Burden and age-specific trends of atrial fibrillation/atrial flutter from 1990 to 2023: a growing challenge among younger and middle-aged adults

Abstract

Aims

The burden of atrial fibrillation and atrial flutter (AF/AFL) has increased, but age-specific patterns across World Bank income levels (WBILs) remain unclear.

Methods and results

Using the Global Burden of Disease 2023 estimates, we assessed age- and WBIL-stratified trends in prevalence, incidence, mortality, and disability-adjusted life years (DALYs) for AF/AFL from 1990 to 2023, employing age–period–cohort analysis and joinpoint regression. Mortality attributable to six modifiable risk factors was quantified based on comparative risk assessment estimates. Projections of AF/AFL burden for 2024–48 were generated using Bayesian age–period–cohort models. In 2023, AF/AFL affected 58.99 million prevalent cases, 5.02 million incident cases, 376 862 deaths, and 9.26 million DALYs, with the absolute burden concentrated in adults aged ≥ 65 years. From 1990 to 2023, age-standardized prevalence (ASPR) and incidence rate (ASIR) increased, while mortality rate (ASMR) and disability-adjusted life year rate (ASDR) remained stable. High-income countries showed increases across all metrics, upper-middle-income countries had rising ASPR/ASIR and decreasing ASMR/ASDR, and lower-middle- and low-income countries showed consistent increases across all metrics. Among younger (30–44 years) and middle-aged (45–64 years) adults, all metrics increased, while in older adults, only ASPR rose. High systolic blood pressure was the leading attributable risk factor, with larger contributions from high body mass index, smoking, and alcohol use in younger and middle-aged adults. Projections indicated modest declines in ASPR, stable ASIR, and increasing ASMR/ASDR through 2048.

Conclusion

Despite the concentrated burden in older adults, mortality- and disability-related burden is rising in younger and middle-aged populations, with significant variation across WBILs.

Graphical Abstract

Graphical Abstract.

What’s new?

• From 1990 to 2023, ASPR, ASIR, ASMR, and ASDR of AF/AFL increased in younger (30–44 years) and middle-aged (45–64 years) individuals, while only ASPR increased in older individuals (≥65 years).

• High SBP remained the leading attributable risk factor for AF/AFL-related ASMR and ASDR across all age groups, while high BMI, smoking, and alcohol use contributed more prominently in younger and middle-aged individuals than in older individuals.

• Bayesian age–period–cohort projections revealed declining ASPR and stable ASIR, but increasing ASMR and ASDR through 2048, with rising ASMR and ASDR in younger and middle-aged populations despite declining ASPR and ASIR.

Introduction

Atrial fibrillation and atrial flutter (AF/AFL) are the most prevalent sustained arrhythmias in adults, significantly increasing the risk of stroke, heart failure, and sudden cardiac death.^1,2 Over the past three decades, the burden of AF/AFL has increased across regions and economic settings, driven largely by population aging and the rising prevalence of major cardiometabolic risk factors, including hypertension, diabetes mellitus, and obesity.^3,4 Beyond impairing patients’ quality of life and long-term prognosis, AF/AFL impose considerable economic pressure on healthcare systems worldwide, mainly related to inpatient hospital costs.⁵ In the USA, the incremental annual cost per patient has been estimated at approximately $5218 for hospitalizations and $3596 for outpatient care, translating into national expenditures of $6–26 billion annually and underscoring the need for more effective prevention and management strategies.⁶

Although AF/AFL have traditionally been considered age-related arrhythmias, accumulating evidence indicates an increasing incidence and prevalence in younger populations.⁷ This pattern has been associated with adverse lifestyle and metabolic exposures, including unhealthy dietary patterns, physical inactivity, and excessive alcohol consumption.^8–11 Nevertheless, existing studies continue to focus predominantly on older populations, and the burden, temporal trends, and implications of AF/AFL among younger adults remain insufficiently characterized, particularly in direct comparison with older cohorts. Moreover, despite well-recognized geographic variation in AF/AFL burden, age-specific patterns across socioeconomic contexts remain incompletely understood.¹² Clarifying age- and income-related differences is essential for informing targeted prevention strategies and health-system planning.

Using the most recent data from the Global Burden of Disease (GBD) 2023 study, we aimed to characterize age-specific patterns and temporal trends in AF/AFL burden from 1990 to 2023, with particular emphasis on differences between younger and older populations across World Bank income levels (WBILs). Beyond extending existing GBD analyses to a more recent period and longer projection horizon, this study explicitly focuses on age-related divergence in mortality and disability burden, assessing whether changes in AF/AFL burden reflect shifts in disease incidence or in clinical impact across the life course. By prioritizing income-stratified analyses rather than relying on aggregated estimates, we assessed trends in prevalence, incidence, mortality, and disability-adjusted life years (DALYs), examined the distribution of attributable risk factors, and projected the future burden of AF/AFL through 2048. This age- and income-focused analytical framework is intended to support prevention prioritization, health-system planning, and long-term policy development across diverse economic settings.

Methods

Data sources

The GBD 2023 study provides standardized descriptive epidemiological estimates for 369 diseases and injuries across 204 countries and territories from 1990 to 2023. Each death is assigned to a single underlying cause according to the GBD cause-of-death hierarchy. Age-specific estimates of prevalence, incidence, mortality, and DALYs for AF/AFL were retrieved from the Global Health Data Exchange (GHDx) Results Tool (http://ghdx.healthdata.org/gbd-results-tool).¹³ Age-standardized rates (ASRs) were calculated using the GBD world standard population, and all estimates are reported with corresponding 95% uncertainty intervals (UIs).

Analyses were primarily stratified by WBIL, including high-income, upper-middle-income, lower-middle-income, and low-income settings, to characterize age-specific patterns of AF/AFL burden across socioeconomic contexts (see Supplementary material online, Table S1). This classification is based on gross national income per capita and is commonly used to reflect differences in economic development and health-system capacity. All analyses were conducted using de-identified and aggregated data; therefore, ethical approval and informed consent were waived by the University of Washington Institutional Review Board.

Case definition and outcome measures

AF/AFL cases were identified using International Classification of Diseases (ICD) codes 427.3–427.32 in ICD-9 and I48–I48.92 in ICD-10. Individuals were categorized into three age groups, younger (30–44 years), middle-aged (45–64 years), and older (≥ 65 years), to facilitate age-stratified comparisons and align with prior GBD-based epidemiological analyses.

Years of life lost (YLLs) were calculated by multiplying the number of AF/AFL-related deaths at each age by the corresponding standard life expectancy. Years lived with disability (YLDs) were estimated by multiplying AF/AFL prevalence by disease-specific disability weights used in the GBD framework. DALYs were calculated as the sum of YLLs and YLDs. Age-standardized prevalence rate (ASPR), incidence rate (ASIR), mortality rate (ASMR), and DALY rate (ASDR) were derived by applying the GBD standard population weights to age-specific estimates.

Attributable risk factors

Risk factors attributable to AF/AFL-related mortality were evaluated using the GBD comparative risk assessment framework. Risk factors were included based on the following criteria: (i) evidence supporting a causal association with AF/AFL; (ii) availability of exposure estimates across countries and time periods; and (iii) potential modifiability.

In GBD 2023, six risk factors were identified as contributors to AF/AFL-related deaths: high systolic blood pressure (SBP), high body mass index (BMI), smoking, alcohol use, diet high in sodium, and lead exposure. Detailed definitions and estimation methods have been reported previously.¹⁴

Statistical analysis

AF/AFL burden was quantified using ASPR, ASIR, ASMR, and ASDR across WBIL strata and age groups. The proportions of AF/AFL-related mortality and DALYs attributable to the six risk factors were estimated within each income and age category. Temporal trends from 1990 to 2023 were assessed using two complementary approaches: log-linear regression to evaluate the estimated annual percentage change (EAPC) for overall trends and joinpoint regression to estimate the annual percentage change (APC) for each segment and the average annual percentage change (AAPC) over the entire period. Trends were classified as increasing or decreasing when APC or AAPC estimates were statistically significant (P < 0.05); otherwise, trends were considered stable. Age–period–cohort analyses were performed to evaluate the independent effects of age, period, and birth cohort on AF/AFL prevalence, incidence, mortality, and DALYs. Future trajectories of ASPR, ASIR, ASMR, and ASDR from 2024 to 2048 were projected using Bayesian age–period–cohort (BAPC) models stratified by age group.

All statistical analyses were performed using R software (Version 4.4.1), and figures were generated using GraphPad Prism (Version 9.0).

Results

Overall burden of AF/AFL in 2023

In 2023, AF/AFL accounted for 5 8988 598 (95% UI: 46 445 673–72 696 713) prevalent cases, 5 017 347 (95% UI: 3 980 233–6 339 119) incident cases, and 376 862 (95% UI: 318 449–423 293) deaths. Total DALYs were 9 256 286 (95% UI: 7 510 087–11 650 849) (see Supplementary material online, Table S2). When stratified by age, the absolute burden of AF/AFL was concentrated in older individuals. Younger individuals contributed 1 047 176 (95% UI: 552 308–1 723 338) prevalent cases and 225 854 (95% UI: 127 815–353 998) incident cases, with 1075 (95% UI: 837–1376) deaths and 140 743 (95% UI: 92 338–204 651) DALYs (see Supplementary material online, Table S3). Middle-aged individuals accounted for 13 774 562 (95% UI: 9 221 511–19 317 091) prevalent cases and 1 497 911 (95% UI: 899 564–2 264 512) incident cases, along with 17 759 (95% UI: 15 813–20 097) deaths and 1 683 813 (95% UI: 1 233 191–2 334 002) DALYs (see Supplementary material online, Table S4). Older individuals accounted for the majority of fatal and disabling outcomes, with 358 027 (95% UI: 297 234–406 945) deaths and 7 431 730 (95% UI: 5 880 175–9 264 020) DALYs (see Supplementary material online, Table S5).

Income- and age-specific temporal trends in AF/AFL burden

Temporal trends in age-standardized AF/AFL burden varied across WBILs from 1990 to 2023. In the overall population, ASPR and ASIR increased, with EAPCs of 0.11 (95% CI: 0.08–0.14) and 0.07 (95% CI: 0.04–0.11), respectively. In contrast, ASMR and ASDR were stable, with EAPCs of 0.06 (0.03–0.09) for both metrics (Table 1). Income-stratified analyses indicated that all four age-standardized metrics increased in high-income countries (HICs). In upper-middle-income countries (UMICs), ASPR and ASIR increased, whereas ASMR and ASDR decreased. In lower-middle-income countries (LMICs) and low-income countries (LICs), ASPR, ASIR, ASMR, and ASDR consistently increased over time (Figure 1; Table 1).

Table 1.

Age-standardized AF/AFL rates in 1990 and 2023 and EAPC from 1990 to 2023 for all-age populations by WBIL

WBIL	Prevalence (95% UI)			Incidence (95% UI)			Deaths (95% UI)			DALYs (95% UI)
	ASR per 100 000 (1990)	ASR per 100 000 (2023)	EAPC 1990–2023 (95% CI)	ASR per 100 000 (1990)	ASR per 100 000 (2023)	EAPC 1990–2023 (95% CI)	ASR per 100 000 (1990)	ASR per 100 000 (2023)	EAPC 1990–2023 (95% CI)	ASR per 100 000 (1990)	ASR per 100 000 (2023)	EAPC 1990–2023 (95% CI)
Total	644.3 (490.9–806.37)	649.36 (510.92–796.34)	0.11 (0.08–0.14)	54.46 (42.67–69.4)	54.75 (43.49–68.8)	0.07 (0.04–0.11)	4.17 (3.67–4.5)	4.4 (3.7–4.95)	0.07 (0–0.13)	101.61 (82.81–126.59)	103.83 (84.63–130.25)	0.06 (0.03–0.09)
High income	783.16 (601.88–976.69)	813.13 (667.38–986.5)	0.17 (0.15–0.2)	67.08 (53.05–85.46)	69.22 (57.34–84.69)	0.13 (0.11–0.16)	4.59 (4.06–4.91)	5.1 (4.24–5.58)	0.29 (0.25–0.33)	118.26 (95.36–146.85)	125.11 (102.32–154.62)	0.18 (0.16–0.20)
Upper middle income	537.31 (410.04–673.87)	583.21 (445.5–733.92)	0.45 (0.36–0.54)	45.77 (35.45–57.78)	49.47 (38.34–62.4)	0.38 (0.3–0.47)	4.33 (3.62–4.92)	3.62 (3.02–4.06)	−0.85 (−0.98 to −0.72)	95.13 (79.6–117.14)	88.88 (71.31–113.57)	−0.27 (−0.33 to −0.22)
Lower middle income	552.68 (415.91–707.2)	575.94 (433.39–736.82)	0.13 (0.12–0.14)	48.6 (37.21–62.38)	50.38 (38.52–64.78)	0.11 (0.11–0.12)	2.78 (2.1–3.57)	4.64 (3.43–6.1)	1.67 (1.59–1.76)	77.97 (59.55–100.77)	102.6 (79.66–131.44)	0.89 (0.84–0.93)
Low income	404.45 (302.2–511.87)	420.8 (315.17–531.94)	0.12 (0.11–0.12)	36.18 (27.67–46.38)	37.5 (28.86–47.89)	0.11 (0.1–0.11)	2.65 (1.92–3.4)	3.58 (2.35–4.82)	0.85 (0.73–0.96)	65.19 (51–82.64)	79.02 (58.89–101.27)	0.54 (0.47–0.60)

Abbreviations: AF/AFL, atrial fibrillation and atrial flutter; ASR, age-standardized rate; CI, confidence interval; DALY, disability-adjusted life year; EAPC, estimated annual percentage change; GBD, global burden of disease; UI, uncertainty interval; WBIL, World Bank income level.

Figure 1.

Temporal trends in the age-standardized burden of AF/AFL by WBILs from 1990 to 2023. Abbreviations: AF/AFL, atrial fibrillation and atrial flutter; ASDR, age-standardized disability-adjusted life year rate; ASIR, age-standardized incidence rate; ASMR, age-standardized mortality rate; ASPR, age-standardized prevalence rate; HICs, high-income countries; LICs, low-income countries; LMICs, lower-middle-income countries; UMICs, upper-middle-income countries; WBIL, World Bank income level.

Age-stratified analyses further demonstrated distinct patterns. Among younger individuals, ASPR, ASIR, ASMR, and ASDR increased when all income groups were combined. By income group, ASPR and ASIR increased in UMICs, LMICs, and LICs but were stable in HICs, whereas ASMR and ASDR increased across all income groups, with smaller increases in HICs (see Supplementary material online, Table S6). Among middle-aged individuals, ASPR, ASIR, ASMR, and ASDR increased across all income groups, although ASMR decreased in UMICs (see Supplementary material online, Table S7). Among older individuals, ASPR increased overall, whereas ASIR, ASMR, and ASDR remained stable when all income groups were combined. By income group, ASMR and ASDR increased in HICs, LMICs, and LICs but decreased in UMICs (see Supplementary material online, Table S8).

Regional and national patterns in AF/AFL burden and temporal changes

Substantial geographic heterogeneity was observed in AF/AFL burden and temporal trends from 1990 to 2023. Higher ASPR and ASIR were mainly observed in North America, Europe, South America, Australia, and Russia. The largest increases in ASPR were concentrated in Northeast Africa, the Middle East, China, Russia, and the USA. Similar patterns were observed for ASIR, with notable increases in the Middle East, Australia, China, Russia, and the USA. Higher ASMR and ASDR were mainly observed in Europe and the USA, whereas the largest increases in ASMR and ASDR were observed in the Middle East and South Asia (see Supplementary material online, Figure S1; Supplementary material online, Table S9).

Age-stratified analyses indicated distinct national patterns. Among younger individuals, higher ASPR and ASIR were primarily observed in Russia, the USA, Brazil, and Australia, while the largest increases in ASPR and ASIR were observed in China, the USA, Northeast Africa, and the Middle East (see Supplementary material online, Figure S2A and B; Supplementary material online, Table S10). ASMR was higher in the USA and Southeast Asia, with marked increases over time. Higher ASDR was observed in the USA, Russia, Australia, and South America, with the most pronounced increases in the USA, Central Africa, and Southeast Asia (see Supplementary material online, Figure S2C and D; Supplementary material online, Table S10).

Among middle-aged individuals, higher ASPR and ASIR were mainly observed in Russia, the USA, Brazil, and Australia, while the largest increases were concentrated in China, the USA, Northeast Africa, and the Middle East (see Supplementary material online, Figure S3A and B; Supplementary material online, Table S11). ASMR and ASDR remained highest in Europe and the USA, and the largest increases were observed in the Middle East and South Asia (see Supplementary material online, Figure S3C and D; Supplementary material online, Table S11).

Among older individuals, regional and national patterns were broadly consistent with those in the all-age population. Higher ASPR and ASIR remained concentrated in North America, Europe, South America, Australia, and Russia, with the most notable increases in China, the USA, Northeast Africa, and the Middle East (see Supplementary material online, Figure S4A and B; Supplementary material online, Table S12). ASMR and ASDR remained highest in Europe and the USA, with substantial increases observed in the Middle East and South Asia (see Supplementary material online, Figure S4C and D; Supplementary material online, Table S12).

Age patterns and age–period–cohort effects

In 2023, ASPR, ASIR, ASMR, and ASDR increased with advancing age, with more pronounced age gradients for ASMR and ASDR (Figure 2). Older individuals had higher ASPR, ASIR, ASMR, and ASDR than younger individuals, and these patterns were consistent across sexes (see Supplementary material online, Figure S5).

Figure 2.

Age-specific patterns of AF/AFL burden in 2023. Abbreviations: AF/AFL, atrial fibrillation and atrial flutter; DALY, disability-adjusted life year.

Age–period–cohort analyses indicated age- and period-dependent variations in AF/AFL burden during 1990–2023 (Figure 3). Local net drift estimates indicated increases in ASPR among individuals aged 47.5–77.5 years and increases in ASIR among those aged 47.5–72.5 years. In contrast, ASMR decreased among individuals aged 57.5–92.5 years, and ASDR decreased among those aged 67.5–97.5 years. Longitudinal age curves demonstrated that ASPR, ASIR, ASMR, and ASDR increased with age within the same birth cohort. Period effects indicated that the relative risks of ASPR and ASIR increased, followed by a decline and a subsequent increase. ASMR decreased before 2007 and then increased sharply, whereas ASDR continued to rise throughout the study period.

Figure 3.

Age–period–cohort analysis of AF/AFL burden from 1990 to 2023. Abbreviations: AF/AFL, atrial fibrillation and atrial flutter; DALY, disability-adjusted life year; RR, relative risk.

Risk-factor attribution and income-specific patterns

The attributable burden of AF/AFL-related mortality and DALYs in 2023 varied across risk factors, age groups, and WBIL strata. Overall, high SBP was the leading attributable risk factor for both ASMR and ASDR across all age groups. High BMI ranked second, followed by smoking and alcohol use, whereas diet high in sodium and lead exposure contributed smaller proportions of the attributable burden (Figure 4).

Figure 4.

Risk factor-attributable proportions of AF/AFL mortality and DALYs by age group and WBIL in 2023. Abbreviations: AF/AFL, atrial fibrillation and atrial flutter; ASDR, age-standardized disability-adjusted life year rate; ASMR, age-standardized mortality rate; DALY, disability-adjusted life year; HICs, high-income countries; LICs, low-income countries; LMICs, lower-middle-income countries; UMICs, upper-middle-income countries; WBIL, World Bank income level.

Income-stratified analyses revealed distinct risk-factor profiles across socioeconomic settings. Compared with LMICs and LICs, HICs and UMICs exhibited higher attributable proportions of ASMR and ASDR related to high BMI, smoking, alcohol use, and diet high in sodium, while the attributable proportion related to lead exposure was lower. Age-specific patterns were also evident. Compared with older individuals, younger and middle-aged groups exhibited a more diverse risk-factor composition, with greater contributions from high BMI, smoking, and alcohol use (Figure 4).

From 1990 to 2023, risk factor-attributable ASRs demonstrated dynamic temporal changes. The attributable ASR related to high BMI increased across age groups, whereas the attributable ASR related to smoking showed an overall decreasing trend. In contrast, alcohol use, diet high in sodium, and lead exposure remained relatively stable during the study period (see Supplementary material online, Figure S6).

Temporal joinpoint regression analysis of AF/AFL burden from 1990 to 2023

Joinpoint regression analysis indicated that ASPR increased from 1990 to 2023, with an AAPC of 0.03 (95% CI: 0.01–0.05; P = 0.004) (see Supplementary material online, Table S13). In age-stratified analyses, ASPR increased in younger individuals (AAPC = 0.06, 95% CI: 0.02–0.11; P = 0.010) and middle-aged individuals (AAPC = 0.12, 95% CI: 0.08–0.15; P < 0.001), whereas no significant change was observed in older individuals (P = 0.879) (see Supplementary material online, Table S13). The period 2000–05 was characterized by a marked increase in ASPR across all age groups, followed by attenuation or decline in later periods (see Supplementary material online, Table S13).

For incidence, ASIR showed no significant overall change during 1990–2023 (P = 0.196) (see Supplementary material online, Table S14). However, increasing trends were observed in younger (AAPC = 0.07, 95% CI: 0.02–0.12; P = 0.010) and middle-aged (AAPC = 0.09, 95% CI: 0.06–0.12; P < 0.001) individuals, whereas older individuals showed a comparatively stable pattern (see Supplementary material online, Table S14).

For mortality-related outcomes, ASMR increased overall, with an AAPC of 0.17 (95% CI: 0.04–0.30; P = 0.013) (see Supplementary material online, Table S15). Age-stratified analyses showed increasing trends in younger (AAPC = 1.04, 95% CI: 0.89–1.19; P < 0.001) and middle-aged (AAPC = 0.20, 95% CI: 0.11–0.29; P < 0.001) individuals, while older individuals exhibited a more stable pattern (AAPC = 0.17, 95% CI: 0.03–0.30; P = 0.018) (see Supplementary material online, Table S15). Similarly, ASDR increased from 1990 to 2023 (AAPC = 0.09, 95% CI: 0.06–0.12; P < 0.001), with the most pronounced increases in younger (AAPC = 0.41, 95% CI: 0.37–0.44; P < 0.001) and middle-aged (AAPC = 0.14, 95% CI: 0.02–0.25; P = 0.018) individuals, whereas older individuals showed smaller changes over time (AAPC = 0.05, 95% CI: 0.02–0.08; P = 0.001) (see Supplementary material online, Table S16).

Forecasted trends in AF/AFL burden using the BAPC model

BAPC projections for 2024–48 are presented in Figure 5. Overall, ASIR was projected to remain relatively stable, whereas ASPR was projected to decline modestly. In contrast, both ASMR and ASDR were projected to increase over the forecast period. Age-stratified projections showed that ASPR and ASIR were projected to decline in younger and middle-aged populations, whereas ASMR and ASDR were projected to increase. Among older individuals, ASPR was projected to decrease, ASMR was projected to increase slightly, and ASIR and ASDR were projected to remain stable.

Figure 5.

Forecasted trends in AF/AFL burden from 1990 to 2048 based on the BAPC model. Abbreviations: AF/AFL, atrial fibrillation and atrial flutter; ASDR, age-standardized disability-adjusted life year rate; ASIR age-standardized incidence rate; ASMR, age-standardized mortality rate; ASPR, age-standardized prevalence rate; BAPC, Bayesian age–period–cohort.

Across all age groups, ASPR was projected to decline (see Supplementary material online, Figure S7). ASIR was projected to decrease in individuals aged 30–55 years and remain stable in those aged ≥ 55 years (see Supplementary material online, Figure S8). ASMR was projected to increase substantially among individuals aged 30–64 years, while remaining stable in those aged ≥ 65 years (see Supplementary material online, Figure S9). ASDR was projected to increase in the 30–64 age group, remain stable in those aged 65–94 years, and decline in individuals aged ≥ 95 years (see Supplementary material online, Figure S10).

Discussion

Utilizing GBD 2023 estimates, we characterized age-specific and income-stratified patterns in AF/AFL burden from 1990 to 2023 and identified substantial heterogeneity across age groups and WBIL strata. Several findings warrant emphasis. First, the absolute burden of AF/AFL in 2023 remained concentrated in older adults, who accounted for the majority of deaths and DALYs. Second, temporal patterns differed by age: ASPR increased in older individuals with comparatively stable ASIR, ASMR, and ASDR in the overall older population, whereas younger and middle-aged populations exhibited consistent increases across ASPR, ASIR, ASMR, and ASDR. Third, high SBP was the leading attributable risk factor for both ASMR and ASDR across all age strata, with high BMI ranking second and smoking and alcohol use contributing more prominently in younger and middle-aged groups. Finally, BAPC projections through 2048 suggested that, despite projected declines in ASPR and ASIR among younger and middle-aged populations, ASMR and ASDR are expected to increase, indicating a potentially worsening clinical impact in these age groups.

From 1990 to 2023, age-standardized metrics of AF/AFL burden generally increased, with the numbers of prevalent cases, incident cases, deaths, and DALYs increasing by 2.5-, 2.4-, 3.3-, and 2.7-fold, respectively, underscoring the growing public health challenge posed by AF/AFL and highlighting the need for strengthened prevention and management strategies. While prior GBD-based studies have documented increasing trends in AF/AFL burden, detailed comparisons across younger, middle-aged, and older populations have been less consistently emphasized.^4,12,14 Recent analyses focusing on early-onset vs. older-age AF/AFL have highlighted a disproportionately faster growth of burden in younger populations, together with distinct age-specific risk profiles.⁷ In the present analysis, among individuals aged ≥ 65 years, ASPR increased, whereas ASIR, ASMR, and ASDR remained relatively stable, indicating that the expanding prevalence in older adults has not been accompanied by parallel increases in ASMR and ASDR. In contrast, among individuals aged < 65 years, ASPR, ASIR, ASMR, and ASDR all increased, supporting a shift towards a more prominent burden in younger and middle-aged populations. Several factors may contribute to the increasing AF/AFL burden in younger and middle-aged groups, including the rising prevalence of cardiometabolic disorders, adverse behavioural exposures, and enhanced detection of subclinical and previously undiagnosed AF/AFL through expanded screening and diagnostic availability, including widespread availability of wearable and smartphone-based electrocardiogram technology, which increases detection of asymptomatic cases but is coupled with insufficient anticoagulation adherence and accessibility.^15–22 These findings advocate for earlier intervention in prevention and care, including systematic risk-factor identification, timely rhythm evaluation, and equitable access to guideline-recommended anticoagulation when appropriate, to reduce preventable deaths and disabilities in younger and middle-aged patients, potentially enhanced by the use of polygenic risk scores.²³

Beyond age heterogeneity, AF/AFL burden showed marked disparities across WBIL strata, likely reflecting differences in risk-factor profiles, diagnostic capacity, and long-term disease management. At the overall level, all four age-standardized metrics increased in HICs, whereas UMICs showed increasing ASPR and ASIR but decreasing ASMR and ASDR; in contrast, LMICs and LICs experienced concurrent increases across all metrics, indicating disparities in AF/AFL burden related to economic development and healthcare infrastructure. Age-stratified analyses revealed that in younger populations, both ASPR and ASIR increased in UMICs, LMICs, and LICs, but remained stable in HICs. Similarly, ASMR and ASDR increased across all income groups, with larger increases observed in lower-income regions. These findings highlight socioeconomic gradients in AF/AFL burden, where limited diagnostic capacity and constrained access to evidence-based long-term management may jointly contribute to higher preventable mortality and disability in lower-income settings. While HICs typically report higher rates of ASPR and ASIR due to better diagnostic capabilities, LMICs and LICs show higher ASMR and ASDR, reflecting deficiencies in secondary prevention and healthcare access. In these regions, delayed diagnoses in younger populations and multimorbidity in older adults contribute to the escalating burden. Reducing these inequities will require strengthening primary-care-based detection, improving affordability and continuity of anticoagulation and rate/rhythm control, and building capacity for longitudinal follow-up and stroke rehabilitation in LMICs and LICs.

Consistent with prior GBD data, elevated SBP remained the leading risk factor for AF/AFL-related mortality and DALYs.^12,24,25 However, attributable profiles differed across WBIL strata: compared with LMICs and LICs, HICs and UMICs showed higher attributable proportions related to high BMI, smoking, alcohol use, and diet high in sodium, particularly among younger and middle-aged groups. High BMI, indicative of visceral adiposity and insulin resistance, facilitated atrial remodelling through chronic inflammation, oxidative stress, and activation of neurohormonal pathways, including the renin–angiotensin–aldosterone system.^26–29 Concurrent insulin resistance exacerbated sympathetic activation, endothelial dysfunction, left atrial enlargement, and electrophysiological vulnerability.^30,31 Smoking may increase AF/AFL susceptibility through multiple pathways, including endothelial dysfunction and atrial microvascular impairment, heightened systemic inflammation and oxidative stress, and autonomic imbalance with increased sympathetic tone. These processes promote atrial fibrosis and structural remodelling, shorten atrial effective refractory periods, and increase ectopic activity, collectively facilitating both AF/AFL initiation and maintenance. Additionally, alcohol exerted dose-dependent and partially synergistic proarrhythmic effects, with binge drinking episodes precipitating acute arrhythmogenic triggers, and chronic heavy drinking inducing cardiomyocyte apoptosis, intracellular calcium overload, and autonomic imbalance.^32,33 Therefore, the intersection of obesity, smoking, and excessive alcohol consumption presented a multifactorial metabolic-neurohumoral insult, significantly contributing to premature AF/AFL-related morbidity and mortality in younger populations, particularly in HICs and UMICs. Public health strategies must evolve beyond isolated hypertension management towards a comprehensive, integrated approach encompassing: (i) population-wide weight-management interventions (e.g. sugar-sweetened beverage taxation, front-of-pack caloric labelling, and active-transportation urban design); (ii) smoking cessation interventions, including systematic screening for tobacco use, brief physician-delivered counselling, and access to evidence-based pharmacotherapy (e.g. nicotine replacement therapy, varenicline, or bupropion); (iii) alcohol harm-reduction policies (e.g. minimum unit pricing, sales restrictions, and brief primary care interventions); and (iv) early metabolic screening (e.g. fasting glucose and glycated haemoglobin) among individuals with overweight or obesity. Without aggressive action on this expanding metabolic–behavioural axis, gains from hypertension control risk are being undermined by the emerging obesity–smoke–alcohol-driven atrial disease epidemic, particularly impacting younger, economically productive populations.

Future projections based on BAPC models indicated that ASPR may decline modestly and ASIR may remain relatively stable in the overall population, whereas both ASMR and ASDR were projected to increase over 2024–48. Among individuals aged < 65 years, despite projected declines in ASPR and ASIR, ASMR and ASDR were expected to increase, suggesting that although the number of AF/AFL cases may decrease, the risks of death and disability may rise in younger patients. This paradox may reflect earlier disease onset driven by obesity, insulin resistance, and hazardous alcohol use; inadequate utilization of rhythm control and guideline-directed anticoagulation therapy in younger populations; and prolonged exposure to atrial structural and electrical remodelling. In contrast, among individuals aged ≥ 65 years, ASPR was projected to decline, while ASIR and ASDR were expected to remain stable and ASMR to increase slightly. This pattern may reflect increasing multimorbidity and competing cardiovascular risks among older adults, even as anticoagulation and rhythm-control strategies improve in higher-resource settings. These nuanced age-specific projections highlight several critical priorities for public health action. First, prevention strategies should be initiated earlier in adulthood by integrating structured weight management, blood pressure optimization, and alcohol reduction programs into routine primary care and workplace health initiatives before midlife. Second, persistent treatment gaps must be addressed, particularly the underuse of oral anticoagulation among middle-aged individuals who already meet guideline-based criteria. Third, even with declining incidence in older populations, the cumulative number of AF/AFL survivors will continue to generate sustained demand for atrial-specific rehabilitation and long-term chronic disease management infrastructure, necessitating strategic capacity planning across cardiology and related specialties. Finally, interpretation of these projections required consideration of two important caveats: the dependence of projections on current metabolic and therapeutic trends, where unexpected shifts in obesity prevalence or novel preventive and therapeutic breakthroughs could significantly alter future trajectories, and the likelihood of underestimating future burdens in lower socioeconomic settings due to limited surveillance capacity, underscoring the necessity of strengthened age-stratified monitoring and targeted implementation research to translate model-based insights into timely, evidence-informed policy decisions.

Limitations

This study had several limitations. First, GBD 2023 estimates are model-based outputs derived from ICD-coded sources, statistical modelling, and imputation rather than uniformly collected clinical registries. Consequently, the accuracy of national estimates, particularly in LMICs and LICs with sparse surveillance, depended on the strength of the underlying statistical priors and may diverge from local epidemiology. Second, silent or paroxysmal AF/AFL frequently remains undiagnosed, even in high-income health systems; routine data sources therefore under-captured these episodes, potentially leading to underestimation of incidence and prevalence. Third, the GBD framework treated AF/AFL as a single entity, without distinguishing paroxysmal, persistent, or permanent forms, nor grading disease severity or accounting for common comorbidities such as hypertension, diabetes mellitus, or chronic kidney disease—all factors that modulated prognosis.³⁴ Fourth, the current GBD publication and authorship policy framework emphasizes coordinated dissemination of certain fully aggregated ‘global’ outputs through designated channels. To ensure full compliance with this policy environment, our analyses and interpretation were intentionally focused on income-stratified and age-specific patterns, rather than reproducing or substituting for GBD core global outputs. While this design may limit direct comparability with some earlier GBD-based studies that primarily highlighted global summaries, it provides a complementary perspective that is more actionable for targeted prevention and health-system planning across socioeconomic settings. Fifth, severity weights were applied uniformly across geography and time, yet therapeutic advances, ranging from widespread adoption of direct oral anticoagulants to greater availability of catheter ablation, may have altered the natural history of AF/AFL in ways not fully reflected by static disability weights. Addressing these limitations would require prospective, subtype-specific registries in under-represented regions, linkage of administrative claims to detailed clinical data, and periodic recalibration of disability weights to capture evolving treatment landscapes.

Conclusions

This study demonstrated pronounced age-related heterogeneity in AF/AFL burden from 1990 to 2023, with a shifting epidemiological profile characterized by increasing mortality- and disability-related burden in younger and middle-aged populations. These findings underscore the need for age-tailored prevention and management strategies, including earlier detection, optimized blood pressure and metabolic risk control, and improved adherence to guideline-recommended anticoagulation when indicated. Public health priorities should target modifiable risk factors, particularly high SBP, high BMI, smoking, and alcohol use, and strengthen equitable access to effective AF/AFL care pathways, especially in resource-limited settings, to mitigate projected increases in AF/AFL-related mortality and disability.

Authors’ contributions

Siyuan Tan (Conceptualization, Writing—original draft), Zixi Zhang (Conceptualization, Writing—original draft), Qiuzhen Lin (Methodology, Software), Tao Tu (Investigation, Resources), Jiabao Zhou (Supervision, Validation), Jiayi Zhu (Supervision, Validation), Gaoming Zeng (Data curation, Investigation), Fanqi Li (Data curation, Investigation), Kangrong Li (Methodology, Resources), Yongguo Dai (Methodology, Resources), Cancan Wang (Methodology, Resources), Chan Liu (Methodology, Resources), Yichao Xiao (Conceptualization, Writing—review & editing), and Qiming Liu (Conceptualization, Writing—review & editing).

Supplementary material

Supplementary material is available at Europace online.

Funding

This work was supported by the National Natural Science Foundation of China (No. 82270337, 82470333, 82300357), the Chinese Society of Cardiology’s Foundation (No. CSCF2024B02), the Scientific Research Program of FuRong Laboratory (2025PT5001), the Hunan Provincial Natural Science Foundation of China (No. 2021JJ30033, 2023JJ30791), the Key Project of Hunan Provincial Science and Technology Innovation (No. 2024JK2119), and the Clinical Medical Technology Innovation Guidance Project of Hunan Science and Technology Agency (No. 2021SK53519).

Conflict of interest

All authors confirm that they have no conflicts of interest to declare, and none has received any personal fees.

Data availability

This study utilized publicly available data from the GBD Study 2023, accessed through the Institute for Health Metrics and Evaluation (IHME) Global Health Data Exchange platform. Ethical approval was not required as the data are anonymized and aggregated. The data from this study can be accessed openly through the GBD 2023 online database.