Trends in concomitant cerebrovascular disease and coagulation disorders–related mortality in the United States, 1999–2020

Abstract

Background

See More

Cerebrovascular disease (CVD) remains a leading cause of death in the United States. Although the role of conventional stroke risk factors is well established, the impact of coagulation disorders—both inherited and acquired—on long-term CVD mortality remains underexplored.

Methods

We conducted a cross-sectional analysis using U.S. death certificate data from the CDC WONDER platform (1999–2020). Adults aged ≥ 25 years with CVD as the underlying cause and any coagulation disorder listed as a contributing cause were included. Age-adjusted mortality rates (AAMRs) were calculated per 100,000 population. Joinpoint regression was used to evaluate temporal trends, estimating Annual Percent Change (APC) and Average Annual Percent Change (AAPC) across subgroups.

Results

Between 1999 and 2020, 54,545 CVD-related deaths occurred among adults with coagulation disorders. The overall AAMR was 1.16, with a significant decline over time (AAPC: − 0.58%, 95% CI: − 0.87 to − 0.28; p = 0.0010). Males had higher mortality than females (1.35 vs. 1.01), but females showed greater declines (AAPC: − 0.96% vs. − 0.32%). Black and American Indian individuals experienced the highest rates but also the steepest improvements (AAPCs: − 1.87% and − 2.02%, respectively). In contrast, Hispanic populations showed early declines followed by recent increases. Only the Northeast region had a statistically significant mortality reduction. Rural residents and those in the South had stagnant trends. Mortality rose sharply with age, peaking in adults ≥ 85 years, although older groups also showed significant declines. Most deaths (77%) occurred in inpatient settings.

Conclusions

Although overall CVD mortality declined modestly, widening disparities by sex, race, region, and age group signal critical gaps in prevention and care.

Introduction

Cerebrovascular disease (CVD), including ischemic and hemorrhagic strokes, remains a leading cause of death and long-term disability in the United States, despite decades of public health efforts and clinical advances [1, 2]. While the influence of conventional cardiovascular risk factors—such as hypertension, diabetes, and atrial fibrillation—is well recognized, the potential role of hematologic abnormalities, particularly coagulation disorders, in shaping stroke mortality patterns has received comparatively limited attention.

Coagulation disorders represent a diverse group of conditions that disrupt the balance between bleeding and thrombosis. Inherited forms such as hemophilia A, hemophilia B, and von Willebrand disease have traditionally been associated with hemorrhagic complications. However, accumulating evidence shows that individuals with these disorders may also experience thrombotic events, including ischemic stroke, particularly as they age and accumulate vascular comorbidities [3, 4]. Acquired coagulation abnormalities—such as those associated with malignancy, sepsis, autoimmune conditions, or chronic liver disease—further complicate this picture, often manifesting with simultaneous risks of thrombosis and bleeding [5].

Pathophysiologic studies have demonstrated that abnormal activation or consumption of clotting factors may contribute to cerebrovascular injury. For example, elevated levels of factor VIII and fibrinogen have been independently associated with increased stroke risk [6], while microvascular damage and platelet activation seen in hemorrhagic diatheses may accelerate vascular injury or exacerbate underlying stroke pathology [7]. These mechanisms suggest that both thrombotic and hemorrhagic states—when occurring in the context of coagulation disorders—may increase vulnerability to fatal CVD outcomes.

Despite these insights, long-term national data examining trends in CVD-related mortality among individuals with coagulation disorders are lacking. Most existing studies focus on clinical outcomes within narrow subgroups or investigate stroke risk among patients with specific bleeding disorders, limiting generalizability to broader populations [8]. In light of this gap, a comprehensive population-level analysis is needed to assess the contribution of coagulation disorders to stroke-related mortality over time.

To address this need, we conducted a 22-year nationwide analysis of U.S. mortality data, focusing on adults with coexisting cerebrovascular disease and underlying coagulation disorders. By evaluating trends across demographic, geographic, and clinical subgroups, this study provides new insights into a high-risk and understudied population, with implications for risk stratification, health policy, and future research.

Methods

Study design

We conducted a cross-sectional analysis of U.S. mortality data using the Centers for Disease Control and Prevention (CDC) Wide-ranging Online Data for Epidemiologic Research (WONDER) platform [9]. This publicly accessible source provides de-identified, population-based death certificate data for all U.S. residents. Our study covered the years 1999 through 2020 and focused on adults aged 25 years and older.

Deaths were selected using the International Classification of Diseases, 10th Revision (ICD-10) [10]. Cerebrovascular disease (CVD) was defined using codes I60–I69, capturing both hemorrhagic and ischemic subtypes. We then identified cases where any coagulation disorder was listed as a contributing cause of death using codes D65 (disseminated intravascular coagulation), D66–D68 (hereditary/acquired coagulation defects), and D69 (purpura and other hemorrhagic conditions). Only records listing a CVD code as the underlying cause of death were included in the final dataset (Supplementary Table S1).

This study relied exclusively on publicly available, de-identified data and was therefore exempt from Institutional Review Board oversight. We followed the STROBE guidelines for reporting observational studies (Supplementary Table S2) [11].

Data extraction

Data were extracted for the total adult population and stratified by sex, race/ethnicity, ten-year age groups (25–34 through ≥ 85), U.S. Census region (Northeast, Midwest, South, West), urbanization level (metropolitan vs. non-metropolitan), state of residence, and place of death.

Race and ethnicity were recorded based on information typically provided by next of kin and reported by funeral directors. Urbanization level followed the 2013 National Center for Health Statistics (NCHS) Urban–Rural Classification Scheme [12]. Counties with ≥ 50,000 population were classified as metropolitan; all others were non-metropolitan. Consistent with CDC WONDER guidance, we excluded annual data cells with fewer than 10 deaths to prevent unstable estimates. In some subgroups—particularly younger age groups and smaller racial/ethnic populations—this suppression limited longitudinal trend analyses.

Statistical analysis

Annual death counts and population estimates were extracted from CDC WONDER. Age-adjusted mortality rates (AAMRs) were calculated using the direct method and standardized to the 2000 U.S. standard population. AAMRs were expressed per 100,000 population, with 95% confidence intervals (CIs) calculated using Poisson distribution assumptions.

Temporal trends were assessed using Joinpoint regression, applying log-linear models to estimate Annual Percent Change (APC) for each identified segment and Average Annual Percent Change (AAPC) across the full study period. The Joinpoint Regression Program (version 5.0.2) from the National Cancer Institute was used to perform the analysis [13]. The model began with a single linear segment and added joinpoints where statistically supported using the Monte Carlo permutation test [14].

The number and location of joinpoints were selected empirically based on best model fit. A p-value < 0.05 was considered statistically significant. Subgroup-specific models were fit separately for sex, race/ethnicity, age group, geographic region, urbanization level, and state of residence. For each segment, APCs and corresponding 95% CIs were reported. Tests of parallelism were used to assess whether trends differed significantly across subgroups. All supplementary analyses and visualizations were conducted in Python (v3.11) using the statsmodels, scipy, and matplotlib libraries.

Complete Joinpoint regression outputs (segment years, APCs with 95% CIs and p-values, and AAPCs) are provided in (Supplementary Tables S3–S7).

Results

Overall trends

From 1999 to 2020, there were 54,545 cerebrovascular disease (CVD)–related deaths recorded among U.S. adults with documented coagulation disorders. The overall age-adjusted mortality rate (AAMR) was 1.16 per 100,000 (95% CI, 1.13–1.19) (Table 1). Joinpoint regression identified three periods: a modest, non-significant decline in 1999–2005 (APC, − 0.72%; 95% CI, − 1.37 to − 0.07; p = 0.0811), followed by slight, non-significant increases in 2006–2012 (APC, 0.70%; 95% CI, − 0.27 to 1.68; p = 0.2151) and 2013–2020 (APC, 0.87%; 95% CI, − 0.69 to 2.46; p = 0.3156). Across the full period, the average annual percent change (AAPC) indicated a significant decline (AAPC, − 0.58%; 95% CI, − 0.87 to − 0.28; p = 0.0010) (Table 2; Fig. 1).

Fig. 1

Trends in cerebrovascular disease–related mortality with coagulation disorders stratified by sex and overall, United States, 1999–2020

Table 1 Demographic characteristics of decedents with cerebrovascular disease as the underlying cause of death and coagulation disorders as contributing causes, united States, 1999–2020

Table 2 Annual percentage changes (APCs) and average annual percentage changes (AAPCs) in cerebrovascular disease–related mortality among adults with coagulation disorders, united States, 1999–2020

Trends by sex

Across the study period, 27,826 deaths occurred in males and 26,719 in females. Males had consistently higher mortality (mean AAMR, 1.35; 95% CI, 1.32–1.38) than females (mean AAMR, 1.01; 95% CI, 0.98–1.05) (Table 1). Mortality declined significantly in both sexes, with a larger decrease in females (AAPC, − 0.96%; 95% CI, − 1.31 to − 0.61; p < 0.0001) than in males (AAPC, − 0.32%; 95% CI, − 0.60 to − 0.03; p < 0.0001). Among females, a significant decline was noted in 1999–2005 (APC, − 0.90%; 95% CI, − 1.63 to − 0.16), followed by non-significant stability (2006–2012) and a slight, non-significant increase (2013–2020). In males, segment-specific APCs were small and non-significant throughout (Table 2; Fig. 1).

Trends by race/ethnicity

Mortality varied by race/ethnicity. Black and White individuals had the highest AAMRs; Asian or Pacific Islander and Hispanic/Latino populations had lower rates (Table 1). Overall trends declined significantly for Black (AAPC, − 1.87%; 95% CI, − 2.51 to − 1.22; p < 0.0001), Asian or Pacific Islander (AAPC, − 1.68%; 95% CI, − 2.49 to − 0.86; p = 0.0007), and White individuals (AAPC, − 0.41%; 95% CI, − 0.71 to − 0.11; p = 0.0147). Among Hispanic/Latino individuals, mortality declined early (APC, − 3.58%; 95% CI, − 5.66 to − 1.45) and rose in later years (APC, 2.73%; 95% CI, 0.36–5.15), yielding an overall significant decline (AAPC, − 0.74%; 95% CI, − 1.33 to − 0.16; p = 0.0217). For American Indian/Alaska Native individuals, data sparsity limited segment analysis, but the overall trend declined significantly (AAPC, − 2.02%; 95% CI, − 3.12 to − 0.91; p = 0.0074) (Table 2; Fig. 2).

Fig. 2

Trends in cerebrovascular disease–related mortality with coagulation disorders stratified by race/ethnicity, United States, 1999–2020

Trends by urbanization status

Non-metropolitan areas had a slightly higher mean AAMR than metropolitan areas (1.21 [95% CI, 1.18–1.24] vs. 1.14[95% CI, 1.11–1.17]) (Table 1). In metropolitan areas, segment-specific changes were not significant, but the overalltrend declined significantly (AAPC, − 0.61%; 95% CI, − 0.93 to − 0.30; p = 0.0010). In non-metropolitan areas, segment APCs and the overall trend were not significant (AAPC, − 0.31%; 95% CI, − 0.78 to 0.16; p = 0.1886) (Table 2; Fig. 3).

Fig. 3

Trends in cerebrovascular disease–related mortality with coagulation disorders stratified by urbanization level (metropolitan vs. non-metropolitan), United States, 1999–2020

Trends by census region

The West had the highest mean AAMR (1.30; 95% CI, 1.27–1.34), followed by the South (1.17), Midwest (1.09), and Northeast (1.02) (Table 1). The Northeast showed a significant early decline (1999–2005; APC, − 2.28%; 95% CI, − 3.56 to − 0.99) and a significant overall decrease (AAPC, − 0.90%; 95% CI, − 1.44 to − 0.36; p = 0.0038). In the Midwest, mortality rose slightly early and declined significantly in 2013–2020; the overall decrease approached significance (AAPC, − 0.44%; 95% CI, − 0.88 to 0.00; p = 0.0507). The South was largely stable with a non-significant overall trend (AAPC, − 0.40%; 95% CI, − 0.94 to 0.13; p = 0.1339). In the West, a significant decline occurred in 2006–2012 (APC, − 1.55%; 95% CI, − 2.88 to − 0.21), while the overall trend was not significant (AAPC, − 0.39%; 95% CI, − 0.97 to 0.19; p = 0.1799) (Table 2; Fig. 4).

Fig. 4

Trends in cerebrovascular disease–related mortality with coagulation disorders stratified by U.S. census region, 1999–2020

Trends by state

State-level AAMRs varied widely. The highest mean rates occurred in Hawaii (1.87; 95% CI, 1.69–2.05), the District of Columbia (1.61; 95% CI, 1.34–1.88), and North Dakota (1.53; 95% CI, 1.29–1.76). The lowest were in Louisiana(0.74; 95% CI, 0.67–0.81), New York (0.82; 95% CI, 0.79–0.86), and Virginia (0.94; 95% CI, 0.89–1.00) (Fig. 5).

Fig. 5

Geographic distribution of cerebrovascular disease–related mortality with coagulation disorders by state, United States, 1999–2020

Trends by ten-year age group

Mortality increased steeply with age. Mean AAMRs were highest in adults ≥ 85 years (7.01; 95% CI, 6.85–7.17), followed by 75–84 years (4.75; 95% CI, 4.67–4.83) and 65–74 years (2.42; 95% CI, 2.38–2.46). The lowest rates were in 25–34 years (0.11; 95% CI, 0.11–0.11) and 35–44 years (0.28; 95% CI, 0.26–0.30) (Table 1). Significant long-term declines were observed for 65–74 years (AAPC, − 1.06%; 95% CI, − 1.42 to − 0.70; p < 0.0001), 75–84 years (AAPC, − 0.71%; 95% CI, − 1.08 to − 0.35; p = 0.0011), and ≥ 85 years (AAPC, − 0.65%; 95% CI, − 1.02 to − 0.27; p = 0.0028). Younger adults (25–54 years) showed no significant long-term change; among 25–34 years, a non-significant increase appeared in 2013–2020 (APC, 4.74%; 95% CI, − 1.40 to 11.26; p = 0.1839) (Table 2; Fig. 6).

Fig. 6

Crude mortality rates for cerebrovascular disease–related deaths with coagulation disorders stratified by ten-year age group and sex, United States, 1999–2020

Trends by place of death

Most deaths occurred in inpatient medical facilities (77.08%). Home was the next most common location (7.00%), followed by outpatient or emergency department settings (3.66%). Small proportions were dead on arrival (0.10%) or had an unknown place of death (0.10%) (Table 1).

Discussion

In this national analysis from 1999 to 2020, we identified 54,545 cerebrovascular disease (CVD)–related deaths among adults with coagulation disorders in the United States. The overall age-adjusted mortality rate (AAMR) was 1.16 per 100,000 population, with a modest but statistically significant decline over time (AAPC: − 0.58%, 95% CI: − 0.87 to − 0.28; p = 0.0010). Joinpoint regression revealed three phases: a decline from 1999 to 2005, followed by stabilization and a gradual increase through 2020, though later trends were not statistically significant.

Sex-based disparities were persistent. Men had higher mortality than women (AAMR: 1.35 vs. 1.01), though both groups improved. The sharper decline in females (AAPC: − 0.96% vs. − 0.32%) may reflect stronger engagement with preventive care and primary care access among women [15,16,17]. This pattern is consistent with broader stroke literature, where women tend to exhibit better health-seeking behavior and adherence to preventive therapies [18, 19]. Racial and ethnic disparities were similarly prominent. Black and American Indian or Alaska Native individuals had the highest mortality rates but also experienced the steepest declines (AAPC: − 1.87% and − 2.02%, respectively). These patterns mirror longstanding disparities in stroke outcomes linked to structural inequities, limited healthcare access, and higher vascular risk [20,21,22,23]. In contrast, Hispanic populations showed early gains but recent worsening—echoing national data on rising stroke incidence in younger Hispanic adults [21]. Asian or Pacific Islander individuals had the lowest mortality and consistent declines over time [24].

Geographic disparities revealed further complexity. The West had the highest overall AAMR (1.30 per 100,000), while the Northeast had the lowest (1.02 per 100,000) and the only significant mortality decline (AAPC: − 0.90%, p = 0.0038). The South—long recognized as the “Stroke Belt”—showed no significant change (AAPC: − 0.40%, p = 0.1339), reflecting persistent challenges in hypertension control, stroke center access, and neurologist density [25, 26]. Medicaid expansion and differential investment in public health likely influence these regional differences [27]. Urban–rural disparities, though numerically small, carried clinical importance. Rural residents had slightly higher AAMRs (1.21 vs. 1.14) and showed no significant improvement over time. In contrast, metropolitan areas experienced significant declines (AAPC: − 0.61%, p = 0.0010), possibly due to better access to neuroimaging, stroke-certified centers, and hematologic expertise [28, 29].

Older adults bore the greatest mortality burden in our cohort, with rates rising steeply across advancing age groups. This age gradient mirrors national stroke patterns in the general population, where risk and mortality accelerate with age [30, 31]. Direct numerical comparison to adults without coagulation disorders is not possible in CDC WONDER because our case definition captures a subset of cerebrovascular deaths in which a coagulation disorder is recorded as a contributing cause. Nevertheless, the convergence of age-related vascular comorbidity with impaired hemostasis plausibly compounds risk and complicates prevention and treatment choices in clinical practice, potentially limiting the translation of system-level stroke improvements to this subgroup [15, 30, 32].

These disparities are compounded by the biological complexity of coagulation disorders. Individuals with hemophilia, thrombocytopenia, or disseminated intravascular coagulation (DIC) are predisposed to hemorrhagic stroke, while those with prothrombotic states—such as antiphospholipid syndrome or malignancy-associated coagulopathy—face heightened ischemic risk [33,34,35]. In older adults with multiple comorbidities, dual risks often coexist, making stroke prevention and treatment highly individualized [36].

Despite therapeutic advances—including recombinant clotting factors, refined transfusion protocols, and direct oral anticoagulants (DOACs)—outcomes remain poor [32, 37]. The underuse of standard stroke therapies in coagulopathic patients may reflect safety concerns and limited clinical guidelines [38, 39]. Over 77% of deaths occurred in inpatient settings, suggesting these patients are often acutely ill and may present late or ineligible for interventions [40].

While national stroke mortality has declined due to improved hypertension control, smoking cessation, and thrombolytic therapy [30, 31], our findings suggest these benefits are not fully realized among those with coagulation disorders. The plateau in mortality after 2005 may reflect rising comorbidity burdens (e.g., cancer, obesity, cirrhosis, and sepsis), all of which are linked to acquired coagulopathy and worse stroke outcomes [41, 42]. Uneven adoption of newer therapies, such as DOACs, and clinical hesitation to use anticoagulation in this group may have further limited progress [39, 43]. Lastly, persistent fragmentation of care and widening social inequalities may dilute gains from earlier stroke initiatives [44]. Efforts to improve stroke recognition, access to specialist care, and guideline development for coagulopathic patients are essential to reduce the continued burden of CVD-related mortality in this vulnerable population.

Beyond being the first nationwide study to specifically evaluate cerebrovascular disease mortality in individuals with coagulation disorders, our findings provide important insights into the intersection of vascular and hematologic risk. Even modest declines or differences in mortality carry clinical and public health significance in a vulnerable population with dual risks of bleeding and thrombosis. By identifying persistent subgroup disparities by sex, race/ethnicity, region, and urbanizationthese results point to structural and biological gaps in prevention and acute care that may blunt population-level gains in stroke outcomes [18, 30]. Collectively, they underscore the need to incorporate hematologic comorbidity into stroke risk stratification, strengthen stroke systems of care, and adapt guidance for patients in whom standard therapies (anticoagulation, thrombolysis, reversal) require nuanced, safety-conscious decision-making [38, 43, 45].

Strengths and limitations

This study offers a comprehensive national assessment of cerebrovascular disease (CVD)–related mortality among adults with coagulation disorders over a 22-year period, using robust data from the CDC WONDER platform. Among its key strengths is the large sample size and population-based design, which enhances generalizability and allows for granular stratification by sex, race/ethnicity, urbanization, region, age group, and place of death. The application of Joinpoint regression enabled the identification of meaningful inflection points in mortality trends, which adds important temporal context to evolving patterns. Furthermore, by focusing on a high-risk and understudied population, this study fills a significant gap in the literature and offers new insights into the intersection of hematologic and vascular disease burden in the U.S.

Despite these strengths, several limitations must be acknowledged. First, the analysis relied on death certificate data, which may be prone to misclassification or underreporting—especially for secondary diagnoses like coagulation disorders [46]. It is possible that some decedents with relevant hematologic conditions were not identified, and conversely, that some recorded cases reflect subclinical or incidental findings. Second, the lack of individual-level clinical data (e.g., laboratory values, medication use, stroke subtype, comorbidities) limits the ability to disentangle causal pathways or assess the appropriateness of treatment. Finally, the use of aggregated ICD-10 codes (D65–D69) captures a wide spectrum of coagulation abnormalities with differing prognoses and mechanisms, which may obscure important subtype-specific trends. However, this broad inclusion also reflects real-world clinical complexity, where multiple overlapping hematologic conditions are common.

Conclusion

Between 1999 and 2020, cerebrovascular disease–related mortality among adults with coagulation disorders declined modestly, yet the trend was uneven across key subgroups. Higher mortality persisted among men, Black and American Indian populations, rural residents, and older adults. In recent years, the overall gains have plateaued or reversed in several groups, signaling emerging gaps in prevention and treatment. These findings highlight the need for more tailored stroke care strategies in individuals with coagulation disorders, especially those at the intersection of social and biological vulnerability. Strengthening access to timely diagnosis, evidence-based interventions, and equitable care delivery will be critical to reducing preventable deaths in this high-risk population.