Stroke remains a leading cause of death and disability worldwide [1]. More than half of stroke survivors are afflicted by serious sequelae such as aphasia, hemiparesis, depression and anxiety which can drastically impair the overall quality of life (QoL) [2, 3]. Several risk factors have been identified for stroke which can be split into several categories: (1) demographic factors, which include age, sex, ethnicity and socioeconomic status. (2) Health related factors such as hypertension, diabetes mellitus, high body mass index (BMI), high lipoprotein (a) and prior occurrence of cardiovascular disease (CVD) [4, 5, 6]. (3) Lifestyle related factors such as alcohol and tobacco use, physical activity, and sedentary time [7, 8, 9, 10, 11]. In addition to these factors, there is a growing selection of literature on the possible effects of personality on the risk of developing stroke. This research between personality and disease has been coined “personological epidemiology” [12].

The Five-Factor Model, also known as the “Big Five personality traits”, is perhaps the most robust and well-known method of characterizing a person’s psychological and behavioral dispositions [12, 13]. The model defines five superordinate bipolar factors, or personality traits: (1) extraversion, (2) agreeableness, (3) conscientiousness, (4) openness to experience, and (5) neuroticism. The five traits tend to remain stable throughout our lifetime and are present in over 50 cultures, which are the key aspects adding to the model’s broad applicability [14, 15]. Prior work has identified conscientiousness as a strong predictor for reduced all-cause mortality, while openness reduces CVD mortality in particular [16]. Neuroticism and extraversion have generated mixed results, with both positive and negative associations with general health effects [16]. However, agreeableness does not appear to have a strong association with health [16].

To date, personological epidemiological studies on stroke have primarily focused on stroke sequelae and post-stroke QoL. For example, neuroticism and individual facets of this trait have been associated with an increased risk for post-stroke depression and an overall reduced QoL [17, 18]. Additionally, stroke survivors with a strong degree of neuroticism have higher levels of fatigue and score lower on the Barthel scale measuring performance in activities of daily living [19]. On the other hand, low extraversion has been associated with an increased post-stroke apathy [18]. In relation to this, the type D (distressed) personality, which is characterized by low extraversion and high neuroticism, has lower scores for health promoting behaviours, higher post-stroke disability scores and lower QoL after stroke [20]. Unfortunately, studies on the association between personality and risk of developing stroke have been limited to cases of stroke mortality, thus excluding stroke morbidity that could potentially severely impact an individual’s QoL [21, 22, 23].

This study aims to examine potential associations between personality and stroke risk, regardless of mortality, both cross-sectionally and longitudinally, using the large-scale UK Biobank cohort (UKB) [24]. The exceptionally detailed UKB contains data from over 500,000 participants and follow-up data over the course of approximately 7 years. As neuroticism is the only Big Five personality trait that is included in the UKB, we aimed to identify phenotypes in the dataset that can be implemented as proxies to represent the other traits. Based on previous literature we expected to find a negative correlation between conscientiousness-like traits and stroke risk. Moreover, existing literature is unclear about the role of extraversion and neuroticism on stroke risk. We expected potential correlations between these traits and stroke risk but were uncertain whether these will be positive or negative. As a subgoal, we also performed a sex specific analysis in order to identify associations that are specific to females or males. This was of interest as none of the previously mentioned personological epidemiological stroke risk studies had assessed sex specific associations. Because personality remains relatively stable throughout one’s life, personality might be used to identify at-risk individuals at an early stage. This could be an important step into taking preventive measures earlier on in life. At the same time, identification of personality traits associated with stroke risk could support in the development of personality-targeted interventions.

After exclusions (Fig. 1), the study included 482,535 participants. Of these participants, 6793 (1.41%) had a stroke before baseline assessment (prevalent cases), and 14,574 (3.02%) had missing data on stroke prevalence and incidence. Participants with prevalent stroke or missing stroke data were excluded from the longitudinal analysis. Thus, the longitudinal analysis contained the remaining 461,168 participants, which included 3312 (0.72%) incident cases of stroke. The dataset consisted of 55.2% female and 44.8% male participants, with a respective mean age of 56.3 ($\pm$ 8.0) and 56.6 ($\pm$ 8.2) years. The number of both prevalent and incidence stroke was higher in males (57.0% and 57.5% respectively). Participant ethnicity was primarily white (94.1%) (Table 2).

3.1 Personality Proxy Associations with Stroke Prevalence

The associations between the Big Five personality trait proxies and stroke prevalence are illustrated in Fig. 2. Model I, adjusting for demographic characteristics, showed that neuroticism [OR = 1.21; (95% confidence interval (CI): 1.18; 1.24); (p = 2.98 $\times$ 10${}^{-44}$)] and its proxy nervousness [OR = 1.24; 95% CI = (1.21; 1.27); (p = 1.86 $\times$ 10${}^{-73}$)], were both associated with an increased risk of having had a stroke event at baseline (Fig. 2). Curiosity [OR = 0.95; 95% CI = (0.93; 0.97); (p = 5.63 $\times$ 10${}^{-5}$)], diligence [OR = 0.77; 95% CI = (0.75; 0.79); (p = 9.14 $\times$ 10${}^{-98}$)], warmth [OR = 0.82; 95% CI = (0.80; 0.84); (p = 2.38 $\times$ 10${}^{-64}$)] and sociability [OR = 0.79; 95% CI = (0.77; 0.81); (p = 3.85 $\times$ 10${}^{-81}$)] were all associated with lower stroke risk at baseline. Model II, which extended model I with health-related factors, revealed the same trends as model I. The odds ratio for stroke at baseline was higher for participants with high neuroticism [OR = 1.20; 95% CI = (1.16; 1.23); (p = 1.44 $\times$ 10${}^{-37}$)] and nervousness [OR = 1.22; 95% CI = (1.19; 1.25); (p = 1.72 $\times$ 10${}^{-58}$)]. Conversely, higher curiosity [OR = 0.95; 95% CI = (0.93; 0.98); (p = 2.45 $\times$ 10${}^{-4}$)], diligence [OR = 0.80; 95% CI = (0.78; 0.82); (p = 7.06 $\times$ 10${}^{-72}$)], warmth [OR = 0.83; 95% CI = (0.81; 0.85); (p = 7.53 $\times$ 10${}^{-57}$)] and sociability [OR = 0.82; 95% CI = (0.80; 0.84); (p = 2.11 $\times$ 10${}^{-59}$)] were associated with lower prevalence of stroke. Model III, the fully adjusted model that extends model II with lifestyle factors, continued to show that neuroticism [OR = 1.16; 95% CI = (1.12; 1.19); (p = 2.33 $\times$ 10${}^{-19}$)] and nervousness [OR = 1.19; 95% CI = (1.15; 1.22); (p = 8.23 $\times$ 10${}^{-33}$)] were associated with a higher stroke occurrence at baseline. Additionally, diligence [OR = 0.83; 95% CI = (0.81; 0.86); (p = 1.16 $\times$ 10${}^{-33}$)], warmth [OR = 0.85; 95% CI = (0.83; 0.88); (p = 1.91 $\times$ 10${}^{-28}$)] and sociability [OR = 0.85; 95% CI = (0.83; 0.88); (p = 1.23 $\times$ 10${}^{-25}$)] are associated with a lower stroke occurrence at baseline. Contrary to the results adjusted by model I and II, curiosity [OR = 0.99; 95% CI = (0.96; 1.02); (p = 0.52)] was not significantly associated with stroke risk at baseline in model III.

Fig. 2.

Odds ratios (OR) and hazard ratios (HR) for prevalent and incident stroke cases and each of personality trait proxies. ORs and HRs were obtained by performing logistic regression and Cox’s proportional hazard regression respectively. The three models adjusted for demographic characteristics, health-related measurements and lifestyle factors (see methods). The UKB variable Neuroticism is included to validate our related “Nervousness proxy”.

Sex-Stratified Personality Proxy Associations with Stroke Prevalence

To examine potential sex-specific differences in the associations of the personality traits and stroke prevalence, a sex-stratification was performed (Fig. 3). In the fully adjusted model, females and males showed similar trends. In females, both neuroticism [OR = 1.16; 95% CI = (1.10; 1.22); (p = 9.72 $\times$ 10${}^{-9}$)] and nervousness [OR = 1.20; 95% CI = (1.14; 1.25); (p = 2.88 $\times$ 10${}^{-15}$)] were associated with increased stroke prevalence. Diligence [OR = 0.83; 95% CI = (0.79; 0.87); (p = 2.93 $\times$ 10${}^{-15}$)], warmth [OR = 0.86; 95% CI = (0.83; 0.90); (p = 2.93 $\times$ 10${}^{-11}$)] and sociability [OR = 0.84; 95% CI = (0.81; 0.88); (p = 3.38 $\times$ 10${}^{-13}$)] were associated with a decreased stroke prevalence. Curiosity [OR = 0.98; 95% CI = (0.94; 1.03); (p = 0.393)] was not significantly associated with stroke prevalence. In males, both neuroticism [OR = 1.16; 95% CI = (1.11; 1.20); (p = 7.58 $\times$ 10${}^{-12}$)] and nervousness [OR = 1.18; 95% CI = (1.14; 1.22); (p = 1.09 $\times$ 10${}^{-18}$)] were associated with an increased risk of stroke prevalence. Diligence [OR = 0.84; 95% CI = (0.80; 0.87); (p = 4.98 $\times$ 10${}^{-20}$)], warmth [OR = 0.85; 95% CI = (0.81; 0.88); (p = 1.17 $\times$ 10${}^{-18}$)] and sociability [OR = 0.86; 95% CI = (0.83; 0.90); (p = 4.71 $\times$ 10${}^{-14}$)] were associated with a decreased risk of stroke at baseline whereas no associations were observed for curiosity [OR = 1.00; 95% CI = (0.96; 1.04); (p = 0.907)].

Fig. 3.

Odds ratios (OR) and hazard ratios (HR) for prevalent and incident stroke stratified by sex (female (F), male (M)). A notable difference, when compared to the non-stratified results, is that nervousness might pose a potential risk factor in stroke incidence in females but not in males.

To our knowledge, this is the first UK Biobank (UKB) study to investigate the effect of personality traits on the risk of experiencing non-fatal stroke events. We found that diligence and sociability were associated with a lower risk for stroke. However, no significant associations were observed between the remaining personality traits, including neuroticism, and the risk of developing stroke.

Personality is tightly interlinked with several health-related behaviours, such as exercise and alcohol consumption, which are also deemed strong determinants for stroke risk [7, 39, 40]. Thus, it is plausible that personality traits themselves can influence the risk of stroke throughout an individual’s lifetime. In both the cross-sectional and longitudinal analyses, diligence was associated with a lower risk for stroke. This parallels prior findings by Jokela et al. [23] where conscientiousness was protective against stroke mortality in the pooled results from US-based cohorts of approximately 25,000 participants. Conscientious individuals are often self-disciplined, cautious, persistent, and have been shown to be more physically active, have healthier eating habits, smoke less and engage in fewer injury prone behaviours (e.g., risky driving, violence) [41, 42]. As we accounted for all known conscientiousness-related behaviours that could affect stroke risk in UKB, with the exception of injury prone behaviours, it is plausible that high conscientiousness confers psychosocial or psychobiological protections against stroke [43, 44].

When it comes to extraversion, characterised by outgoing, talkative and energetic qualities, the general health related benefits are inconsistent between studies [16]. In the current analysis, the sociability proxy revealed a lower risk for both prevalent and incident stroke. However, neither Nakaya et al. [22] nor Shipley et al. [21] found effects of extraversion on stroke mortality in Japanese and British populations, respectively. In contrast, Jokela et al. [23] deemed extraversion to be a risk factor for stroke mortality. A possible explanation for these discrepancies may be the use of stroke mortality as the outcome, rather than the current approach including non-fatal stroke events. The inclusion of non-fatal stroke events results in more cases and, in turn, a higher statistical power to detect potential associations. The contradicting results may also be due to cultural differences between the study samples, as extraversion may lead to differing social effects in extraverted or introverted nations [45, 46]. According to Hans Eysenck, extraverts and introverts seek different levels of social interaction due to their differences in naturally occurring arousal levels [47, 48]. Social activity has been associated with positive cardiovascular effects, lower stroke risk, and higher overall survival rates in elderly [49, 50, 51, 52]. Thus, it is likely that the possible beneficial effects of extraversion are related to social behaviours.

Just as certain personality traits may be protective against stroke, the extent to which neuroticism has adverse health outcomes has been debated in the realm of stroke and cardiovascular disease [53]. Neuroticism, and its proxy nervousness, were significantly associated with an elevated prevalence of stroke, but not with the incidence of stroke in the fully adjusted model. Previous work has reported neuroticism to be a risk factor for coronary heart disease, but this has not been established for stroke [21, 22, 23]. Given that neuroticism, which encompasses strong negative affect and social inhibition, may increase in response to traumatic events [54], it may be of more interest after a stroke has occurred [55, 56]. Indeed, high neuroticism has been linked with decreased QoL after stroke, fewer health promoting behaviours, and increased risks of post-stroke depression, fatigue and apathy [17, 18, 19, 20].

The strengths of this study included the extensive data contained in the UKB, including demographic information, official medical records, lifestyle information, various health and anthropomorphic measurements and personality related questionnaires. Particularly, the official medical records allowed us to reliably identify stroke cases, but also participants with other conditions such as diabetes or certain psychological disorders. The UKB consisted of information from over half a million participants and contained 3312 confirmed cases of stroke incidence over a median follow-up time of 7 years. This provided high statistical power and the possibility to perform additional sex-specific analyses. In addition, while prior studies focused primarily on stroke mortality our study included stroke morbidity as well. Finally, using proxies for all Big Five personality traits allowed us to cover various aspects of human personality, albeit in a fairly broad manner.

One of the limitations of this study was the inference of the Big Five personality traits using proxies. Although great care was taken into the development of these proxies, their association to the defined Big Five personality traits has not been extensively verified. However, the construction of the proxies followed those of other established short-form personality trait questionnaires [57]. For example, the Midlife Development Inventory [58] and the Big Five Inventory [59], consist of 4 to 5 questions to characterise each trait. Although some neuroticism related variables were utilized for all the proxies, their usage has been adapted (i.e., by using the inverse answer option). By combining these adapted variables with other relevant variables, we believe the proxies approximate the respective traits. The broadness of the Big Five personality traits can be considered another limitation of the study. Nevertheless, it allowed us to pinpoint conscientiousness-like and extraversion-like traits as protective factors against stroke, thus suggesting future studies to focus on these traits. Finally, a limitation within the UKB cohort is the high prevalence of the white ethnicity (~94%).

To conclude, analysis of over 3300 stroke cases within the UK Biobank revealed that diligent and sociable individuals were associated with a lower risk of stroke. The potential protective effects of diligence and sociability were not fully explained by health or lifestyle factors alone. Clarifying which particular facets of conscientiousness and extraversion may influence the risk of stroke, could improve support programs and help prevent the devastating effects of stroke.

BFI, big five inventory; BMI, body mass index; CI, confidence interval; CVD, cardiovascular disease; DBP, diastolic blood pressure; EPQ, Eysenck personality questionnaire; HR, hazard ratio; ICD, international classification of disease; MET, metabolic equivalent per time; MVPA, moderate-to-vigorous physical activity; OR, odds ratio; QoL, quality of life; SBP, systolic blood pressure; SD, standard deviation; TDI, Townsend deprivation index; UK, United Kingdom; UKB, UK Biobank.

MJTR, ADD, GR and HBS designed the research study. MJTR performed the research. ADD, GR and HBS provided help and advice on methodology. MJTR analysed the data and wrote the manuscript. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript.

The UKB study was approved by the North West Multi-Centre Research Ethics Committee, the Regional Ethics Committee of Uppsala, Sweden and all participants provided written informed consent (approved UKB application no. 25308 and 30172).

We thank the participants of UK Biobank that contributed to this valuable resource.

GR was supported by Svenska Sällskapet för Medicinsk Forskning (Postdoctoral fellowship 2019), and HBS was supported by grants from AFA Insurance Foundation (160345) and Swedish Research Council (VR 2014-02912).

The authors declare no conflict of interest.