Temporal trends and disparities in stroke mortality among adults with chronic kidney disease in the United States, 1999-2020

Abstract

BACKGROUND

Chronic kidney disease (CKD) is a major independent stroke risk factor. This study characterizes 22-year national trends and disparities in stroke mortality among United States adults with CKD.

AIM

To evaluate 22-year national trends and demographic disparities in stroke mortality among United States adults with CKD to inform targeted strategies for reducing cerebrovascular risk in this vulnerable population.

METHODS

Using Centers for Disease Control and Prevention Wide-Ranging Online Data for Epidemiologic Research Multiple Cause-of-Death data (1999-2020), we analyzed stroke deaths (underlying cause) with CKD (contributing cause) among adults ≥ 25 years. Age-adjusted mortality rates per 100000 population were calculated. Joinpoint regression estimated annual percentage changes (APCs) and average APCs with 95% confidence intervals, stratified by sex, race/ethnicity, region, and urbanization.

RESULTS

Among 37308 stroke deaths with CKD, the overall age-adjusted stroke mortality rates (AAMR) declined from 1.08 (95%CI: 1.03-1.13) in 1999 to 0.71 (95%CI: 0.68-0.75) in 2020 (average annual percent change: -1.79%). Significant trends included a decline from 1999-2009 (APC: -4.25%), followed by an increase from 2009-2012 (APC: 23.25%), a sharp decline from 2012-2015 (APC: -28.10%), and another increase from 2015-2020 (APC: 8.72%). Males had higher mortality than females (AAMR 0.79 vs 0.71). Non-Hispanic Black individuals had the highest AAMR (1.95), followed by Hispanic (0.87) and Non-Hispanic White individuals (0.63). Regionally, the West had the highest AAMR (0.89). State-level mortality varied more than three-fold (District of Columbia: 1.27 vs Arizona: 0.38). Small metropolitan areas had the highest urbanization-stratified AAMR.

CONCLUSION

While stroke mortality among United States adults with CKD significantly declined over two decades, reflecting improvements in prevention and management, substantial disparities persist. The findings underscore the critical need for targeted public health interventions to address underlying biological, structural, and systemic determinants of cerebrovascular risk in this vulnerable population.

Key Words: Stroke; Chronic kidney disease; Epidemiology; Mortality; CDC WONDER

Core Tip: This national study examines 22-year trends in stroke mortality among United States adults with chronic kidney disease (CKD), a high-risk population often overlooked in cerebrovascular research. Using population-level mortality data, we identified periods of both decline and resurgence in mortality, along with striking disparities by sex, race/ethnicity, region, and urbanization. Non-Hispanic Black individuals and residents of small metropolitan areas experienced the greatest burden. These findings highlight evolving epidemiologic patterns and the urgent need for targeted, equity-driven interventions to reduce stroke mortality in adults with CKD.

INTRODUCTION

Chronic kidney disease (CKD) is a major public health crisis, affecting over 35 million adults in the United States, equivalent to 15% of the adult population with alarmingly low diagnosis rates exceeding 90%[1]. This condition extends beyond renal impairment, functioning as a potent multiplier of cardiovascular and cerebrovascular risk. Stroke, the fifth leading cause of United States mortality in 2021, claims approximately 795000 lives annually while incurring economic costs surpassing $56 billion[2]. Crucially, CKD independently elevates stroke susceptibility through non-traditional pathways, including uremic toxin-mediated endothelial dysfunction, chronic inflammation, and disrupted cerebral autoregulation[3,4].

Robust epidemiological evidence confirms a dose-dependent relationship between CKD severity and stroke incidence. For every 10 mL/minute/1.73 m² decline in estimated glomerular filtration rate, stroke risk increases by 7%[5]. Similarly, each 25 mg/mmoL rise in albumin-creatinine ratio corresponds to an 11% higher stroke hazard[6], persisting after adjustment for hypertension, diabetes, and other conventional risk factors[7]. These associations underscore CKD as a distinct neurovascular risk entity rather than a mere comorbidity.

Despite these established linkages, national mortality assessments in CKD populations remain disproportionately focused on aggregate cardiovascular outcomes. Recent analyses describe only modest 7.1% declines in overall cardiovascular mortality from 1999-2020[8], obscuring stroke-specific trajectories. Furthermore, critical gaps persist regarding how stroke mortality trends vary across racial, geographic, and socioeconomic strata, despite CKD’s disproportionate burden in marginalized communities, including Non-Hispanic Black and rural populations[9].

To address these limitations, we conducted the first nationwide, 22-year analysis (1999-2020) of stroke-specific mortality trends exclusively in CKD patients. Leveraging Centers for Disease Control and Prevention Wide-Ranging Online Data for Epidemiologic Research (CDC WONDER) comprehensive mortality database, this study uniquely quantifies age-adjusted stroke mortality rates (AAMRs) across sex, racial/ethnic, and urbanization subgroups; Identifies inflection points in mortality trends using Joinpoint regression; Examines state-level disparities to guide targeted public health interventions. By elucidating heretofore uncharacterized disparities, our findings aim to catalyze precision interventions for this high-risk population.

MATERIALS AND METHODS

Study setting and population

This study analyzed stroke-related mortality among United States adults aged 25 years and older with CKD from 1999 to 2020 using data from the CDC WONDER platform. Specifically, records were retrieved from the Multiple Cause-of-Death Public Use Files, where stroke was documented as the underlying cause and CKD as a contributing cause of death. CDC WONDER is a comprehensive, publicly accessible source of United States health data available since 1999[10].

Stroke and CKD deaths were identified using the International Classification of Diseases, 10^th Revision, Clinical Modification (ICD-10-CM) codes I60.x, I61.x, I63.x, I64, I69.0, I69.1, I69.3, I69.4, and N18, respectively. Individuals under 25 years of age were excluded due to small case counts in this group. The study followed the Strengthening the Reporting of Observational Studies in Epidemiology guidelines and did not require Institutional Review Board approval, as the data were de-identified and publicly available[11].

Data abstraction

Variables extracted included year of death, sex, race/ethnicity, geographic region, place of death, urbanization category, and United States state. Death locations were classified into four categories: Hospitals, homes, hospices, and nursing home/Long-term care facilities. Urbanization levels were defined using the National Center for Health Statistics urban-rural classification system, including large central metro, large fringe metro, medium metro, small metro, micropolitan, and noncore (non-metro) areas.

Race/ethnicity was categorized as NH Asian or Pacific Islander, Non-Hispanic (NH) White, NH Black or African American, and Hispanic or Latino. NH American Indian or Alaska Native were excluded due to low counts or unreliable data. Regional classification followed the United States Census Bureau definitions for Northeast, Midwest, South, and West.

Statistical analysis

Crude mortality rates per 100000 were calculated annually by dividing the number of stroke deaths among CKD patients by the United States population for that year. AAMRs were standardized using the 2000 United States standard population to allow valid comparisons across time and demographic groups by minimizing age-related confounding.

Temporal trends in AAMRs were assessed using the Joinpoint Regression Program (version 5.0.2), which fits log-linear models to determine annual percentage changes (APCs) and average annual percentage changes, with 95% confidence intervals. The Monte Carlo permutation method was used to identify significant trend changes (joinpoints). An APC was considered statistically significant if its slope differed from zero with a two-sided P value ≤ 0.05[12].

RESULTS

Between 1999 and 2020, a total of 37308 deaths were recorded due to stroke among individuals aged 25 years and older with CKD. Of these, 19440 deaths occurred in females (52.11%) and 17868 in males (47.89%) (Supplementary Table 1).

Annual trends

The age-adjusted mortality rate (AAMR) for stroke among individuals with CKD showed fluctuations over the 22-year study period, declining from 1.08 deaths per 100000 population in 1999 (95%CI: 1.03-1.13) to 0.71 in 2020 (95%CI: 0.68-0.75). The overall average annual percent change was -1.79 (95%CI: -5.85 to 2.45).

Between 1999 and 2009, the AAMR experienced a statistically significant reduction, with an APC of -4.25 (95%CI: -5.43 to -3.06). This was followed by a significant upward trend between 2009 and 2012 (APC: 23.25; 95%CI: 1.24 to 50.05). From 2012 to 2015, a sharp statistically significant decline occurred (APC: -28.10; 95%CI: -44.14 to -7.46). Finally, between 2015 and 2020, the rate increased significantly (APC: 8.72; 95%CI: 3.66-14.04) (Figure 1 and Supplementary Table 2).

Figure 1 Annual trends in stroke age-adjusted stroke mortality rates in patients with chronic kidney disease. APC: Annual percent change; AAMR: Age-adjusted stroke mortality rates.

Gender-stratified trends

Throughout the study period, males exhibited consistently higher AAMRs for stroke with CKD compared to females. The overall AAMR among males was 0.79 per 100000 (95%CI: 0.78-0.80), whereas for females it was 0.71 (95%CI: 0.70-0.72).

For females, the AAMR declined significantly from 1.00 in 1999 (95%CI: 0.94-1.06) to 0.60 in 2020 (95%CI: 0.56-0.64), with a statistically significant overall decreasing trend across the entire period (APC: -2.26; 95%CI: -4.21 to -0.27).

For males, the AAMR decreased from 1.25 in 1999 (95%CI: 1.17-1.34) to 0.92 in 2020 (95%CI: 0.91-0.94), also showing a statistically significant overall decline (APC: -1.89; 95%CI: -3.60 to -0.15). No joinpoints were identified in either group, indicating consistent monotonic trends throughout 1999-2020 (Figure 2 and Supplementary Table 3).

Figure 2 Trends in stroke age-adjusted stroke mortality rates in patients with chronic kidney disease stratified by sex. APC: Annual percent change; AAMR: Age-adjusted stroke mortality rates.

Race-stratified trends

When stratified by race and ethnicity, Black or African American individuals exhibited the highest overall AAMR for stroke with CKD at 1.95 per 100000 population (95%CI: 1.91-2.00), followed by Hispanic or Latino individuals at 0.87 (95%CI: 0.84-0.91), Asian or Pacific Islander individuals at 0.85 (95%CI: 0.81-0.90), and White individuals at 0.63 (95%CI: 0.63-0.64). Asian or Pacific Islander individuals demonstrated a statistically significant monotonic decline over the entire 22-year period (APC: -3.66; 95%CI: -5.97 to -1.29), as did Black or African American individuals (APC: -4.43; 95%CI: -5.93 to -2.91).

In contrast, White individuals experienced significant fluctuations: A decline from 1999-2009 (APC: -3.78; 95%CI: -4.95 to -2.60), followed by an increase during 2009-2012 (APC: 26.23; 95%CI: 6.14-50.13), a sharp decline from 2012-2015 (APC: -28.11; 95%CI: -42.51 to -10.10), and finally an increase from 2015-2020 (APC: 9.19; 95%CI: 4.80-13.78).

Similarly, Hispanic or Latino individuals showed a significant decline (1999-2009: APC: -4.49; 95%CI: -6.12 to -2.83), followed by an increase (2009-2012: APC: 22.19; 95%CI: 1.73-46.76), a steep decline (2012-2015: APC: -37.33; 95%CI: -51.16 to -19.57), and concluded with an increase (2015-2020: APC: 14.49; 95%CI: 9.44-9.78) (Figure 3 and Supplementary Table 4).

Figure 3 Trends in stroke age-adjusted stroke mortality rates in patients with chronic kidney disease stratified by race. APC: Annual percent change; AAMR: Age-adjusted stroke mortality rates.

State and region-based differences

AAMRs demonstrated substantial variation across United States. The highest AAMR was observed in the District of Columbia at 1.27 per 100000 population (95%CI: 1.03-1.52), followed by South Carolina (1.24; 95%CI: 1.16-1.33), West Virginia (1.17; 95%CI: 1.05-1.29), and Mississippi (1.11; 95%CI: 1.01-1.21). In contrast, the lowest rates were recorded in Arizona (0.38; 95%CI: 0.34-0.42), New York (0.40; 95%CI: 0.38-0.42), and Nevada (0.42; 95%CI: 0.35-0.48) (Figure 4 and Supplementary Table 5). A choropleth map (Figure 4) illustrates these geographic disparities.

Figure 4 State-wise map highlighting the states with highest mortality rates in the United States from 1999 to 2020. Source: https://wonder.cdc.gov/.

At the regional level, the West exhibited the highest overall AAMR (0.89; 95%CI: 0.87-0.90), followed by the South (0.84; 95%CI: 0.82-0.85), Midwest (0.81; 95%CI: 0.80-0.83), and Northeast (0.53; 95%CI: 0.52-0.54). Statistically significant declining trends occurred in the Northeast (APC: -2.21; 95%CI: -3.84 to -0.55) and South (APC: -2.36; 95%CI: -4.06 to -0.64). The Midwest and West showed non-significant downward trends (Midwest APC: -1.64, 95%CI: -3.41 to 0.17; West APC: -2.11, 95%CI: -4.29 to 0.11) (Figure 5 and Supplementary Table 6).

Figure 5 Trends in stroke age-adjusted stroke mortality rates in patients with chronic kidney disease stratified by region. APC: Annual percent change; AAMR: Age-adjusted stroke mortality rates.

Urbanization-stratified

AAMRs for stroke among individuals with CKD varied by urbanization level. The highest AAMR was observed in Small Metropolitan areas at 0.87 per 100000 population (95%CI: 0.84-0.90), followed by Micropolitan (Nonmetro) areas (0.85; 95%CI: 0.82-0.88), NonCore (Nonmetro) areas (0.83; 95%CI: 0.80-0.85), Medium Metro areas (0.82; 95%CI: 0.81-0.84), Large Central Metro areas (0.81; 95%CI: 0.80-0.83), and Large Fringe Metro areas (0.64; 95%CI: 0.62-0.65). Statistically significant declining trends occurred in Large Central Metro areas (APC: -2.54; 95%CI: -4.49 to -0.55), Large Fringe Metro areas (APC: -2.26; 95%CI: -3.85 to -0.65), and NonCore (Nonmetro) areas (APC: -1.58; 95%CI: -3.09 to -0.05). Non-significant downward trends were observed in Medium Metro (APC: -1.65; 95%CI: -3.48 to 0.22), Small Metro (APC: -1.80; 95%CI: -3.89 to 0.33), and Micropolitan areas (APC: -1.60; 95%CI: -3.32 to 0.15) (Figure 6 and Supplementary Table 7).

Figure 6 Trends in stroke age-adjusted stroke mortality rates in patients with chronic kidney disease stratified by urbanization. APC: Annual percent change; AAMR: Age-adjusted stroke mortality rates.

Place of death

Among the 37308 recorded stroke-related deaths among individuals with CKD, the majority (52.78%) occurred in medical facilities where the decedent was an inpatient. Nursing homes and long-term care settings accounted for 24.58% of deaths, while 11.35% occurred at home. Hospice facilities were the site of 5.05% of deaths. Outpatient and emergency department settings contributed to 2.68% of cases, and a small number were either dead on arrival (0.17%) or died in medical facilities with an unknown status (0.14%). Additionally, 3.00% of deaths occurred in other locations, and in 0.25% of cases, the place of death was not specified (Supplementary Table 8).

DISCUSSION

Our nationwide analysis of 22-year stroke mortality trends among United States adults with CKD reveals both encouraging improvements and persistent challenges. The overall decline in AAMR from 1.08 to 0.71 per 100000 between 1999 and 2020 aligns with broader cerebrovascular mortality reductions observed in the general population during this period[13]. This trend likely reflects advancements in hypertension management, statin utilization, and smoking cessation initiatives that began in the late 1990s[14]. However, our stratified analyses reveal that these benefits have been unequally distributed, with concerning disparities persisting across demographic and geographic dimensions.

The accelerated mortality reduction between 1999 and 2009 (APC: -4.25; 95%CI: -5.43 to -3.06) coincides with national quality improvement initiatives targeting cardiovascular risk factors, including widespread implementation of the Kidney Disease Outcomes Quality Initiative guidelines[15]. However, subsequent fluctuations—including a significant increase from 2009 to 2012 (APC: 23.25; 95%CI: 1.24-50.05), a sharp decline from 2012 to 2015 (APC: -28.10; 95%CI: -44.14 to -7.46), and a renewed increase from 2015 to 2020 (APC: 8.72; 95%CI: 3.66-14.04)—suggest complex underlying dynamics. These may reflect competing influences such as the rising prevalence of diabetes[16], obesity[17], and evolving therapeutic limitations specific to CKD, including altered drug metabolism, accelerated atherosclerosis, and the complexities of anticoagulation[18].

Standard stroke prevention protocols may underperform in CKD populations due to pathophysiological differences like uremia-induced platelet dysfunction[19]. Furthermore, ongoing debates regarding optimal blood pressure targets in CKD continue to complicate risk management[20].

Gender disparities remained persistent, with males consistently exhibiting higher stroke-related mortality than females throughout the study period (AAMR: 0.79 vs 0.71 per 100000). Notably, both genders experienced significant declines—males (APC: -1.89; 95%CI: -3.60 to -0.15) and females (APC: -2.26; 95%CI: -4.21 to -0.27)—but without any joinpoints, indicating consistent monotonic trends. These findings mirror broader stroke epidemiology patterns[21] yet highlight unique barriers in CKD, such as delayed care engagement and lower transplant access for males[22].

Racial disparities were stark. Black or African American individuals exhibited the highest overall AAMR (1.95 per 100000; 95%CI: 1.91-2.00), nearly three times that of White individuals (0.63; 95%CI: 0.63-0.64). Despite a significant declining trend among Black individuals (APC: -4.43; 95%CI: -5.93 to -2.91), the absolute burden remains disproportionately high. Similarly, Hispanic or Latino individuals and Asian or Pacific Islander individuals showed significant overall declines (APC: -4.49 and -3.66, respectively), but experienced concerning fluctuations after 2009—including sharp increases and decreases—possibly indicating instability in access to preventive care and stroke services. These patterns reflect structural inequities in access to neurology-nephrology co-management[23,24], medication affordability[25], and underutilization of renal replacement therapy among minority populations[26].

Geographic variation was profound. The highest AAMRs occurred in the District of Columbia (1.27), South Carolina (1.24), and West Virginia (1.17), while the lowest were in Arizona (0.38) and New York (0.40). The 3.3-fold state-level disparity underscores the intersection of the "Stroke Belt" phenomenon[27] with regional gaps in nephrology workforce and quality of care[28,29]. Regional analysis revealed that the West (AAMR: 0.89) and South (0.84) bore the greatest burden, while the Northeast had the lowest AAMR (0.53) and exhibited a statistically significant decline (APC: -2.21; 95%CI: -3.84 to -0.55). These findings underscore the importance of location-specific stroke prevention and CKD management initiatives.

Urbanization-stratified analysis revealed that Small Metropolitan (AAMR: 0.87) and Micropolitan areas (0.85) had the highest burden, while Large Fringe Metro areas (0.64) had the lowest. Declining trends were significant in Large Central Metro (APC: -2.54), Large Fringe Metro (-2.26), and Noncore rural areas (-1.58), suggesting that improvements are occurring but not uniformly. The higher AAMRs in more urbanized areas may reflect pollution-related vascular risks[30], noise exposure[31], and comorbidity clustering. Meanwhile, rural areas face reduced access to stroke centers and subspecialty care[32,33].

Place of death patterns provide further insight into healthcare delivery. Over half of the deaths (52.78%) occurred in inpatient settings, while long-term care facilities accounted for nearly a quarter (24.58%). Notably, 11.35% of deaths occurred at home and 5.05% in hospice settings—indicating potential gaps in access to timely acute stroke intervention and appropriate end-of-life care planning. The relatively small proportion of deaths occurring in outpatient or emergency settings (2.68%) may point to under recognition or delayed care-seeking behaviors in this population.

Together, these findings highlight the need for three paradigm-shifting approaches. First, risk stratification should incorporate CKD-specific biomarkers—such as albuminuria progression and fibroblast growth factor-23 levels[34]—alongside traditional cardiovascular tools. Second, targeted disparity-reduction interventions are crucial: Community screening programs in high-burden regions, telehealth hypertension management in rural areas[35], and culturally co-developed educational initiatives for Black communities[36]. Third, integrative care models must embed neurologists in nephrology clinics, with standardized atrial fibrillation screening and individualized antithrombotic protocols tailored to CKD-specific bleeding risks[37,38].

Study limitations must be acknowledged. Death certificate data may underestimate CKD as a contributing cause[39] and lack granularity on disease stage or stroke subtype. Our inability to isolate dialysis-dependent individuals likely underrepresents a high-risk group with unique stroke mechanisms[40]. Moreover, potential confounding by unmeasured factors such as socioeconomic status and medication adherence remains a concern. Future research should employ linked registry data evaluate CKD-specific stroke biomarkers[41], and test community-based interventions in high-disparity settings[42].

In conclusion, while stroke mortality has declined among CKD patients since 1999, profound disparities persist across racial, gender, and geographic lines. These inequities demand precision public health approaches that move beyond uniform interventions to address region-specific, race-specific, and place-specific determinants of cerebrovascular risk. CKD must be reconceptualized not merely as a comorbidity but as a cerebrovascular risk multiplier requiring integrated nephrology-neurology care models.

Limitations

This study has several limitations. Its retrospective nature restricts our ability to control for data accuracy and may introduce bias associated with the use of existing administrative records. Dependence on ICD coding for identifying stroke and CKD could result in misclassification or underreporting, and temporal changes in coding practices may affect the consistency of observed trends. Additionally, the lack of granular clinical information, such as CKD stage, duration, or treatment limits the depth of clinical interpretation. Important variables such as socioeconomic status, comorbidities, and access to healthcare were not captured, potentially introducing unmeasured confounding. Furthermore, as this is an observational study, causal relationships cannot be established. Variability in healthcare practices across regions may also affect the generalizability of our findings.

CONCLUSION

Over the 22-year period from 1999 to 2020, stroke-related mortality among individuals with CKD demonstrated an overall decline in AAMRs, reflecting potential improvements in preventive care, clinical management, and public health strategies. This downward trend was most pronounced between 1999-2008 and 2011-2020, though short periods of non-significant increases were also noted.

Despite these encouraging trends, important disparities persist. Mortality rates remained higher among males, non-Hispanic Black individuals, and residents of southern United States. Additionally, urban-rural and regional differences were evident, with higher AAMRs observed in central metropolitan areas and southern regions. Gender- and race-specific patterns revealed unique temporal variations, underscoring the influence of demographic and social determinants on health outcomes.

The findings highlight the need for targeted public health efforts to address these persistent disparities. Future research should focus on identifying the underlying biological, structural, and systemic factors contributing to these differences and evaluating interventions aimed at reducing stroke mortality among CKD populations. Enhanced surveillance, equitable access to care, and culturally informed prevention strategies will be critical to advancing health equity in this high-risk group.

ACKNOWLEDGEMENTS

The authors would like to express their sincere gratitude to Dr. Raheel Ahmed, MBBS, MRCP, PhD (Imperial College London, United Kingdom) for his professional English language editing and refinement of the manuscript to meet international academic publication standards (Grade A). The authors also acknowledge Dr. Bilal Ahmed, MBBS, Mph, PhD (Biostatistics) for his expert review and verification of the statistical analyses and methodologies employed in this study.