Physical activity, muscle strength, sedentary behavior, sleep, and genetic risk of stroke and dementia: findings from a large cohort study

Abstract

Background

Physical activity, grip strength, sedentary behaviors, and sleep duration were found to be associated with risk of developing stroke and dementia. However, the combined influence of these factors on stroke and dementia remains unclear.

Objective

To investigate the combined influence of these multiple lifestyle and functional factors on risk of stroke and dementia and their subtypes and to investigate the potential interaction between combined factors and the apolipoprotein E gene ε4 allele (APOE ε4).

Methods

Data were obtained from the UK Biobank, including 474,983 participants. A score ranging from 0 to 4 was assigned based on adherence to healthy factors: meeting physical activity recommendations, grip strength above the sex-specific median, sleep duration of 7–8 h/day, and sedentary time < 6 h/day. Cox proportional hazards models estimated hazard ratios (HRs) and 95% confidence intervals (CIs) for incident stroke and dementia, adjusting for potential confounders.

Results

Over a median follow-up of 10.1 years, 4,992 incident strokes and 2,120 dementias were recorded. Compared with participants with 0–1 healthy factor, adjusted HRs (95% CIs) for total stroke were 0.85 (0.79–0.92), 0.71 (0.66–0.77), and 0.65 (0.59–0.72) for those with 2, 3, and 4 healthy factors, respectively (P-trend < 0.001). Similar inverse associations were observed for ischemic stroke and intracerebral hemorrhage but not subarachnoid hemorrhage. For dementia, HRs (95% CIs) were 0.74 (0.66–0.83), 0.64 (0.56–0.71), and 0.43 (0.39–0.51) across increasing numbers of healthy factors (P-trend < 0.001), with consistent results for Alzheimer’s disease and vascular dementia. All four factors independently predicted lower risk of all-cause stroke and all-cause dementia. For stroke subtypes, associations varied by factor. Regular physical activity and higher grip strength were both associated with lower risk of intracerebral hemorrhage, while none of the healthy factors had associations with subarachnoid hemorrhage. For dementia subtypes, all healthy factors, except for physical activity, were associated with a lower risk of both Alzheimer’s disease and vascular dementia. In addition, the association between combined healthy factors and total stroke or all-cause dementia was independent of APOE ε4 carrying status.

Conclusions

The cumulative association of multiple healthy factors with reduced risk of stroke and dementia highlights the importance of adopting a lifestyle with more elements of healthy factors for the prevention of these neurological diseases.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-25305-4.

Introduction

Neurological diseases are the leading cause of disability and the second leading cause of death worldwide [1]. Stroke and dementia are two predominant neurological disorders which caused about 8.2 million deaths in 2019 and accounted for 87.2% of neurological deaths [2]. Stroke and dementia increase the risk of each other and share some common risk factors at a population level [2]. The reciprocal association between stroke and dementia is well-established [3]. Due to their intertwined nature driven by common risk factors, the establishment of more comprehensive, efficient and integrative prevention and intervention strategies to alleviate their collective burden should be given high priority.

Modifiable lifestyle factors, including physical activity (both aerobic and resistance exercise), sedentary behavior, and sleep patterns, have been linked to the risk of stroke and dementia. For example, regular physical activity and greater grip strength (a partial indicator of resistance training) have been associated with lower risk of stroke and dementia [4–8]. Conversely, prolonged sitting time and both insufficient and excessive sleep duration (compared to 7–8 h) have been linked to an increased risk of stroke and dementia [9–12]. From a movement perspective, a 24-hour day can be regarded as a continuum of behaviors ranging from sleep to sedentary activity and physical activity. In addition to their well-established associations with stroke and dementia risk, the four factors (physical activity, grip strength, sleep duration, and sedentary time) were selected to represent this spectrum of movement intensity. This selection allows for a comprehensive assessment of how the full range of daily movement behaviors may collectively influence neurological health. Nevertheless, most previous studies have focused on only one factor at a time without accounting for their correlations. Since behavioral risk factors often interact and do not operate in isolation, it is important to examine their combined influence. To date, few studies have explored the joint impact of multiple behaviors on the risk of stroke and dementia. Understanding how these behaviors interact could inform more holistic and effective preventive strategies. In addition, previous studies have been largely focused on these factors and total stroke/dementia, with limited attention to specific subtypes and yielded inconsistent results [13–20]. Given the distinct etiological mechanisms underlying stroke and dementia subtypes, such as ischemic versus hemorrhagic stroke and Alzheimer’s disease versus vascular dementia [21, 22], and evidence from previous studies suggesting that different factors may may have varying associations with disease subtypes [17, 19], it is essential to investigate how specific above mentioned factors are associated with these subtypes. This would offer targeted insights into individualized preventive and therapeutic strategies.

To address these gaps, our study utilized data from the UK Biobank, a large population-based cohort, to examine the associations between four key factors-regular physical activity, grip strength, sleep duration, and sedentary time-and the risk of stroke and dementia, including their major subtypes. Additionally, we explored the potential modification of these relationships by the APOE ε4 genotype, a well-established genetic risk factor for dementia [23] and possibly stroke [24]. By delving into these multifaceted associations, our research aims to contribute novel insights that inform more effective strategies for stroke and dementia prevention.

Methods

Study population

The UK Biobank is a large prospective observational study established to provide a resource for investigation of the genetic, environmental, and lifestyle factors associated with a wide range of diseases [25]. In brief, between 2006 and 2010, approximately 500,000 ethnically diverse men and women aged 37–73 years were recruited from 22 centers across England, Wales, and Scotland. At recruitment, participants provided a wide range of information on health and diseases and underwent various physical measurements. For the present study, we excluded participants who had stroke or other cerebrovascular diseases (n = 9,551) or dementia (n = 209) at baseline, participants without data on one or more of the examined factors (n = 17,772). Finally, a total of 474,983 participants remained for the main analysis, and 436,511 participants with genotyped data who self-reported as having British White ethnicity were included in the analyses that involved APOE genotype related analysis. (Supplementary Fig. 1).

Ethics considerations

The UK Biobank (UKB) obtained ethical approval from the North West Multi-center Research Ethics Committee (reference 21/NW/0157). This approval means that researchers do not require separate ethical clearance and can operate under the above approval. Prior to enrollment in the study, all participants furnished written informed consent, and the research adhered to the principles outlined in the Declaration of Helsinki.

Measurements of physical activity, grip strength, sedentary time and sleep

Information on physical activity was collected by the International Physical Activity Questionnaire (IPAQ) [26]. Regular physical activity was defined as engaging in at least 150 min/week of moderate activity or 75 min/week of vigorous activity per week (or an equivalent 150 combination), or moderate activity at least 5 days a week or vigorous activity at least 3 days a week (≥ 10 min continuously at a time) [27]. Grip strength was measured by a Jamar J00105 hydraulic hand dynamometer and the mean values of grip strength expressed in kilograms was generated from the right and left hand. Sedentary time was the sum of self-reported time spent watching television, using computer and driving, and daily sitting of no less than 6 h was deemed to be sedentary behavior [28]. Sleep duration was reported as the hours of sleep in every 24 h (including naps), and appropriate sleep duration was defined as sleeping 7–8 h per day as previously suggested [29]. According to previous studies for scoring the healthy lifestyles [30–32], we assigned each healthy parameters with one point. The healthy factors included regular physical activity, higher grip strength (above sex-specific median), lower sedentary time (< 6 h/day), and appropriate sleep duration (7–8 h/day), and total score of the healthy factors ranged between 0 and 4 (Supplementary Table 1).

Ascertainment of stroke and dementia

The outcomes (stroke and dementia) were ascertained by the algorithmic combinations of coded information from hospital admissions (diagnoses and procedures), death registries, and self-reported medical condition. More information, such as the defined codes for stroke and dementia based on International Classification of Diseases-10th revision (ICD-10) can also be found on the website of the UK Biobank (https://biobank.ndph.ox.ac.uk/ukb/refer.cgi?id=460). The detailed field ID information of stroke and dementia from the UK Biobank was provided in Supplementary Table 2. In this study, total stroke and total dementia were designated as the primary outcomes, while the disease subtypes—including ischemic stroke, intracerebral hemorrhage, subarachnoid hemorrhage, Alzheimer’s disease, and vascular dementia—were considered secondary outcomes. Follow-up time was calculated as the time interval between the date of baseline assessment and the date of censoring, diagnosis, death, or lost to follow-up, whichever came first.

APOE genotyping

Genotypes were assayed among nearly one-half million participants in the UK Biobank by 2 very similar genotyping arrays manufactured by Affymetrix: about 50,000 participants by BiLEVE Axiom array and about 450,000 participants by Biobank Axiom array [33]. The genotyping quality control was performed by centralized analysis of the genetic data. Poor quality samples with unusually high heterozygosity and missingness (>5%) and mismatches between reported sex and genetic sex were excluded in genotype-related analyses. The APOE haplotypes (ε2/ε3/ε4) were determined by 2 genetic variants, rs429358 and rs7412. Participants with 1 or 2 ε4 alleles were defined as APOE ε4 carriers and otherwise as APOE ε4 noncarriers.

Covariates

The adjusted covariates comprised demographic and socioeconomic factors (age, sex, ethnic, Townsend deprivation index), medical history (hypertension, diabetes, dyslipidemia, and BMI), lifestyle factors (smoking status, alcohol consumption and diet). These variables were chosen based on their known relevance to stroke and dementia risk in prior literature.

The Townsend deprivation index was generated based on 4 socioeconomic variables (unemployment, non-car ownership, non-home ownership, and household overcrowding) [34]. Information on usual dietary factors was collected via a touchscreen food-frequency questionnaire which included 29 questions concerning average intake of major foods or food groups over the past year. Following recommendations on dietary priorities for cardiovascular health [35]. Healthy diet was defined based on consumption of 5 commonly eaten food groups: higher consumption of fruit and vegetables, whole grains, and fish, lower consumption of red and processed meat, and refined grain. Each favorable dietary factor was assigned a score of 1 point, resulting in a total diet score ranging from 0 to 5 (Supplementary Table 3). Body mass index (BMI) was calculated using measured weight and height (kg/m²). Diabetes was defined using self-report, diagnosis codes and a glycated haemoglobin (HbA1c) level ≥ 48 mmol/mol (6.5%). Baseline blood pressure data were obtained by a digital blood pressure monitor (Omron HEM-7015IT; OMRON Healthcare Europe B.V., Hoofddorp, Netherlands). Mean systolic blood pressure (SBP) and diastolic blood pressure (DBP) were calculated from two automated or two manual measurements by trained nurses [36]. Hypertension was defined as measured systolic/diastolic blood pressure of 140/90 mmHg or above, self-reported physician’s diagnosis, or self-reported use of antihypertensive medications. Dyslipidemia was defined as self-reported physician’s diagnosis or self-reported lipid-lowering medication use.

Statistical analysis

Baseline sociodemographic, lifestyle, dietary quality score, prevalence of chronic diseases (diabetes, hypertension and dyslipidemia), and APOE genotype were summarized and stratified by the score of the healthy factors. Among incident cases, total stroke and its subtypes (ischemic stroke, intracerebral hemorrhage and subarachnoid hemorrhage), all-cause dementia and its subtypes (Alzheimer’s disease and vascular dementia) were treated as separate outcomes. Kaplan-Meier curves were constructed to depict the relationship between the scores and the risk of stroke and dementia. The associations between score of healthy factors and incident stroke or dementia were explored by multivariate Cox proportional hazards models and the group with lowest score (0–1) was used as the reference. The proportional hazards assumption for the Cox models was assessed by examining Schoenfeld residuals and by testing for a non-zero slope in a generalized linear regression of the scaled Schoenfeld residuals on time. No major violations of the assumption were detected. For the reference group, we opted for those with 0 or 1 point to secure a satisfactory number of events. This selection was based on the limited subset of participants (8,837 out of 474,983) who had none of the assessed healthy factors, resulting in 185 incident stroke cases and 88 incident dementia cases. Hazard ratios (HR) with 95% confidence intervals (CI) were reported for all analyses.

Three models were applied in our analyses. The minimally adjusted model (Model 1) was adjusted for age at baseline, sex, self-reported ethnicity (White, Asian/Asian British, Black/Black British, mixed/unknown). The multivariable-adjusted model (Model 2) was additionally adjusted for socioeconomic status (Townsend deprivation index), smoking status (never, former, current), pack-years of smoking (for current smokers), alcohol drinking (never, former, current: <1, 1 to < 3, 3 to < 7, ≥7 drinks/week), and diet quality score (0–5). We used an exploratory model (Model 3) which was additionally adjusted for BMI, hypertension, dyslipidemia and diabetes, because these metabolic factors may be on the pathways from the healthy factors to the development of stroke or dementia. Due to the exploratory nature of Model 3 and the possibility that the additional variables included in this model may serve as intermediates, results from Model 2 were treated as main findings. When examining the associations of individual healthy factors with risk of stroke, dementia and their major subtypes, we conducted additional adjustments, accounting for the remaining 3 individual healthy factors after the multivariable adjustment.

Subgroup analyses were performed to examine score of the healthy factors in relation to all-cause stroke or dementia according to age at baseline, sex, smoking status, BMI, hypertension, dyslipidemia, and diabetes. Interactions were assessed using the likelihood ratio test.

To address the possibility of reverse causation, sensitivity analyses were performed to examine the association of the combined healthy factors (score) with risk of stroke or dementia. This was achieved by excluding participants who had coronary heart disease or cancer at baseline, or participants who developed incident stroke or dementia within the first 4 years of follow-up. Additionally, to evaluate the potential confounding effect of education, we conducted a sensitivity analysis by further adjusting for education level, categorized as College or University degree, A levels/AS levels or equivalent, O levels/GCSEs or equivalent, and others, to assess the robustness of the associations.

We also examined individual associations between the 4 healthy factors and risk of stroke and dementia and their major subtypes, after multivariable adjustment and further adjustment for the other 3 individual healthy factors.

Finally, to examine potential effect modification by APOE ε4 carrier status, we included a multiplicative interaction term between the number of healthy factors (continuous, range 0–4) and APOE ε4 status (binary: carrier vs. non-carrier) in the Cox proportional hazards models. The statistical significance of the interaction was assessed using likelihood ratio tests by comparing models with and without the interaction term. All analyses were performed with Stata (version 15.1; StataCorp). All P values were 2-sided and P <.05 was considered statistically significant.

Patient and public involvement

No patients were involved in setting the research question or the outcome measures, nor were they involved in the design and implementation of the study. No plans exist to disseminate the results to study participants.

Results

Baseline characteristics

Overall, 12.2%, 27.5%, 38.9% and 22.4% participants had 0–1, 2, 3 and 4 healthy factors, respectively. Compared with participants with a lower score of the healthy factors, those with a higher score were younger, had better socioeconomic status, were more likely to be female, ethnically white, nonsmokers, and moderately alcohol drinkers (Table 1). Participants with a higher score also ate healthier and were less likely to have diabetes, hypertension or dyslipidemia. Participants with a higher score had a higher proportion of APOE ε4 genotypes.

Table 1.

Baseline characteristics of participants from the UK Biobank

	Score of healthy factors				P value
	0–1 (n = 58,040)	2 (n = 130,834)	3 (n = 179,825)	4 (n = 106,284)
Age (years), mean ± SD	58.0 ± 7.7	57.4 ± 7.9	56.4 ± 8.1	54.2 ± 8.1	< 0.001
Male sex, %	52.7	48.7	44.7	39.3	< 0.001
Ethnicity, %
White	92.2	93.2	95.0	96.7	< 0.001
Asian/Asian British	0.6	0.6	0.6	0.5
Black/Black British	2.8	2.4	1.6	0.8
Mixed/unknown	4.4	3.8	2.8	2.0
Townsend deprivation index, %					< 0.001
Q1	16.0	19.1	21.2	23.0
Q2-Q4	57.2	59.7	61.2	61.8
Q5	26.8	21.3	17.6	15.2
Smoking status					< 0.001
Never	48.1	53.3	56.3	58.9
Former	37.3	35.5	34.3	32.7
Current	14.6	11.2	9.4	8.4
Alcohol drinking, %					< 0.001
Never	5.7	5.0	3.9	2.8
Former	5.8	3.9	2.9	2.3
Current < 1 drink/week	27.2	24.0	21.6	19.4
Current 1 to < 3 drinks/week	24.5	25.7	26.0	26.7
Current 3 to < 7 drinks/week	18.5	21.4	24.3	27.2
Current ≥ 7 drinks/week	18.3	20.0	21.3	21.6
Diet quality score, %					< 0.001
0–1	39.3	33.6	29.1	25.6
2–3	51.5	54.2	56.4	58.0
4–5	9.2	12.2	14.5	16.4
Body mass index, kg/m2	29.2 ± 5.6	27.9 ± 4.9	27.0 ± 4.5	26.3 ± 4.2
Diabetes, %	11.7	7.2	4.8	2.9	< 0.001
Hypertension, %	63.2	58.9	54.5	48.5	< 0.001
Dyslipidemia, %	26.9	20.7	16.0	10.9	< 0.001
Apolipoprotein E genotype*, %
ε2/ε2, ε2/ε3	13.5	12.9	12.9	12.7	< 0.001
ε3/ε3	58.6	59.0	58.4	58.3
ε2/ε4, ε3/ε4, ε4/ε4	27.9	28.1	28.7	29.0

*For 436,511 European white individuals only

Kaplan-Meier curves and Cox proportional hazards models

During a median follow-up of 10.1 years, 4,992 incident cases of all-cause stroke occurred, among which there were 3883 ischemic stroke, 830 intracerebral hemorrhage stroke and 563 subarachnoid hemorrhage stroke cases. Supplementary Fig. 2 showed the Kaplan-Meier curves for incident stroke and dementia by score of the healthy factors. These curves depict the cumulative probability of survival without stroke or dementia over the follow-up time, according to the number of healthy factors. These curves showed that participants with more elements of healthy factors consistently had a higher cumulative probability of survival without stroke and dementia over the follow-up period. Table 2 showed the results of Cox proportional hazards models. Irrespective of the degree of adjustment for potential confounders, having a higher score of healthy factors was associated with progressively lower risk of all-cause stroke, ischemic stroke, and intracerebral hemorrhage, but not risk of subarachnoid hemorrhage (Table 2). In the fully-adjusted model (Model 2), compared with participants with 0–1 healthy factors, HRs (95% CI) of all-cause stroke for those with 2, 3 and 4 healthy factors were 0.85 (0.79–0.92), 0.71 (0.66–0.77), and 0.65 (0.59–0.72) (P for trend < 0.001), respectively (Table 2). This association was only slightly attenuated after further adjustment for BMI, diabetes, hypertension, and dyslipidemia (Model 3, P for trend < 0.001). Similar inverse associations were observed between score of the healthy factors and risk of ischemic stroke and intracerebral hemorrhage (Model 2: both P for trend values < 0.001) but not subarachnoid hemorrhage (P for trend = 0.48) (Table 2). We then assessed the association of healthy factors with all-cause dementia (2120 incidents) and its subtypes (820 cases of Alzheimer’s disease and 412 cases of vascular dementia). After multivariable adjustment, a higher score of healthy factors was associated with lower risk of all-cause dementia. Compared with those with 0–1 healthy factor, HRs (95% CI) of all-cause dementia for those with 2, 3 and 4 healthy factors were 0.74 (0.66–0.83), 0.64 (0.56–0.71), and 0.43 (0.39–0.51) (P for trend < 0.001), respectively (Table 3). Similar inverse associations were observed for Alzheimer’s disease and vascular dementia. The associations between combined healthy factors and risk of all-cause dementia, Alzheimer’s disease or vascular dementia were similar after further adjustment for BMI, diabetes, hypertension, and hyperlipidemia (Model 3 in Table 3).

Table 2.

Association of healthy factors with risk of stroke in participants from the UK Biobank

	Score of healthy factors				P for trend
	0-1 (n = 58,040)	2 (n = 130,834)	3 (n = 179,825)	4 (n = 106,284)
All stroke
No. of cases	957	1645	1668	722
Model 1 (HR [95% CI])	1.00 (Referent)	0.79 (0.73-0.86)	0.63 (0.59-0.69)	0.56 (0.51-0.62)	<.001
Model 2 (HR [95% CI])	1.00 (Referent)	0.85 (0.79-0.92)	0.71 (0.66-0.77)	0.65 (0.59-0.72)	<.001
Model 3 (HR [95% CI])	1.00 (Referent)	0.89 (0.82-0.96)	0.76 (0.70-0.83)	0.71 (0.64-0.78)	<.001
Ischemic stroke
No. of cases	790	1288	1271	534
Model 1 (HR [95% CI])	1.00 (Referent)	0.76 (0.69-0.83)	0.60 (0.55-0.65)	0.52 (0.47-0.58)	<.001
Model 2 (HR [95% CI])	1.00 (Referent)	0.82 (0.75-0.90)	0.68 (0.62-0.74)	0.62 (0.55-0.69)	<.001
Model 3 (HR [95% CI])	1.00 (Referent)	0.87 (0.79-0.95)	0.74 (0.68-0.81)	0.69 (0.62-0.78)	<.001
Intracerebral hemorrhage
No. of cases	155	272	287	116
Model 1 (HR [95% CI])	1.00 (Referent)	0.81 (0.67-0.99)	0.68 (0.55-0.82)	0.56 (0.44-0.71)	<.001
Model 2 (HR [95% CI])	1.00 (Referent)	0.84 (0.69-1.03)	0.72 (0.59-0.87)	0.60 (0.47-0.77)	<.001
Model 3 (HR [95% CI])	1.00 (Referent)	0.85 (0.69-1.04)	0.72 (0.59-0.89)	0.61 (0.48-0.79)	<.001
Subarachnoid hemorrhage
No. of cases	69	188	196	110
Model 1 (HR [95% CI])	1.00 (Referent)	1.19 (0.90-1.57)	0.90 (0.68-1.19)	0.87 (0.65-1.19)	.069
Model 2 (HR [95% CI])	1.00 (Referent)	1.27 (0.96-1.68)	1.01 (0.76-1.33)	1.02 (0.75-1.38)	.48
Model 3 (HR [95% CI])	1.00 (Referent)	1.23 (0.93-1.62)	0.95 (0.72-1.26)	0.94 (0.69-1.29)	.23

Model 1 was adjusted for age, sex, and ethnic background (White, Asian/Asian British, Black/Black British, mixed/unknown)

Model 2 was further adjusted for Townsend deprivation index (in quintile), smoking status (never, former, current), pack-years of smoking (for current smokers), alcohol drinking (never, former, current: <1, 1 to <3, 3 to <7, ≥7 drinks/week), and diet quality score (0-5)

Model 3 was further adjusted for body mass index (kg/m²), diabetes, hypertension, and hyperlipidemia

Table 3.

Association of healthy factors with risk of dementia in participants from the UK Biobank

	Score of healthy factors				P for trend
	0–1 (n = 58,040)	2 (n = 130,834)	3 (n = 179,825)	4 (n = 106,284)
All dementia
No. of cases	477	712	722	209
Model 1 (HR [95% CI])	1.00 (Referent)	0.69 (0.62–0.78)	0.58 (0.51–0.65)	0.38 (0.32–0.45)	< 0.001
Model 2 (HR [95% CI])	1.00 (Referent)	0.74 (0.66–0.83)	0.64 (0.56–0.71)	0.43 (0.39–0.51)	< 0.001
Model 3 (HR [95% CI])	1.00 (Referent)	0.74 (0.66–0.83)	0.64 (0.56–0.72)	0.44 (0.37–0.51)	< 0.001
Alzheimer’s disease
No. of cases	171	278	286	85
Model 1 (HR [95% CI])	1.00 (Referent)	0.75 (0.62–0.91)	0.63 (0.52–0.77)	0.44 (0.34–0.57)	< 0.001
Model 2 (HR [95% CI])	1.00 (Referent)	0.80 (0.66–0.96)	0.69 (0.57–0.84)	0.49 (0.38–0.64)	< 0.001
Model 3 (HR [95% CI])	1.00 (Referent)	0.78 (0.64–0.94)	0.67 (0.55–0.81)	0.47 (0.36–0.62)	< 0.001
Vascular dementia
No. of cases	107	144	124	37
Model 1 (HR [95% CI])	1.00 (Referent)	0.64 (0.49–0.82)	0.45 (0.35–0.59)	0.32 (0.22–0.47)	< 0.001
Model 2 (HR [95% CI])	1.00 (Referent)	0.69 (0.54–0.89)	0.52 (0.40–0.68)	0.39 (0.26–0.56)	< 0.001
Model 3 (HR [95% CI])	1.00 (Referent)	0.73 (0.57–0.94)	0.57 (0.44–0.75)	0.43 (0.29–0.64)	< 0.001

Model 1 was adjusted for age, sex, and ethnic background (White, Asian/Asian British, Black/Black British, mixed/unknown)

Model 2 was further adjusted for Townsend deprivation index (in quintile), smoking status (never, former, current), pack-years of smoking (for current smokers), alcohol drinking (never, former, current: <1, 1 to < 3, 3 to < 7, ≥7 drinks/week), and diet quality score (0–5)

Model 3 was further adjusted for body mass index (kg/m²), diabetes, hypertension, and hyperlipidemia

Subgroup and sensitivity analyses

The inverse associations between score of the healthy factors and all-cause stroke or dementia were largely consistent in various population subgroups (Supplementary Tables 4 and 5), except for a stronger association for all-cause stroke observed among individuals with dyslipidemia (P for interaction = 0.008). The results were similar after excluding participants with coronary heart disease or cancer at baseline, or participants who developed incident stroke or dementia within the first 4 years of follow-up (Supplementary Table 6). In addition, sensitivity analyses adjusting for education level (categorized as College or University degree, A levels/AS levels or equivalent, O levels/GCSEs or equivalent, and others) showed consistent results, indicating that the observed associations were not materially influenced by differences in educational attainment (Supplementary Tables 7 and 8).

To identify which combination of healthy factors is more strongly associated with a reduced risk of both stroke and dementia. We sequentially removed one of the healthy factors and studied the combined impact of the remaining 3 heathy factors on the risk of stroke and dementia. The results showed that an exclusion of any one of the healthy factors yielded quite similar results, both for stroke and dementia. The data were provided in Supplementary Table 9.

Association of individual healthy factors with risk of stroke and dementia

With full adjustment for potential confounders (Model 2) and mutual adjustment for other components of the healthy factors, each individual healthy factor was independently associated with lower risk of all-cause stroke and all-cause dementia (Table 4). Among the 4 healthy factors, the strongest association for stroke was observed for lower sedentary time (HR: 0.84; 95% CI: 0.79–0.89) and the strongest association for dementia was for higher grip strength (HR: 0.63; 95% CI: 0.58–0.70). Analysis for disease subtypes showed that each component of the healthy factors with risk of stroke was independently associated with lower risk of ischemic stroke (all P values ≤ 0.009) (Supplementary Table 10). Regular physical activity (P =.004) and higher grip strength (P =.007) but not lower sedentary time or appropriate sleep duration (both P values > 0.05) were associated with lower risk of intracerebral hemorrhage. None of the 4 healthy factors were associated with subarachnoid hemorrhage (Supplementary Table 10).

Table 4.

Association of healthy factors with risk of stroke and dementia in participants from the UK Biobank

Healthy Factors	HR (95% CI)	P value
All stroke
Regular physical activity	0.91 (0.85-0.96)	.003
Higher grip strength (above sex-specific median)	0.85 (0.80-0.90)	<.001
Lower sedentary time (<6 hours/day)	0.84 (0.79-0.89)	<.001
Appropriate sleep duration (7-8 hours/day)	0.87 (0.82-0.92)	<.001
All dementia
Regular physical activity	0.87 (0.79-0.96)	.005
Higher grip strength (above sex-specific median)	0.63 (0.58-0.70)	<.001
Lower sedentary time (<6 hours/day)	0.88 (0.81-0.97)	.008
Appropriate sleep duration (7-8 hours/day)	0.80 (0.73-0.87)	<.001

Results were adjusted for age, sex, ethnic background (White, Asian/Asian British, Black/Black British, mixed/unknown), Townsend deprivation index (in quintile), smoking status (never, former, current), pack-years of smoking (for current smokers), alcohol drinking (never, former, current: <1, 1 to <3, 3 to <7, ≥7 drinks/week), diet quality score (0-5), and the remaining three factors

In terms of incident Alzheimer’s disease and vascular dementia, with the exception of regular physical activity, the remaining 3 healthy factors were significantly associated with lower risk of both dementia subtypes despite of varying magnitudes in the associations (Supplementary Table 11).

Role of apolipoprotein E ε4 carrying status

The stratified analyses by APOE ε4 carrying status showed that a higher score of healthy factors was associated with lower risk of stroke and dementia regardless of APOE ε4 carrying status (P for interaction = 0.30 for stroke and 0.67 for dementia) (Fig. 1). As carrying APOE e4 is a well-known genetic risk factor for Alzheimer’s disease, we further performed analyses to investigate the associations between healthy factors with Alzheimer’s disease by APOE ε4 carrying status, and found no evidence of interactions (Supplementary Fig. 3, P for interaction = 0.84).

Fig. 1.

Association of healthy factors with risk of stroke and dementia according to apolipoprotein E genotype in participants from the UK Biobank. Analyses were conducted among 436,511 European white individuals who had genotyping data. Results were adjusted for age, sex, Townsend deprivation index (in quintile), smoking status (never, former, current), pack-years of smoking (for current smokers), alcohol drinking (never, former, current: <1, 1 to <3, 3 to <7, ≥7 drinks/week), diet quality score (0-5), and the first 10 principal components of ancestry. P values for interaction with apolipoprotein E genotype were .30 for stroke and .67 for dementia. n/N indicates the number of cases/participants

Discussion

Principal results

In this large prospective cohort study of middle-aged and older participants in the UK, a combination of healthy factors (meeting physical activity recommendations, grip strength above the sex-specific median, sleep duration of 7–8 h/day, and sedentary time < 6 h/day) was associated with lower risk of stroke and substantially lower risk of dementia. Participants with 4 of healthy factors had 35% lower risk of stroke and 57% lower risk of dementia. Both associations were independent of various sociodemographic factors, other lifestyles and major chronic diseases and did not appear to be modified by APOE ε4 carrying status. An increasing number of healthy factors also was associated with lower risk of stroke subtypes including ischemic stroke and intracerebral hemorrhage (but not subarachnoid hemorrhage) and major subtypes of dementia including Alzheimer’s disease and vascular dementia. Further analyses found that these healthy factors were associated with risk of stroke and dementia independently of each other.

Interpretations, implications, and comparisons with existing literature

Due to global aging, the prevalence of major disabling neurological disorders, especially stroke and dementia, sharply increases [37]. Governments will face increasing burden related to treatment, rehabilitation, and support services for stroke and dementia. The scarcity of established modifiable risk factors for dementia demonstrates that new knowledge is required to develop effective prevention and treatment strategies [1].

Among modifiable lifestyle factors, from a movement perspective, a 24-hour day comprises a sequence of time periods spent in movement behaviors that fall on a no/low-high intensity continuum, ranging from sleep to sedentary behavior and physical activity. Regular physical activity is a widely acknowledged protective factor for ischemic stroke [4, 6, 7], although evidence is much more limited for hemorrhagic stroke. The association of physical activity with dementia is also inconsistent [38, 39]. There is accruing evidence for a positive association between sedentary behavior and dementia by a recent meta-analysis [9]. Nonetheless, it needs to point out that none of the studies included in the abovementioned meta-analysis adjusted for physical activity and sleep duration, the two very important confounding factors.

Most previous studies have typically examined a single behavior with respect to the association with stroke and dementia [7–11, 39–41], and whether these behaviors are associated with stroke and dementia independently of each other remains unclear. To the best of our knowledge, no study to date has considered these above-mentioned behaviors comprehensively to investigate their associations with stroke and dementia and their major subtypes. Our findings showed that every individual component of the healthy factors was independently associated with lower risk of all-cause stroke and all-cause dementia. Moreover, a combination of regular physical activity, greater grip strength, appropriate sleep duration, and reduced sedentary time demonstrated a stronger and more robust association with lower risk of stroke and dementia than any one of these individual healthy factors in isolation. This observation highlights the potential cumulative benefits of adhering to multiple healthy factors in relation to reduced risk of stroke and dementia.

Due to substantial etiological differences between stroke and dementia subtypes, the disparities of healthy factors on these specific disease subtypes also need to be explored. We found that an increasing number of healthy factors were associated with lower risk of ischemic stroke and intracerebral hemorrhage but not subarachnoid hemorrhage. Although ischemic stroke and hemorrhage stroke share some risk factors, ischemic stroke is often related to atherosclerosis and thromboembolism, while hemorrhagic stroke is more influenced by hypertension, cerebral small vessel disease (intracerebral hemorrhage) [42] or aneurysm rupture (subarachnoid hemorrhage, in approximately 85% of cases) [43]. Our present study found that healthy factors may be particularly beneficial for the prevention of ischemic stroke and intracerebral hemorrhage, while the potential mechanisms underlying such differences in the associations for stroke subtypes need further investigation. We found that the 4 healthy factors were jointly associated with both Alzheimer’s disease and vascular dementia. When examining individual behaviors, the strongest associations were observed for higher grip strength, while regular physical activity showed the weakest associations. Grip strength serves as an important marker of overall muscle strength. A longitudinal study showed that people who maintained/increased time spent in moderate-to-vigorous physical activity did not experience any benefit in grip strength and it suggested that interventions with incorporation more muscle-strengthening exercises to gain the independent benefits of increased muscle strength [44]. Skeletal muscle has a well-recognized secretory function, releasing cytokines and peptides such as brain-derived neurotrophic factor (BDNF), interleukin-6 (IL-6), IL-8, and IL-15, which are involved in systemic inflammation and neuroplasticity. Declining muscle strength may lead to reduced expression of BDNF and insulin-like growth factor-1 (IGF-1), both of which are critical for learning, memory, and neural resilience [45–48]. Furthermore, grip strength has been linked to brain structure and function, possibly reflecting shared underlying mechanisms including chronic inflammation, vascular integrity, and neurodegeneration [49]. Therefore, higher grip strength may serve as a proxy for the integrity of the neuromuscular and central nervous systems and reflect a greater resistance to oxidative stress and systemic inflammation, ultimately contributing to preserved cognitive function. These findings suggest that different factors may have distinct associations with stroke and dementia, highlighting the need for subtype-specific prevention strategies that consider both the nature of the disease and the specific type of above-mentioned healthy factors.

The ε4 allele of APOE is a well-known genetic risk factor that contributes to the development of dementia [50, 51] and possibly the development of stroke [52]. However, this risk may be mitigated for most when adhering to a healthy lifestyle [53]. Our current study confirmed that the combination of healthy factors was associated with lower risk of stroke and dementia, even in individuals with APOE ε4 genotype.

Limitations

This study has several limitations. First, unlike the genetic variants, adherence to healthy factors was not randomly assigned, and residual confounding and reverse causation cannot be fully ruled out, despite adjustment for a range of covariates. Second, stroke and dementia outcomes were ascertained through linkage to hospital inpatient data, death registries, and primary care records, which may miss milder or undiagnosed cases, particularly for dementia. Individuals with healthier behaviors might also be less likely to seek medical care, potentially introducing detection bias that may attenuate the observed associations. Third, although the UK Biobank provides a large and rich dataset, it is a volunteer-based cohort, and participants are generally healthier and more socioeconomically advantaged than the general population. As such, the generalizability of our findings may be limited. Fourth, although stroke, dementia, and death are biologically interrelated, we treated death as a censoring event to align with prior literature. This approach may overlook competing risk effects, which future studies could address using alternative models such as Fine-Gray subdistribution hazards. Fifth, these healthy factors were measured only at baseline. The lack of repeated assessments limits our ability to examine behavioral changes or long-term trajectories over time. Sixth, dichotomizing continuous variables (e.g., grip strength) may reduce statistical power and obscure dose–response relationships. However, this approach is common in epidemiological research and facilitates interpretability. Future work could model these variables continuously to capture more nuanced associations.

Conclusions

Findings from the present study showed that an increasing number of healthy factors was associated with a further reduced risk of stroke and dementia, including their major disease subtypes, and these associations were independent of APOE ε4 risk. These findings suggest that future prevention strategies for stroke and dementia are needed to comprehensively take into account types of healthy factors and disease subtypes. Overall, this study encourages future health policies to promote a lifestyle with more elements of healthy factors for the prevention of stroke and dementia.

Abbreviations

APOE

apolipoprotein E

BMI

body mass index

ICD-10

International Classification of Diseases 10th revision

SBP

systolic blood pressure

DBP

diastolic blood pressure

Authors’ contributions

All authors contributed significantly to the conception, design, and analysis or interpretation of data. Tong Liu and Haili Tian contributed equally to this work and are joint senior authors. Tong Liu and Haili Tian supervised and coordinated the study. Li-Hua Chen, Cai-Long Chen and Yan Hong contributed equally to this work. Li-Hua Chen take responsibility for the integrity of the data and the accuracy of the data analysis.Concept and design: Li-Hua Chen, Cai-Long Chen, Yan Hong, Xi Yin, Fu-Rong Li, Guo-Chong Chen, Tong Liu and Haili TianAcquisition, analysis, or interpretation of data: Li-Hua Chen, Cai-Long Chen, Yan Hong, Xi Yin, Fu-Rong Li, Guo-Chong Chen, Tong Liu and Haili TianDrafting of the manuscript: Li-Hua Chen, Cai-Long Chen, Yan Hong, Xi Yin, Fu-Rong Li, Guo-Chong Chen, Tong Liu and Haili TianCritical revision of the manuscript for important intellectual content: All listed authors.

Funding

This study was supported by National Natural Science Foundation of China (grant number 81870874, 82171229, 81803307); 2017 Chinese Nutrition Society (CNS) Nutrition Research Foundation-DSM Research Fund (grant number 2017-040); Jiangsu Key Laboratory of Preventive and Translational Medicine for Sex Differences in Physical Activity and Incident StrokeMajor Chronic Non‐communicable Diseases (grant number: KJS2439).

Data availability

Researchers registered with UK Biobank can apply for access to the database by completing an application. (https://www.ukbiobank.ac.uk/enable-your-research/apply-for-access) The code is available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study is based on data from the UK Biobank study that received approval from the National Information Governance Board for Health and Social Care and the National Health Service North West Multicenter Research Ethics Committee. All participants gave written informed consent before enrollment in the study, which was conducted in accord with the principles of the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.