Semmelweis University - School of PhD Studies Mental Health Sciences Behavioural Sciences

BIOPSYCHOSOCIAL AND BEHAVIOURAL CORRELATES OF CORONARY HEART DISEASE

Réka Baranyai, M.D.

Supervisor and Head of Doctoral Program: Head of Doctoral School:

Research Setting:

Prof. Mária Kopp, M.D., D. Sc. Prof. István Bitter, M.D., D. Sc.

Institute of Behavioural Sciences Semmelweis University Budapest Central Institute of Mental Health University of Heidelberg

Final Examination Board:

Prof. László Tringer, M.D., Ph.D. Ádám Tahy, M.D., Ph.D. Szabolcs Kéri, M.D., Ph.D.

Official Reviewers:

Ádám Tahy, M.D., Ph.D. János Réthelyi, M.D., Ph.D.

Budapest 2006

TABLE OF CONTENTS 1. INTRODUCTION: IMPORTANCE OF THE TOPIC: KEY POINTS..............................………………… 2. OVERVIEW OF THE LITERATURE 2.1. CHARACTERISTICS OF PHYSICAL AND MENTAL WELL-BEING…………………………….… 2.2. EVIDENCE OF CONNECTIONS BETWEEN CARDIOVASCULAR DISEASE AND PSYCHOSOCIAL FACTORS…………………………………………………………………………...……… 2.2.1. Psychological Factors 2.2.1.1. Psychological Risk Factors…………………………………………….…… 2.2.1.2. Positive Psychological Factors..................................................….................. 2.2.2. Social Factors…………………………………………………………………..…….. 2.3. POSSIBLE PATHWAYS………………………………………………………………………… 2.3.1. Health Behaviour Pathway………………………………………………………......... 2.3.2. Physiological Pathway of Psychosocial Symptoms: Focusing on Platelet Activation.. 3. AIMS ……………………………………………………………………………………………… 3.1. STUDY 1. HEALTH BEHAVIOUR, HEALTH BELIEFS, RISK AWARENESS AND PSYCHOSOCIAL WELL-BEING IN UNIVERSITY AND COLLEGE STUDENTS IN HUNGARY…………..……….. 3.2. STUDY 2. POSSIBLE TARGETS OF INTERVENTIONS AFTER CORONARY EVENTS – LESSONS LEARNED FROM HUNGAROSTUDY 2002 …………………………………………………. 3.3. STUDY 3. PLATELET REACTIVITY IN DEPRESSED PATIENTS……………….………………… 4. METHODS 4.1. STUDY 1. HEALTH BEHAVIOUR, HEALTH BELIEFS, RISK AWARENESS AND PSYCHOSOCIAL WELL-BEING IN UNIVERSITY AND COLLEGE STUDENTS IN HUNGARY……………………. 4.1.1. Study Design and Subjects…..………………………………………………………... 4.1.2. Questionnaires….………………………………………...……………………….….. 4.1.2.1. Health Behaviour……………………………….…………………………. 4.1.2.2. Health Beliefs……………………………...………………………………. 4.1.2.3. Risk Awareness……………………….…………………………………… 4.1.2.4. Social Well-Being……………………………………...………………..….. 4.1.2.5. Psychological Well-Being………………………………..………………... 4.1.3. Statistical Analyses…………………..……………………………………….………. 4.2. STUDY 2. POSSIBLE TARGETS OF INTERVENTIONS AFTER CORONARY EVENTS – LESSONS LEARNED FROM HUNGAROSTUDY 2002.……………………..…………………………... 4.2.1. Subjects ……………………………….………………………...……………………. 4.2.2. Questionnaires.………………………...……………………………………………. 4.2.2.1. Health Behaviour……...………...………………………………………… 4.2.2.2. Social Well-Being ……………………….………………………………… 4.2.2.3. Psychological Well-Being ………………………………….…………….. 4.2.3. Statistical Analyses…………………..………………………….………………….… 4.3. STUDY 3. PLATELET REACTIVITY IN DEPRESSED PATIENTS…..…….………………….…… 4.3.1. Subjects………………………………………….………………………………….. 4.3.2. Study Protocol…………………….………………………………………………….. 4.3.3. Statistical Analyses.…………………………………………………………………… 5. RESULTS 5.1. STUDY 1. HEALTH BEHAVIOUR, HEALTH BELIEFS, RISK AWARENESS AND PSYCHOSOCIAL WELL-BEING IN UNIVERSITY AND COLLEGE STUDENTS IN HUNGARY 5.1.1. Demographic Characteristics………………………………………………………… 5.1.2. Behavioural Factors…………………………………………………………………… 5.1.2.1. Prevalence of Health and Risk Behaviours…….……………...………….. 5.1.2.2. Intentions to Improve Health Behaviours….……………………………… 5.1.3. Beliefs in Lifestyle and Health……………………………...………………………… 5.1.3.1. Prevalence of Beliefs ………………………………….……...…………… 5.1.3.2. Prevalence of Beliefs by Risk or Health Behaviours……………………… 5.1.4. Awareness of the Influence of Lifestyle on Heart Disease……..…………………….

-1-

3 7 9 9 19 23 24 24 27 31 33 33 33

34 34 35 36 37 38 38 39 41 42 43 43 43 43 44 47 48 48 49 51

52 53 53 57 58 58 58 59

5.1.4.1. Prevalence of Risk Awareness…….……………………………………… 5.1.4.2. Prevalence of Risk Awareness by Risk or Health Behaviour …………..… 5.1.5. Psychological Well-Being of Students………………………………………………. 5.1.5.1. Risk Factors …………………………………………………………….… 5.1.5.2. Protective Factors………………………………………………………… 5.1.6. Social Well-Being of Students……………………………………………………….. 5.1.6.1. Social Support…………………………………………………………….. 5.1.6.2. Socioeconomic Status………….…………………….……………………. 5.1.7. Relationship between Health Beliefs or Risk Awareness and Behaviour ……………. 5.1.7.1. Gender differences in the relationship between health beliefs and behaviour…………………………………………………………………………… 5.1.7.2. Gender differences in the relationship between risk awarenesss and behaviour…………………………………………………………………………… 5.1.8. Relationship between Psychosocial Factors and Behaviour……………………….…. 5.1.9. The Final Model: Beliefs, Risk Awareness and Relevant Psychosocial Factors……... 5.1.10. International Comparison of Behaviour, Health Beliefs and Risk Awareness……… 5.1.11.Trends in Smoking, Physical Exercise, Diet, Health Beliefs and Risk Awareness in Hungarian University and College Students, 1990-2000……...…………………….. 5.2. STUDY 2. POSSIBLE TARGETS OF INTERVENTIONS AFTER CORONARY EVENTS – LESSONS LEARNED FROM HUNGAROSTUDY 2002 5.2.1. Clinical Characteristics…………………………..……………………………………. 5.2.2. Behavioural Risk Factors………..……………………………………………………. 5.2.3. Psychological Well-Being………………….…………………………………………. 5.2.3.1. Psychological Risk Factors…………..……………………………………. 5.2.3.2. Protective Psychological Factors……….………………………………… 5.2.4. Social Factors..……………………………………..……………………………….… 5.2.4.1. Social Support………….……………...……………………………………. 5.2.4.2. Socioeconomic Status ……..…………………………..………………….. 5.2.5. Patient Characteristics: Psychosocial Factors and Depressive Symptoms …………… 5.2.6. Patient Characteristics: Psychosocial Factors and Life Satisfaction …………………. 5.2.7. Patient Characteristics: Psychosocial Factors and Self-Rated Health ………..………. 5.3. STUDY 3. PLATELET REACTIVITY IN DEPRESSED PATIENTS 5.3.1. Clinical Characteristics…………………………………………………..……………. 5.3.2. Changes of Cardiovascular Parameters in Patients and in Controls…………..……… 5.3.3. Severity of Depression and Platelet Activation Markers……………………………... 5.3.4. Activation of Platelets during Mental and Physical Stress……………………………. 5.3.5. Effect of Treatment on Platelet Activation…………………………………………….

59 60 61 62 62 63 63 63 64

6. DISCUSSION………………………………………………….…………………………………. 6.1. STUDY 1. HEALTH BEHAVIOUR, HEALTH BELIEFS, RISK AWARENESS AND PSYCHOSOCIAL WELL-BEING IN UNIVERSITY AND COLLEGE STUDENTS IN HUNGARY……………………. 6.2. STUDY 2. POSSIBLE TARGETS OF INTERVENTIONS AFTER CORONARY EVENTS – LESSONS LEARNED FROM HUNGAROSTUDY 2002…………………………………………...………. 6.3. STUDY 3. PLATELET REACTIVITY IN DEPRESSED PATIENTS…..…………………………….. 6.4. LIMITATIONS OF THE STUDIES…………………………………………………….…………. 6.4.1. Limitations of the Epidemiological Studies…………………….……………………. 6.4.2. Limitations of the Laboratory Research………………………………………………

98

65 67 67 69 73 75

78 79 82 82 83 87 87 89 90 91 92 93 94 96 96 97

99 107 115 117 117 118

7. CONCLUSIONS, RECOMMENDATIONS………………….………………………………… 119 8. SUMMARY OF THESIS………………………………………………………………………… 126 9. REFERENCES………..…………………………………………………………………………. 128 10. PUBLICATIONS…………….……………………………………………………………….…. 141 11. ACKNOWLEDGEMENTS……………………………………………..………..…………….. 143 12. APPENDIX…………...…………………………………………………………………….…..... 148

-2-

1. INTRODUCTION Importance of the Topic: Key Points Cardiovascular disease is the leading cause of death worldwide. According to World Health Report 2003, 16.7 million people died due to cardiovascular disorder in 2002, which accounts for 29.2% of total global deaths. The most common cardiovascular disorder is coronary heart disease (CHD): out of the nearly 17 million deaths 7.2 million are due to ischaemic heart disease, being responsible for 13% of deaths worldwide (1). Furthermore, at least 20 million people survive heart attacks every year, a significant proportion of them requiring costly clinical care. Coronary heart disease affects people in their mid-life years, often leading to a worsening of the socioeconomic situation and of psychosocial well-being. It influences not only the affected individuals, but their families as well, and by putting a huge burden on care resources, it has even national consequences. Although recently there has been a drop in age specific CHD incidence as well as death rates in most parts of the developed world and its prognosis has improved markedly (2, 3), CHD still remains the leading cause of deaths in developed countries. In contrast to the decrease of CHD incidence and death rates seen in Western and Northern European countries, there is a morbidity and mortality crisis in Central and Eastern European societies, representing a challenge to behavioural sciences, public health, medicine and biology. CHD and cardiovascular death are by far the leading causes of death in these societies: in Hungary in 2001 the death rate of coronary heart disease was 294.1/ 100,000 population (4). Furthermore, an even higher prevalence of coronary heart disease is expected in future. It has been projected that cardiovascular disease will climb worldwide to the first most common cause of death, with more than 36 percent of all deaths in 2020 (5). The analysis of the social, psychological and biological factors that contribute to these huge number of patients with CHD are of special interest not only for identifying risk and prognostic variables but also in pointing out possible targets of subsequent health promotion and intervention programmes. Much has been learned about the importance of traditional risk factors such as smoking, lack of exercise, overweight/obesity, hypertension or hypercholesterolemia. However, the

-3-

practical application of the research findings, especially regarding behavioural risk factors (such as smoking, lack of physical activity and overweight) still lacks sufficient public and personal support. Moreover, it has been suggested that not all patients with coronary heart disease have one of the established coronary risk factors mentioned above – 10 to 35 percent lack any of them (6). These findings have led to a search for potential novel markers and other non-traditional risk factors in order to provide a better assessment of the cardiovascular risk (Table 1.1). Among the most important novel and non-conventional risk and prognostic indicators were psychosocial variables such as depression, hostility, lack of social support, work or marital stress, low socioeconomic status as well as biological markers, such as altered autonomic function (reduced heart rate variability), inflammation or platelet activation. Recently there is a growing body of literature pointing out that identifying psychosocial characteristics that could attenuate the negative effect of risk indicators, would be of utmost importance, as it is not easy to modify detrimental habits or behaviours. Well-being, selfefficacy, spirituality, purpose in life, self-rated health have been proposed among others as possible protective factors. However, the vast majority of studies exploring the relationship of psychosocial variables and CHD have been carried out in western countries. There is a long-lasting debate on how far can results of studies conducted in western countries be applied in other regions of the world. In order to combat the morbidity and mortality crisis in Central and Eastern Europe it is crucial to investigate which factors are important in our societies and subsequently target health promotion and intervention programmes on these particular areas. Based on these findings, the PhD research was built on two pillars: epidemiologic studies were carried out to assess psychosocial and behavioural factors associated with myocardial infarction (MI). We aimed at improving the efficacy of both primary and secondary prevention strategies, and to provide clinicians, other health care professionals and policy makers specific hints on the special characteristics of the targeted population, pointing out possible drawbacks and pitfalls of the interventions. On the other hand we conducted laboratory research, examining a potential mechanism between depression and heart disease, the role of platelet activation, which could become in the future a part of the pharmacological treatment.

-4-

The first study therefore evaluated health behaviour, health beliefs, risk awareness and psychosocial factors among university and college students in Hungary. The results were compared with European data in order to be able to examine similarities or differences. Furthermore, we yielded an estimation of changes in the examined measures between 1990 and 2000 in Hungary. The second study involved patients who suffered a heart attack, in order to identify and monitor those behavioural and psychosocial factors that could play a role in the recovery of Hungarian patients, as well as to detect possible drawbacks and pitfalls, that could hinder either patients or clinicians and health care professionals, to make best use of possibilities. The third study examined a postulated mechanism, platelet reactivity in depressed patients. Platelets, playing a central role in hemostasis, atherosclerosis, thrombosis and acute coronary syndromes, have been proposed to contribute to the higher risk of ischemic heart disease in patients with depressive disorders, as well as to the increased morbidity and diminished survival of depressed patients after a heart attack. Alterations of platelet reactivity related to depressive symptoms have been investigated.

-5-

TABLE 1. Risk factors of coronary heart disease (CHD) – modified after The atlas of heart disease and stroke, WHO (162).

MAJOR MODIFIABLE RISK FACTORS Tobacco use Increases risks of CHD, especially in people who started young, and heavy smokers. Passive smoking an additional risk. Physical inactivity Increases risk of heart disease by 50%. Obesity Major risk for CHD and diabetes. Unhealthy diets Low fruit and vegetable intake is estimated to cause about 31% of CHD worldwide; high saturated fat intake increases the risk CHD through its effect on blood lipids and thrombosis.

High blood pressure Major risk for heart attack and the most important risk factor for stroke. Diabetes mellitus The risk of heart attack is three times higher compared to those without DM. Abnormal blood lipids High total cholesterol, LDL-cholesterol and triglyceride levels, and low levels of HDLcholesterol increase risk

OTHER MODIFIABLE RISK FACTORS Poor mental health Depression is associated with a 2 to 4-fold increased risk of CHD. Psychosocial stress & personality Lack of social support, chronic life stress (e.g. work stress or marital stress), anxiety as well as hostility and type-D personality increase the risk of CHD. Low socioeconomic status (SES) Consistent inverse relationship with risk of heart disease.

Alcohol use One to two drinks per day may lead to a 30% reduction in heart disease, but heavy drinking damages the heart muscle. Use of certain medication Some oral contraceptives and hormone replacement therapy increase risk of CHD. Lipoprotein(a) Increases risk of heart attacks especially in presence of high LDL-cholesterol.

NON-MODIFIABLE RISK FACTORS Advancing age Most powerful independent risk factor cardiovascular disease; risk of stroke doubles every decade after age 55.

Heredity or family history for Increased risk if a first-degree blood relative has had CHD or stroke before the age of 55 years (for a male relative) or 65 years (for a female relative). Gender Ethnicity or race Higher rates of coronary heart disease among men Increased CHD deaths noted for South Asians and compared with women (premenopausal age). American Blacks in comparison with Whites.

NOVEL RISK FACTORS Abnormal blood coagulation - platelet activation Increased platelet acitvation, elevated blood levels of fibrinogen and other markers of blood clotting increase the risk of cardiovascular complications.

Inflammation Several inflammatory markers are associated with increased cardiovascular risk, e.g. elevated Creactive protein (CRP). Excess homocysteine in blood High levels may be associated with an increase in cardiovascular risk. Altered autonomic function Lower heart rate variability predicts poor prognosis in CHD.

-6-

2. OVERVIEW OF THE LITERATURE 2.1. Characteristics of Physical and Mental Well-Being Physical and mental health determine essentially the overall well-being of individuals, societies and countries. Although both of them play a vital role, curiously in most parts of the world, mental health does not get the same attention as physical health does. However, mental and behavioural disorders are estimated to account for 12% of the global burden of disease for the year 2000, and it is projected that by the year 2020 it will increase to 15% (7). Taking the disability component of burden alone, global burden of health 2000 estimates showed that mental and neurological conditions accounted for 30.8% of all years lived with disability, with depression causing the largest amount (accounting for almost 12%) of disability. In accordance with these findings, the WHO defined “health” in the WHO Constitution as “not merely the absence of disease or infirmity”, but rather, “a state of complete physical, mental and social well-being” (8). This definition has several consequences. First, health promotion must aim for improving all these areas. Second, these areas are all to be considered not only in primary prevention, but also in the rehabilitation of patients in order to be able to achieve the best possible “health” of patients. Furthermore, this concept implicates that besides risk factors protective characteristics are also relevant, pointing out the crucial importance of understanding and reinforcing human strengths and factors that allow individuals, communities, and societies not just to endure and survive but also to flourish (9). This concept presented by the WHO is at the same time a real challenge to medicine, behavioural sciences and biology. Engel argued for a need for a new medical approach already in 1977, proposing a biopsychosocial approach (10). Most illnesses, both mental and physical, seem to be influenced by a combination of biological, psychological and social factors (see Figure 2.1). In the medical setting, however, the emphasis is sadly still almost solely on the biological aspects (that is also seen in the practice of taking a history of the patient: psychological and social factors are most of the time not assessed). However, there is an emerging literature and also a widening of the focus of attention of clinicians from the one and only biological approach to a more holistic one, taking into

-7-

consideration psychological as well as social risk and protective factors besides biological markers.

Physical illness

Figure 2.1. Interaction between biological, psychological, social factors and mental as well as behavioural disorders and physical illness - modified from WHO Report 2001 (7).

Several studies have investigated the biological markers contributing to the establishment and worsening of coronary heart disease. In addition, in the last two decades a growing body of literature has focused on the psychological and social characteristics. The results of these studies indicate that psychosocial factors strongly influence the course of CHD. In the next chapter I will summarise the major findings in the literature concerning the relationship between psychosocial correlates and coronary heart disease.

-8-

2.2. Evidence of Connections between Cardiovascular Disease and Psychosocial Factors There are several psychosocial measures that have been connected with CHD. Besides identifying risk factors for the development of coronary heart disease and the worsening of prognosis after a myocardial infarction, I would like to emphasise the importance of identifying those psychosocial characteristics, which could be protective and could attenuate the negative effect of the risk indicators. I will first present the evidence for psychological and social risk factors, followed by an overview of suggested protective indicators.

2.2.1. Psychological Factors 2.2.1.1. Psychological Risk Factors Depressive Symptoms According to world health statistics, in 2000 depression was the fourth most common disease worldwide, and had the leading position in the ranking of the years lived with disability (8). In the age group of 15-44 depression was responsible for the second highest burden. It was estimated that by the year 2020, if current trends for demographic and epidemiological transition would continue, the burden of depression will increase even more and will be worldwide the second only to ischaemic heart disease for DALYs lost for both sexes. Disability-adjusted life years (DALYs) lost can be thought of as “healthy years of life lost”. They indicate the total burden of a disease, as opposed to simply the resulting deaths. In the developed regions, depression will then be the highest ranking cause of burden of disease (5, 7). The point prevalence of unipolar depressive episodes was 1.9% for men and 3.2% for women in 2000, and it was estimated that 5.8% of men and 9.5% of women would experience a depressive episode in a 12-month period (8). Depression is characterised by loss of confidence and self-esteem, inappropriate guilt, sadness, loss of interest in activities, decreased energy, diminished concentration, disturbance of sleep and appetite as well as thoughts of death and suicide. A variety of somatic symptoms may also be present. Major depression is diagnosed when depressive symptoms reach a threshold and last for at least 2 weeks. In the clinical practice two general classifications are in use, the ICD-10, Tenth Revision of the International Classification of

-9-

Diseases (11) and the DSM-IV, Diagnostic and Statistical Manual of Mental Disorders, developed by the American Psychiatric Association (12). Nevertheless, in the cardiovascular research a heterogeneity of the definitions of depression can be observed: some researchers have measured depressive symptoms, while others have investigated depressive syndromes. Besides, a variety of measures have been used to assess depressive symptoms: 1 or 2 questions (e.g. “Have you felt sad in the past 2 weeks?”), clinical interviews (e.g. Diagnostic Interview Schedule), external-rated instruments (e.g. Hamilton Depression Rating Scale), and self-report questionnaires, such as Beck Depression Inventory and Zung Self-rating Depression Scale. Although there were no standard approaches to definition and measurement it is quite remarkable, that even with such a variety of measures, several etiologic and prognostic studies have shown significant correlations between depression and coronary heart disease outcomes. This affirms the robustness of the underlying concept of depression. Depression can vary in severity from mild to very severe. Although subjects with more severe depressive symptoms show a higher risk of adverse cardiovascular events, no consistent level could be identified at which depressive symptoms begin to exert cardiotoxic effects. Even mild depressive symptoms have proved to influence the course of coronary heart disease (13). This finding emphasises the importance of considering psychosocial factors, even if they appear in a very mild form. As depression and heart disease share a number of symptoms, such as tiredness, loss of energy, inability to perform different activities, it is an important point to discriminate whether depression scores predict outcome in CHD only because they reflect symptoms of an underlying heart disease or it is an independent risk factor representing a not negligible effect of the affective, emotional aspects of depression. There is evidence that scores computed only from the negative affect items of the depression scales predict survival independent of disease severity and somatic symptoms (14), suggesting therefore that depression is an independent factor.

- 10 -

A. Depression as a Primary Risk Factor for Coronary Heart Disease Several etiologic studies have revealed that patients with depressive symptoms are at increased risk of developing and dying of coronary heart disease (15-20) even after controlling for smoking and other known risk factors for cardiovascular disease (gender, weight, physical activity, high blood pressure, cholesterol). The same relationship has been described for cerebrovascular disease (21, 22). Some investigators reported of a relationship for hypertension, too (23, 24). An overview of the literature is provided in Table 2.1a and Table 2.1b. Only prospective cohort studies on initially healthy subjects were included with depression as the predictor. In these studies depression was defined as either unipolar clinical depression assessed by clinical procedures or depressive mood measured by standardised, well-known psychometric scales. As already mentioned, several studies have shown that the two concepts are closely related and that high scores on depressive mood scales strongly correlate with the presence of unipolar depression (25, 26). The following diagnoses were used as outcome measures: fatal or non-fatal myocardial infarction, coronary death or cardiac death. Relative risks are an easy and straightforward way to compare results of different studies. Most of the studies presented in the tables below provide relative risks (RRs) (27). In the recent years two meta-analyses have been published, and although they differed slightly in their inclusion criteria (resulting in having 7 common studies, 3 and 4 different ones, respectively) (28, 29), the calculated relative risks were very similar in both papers. The relative risk was found to be 1.64 (95%CI: 1.14-1.90, p < .001) in the first metaanalysis. Rugulies reported of a relative risk of 1.64 as well, although with slightly different confidential intervals (95%CI: 1.29-2.08, p < .001). He also carried out a sensitivity analysis, showing that clinical depression was a stronger predictor than depressive mood (RR: 2.69; 95%CI: 1.63-4.43, p < .001 and RR: 1.49; 95%CI: 1.16-1.92, p = .02, respectively).

- 11 -

TABLE 2.1a. Overview of studies examining depression as a primary risk factor for coronary heart disease Study

Year Sample Depression Comparison Groups Outcome Controlled for RR (95%CI) of Women/Men (%) Measure Measures publ. Age group (years) Follow-up (years) Anda (30) 1993 2832 US adults General Well- Depressed ( >12 on CHD death Age, gender, race, marital status, 1.5 (1.1-2.4) (52/48) Being Scale depressed subscale) education, smoking, BMI, 45-77 vs non-depressed exercise, cholesterol, blood 17-y f/u pressure, alcohol Aromaa (31) 1994 5355 Finnish adults Present State Depresssed CHD death Age, sex, education, “cardiac risk Men: 52/48 Exam vs non-depressed factors”, chronic physical illness 3.45 (1.76-6.76) 40-64 Women: 6.6 years 2.59 (1.12-5.99) Pratt (32) 1996 1551 US adults DIS Depresssed Non-fatal MI Age, gender, marital status, Major depression: 62/38 vs non-depressed education, diabetes, smoking, 4.54 (1.65-12.45) 91% < 61 blood pressure, Dysphoria: 13-y f/u 2.07 (1.62-2.37) MMPI Depresssed Fatal and non-fatal Age, gender, smoking, systolic Continuous analysis: Barefoot (33) 1996 730 Danish adults vs non-depressed MI blood pressure, triglycerides, 1.7 (1.23-2.34) 44/56 Dichotomous analysis: 50 or 60 (birth: sedentary work and sedentary 1.91 (1.23-2.97) 1914) leisure 17 & 27-y f/u Depresssed Non-fatal AMI - Age, BMI, smoking, blood All endpoints combined: Sesso (34) 1998 1305 US men MMPI-2 D MMPI-2 DEP (highest tertile) CHD death and pressure, cholesterol. alcohol, MMPI-2 DEP: 0/100 SCL-90 vs non-depressed angina family history of CHD 2.07 (1,13-3,81) 40-90 SCL-90: ns. 7-y f/u Depressed Non-fatal and fatal Graduation age, baseline Clinical depression: Self-report of Ford (35) 1998 1190 US men vs non-depressed AMI cholesterol levels , family history 2.12 (1.11-4.06) an episode of 0/100 clinical of AMI, smoking, incident HT US students depression and DM, exercise 40-y f/u RR = relative risk; CI = confidence interval; f/u = follow-up period; US = american; CHD = coronary heart disease; AMI = acute myocardial infarction; HT= hypertension; DM= diabetes mellitus; BMI = Body Mass Index DIS: modified version of the National Institute of Mental Health Diagnostic Interview Schedule; MMPI: Minnesota Multiphasic Personality Inventory; MMPI-2 D: depression subscale; MMPI-2 DEP: depressive thought without somatic symptoms; SCL-90: 90-item Symptom Check List

- 12 -

TABLE 2.1b. Overview of studies examining depression as a primary risk factor for coronary heart disease (continued) Study

Year Sample of Women/Men (%) publ. Age group (years) Follow-up (years) Penninx (36) 1998 3701 US adults 66.2/33.8 70-103 4-y f/u Mendes de 1998 2391 US adults Leon (37) 61/39 65-99 9-y f/u Ferketich (38) 2000 7894 US adults 63/37 not reported -only mean age: w 53.7, m 55.9 10-y f/u Penninx (39) 2001 2847 US adults 450 CHD patients, 2397 healthy 55-85 4-y f/u

Depression Comparison Groups Outcome Measure Measures

Controlled for

CES-D

Depressed (CES-D>19) vs non-depressed

Non-fatal and fatal AMI

Age, sex, smoking, alcohol, BMI, HT, Only in men sign., in case of history of DM/stroke/cancer, physical new depression: disability 2.03 (1.28 –3.24)

CES-D

Continuous- 5 groups

CHD

Age, education, smoking, diabetes, blood pressure

CES-D

Depressed (CES-D>16) vs non-depressed

Non-fatal Age, race, smoking, HT, DM, BMI, CHD non-fatal CHD during follow-up, events and poverty CHD death

DSM-III CES-D

RR (95%CI)

Women: 1.02 (1.00-1.04) Men: 0.99 (0.96-1.02) Women: only non-fatal sign. 1.73 (1.11-2.68) Men: non-fatal 1.71 (1.14-2.56) CHD death: 2.34 (1.54-3.56)

Major depression & CHD death Demographics, smoking, alcohol use, Cardiac patients minor depression blood pressure, BMI and comorbidity mD: 1.6 (1.0-2.7) (CES>16) MD: 3.0 (1.1-7.8) vs non-depressed Without cardiac disease in CHD patients and mD: 1.5 (0.9-2.6) MD: 3.9 (1.4-10.9) in non-CHD patients 1 SD elevation in Rowan (40) 2005 1302 US adults CES-D Continuous CHD age, sex, BMI, physical activity , 45 or older family history of premature CHD, CES-D score: 4 -y f/u diastolic blood pressure, lipids, 1.32 (1.01-1.71) smoking, alcohol use, diabetes, education level RR = relative risk; CI = confidence interval; publ = publication; f/u = follow-up period; w = women; m = men; US = american CHD = coronary heart disease; AMI = acute myocardial infarction; DM = diabetes mellitus, HT = hypertension; BMI = Body Mass Index; mD = minor depression; MD = major depression CES-D: Center for Epidemiological Studies-Depression Scale

- 13 -

B. Depression as a Prognostic Variable in Patients with Established Coronary Heart Disease/ Post-Acute Myocardial Infarction (AMI) Several prognostic studies showed that the presence of depressive symptoms has a negative impact on the prognosis of patients with established coronary artery disease even after the usual medical and social risk factors had been taken into account. Since this thesis focuses on the bio-psycho-social factors of myocardial infarction, I present only briefly the literature regarding studies examining the impact of cardiovascular events other than AMI and I will focus on studies that included post-infarction patients. Authors of the former group studied patients undergoing percutaneous transluminal coronary angioplasty (PTCA) (41, 42), others patients after a coronary artery bypass grafting procedure (CABG) (43, 44) or included patients after PTCA, CABG and AMI (45). The effects of initial depressive symptoms on the prognosis and mortality after CABG were also assessed (46, 47) and were found significant: the point estimates of the odds ratios in these studies ranged from 1.26 to 2.69. For a detailed review on these studies see Barth et al. (48). The studies that investigated the effects of depressive symptoms or depressive disorder on the prognosis of patients who suffered from myocardial infarction are presented in detail in Tables 2.2 to 2.4. Studies providing data concerning the following outcome measures are listed below: all-cause mortality, cardiac mortality and cardiovascular events (e.g. myocardial infarction, unstable angina, need for revascularisation, arrhytmia). Both the bivariate and the multivariate odds ratios are demonstrated if information of them was accessible. Only articles assessing depression by standard self-reported questionnaires or by standardised psychiatric interviews for the assessment of depressive disorder were included. Self-reported questionnaires had to be validated for depressive symptomatology, such as BDI = Beck Depression Inventory, HADS = Hospital Anxiety and Depression Scale, Zung = Zung Self-rating Depression Scale, MADRS = Montgomery Asberg Depression Rating Scale, KSb = Klinische Selbstbeurteilungsskalen aus dem Münchner psychiatrischen Informationssystem. Most commonly implemented standardised psychiatric interviews were the following: SCID = Structured Clinical Interview for DSM-IIIR or DSM-IV, DISH = Depression Interview and Structured Hamilton; DIS = modified version of the National Institute of Mental Health Diagnostic Interview Schedule).

- 14 -

All-cause Mortality Table 2.2 presents the effect of depressive symptoms on all-cause mortality. Some studies adjusted for other risk factors such as hypertension, diabetes mellitus, smoking, left ventricular ejection fraction, prior MI or CABG (49-51). In a meta-analysis including nine studies with a total of 3082 patients the pooled odds ratio of all-cause mortality after MI in 952 depressed patients was 2.38 (CI: 1.76-3.22; p < .001) when compared with 2130 nondepressed patients (52). Cardiac Mortality Nine publications reporting of 6 cohorts with a total of 3343 patients were found, in 2 cases (53-55, 56-57) with varying follow-up periods from the same group of patients (see Table 2.3). The pooled odds ratio of cardiac mortality of 1091 depressed patients compared to the 2252 non-depressed patients was 2.59 (CI 1.77-3.77; p < .001) (52). Cardiovascular Events Since the definition of cardiovascular events varied among studies, the exact end-points for each study can be found in Table 2.4. Nevertheless, all authors reported on recurrent MI, cardiac arrest and cardiac death. Nine studies have included a total of 3401 patients. The final pooled odds ratio was 1.95 (CI:1.33-2.85; p < .001) (52). Since depression can present itself in a large variety of symptoms, both physical and affective, that are not necessarily seen together, several psychological explanatory models have been described. Several investigators proposed that some aspects of depressive symptoms might be more important than others for patients with cardiovascular disease. Hostility and anger, characteristic for type A personality have been considered to have an influence on the cardiovascular system (72, 73). Denollet reported that patients with type D behaviour, representing both high social inhibition and negative affectivity show an increased mortality rate compared to those without these personality traits (74, 75). Others suggested feelings of fatigue and demoralisation, the state of vital exhaustion to play an important role (76, 77, 78). Feeling of hopelessness and loss of control were also shown to predict (79, 80, 81) the genesis and worsening of coronary artery disease.

- 15 -

TABLE 2.2. Effect of depressive symptoms/disorders on all-cause mortality in prognostic studies with patients after MI Study

Year of Selection of publ. MI pts (Yes/No) Silverstone (58) 1987 N Lesperance (59) 1996 N Irvine (60) 1999 Y† Kaufmann (49) 1999 N

Mean Age (years) 63 60 64 65

Women (%)

Depression Measure

25 22 18 34

MADRS DIS BDI DIS

Welin (61) Mayou (62) Bush (50)

2000 2000 2001

Y** N N

NA 63 65

16 27 42

Zung HADS BDI SCID

Lauzon (63) Strik (64) Carney (51)

2003 2003 2003

N Y‡ N

60 59 59

21 24 40

BDI SCID DISH

Time Post MI (days) 1 5-15 6-45 7

Depr. (%)

F/u (mo)

Depression defined by

Adjusted for

44 16 33 27

0.25 6/12/1 24 6/12

Symptoms Disorder Symptoms Disorder

30 25: severe depressive symptoms (194). Hostility was examined by the shortened Cook and Medley Hostility subscale of the MMPI (195-198). This scale primarily measures suspiciousness, resentment, and cynical mistrust (199). We included 4 items in the questionnaire: 1.

People are honest because they fear from the exposure

2.

The best is if you distrust anybody

3.

If I have heard the success of a friend of mine, I feel I am frustrated

4.

People are generally dishonest and selfish and they want only to take advantage of others

Response categories ranged from 0 = strongly disagree to 4 = strongly agree. The scores were summed, and high hostility was defined as scoring 7 points and above. Statement number 4 was used as a measure of social distrust separately (200). The variable was dichotomised to compare the strongly agree and slightly agree group with the neutral, slightly and strongly disagree groups. Psychological Protective Factors Life Satisfaction The following question was applied: “All things considered, how satisfied are you with your life as a whole? The options provided were: 5 = very satisfied; 4 = moderately satisfied; 3 = no feelings either way; 2 = moderately dissatisfied; 1 = very dissatisfied. For descriptive statistics being satisfied with life was defined as moderately and very satisfied.

- 39 -

Purpose in Life The shortened Hungarian version of the Purpose in Life Questionnaire was administered (201, 188). The following four questions out of the twenty were selected: 1.

I am reliable

2.

I have no goals in life

3.

Every single day is new and different

4.

I am generally bored

Answer categories ranged from 0 = not at all characteristic to 3 = totally characteristic. The scores were summed for descriptive analyses, respondents with a sum higher than 7 were considered to have high purpose in life. Self-Efficacy The General Self-Efficacy Scale was originally developed in Germany. It can be used in both adult and adolescent populations, and very similar characteristics have been found in many cultures (202, 203). The tenth item of the scale, namely “I can usually handle whatever comes my way“ was used to measure general self-efficacy. This item showed the highest item-residual in the Hungarian population, 0.73 (204). Self-efficacy was coded as 0 - not at all/scarcely and 1 - characteristic/totally characteristic for descriptive statistics. Religiosity was assessed by 4 questions: 1. Do you consider yourself a believer? yes/no 2. If yes, how important is your faith to you: 1 = if problems arise; 2 = a bit; 3 = not too much, but still important; 4 = very important 3. How important is your faith in your daily life? 0 = not at all; 1 = sometimes; 2 = quite; 3 = very important 4. Which role does your faith play in setting your goals for your life? 0 = no role; 1 = sometimes it plays a role; 2 = it defines everything

Respondents considering themselves believers and attaching importance to their faith in general were classified as participants to whom spirituality is of value. Coping strategies were assessed using the Ways of Coping Questionnaire developed by Folkman and Lazarus (115). Both emotion-focused and problem-focused coping were investigated. The sum of scores was computed, and for descriptive statistics the final scores were dichotomised as low or high emotional and problem-focused coping (cut-off points were 7 and 13, respectively). Self-Rated Health According to the recommendations of Eriksson et al. the following question was included: “In general, would you say that your health is: excellent; very good;

- 40 -

good; fair or very poor”? (205). In the descriptive analyses poor self-rated health was defined as fair or very poor.

4.1.3. Statistical Analyses Data were analysed using SPSS 9.0 for Windows (Statistical Package for Social Sciences, 207). Descriptive statistics of the prevalence of health behaviours, intentions, beliefs and risk awareness were carried out. Chi-square and (in case of multilevel variables) MannWhitney-U-tests were used to test whether variables of interest differed across gender. Age and gender adjusted odds ratios of beliefs and risk awareness with health behaviours were calculated via binary logistic regression models. The prevalences of psychosocial risk and protective factors and gender-dependent differences were established in the same manner, using crosstabs and chi-square tests. Selection of behaviour-relevant psychosocial factors was carried out in a forward conditional manner (selection cut-off of p ≤ .05 was used) after adjusting for age and gender, taking into consideration possible gender related interactions in sex-specific models. Psychosocial measures were considered as continuous variables in the models. In the next step the selected variables (age, gender, beliefs, risk awareness and selected psychosocial factors) were examined in univariable regression analyses. Finally, multivariable logistic regression models were constructed to investigate the major correlates of the specific health behaviour pattern. In these analyses age, gender, health beliefs and risk awareness (only if significant in univariate model) were first entered (Model 1), followed by selected psychosocial factors with at least statistical tendency in univariable analyses (p ≤ .1) (Model 2), and significant interactions with gender (Model 3). The psychosocial measures were included as continuous variables in the logistic regression models. Following the recommendations of Baron, we explored the possible mediating effect of gender differences in beliefs on gender differences in health behaviour applying logistic regression models (208). For international comparisons and the Hungarian 10-year comparison we adjusted for age using general linear model. The prevalence of behaviours and risk awareness are presented as percentages with 95% confidence intervals (CI). Beliefs in health benefits and risk awareness adjusted for age are presented as means with 95% confidence intervals. A pvalue of less than .05 was considered indicating statistical significance.

- 41 -

4.2. Study 2. Possible Targets of Interventions after Coronary Events – Lessons Learned From Hungarostudy 2002 Type of Study: cross-sectional epidemiologic study Data were obtained from the Hungarostudy 2002, a national, cross-sectional survey that is representative of the Hungarian population older than 18 years according to sex, age, and the 150 subregions in the country. The health survey was conducted by the Institute of Behavioural Sciences, Semmelweis University Budapest in 2002. Sample A clustered, stratified sampling procedure was developed by experts at the National Population Register. All communities with a population of more than 10,000 inhabitants were included in the sample as well as a randomly selected sample of smaller villages. In a second step, single persons were selected from this database in an age and gender distribution that was comparable to that existing in the specific county or settlement size from which they were drawn. The final sample therefore reflected the gender, age, and settlement size characteristics of each given county. In order to allow for replacements of individuals who refused to participate, two random samples of 13,000 persons were generated. Persons living in chronic psychiatric institutions and those living in elderly homes were excluded from the study-population, because of the household-based nature of the sample. In 2002, 12,643 persons were interviewed in their homes (209, 210) representing 0.16% of the population older than age 18. The refusal rate was 17.7%. There were differences in refusals based on residence (refusals tended to be higher in large cities than in small villages) and on gender (refusals were 4% higher in men than in women). For each refusal, another person was selected of the same age and sex from the same community. This replacement procedure did not result in any significant selection bias. The final sample corresponded well to the population description of the Central Statistical Office. When comparing the distributions of selected important variables in the final dataset and in the original, the sampling error in each case was within statistically acceptable limits with the highest estimated stratification error being 2.2% in men aged 18-39. This error is within the limits of the permitted statistical deviation (210). The interviewers in this study were district nurses, and they spent approximately 1 hour with each respondent administering the questionnaires. Because of their education in health, district nurses were selected as the most competent persons for interviewing participants. These nurses were intensively trained for their duties.

- 42 -

4.2.1. Subjects Respondents above 45 years reporting of a history of acute myocardial infarction were selected as patients after an AMI. As above 90 there were no patients, we decided to limit the inclusion criteria to the age-group of 45 to 90 years. Participants above the age of 45 with no history of a heart attack, cerebrovascular or peripheral artery or venous disease were taken as controls.

4.2.2. Questionnaires The questionnaire applied in the survey consisted of 701 questions assessing sociodemographic variables and current physical and mental health status. Traditional (gender, age, smoking, hypertension, weight, lack of physical activity), psychological and social risk as well as protective factors were included.

4.2.2.1. Health Behaviour Physical Activity Physical exercise was measured with a question: “How often do you exercise?”. The options provided were: never; irregularly; less than once a month; once a month; once a week; several times a week; daily. Being physically active was defined as several times a week or daily according to the guidelines of the WHO (162). Overweight - Obesity Respondents were asked to report their weight and height, and we calculated the BMI values. Since the ideal BMI changes with age, according to the recommendations of the WHO we regarded a BMI value of 30 and above as adipositas (165). This is the value where there is a need for therapy even if no other illnesses, such as high blood pressure, diabetes, elevated cholesterol-levels or arthrosis are present. Substance Use – Smoking Regarding smoking the following question was applied: “Do you currently smoke?” The options were: no, I never did; no, but in the past I did; yes). We considered only those who were currently smoking as smokers. To explore the habits of smokers, we also asked for the number of cigarettes per day, for how many years they had been smoking, and also at which age they started smoking.

4.2.2.2. Social Well-Being Social Support The same questionnaire was applied as in the first survey (see under section 4.1.2.4.). Subjects reporting of strong support from at least 1 category of the relationships

- 43 -

(parents/children, spouse/partner, friends or relatives), were regarded as participants with high social support. Marital Status Categories were defined as single, married (living together with spouse), married (living apart from spouse), divorced, widowed, and living with partner. Socioeconomic Status Educational Level Categories of education were the following: less than 8 elementary school, finished 8 elementary, skilled labor training, high school diploma, and college/university degree. Low educational level was considered and dichotomised in the analyses if the individual had skilled labor training or less. Monthly Income The monthly income of the household was assessed. The options provided were: less than 50,000 HUF per month, between 51,000 and 100,000 HUF/month, between 100,001 and 150,000 HUF/month, between 151,000 and 200,000 HUF/month and above 201,000 HUF/month. Subjective Socioeconomic Status The question ”On a scale ranging from 0 to 10, how would you describe your financial status?” was used as a subjective rating of economic situation. Participants ranking their financial state as less than average were considered reporting of low self-perceived socioeconomic status. Employment Status. The variable was defined as employee, entrepreneur, occasional worker, unemployed, retired or on disability pension, homemakers or others.

4.2.2.3. Psychological Well-Being Psychological Risk Factors Depressive Symptoms The Hungarian version of the Shortened Beck Depression Inventory (BDI) was applied to assess depressive symptoms (211, 212). It was confirmed to be valid and reliable in a Hungarian population sample (213). The shortened version contains the 9 items with the highest Cronbach’s α from the 21-item version of the BDI, having an overall Cronbach’s α of 0.85 (210). The questionnaire contained the following 9 items: 1. I have lost all of my interest in other people 2. I cannot make decisions at all any more 3. I wake up several hours earlier than I used to and cannot get back to sleep 4. I am too tired to do anything 5. I am so worried about my physical problems that I cannot think about anything else 6. I cannot do any work at all 7. I feel that the future is hopeless and things cannot improve

- 44 -

8. I am dissatisfied or bored with everything 9. I feel guilty all of the time

As the sum of the shortened version can be reliably transformed into the full score (188), for statistical analyses the scores of the shortened Beck Depression Inventory were transformed into the original score by multiplying the total score with 2.3. The scores were grouped in the following categories: normal: 0-9 points; mild: 10-18 points; moderate: 19-25 points; severe: 26 points or higher. Measures were dichotomised: participants with moderate to severe depressive symptomatology were compared with the rest of the sample. Hostility was measured by the shortened Cook and Medley Hostility subscale of the MMPI (195, 196). This scale primarily measures suspiciousness, resentment and cynical mistrust (199). The following questions were administered, with answers ranging from 0 = not at all to 3 = totally characteristic: 1. People are honest because they fear from the exposure 2. My every relative is well meaning with me 3. Nobody takes care of the others 4. The best is if you distrust anybody 5. If I have heard the success of a friend of mine, I feel I am frustrated 6. People are generally dishonest and selfish and they want only to take advantage of others

The scores were summed, and high hostility was defined as scoring above 7 points. Statement number 6 was used as a measure of social distrust separately (199). Psychological Protective Factors Sense of Well-Being The sense of well-being was measured by the WHO Well-Being 5 index (214). This questionnaire is a brief, self-rating instrument. The reliability of the Hungarian version of the well-being index was high, Cronbach’s α was 0.84 (210). The scores ranged from 0 = lowest score, 15 = maximal score. The questions concerning the past two weeks were the following: 1. I have felt cheerful and in good spirits 2. I have felt calm and relaxed 3. I have felt active and vigorous 4. I woke up feeling fresh and rested 5. My daily life has been filled with things that interest me

The measure was dichotomised by grouping respondents with under 8 points (defined as having low sense of well-being) and 8 or more points (having high sense of well-being).

- 45 -

Self-Efficacy 4 items from the General Self Efficacy Scale were selected: 1. It is easy for me to stick to my aims and accomplish my goals 2. I am confident that I could deal efficiently with unexpected events 3. I can solve most problems if I invest the necessary effort 4. I can usually handle whatever comes my way

Cronbach’s α for this four-item scale in this study was 0.83 (210); the lowest score was 0 and the maximal score 12. Participants with a sum of 7 or more points were classified as highly self-efficacious. Life Meaning Life meaning was assessed by the Brief Stress and Coping Inventory (216). Cronbach’s α for this scale was 0.69 in this study, indicative of acceptable internal consistency (208). Answers: 0 = no, 1 = sometimes, 2 = frequently 1.

I feel my life is part of a larger plan

2.

Many things in life give me great joy

3.

I am able to forgive myself and others

4.

I doubt that my life makes a difference (reverse coded)

5.

My values and beliefs guide me daily

6.

I feel in tune with people around me

7.

I am at peace with my place in life

The scores were summed, and the measure was included in the analyses as high life meaning (8 or more points) and low life meaning (under 8 points).

Spirituality Involvement in religion was measured by the question: “Are you religious? If yes, what is the form of your worship?” Responses were scored 0 for “I am not religious,” 1 for “I don’t worship,” 2 for “I worship in my own way,” 3 for “I worship rarely in my church,” and 4 for “I worship regularly in my church.” The importance of religion was also estimated by asking: “How important is your faith to you?”. Responses were scored: 0 for “Not at all,” 1 for “Slightly,” 2 for “Very important,” and 3 for “It influences my every action.” The measure was then dichotomised, comparing “not at all” and “slightly” with “very important” and “it influences my every action”. Coping Strategies Coping strategies were assessed using the Ways of Coping Questionnaire developed by Folkman and Lazarus (115). Both emotion-focused and problem-focused coping were

- 46 -

investigated. The sum of scores was computed, and for descriptive statistics the final scores were dichotomised as low or high emotion and problem-focused coping (cut-off points were 10 and 13, respectively). Self-Rated Health (SRH) SRH was measured with the question “How do you rate your health in general?” We provided five options: very good, good, fair, poor, and very poor. Participants with poor and very poor health were considered to report low self-perceived health.

4.2.3. Statistical Analyses SPSS 9.0 (Statistical Package for Social Sciences) program for statistical data analyses was used for the evaluation of the questionnaires (207). The general characteristics of patients and controls were examined via crosstabulation. Since age and gender differences were observed age and gender adjusted values were also calculated with general linear models. Statistical tests within gender (patients vs controls) and between gender groups were calculated using the adjusted values. The mental health of postinfarction patients was explored via multivariable linear regression models. Demographic and psychosocial measures were entered as independent variables. The assessed dependent variables were the following: depression, self-rated health and well-being. Following the recommendations of Babyak, we selected covariates according to our hypotheses a priori (217). Psychosocial measures were included as continuous variables. After controlling for age and gender, level of education, annual income and level of social support were added to the model (first model), followed by putative protective psychological factors (second model). Putative psychological risk factors were then finally included in the model (third model). A p-value under .05 was considered statistically significant.

- 47 -

4. 3. Study 3. Platelet Reactivity in Depressed Patients Type of Study: basic research (in humans)

4.3.1. Subjects Depressed Patients We included ten patients who had been admitted to the Clinic of Psychiatry and Psychotherapy at the Central Institute of Mental Health, Mannheim, University of Heidelberg, Germany. Inclusion criteria: 1) an episode of major depression according to DSM-IV criteria (American Psychiatric Association, 1994) 2) a score of at least 18 on the 21-item Hamilton Depression Rating Scale (HAM-D, 218) 3) no history of psychosis 4) no substance dependence 5) no bleeding disorders 6) no other medical conditions known to influence platelet function (219) such as chronic renal failure, defined as a serum creatinin concentration greater than 2 mg/dl, and known diabetes mellitus.

7) no ingestion of drugs known to influence platelet function (when taking antiplatelet drugs such as aspirin, clopidogrel, dipyridamol subjects were excluded from the study)

Patients were studied after a five-day period without anti-depressant medication (HAM-D: 19.4 ± 4.1) to provide at least a 5-day wash-out period. We decided to carry out the measurements after a five-day interval without antidepressant medication after taking into consideration, that the subjects were severely depressed patients, who might have easily rushed into a life-threating condition unless receiving the proper therapy. Controls For each patient, a German age- and sex-matched healthy comparison subject was selected. Inclusion criterias: ¾

absence of either a past or present personal history of major psychiatric disorders

¾

subjects fulfilled all the other conditions mentioned above

Coexisting medical conditions were found in five patients (hypercholesterolemia in three, as well as hypertension, peripherial arterial disease and goiter in one patient each) and in six comparison

subjects

(two

patients

suffered

from

migraine,

and

one

from

hyercholesterolemia, scleroderma, tinnitus, and both goiter and hypotension, respectively). There were five post-menopausal women in each study group. Regular medication in the

- 48 -

patients’ group included statins (2), thyroxin (1) and estrogen-replacement therapy (1), in the healthy control group five women were on estrogen-replacement therapy. All subjects had given written informed consent before being included in the study. The study protocol had been approved by the ethics committee of the University of Heidelberg.

4.3.2. Study Protocol Subjects were asked to refrain from major physical activity two days prior to the study, and not to smoke or eat after 10:00 pm the evening before the stress tests. On the study day, water was allowed in small amounts, but coffee, tea or alcoholic beverages were strictly forbidden, and all medication was withheld until completion of testing. Participants were asked to rest in their beds until they were picked up from the wards. At 8:00 am the protocol was explained again and subjects had the chance to decline if they wished to. Taking the elevators patients were then brought to the 1st floor of the building, walking only maximum 50m in a very slow speed. The study was conducted in a quiet, partially darkened room with a temperature setting of 20-25°C. After placing the electrodes for cardiovascular measures, subjects rested in a recumbent position for 20 minutes to reach an equilibrium for baseline measurements. To aid relaxation, subjects were listening to soft music and seaside sounds provided over headphones. Both mental and physical stressors were administered on the same day in a fixed order, as described below, with a time interval of 15 minutes between exposure. Measurements Done after Each Study Segment (Baseline, Mental Stressor, Physical Activity) ¾

Blood was drawn from an antecubital vein by a separate smooth venipuncture, using a 20-gauge needle and minimal stasis (maximal cuff pressure, 40 mmHg). As there is a known rapid recovery of most platelet activation markers after physical activity (220), no additional blood was drawn after patients had rested following the treadmill test.

¾

Blood pressure was measured with a calibrated aneroid sphygmomanometer and mean arterial blood pressure was calculated as the diastolic blood pressure plus one third of the difference between the systolic blood pressure and the diastolic blood pressure.

¾

Impedance cardiography (Cardioscreen, Medis, Germany) was used to monitor continuously both cardiac index and pre-ejection period during the relaxation period and the mental stress test (221).

¾

Platelet counts were determined by a Coulter AcTdiff (Beckman Coulter Inc., Brea, USA).

A scheme of the study protocol is provided in Table 4.1.

- 49 -

TABLE 4.1. Scheme of study protocol

Cardiovascular measures Heart rate Arterial blood pressure Cardiac index Pre-ejection period Platelet measures P-selectin GP53 Platelet count

Relaxation period Minute 1 19

Mental stress Minute 1 19

Physical activity Time point start peak

x x x x

x x x x

x x

x x x x x x x

x x x x x x x

x x

x x x

Relaxation period and mental stress lasted 20 minutes each. The duration of the physical activity varied.

Mental Stress Test and Physical Activity As mental stressor a 20-minute computer-based detection task (222), was performed while subjects were sitting. Subjects had to detect two figures, either appearing alone or in combination, or note their absence, in a number of consecutive matrices. In order to create a feeling of pressure, frustration and failing, the speed of the task was regulated so that subjects were only allowed to succeed in approximately 80% of trials at the beginning and 60% at the end of the test. Physical activity was standardised by administering a recumbent treadmill test using a modified Bruce protocol (223) with a load increase of 25 Watt every two minutes. Subjects were asked to cycle until exhaustion. Flow Cytometry and Platelet Counts I have used flow cytometry by modifying the assays described by Tschoepe (224): ¾

blood was sampled directly into a platelet-stabilising medium (phosphate-buffered saline [PBS] containing 0.17 M sodium citrate and 5.55 M acetylsalicylic acid, one part reagent mixed with nine parts blood)

¾

it was immediately fixed by adding an equal volume of formaldehyde (1%, wt/vol)

¾

after a fixation period of 15 minutes, samples were diluted with PBS 1:5 (vol/vol)

¾

aliquots were prepared (10 μl antibody and 50 μliter blood) and they were incubated for 30 minutes. The following monoclonal antibodies (mab) were applied: o

CD41 - directed against a complex of glycoprotein IIb/IIIa

o

either P-selectin (CD 62) or lysosomal GP53 (CD63) for detecting platelet activation

o

Isotype-matched mouse IgG was used for detecting non-specific binding

CD41-mab was conjugated with anti-mouse Phycoerythrin (PE), CD62-mab and CD63-mab with anti-mouse F(ab)2-fluorescein isothiocyanate (FITC). All antibodies were from Becton Dickinson, San Jose, USA. Fluorescence-activated cell sorter analysis (FACS Calibur,

- 50 -

Becton Dickinson, San José, USA) was run with 10 000 platelets per gate and a flow rate of 1 000 platelets/sec. The extent of fluorescence was obtained through channel-by-channel substraction of non-specific mouse-IgG binding from specific binding. All measurements were run in duplicate within 30 minutes after staining. Quality Control In our laboratory, the coefficient of variation for analysis of P-selectin was 5% and 4% for GP53. Contribution to the Study I was reponsible for establishing and validating the protocol of platelet measurement in our institute by modifying the assay described by Tschoepe for flow cytometry. Furthermore, I worked out the study protocol of the measurements with applying mental and physical stressors. I was also involved in developing the study design. I carried out the flow cytometry measurements as well as all the data input and statistical evaluations (flow cytometry, cardiovascular parameters, impedance cardiography, blood counts, general information).

4.3.3. Statistical Analyses Two different statistical approaches were used for data analyses. 1. A two-way ANOVA (rm-ANOVA) with independent groups (patients, normal controls) and repeated measures for task (baseline, mental stress, physical activity) seemed appropriate to explore the effect of stress on platelet activity. 2. However, rm-ANOVA explores the effect of stress as a whole, thus neglecting the influences of the individual stressors. To further differentiate the weight of the different tasks (mental stress, physical activity), t-tests were used: independent samples t-test for comparison of healthy controls to depressed patients and paired samples t-test for comparison of activation after mental and after physical stress.

We also examined the interrelationship between platelet baseline measures and clinical parameters via correlation analysis. All measures are reported as means + standard deviation, except for the flow cytometry values, which are given as geometrical means + standard deviation according to the recommendations of the developing company (Becton Dickinson, San Jose, USA). In every analysis we included the highest possible number of subjects. A p-value of less than .05 was considered indicating statistical significance.

- 51 -

5. RESULTS 5.1. Study 1. Health Behaviour, Health Beliefs, Risk Awareness and Psychosocial Well-Being in University and College Students in Hungary 5.1.1. Demographic Characteristics The sample characteristics are detailed in Table 5.1. TABLE 5.1. Sample characteristics of students aged 19-28 (n = 590) Men % (n) 40.2% (237) 21.7 ± 1.9 2.6% (6)

Women % (n) 59.8% (353) 21.9 ± 1.8 0.9% (3)

Total % (n) 100% (590) 21.8 ± 1.8 1.6% (9)

54.9% (130) 7.6% (18) 14.8% (35) 8.4% (20) 8.9% (21) 3.8% (9) 1.7% (4)

10.5% (37) 21.5% (76) 12.2% (43) 1.1% (4) 11.0% (39) 17.0% (60) 14.2% (50) 12.5% (44)

28.3% (167) 15.9% (94) 13.3% (78) 0.7 (4) 10.0% (59) 13.7% (81) 10.0% (59) 8.1% (48)

48.1% (114) 51.9% (123)

38.5% (136) 61.5% (217)

42.4% (250) 58.0% (340)

27.4% (65) 69.6% (165)

29.0% (102) 66.2% (233)

28.4% (167) 67.6% (398)

2.1% (5)

4.5% (16)

3.6% (21)

Educational level of mother Primary school High school College /university I do not know

6.9% (16) 45.5% (105) 47.2% (109) 0.4 % (1)

9.2% (32) 47.4% (165) 43.4% (151)

8.3% (48) 46.6% (260) 44.9% (260) 0.2% (1)

Educational level of father Primary school High school College /university I do not know

5.3% (12) 45.3% (102) 48.9% (110) 0.4% (1)

6.4% (22) 42.9% (147) 49.0% (168) 1.7% (6)

6.0% (34) 43.8% (249) 48.9% (278) 1.2% (7)

Gender Age (mean ± SD) Married Field of study Engineering and technology Liberal Arts Natural Sciences Fine Arts Law and state administration Economics Teachers’ training (higher grade) Teachers’ training (kindergarten & lower grade) Place of study Budapest Other cities During college term living At home with parents In college accomodation/rented room Own appartement

- 52 -

604 questionnaires were returned. 10 students had to be excluded because of missing data on age (1.7%) and 1 student because of missing data on gender (0.2%). As there were only 1 student/year above the age of 28, we also excluded students who were above 28 years (n = 3, 0.5%). The mean age of the sample was 21.8 years with a standard deviation of 1.81 years. There was no difference among the mean age of women (21.9 ± 1.8) and men (21.7 ± 1.9).

5.1.2. Behavioural Factors 5.1.2.1. Prevalence of Health and Risk Behaviours The prevalence of the health behaviours showed a varied picture, ranging from 12.8% (effort to eat high-fibre food in men) to 85.2% (doing exercise at least once in 2 weeks in men; see Table 5.2). A detailed description of the health behaviours follows after the table below. Furthermore, gender differences will also be presented. TABLE 5.2. Prevalence of health and risk behaviours (n = 590) Men

Women

Total

19.5%

19.3%

19.4%

ns

59.3%

51.6%

54.7%

.064

85.2%

83.2%

84.0%

ns

15.7%

5.4%

9.5%