Age at Coronary Heart Disease Onset, Genetic Susceptibility, and Subsequent Risk of Stroke

Abstract

Background

Coronary heart disease (CHD) has been established as a risk factor for stroke. The study aimed to assess how age at CHD onset influences stroke risk and interacts with genetic predisposition to stroke.

Methods

This prospective cohort study included UK Biobank participants without baseline or pre‐CHD stroke. Polygenic risk scores assessed genetic susceptibility. Cox models and propensity score matching investigated the association between CHD onset age and stroke. The interaction analysis was performed using the bootstrap method and likelihood tests. Sensitivity analyses were conducted to ensure robustness.

Results

Over a median follow‐up of 15.1 years, 11 180 participants experienced incident stroke. Patients with CHD exhibited higher risks of developing stroke, hemorrhagic stroke, and ischemic stroke than participants without CHD. Notably, each 10‐year decrease in onset age correlated with an increased risk of stroke (hazard ratio [HR], 1.27 [95% CI, 1.23–1.32]; P<0.001), hemorrhagic stroke (HR, 1.31 [95% CI, 1.20–1.43]; P<0.001), and ischemic stroke (HR, 1.28 [95% CI, 1.23–1.33]; P<0.001). Propensity score matching revealed that, compared with their matched counterparts with later CHD onset, participants with earlier CHD onset had an increased risk of subsequent stroke: <57 versus ≥57 years (HR, 1.33 [95% CI, 1.20–1.48]; P<0.001) and <66 versus ≥66 years (HR, 1.57 [95% CI, 1.41–1.75]; P<0.001), as well as subsequent hemorrhagic stroke and ischemic stroke. However, no evidence supported an interaction between CHD onset age and genetic risk for stroke.

Conclusions

Earlier CHD onset is independently associated with greater risks of stroke, hemorrhagic stroke, and ischemic stroke, regardless of genetic predisposition. Earlier onset of CHD may serve as a strong independent marker of elevated stroke risk.

Nonstandard Abbreviations and Acronyms

HS

hemorrhagic stroke

IS

ischemic stroke

PRS

polygenic risk score

UKB

UK Biobank

Clinical Perspective.

What Is New?

• Using data from the UK Biobank, we demonstrated that an earlier onset of coronary heart disease confers a markedly higher risk of subsequent stroke, including ischemic and hemorrhagic subtypes, with no evidence that genetic susceptibility modifies this association.

What Are the Clinical Implications?

• Clinicians should more closely monitor cerebrovascular health and initiate targeted preventive strategies in patients with early‐onset coronary heart disease, given their elevated long‐term stroke risk.

• Our findings suggest a potential need for increased awareness of stroke risk in young patients with coronary heart disease (aged ≤40 years). Given the limited number of stroke events in this subgroup, these observations should be interpreted with caution and require confirmation in other younger populations.

Stroke remains a critical global health concern, marked by both its high incidence and significant mortality. ¹ Data from the 2021 GBD (Global Burden of Disease) Study report indicate that 7.3 million stroke‐related deaths occurred worldwide, accounting for 10.7% of all fatalities. This makes stroke the third leading cause of death globally and the fourth largest contributor to disability‐adjusted life years. ² Irrespective of stroke severity, all forms are associated with a permanent reduction in quality of life. ³ , ⁴ , ⁵ Compelling evidence continues to emerge, suggesting that most strokes are preventable through targeted risk factor interventions. ⁶ , ⁷

Coronary heart disease (CHD) represents a significant contributor to global mortality and morbidity, frequently coexisting with stroke, ⁸ , ⁹ , ¹⁰ and potentially linked to it through various pathways. ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ Recent studies have shown that the risk of subsequent stroke is elevated in patients with CHD. Two cohort studies provide evidence that the risk of stroke in patients with CHD increases by 2‐fold and 1‐fold, respectively. ¹⁶ , ¹⁷ Considering the extensive age range at which CHD may manifest, combined with its progressive trajectory, prolonged duration often leads to detrimental health outcomes. ¹⁸ , ¹⁹ Therefore, it is reasonable to postulate that an earlier onset age of CHD correlates with an elevated risk of subsequent stroke.

To date, the relationship between the onset age of CHD and the risk of subsequent stroke has yet to be thoroughly investigated. Furthermore, prior research has suggested that genetic predisposition may interact with various factors in shaping health outcomes. ¹¹ , ²⁰ However, the influence of age at the onset of CHD on the modulation of genetic susceptibility to stroke remains largely unexamined.

Determining the association between the age at onset of CHD and subsequent stroke, as well as the interaction between CHD onset age and genetic predisposition to stroke, may provide a theoretical foundation for early intervention and management in patients with CHD. Therefore, using data from the UK Biobank (UKB), which includes long‐term follow‐up information on CHD onset age, genetic risk for stroke, and stroke incidence, we aim to investigate whether an earlier age of CHD onset is associated with an increased risk of stroke and to explore the interaction between CHD onset age and genetic predisposition to stroke.

METHODS

The UKB data are available online at https://www.ukbiobank.ac.uk. The UKB was approved by the North West Multicenter Research Ethics Committee (reference: 11/NW/0382). All written informed consent was obtained from participants before enrollment in the study. The present research was conducted under application number 91090. The data supporting the findings of this study are available from the corresponding author on reasonable request.

Study Design and Participants

The UKB is a large prospective cohort study consisting of >500 000 participants recruited from 22 assessment centers across the United Kingdom between 2006 and 2010. More detailed information about the design of the UKB is available elsewhere. ²¹

As shown in Figure 1, a total of 502 394 participants were assessed at baseline. Among these, 7969 individuals with a history of stroke at baseline were excluded. The remaining cohort of 494 425 participants was included in the primary analysis to explore the association between CHD status and stroke risk. Furthermore, a subset of 400 630 participants with available polygenic risk score (PRS) data underwent additional adjustments for PRS effects. Subsequently, 59 636 participants with documented CHD onset age were selected to examine the association between age at CHD onset and stroke occurrence. Participants who experienced stroke before CHD diagnosis during follow‐up (n=720) were excluded from this analysis to ensure a correct temporal relationship between CHD onset and stroke. Moreover, within the population with CHD onset age, propensity score matching was performed based on age at onset (<57 versus ≥57 years and <66 versus ≥66 years) to further assess the association between the age at CHD onset and stroke events. Last, a combined analysis of CHD onset age and genetic risk for stroke was conducted among 47 743 participants who possessed both PRS data and information on CHD onset age.

Figure 1. Flowchart of participant selection for this study.

CHD indicates coronary heart disease; and PRS, polygenic risk score.

Ascertainment of CHD and CHD Onset Age

CHD was identified using self‐reported data, hospital inpatient records, and mortality register data with the International Classification of Diseases, Tenth Revision (ICD‐10), codes of I20 to I25, including angina pectoris, acute myocardial infarction, subsequent myocardial infarction, certain current complications following acute myocardial infarction, other acute ischemic heart diseases, and chronic ischemic heart disease. The onset age of CHD was ascertained either through self‐reported data or by calculating the interval between the participant's birth date and the date of the first recorded CHD event. Comprehensive details are available in Table S1.

PRS for Stroke

Genotyping, quality control, and risk prediction methods were meticulously conducted by the expert team at the UKB, with comprehensive accounts delineated in prior literature. ²² A total of 87 independent single‐nucleotide polymorphisms were carefully selected, ²³ and detailed information on these selected single‐nucleotide polymorphisms is presented in Table S2. Each participant's PRS for stroke was derived by multiplying the posterior effect size of each variant by its corresponding allele count, followed by the aggregation of these results. Participants' genetic risk profiles were stratified into low, medium, and high categories based on the tertiles of the PRS.

Ascertainment of Stroke

In the UKB study, stroke was ascertained through a combination of self‐reported information, hospital admission records, and mortality registry data. ²⁴ All‐cause stroke, hemorrhagic stroke (HS), and ischemic stroke (IS) were considered as outcomes in the present analyses. The ICD‐10 codes used to define stroke outcomes are comprehensively outlined in Table S3. All outcomes were monitored through April 30, 2024.

Covariates

Covariates encompassed age, sex, race (categorized as White or other [the "other" category refers to non‐White participants, specifically including Mixed, Asian, Black, Chinese, and other ethnic groups]), educational attainment (higher versus nonhigher education), obesity status (yes or no), hypertension, diabetes, low‐density lipoprotein cholesterol, total cholesterol, current smoking status (yes or no), alcohol consumption (defined as consuming alcohol once per week or more), exercise, depressive symptoms, statin use, and atrial fibrillation (AF). The selection of covariates was consistent with the relevant research. ²⁴ , ²⁵ Higher education was specified as holding a college or university degree, or an equivalent professional qualification. Exercise was defined as engaging in moderate or vigorous leisure‐time physical activity for >10 minutes at least twice per week, excluding occupational physical activity. Depressed mood was defined as reporting to feel down, depressed, or hopeless nearly every day or more than half the days during the past 2 weeks. Obesity was classified as a body mass index of ≥30 kg/m². The details of the covariates are summarized in Table S4.

Detailed information on missing values is provided in Table S5. Missing data for continuous variables were addressed using multiple imputation, whereas missing data for categorical variables were defined as a new category. This method involved generating 5 imputed data sets using the MICE package in R, a well‐established approach that accounts for the inherent uncertainty in missing data. ²⁶

Statistical Analysis

Baseline characteristics are reported as mean±SD for continuous variables and as frequencies (percentages) for categorical variables. Comparisons of baseline characteristics between participants with and without CHD were performed using t tests for continuous variables or χ² tests for categorical variables. Among participants with CHD, variations in baseline characteristics across age groups at CHD onset (divided by tertiles: <57, 57–66, and ≥66 years) were assessed using ANOVA for continuous variables or χ² tests for categorical variables.

Cox proportional hazards models were used to assess the associations between CHD, its age at onset, and the risk of incident stroke, following previously reported methods. ²⁷ , ²⁸ For analyses of CHD status, CHD was modeled as a time‐varying exposure to account for the occurrence of CHD during follow‐up. For analyses of age at CHD onset, conventional Cox proportional hazards models were applied among participants with CHD. The proportional hazards assumption was assessed by examining Schoenfeld residuals, and no evidence of violation was observed. ²⁹ , ³⁰ , ³¹ The time scale for analyses of CHD status and incident stroke was defined as the duration from baseline to the occurrence of incident stroke, death, or the end of follow‐up, whichever came earlier. ²⁵ For participants who experienced both stroke subtypes in succession, the follow‐up period associated with the initial stroke subtype was used as the time scale when defining all‐cause stroke as the outcome. Three models were used for adjustment: model 1 adjusted for age, sex, race, and education; model 2 incorporated adjustments for age, sex, race, education, obesity, systolic blood pressure, diastolic blood pressure, low‐density lipoprotein cholesterol, total cholesterol, current smoking, current drinking, exercise, depressed mood, hypertension, diabetes, statin use, and AF; and model 3 included additional adjustments for PRS, alongside the covariates in model 2. First, we investigated the association between CHD and stroke in the overall cohort (n=494 425) and in the subset of participants with available PRS data (n=400 630), using CHD as a time‐varying exposure. Second, for analyses of age at CHD onset, the time scale was defined as the duration from CHD diagnosis to the occurrence of stroke, death, or end of follow‐up. Participants with CHD were stratified into 3 groups based on their age at onset: <57, 57–66, and ≥66 years. Age at CHD onset was divided into 3 groups, consistent with previous studies on disease onset age. ²⁴ , ²⁵ We assessed the relationship between age at CHD onset and stroke among participants diagnosed with CHD (n=59 636). Third, we further explored the relationship between age at CHD onset and the subsequent risk of stroke by performing 1:1 nearest neighbor propensity score matching between the age groups <57 versus ≥57 years (n=29 644) and <66 versus ≥66 years (n=30 644). This matching was adjusted for age, sex, race, education, obesity, systolic blood pressure, diastolic blood pressure, low‐density lipoprotein cholesterol, total cholesterol, current smoking, current drinking, exercise, depressed mood, hypertension, diabetes, statin use, and AF.

To evaluate interactions, participants were categorized into 9 groups based on the combination of genetic risk for stroke and age at CHD onset. After full adjustment (model 2), the hazard ratios (HRs) and their 95% CIs for subsequent stroke, HS, and IS were estimated using the group with low genetic risk for stroke and high age at CHD onset as the reference. The relative excess risks attributable to interaction were calculated using the bootstrap method, whereas the likelihood tests were used to compare models with and without interaction terms, assessing associations on additive and multiplicative scales, respectively.

Sensitivity Analyses

Several sensitivity analyses were performed to ensure the robustness of our findings. First, subgroup analyses were conducted to explore potential effect modification by covariates on the relationship of CHD and its onset age with incident stroke. Second, competing risk models were used to account for death as a competing event in the association between CHD onset age and stroke. ³² Third, we excluded stroke events occurring within 5 years of baseline and repeated the primary analyses to mitigate potential reverse causality. Fourth, analyses were restricted to participants aged ≥50 years, given the lower prevalence of CHD in younger individuals. Fifth, the follow‐up cutoff was set to December 31, 2019, and the main analyses were repeated to address potential biases introduced by the COVID‐19 pandemic, as disruptions to hospital admissions and primary care services may have affected the diagnosis of both CHD and stroke during this period. Sixth, patients in the younger onset age group exhibited a higher prevalence of antidiabetic drug and antihypertensive drug use (P<0.001); thus, adjustments for these drugs were incorporated in our principal analysis to account for their potential confounding effects. Seventh, we calculated the cumulative incidence of stroke at 1, 5, and 10 years after baseline across different CHD onset age groups to examine the patterns of stroke occurrence during follow‐up. Eighth, to account for potential differences in stroke risk by the age at CHD onset, we further stratified participants with CHD into 4 groups (<40, 40–57, 57–66, and ≥66 years) and reassessed stroke risk using ≥66 years as the reference group. Ninth, we repeated the primary analyses after excluding participants with AF at baseline. Tenth, we repeated the primary analyses using multiple imputation for missing categorical covariates. Finally, we fitted Cox models restricted to participants who developed CHD during follow‐up.

All analyses were conducted using R version 4.3.3. Two‐sided tests were used for all analyses, and P<0.05 was considered statistically significant.

RESULTS

Baseline Characteristics

A total of 494 425 participants (54.6% women, n=269 976; mean±SD age, 56.5±8.1 years) were included in the analysis, of whom 59 929 (12.1%) had CHD. As shown in Table 1, compared with the other 2 groups with older onset age of CHD, those diagnosed with CHD at <57 years of age were younger at study baseline, and had a lower proportion of women, higher educational attainment, and exercise, while displaying higher rates of obesity, current smoking, depressed mood, diabetes, and statin use. Additionally, baseline characteristics stratified by CHD status are detailed in Table S6.

Table 1.

Characteristics of Study Participants With CHD, Stratified by CHD Onset Age Group

Characteristic	Onset age groups of CHD, y			P value
	<57	57–66	≥66
	(n=19 877)	(n=19 188)	(n=20 571)
Age, y	57.0±7.8	60.7±5.7	64.4±3.5	<0.001
Women, n (%)	6011 (30.2)	6728 (35.1)	8212 (39.9)	<0.001
White, n (%)	18 133 (91.2)	18 023 (93.9)	19 707 (95.8)	<0.001
Higher education, n (%)	7128 (35.9)	7209 (37.6)	7779 (37.8)	<0.001
Obesity, n (%)	7916 (39.8)	6388 (33.3)	6168 (30.0)	<0.001
SBP, mm Hg	137.5±18.3	141.3±18.5	146.4±19.0	<0.001
DBP, mm Hg	81.2±10.7	82.0±10.7	83.2±10.2	<0.001
LDL‐C, mmol/L	3.1±0.9	3.3±1.0	3.5±0.9	<0.001
TC, mmol/L	5.0±1.2	5.3±1.3	5.6±1.2	<0.001
Current smoking, n (%)	3356 (16.9)	2475 (12.9)	2192 (10.7)	<0.001
Current drinking, n (%)	12 347 (62.1)	12 535 (65.3)	13 901 (67.6)	<0.001
Exercise, n (%)*	14 566 (73.3)	14 736 (76.8)	16 157 (78.5)	<0.001
Depressed mood, n (%)	1819 (9.2)	1207 (6.3)	870 (4.2)	<0.001
Hypertension, n (%)	11 580 (58.3)	12 303 (64.1)	14 460 (70.3)	<0.001
Diabetes, n (%)	4083 (20.5)	3000 (15.6)	2871 (14.0)	<0.001
Statin use, n (%)	12 137 (61.1)	9110 (47.5)	6160 (29.9)	<0.001
AF, n (%)	1273 (6.4)	1335 (7.0)	1031 (5.0)	<0.001

Data are presented as mean±SD or number (percentage). AF indicates atrial fibrillation; CHD, coronary heart disease; DBP, diastolic blood pressure; LDL‐C, low‐density lipoprotein cholesterol; SBP, systolic blood pressure; and TC, total cholesterol.

^*Exercise was defined as engaging in moderate or vigorous leisure‐time physical activity for >10 minutes at least twice per week, excluding occupational physical activity.

Association Between CHD and Incident Stroke in the Overall Study Population

Over a median follow‐up of 15.1 years (interquartile range, 14.3–15.9 years), a total of 11 180 all‐cause stroke events were documented, including 2648 cases of HS and 9114 cases of IS, reflecting the occurrence of both stroke subtypes in some participants over time. Individuals with CHD had a markedly elevated risk of all‐cause stroke (HR, 1.80 [95% CI, 1.72–1.90]; P<0.001), HS (HR, 1.58 [95% CI, 1.42–1.77]; P<0.001), and IS (HR, 1.85 [95% CI, 1.75–1.95]; P<0.001) in the fully adjusted model (Table 2). Among the subset with available PRS data (n=400 630), participants with CHD persistently exhibited markedly elevated risks of all‐cause stroke (HR, 1.79 [95% CI, 1.69–1.90]; P<0.001), HS (HR, 1.61 [95% CI, 1.43–1.83]; P<0.001), and IS (HR, 1.87 [95% CI, 1.76–1.98]; P<0.001).

Table 2.

Associations Between CHD and Stroke Among the Total Study Population

Outcome	HR (95% CI) for CHD vs non‐CHD	P value
All‐cause stroke
Model 1*	2.03 (1.93–2.12)	<0.001
Model 2†	1.80 (1.72–1.90)	<0.001
Model 3‡	1.79 (1.69–1.90)	<0.001
HS
Model 1*	1.70 (1.53–1.89)	<0.001
Model 2†	1.58 (1.42–1.77)	<0.001
Model 3‡	1.61 (1.43–1.83)	<0.001
IS
Model 1*	2.11 (2.01–2.22)	<0.001
Model 2†	1.85 (1.75–1.95)	<0.001
Model 3‡	1.87 (1.76–1.98)	<0.001

CHD indicates coronary heart disease; HR, hazard ratio; HS, hemorrhagic stroke; and IS, ischemic stroke.

^*Adjusted for age, sex, race, and education.

^† Further adjusted for obesity, systolic blood pressure, diastolic blood pressure, low‐density lipoprotein cholesterol, total cholesterol, current smoking, current drinking, exercise, depressed mood, hypertension, diabetes, statin use, and atrial fibrillation.

^‡ Further adjusted for polygenic risk score.

Association Between CHD Onset Age and IS in Participants With CHD

A total of 59 636 participants with CHD were included in this analysis. As presented in Table 3, after adjusting for multiple variables, the risk of all‐cause stroke, HS, and IS increased progressively with younger age at CHD onset. Participants diagnosed with CHD at an earlier age exhibited a higher risk of all‐cause stroke (per 10‐year decrease in onset age: HR, 1.27 [95% CI, 1.23–1.32]; P<0.001), HS (per 10‐year decrease: HR, 1.31 [95% CI, 1.20–1.43]; P<0.001), and IS (per 10‐year decrease: HR, 1.28 [95% CI, 1.23–1.33]; P<0.001).

Table 3.

CHD Onset Age and Stroke Association in Patients With CHD

Outcome	Stroke cases	HR (95% CI)	P value*
All‐cause stroke
≥66 y (n=20 571)	823	Reference	…
57–66 y (n=19 188)	866	1.40 (1.27–1.55)	<0.001
<57 y (n=19 877)	946	1.75 (1.57–1.95)	<0.001
Per 10‐y decrease†	2635	1.27 (1.23–1.32)	<0.001
HS
≥66 y (n=20 571)	145	Reference	…
57–66 y (n=19 188)	161	1.48 (1.16–1.87)	0.001
<57 y (n=19 877)	175	1.91 (1.47–2.46)	<0.001
Per 10‐y decrease†	481	1.31 (1.20–1.43)	<0.001
IS
≥66 y (n=20 571)	717	Reference	…
57–66 y (n=19 188)	753	1.41 (1.26–1.57)	<0.001
<57 y (n=19 877)	828	1.76 (1.56–1.98)	<0.001
Per 10‐y decrease†	2298	1.28 (1.23–1.33)	<0.001

CHD indicates coronary heart disease; HR, hazard ratio; HS, hemorrhagic stroke; and IS, ischemic stroke.

^*Adjusted for age, sex, race, education, obesity, systolic blood pressure, diastolic blood pressure, low‐density lipoprotein cholesterol, total cholesterol, current smoking, current drinking, exercise, depressed mood, hypertension, diabetes, statin use, and atrial fibrillation.

^† Continuous variable of onset age of CHD, which is equal to: (CHD onset age/10).

Following propensity score matching, 2 separate matched cohorts were constructed for the comparisons of CHD onset age <57 versus ≥57 years (n=29 644) and <66 versus ≥66 years (n=30 644), respectively. As indicated in Tables S7 and S8, the baseline characteristics were well balanced between participants with earlier‐onset CHD and their matched counterparts with later‐onset CHD. A diagnosis of CHD before the age of 57 years is associated with an elevated risk of IS compared with those diagnosed at or beyond 57 years (HR, 1.33 [95% CI, 1.20–1.48]; P<0.001) (Table 4). Similarly, CHD onset before 66 years of age confers a greater risk of stroke compared with diagnosis at or beyond 66 years (HR, 1.57 [95% CI, 1.41–1.75]; P<0.001). Furthermore, individuals with earlier CHD onset exhibit consistently higher HRs for both HS and IS relative to their counterparts with later CHD onset.

Table 4.

CHD Onset Age and Stroke Association in Patients With CHD Following Propensity Score Matching

Outcome	Stroke cases	HR (95% CI)	P value*
All‐cause stroke
<57 vs ≥57 y (n=29 644)	1378	1.33 (1.20–1.48)	<0.001
<66 vs ≥66 y (n=30 644)	1458	1.57 (1.41–1.75)	<0.001
HS
<57 vs ≥57 y (n=29 644)	248	1.49 (1.15–1.91)	0.002
<66 vs ≥66 y (n=30 644)	259	1.90 (1.47–2.46)	<0.001
IS
<57 vs ≥57 y (n=29 644)	1199	1.34 (1.20–1.50)	<0.001
<66 vs ≥66 y (n=30 644)	1282	1.55 (1.39–1.73)	<0.001

CHD indicates coronary heart disease; HR, hazard ratio; HS, hemorrhagic stroke; and IS, ischemic stroke.

^*Adjusted for age, sex, race, education, obesity, systolic blood pressure, diastolic blood pressure, low‐density lipoprotein cholesterol, total cholesterol, current smoking, current drinking, exercise, depressed mood, hypertension, diabetes, statin use, and atrial fibrillation.

Interaction Between Genetic Risk for Stroke and CHD Onset Age

The integrated analysis of genetic risk for stroke and CHD onset age in the subset of participants with available PRS data (n=47 743) is presented in Figure 2. In this combined analysis, an older age at CHD onset was consistently associated with a lower risk of stroke across individuals with low, medium, and high genetic risk profiles. Compared with participants with low genetic risk and CHD onset at age ≥66 years, those with high genetic risk and CHD onset before the age of 57 years showed a markedly elevated risk of stroke: all‐cause stroke (HR, 2.12 [95% CI, 1.75–2.59]; P<0.001), HS (HR, 2.37 [95% CI, 1.51–3.70]; P<0.001), and IS (HR, 2.18 [95% CI, 1.76–2.69]; P<0.001). However, although no statistically significant interaction was found between genetic risk for stroke and age at CHD onset on additive or multiplicative scales, we found that across low, medium, and high stroke genetic risk levels, different CHD onset age groups exhibited elevated trends in the risks of stroke, HS, and IS.

Figure 2. Interaction between genetic risk for stroke and age at CHD onset.

A, Interaction between genetic risk and age at onset of CHD in stroke. B, Interaction between genetic risk and age at onset of CHD in hemorrhagic stroke. C, Interaction between genetic risk and age at onset of CHD in ischemic stroke. Models were adjusted for age, sex, race, education, obesity, SBP, DBP, LDL‐C, TC, current smoking, current drinking, exercise, depressed mood, hypertension, diabetes, statin use, and AF. AF indicates atrial fibrillation; CHD, coronary heart disease; DBP, diastolic blood pressure; HR, hazard ratio; LDL‐C, low‐density lipoprotein cholesterol; RERI, relative excess risk attributable to interaction; SBP, systolic blood pressure; and TC, total cholesterol.

Subgroup and Sensitivity Analysis

As illustrated in Figures S1 through S6, our subgroup analysis provided valuable insights, demonstrating that, within the overall study cohort, the association between CHD and all‐cause stroke revealed significant interactions with sex, current smoking, obesity, hypertension, and AF. Moreover, the association between CHD and HS demonstrated notable interactions with diabetes. Similarly, the relationship between CHD and IS revealed significant interactions with sex, obesity, hypertension, and AF. Furthermore, among individuals with CHD, each 10‐year decrease in onset age was more strongly associated with HS in current alcohol consumers (interaction P<0.05).

As outlined in Table S9, after accounting for the competing risk of death, the findings remained consistent. Moreover, the results remained robust after excluding participants diagnosed with stroke within the first 5 years of baseline, excluding participants aged <50 years at baseline, setting the follow‐up cutoff to December 31, 2019, further adjusting for antihypertensive and antidiabetic drug use, excluding participants with AF at baseline, and repeating the primary analyses using multiple imputation for missing categorical variables (Tables S10–S13, S16, and S17). The cumulative incidence of stroke increased over time in all CHD onset age groups. Moreover, participants with earlier CHD onset exhibited higher cumulative incidence compared with those with later onset (Table S14). When CHD onset age was further divided into 4 categories, individuals who developed CHD before the age of 40 years exhibited the highest risks of all‐cause stroke and IS. For HS, the HR in the <40‐year group was slightly lower than that in the 40‐ to 57‐year group (Table S15). When analyses were restricted to participants who developed CHD during follow‐up, the association between the youngest CHD onset group and HS did not reach statistical significance, although the direction of effect was consistent (Table S18).

DISCUSSION

Using data from the prospective longitudinal UKB cohort, this study observed that individuals with CHD were at a higher risk of stroke compared with those without CHD. Notably, among individuals with CHD, an earlier age at onset was associated with an elevated risk of stroke, even after comprehensive adjustment for potential confounders. This association was consistently observed across diverse subgroups and remained robust under a comprehensive series of sensitivity analyses. Furthermore, participants with an earlier CHD onset age and high genetic predisposition to stroke had a higher observed risk of stroke.

Extensive prior research has shown that CHD significantly increases the risk of stroke. Evidence suggests that myocardial infarction markedly elevates stroke risk, particularly in the early postevent period. ³³ , ³⁴ , ³⁵ In a large cohort study involving 78 195 participants with a median follow‐up of 36 years (interquartile range, 17–60 years), CHD was found to be associated with a higher risk of IS. ³⁶ The FHS (Framingham Heart Study), along with a cohort study tracking 64 977 participants over the course of 1 year, revealed that CHD elevates the risk of stroke by 2‐ and 1‐fold, respectively. ¹⁶ , ¹⁷ These findings are consistent with those of the present analysis, in which CHD, modeled as a time‐varying exposure, was associated with an HR of 1.80 for all‐cause stroke, 1.58 for HS, and 1.85 for IS. Previous cohort studies leveraging data from the UKB have demonstrated that an earlier onset of CHD is associated with an elevated risk of subsequent dementia, ²⁵ whereas an earlier onset of orthostatic hypotension correlates with an augmented risk of stroke. ²⁴ Grounded in precise data on the age at CHD onset from the UKB and bolstered by a substantial sample size, we further observe an association between earlier CHD onset and an increased subsequent risk of stroke and its subtypes.

A previous study suggested that no interaction exists between genetic susceptibility to β‐cell dysfunction and an earlier onset age of type 2 diabetes in glycemic progression. ³⁷ Although the focus of the 2 studies differs, both the previous study and the current one failed to identify an interaction between earlier disease onset and genetic risk. This lack of interaction may suggest that, once CHD develops, acquired and lifestyle‐related factors, such as vascular remodeling, inflammation, and metabolic disturbances, could play a more dominant role in determining subsequent cerebrovascular outcomes than inherited susceptibility alone. Furthermore, other studies exploring the interaction between health behaviors, muscle strength, and genetic susceptibility to stroke have demonstrated that adherence to a healthy lifestyle and greater muscle strength are associated with a reduced stroke risk, independent of genetic susceptibility to stroke. ²³ , ³⁸ Notably, our study found that participants with high genetic risk for stroke but later onset of CHD (≥66 years) exhibited a lower risk of subsequent stroke, HS, and IS compared with those with low or medium genetic risk but earlier onset of CHD (<57 years). A UKB‐based study found that individuals with hypertension and high genetic risk who adhered to a healthy lifestyle had a generally lower risk of both CHD and stroke than those with lower or moderate genetic risk but an unhealthy lifestyle, ³⁹ a finding consistent with our own. This pattern suggests that the age at CHD onset may be an important marker of subsequent stroke risk, even among individuals with high genetic susceptibility.

Our study further revealed that CHD exerted a more pronounced effect on stroke risk in women, noncurrent smokers, individuals devoid of hypertension, diabetic individuals, individuals without AF, and those who were nonobese. Moreover, our study also revealed a more marked effect of a 10‐year earlier onset of CHD on HS in current drinkers. This may have been attributable to the relatively higher baseline risks in male participants, current smokers, individuals with hypertension, individuals with AF, and obese people, as well as a survival effect, where participants with higher mortality risks may have already died before enrollment. ⁴⁰ These findings highlight the potential importance of enhanced stroke risk assessment and closer clinical monitoring in patients with established CHD, specifically targeting high‐risk subgroups, including women, noncurrent smokers, nonhypertensive individuals, diabetic patients, individuals without AF, and nonobese individuals, as well as current drinkers among those with early CHD onset. In sensitivity analyses incorporating finer stratification of CHD onset age, individuals with early‐onset CHD (aged <40 years) had the highest risks of all‐cause stroke and IS. The slightly lower HR for HS in this subgroup, despite remaining statistically significant, may reflect the relatively small number of participants (n=1711) and HS events (n=14), leading to reduced precision of the risk estimates. The nonsignificant association for HS in the youngest CHD onset group in this sensitivity analysis is likely attributable to the small number of events (n=14).

Although the precise pathophysiological mechanisms underlying the association between the age at onset of CHD and stroke remain unclear, several potential pathways may explain our findings. First, observations indicate that those diagnosed with CHD before the age of 57 years typically have lower education, are less likely to engage in exercise, and exhibit a higher susceptibility to obesity, smoking, depression, and diabetes comorbidities, all of which are established risk factors for stroke. ¹¹ , ⁴¹ , ⁴² , ⁴³ Furthermore, studies have shown that patients with early‐onset CHD present a higher prevalence of modifiable risk factors, yet these factors are frequently poorly controlled. ⁴⁴ Second, CHD shares multiple risk factors with stroke, ¹¹ , ⁴¹ , ⁴² which contribute to vascular narrowing and impaired blood flow by promoting atherosclerosis. ⁴⁵ In addition, patients with early‐onset CHD often exhibit endothelial dysfunction and systemic microvascular abnormalities. ⁴⁶ , ⁴⁷ The prolonged cardiovascular burden may lead to microvascular disease, which plays a crucial role in the initiation and progression of stroke. ¹⁴ Moreover, inflammation is involved in the formation, growth, and rupture of atherosclerotic plaques, significantly influencing the mechanisms underlying both CHD and certain types of IS. ¹⁵ The prolonged inflammatory response in patients with early‐onset CHD may further elevate the subsequent risk of stroke, and studies investigating anti‐inflammatory therapies for patients with CHD have reported improvements in prognosis, highlighting the critical role of inflammation in stroke development. ⁴⁸ , ⁴⁹ These findings highlight the need for clinicians to pay particular attention to patients with early‐onset CHD. Such individuals may benefit from more intensive control of vascular risk factors, and from more frequent neurologic assessments to facilitate early detection and prevention of stroke.

This study has several advantages. First, the large sample size provides sufficient statistical power to examine the relationship between the age at onset of CHD and stroke, as well as its subtypes. Second, the results derived from algorithms within the UKB use a standardized method to define health outcomes, which enables the identification of specific dates for health outcomes recorded from self‐reported data, hospital inpatient records, and mortality register data, thereby ensuring a high positive predictive value for stroke. Finally, the implementation of propensity score matching helped control for a wide variety of risk factors, thereby mitigating confounding bias.

Despite these advantages, several limitations should be highlighted. First, as an observational investigation, it cannot definitively establish causal relationships. Second, reliance on baseline self‐reported data may produce recall bias, potentially resulting in the misclassification of critical variables. Furthermore, the absence of long‐term follow‐up raises the possibility that these variables underwent changes during the study period. Third, although our research accounted for numerous potential confounding factors, the effects of residual confounding cannot be completely ruled out. Fourth, as >94% of our study population were White, the findings may not be generalizable to other racial or ethnic groups. Future studies including more diverse populations are warranted to validate these associations. Fifth, the presence of missing data in this study may have introduced potential biases that could influence the findings. To address this limitation, we used multiple imputation techniques to maximize the use of available information. However, this approach may not accurately reflect the true nature of the missing data, highlighting the need for further research to overcome these limitations. Sixth, although only 1711 participants had CHD onset before the age of 40 years, the relatively small number of participants and HS events may have reduced the precision of the HR estimates. Nevertheless, the associations with all‐cause stroke, HS, and IS remained statistically significant, suggesting that elevated cerebrovascular risk may also be observed among individuals with early‐onset CHD. Further studies in younger cohorts are warranted to confirm these findings.

CONCLUSIONS

This study elucidates that earlier onset of CHD may serve as a strong independent marker of elevated stroke risk. These findings carry substantial public health implications, highlighting the critical need for early intervention and vigilant monitoring in patients with early‐onset CHD.

Sources of Funding

This work was supported by grants from the National Natural Science Foundation of China (82270254, 82271609, and 82371573), the Guangzhou Science and Technology Plan Project (2023B01J1011), and the Sun Yat‐sen Pilot Scientific Research Fund (YXQH202426).

Disclosures

None.

Supporting information

Tables S1–S18

Figures S1–S6

STROBE Checklist