to explore trends in cardiomyopathy-related disease ndings as a function of time.

Data sources
The present study was conducted using data derived from the GBD study coordinated by the Institute for Health Metrics and Evaluation (IHME). Each annual GBD study update entails the re-estimation of the entire data time series based on new advances in medical knowledge, modeling, estimation approaches, and data in an effort to ensure the continuous improvement of study quality. The GBD 2019 study was published in October 2020 and included epidemiological data pertaining to 369 diseases/injuries, 286 causes of death, and 87 risk factors in 204 nations and regions, with sub-national estimates additionally being reported for a subset of countries. For the present study, we obtained published estimates pertaining to the prevalence rates, mortality rates, and disabilityadjusted life years (DALYs) associated with cardiomyopathy (9) . As the GBD 2019 study consists of aggregated, de-identi ed data, the University of Washington Institutional Review Board approved a waiver of informed consent with respect to the use of these data for research purposes.
Estimates of mortality rates were obtained from vital registration data sources for both OC and AC and from verbal autopsy data sources for AC using the ICD codes detailed above. The standard Cause of Death Ensemble model (CODEm), which incorporates a range of covariates and individual models (including linear mixed-effects regression and spatiotemporal Gaussian process regression models) to generate cause of death (COD) predictions, was used to estimate rates of AC-and OC-related mortality. Out-of-sample predictive validity tests were employed to evaluate utilized individual and ensemble models, with experts in appropriate disease elds having vetted these models in addition to their having been validated by the IHME and international collaborators. Estimated mortality rates were scaled based on other COD-related estimates to yield all-cause mortality estimates summing to 100% within individual year, age, sex, and location groups (10) . As mortality-related data availability was limited for some nations, it was converted from incidence data through mortality-to-incidence ratio modeling.
The Bayseian mixed-effects DisMod-MR 2.1 meta-regression tool, which was designed to analyze GBD data, was used to estimate the prevalence of OC and AC (11) . Consistent prevalence estimates for all locations were generated by integrating the incidence, remission, and mortality rates for particular causes. Prevalence estimates did not incorporate non-literature data sources other than hospital or claim data. Outpatient data were excluded from these analyses as they were available at substantially lower levels than corresponding inpatient and claim data from matched locations. For high-income nations, inpatient hospital data found to be either 2-fold above or 0.5-fold below the median absolute deviation value for the corresponding age-sex group were additionally excluded (11) . DALYs were measured by summing both years lost in life (YLLs) and years lost due to disability (YLDs), providing a metric for health losses associated with a particular disease state. YLLs were quanti ed by multiplying estimated numbers of deaths by age with the corresponding standard life expectancy for individuals in that age group. YLDs were quanti ed by multiplying the number of cases of a given disease by the disability weight (ranging from 0-1, with 0 and 1 respectively corresponding to normal health and death), thus providing a metric proportional to the degree of health lost as a consequence of a given disease state (12) . Total DALYs within a given population offer substantial value as a means of measuring the overall disease burden experienced by that population.
Data from the GBD are subdivided into 21 global regions based on characteristics including geographical proximity and epidemiological similarity. The GBD study additionally reports the socio-demographic index (SDI) values for included regions at the state level, corresponding to a metric for average per-capita income, total fertility rates, and educational attainment (13)(14)(15)(16) . SDI values range from 0 -1, with larger values corresponding to higher levels of sociodemographic development such that nations are grouped into SDI-based development quintiles (high, high-middle, middle, low-middle, and low) (11) .
Age-standardized rate (ASR) values were established based upon the world standard population developed for the GBD study. GBD disease burden estimates were reported with 95% uncertainty intervals (UIs) that take into account differences and uncertainties in parameter estimation, model selection, data collection, and other factors.

Data analysis
The overall burden of cardiomyopathy was assessed by combining AC-and OC-related data. As no published UI estimates corresponding to cardiomyopathy were available, 95% con dence intervals were calculated based on standardized error values determined based upon the width of the 95% UI divided by 1.96×2.
Trends in ASRs correspond to shifts in disease-related patterns within particular populations, providing insight into potential underlying shifts in risk factor prevalence (17) . Trends in cardiomyopathy incidence, prevalence, and DALYs were evaluated based on percentage changes calculated based upon available estimates for the selected timepoints. An ASR was considered to be increased or decreased if the entirety of the 95% UI was above or below 0, respectively. If the 95% UI included 0, then the ASR was deemed stable. Estimated annual percentage changes served as a measure of ASR trends over particular time intervals, and were calculated as previously reported by Hankey et al (18) . R (v 3.4.4, R core team) was used for all statistical analyses, with P < 0.05 as the threshold of signi cance.

Cardiomyopathy prevalence in 2019
Globally, there were an estimated 0.7 million (95% UI: 0.5 -0.9) AC cases and 3.7 million (95% UI: 2.9 -4.7) OC cases in 2019. Respective age-standardized cardiomyopathy, AC, and OC prevalence rates (per 100,000 persons) in 2019 were 56.0 (95% CI: 43.8 -71.2), 8.5 (95% UI: 6.6 -11.0), and 47. 5   The proportion of cardiomyopathy cases attributed to OC and AC at the global and regional levels in 1990 and 2019 are summarized in Figure 2. Over half of cardiomyopathy cases at the global level were caused by OC, with these proportions having largely remained stable over time with the exception of in Eastern Europe, where a signi cant change in this proportion was observed when comparing these time points. The proportion of AC in Eastern Europe was outnumbered that of OC in both 1990 and 2019 (70.8%, 68% respectively). The gap between these two etiological forms of cardiomyopathy was also less in South Asia than in other surveyed regions.

Cardiomyopathy-Associated Mortality Rates in 2019
In total, the respective numbers of global deaths attributed to AC and OC were 0.07 million (95%UI: 0.06 -0.08) and 0.24 million (95%UI: 0.19 -0.26). The agestandardized mortality rate for cardiomyopathy in 2019 was 4 (95%CI: 3.3 -4), with respective mortality rates of 0.9 (95%UI: 0.7 -1) and 3.1 (95%UI: 2.6 -3.4) for AC and OC. When comparing different SDI groups, the highest mortality rates were observed in the high-middle SDI group in all categories, followed by the high SDI group, with only slight differences among the three other groups (   Table S3, Supplementary Figure S2).
Proportions of cardiomyopathy-related deaths attributed to AC and OC at the global and regional levels in 1990 and 2019 are summarized in Supplementary Figure S3. At the global level, over 50% of cardiomyopathy-related deaths were attributed to OC and these proportions were stable over time, but this trend was also totally adverse in Eastern Europe. The proportion of death caused by AC both exceeded that of OC in 1990 and 2019 (75.5%, 63.1% respectively). As above, the gap between these two etiologies was again found to be reduced in South Asia relative to other surveyed regions.

Cardiomyopathy-related DALY rates in 2019
At the global level in 2019, 2.4 million (95%UI: 2.0-2.8) DALYs were attributed to AC, while 5.7 million (95%UI: 4.9-6.3) DALYs attributed to OC. As shown in  (Table 3). At the national level, age-standardized DALY rates varied by 43.6-fold across countries, ranging from a low of 19. As a function of SDI quintile, the greatest drop in cardiomyopathy age-standardized prevalence rates was observed in the high-middle SDI group, followed by the high SDI group. No decreases were observed in the middle, low-middle, or low SDI groups with respect to AC. Similar decreases were observed in all SDI quintiles other than the low SDI group when assessing OC age-standardized prevalence rates (Table 1,2,3).
In line with the above data, age-standardized AC and OC mortality rates both declined by -0.4% (95%UI: -0.4 to -0.3), despite 24.8% (95%UI: 11.2-41.5) and 30.2% (95%UI: 23.9-52.3) increases, respectively, in the numbers of AC-and OC-related deaths. Patterns of changing age-standardized death rates across SDI quintiles were distinct from those observed for age-standardized prevalence rates ( Table 2,3). In addition, no clear patterns in age-standardized cardiomyopathy-related mortality rates were observed across the 21 surveyed regions of the world or as a function of SDI over the 1990-2019 period (Figure 3, S5, S6). For further details regarding time-dependent changes in AC and OC burden over the analyzed period, see Tables 1-3 and Tables S1-3.
Herein, we leveraged the GBD 2019 modeling framework to estimate the global burden of AC and OC, with data being strati ed by year, SDI, and geographic region. While one recent report discussed the prevalence rates, mortality rates, YLDs, and YLLs associated with both AC and OC in 2017 and national, regional, and global levels, it did not analyze associated DALY-related data. The present study is the rst to our knowledge to have systematically assessed the burden of AC and OC as of 2019 at the national, regional, and global levels. Through these analyses, we determined that there were approximately 0.7 million and 3.7 million worldwide AC and OC cases, respectively, in 2019. We further determined that there were 2.4 million and 5.7 million DALYs attributable to AC and OC, respectively, in 2019 at the global level. Together these results provide an epidemiological foundation that can guide public health efforts and policymakers as they seek to better understand, prevent, detect, and treat cardiomyopathy in a manner that better optimizes health system resource allocation.
We additionally evaluated differences in global AC and OC prevalence as a function of SDI and region. At all time points, AC prevalence were higher in regions with high SDI values irrespective of sex, while OC prevalence rates were higher in areas with low SDI values. The greatest decreases in overall cardiomyopathy prevalence were observed in high-middle SDI regions, while the largest increase in AD prevalence was observed in low-middle SDI regions, and no increasing trends were observed for OC. The primary components of SDI include gross fertility rates, per capita income levels, and years of education. As such, social factors can in uence cardiomyopathy prevalence rates and associated changes. One possible explanation for this may be that healthcare surveillance systems in high SDI regions are more comprehensive, while there is signi cant room for improving healthcare systems in low, low-middle, and middle SDI regions in order to improve the prevention and diagnosis of both AC and OC, particularly among individuals with asymptomatic disease.
We found that cardiomyopathy burden in 2019 varied substantially among geographic regions, with the highest age-standardized rates of overall cardiomyopathy and OC being evident in Eastern Sub-Saharan Africa, likely owing to the more haphazard medical conditions in many African nations and a consequent lack of surveillance and detection systems. In addition, rates of communicable diseases and peripartum cardiomyopathy are particularly high in Africa (19,20) . The highest age-standardized AC prevalence rates were observed in Eastern Europe, potentially due to high levels of alcohol intake and types of alcoholic beverage consumed in this region (21) . Alcoholism and associated issues have been reported to impose a major burden on public health in Eastern Europe (22) . While many nations in this region have enacted policies aimed at curbing alcohol intake resulting in lower rates of alcohol-associated mortality, rates of associated morbidity nonetheless remain relatively high (23) .
The analyses conducted herein revealed that Eastern Europe exhibited the highest age-standardized prevalence of AC as well as the highest age-standardized cardiomyopathy-related mortality rate. This emphasizes the importance of alcohol intake as a major risk factor associated with global disease burden that results in substantial health-related losses (24) . Higher frequencies of alcohol use likely coincide with higher AC prevalence. Given the importance of alcohol intake as a driver of AC development, further studies of national trends pertaining to alcohol consumption and associated AC surveillance are warranted to guide the design of targeted prevention strategies.
Age-standardized DALYs were used to quantify the overall burden of cardiomyopathy, which decreased over the study period. However, owing to the overall growth and aging of the global population, a larger number of DALYs were lost due to cardiomyopathy over time during the analyzed period. Moreover, higher age-standardized DALYs were observed in SDI regions exhibiting higher prevalence and mortality rates. At the regional level, DALYs associated with cardiomyopathy declined in most areas over the study period. In contrast, however, these DALYs rose Eastern Europe and Central Asia over this duration. This may be attributable to alcohol intake in these regions, to the effective treatment of complications such as heart failure, or to changes in rates of successful heart transplantation, all of which vary unequally as a function of socioeconomic status (25) .
With respect to time-dependent trends in disease burden observed from 1990 -2019, the overall global prevalence of AC and OC have signi cantly risen, as have overall mortality rates associated with these two forms of cardiomyopathy. Conversely, the age-standardized prevalence and mortality rates for these conditions have declined over this same interval. While the precise basis for these trends remains unclear, it may be in part due to the growth and aging of the global population (26) . We did not observe any clear patterns in age-standardized cardiomyopathy-related mortality rates across the 21 surveyed regions of the world over this 30-year period, suggesting that any such relationships that may exist are complex and non-linear. Consistent with this observation, the burden of AC and OC was not restricted to nations with particular levels of socioeconomic development, instead being highly prevalent in nations with a range of SDI values.

Limitations
While these GBD-based epidemiological estimates provide insight regarding the global burden of AC and OC that would otherwise be di cult or impossible to quantify, there are nonetheless certain limitations to this analysis. For one, data availability for certain countries was limited. While appropriate statistical efforts were employed to account for such data scarcity and associated uncertainty, the resultant data are primarily based upon trends in neighboring countries and/or covariates related to these diseases. Secondly, while the GBD study seeks to achieve maximal reliability and comparability with respect to the included data, there are inevitable instances of delayed/inaccurate reporting, variable data collection/source quality, misclassi cation, and coding deviations among countries. Lastly, the diagnostic criteria for speci c conditions can change over time and be re ective of regional coding trends, with varied nomenclature and classi cations for these diseases that can be vague or contradictory having been reported in the literature (27) .

Conclusions
In conclusion, the effective management of cardiomyopathy remains a pressing global public health challenge. The heterogeneous cardiomyopathy burden of cardiomyopathy observed across different regions of the world was primarily attributable to uneven socioeconomic development, suggesting that efforts to implement systematic surveillance systems should be implemented, particularly in middle-and low-income regions. Such monitoring would support. efforts to better manage this disease, as well as associated risk factors and complications therefore. Public health policymakers and other o cials should seek to develop regionally-appropriate efforts with the goal of counteracting and mitigating the future burden of cardiomyopathy. Declarations a. Ethics approval and consent to participate: All included patients gave their oral and written informed consent and we con rmed that this experiment was performed in accordance with relevant guidelines and regulations. The study was approved by the Ethics Committee (full name: Ethics Committee of The A liated Hospital of Nanjing University Medical School) (reference number:2021-190-01 ) to the department of ultrasound diagnosis, The A liated Hospital of Nanjing University Medical School, Nanjing, China.

b. Consent for publication
Not applicable c. Availability of data and materials: All data generated or analyzed during this study are included in this published article. The datasets generated during and/or analyses during the current study are available in the global burden of disease.

d. Competing interests
The authors declare that they have no competing interests.