Trends in diabetes and stroke-related mortality in the United States, 1999–2024: a population-based analysis

Abstract

Introduction:

Stroke remains a leading cause of cardiovascular mortality and disability in the United States, with diabetes serving as a major comorbidity that exacerbates outcomes and economic burden. Individuals affected by both diabetes and stroke experience heightened risks of death, disability, and healthcare utilization. This study examines national trends in diabetes and stroke-related mortality from 1999 to 2024, stratified by demographic and geographic factors.

Methods:

Mortality data were extracted from the Centers for Disease Control and Prevention Wide-Ranging Online Data for Epidemiologic Research Multiple Cause of Death database. Deaths were identified using International Classification of Diseases, 10th Revision codes for diabetes mellitus and stroke, defined as deaths in which both conditions were listed on the death certificate, either as underlying or contributing causes. Both crude and age-adjusted mortality rates (AAMRs) per 100 000 were calculated, standardized to the 2000 US population. Temporal patterns were assessed using Joinpoint regression to estimate annual percent change with 95% confidence intervals. Analyses were further stratified by age, sex, race, ethnicity, urbanization, and census region.

Results:

Between 1999 and 2024, 617 059 deaths were attributed to the combined burden of diabetes and stroke. National AAMR declined from 10.4 to 7.0 per 100 000, showing significant decreases until 2018, followed by an increase during 2018–2021 and subsequent decline through 2024. Mortality remained higher among men, non-Hispanic Black individuals, rural populations, and residents of Southern states.

Conclusion:

Although long-term mortality declined, recent reversals and enduring disparities emphasize the urgent need for integrated prevention and equitable public health strategies.

Introduction

In the United States, stroke is among the most common causes of cardiovascular death and disability. Stroke accounted for about one in six deaths from cardiovascular disease (CVD) (17.5%) in 2022^[1]. Strokes are a leading cause of severe long-term disability in the United States, occurring once every 40 s. Over half of stroke survivors who are 65 years of age or older report having less mobility. Even though stroke risk rises with age, it is not just a problem for older adults; in 2014, 38% of stroke hospitalizations occurred in people under 65^[2].

Another significant risk factor for neurovascular complications and CVD is diabetes^[3,4]. An estimated 38.4 million Americans had diabetes in 2021, of whom 29.7 million had a diagnosis and 8.7 million were not diagnosed^[5]. Between 1990 and 2021, the age-standardized prevalence of diabetes (types 1 and 2) rose by 141% to 9001 per 100 000 people^[6]. In addition to its negative health effects, diabetes has a significant financial impact; its estimated $327 billion in 2017 costs included $90 billion in lost productivity and $237 billion in direct medical expenses^[7].

There are significant clinical and economic impacts when diabetes and stroke coexist. In addition to carrying a higher risk of having a stroke, patients with diabetes also exhibit higher stroke-related mortality, more disability, and more difficult management issues than patients without diabetes^[8]. The additional burden of hyperglycemia, cardiovascular comorbidities, and renal dysfunction in diabetic stroke patients results in significantly higher hospitalization costs^[9].

Although stroke and diabetes have been extensively studied as separate drivers of cardiovascular morbidity and mortality, their combined impact remains underexplored in longitudinal population-level analyses. This study addresses that gap by jointly examining diabetes- and stroke-related mortality trends in the United States from 1999 to 2024. By integrating these conditions into a unified analytic framework and stratifying results by age, sex, race, and region, the investigation reveals intersectional disparities that are often overlooked in disease-specific research. This approach offers a more comprehensive understanding of how overlapping cardiometabolic risks shape mortality patterns across diverse demographic groups.

Methods

Data source and study population

We conducted a population-level analysis using the Centers for Disease Control and Prevention (CDC) Wide-ranging Online Data for Epidemiologic Research (WONDER) system. Mortality records were retrieved from the Multiple Cause of Death Public Use database for 1999–2024. To capture the combined burden of diabetes and stroke, we identified deaths in which both diabetes mellitus (ICD-10 codes E10–E14) and stroke, including subarachnoid hemorrhage (I60.0–I60.9), intracerebral hemorrhage (I61.0–I61.9), cerebral infarction (I63.0–I63.9), or stroke not specified as hemorrhage or infarction (I64), were listed on the death certificate, whether as underlying or contributing causes of death. No age restrictions were applied. Consistent with the TITAN 2025 recommendations on responsible disclosure of artificial intelligence in academic work, no AI systems contributed to study conception, data retrieval, statistical evaluation, or interpretation of findings; its role was confined exclusively to stylistic editing of the manuscript’s text^[10]. As the dataset is publicly available and fully deidentified, ethical approval from an institutional review board was not required.

Variables and data extraction

Information abstracted from the database comprised year of death, national and state population counts, demographic characteristics, urban–rural classification, place of death, and geographic region. Demographic indicators included age, sex, race, and ethnicity. Age was analyzed both as a continuous variable and in predefined age strata: <1, 1–4, 5–14, 15–24, 25–34, 35–44, 45–54, 55–64, 65–74, 75–84, and ≥85 years. Racial and ethnic groups were categorized as non-Hispanic (NH) White, NH Black or African American, Hispanic or Latino, NH American Indian or Alaska Native, and NH Asian or Pacific Islander. Location of death was classified into hospital, home, hospice, nursing facility, or long-term care setting. Geographical categorization followed the US Census Bureau’s regional division into Northeast, Midwest, South, and West^[11]. Urban–rural status was defined according to the 2013 National Center for Health Statistics classification, distinguishing large metropolitan areas (≥1 million residents), medium/small metropolitan areas (50 000–999 999 residents), and rural counties (<50 000 residents)^[12]. Although the CDC WONDER database provides mortality counts by urban–rural classification through 2024, age-adjusted mortality rates (AAMRs) are only available up to 2020. Therefore, the urbanization analysis was restricted to AAMRs from 1999 to 2020.

HIGHLIGHTS

• Mortality from diabetes and stroke declined overall in the United States between 1999 and 2024.

• Age-adjusted mortality rates increased sharply during 2018–2021 before falling again.

• Men and non-Hispanic Black populations experienced consistently higher mortality.

• The South and rural regions carried the greatest mortality burden across the study period.

• Findings emphasize the need for targeted, equity-driven interventions in high-risk groups.

Statistical analysis

Mortality rates were computed as both crude and AAMR per 100 000 population for the years 1999–2024. Crude rates were derived by dividing the number of deaths by the respective annual US population. AAMRs were standardized to the age distribution of the US population in the year 2000^[13]. Temporal patterns in diabetes- and stroke-related mortality were evaluated using Joinpoint regression software (version 5.4.0.0; National Cancer Institute). This approach applies log-linear modeling to calculate annual percent change (APC) and 95% confidence intervals (CIs), identifying inflection points where statistically significant shifts in trends occur^[14]. Joinpoint regression was selected because it enables the identification of statistically significant time points where mortality trends change direction or magnitude. Unlike conventional linear regression, which assumes a single uniform trend, the Joinpoint approach models data as a series of connected log-linear segments, allowing for flexible detection of non-linear temporal patterns. This method is widely adopted in epidemiological and surveillance research, including by the US National Cancer Institute and CDC, to assess changes in mortality and disease incidence trends over time. An APC was regarded as increasing or decreasing when the slope differed significantly from zero based on two-tailed testing, with statistical significance set at P < 0.05.

Results

Between 1999 and 2024, a cumulative total of 617 059 deaths were attributed to the combined burden of diabetes mellitus and stroke across all age groups (Table 1; Supplemental Digital Content Table S1, available at: http://links.lww.com/MS9/B65). Many deaths occurred in hospitals or other healthcare institutions (279 272; 45%), followed by nursing homes or long-term care facilities (175 107; 28%) and private residences (115 191; 19%), with smaller proportions in hospice centers and other settings (Supplemental Digital Content Table S2, available at: http://links.lww.com/MS9/B65).

Table 1.

Characteristics of deaths attributed to diabetes-related stroke in the United States across all age groups, 1999–2024

Variable	Diabetes mellitus and stroke-related deaths	Average AAMRs per 100 000	CMRs per 100 000
Overall population	617 059	7.0 (6.9–7.1)	–
Sexa
Male	289 397	7.8 (7.7–8.0)	–
Female	327 662	6.4 (6.3–6.5)	–
Census regiona
Northeast	93 453	5.4 (5.2–5.6)	–
Midwest	137 850	7.0 (6.8–7.2)	–
South	249 176	7.7 (7.6–7.9)	–
West	136 680	7.3 (7.1–7.5)	–
Race/ethnicitya
NH American Indian or Alaska Native	4597	10.1 (8.5–11.6)	–
NH Asian or Pacific Islander	22 990	7.2 (6.7–7.8)	–
NH Black or African American	109 854	14.1 (13.6–14.5)	–
NH White	422 180	6.0 (5.9–6.1)	–
Hispanic or Latino	56 078	8.8 (8.4–9.2)	–
Ageb
5–14 years	16	–	0.004
15–24 years	17	–	0.01
25–34 years	82	–	0.1
35–44 years	751	–	0.52
45–54 years	3472	–	2.2
55–64 years	11 806	–	7.5
65–74 years	36 520	–	21.8
75–84 years	84 374	–	57.8
85+ years	134 310	–	112.7
Urbanizationa,d
Metropolitan	483 896	6.8 (6.7–6.9)	–
Nonmetropolitan	133 163	8.6 (8.4–8.9)	–
Place of deathc
Medical facility	279 272	–	–
Decedent’s home	115 191	–	–
Hospice facility	27 953	–	–
Nursing home/long-term care facility	175 107	–	–
Others	19 536	–	–

AAMRs, age-adjusted mortality rates; CMRs, crude mortality rates.

^aAAMRs per 100 000 population.

^bCMRs used for age-specific groups.

^cAAMRs and CMRs not applicable for place-of-death categories.

^dAAMRs for urbanization represent 1999–2020 data.

Trends in age-adjusted mortality rates (AAMR)

At the national level, the AAMR per 100 000 for deaths attributed to diabetes and stroke declined substantially over the 26-year study period, falling from 10.4 per 100 000 in 1999 to 7.0 in 2024 (Fig. 1; Supplemental Digital Content Tables S3–4, available at: http://links.lww.com/MS9/B65). The reduction was not uniform across time: mortality declined modestly between 1999 and 2002, sharply from 2002 to 2009, and more gradually from 2009 to 2018. A reversal occurred during 2018–2021, when mortality increased, followed by another decline through 2024. Despite these fluctuations, the overall trajectory reflected substantial long-term improvement in population-level mortality.

Figure 1.

AAMRs per 100 000 for diabetes–stroke mortality in the United States, 1999–2024, by sex. Data source: CDC WONDER MCOD, 1999–2024; deaths listing both diabetes (E10–E14) and stroke (I60–I64). Asterisk indicates significant APC (α = 0.05). AAMRs, age-adjusted mortality rates; CDC WONDER MCOD, Centers for Disease Control and Prevention Wide-Ranging Online Data for Epidemiologic Research Multiple Cause of Death; APC, annual percent change.

Sex-specific patterns

Throughout the study period, 253 287 deaths occurred among men and 258 203 among women. Men consistently exhibited higher AAMR per 100 000 than women (7.8 vs. 6.4 per 100 000; Fig. 1; Supplemental Digital Content Tables S3–4, available at: http://links.lww.com/MS9/B65). In men, mortality declined rapidly through 2011, then more slowly until 2018, followed by a sharp rise during 2018–2021 and a renewed decrease thereafter. Among women, mortality fell steadily through 2018, most prominently between 2002 and 2009, rose briefly in 2018–2021, and again declined through 2024. Overall, both sexes followed similar temporal trajectories: two decades of improvement, a short-term interruption, and recovery in recent years.

Race/ethnicity-specific patterns

Marked disparities were evident across racial and ethnic groups (Fig. 2; Supplemental Digital Content Tables S3 and S5, available at: http://links.lww.com/MS9/B65). A total of 95 168 deaths occurred among NH Black or African American individuals, who experienced the highest overall mortality burden (AAMR per 100 000 is 14.1 per 100 000), followed by 6250 deaths among NH American Indian or Alaska Native (AAMR per 100 000 is 10.1), 66 121 among Hispanic or Latino (AAMR per 100 000 is 8.8), 17 894 among NH Asian or Pacific Islander (AAMR per 100 000 is 7.2), and 323 974 among NH White individuals (AAMR per 100 000 is 6.0). Temporal trends were broadly similar across groups, with steady declines through 2017–2018, a sharp increase around 2018–2021, and subsequent declines thereafter. NH Black and NH American Indian populations exhibited the steepest short-term increases but also a notable post-2021 recovery. Despite overall convergence, these two groups continued to experience the highest mortality by 2024.

Figure 2.

Trends in AAMRs per 100 000 for diabetes–stroke mortality in the United States, 1999–2024, by race/ethnicity. Data source: CDC WONDER MCOD, 1999–2024; deaths listing both diabetes (E10–E14) and stroke (I60–I64). Asterisk indicates significant APC (α = 0.05). NH, non-Hispanic; AAMRs, age-adjusted mortality rates; CDC WONDER MCOD, Centers for Disease Control and Prevention Wide-Ranging Online Data for Epidemiologic Research Multiple Cause of Death; APC, annual percent change.

Geographic variation

Substantial geographic variability was observed across the United States (Fig. 3; Supplemental Digital Content Table S6, available at: http://links.lww.com/MS9/B65). Between 1999 and 2020, AAMR per 100 000 ranged from 3.9 in Arizona to 12.1 in Mississippi, and between 2.7 in Connecticut and 16.1 in Oklahoma during 2021–2024, indicating a persistent multi-fold difference between states.When stratified by census region, the South accounted for 209 251 deaths, representing the greatest mortality burden (AAMR per 100 000 is 7.7; 95% CI: 7.6–7.9), followed by 118 500 deaths in the West (AAMR per 100 000 is 7.3), 98 458 in the Midwest (AAMR per 100 000 is 7.0), and 85 281 in the Northeast (AAMR per 100 000 is 5.4) (Fig. 4; Supplemental Digital Content Tables S3 and S7, available at: http://links.lww.com/MS9/B65). Each region experienced long-term declines through 2018, a brief increase during 2018–2021, and renewed declines through 2024. The South showed the highest absolute mortality throughout, while the Northeast maintained the lowest, demonstrating persistent regional inequality.

Figure 3.

State-level AAMRs per 100 000 for diabetes–stroke mortality: (A) United States, 1999–2020; (B) United States, 2021–2024. Data source: CDC WONDER MCOD; deaths listing both diabetes (E10–E14) and stroke (I60–I64). AAMRs, age-adjusted mortality rates; CDC WONDER MCOD, Centers for Disease Control and Prevention Wide-Ranging Online Data for Epidemiologic Research Multiple Cause of Death.

Figure 4.

Regional AAMRs per 100 000 for diabetes–stroke mortality in the United States, 1999–2024, by US Census region. Data source: CDC WONDER MCOD, 1999–2024; deaths listing both diabetes (E10–E14) and stroke (I60–I64). Asterisk indicates significant APC (α = 0.05). AAMRs, age-adjusted mortality rates; CDC WONDER MCOD, Centers for Disease Control and Prevention Wide-Ranging Online Data for Epidemiologic Research Multiple Cause of Death; APC, annual percent change.

Urbanization

When examined by urban–rural classification, 424 686 deaths occurred in metropolitan areas and 86 804 in nonmetropolitan areas. Mortality was consistently higher in nonmetropolitan regions (AAMR per 100 000 is 8.6) compared with metropolitan ones (6.8) (Fig. 5; Supplemental Digital Content Tables S3 and S8, available at: http://links.lww.com/MS9/B65). Nonmetropolitan areas showed a steep decline between 2002 and 2010, slower improvement through 2018, a sharp rise during 2018–2021, and renewed decreases thereafter. Metropolitan areas followed a comparable pattern, with the most substantial reductions from 2002 to 2009, an uptick during 2018–2021, and a subsequent decline. Despite these parallel trajectories, the mortality gap between urban and rural populations persisted throughout the study period, underscoring enduring disparities in outcomes related to diabetes–stroke comorbidity.

Figure 5.

AAMRs per 100 000 for diabetes–stroke mortality in the United States, 1999–2024, by urbanization level. Data source: CDC WONDER MCOD, 1999–2024; deaths listing both diabetes (E10–E14) and stroke (I60–I64). Asterisk indicates significant APC (α = 0.05). AAMRs, age-adjusted mortality rates; CDC WONDER MCOD, Centers for Disease Control and Prevention Wide-Ranging Online Data for Epidemiologic Research Multiple Cause of Death; APC, annual percent change.

Discussion

Mortality from diabetes mellitus and stroke declined steadily from 1999 to 2018, followed by a sharp increase between 2018 and 2021, and a partial reversal through 2024. This temporal pattern was consistent across most demographic and geographic subgroups, though the magnitude of change varied. Rather than restating subgroup-specific trends, we focus here on interpreting the underlying drivers of these disparities.

Both ischemic and hemorrhagic subtypes of stroke are impacted by diabetes, which is a known modifiable risk factor for the condition^[15,16]. Large-artery atherosclerosis, peripheral arterial disease, and obesity are especially linked to type II diabetes, which accounts for about 90% of all cases of the disease. On the other hand, peripheral arterial disease and coronary heart disease (CHD) are typically more closely associated with type I diabetes^[17]. Data from the Emerging Risk Factors Collaboration underscore the scale of this association, showing hazard ratios of 2.27 for ischemic stroke, 1.56 for hemorrhagic stroke, and 1.84 for unclassified stroke among individuals with diabetes^[18]. These findings emphasize the biological pathways through which chronic hyperglycemia accelerates vascular injury and thrombosis, thereby magnifying stroke risk.

The early 2000s decline in stroke mortality is considered one of the decade’s major public health successes. Stroke dropped from the third to the fourth most common cause of death between 2002 and 2009^[19]. This accomplishment was fueled by improvements in acute treatments, medications, and the general quality of care, as well as a decrease in smoking, high cholesterol, and uncontrolled hypertension^[20,21]. Increased statin use, better hypertension control through population screening, and increased awareness of stroke warning signs through national campaigns were additional factors^[22]. The sustained success of these interventions illustrates how comprehensive prevention strategies, when equitably implemented, can significantly alter population-level outcomes. However, uneven access to these benefits across communities limited the overall equity of these achievements. Disparities persisted, though, in spite of these advancements at the population level. Despite a 23% decrease in age-adjusted rates, stroke remained the second most common cause of death among Black adults, while among White adults, it dropped to the fourth most common cause of death after a 26% decline^[23]. This contrast underscores how social and structural inequities continued to shape health outcomes despite overall national progress. Stroke became the fifth most common cause of death in the country by 2023^[2].

There were clear and complex sex differences. AAMR per 100 000 were consistently higher in men, even though women had a higher absolute number of deaths due to their longer average life expectancy. Despite lower relative rates, women make up a larger share of the oldest age groups, which helps explain why they have higher overall death rates. Male-to-female stroke mortality rate ratios are roughly 1.5–2 across the majority of adult age groups, according to numerous studies that have demonstrated that stroke mortality rates are higher in men than in women until very advanced age^[24]. Despite being relatively protected during midlife, women are more likely to experience stroke later in life, while men typically develop CVD at a younger age and have a higher propensity to develop CHD^[25,26]. Therefore, although middle-aged men have higher mortality rates from stroke and CHD than women, women’s risk increases significantly as they age, reducing – but not eliminating the sex gap in older populations. These observations likely reflect both biological mechanisms, such as hormonal influences and vascular aging, and gendered differences in risk factor exposure, treatment adherence, and healthcare utilization.

Beyond these variations, social and structural factors significantly influenced the course of strokes. A higher prevalence of risk factors and worse disease management were consistently linked to adverse social determinants of health (SDOH), which include racism, low income, low educational attainment, neighborhood disadvantage, and limited access to healthcare^[27]. The southern US region known as the Stroke Belt has long had the highest rates of stroke mortality. The Diabetes Belt, which has the highest national prevalence of diabetes, and this region has substantial overlap^[28,29]. The dual burden of diabetes and stroke in these regions is increased by shared risk factors like obesity, high blood pressure, and limited access to healthcare. These overlapping geographies reflect the compounded effects of poverty, limited preventive infrastructure, and historical inequities in public health investment, illustrating that disparities are deeply rooted in place-based determinants rather than individual behaviors alone. Similarly, compared to urban dwellers, mortality rates were consistently higher in rural populations. The higher rates of obesity and metabolic syndrome in rural areas^[30,31], as well as the lack of access to healthcare, food security, and diabetes self-management education programs^[32], can all be blamed for this discrepancy.

The COVID-19 pandemic temporally aligned with a reversal in mortality trends from 2018 to 2021. During this period of disruption, diabetes, particularly type II, was frequently present among hospitalized COVID-19 patients and was linked to increased risk of adverse outcomes and death^[33]. Uncontrolled hyperglycemia, defined as blood glucose levels exceeding 180 mg/dl within a 24-h period, was associated with longer hospital stays and elevated mortality^[34]. Hypertension, a common comorbidity of diabetes, was also correlated with an estimated ~2.5-fold increased risk of severity and death among COVID-19 patients^[35]. These overlapping risk factors and comorbidities may help contextualize the elevated mortality observed among individuals vulnerable to both stroke and diabetes during this time. Importantly, the pandemic period coincided with a widening of racial disparities, further compounding long-standing inequities.

Age-adjusted stroke mortality rates among NH Black or African American (Black) adults were still higher than those among NH White (White) adults, despite a decrease in stroke mortality rates since the 1950s^[1]. Disparities still exist despite intervention efforts to lessen racial disparities in stroke prevention and treatment by lowering risk factors, raising awareness of stroke symptoms, and enhancing access to care and treatment for stroke^[36]. NH whites benefited more from these trends than people of other races and ethnicities, according to evidence^[37].

Not only did the number of stroke deaths from diabetes rise during the pandemic, but racial groups were disproportionately affected. The absolute difference in age-adjusted stroke death rates between Black and White adults during the pandemic increased by 21.7% from 31.3 to 38.0 deaths per 100 000, according to an analysis of National Vital Statistics System data for adults aged ≥35 years^[38]. These results imply that the pandemic worsened pre-existing racial disparities in addition to increasing overall stroke mortality; this pattern is reflected in the analysis of combined diabetes–stroke mortality. Many stroke risk factors, including obesity, diabetes mellitus, smoking, and hypertension, are most prevalent in American Indian or Alaskan Native populations^[39].

Despite these figures, public health studies on the incidence and mortality of strokes do not adequately include American Indian or Alaska Native people^[40]. However, compared to other racial and ethnic groups in the United States, Asian Americans have lower mortality rates, which can be partially attributed to lower rates of smoking and obesity. Vegetables, soy, and fish are the mainstays of traditional Asian diets, which are linked to lower cardiovascular risk and generally better health outcomes. Furthermore, Asian Americans who immigrate frequently have better health profiles than their counterparts who were born in the United States, which is a result of both selection and healthier lifestyle choices. Because of this, Asian Americans have the longest life expectancy of any racial or ethnic group in the United States^[41–43].

This study extends national mortality surveillance by jointly analyzing diabetes- and stroke-related deaths over a 26-year period using Joinpoint regression, a method that identifies statistically significant inflection points in long-term trends. Unlike prior national reports that typically present descriptive summaries, this analysis captures periods of acceleration, stagnation, and reversal in mortality trajectories, particularly the sharp rise during 2018–2021 and the subsequent recovery through 2024. By incorporating data through the post-pandemic era and stratifying results by sex, race/ethnicity, geography, and urbanization, the study provides the most up-to-date and comprehensive picture of evolving disparities in cardiometabolic mortality across the United States.

We present a comprehensive evaluation of mortality trends over a 26-year period using a large nationwide mortality database, with stratification by demographic and geographic variables. Nevertheless, certain limitations warrant consideration. First, as this is an ecological analysis based on aggregated data, causal inferences cannot be drawn, and within-group variability may not be fully captured. Given the reliance on a population-level database, individual-level data were unavailable, thereby precluding a more refined assessment of risk factors and the development of personalized treatment strategies. Second, the inability to adjust for individual-level factors such as smoking, blood pressure control, or medication adherence may have introduced residual confounding, potentially biasing estimates in either direction.

Third, several sources of potential misclassification should be acknowledged. Because this analysis relied on ICD-coded cause-of-death data from the CDC WONDER system, there is a risk of misclassification between diabetes listed as an underlying versus contributing cause of death, as well as inaccuracies in the identification of stroke as the primary cause. Race and ethnicity reporting on death certificates is also subject to misclassification, which may vary across regions and demographic groups. Additionally, shifts in population denominators following the 2020 Census could have influenced rate calculations, particularly for smaller subgroups or geographic areas. Changes in the distribution of deaths by place of occurrence during the COVID-19 pandemic (e.g., hospital vs. home) may further have affected classification and reporting accuracy.

In addition, data corresponding to the COVID-19 era should be interpreted with caution, as the disruption of healthcare access during that time may have led to underreporting of cases. Coding practices may vary across states and over time, leading to differential classification of primary and contributing causes of death. Regional differences in death certification accuracy and reporting completeness could also have introduced variability, particularly between metropolitan and nonmetropolitan areas. Despite these limitations, the use of a standardized national dataset ensures consistency in data collection and comparability across years.

Lastly, sensitivity analyses excluding the COVID-19 pandemic years were not performed. The primary objective was to characterize overall national mortality trends across the full 1999–2024 period. Because Joinpoint regression identifies statistically significant inflection points, any deviations in trend associated with pandemic-related disruptions would be reflected as changes in slope or segment without the need for separate exclusionary analyses. Excluding these years could compromise temporal continuity and obscure the broader long-term patterns of mortality under study.

Conclusion

When combined, these results highlight how significant and fragile the progress made between 1999 and 2018 was. Long-standing gains can be quickly reversed when systemic inequities intersect with emerging health risks, as demonstrated by the COVID-19 pandemic. Persistent disparities by sex, race, socioeconomic status, and geography underscore the urgent need for multimodal interventions that integrate medical management of diabetes and stroke with policies addressing structural inequities and broader SDOH. Strengthening primary prevention through community-based screening, affordable access to antihypertensive and glucose-lowering therapies, and public health campaigns promoting healthy lifestyles remain key priorities. Expanding preventive and specialty care infrastructure in underserved and rural areas, investing in workforce development to support equitable care delivery, and ensuring the inclusion of high-risk populations in national surveillance systems are also essential. Finally, policy actions aimed at reducing poverty, improving food and housing security, and expanding Medicaid coverage could yield sustained improvements in cardiovascular and metabolic outcomes nationwide.

Ethical approval

This investigation utilized anonymized, publicly available data from the CDC WONDER database. As no direct human participation was involved, review by an institutional ethics committee and informed consent were not required.

Consent

Because the dataset contained no identifiable or individual-level information, the need for patient consent for publication does not apply.

Sources of funding

The research was carried out without external funding support from governmental agencies, commercial entities, or nonprofit organizations.

Author contribution

I. M.: Conceptualization, Methodology, Data Curation, Formal analysis, Writing - Original Draft, Supervision, and Visualization. K.K.: Conceptualization, Methodology, Data Curation, Formal analysis, Writing - Original Draft. H.K.: Conceptualization, Methodology, Data Curation, Formal analysis, and Writing - Original Draft. N.B.: Methodology, Data Curation, Formal analysis, and Writing - Original Draft. F.R.: Data Curation, Formal analysis, and Writing - Original Draft. S.K: Data Curation, Formal analysis, and Writing - Original Draft. F.S.: Data Curation and Writing - Original Draft. H.A.: Data Curation and Writing - Original Draft. F.S.: Data Curation and Writing - Original Draft. S.R.: Data Curation and Writing - Original Draft. S.S.S.: Data Curation and Writing - Original Draft. S.A.Y.: Data Curation and Writing - Original Draft.

Research registration unique identifying number (UIN)

This investigation utilized only publicly accessible, de-identified data obtained from the CDC WONDER (Centers for Disease Control and Prevention Wide-Ranging Online Data for Epidemiologic Research) database. Because it did not involve direct interaction with human participants or the collection of personally identifiable information, it does not meet the criteria for human subjects research as defined by the U.S. Department of Health and Human Services. Accordingly, registration in clinical trial or human subjects research registries is not applicable, and assignment of a Unique Identifying Number (UIN) is neither required nor relevant to this study.

Guarantor

Salih Abdella Yusuf.

Conflicts of interest disclosure

The authors confirm that they have no financial relationships, personal ties, or academic interests that could be perceived as having influenced the study’s design, execution, or reporting.

Provenance and peer review

This article was submitted independently and has undergone external peer review.

Data availability statement

All data employed in this study, together with the complete analytical framework, are incorporated within the manuscript and its supplementary materials.