Global burden of atrial fibrillation/atrial flutter and its attributable risk factors from 1990 to 2021

1What is already known?

2Previous studies have estimated the risk factors that influence the burden of multiple cardiovascular diseases. Among cardiovascular diseases, atrial fibrillation (AF) and atrial flutter (AFL) are often overlooked by the public because of their relatively low incidence, but they often lead to life-threatening and serious complications that affect public health.

3Among them, some studies related to the Global Burden of Disease (GBD) database on AF/AFL data suggest that the burden of AF/AFL is not limited to developed or less developed countries and a high burden of these diseases was seen in countries with various socio-demographic indices.

4Finally, a comparison of the years of life lost and years lived with disability in AF/AFL by gender leads to the conclusion that more emphasis should be placed on reducing mortality in female AF/AFL patients in the future.

5What does this study add?

6This study mainly focused on the latest data of AF/AFL belong to GBD and analysed multiple indicators from national, regional, and age-standard dimensions to identify potential risk factors including prevalence, deaths, years lived with disability-adjusted life years (DALYs) that influence the development of AF/AFL.

7We have refined the study of AF/AFL to each country in the world, based on the geographical information of the GBD database, and have obtained age-standardized incidence and age-standardized mortality rates for AF/AFL for 204 countries worldwide.

8In addition, we calculated DALYs, which is a more comprehensive indicator to assess disease burden, enabling a better evaluation of disease risk and quality of life of the population. By calculating them, we found that although age-standardized DALYs decreased over the study period, a larger number of DALYs were lost due to AF/AFL over time during the analysed period.

9What do the new findings imply?

10Accurate assessment of the burden of AF/AFL is essential for formulating effective preventative prevention and treatment programmes and optimizing health system resource allocation.

11The number of cases, deaths, and DALYs of AF/AFL showed an overall increasing trend and obvious geographical differences in the past three decades, despite the decreases in age-standardized rates. The burden of cardiomyopathy remains a persistent threat to global public health. These results provide an epidemiological foundation that can guide public health efforts and policymakers.

12Atrial fibrillation (AF) and atrial flutter (AFL) are prevalent tachyarrhythmias that can lead to severe complications, including heart failure and stroke, significantly impacting patients’ physical and mental health and quality of life. The aetiology of AF/AFL is complex, involving risk factors such as hypertension, smoking, alcohol consumption, high-sodium intake, obesity, and socio-demographic variables. Recent trends indicate a steep rise in AF-related hospitalizations, imposing heavy burdens on patients and the global healthcare system. Regular assessments of the current AF burden and its risk factors are crucial for developing effective prevention and control strategies.

13The global prevalence of AF has doubled from 1990 to 2019, reaching 59.7 million cases in 2019. This trend is expected to continue. In the USA, AF/AFL affects 3–5 million individuals and is projected to increase to over 8 million by 2050 due to population aging. In Europe, the prevalence is anticipated to surge from 8.8 million to 18 million by 2060.

14Economic globalization has led to disparities in morbidity and mortality among countries. In less developed regions, low cultural and educational levels contribute to a higher disease burden. However, in economically advanced countries with better healthcare resources and education, while AF/AFL incidence is higher, advanced diagnosis and treatment measures improve survival rates.

15A previous study evaluated the burden of AF from 1990 to 2019, utilizing data from the Global Burden of Disease (GBD) 2019 study. However, given the advancements in research and the emergence of new data, an updated assessment is paramount. Notably, since the Global Burden of Disease, Injury, and Risk Factor Study 2021 (GBD 2021) update, there has been a lack of a systematic study comprehensively updating the epidemiological trends in AF/AFL. Furthermore, while previous studies have examined the epidemiological aspects of AF in specific time periods within a particular country or region, there is currently no available research that utilizes the recently updated GBD database to analyse the trends of AF/AFL over the span of ∼30 years from 1990 to 2021, as well as the specific epidemiological situation of AF/AFL in 2021. In general, our research is conducted on a broader scale and over a longer duration, using the latest information. Therefore, leveraging the GBD 2021 data, this study aims to conduct a thorough and systematic analysis of the disease burden of AF/AFL. Specifically, it will evaluate the global, regional, and national trends in AF/AFL prevalence, incidence, death, disability-adjusted life years (DALYs), and risk factors. This analysis holds significant value, providing crucial epidemiological information that can inform policy decisions and interventions.

16The data for this study were sourced from the Global Burden of Diseases, Injuries, and Risk Factors (GBD) 2021 database, a refined version of the GBD 2019 iteration. This comprehensive data set incorporates 100 983 diverse sources to rigorously evaluate the disease burden of 371 diseases and injuries. Among these sources, vital registration systems, cancer registries, and verbal autopsy data play pivotal roles, guaranteeing the utmost accuracy and reliability of the data. Furthermore, to gain deeper insights into the underlying causes and determinants of AF and AFL, GBD 2021 incorporates population-based surveys, epidemiological studies, and published literature. Covering 204 countries across all five socio-demographic index (SDI) regions and 54 GBD regions, GBD 2021 employs a standardized and comparable approach to estimate population size, fertility rates, incidence, prevalence, mortality, and DALYs. Leveraging this robust data set, our study aimed to thoroughly investigate the global burden of AF and AFL. Specifically, we relied on the GBD 2021 database to accurately estimate the global, regional, and national incidence, prevalence, mortality, and DALYs attributable to AF and AFL from 1990 to 2021. The waiver of informed consent was approved by the University of Washington's Institutional Review Board due to the use of de-identified aggregated data.

17In the GBD study, we included all forms of AF and AFL, whether paroxysmal, persistent, or permanent/chronic.

18The case definition of input data of AF/AFL was offered by GBD 2021 Methods Appendices: (i) irregularly irregular RR intervals (in the absence of complete AV block); (ii) no distinct P waves on the surface ECG; and (iii) an atrial cycle length (when visible) that is typically variable and not exceeding 200 ms. The International Classification of Disease (ICD) codes used for inclusion of hospital and claims are I48I48.9 for ICD-10 and 427.3 for ICD-9 (https://www.healthdata.org/gbd/methods-appendices-2021/atrial-fibrillation-and-flutter).

19Years of life lost (YLLs) were calculated by multiplying standard life expectancy at death by the number of deaths. Years lived with disability (YLDs) were estimated as YLLs converted to health status, weighted by disability severity. The sum of YLLs and YLDs comprised DALYs.

20Risk factors for AF/AFL were identified based on causation evidence, exposure data availability, and modification potential., The GBD 2021 identified six risk factors: high body mass index (BMI), elevated systolic blood pressure, smoking, excessive alcohol consumption, high-sodium diet, and lead exposure. Their precise definitions and contribution to AF/AFL-related death are reported in previous studies.

21In this research, age-standardized rates (ASRs) were employed to quantify the occurrence, prevalence, death, and DALYs associated with AF and AFL. Additionally, we calculated the estimated annual percentage change (EAPC) in prevalence to evaluate the global burden of AF/AFL. Age-standardized rate was determined by adjusting the global age distribution to a standard population, which is crucial for comparing populations with varying age structures or analysing changes in the age structure of a single group over time. Standardization enables a comparison per 100 000 individuals, offering valuable insights into changes in disease patterns and risk factors within a community. Therefore, ASR serves as a tool for developing more focused and precise approaches for prevention and therapy.

22Uncertainty intervals (UIs) were computed based on 1000 draw-level estimates for each parameter, with 95% UI defined as the 25th and 975th values across all draws. Estimated annual percentage change serves as a concise metric for assessing trends in ASR over specific time periods. The regression line follows the natural logarithm of the ratio, y = α + βx + ɛ, where y = ln(ASR), x = calendar year, and ε represents the error. Estimated annual percentage change is reported with a 95% confidence interval (CI). An increase in ASR is indicated when EAPC estimates and their lower 95% CI exceed 0, whereas a decrease is suggested when EAPC estimates and their upper 95% CI are less than 0. Furthermore, to investigate the implications of EAPC, we analysed its correlation with ASR 2021 and HDI 2021 nationally.

23Mortality rates were estimated using population registry data coded with ICD or household surveys. We improved data comparability using statistical methods, including code reclassification, noise reduction, and Bayesian geospatial regression (CODEm), which leveraged age, space, and time to create smooth trends for 204 countries/territories. Disease incidence and prevalence were derived from diverse population-representative sources, including cohort/registry reports, surveys, registry/cohort microdata, and health system data. Consistent estimates were generated using DisMod-MR and MR-BRT, which were adjusted for methodological and definitional differences.

24The GBD aims to estimate all quantities for each location, even in data scarcity. Best estimates and uncertainties are provided, motivating the search for better global data and statistical methods. This commitment to best estimates distinguishes the GBD from other efforts. Emphasizing comparability across time and geography is crucial for policymakers to understand disease burdens and trends. To ensure comparability, the entire historical time series is recalculated with each GBD release, preventing spurious comparisons. The principles of data inclusion and use in the GBD database ensure data accuracy.

25The estimation of AF/AFL-related disease burden involved several key steps. Firstly, we described the disease burden of AF/AFL globally and in several subtypes, including sexes, age groups, SDI regions, GBD regions, and countries based on the available data. Secondly, we determined the changing trend of AF/AFL by using the linear regression model. To gain further insights into the evolving patterns across diverse GBD regions, we employed a hierarchical cluster analysis based on the EAPC values. Finally, we delved into the risk factors that contribute to the disease burden of AF/AFL.

26All statistical analyses were conducted using R software, version 4.3.1, and a two-sided P-value of less than 0.05 was deemed statistically significant.

27The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). All information from the GBD programme is available and free for the public, so the agreement of the medical ethics committee board was not necessary.

28In 2021, AF/AFL affected 52.55 million individuals worldwide, a 137% increase from 1990. Incidence cases rose 124.0%, yet the age-standardized incidence rate (ASIR) decreased by 7% from 1990 to 2019, with the most substantial decreases in high–middle SDI countries (Table 1).

29Atrial fibrillation/AFL-related deaths totalled 0.34 million in 2021, a 196.0% increase since 1990, with the greatest increases in low–middle SDI countries. Atrial fibrillation/AFL contributed to 8.36 million DALYs in 2021, a 149.0% increase from 1990. Despite this, the age-standardized DALYs rate remained stable, with the most significant decreases in high–middle SDI countries (Table 4).

30Furthermore, our study revealed a consistent trend from 1990 to 2021: the ASR of incidence, prevalence, deaths, and DALYs associated with AF/AFL all displayed a gradual upward trajectory (Figure 1).

31Age groups 70–79 exhibited the highest AF/AFL prevalence, while 65–74 showed higher incidence rates (see Supplementary material online, Figure S1; Tables 1 and 2). Unexpectedly, these groups had lower death rates than those over 80. The 80–89 age groups had higher death numbers (see Supplementary material online, Figure S1; Table 3), while the 80–84 age group exhibited the highest DALYs (see Supplementary material online, Figure S1; Table 4). The ASRs for AF/AFL prevalence, incidence, death, and DALYs were generally consistent, except for the >95 age group, which had the highest ASRs for death and DALYs, as well as higher ASRs for prevalence and incidence (see Supplementary material online, Figure S1; Tables 1–4).

32From 1990 to 2021, the incidence, prevalence, deaths, and DALYs for AF/AFL increased for both sexes. Males had a slightly higher disease burden than females, with their prevalence and incidence exceeding women post-2000. However, females consistently had higher death and DALY counts. Age-standardized rate analysis showed male ASRs for prevalence, incidence, and DALYs surpassed females’ throughout the period. Conversely, before 2005, female ASR for death exceeded males’, but since 2005, male ASR for death has increased faster (see Supplementary material online, Figure S2).

33Age-standardized rates and case counts for AF/AFL incidence, DALYs, prevalence, and death were higher in upper SDI quintiles from 1990 to 2021. Low–middle SDI regions had higher ASRs for incidence than high–middle and middle SDI regions. High SDI quintile consistently had the highest values. Trends in high–middle and middle SDI regions were similar (see Supplementary material online, Figure S3; Tables 1–4). Cluster analysis revealed distinct patterns, with Southern Latin America showing a significant increase. Thirty-eight regions, including World Bank High Income, fell into the minor increase category. Fourteen regions, such as World Bank Lower Middle Income, remained stable or had minor decreases. High-income Asia Pacific showed a significant decrease (see Supplementary material online, Figure S4).

34From 1990 to 2021, the United Arab Emirates (971% increase), Qatar (907%), and Jordan (971%) had the highest increases in AF/AFL incidence cases. These three countries also showed the greatest prevalence increases, with the UAE (942%), Qatar (908%), and Jordan (557%). For mortality, Kuwait (606%), Maldives (590%), and Bhutan (571%) had the highest death increases. In terms of DALYs, the UAE (660%), Qatar (557%), and Kuwait (460%) showed the most substantial increases. Romania (−6%), Georgia (−4%), and Niue (−2%) had the largest decreases in incidence cases. Niue uniquely showed decreasing trends in prevalence (−1%), mortality (−24%), and DALYs (−12%) (Figure 2). Using the EAPC, Austria (2.21), Czechia (1.50), and Israel (1.25) showed the highest incidence growth rates, while Argentina (−1.31), Finland (−1.16), and Romania (−1.21) had the largest decreases. For prevalence, Austria (2.33), Israel (1.46), and Czechia (1.44) led growth, and Romania (−1.17), Finland (−1.12), and Argentina (−1.03) showed the greatest reductions. In mortality, the United Arab Emirates (2.97), Sweden (2.57), and Lesotho (2.28) had the highest growth; Guam (−3.00), Qatar (−2.71), and Saint Lucia (−2) decreased the most. For DALYs, the United Arab Emirates (1.96), Sweden (1.61), and Austria (1.46) grew the most, while Cyprus (−1.71), Qatar (−1.66), and Finland (−1.47) decreased most (Figure 3; Supplementary material online, Tables S1–S4).

35Worldwide, the highest incidence and prevalence of AF were observed in the 70–74 age group for both sexes. However, when considering age-specific incidence rates, both sexes exhibited peak values in the 95+ age group (see Supplementary material online, Figure S5; Tables 1 and 2). Mortality due to AF peaked in the 75–79 age group for both sexes. In contrast, the global number of DALYs peaked in the 85–89 age group for males and the 80–84 age group for females. Nonetheless, the age-specific rates of mortality and DALYs for both sexes still peaked in the 95+ age group (see Supplementary material online, Figure S5; Tables 3 and 4). Notably, the global number of deaths and DALYs attributed to AF was lower among males compared with females across all age groups. However, other indicators, such as those depicted in Supplementary material online, Figure S6 and Tables 1–4, exhibited greater values in males compared with females across the entire age range.

36In 2021, the high SDI region had higher incidence, prevalence, mortality, and DALYs than other regions, regardless of ASR or absolute cases. As SDI decreased, ASRs declined gradually. However, middle SDI countries had higher absolute case numbers than high–middle SDI countries (see Supplementary material online, Figure S7; Tables 1–4). Across 204 countries, a positive correlation was found between SDI and ASRs of incidence, prevalence, mortality, and DALYs. Some countries showed significantly higher ASRs than expected based on their SDI values (see Supplementary material online, Figure S8).

37Among 54 GBD regions, Asia, Basic Health System, and Advanced Health System had the highest incidence cases, with similar trends for prevalence, mortality, and DALYs (see Supplementary material online, Figure S9; Tables 1–4). However, ASRs varied. High-income North America, North America, and the Region of the Americas had the highest incidence ASRs. Australasia, the Region of the Americas, and America led in prevalence ASRs. Australasia, Western Europe, and Commonwealth high-income regions had the highest death ASRs. Andean Latin America, Central Sub-Saharan Africa, and World Bank Lower Middle Income regions showed the highest DALYs ASRs (Figure S9; Tables 1–4).

38Globally, China, India, and the USA have the highest incidence, prevalence, and DALYs. However, for mortality, China, India, and Germany have the most deaths. In terms of ASRs, Sweden, Austria, and Germany lead in incidence and prevalence, while Montenegro, Nauru, and Sweden have the highest death and DALYs ASRs (Figure 4; Supplementary material online, Tables S1–S4).

39The EAPC was significantly associated with the ASRs of incidence, prevalence, and deaths, but not with the ASR of DALYs. Specifically, when the ASRs surpassed 75/100 000 for incidence, 800/100 000 for prevalence, and 5/100 000 for deaths, a positive correlation with EAPC was observed. However, the EAPC of deaths and DALYs in middle HDI regions was higher, forming an inverted U-shape pattern. Conversely, the EAPC for incidence and prevalence showed a slight downward trend with increasing HDI. Overall, higher HDI countries tend to experience a decline in AF/AFL-related ASRs (see Supplementary material online, Figure S10).

40The GBD 2021 classifies risk factors into three groups: environmental/occupational, behavioural, and metabolic. Among them, metabolic risks contribute the most to attributable risks for ASRs, deaths, and DALYs in 1990 and 2021, followed by behavioural risks and then environmental/occupational risks (see Supplementary material online, Figure S11). Subgroup analysis shows that high systolic blood pressure is the leading risk factor for ASR, DALYs, and deaths in both years. In regions with higher SDI, high BMI is the second most prevalent risk factor after high systolic blood pressure (see Supplementary material online, Figure S12 and Table S5). In lower SDI regions, other environmental risks rank second after high systolic blood pressure. Among subgroups, smoking had the most significant impact on deaths and age-standardized death rates (ASDRs) in 2021, accounting for 35% and 34.4%, respectively. The disparity with high-sodium diet and lead exposure was minor. Similarly, smoking contributed most to DALYs and age-standardized DALYs rates, standing out at 43.9% and 43.3%. In high–middle and middle SDI regions, a high-sodium diet contributed marginally more to AF/AFL deaths than smoking in 2021 (see Supplementary material online, Figure S13).

41This study offers a comprehensive epidemiological analysis of AF/AFL's global incidence, prevalence, mortality, and DALYs. Our results show AF/AFL is a major global health issue, with 1.37-fold prevalence and 1.24-fold incidence increases over 31 years. In 2021, there were 4.48 million new cases, 52.55 million prevalent cases, 338 947 deaths, and 8 358 894 DALYs. Despite declining ASRs, absolute cases doubled due to population growth, aging, improved diagnostics, and increased awareness. Notably, DALYs and deaths are increasing faster than prevalent and incident cases, and the age-standardized mortality rate is the only upward-trending indicator. This underscores the need for early detection and urgent development of more effective AF/AFL management strategies.

42Our findings align with the GBD 2019 study, indicating a higher burden of DALYs and deaths due to AF/AFL among females compared with males in 2021. Conversely, the incidence and prevalence trends differ. Women experience higher morbidity and complications related to AF/AFL, likely due to underutilization of rhythm control strategies and less oral anticoagulant treatment., This contributes to poor prognosis in the female population. An analysis of AF-related mortality in Europe from 2008 to 2019 revealed a higher total number of deaths in women despite a greater increase in age-standardized mortality for men. This suggests potential inequalities in medical care or biological differences between genders. Hence, prioritizing rhythm control techniques and complication management for female AF/AFL patients is crucial.

43The long-standing Framingham Heart Study emphasizes aging as the paramount risk for AF, surpassing other factors. Aging is linked to oxidative stress, mitochondrial dysfunction, cardiomyocyte hypertrophy, and ion channel inactivation, fostering atrial structural and electrical remodelling, heightening AF risk., Our study reveals a steep rise in AF/AFL incidence, prevalence, deaths, and DALYs among the elderly, with an increasing trend in ASDR and DALYs. This underscores the urgency for improved AF/AFL management in the elderly.

44Analogous to the GBD 2019 analysis of AF/AFL, considerable variations in the ASRs and case numbers of incidence, prevalence, deaths, and DALYs were observed across different SDI quintiles. The SDI, comprising years of education, per capita income, and total fertility rate, significantly impacts these metrics and their temporal changes in AF/AFL. Consequently, each region must devise tailored solutions grounded in its unique socio-economic context. Regions with high SDI exhibited higher ASIRs and case numbers, whereas low SDI regions had lower rates and case numbers. This suggests that a higher percentage of individuals in low SDI locations may have undetected AF/AFL. However, in high SDI regions, a greater number of AF/AFL patients, particularly those with paroxysmal or asymptomatic AF/AFL, are being identified at an early stage with high accuracy. The EAPC values for age-standardized incidence, death, and DALY rates in low SDI areas are positive, indicating an increasing trend during the study period. Conversely, high SDI areas showed a decreasing trend, despite having higher incidence and death cases. This suggests that the overall medical treatment effect for AF/AFL is superior in high SDI regions. In general, the burden of AF/AFL was higher in high SDI regions than in low SDI regions, which is similar to the findings of a study examining the impact of socio-economic factors on trends in AF incidence and mortality in Europe from 1990 to 2017. The observed phenomenon might stem from disparities in the application of therapeutic strategies, encompassing the utilization of novel oral anticoagulants, left atrial appendage closure procedures, and catheter-based or cryogenic balloon ablation techniques. These interventions are not uniformly accessible across diverse socio-economic and geographical landscapes. While significant progress has been made in addressing the AF/AFL burden in high SDI areas in recent years, there has been a notable oversight in managing AF/AFL in low SDI areas. We speculate that because of the better medical conditions in high SDI areas, AF/AFL patients can survive long enough and may develop more comorbidities or worsen their AF/AFL condition. Hence, there is a pressing need to bolster efforts towards the prevention, diagnosis, and treatment of AF/AFL in low SDI regions in the future.

45It is worth noting that in some studies, because the death data used come from relevant publicly accessible World Health Organization databases, there are certain differences from our research. For example, in the above-mentioned analysis study on the trend of AF-related mortality in Europe from 2008 to 2019, the study believed that any European region has a certain increase in the age-standardized mortality rate of AF and the main increase is concentrated in Eastern European countries. Our study believes that the occurrence and development of national age-standardized mortality rates in the European region are inconsistent and the increase mainly occurs in economically developed areas in Western Europe. The reason for this difference is that the death data in the World Health Organization database and the GBD database have different sources and different adjustment methods.

46The current study reveals substantial geographic heterogeneity in AF burden. High-income North America showed the highest ASIR, while Australasia led in ASDR and ASPR. Across 204 countries, the ASIR, ASDR, and age-standardized DALY rate positively correlated with SDI, indicating a heavier AF burden in higher SDI countries. This aligns with prior research. Country-level variations in AF burden may be due to disparities in aetiology, detection, ethnicity, health resources, and treatment. Better medical treatment and public awareness in high SDI regions likely contribute to increased AF identification and reporting. However, data collection limitations in regions without robust medical systems may underestimate the overall AF burden. Among AF patients, a significant subpopulation comprises asymptomatic individuals. Research suggests that ∼15–50% of clinical AF patients exhibit asymptomatic manifestations, varying with the study population, AF type (paroxysmal, persistent, or permanent), and detection method., Notably, fewer asymptomatic AF patients are women, likely due to women's tendency to promptly seek medical attention for elevated heart rates. While rhythm control strategies, such as antiarrhythmic drugs, catheter ablation, and electrical cardioversion, are predominantly used in symptomatic AF patients, their utilization in asymptomatic patients is less frequent. However, asymptomatic AF patients often face a higher stroke risk due to inadequate evaluation of anticoagulant use. This has sparked controversy regarding the relationship between asymptomatic AF and prognosis, as previously, asymptomatic AF was considered benign, with symptom control being the primary treatment goal. Recent studies, including Boriani et al.'s work, have demonstrated that asymptomatic AF patients exhibit a two-fold higher 1-year mortality compared with symptomatic patients. Additionally, a Chinese cross-sectional study found asymptomatic AF patients to have a higher incidence of clinical events leading to hospitalization. Further clinical investigations are imperative to explore the characteristics of this subpopulation and devise effective management strategies.

47Literature reviews suggest that managing modifiable factors like hypertension, obesity, smoking, and excessive alcohol consumption can prevent AF/AFL and improve prognosis., This study analysed modifiable risk factors contributing to AF/AFL cases and rates, including high systolic blood pressure, BMI, alcohol, smoking, dietary risks (high sodium), and environmental risks (lead exposure). High systolic blood pressure is the primary risk factor, with a 10 mmHg increase elevating AF risk by 19%. Effective blood pressure control significantly reduces AF/AFL incidence., The second most significant risk factor is a high BMI, with each unit increase associated with a 4–5% rise in AF risk. This aligns with the Western diet of high calories, fat, and sugar. Smoking's attributable share in AF/AFL is declining due to tobacco control and health promotion. Lead exposure, although rarely linked to AF directly, contributes to hypertension and cardiovascular disease, both risk factors for AF/AFL. Managing risk factors like smoking cessation, alcohol abstinence, low-salt/fat diet, exercise, and healthy lifestyle practices can reduce AF/AFL incidence. However, tailored prevention strategies are needed considering the varying risk factors across different socio-economic levels.

48In assessing AF risk factors, the impact of the COVID-19 pandemic, which has had worldwide prevalence in recent years, must be considered. A clinical study found a significant 14.8% occurrence of AF in COVID-19 patients, accompanied by a heightened risk of mortality, ischaemic stroke, and cardiac arrest. This indicates that not only is AF more prevalent in COVID-19 patients but it also independently correlates with a higher mortality rate after adjusting for confounders. A survey conducted by the Italian Association of Arrhythmology and Cardiac Pacing revealed that during the COVID-19 pandemic in Italy, radiofrequency ablation surgeries for arrhythmias decreased by 70% during the first wave (March to May 2020), with incomplete recovery in subsequent periods compared with other surgeries. This suggests that inadequate AF evaluation and treatment during the pandemic contributed to the global AF mortality rate during this time. A nationwide US study of patients under remote monitoring during the early COVID-19 pandemic (21 January to 29 April 2020) found a significant association between COVID-19 outbreaks and AF prevalence. States with higher COVID-19 infection rates had a more significant rise in AF, likely due to social unrest affecting medication adherence, dietary and behavioural patterns, and negative emotions. Meta-analysis revealed common risk factors between AF and COVID-19, suggesting a close link between AF and infectious diseases due to oxidative stress, inflammatory mediators, and haemodynamic disorders. Notably, even recovered COVID-19 patients showed a higher incidence of AF compared with uninfected patients during the same follow-up period, emphasizing the need for suspicion of AF among recovered COVID-19 patients with relevant symptoms.

49Beyond COVID-19 and lung inflammation, numerous other risk factors, including cardiomyopathy, obstructive sleep apnoea syndrome, coronary artery disease, and lung disease, contribute to the incidence and mortality of AF, yet are not fully captured in the GBD database. Cardiomyopathy, particularly hypertrophic cardiomyopathy, is a prominent example. This condition often leads to AF due to increased left atrial pressure and atrial myopathy resulting from unexplained left ventricular hypertrophy and outflow tract obstruction. The prevalence of AF in hypertrophic cardiomyopathy patients, when age-matched, is significantly higher (20–30%) with an annual incidence of 2–4%, far exceeding the general population rates. Moreover, patients with hypertrophic cardiomyopathy have a lower tolerance to AF, and those with AF face a notably higher risk of embolic events, aberrant ICD discharges, and all-cause mortality. The interplay between coronary heart disease (CHD) and AF is intricate, as the two conditions can coexist. Atrial fibrillation can exacerbate inflammatory responses, hastening atherosclerosis, while CHD heightens the risk of AF by enhancing focal ectopic activity in the left atrium and altering micro-re-entry patterns. Prior studies reveal that CHD prevalence in AF patients ranges from 18% to 46.5%, acting as an independent risk factor for AF development. Moreover, AF frequently coincides with non-ST-segment elevation myocardial infarction (NSTEMI), with AF patients exhibiting a 39% higher risk of NSTEMI and double the likelihood of cardiovascular events or mortality compared with non-AF patients.,

50Global treatment strategies have shifted, and AF mortality and rehospitalization are increasingly influenced by non-cardiovascular causes. Obstructive sleep apnoea syndrome contributes to higher recurrence of AF post-treatment and reduces radiofrequency ablation success, impacting mortality. A study of over 4000 AF patients in five Asian countries revealed that despite good anticoagulant use, most deaths were due to non-cardiovascular diseases (72.3%). Multivariate analysis identified obstructive pulmonary disease as an independent predictor of mortality and the primary cause of re-admission. Similarly, data from the German Atrial Fibrillation Network highlight chronic obstructive lung disease as a contributor to higher mortality rates among AF patients.

51Heart failure, a terminal cardiovascular disease outcome, profoundly influences AF epidemiology. A Danish study showed that heart failure preceding AF led to a higher mortality risk than the reverse or concurrent conditions, likely due to the irreversibility of pre-existing chronic disease. A US study spanning 1999–2020 revealed a 286.4% increase in age-adjusted heart failure mortality among AF patients, with men and younger individuals (<65 years) experiencing a steeper rise. This trend is attributed to the rising prevalence of comorbidities like diabetes. Geographically, Whites and those in the Midwestern USA have higher heart failure-related mortality rates. The peak in heart failure mortality among AF patients between 2011 and 2012 is likely due to improved AF awareness and diagnostic accuracy following US guidelines issued in 2011. Anticoagulant choice in AF patients with heart failure affects mortality, as heart failure heightens anticoagulation risks and AF mortality, even with new oral anticoagulants.

52While the current findings are significant, several limitations must be acknowledged. Firstly, data bias from confounding factors is a concern, particularly in countries with incomplete death certification systems, where verbal autopsy data are often utilized despite its limitations. Additionally, despite our efforts to introduce a risk-confounding star rating system in the GBD assessment, uncertainties in measurements may still be overlooked or misunderstood. The diverse data sources with potential biases and missing data for specific locations and years further complicate the GBD analysis. In some cases, the data patterns or relative risk functions may differ from expert consensus. The GBD follows a rule-based evidence synthesis approach, which can lead to discrepancies with other assessments that rely more on expert opinion. Finally, there are two main points of the limitations for the linear regression model including under-fitting problems and sensitive to outliers.

53Furthermore, the GBD studies have yet to incorporate other risk factors, such as diabetes, physical inactivity, and hyperthyroidism, into their analyses. Lastly, it is noteworthy that there exist different types of AF/AFL, including paroxysmal, persistent, and permanent AF/AFL, yet the GBD database does not recognize this complexity and merely identifies whether a population falls under the AF/AFL classification.

54Our study highlights the varying patterns of AF/AFL prevalence from 1990 to 2021 across gender, SDI quintiles, and regions. Despite progress in prevention and treatment in higher SDI areas, AF/AFL prevalence is rising in lower SDI regions. Given its preventable and curable nature, urgent adoption of cost-effective techniques targeting modifiable risk factors is crucial in high- or increasing-prevalence areas.