The Global Burden of Diseases (GBD) 2021 study, conducted by the Institute for Health Metrics and Evaluation (IHME), provides comprehensive estimates for 369 diseases and 88 risk factors across 204 countries or regions. Study design and methods are detailed in prior GBD literature (2, 15). The University of Washington IRB waived informed consent for GBD data access (16). This study adhered to the Guidelines for Accurate and Transparent Health Estimates Reporting (GATHER) (17).

Temperature estimates were derived from the ERA5 reanalysis dataset, provided by the European Center for Medium-Range Weather Forecasts, which offers high-resolution (0.25° × 0.25°) temperature data from 1979 to the present. Non-optimal temperatures were defined as temperatures exceeding or falling below the theoretical minimum risk exposure level, which corresponds to daily temperatures associated with the lowest mortality rates for specific locations. To mitigate the influence of extreme temperatures and noise, data were truncated, preventing extrapolation of temperatures beyond the range of mortality data. Temperature ranges were truncated at the 1st and 99th percentiles of the mortality dataset, and daily mean temperatures were similarly truncated within each temperature range (13, 18).

Non-optimal temperatures, including both extreme high and low temperatures, are identified as risk factors for myocarditis/cardiomyopathy (referred to as myocardial disease; MD) and ischemic heart disease (IHD) in the GBD database. Cause-specific relative risks of non-optimal temperature on daily mortality in the Global Burden of Disease Study were estimated using methods by Burkart et al. (19). Firstly, a Bayesian meta-regression framework with a two-dimensional spline model was applied to analyze the relative risks of 176 causes of death related to daily temperature. These relative risks were then applied globally, using daily temperature data combined with Global Burden of Disease Study data to calculate disease burden for each region.

Burden was measured using mortality and disability-adjusted life years (DALYs), comprising years of life lost and years lived with disability (YLD). Age-standardized rates were calculated using the direct method with the GBD 2019 standard population as a reference, expressed per 100,000 population (ASMR for mortality, ASDR for DALYs). Age standardization was applied to eliminate the confounding effects of different population age structures across regions. Trends from 1990 to 2021 were assessed using Estimated Annual Percent Change (EAPC); Total Percentage Change (TPC) was used for regions with missing or negative values. Analyses were stratified by sex, 5-year age groups, and socio-demographic index (SDI) categories to capture burden patterns at global, regional, and national levels.

A positive 95% confidence interval indicated increasing trends, negative values indicated decreases, and intervals including zero suggested stability. All analyses were performed using R (version 4.2.1), with a two-sided P-value < 0.05 considered significant.

From 1990 to 2021, age-standardized mortality rates (ASMR) and age-standardized DALY rates (ASDR) attributable to high temperatures for myocardial disease (MD) and ischemic heart disease (IHD) generally increased, although with fluctuations. By 2021, the ASDR for heat-attributable MD was 1.97 per 100,000 (95% CI: −0.80 to 4.76), corresponding to 158,029 DALYs, while the ASMR reached 0.06 per 100,000 (95% CI: −0.02 to 0.15), with 5,224 deaths. For IHD, the ASDR was 30.57 per 100,000 (95% CI: 5.41 to 66.96), amounting to 2,623,940 DALYs, with an ASMR of 1.34 per 100,000 (95% CI: 0.20–3.07), corresponding to 112,390 deaths. Over 30 years, the estimated annual percent change (EAPC) in DALYs attributable to high temperatures was 1.26% (95% CI: 0.78% −1.75%) for MD, and 1.68% (95% CI: 1.33% −2.03%) for IHD. Mortality rates showed similar trends, with EAPCs of 1.35% (95% CI: 0.86% −1.85%) and 1.66% (95% CI: 1.31% −2%), respectively (Figure 1).

In contrast, ASMR and ASDR attributable to low temperatures, although historically higher, showed consistent reductions over time. By 2021, the ASDR for cold-attributable MD was 7.01 per 100,000 (95% CI: 5.35–9.03), corresponding to 578,936 DALYs, with an ASMR of 0.27 per 100,000 (95% CI: 0.21–0.34), reflecting 22,248 deaths. For IHD, the ASDR was 117.40 per 100,000 (95% CI: 101.19–144.86), accounting for 9,961,231 DALYs, and the ASMR was 6.14 per 100,000 (95% CI: 5.23–7.53), corresponding to 505,298 deaths. The EAPC in DALYs attributable to low temperatures was −1.87% (−2.17% −1.58%) for MD, and −1.73% (95% CI: −1.85% −1.6%) for IHD, with similar decreasing trends observed in ASMR (Figure 1 and Table 1).

Over the past three decades, men consistently exhibited a higher burden from extreme temperatures than women. For heat-attributable MD, the ASDR increased more rapidly in men (EAPC: 1.47%, 95% CI: 0.97% −1.97%) than in women (EAPC: 0.93%, 95% CI: 0.45% −1.41%), reaching 2.53 vs. 1.42 per 100,000 in 2021, respectively. A similar pattern was observed for heat-attributable IHD, with ASDR EAPCs of 1.83% (95% CI: 1.48% −2.18%) in men and 1.40% (95% CI: 1.04% −1.76%) in women, and 2021 ASDRs of 40.34 vs. 21.65 per 100,000.

In contrast, for cold-attributable burden, both sexes showed decreasing trends, but women experienced steeper declines. For MD, the ASDR EAPC was −2.38% (95% CI: −2.59% to −2.16%) in women vs. −1.58% (95% CI: −1.92% to −1.24%) in men; corresponding 2021 ASDRs were 4.47 in women and 9.62 in men. For IHD, EAPCs were −1.98% (95% CI: −2.10% to −1.86%) in women and −1.55% (95% CI: −1.67% to −1.42%) in men, with ASDRs of 84.78 vs. 153.52 per 100,000, respectively (Figure 1).

From 1990 to 2021, the heat-attributable CVD burden increased across most socio-demographic index (SDI) tiers, particularly in low-SDI regions. For MD, high-SDI regions showed a slight decline (ASDR EAPC −0.29%), while low-SDI regions experienced significant rises (ASDR EAPC 2.8%). IHD burdens also grew more sharply in lower SDI areas. By 2021, low-middle SDI regions recorded the highest ASDRs for both MD (3.93 per 100,000) and IHD (78.49 per 100,000), contributing 64,292 DALYs for MD and 1,169,778 DALYs for IHD attributable to heat (Figure 1, Table 1).

For low temperatures, burdens declined across all SDI tiers, with the steepest reductions in high-SDI regions (MD: ASDR EAPC −2.93%; IHD: ASDR EAPC −3.66%). Declines were less pronounced in low and low-middle SDI regions. By 2021, middle-high SDI regions had the highest ASDRs for both MD and IHD, contributing ~2,833,786 DALYs for IHD and 234,094 DALYs for MD, highlighting persistent cold-related risks even in relatively developed areas. ASMR trends followed similar patterns across SDI levels (Table 1).

Significant regional variation were noted. For heat-attributable MD, Eastern Europe saw the largest increase (ASDR TPC 20.01%), and for IHD, Southeast Asia had the largest rise (ASDR TPC 7.59%). For cold-attributable MD, only Central Asia and Eastern Europe exhibited slight increases, while other regions declined. By 2021, Central Asia, South Asia, and North Africa/Middle East bore the highest heat-attributable burdens, while Eastern Europe and Central Asia had the highest cold-attributable ASDRs (Figure 2 and Table 2).

Temperature-related burdens followed SDI patterns: high-temperature burdens increased in regions with SDI below 0.5, but declined above this threshold. For cold-attributable IHD, burdens increased sharply between SDI 0.6 and 0.75 and then declining at higher SDI levels. For cold-attributable MD, burdens increased below SDI 0.45, fluctuated between 0.45 and 0.75, and declined beyond 0.75 (Figure 2).

Between 1990 and 2021, heat-attributable cardiovascular trends varied widely by country. MD burden rose in 38% of countries, led by Kazakhstan (total percentage change, TPC = +45.84%), whereas Bangladesh showed the largest decline (−104.11%). For IHD, 39% of countries registered increases, headed by Jamaica (+33.78%), with Zambia recording the greatest reduction (−44.36%). In 2021, Turkmenistan had the highest heat-attributable MD burden (ASDR 41.22; ASMR 1.22 per 100,000) and Iraq the highest IHD burden (ASDR 467.99; ASMR 24.70 per 100,000; Figure 3; Supplementary Tables S1, S2).

For cold-attributable burden, most countries improved: only ~15% saw increases. Kazakhstan again recorded the largest MD rise (+17.59%), whereas Italy showed the steepest fall (−0.87%). In 2021 Kazakhstan also had the highest cold-attributable MD burden (ASDR 75.50; ASMR 2.65 per 100,000). For IHD, Kiribati posted the greatest increase (+7.97%), while the Maldives showed the largest drop (−0.88%). Turkmenistan remained highest for cold-attributable IHD (ASDR 556.82; ASMR 29.72 per 100,000). The Cook Islands and Marshall Islands had the lowest burdens for both diseases.

Heat-attributable burden: MD ASDR rose steadily with age—from 3.93 (60–64 years) to 7.40 (80–84 years), peaking at 18.26 per 100 000 beyond 95 years; children < 5 year were also vulnerable (3.45). Heat-attributable IHD rose sharply from mid-life, reaching 320.06 per 100,000 beyond 95 years.

Cold-attributable burden: MD ASDR climbed from 16.58 (60–64 years) to 35.69 (80–84 years) and 124.39 beyond 95 year, with children < 5 year again relatively affected (5.23). Cold-attributable IHD accelerated after 80 years, rising from 340.03 (60–64 years) to 1 130.74 (80–84 years) and 2 734.64 per 100,000 beyond 95 years (Figure 4; Supplementary Tables S3–S10).

The three main contributors to the increased global burden of MD and IHD attributable to non-optimal temperatures were aging, population change, and epidemiological progress. For heat-attributable MD, population change contributed 54.79% to the DALY increase, with epidemiological shifts accounting for 27.87% and aging for 17.34%. For heat-attributable IHD, aging accounted for 30.70%, population growth for 33.53%, and epidemiological change for 35.77%. Cold-attributable diseases, aging and population change were the dominant factors, especially in low-SDI regions where population growth was the primary driver (Figure 5 and Supplementary Table S11).

From 2021 to 2040, the global heat-attributable burden of MD and IHD is projected to rise. MD DALYs and mortality are expected to increase steadily, with the global ASDR reaching 2.37 per 100,000 and the ASMR 0.077 per 100,000 (≈10,117 deaths) by 2040. For IHD, the heat-attributable ASDR is projected to climb to 35.24 per 100,000 and the ASMR to 1.523 per 100,000 (≈215,318 deaths), peaking around 2034 before a slight decline (Figure 6; Supplementary Table S12).

For cold-attributable burden, global DALYs and mortality are projected to continue declining. By 2040, the cold-attributable ASDR for MD is expected to fall to 5.56 per 100,000, with an ASMR of 0.217 per 100,000 (≈29,422 deaths). For IHD, the ASDR is forecast to decrease to 137.18 per 100,000, with an ASMR of 5.193 per 100,000 (≈756,304 deaths; Figure 6; Supplementary Table S12).

Our study confirms that cold temperatures still account for a greater share of global cardiovascular mortality compared to heat. However, while cold-related burdens have steadily decreased, heat-related burdens have continued to rise since 1990. This divergence parallels the accelerating pace of anthropogenic warming and intensified urban heat-island effects highlighted by the IPCC and Lancet Countdown reports (23, 24). Additionally, incomplete classification and underreporting of heat-related deaths, particularly in regions with inadequate mortality registration systems, may further underestimate the true impact (10).

Our findings highlight distinct differences between IHD and MD in their susceptibility and mechanisms underlying temperature-related cardiovascular burdens. IHD demonstrates particular vulnerability to high-temperature exposures through mechanisms including increased vascular inflammation, plaque instability, hypercoagulability, and dehydration-induced hemoconcentration, all exacerbating existing atherosclerotic conditions (25). In contrast, MD—encompassing diverse etiologies such as viral myocarditis, autoimmune reactions, and genetic cardiomyopathies—is typically exacerbated by heat-induced systemic inflammation, oxidative stress, and direct myocardial injury rather than primarily through vascular or thrombotic pathways (26, 27). These pathophysiological differences help explain why IHD and MD show distinct epidemiological trends under temperature extremes and emphasize the need for disease-specific prevention and management strategies.

The age-related vulnerabilities to extreme temperatures involve specific physiological mechanisms. Children under five exhibit heightened sensitivity due to immature thermoregulatory systems and diagnostic delays (28, 29). Conversely, older adults face increased risks from impaired vasodilation, reduced sweating capacity, and diminished autonomic control exacerbated by chronic conditions like diabetes and hypertension (30). Gender disparities across socioeconomic development indices (SDI) showed men consistently having higher burdens, likely due to occupational exposure, lifestyle factors (e.g., smoking, alcohol), and less protective estrogenic vascular effects compared to women (31–33).

High-temperature burdens have risen fastest in low-SDI regions, where weak health systems, minimal cooling infrastructure, and widespread outdoor work intensify exposure. Each 1°C rise in temperature is associated with an ~9% increase in cardiovascular mortality in lower-middle-income countries compared to high-income settings (34). While high-SDI regions have largely stabilized or reduced heat- and cold-attributable burdens through better housing, air-conditioning, and heat-alert systems, several upper-middle-SDI economies are now seeing renewed increases, underscoring that economic growth alone does not guarantee resilience (35, 36).

Against this broader pattern, Eastern Europe has experienced the steepest rise in heat-attributable MD (ASDR +20%), likely driven by high rates of alcohol-related cardiomyopathy (37), rapid increases in extreme-heat days coupled with low air-conditioning coverage, and population aging combined with high smoking and hypertension prevalence (38, 39). Southeast Asia records the largest surge in heat-related IHD (ASDR +7.6%), reflecting urban heat islands, limited affordable cooling, and increasing rates of hypertension and diabetes (19, 40). Although cold-attributable MD has declined globally, Central Asia and Eastern Europe still show slight increases, likely due to poorly insulated housing, low indoor winter temperatures, and persistent high smoking rates (41). By 2021, heat-related cardiovascular mortality peaks had shifted to Central Asia, South Asia, and North Africa/Middle East, where ≥40°C heatwaves coincide with limited cooling infrastructure, whereas cold-related peaks remained concentrated in Eastern Europe and Central Asia (42).

Overall, heat-related mortality has been rising while cold-related mortality has decreased. Yet, three exceptional seasons disrupted these trends: the western Russia heatwave in 2010 (43) and the trans-Eurasian heatwave of 2020 (44) each triggered significant spikes in cardiovascular deaths, briefly reversing earlier gains from milder summers, heat-action plans, and expanded cooling access. Conversely, the 2005 winter saw an unusual cold-related spike linked to a strong negative Arctic Oscillation (45, 46), a resurgence of solid-fuel heating, and an influenza B outbreak (47). These episodes illustrate how single-season extremes can temporarily override long-term trends, and how recent adaptation gains are increasingly challenged by rising heat intensity and an aging global population.

Our findings extend previous work by providing a disease-specific breakdown of temperature-related cardiovascular burdens. Unlike earlier multi-country studies that aggregated all cardiovascular outcomes (48, 49), our analysis separately quantified IHD and MD burdens across diverse demographics and socioeconomic contexts. Previous GBD analyses primarily considered older populations without detailed sex or regional stratification (50), whereas our study highlights vulnerabilities across all ages and explicitly quantifies gender-specific risks. Additionally, previous single-country studies noted seasonally increased cardiomyopathy events without comprehensive mortality data (51, 52); our global perspective fills this gap and underscores sustained heat-related increases in MD. Our future projections, which incorporate updated demographic and epidemiological data, contrast with prior optimistic forecasts based solely on technological adaptation (53, 54).

Our decomposition analysis revealed that population growth significantly contributed (over 54% for MD and 30% for IHD) to rising heat-related cardiovascular burdens, particularly in rapidly urbanizing lower-SDI regions (55). Population aging accounted substantially for both MD and IHD burden increases (17.3 and 30.7%, respectively), reflecting age-related physiological vulnerabilities and chronic disease prevalence (30, 56). Epidemiological transitions, especially increased prevalence of obesity and diabetes, contributed more significantly to IHD (35.8%) compared to MD (27.9%), highlighting IHD's stronger association with modifiable risk factors (57). The declining cold-related burdens primarily reflect improved housing, heating infrastructure, and healthcare, though aging populations still sustain significant residual risks.

This study provides the first comprehensive, disease-specific global assessment of temperature-related cardiovascular burdens, incorporating long-term trend analyses and future projections with decomposition of demographic and epidemiological drivers. However, it shares inherent limitations associated with GBD methodologies, including potential misclassification of deaths, incomplete data coverage (particularly in low-SDI settings), and limited assessment of delayed temperature effects or adaptation behaviors. Future research should address these gaps by integrating finer-scale data and evaluating the impacts of adaptation measures more systematically.