This study evaluated data from the Hungarian Epidemiological Panel Survey, a prospective study with a four-year time lag focusing on the biopsychosocial determinants of chronic diseases of public health importance among Hungarian adults. The study protocol was approved by the Ethical Committee of Semmelweis University. In case of the first wave of data collection, the sampling frame was the National Population Register. Municipalities with a population over 10,000 inhabitants, as well as a representative random sample of smaller communities throughout the country were included. At follow-up, approximately 4 years later, 4524 adult individuals were again interviewed face-to-face. At this time, the sampling frame was limited to those individuals who had given informed consent at baseline to be re-interviewed at follow-up; the refusal rate was 23.7%. The representativeness of the wave 2 sample by age, gender, and sub-region was obtained by a weighting process. A detailed description of the study methodology is available elsewhere (Susánszky et al., 2007).

From the entire sample, the data of 816 individuals (65.3% female, Mage=43.2 years, SD=14.7 years) who reported at baseline that they had been treated for a vascular event were selected for this study. Among those selected, 395 (48.4%) participants reported treatment for at least one additional vascular event during the four-year follow-up period.

The occurrence of cardiovascular events, as well as other vascular events, at baseline was assessed by three questions: “Have you ever been treated for (1) a cerebrovascular disease; (2) heart attack; (3) other heart or vascular disease?” Initial respondent feedback suggested that more examples would be beneficial in helping participants answer the questions; as such, the wording was modified for the follow-up assessment. At this point, the reoccurrence of vascular events was assessed by two questions: “Since the previous data collection period, have you been treated for (1) cardiac diseases such as coronary artery disease, cardiomyopathy, ischemic heart disease, chest pain (angina), congestive heart failure, or myocardial infarction; (2) other vascular diseases such as transient ischemic attack, apoplexy, thrombosis, embolism, ischemic stroke, or hemorrhagic stroke?” The different vascular event groups were collapsed in the analyses to increase statistical power.

All independent variables of the present analyses were assessed at baseline. Vital exhaustion was assessed by the shortened Hungarian version (MQ-S; Kopp et al., 2010) of the Maastricht Questionnaire. With respect to validity, the shortened Hungarian version was reported to correlate very strongly with the full-length, 21-item questionnaire (r=.94; Kopp, Falger, Appels, & Szedmák, 1998). The internal reliability coefficient of this nine-item assessment tool was good in the present sample (Cronbach's α of .77).

Depressive symptomatology was measured by the shortened Hungarian version (Rózsa, Szádóczky, & Füredi, 2001) of the Beck Depression Inventory (BDI). As reported by Kopp and colleagues, this adaptation contained 9 items (symptoms) from the 21-item version of the BDI and showed a very strong correlation with total BDI scores (r=.92; Kopp et al., 1998). In the present sample, internal consistency of the measure was excellent (Cronbach's α=.83).

Anxiety was measured by the seven items related to anxiety from the Hungarian version (Muszbek et al., 2006) of the Hospital Anxiety and Depression Scale. Internal consistency of the measure was good in the present sample (Cronbach's α=.75). Hostility was assessed by the Hungarian version (Kopp et al., 1998) of the Shortened Cook-Medley Hostility Scale. Internal consistency of this five-item tool was suboptimal in the current sample (Cronbach's α=.64).

A number of conventional risk factors for vascular diseases were also measured, including age (as a continuous variable), educational attainment (six categories), hazardous drinking, smoking status (smoker versus non-smoker), physical activity, and body mass index. Hazardous drinking was assessed by the Hungarian version (Gerevich, Bácskai, & Rózsa, 2006) of the ten-item Alcohol Use Disorder Identification Test (AUDIT), the scores of which were dichotomized using the standard cut-off score of eight. Physical activity was assessed by a single question (“How often are you involved in doing sports; e.g., swimming, jogging, cycling, football, aerobic, etc.?”), which best represented a larger number of questions on physical activity asked in a previous wave of the survey. Response options were 0=never, 1=irregular, once in a while; 2=less than monthly; 3=once in a month; 4=once per week; 5=several times per week; and 6=daily. These were transformed into a dichotomous variable (“0–3” versus “4–6”) in accordance with recent literature pointing out that physical activity, even on a weekly basis, can significantly decrease cardiovascular risk (Després, 2016).

All analyses were performed using the SPSS 23.0 statistical package (SPSS Inc., Chicago, Illinois). To investigate the role of vital exhaustion in the reoccurrence of vascular events, a set of hierarchical multiple logistic regression analyses were run. In addition to vital exhaustion, the first model contained the traditional sociodemographic and lifestyle factors relevant for vascular morbidity, followed by depression (second model), and anxiety and hostility (third model).