Health behavior of young patients with ischemic stroke in Estonia: A score of five factors

Abstract Background Behavioral risk factors are common among young patients with stroke. This study aimed to compare the health behavior of patients and healthy controls and develop a combined risk score of health behavior. Methods The health behavior of patients aged 18–54 years who suffered an ischemic stroke from 2013 to 2020 in Estonia was compared to the Health Behavior among Estonian Adult Population 2014 study sample. We chose five risk factors for comparison: smoking status, body mass index, physical exercise, diet (salt use and vegetable consumption), alcohol intake (quantity and frequency), and composed a summary score. Results Comparing 342 patients and 1789 controls, daily smoking (49.0% vs. 22.7%), obesity (33.4% vs. 15.9%), low physical activity (< twice/week) (72.2% vs. 60.5%), excessive salt use (8.6% vs. 4.5%), and frequent alcohol use (≥ weekly) (39.9% vs. 34.0%) were more prevalent among patients. The differences in infrequent vegetable consumption (<6 days/week) and excessive alcohol consumption (7 days, >8 units/females, >16 units/males) were not significant. The observed differences were similar for age groups 18–44 years and 45–54 years. The average Health Behavior Stroke Risk Score (0–10) was 4.6 points (CI 4.4–4.8, SD ± 1.97) for patients and 3.5 points (CI 3.4–3.6, SD ± 1.90) for controls. Conclusions Before stroke, young patients displayed significantly worse health behavior than the general population. The largest differences were found for smoking and obesity, and a cumulation of risk factors was observed via the HBSR score.


INTRODUCTION
Lifestyle choices have a significant impact on health, and many studies have emphasized the connection between modifiable risk factors and stroke. It has been demonstrated that more than 80% of strokes are related to hypertension, current smoking status, obesity, unhealthy diet, and physical inactivity (O'Donnell et al., 2010) and that 74% of the stroke burden is due to behavioral factors (Feigin et al., 2016). These traditional stroke risk factors are also common in young patients with stroke (Putaala, 2016), and they are surprisingly prevalent regardless of the stroke etiology (Maaijwee et al., 2014). This may be related to the reported increase in stroke incidence in young adults over the last decades (Ekker et al., 2019). In Estonia, the incidence and case fatality of stroke in young adults are higher than in other high-income countries , and the traditional risk factors are common (Schneider et al., 2017;Vibo et al., 2021). Previous studies on young patients with stroke mainly focused on clinical risk factors, and only a few have evaluated health behavior. More complex topics, for example, diet, are often not included, and there is no standard evaluation of all the factors.

OBJECTIVES
The present study aimed to assess the health behavior of young patients with stroke, compare this to the general population, and develop a simple health behavior summary score to evaluate composite behavioral stroke risk.

METHODS
The Estonian Young Stroke Registry is a prospective hospital-based ongoing database, which comprises all consecutive patients with a discharge diagnosis of ischemic stroke (94.9% first-ever, 6.1% recurrent) and aged 18−54 years, as described previously . While hospitalized for acute stroke, written informed consent was obtained from the participants, and they completed a self-report questionnaire about their health behavior before the stroke. The questionnaire was designed to match the Health Behavior among Estonian Adult Population (HBEAP) study of 2014. The HBEAP study is conducted every 2 years by the National Institute for Health Development (NIHD) in Estonia, and this is a postal questionnaire with a stratified random sample of 5000 people aged 16−64 years (Tekkel & Veideman, 2015 with a score of 10 corresponding to the highest risk. The description of the score is shown in Table S1. For a binary variable, scores were divided into low (0−5) and high (6−10).

Statistical methods
Comparisons between patients with stroke and controls were performed using the Z-test, and Bonferroni correction was used for multiple comparisons. The odds ratios (ORs) for each behavioral factor were calculated using logistic regression, presented in both crude and adjusted models, expressed with 95% confidence intervals (CI). The OR for having a high HBSR score was calculated using the same method, and the mean scores were compared using a two-tailed t-test. All statistical analyses were performed using Stata version 16.1 (StataCorp, College Station, TX, USA).

RESULTS
In all, 436 young patients with ischemic stroke were recruited in the registry between January 1, 2013 and December 31, 2020. Of these,

Health behavior stroke risk score
In young patients with stroke, the average HBSR score was significantly higher than that in the general population ( (Table S2).

DISCUSSION
Our prospective study showed that young patients displayed significantly worse pre-stroke health behavior than the general population in most of the studied domains. Following adjustments for confounding factors (age, sex, education, and marital status), smoking and obesity were more prevalent among patients with stroke than controls. The HBSR score provides comprehensive behavioral stroke risk assessment through an aggregate score that was higher in patients than in controls, indicating worse health behavior.
There are minimal reports that have focused on the health behavior of young patients with ischemic stroke, and there is a lack of casecontrol studies. Our results are in line with those of previous studies reporting smoking and physical inactivity as the most common behavioral risk factors among young patients with stroke (Putaala, 2016).
However, while the earlier studies frequently assessed smoking status (Aigner et al., 2017;Goeggel Simonetti et al., 2015;Kivioja et al., 2018;Mitchell et al., 2015;Putaala et al., 2012;Renna et al., 2014;Von Sarnowski et al., 2013), physical inactivity was rarely reported. The most common risk factor in our study was physical inactivity, with more than 72% of the patients and 61% of controls engaging in exercise less than twice a week. Though this criterion is even less strict than that presented in stroke primary prevention guidelines (Meschia et al., 2014), it only involves physical activity during exercising and does not account for action at work or during the commute. In Germany, 49% of young patients with stroke had low levels of physical activity (20−30 min <3 times a week), compared to 19% of controls (Aigner et al., 2017). In addition, a recent study found that excess sedentary leisure time (≥ 8 hours/day) was associated with an increased risk of long-term stroke in young adults (Joundi et al., 2021). In this study, almost half of the patients were daily smokers compared with less than a quarter of controls. The difference between our patients and controls was higher than that in similar studies from Germany (48% vs. 35%) (Aigner et al., 2017), Finland (44% vs. 31%) (Kivioja et al., 2018), and the United States (45% vs. 29%) (Mitchell et al., 2015). In the general population, the prevalence of smoking has declined by almost half during the last two decades in Estonia, as it was 18% in 2020 (Reile & Veideman, 2021); however, it remains high in young patients with stroke . Based on the high prevalence of smoking among patients, regular documentation of smoking status and cessation support is vital for both primary and secondary prevention.
The weekly alcohol consumption was observed in 40% of the patients and 34% of controls; the difference was statistically significant only before adjusting for age and sex. While differences in the frequency of alcohol intake were noted, the average amount of alcohol consumed (in the last 7 days) was similar in the two groups. The definition of excessive alcohol consumption differs between studies.
When defined as >5 alcoholic drinks per day or occasion at least once a month, heavy episodic consumption was recorded in 33% of patients F I G U R E 2 Distribution of the Health Behavior Stroke Risk Score in young patients with stroke and the general population of Estonia.
versus 18% of controls in a German sub-cohort of the SIFAP1 study, and the overall prevalence was very similar (33%) in the multinational SIFAP1 study (Aigner et al., 2017). Heavy drinking, defined as an estimated intake of >200 g of pure alcohol per week, has been reported in 14% of patients with stroke in Finland (Putaala et al., 2009). Similarly, in our previous study, 16% of patients reported the constant use of alcohol .
The obesity rate of our patients (33%) was twice as high as that of controls, and it was higher than in most earlier studies of young patients with stroke (11%−22%) (Aigner et al., 2017;Putaala et al., 2012;Renna et al., 2014;Von Sarnowski et al., 2013). To date, only one study has shown a higher prevalence (39.5%), but there was also notably more obesity among controls (29%) (Mitchell et al., 2015). The rate of obesity may reflect the proportion of obese population in a specific country, and the prevalence of obesity is increasing in Estonia (Reile et al., 2020). While BMI is not a behavioral factor in itself but only reflects excessive caloric consumption, we decided to use this factor in our study because it seems to be critical considering the increase in obesity and there is lack of standards for measuring excessive caloric consumption itself.
It is challenging to evaluate diet using only a few questions. We analyzed frequent vegetable and low salt consumption as indicators of a healthy diet. The lack of differences in vegetable consumption between patients and controls was unexpected, as consuming more vegetables has been associated with lower stroke risk in long-term prospective studies (Aune et al., 2017). It is possible that recall bias may lead to errors in reporting vegetable consumption frequency and that social desirability bias causes patients to report their behavior toward healthier at the hospital (compared to anonymous replies mailed by controls).
As for salt use, we found no studies that focused on the habit of adding salt to ready-made meals or any studies on salt consumption, specifically in young patients with stroke. As per an estimate, patients who add salt to food before tasting consume approximately 10 g of salt per day, while the body requires ∼0.5 g (Spence, 2019). The stroke primary prevention guidelines recommend a diet low in sodium and rich in fruits and vegetables (Meschia et al., 2014). Considering that a healthy diet could notably lower stroke risk (Spence, 2019) (Manuel et al., 2015). That score included the same health behavior aspects as our study, but the variables differed. Health behavior and stroke history were documented during a national survey, and individuals were followed for 5-year stroke incidence, while we used patients with stroke from the hospital who reported their recent health behavior at the time of stroke. Additionally, they did not focus on young patients with stroke, including 20-to 83-year olds (Manuel et al., 2015).
One of the strengths of our study is recording health behavior during the initial days after stroke onset; therefore, the results reflect the behavior immediately before stroke. Our sample is representative, including all consecutive patients from 8 years from a tertiary stroke center and using a population-based control group. The aggregate score allows comparisons of overall health behaviors and assesses patients' lifestyles.
A limitation of our study is that we did not use matched controls; however, we used adjusted models to correct for confounding factors.
As patients in our sample had milder strokes than non-respondents, our results might not represent patients with more severe strokes, but this is more likely to decrease the differences between patients and controls than increase them. Social desirability could have affected the answers, as the control subjects answered the questionnaire anonymously, but the patients with stroke did so during their hospital stay.
In choosing our indicator variables, we used questions that are easy to answer and identical to the Estonian HBEAP study; however, these are also more robust and not specifically tested in stroke prevention.
In conclusion, our study demonstrated that young patients with stroke displayed significantly worse health behaviors than the general population in Estonia. As young patients have many decades to benefit from behavior change, more emphasis should be placed on informing patients of these modifiable risk factors and helping enforce these changes (e.g., medication and counselling for smoking cessation, guidelines for exercise and diet). The HBSR score can be used to provide feedback to individuals about the degree of modifiable stroke risk factors in their lifestyle or to compare groups. Further studies are required to evaluate the relationship between HBSR score and stroke risk.