The GBD 2019 study provides comprehensive descriptive epidemiology for 369 diseases and injuries across 204 countries and regions during 1990–2019. Each death is attributed to a single underlying cause chosen from a mutually exclusive and collectively exhaustive list of diseases and injuries. The GBD 2019 study team obtained ethics approval from the University of Washington Institutional Review Board Committee. Detailed GBD protocols and all data are available online and can be requested through the EPHI-IHME Office (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9416661/). The GBD 2019 data are publicly accessible and contain no identifiable personal information (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9897059/).

We conducted secondary analysis of IHD mortality data from the GBD 2019 study, which can be accessed via the Global Health Data Exchange (GHDx) website (http://ghdx.healthdata.org/). This is a public database platform supported by researchers from 162 countries and territories since 1990. The GBD project continuously updates its disease lists, data sources, and methods [5]. GBD 2019 provides comprehensive mortality estimates from 1990 to 2019 at precise geographical locations, encompassing 204 countries and territories [5]. Cause of death (CoD) is coded using the International Classification of Disease, 10th Revision (ICD-10). IHD was defined in this study as ICD-10 codes I20–I25. GBD 2019 imported data from vital registration systems and verbal autopsy studies, and incorporated new dietary covariates into the existing standard Cause of Death Ensemble modeling (CODEm) approach. Detailed methodological descriptions can be found in Methodology Appendix 1 [5].

We employed crude and age-standardized mortality rates (ASMRs) to quantify the global burden of IHD deaths. Mortality rates were expressed as deaths per 100,000 person-years. Age standardization was performed using the GBD 2019 world population age standards [5]. We also calculated mortality ratios by comparing mortality rates to a mid-period reference. Successive birth cohorts (1895–2004, at 10-year intervals) were examined to explore temporal patterns. Additionally, we assessed the impact of 26 risk factors for IHD provided in GBD 2019 on changes in ASMRs.

GBD 2019 incorporated the socio-demographic index (SDI) in its analysis. This composite socio-economic development indicator is based on income per capital, average years of schooling, and fertility in females aged 25 years. Higher SDI values denote greater socio-economic development. Although GBD 2019 scaled the SDI value from 0 to 100, we utilized the earlier GBD 2017 SDI scale that ranged from 0 to 1. All countries were categorized into five SDI groupings (high, high-middle, middle, low-middle, and low) according to their SDI quintiles.

We implemented an APC model to disentangle the effects of age, period, and birth cohort on IHD mortality trends. APC model analysis is employed frequently in the descriptive epidemiology of chronic diseases. It enables examination of the individual effects of biological aging, historical periods, and generational cohorts on the disease rate. Moreover, it separates the contributions of technological and social factors from age-associated biological influences [6]. We used the APC model in R (version 3.6.3, R Foundation for Statistical Computing, Vienna, Austria), as described previously [7]. In a typical APC model, age and period are divided into equal time intervals. In this study, we set 16 age groups from 15 to 94 years in 5-year increments (15–19, …, 90–94 years). We also set 11 birth cohorts in 10-year increments (1895–1904, …, 1995–2004), with the mid-period (1945–1954) set as the reference. Period cohorts were defined as 6 five-year calendar periods (1990–1994, …, 2015–2019), with the reference period being 2000–2004. Using unconstrained parameters, we estimated the APC effects while mitigating the standard identification problem [6, 7].

The two most important APC parameters were net drift and local drift [7]. Net drift is defined as the overall annual percentage change in the age-standardized mortality rate across the study period, reflecting the combined time trend from period and cohort effects that is common to all ages. Local drift is defined as the age-specific annual percentage change in mortality over time for each age group [7]. The U.S. Healthy People 2020 Final Progress Table defines an objective that has moved 10% relative to its baseline over a 10-year period as “Little or No Detectable Change”, whereas an objective that has moved 10% or more is defined as “Got Worse or Improved” [8]. Therefore, a net or local drift in the present study of $\pm$1% per year or more was considered substantial, and approximates to changes of $\pm$10%, $\pm$18%, and $\pm$26% in fitted rates over 10-, 20-, and 30-years, respectively. APC model outputs also included fitted longitudinal age-specific rates adjusted for period deviations (representing age effects), and period- and cohort-specific relative risks (reflecting period and cohort effects). Period and cohort rate ratio curves incorporated the entire value of the net drift [7].

Mortality data were presented as absolute numbers, all-age mortality rates, and ASMRs with 95% uncertainty intervals (UIs). Percentage changes were calculated as ([new value – old value] / old value) $\times$100. Relative risks (RRs) were expressed as the ratios of age-specific rates for a given period (or birth cohort) to the corresponding rate in the designated reference period (or cohort). Trends in annual percentage change derived from the APC model were evaluated using Wald chi-square tests. All tests were two-sided, and statistical significance was set at p 0.05. The proportion of deaths refers to the percentage of all IHD deaths in the corresponding group. Rates were depicted as per 100,000 person-years. All analyses were performed in R (version 3.6.3).

Trends in population size, total number of IHD deaths, all-age mortality rates, ASMR, and the net drift in mortality by SDI quintiles from 1990 to 2019 are shown in Table 1, Supplementary Table 1 and Supplementary Figs. 1,2. Globally, annual IHD deaths increased from 5.70 million (95% UI: 5.41–5.90) in 1990 to 9.14 million (95% UI: 8.40–9.74) in 2019, reflecting a 60.43% rise. During the same period, the world population grew by approximately 45%, from 5.35 billion (95% UI: 5.23–5.46) to 7.74 billion (95% UI: 7.49–7.93).

Between 1990 and 2019, high-SDI countries experienced a 14.30% (95% UI: –19.35% to –9.59%) reduction in IHD-related deaths. In contrast, marked increases occurred in other SDI groups: 42.08% (95% UI: 32.95%–50.53%) in high-middle SDI countries, 145.37% (95% UI: 122.94%–167.00%) in middle-SDI countries, 130.96% (95% UI: 104.70%–156.95%) in low-middle SDI countries, and 106.81% (95% UI: 78.43%–134.17%) in low-SDI countries. Consequently, the proportion of global IHD deaths in high-SDI countries declined from 29.65% in 1990 to 15.84% in 2019, whereas in middle-SDI countries they accounted for a substantial 30.91% of total IHD deaths in 2019.

All-age mortality rates reflected similar patterns. The global all-age mortality rate increased from 106.70 (95% UI: 101.03–110.20) to 118.10 (95% UI: 108.51–125.93) per 100,000 person-years. Across SDI quintiles, only high-SDI and low-SDI countries saw decreases in all-age mortality. High-SDI countries demonstrated a –30.49% (95% UI: –34.58% to –26.67%) reduction, while low-SDI countries exhibited a modest –3.23% (95% UI: –16.51% to 9.58%) decrease. In contrast, high-middle SDI countries showed the highest all-age mortality rate, reaching 185.84 (95% UI: 168.61–198.02) per 100,000 person-years in 2019.

ASMR, however, declined consistently across all SDI groups over the past 30 years. Globally, ASMR decreased from 170.45 (95% UI: 159.61–176.94) to 117.95 (95% UI: 107.83–125.92) per 100,000 person-years. High-SDI countries experienced the largest relative reduction in ASMR of –58.68% (95% UI: –60.30% to –56.69%), followed by –35.28% in high-middle SDI countries (95% UI: –39.01% to –31.69%). In 2019, high-SDI countries recorded the lowest ASMR of 67.10 (95% UI: 60.07–71.54) per 100,000 person-years. Other SDI groups had similar ASMRs, ranging from 144.21 (95% UI: 132.05–156.58) to 136.59 (95% UI: 122.96–149.50) per 100,000 person-years.

Net drift estimates from the APC model mirrored the ASMR trends. From 1990 to 2019, the global net drift in annual mortality declined by –1.10% (95% CI: –1.17% to –1.04%) per year. High-SDI countries witnessed the steepest decline of –2.84% (95% CI: –3.05% to –2.64%), while low-middle SDI countries showed the smallest decrease of –0.26% (95% CI: –0.38% to –0.15%).

National-level trends varied significantly, and being in a higher SDI quintile did not always mean lower IHD deaths (Supplementary Fig. 3 and Supplementary Table 2). In 2019, 15 countries and regions recorded at least 100,000 IHD deaths, with the highest being in China (1,874,000, 95% UI: 1,612,000–2,132,000), India (1,519,000, 95% UI: 1,311,000–1,746,000), the Russian Federation (563,000, 95% UI: 489,000–633,000), the United States (558,000, 95% UI: 497,000–594,000), and Ukraine (326,000, 95% UI: 285,000–372,000), collectively accounting for 52.90% of global IHD deaths.

Between 1990 and 2019, disturbingly high increases in all-age IHD mortality were observed in the Northern Mariana Islands (267.09%, 95% UI: 189.90%–357.40%), Philippines (169.25%, 95% UI: 96.66%–227.81%), Timor-Leste (164.87%, 95% UI: 98.31%–242.31%), Albania (158.39%, 95% UI: 102.31%–226.40%), and China (156.61%, 95% UI: 115.61%–205.68%). Notably, Uzbekistan was the only country in which ASMR increased by more than 100% (119.01%, 95% UI: 96.32%–143.33%).

Regarding net drifts, 64 countries exhibited stable or rising mortality trends (net drift $\ge$–0.50%), despite global declines. Nine countries showed net drift of $\ge$1%, including the Philippines (3.60%, 95% CI: 3.33%–3.86%), Lesotho (2.63%, 95% CI: 2.02%–3.25%), Uzbekistan (1.99%, 95% CI: 1.74%–2.24%), Mozambique (1.78%, 95% CI: 1.58%–1.99%), Zimbabwe (1.61%, 95% CI: 1.35%–1.86%), Dominican Republic (1.51%, 95% CI: 1.30%–1.73%), Timor-Leste (1.48%, 95% CI: 0.77%–2.19%), Kenya (1.37%, 95% CI: 1.15%–1.58%), and Guinea (1.17%, 95% CI: 0.90%–1.45%). A 1% annual increase in net drift predicts increases in the mortality rate of 10%, 18%, and 26% over the next 10-, 20-, and 30-years, respectively. Targeted interventions are therefore urgently needed in these regions.

From 1990 to 2019, IHD mortality shifted towards older populations ($\ge$70 years), particularly in high- and high-middle SDI regions (Fig. 1A). In these areas, women aged $\ge$70 years accounted for over 75% of all IHD deaths in females. This proportion was higher than that observed among men in the same age group. In contrast, more than half of IHD deaths in low-middle and low-SDI countries occurred in individuals aged 70 years, potentially foreshadowing increased future mortality (Supplementary Fig. 4).

Fig. 1.

Local drifts in IHD mortality, and age distribution of deaths by SDI quintiles, 1990-2019. (A) Temporal change in the age distribution of IHD, shown as the proportion of deaths in each age group (0–49, 50–59, 60–69, 70–79, 80–89, 90+ years). (B) Local drifts in IHD mortality, estimated from APC models, for 16 age groups with 5-year increments (15–19 to 90–94 years). The dots and vertical lines indicate the annual percentage change in mortality (% per year) and the corresponding 95% CIs. APC, age–period-cohort; IHD, ischemic heart disease; SDI, socio-demographic index.

APC modeling by age group (local drift) revealed global declines in IHD mortality for all ages, with the greatest reduction of –4.63% (95% CI: –4.75% to –4.51%) observed among older women aged 70–74 years in high-SDI countries (Fig. 1B). No significant sex differences were observed in high- and high-middle SDI countries, indicating uniform improvements in both sexes. However, in the middle-, low-middle, and low-SDI countries, nearly 50% of age groups showed stable or even increasing mortality rates. Interestingly, female mortality in these regions improved more rapidly than male mortality.

APC models separated the individual contributions of age, period, and cohort effects on IHD mortality trends (Fig. 2). Advancing age was consistently associated with increased mortality risk. Significant sex differences emerged after the 20–24 age group (RR = 1.32, z = 3.95, p 0.001) and biggest difference at 45–49 age group (RR = 2.67, z = 46.16, p 0.001), and peaking difference in rates occurs in the 90–94 age group (Fig. 2A). A general decrease in IHD mortality risk was observed across all SDI quintiles as time progressed. However, male populations in middle- and low-middle SDI countries experienced fluctuations above a relative risk of one after 2005 (Fig. 2B). Cohort effects manifested as long-term reductions in most settings, except for persistently elevated risk (RR $>$1) in males from middle-SDI countries until the 2000 birth cohort (RR = 0.95, 95% CI: 0.70–1.29), and in low-middle SDI countries until the 1990 cohort (RR = 0.97, 95% CI: 0.81–1.16) (Fig. 2C). These findings indicate delayed improvement in IHD mortality among men in middle- and low-middle SDI regions.

Fig. 2.

Age, period, and cohort effects on IHD mortality by SDI quintiles. (A) The fitted longitudinal curves show the age effects on mortality, adjusted for period effects. (B) Period effects on mortality are shown by the ratio of age-specific mortality rates, from 1990–1994 to 2015–2019, with 2000–2004 being the reference period. (C) Cohort effects on mortality are shown by the ratio of age-specific mortality rates, from the 1895 cohort to the 2004 cohort, with the 1945–1954 cohort being the reference. The mortality rates or RRs and their corresponding 95% CIs are presented. IHD, ischemic heart disease; SDI, socio-demographic index; RR, rate ratio.

Fig. 3 illustrates APC effects in 10 representative countries spanning the SDI spectrum over the past 30 years. High-SDI countries, such as the United States and Japan, exhibited pronounced shifts in IHD mortality towards older age groups ($\ge$70 years), and consistent reductions in mortality rates across periods and cohorts. High-middle SDI countries, such as Italy and Israel, showed similar patterns. Among middle-SDI countries, China exhibited a trend similar to high- and high-middle SDI countries, while IHD mortality in Indonesia remained more evenly distributed across age groups. Middle-SDI countries exhibited similar trends in local drift, age, and period effects, although sex differences were smaller compared to high- and high-middle SDI countries. Although both countries recorded overall declines, period effects were generally higher for men. Cohort effects in China followed an inverted U-shape, while Indonesia exhibited similar patterns but mainly in females, suggesting heterogeneous improvements.

Fig. 3.

APC effects on mortality in selected countries. APC, age–period–cohort.

Conversely, in low-middle and low-SDI countries, no substantial shift in IHD mortality towards older age groups was observed. The trend even reversed in some cases. For example, local drift trends were consistently down in the Philippines and Pakistan, but up in India and Nepal, indicating rising mortality rates. Birth cohort effects varied widely: the Philippines and Pakistan exhibited upward trends, whereas Nepal showed an “inverted U-shaped” pattern, and India demonstrated a downward trend.

Fig. 4 and Supplementary Fig. 5 show risk factor distributions and their contribution to IHD mortality or DALYs across SDI countries. Over the past 30 years, metabolic factors (notably high low-density lipoprotein (LDL) cholesterol and hypertension) have remained the dominant contributors, despite overall declines in ASMR. Environmental risks, including ambient particulate matter pollution and low temperatures, declined worldwide but remained substantial. Low temperatures emerged as the leading environmental risk in high-SDI countries, while ambient pollution from particulate matter stood out as a key environmental concern in middle- and low-SDI countries. Behavioral factors, such as smoking, unhealthy dietary habits, and low physical activity, continue to contribute to mortality. However, smoking-related mortality declined in high- and high-middle SDI quintiles. Middle- and low-middle SDI countries saw mixed progress, with improved control of some behavioral risks, but sustained burdens of metabolic risks.

Fig. 4.

Major risk factors for global age-standardized IHD death, 1990–2019. (A) Age-standardized mortalities by different risk factors in 1990. (B) Age-standardized mortalities by different risk factors in 2019. Every factor represents a category of risk factor. IHD, ischemic heart disease; LDL, low-density lipoprotein.

This study has several limitations. First, GBD 2019 excludes data before 1990, which limits the assessment of younger cohorts. Second, information on CVD morbidity in low- and middle-income nations has still not been completed to GBD data standards. Third, this analysis did not focus on non-fatal outcomes, which are critical consequences of IHD. Finally, this study utilized GBD 2019 data rather than the most recent GBD database, which extends into the COVID-19 era. The rationale was to avoid the impacts of the COVID-19 pandemic. Previous studies have indicated that COVID-19 influenced the post-2019 cardiovascular mortality risk, altered the mortality patterns, and changed the cause-of-death attribution [31, 32]. By restricting our analyses to the GBD 2019 cycle, we minimized pandemic-related period shocks, methodological complexities, and challenges in longitudinal comparability. Consequently, our findings reflect pre-pandemic trends. Nevertheless, it is likely that COVID-19 reshaped cardiovascular mortality dynamics in the subsequent years, and our future work will aim to replicate and extend the current analyses using post-2019 GBD data to capture potential pandemic-related shifts.