Depressive symptoms as risk factor of cardiovascular mortality in older European men: the Finland,Italy and Netherlands Elderly (FINE) study

Abstract

Background

Depressive symptoms have been suggested to increase the risk of cardiovascular diseases, but this may reflect reversed causality. We investigated to what extent depressive symptoms are a true risk factor for cardiovascular mortality in elderly men.

Design

The Finland, Italy and Netherlands Elderly (FINE) study is a prospective cohort study conducted in Finland, Italy and The Netherlands.

Methods

Depressive symptoms were measured with the Zung self-rating Depression Scale in 799 elderly men, aged 70–90 years, free from cardiovascular diseases. Using Cox models, hazard ratios (HRs) were calculated for specific cardiovascular mortality endpoints. The analyses were adjusted for potential confounders, stratified on country and repeated after exclusion of men who died from cardiovascular diseases up to 5 years after baseline.

Results

During 10-years of follow-up 224 (28%) men died from cardiovascular diseases. The adjusted hazard for a five-point increase in depressive symptoms was 1.15 [95% confidence interval (CI) 1.08–1.23] for cardiovascular mortality. This risk was stronger for mortality from stroke (HR 1.35; 95% CI 1.19–1.53) and heart failure (HR 1.16; 95% CI 1.00–1.35) in comparison with mortality from coronary heart disease (HR 1.08; 95% CI 0.97–1.20) and other degenerative heart diseases (HR 1.06; 95% CI 0.91–1.23). Exclusion of men who died from cardiovascular diseases within 5 years after baseline did not change the strength of the associations. There were no significant differences in HRs between northern and southern Europe.

Conclusions

This study provides further and more convincing prospective evidence for depressive symptoms as a risk factor for cardiovascular mortality in elderly men.

Keywords

cardiovascular diseases cohort depression elderly men mortality

Introduction

Cardiovascular diseases (CVDs), including coronary heart disease (CHD) and stroke, are still the major causes of death in most European countries and account for 32% of all deaths [1]. There is increasing evidence that besides classical risk factors, such as hypertension, hypercholesterolemia, diabetes and smoking, psychosocial factors and depression also play a role in the occurrence of cardiovascular diseases [2].

In patients with CVD depression is common. For instance, one in five patients with newly diagnosed coronary disease is depressed [3]. Depression is not only common in patients with cardiovascular disease, it may also worsen the prognosis and may lead to a higher mortality from CVD [4, 5]. Several explanations have been put forward for a worse prognosis associated with depression, such as higher disease severity, unhealthy behavior, non-adherence to cardiac treatment, antide-pressant cardiotoxicity, more cardiovascular risk factors, inflammation, autonomic dysfunction and increased platelet activation [6].

More recently, it has been hypothesized that depression not only leads to poor prognosis in patients with established CVD, but also increases the risk of CVD in initially healthy persons. Recent reviews show evidence for the hypothesis that depression is a risk factor for CVD [2, 7]. Due to heterogeneity between studies, however, results of these studies are difficult to compare. For instance, most of the studies reviewed used different cardiovascular endpoints. In addition, most of the studies were performed in middle-aged persons. In elderly men the results are less conclusive. For instance, mild and major depression have been associated with increased total mortality [8] and CVD mortality [9–11] but the results are not conclusive for CHD [9, 12], stroke [13] and heart failure incidence [14, 15]. Also, depression may still be a consequence of subclinical disease, thus indicating reversed causality [16–19].

We investigated the relation between depressive symptoms and risk of cardiovascular mortality in a population-based sample of older men in three European countries. We investigated specific cardiovascular mortality endpoints, and excluded men who died from CVD in the first 5 years after baseline to minimize the possibility of reversed causality.

Methods

The Finland, Italy and Netherlands Elderly study

The Finland, Italy and Netherlands Elderly (FINE) study is a prospective population-based cohort study on risk factors and health in elderly men. The study design and measurements have been described in detail elsewhere [20]. In brief, the FINE study started in 1984 as a continuation of the Seven Countries Study (SCS), which was originally initiated in 1958 by Keys [21] as a cardiovascular risk factor survey among 12 763 middle-aged men, born between 1900 and 1920. In The Netherlands, the surviving cohort was enlarged in 1985 to an elderly cohort of 939 men, by inviting a random sample of men born between 1900 and 1920 who had not participated in the study before. In total, 2285 men from Finland (n = 716), Italy (n = 682) and The Netherlands (n = 887) participated in the baseline examination of the study in 1985. Informed consent was obtained from all study participants. Data collection followed the international protocol used in previous surveys of the SCS [21]. In 1989–1991 the second round of the FINE study took place. In this round, measures on depression, functional status, cognitive function and self-reported health were added. Around 1995 and 2000 the third and fourth round of examinations were carried out. Mortality data were collected until the year 2000.

Depressive symptoms

Depressive symptoms were measured using the self-rating depression scale (SDS), developed by Zung [22]. This scale was used to assess depression among patients admitted to a psychiatric hospital, but also for non-institutionalized elderly [23] and was highly comparable among different countries [24]. The reproducibility of the SDS is good in elderly men (Cronbach α 0.75) [25] and has been validated repeatedly with other questionnaires on depressive symptoms, such as the Center for Epidemiological Studies Depression Scale (r = 0.69) [26], the Geriatric Depression Scale (r = 0.59) [27] and the Hamilton Depression scale (r = 0.80) [28]. The questionnaire contained 20 either positively or negatively formulated items, based on clinical diagnostic criteria commonly used to diagnose depressive disorders. The answers on those items were coded on a four-point Likert-type scale varying from ‘none or sometimes’ to ‘most or always'. Negative items were converted into positive ones with total scores ranging from 20 to 80. Men with more than two missing items were excluded for the analysis. If one or two items were missing, the mean score of the valid items of the subject was calculated and the missing items were replaced by this mean score. An index for the SDS was derived by dividing the sum of the answers by 80 times 100 (range 25–100), with a higher score indicating more depressive symptoms. The original clinical cut-off values are: no depression (< 50); mild depression (50–59); and moderate-severe depression (≤60) [23]. To increase power a continous measure of the SDS was used (five-point increase) and the SDS was categorized into country-specific tertiles.

Cardiovascular endpoints

Mortality data were collected during 10 years of follow-up and obtained through official death certificates. One person was lost to follow-up. He was included in the analyses, but censored at the date of the examination round that he was lost to follow-up (after 3.6 years). Men who died from other causes were censored at date of death; for men who were still alive in 2000 the censor date was set at the examination date of the last round. In Finland, information on the causes of death was obtained from Finnish death registers; in Italy and The Netherlands the information was obtained from hospital registries or general practitioners. Original coding was done locally and checked by one clinical epidemiologist, who was blinded to the risk factor status of the subject.

All diagnoses were coded according to the International Classification of Diseases (ICD), ninth revision. The following ICD codes were used: CVD (390–459), coronary heart disease (CHD) (410–414), stroke (430–438), heart failure (428) and other degenerative heart disease (oDHD) (401–405, 426-427 and 429). For the analyses, both primary and secondary causes of death were considered.

Cardiovascular risk factors

The self-administered questionnaire contained questions on demographic characteristics, educational level, lifestyle habits, past and current morbidity. Participants were classified as current, past or never smokers according to their smoking habits. Alcohol intake was derived from the cross-check dietary history in The Netherlands, and in Finland and Italy assessed with self-administered questionnaires [29]. A proportion of the Finnish men (n = 62) had no information on alcohol intake, because in 1990 a reduced questionnaire was used. For these men the alcohol intake from 1985 was used. Alcohol intake was expressed as grams of alcohol per day, and classified into three categories (0, 1–29 and ≤30 g/day) [29]. Body mass index (BMI) (kg/m²) was calculated from weight and height, which were measured while the participant was standing in light clothing without shoes. Physical activity was assessed with a self-administered validated questionnaire designed for retired men [30] and expressed as minutes of total physical activity per week (min/week). Self-reported disability was measured using a standardized questionnaire about routine daily activities and classified as moderate/severe disabilities, compared to no/mild disabilities, following a hierarchical disability coding scheme [31].

Arterial blood pressure was measured twice on the right arm after 5 min of rest, with the man in a supine position. In Finland and Italy standard mercury sphygmomanometers were used, while in The Netherlands a random zero sphygmomanometer was used. The average of two readings of both systolic and diastolic blood pressure (fifth Korotkoff phase) was calculated. Venous blood samples (fasting in Finland and non-fasting in Italy and The Netherlands) were taken and total and high-density lipoprotein (HDL) cholesterol (mmol/l) were determined using standardized procedures according to the criteria of the World Health Organization's Lipid Reference Laboratories in Prague, Czech Republic, or Atlanta, Georgia [29]. History of myocardial infarction was obtained using the London School of Hygiene and Tropical Medicine questionnaire [32], verified by information from general practitioners or hospital registries. The clinical history of stroke, heart failure and diabetes was based on the doctor's conclusion using questionnaire information and the results of the physical examination.

Study sample

Depressive symptoms were measured for the first time during the second round of the FINE study, between October 1989 and November 1991, when in total 1416 men (82%) participated of the 1734 men still alive. At that time, 909 (64%) were free of CVD and diabetes. We excluded 85 men with more than two missing items in the SDS as well as 25 men with missing values in covariates. Thus, a study sample of 799 men remained for analysis, 229 men from Finland, 332 men from The Netherlands and 238 men from Italy.

Statistical analysis

Means and standard deviations (SDs) were computed for continuous baseline variables, and medians and 10–90 percentiles for continous variables with a skewed distribution. Frequency distributions were given for categorical variables according to country-specific tertiles of the SDS index score. Differences between SDS tertiles were tested for each country, with ANOVA, Kruskal-Wallis (in case of skewed distribution) or X²-test.

To investigate the association between depressive symptoms and CVD mortality, Cox proportional hazard analyses were performed. By censoring other types of events we could analyze different CVD mortality endpoints separately. Age was used as the time scale in the proportional hazards models to optimally adjust for age [33]. Age at entry was assigned as the age at the date of examination in 1990. Censoring age was defined as age at death or end of study. A hazard curve of the follow-up years versus the percentage of subjects who died of CVD was made, according to tertiles of depressive symptoms. With a log versus log minus log plot the assumptions of proportional hazards were checked. The assumptions of proportionality were not violated. Analyses were done for five-point increase and tertiles of baseline depressive symptoms. In the first model we adjusted for age (in the time-axis). In the second model analyses we also adjusted for possible confounders: country (Finland, The Netherlands, Italy), years of education, body mass index (kg/m²), smoking (never, past, current), alcohol intake (0, 1–29/ ≤30 g/day), SBP (mmHg), total and HDL cholesterol levels (mmol/l) and physical activity (min/week). Analyses were additionally adjusted for disabilities and repeated with clinical categories of depressive symptoms (no, < 50; mild, 50–59; moderate-severe, ≤60). To evaluate whether country modified the association between depressive symptoms and cardiovascular mortality, analyses were also done after stratification by country and interaction terms of country and the SDS index score were included. Analyses were repeated after exclusion of men who died from CVD in the subsequent 5 years after baseline.

All analysis were performed with SPSS version 12.0.1 (SPSS Inc, Chicago, Illinois, USA) and the SAS statistical software package, version 8.2 (SAS Institute, Cary, North Carolina, USA) [34]. Point estimates are given with corresponding 95% confidence intervals (CIs).

Results

After 10 years of follow-up 396 (50%) men had died and 403 men were still alive. Two-hundred and twenty-four (28%) men died from CVD, 93 from CHD, 66 from stroke, 45 from heart failure and 46 from oDHD. The total number of person-years was 5893 and the mean follow-up time was 7.4 years (SD = 3.0). The average SDS score was 49.2 (SD = 11) in Italy, 45.4 (SD = 10) in Finland and 42.6 (SD = 10) in The Netherlands. Table 1 presents the baseline characteristics of the men (n = 799) for each country and for the whole study population, according to country-specific tertiles of the SDS score. Men with more depressive symptoms had received shorter education, were more disabled and less physically active. Although not significant, men from Finland and Italy with more depressive symptoms seemed to have lower total and HDL cholesterol levels. One-hundred and ten men (8%) who were excluded from the analyses due to missing values in the SDS or other covariates were older, less educated and less physically active, more disabled and had a higher mortality rate in comparison with men who were not excluded (results not shown).

Table 2 presents hazard ratios for a five-point increase and tertiles of depressive symptoms and different CVD mortality endpoints. The age-adjusted hazard ratio for the continuous SDS score and CVD mortality was 1.14 (95% CI 1.08–1.22), indicating that for each increase of five points on the SDS (range 25–100) the risk of mortality from CVD increased with 14%. This hazard ratio did not change after additional adjustments for demographical and other cardiovascular risk factors (model 2). In addition, the association was slightly attenuated after adjustment for disabilities (HR 1.12; 95% CI 1.05–1.20). The associations were stronger for mortality due to stroke (HR 1.35; 95% CI 1.19–1.53) and heart failure (HR 1.16; 95% CI 1.00–1.35) in comparison with mortality due to CHD (HR 1.08; 95% CI 0.97–1.20) and oDHD (HR 1.06; 95% CI 0.91–1.23). Figure 1 illustrates a fully adjusted cumulative hazard curve for CVD mortality according to country-specific tertiles of depressive symptoms. The death rate increased with the number of depressive symptoms. In addition, when using the clinical cutoff scores, the adjusted hazard ratios for CVD mortality were 1.72 (95% CI 1.25–2.37) for men with mild depression (50–59) and 1.87 (95% CI 1.25–2.80) for moderate-severe depression (≤60), compared to men with no depression (< 50). Analyses with the primary death causes showed the same estimates with wider CIs (results not shown).

Exclusion of early CVD deaths did not materially change the results (Table 3). After exclusion of 144 men who died within 5 years after baseline, the hazard ratio for a five-point increase was 1.16 (95% CI 1.03–1.30).

After stratification for country the adjusted hazard ratios for CVD mortality did not differ between countries and none of the interaction terms were statistically significant (results not shown).

Discussion

The results of this prospective study among elderly men without CVD at baseline show that depressive symptoms are associated with an increased risk of future cardiovascular mortality. The risk is highest for mortality due to stroke and heart failure. There are no risk differences observed between northern and southern Europe.

A major strength of this study is its prospective longitudinal design, which made it possible to determine depressive symptoms before onset of CVD. Furthermore, the long follow-up period and large sample size enabled us to examine specific causes of death and to exclude subjects who died from CVD in the first 5 years after baseline, making the possibility of reversed causality even more unlikely. In addition, the estimated effects were not disturbed by a selective loss to follow-up, because the mortality follow-up was complete. Finally, to investigate the independent effect of depressive symptoms, we were able to adjust for many potential confounding variables, in particular, lifestyle and CVD risk factors.

We observed a dose-response relationship between depressive symptoms at baseline and CVD mortality providing further evidence of a causal relationship. Also, we extensively adjusted for possible confounding variables, in particular, lifestyle and CVD risk factors. These factors explained only a small part of the association between depressive symptoms and CVD mortality. Additional adjustments for disabilities attenuated the strength of the association a little further, but there was still an independent effect of depressive symptoms.

Our results provide further and more convincing evidence of depression as a true risk factor for CVD [2, 7]. The strength of the association for CHD mortality that we observed is highly similar to the risk ratio of 1.64 that was found in two systematic reviews [17, 35]. In the present study we found stronger associations for depressive symptoms and mortality from stroke and heart failure, in comparison with mortality from CHD and oDHD. To date, in the elderly, results on specific cardiovasular endpoints such as CHD [9, 12] and stroke [13] have been inconclusive. In addition, increased risks on heart failure incidence were observed in elderly women and hypertensive patients [14, 15]. Our study is the first to observe an increased risk for heart failure mortality in elderly men without CVD at baseline.

Table 1

Baseline characteristics per country and tertiles of the Zung self-rating depression scale index score (n = 799)

	Finland			The Netherlands			Italy			All countries
	Depressive symptoms			Depressive symptoms			Depressive symptoms			Depressive symptoms
	Low	Middle	High	Low	Middle	High	Low	Middle	High	Low	Middle	High
	< 42	42–49	< 49	< 37	37–45	< 45	< 43	43–53	>53
	n=79	n=74	n=76	n=104	n=123	n=105	n=78	n=81	n=79	n=261	n=278	n=260
Age (years); mean (SD)	75 (5)	77 (5)	77 (5)	75 (4)	75 (4)	76 (5)	77 (4)	78 (4)	78 (4)	76 (4)	76 (4)	77 (5)
Education (years) Mean (SD)	4.6 (3.0)	4.9 (3.6)	3.8 (2.4)	11.1 (4.3)	11.1 (4.7)	9.2 (3.3)∗	5.6 (3.2)	4.9 (2.6)	4.1 (1.7)∗	7.5 (4.7)	7.6 (5.0)	6.1 (3.7)
Alcohol
0 g/day (%)	41	26	24	31	22	23	19	17	23	31	22	23
1–29 g/day (%)	56	67	71	48	59	65	50	52	35	51	59	58
≤30 g/day (%)	3	7	5	21	19	12	31	31	42	18	19	19
Smoking
Current (%)	13	14	12	24	23	23	18	11	22	19	17	20
Past (%)	47	58	54	58	58	60	44	54	47	50	57	54
Never (%)	40	28	34	18	19	17	38	35	31	31	26	26
Physical activity
Median	630	480	415∗	603	540	420∗∗	978	680	245∗∗	795	540	390∗∗
10–90 percentile (min/week)	210–2760	75–1500	60–1395	180–1335	105–1190	32–910	180–2400	90–1905	0–1830	1180–1985	90–1575	25–1200
Disability (%)	5	10	19∗	5	14	16∗	5	9	33∗∗	5	11	22
BMI	26.1 (3.7)	26.1 (3.9)	26.0 (3.9)	25.5 (2.9)	25.3 (2.9)	26.1 (3.0)	26.3 (3.4)	26.3 (3.6)	25.8 (4.1)	25.9 (3.3)	25.8 (3.4)	25.9 (3.6)
SBP (mmHg) Mean (SD)	153 (20)	161 (23)	158 (22)	150 (19)	147 (20)	153 (21)	164 (19)	164 (19)	159 (19)	155 (20)	156 (220	156 (21)
DBP (mmHg) Mean (SD)	83 (11)	87 (11)	84 (11)	81 (10)	81 (11)	84 (13)	86 (10)	87 (8)	85 (9)	84 (11)	85 (11)	84 (11)
Hypertension (%)	51	64	57	43	35	44	73	69	65	54	53	54
Total cholesterol (mmol/l) Mean (SD)	5.7 (1.1)	5.6 (0.9)	5.6 (1.2)	6.1 (1.2)	6.1 (1.1)	6.1 (1.1)	5.6 (1.2)	5.5 (1.0)	5.4 (0.9)	5.8 (1.2)	5.8 (1.1)	5.7 (1.1)
HDL cholesterol (mmol/l) Mean (SD)	1.1 (0.2)	1.2 (0.3)	1.2 (0.3)	1.2 (0.3)	1.2 (0.3)	1.1 (0.3)	1.38 (0.4)	1.32 (0.3)	1.33 (0.3)	1.2 (0.3)	1.2 (0.3)	1.2 (0.3)

Disability is defined as moderate to severe, compared to no to mild, following the hierarchical disability score. Hypertension was defined as a systolic blood pressure of ≤160 mmHg, or a diastolic blood pressure of ≤95 mmHg or use of medication. SD, standard deviation; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; HDL, high-density lipoprotein. ∗P<0.05; ∗∗P<0.001.

Table 2

Hazard ratios with 95% confidence interval (CI), according to a five-point increase and country specific tertiles of depressive symptoms and mortality from specific cardiovascular diseases

	Cases (person years)	Model 1 hazard ratio (95% CI)	Model 2 hazard ratio (95% CI)
Cardiovascular disease
Depressive symptoms
Five-point increase		1.14 (1.08–1.22)	1.15 (1.08–1.23)
Low	51 (2083)	Reference	Reference
Middle	76 (2087)	1.39 (0.97–1.99)	1.35 (0.84–1.94)
High	97 (1723)	2.15 (1.53–3.03)	2.07 (1.46–2.94)
P for trend		< 0.0001	< 0.0001
Coronary heart disease
Depressive symptoms
Five-point increase		1.07 (0.97–1.17)	1.08 (0.97–1.20)
Low	25 (2083)	Reference	Reference
Middle	31 (2087)	1.18 (0.70–2.01)	1.19 (0.70–2.04)
High	37 (1723)	1.71 (1.03–2.84)	1.68 (0.99–2.84)
P for trend		0.04	0.05
Stroke
Depressive symptoms
Five-point increase		1.30 (1.17–1.45)	1.35 (1.19–1.53)
Low	11 (2083)	Reference	Reference
Middle	21 (2087)	1.78 (0.85–3.72)	1.65 (0.78–3.47)
High	34 (1723)	3.50 (1.77–6.91)	3.41 (1.69–6.90)
P for trend		< 0.0001	0.0003
Heart failure
Depressive symptoms
Five-point increase		1.16 (1.01–1.33)	1.16 (1.00–1.35)
Low	8 (2083)	Reference	Reference
Middle	15 (2087)	1.56 (0.65–3.73)	1.56 (0.64–3.57)
High	22 (1723)	2.91 (1.29–6.56)	2.77 (1.19–6.42)
P for trend		0.006	0.01
Other degenerative heart disease
Depressive symptoms
Five-point increase		1.14 (1.00–1.31)	1.06 (0.91–1.23)
Low	16 (2083)	Reference	Reference
Middle	14 (2087)	0.80 (0.38–1.66)	0.71 (0.34–1.50)
High	16 (1723)	1.14 (0.57–2.28)	0.93 (0.45–1.92)
P for trend		0.73	0.85

Model 1, adjusted for age (in the time-axis). Model 2: Additional adjusted for country (Finland, The Netherlands, Italy), education (years), body mass index (continuously), smoking (never/past/current), alcohol intake (0/1–29/ ≤30 g/day), systolic blood pressure (mmHg), total and high-density lipoprotein cholesterol levels (mmol/l) and physical activity (min/week).

In the present study we did not have clinical diagnoses of depression. As a result, this study cannot draw conclusions on the relation between clinical diagnoses of depression and CVDs mortality. The analyses with depressive symptoms classified according to the original clinical cut-off values, however, showed similar results as the analyses with tertiles of depressive symptom scores. Moreover, the linear and dose-response relationship we observed suggests that there is no clinical threshold to increase the risk of CVD mortality.

Several mechanisms have been hypothesized to explain the relationship between depression and CVD [6]. First, dietary habits or nutrients might play a role. Depressed persons might have a higher saturated-fat intake or a lower Ω-3 fatty acid intake compared with controls [36]. Second, neuro-endocrine changes during depression may accelerate the development of CVD. For instance, dysregulation of the hypothalamic-pituitary-adrenocortical axis, with its elevated cortisol levels is often present in depression and may lead to atherosclerosis-inducing actions such as injury of vascular endothelial cells, hypertension, hypercholesterolemia and inflammation [37]. Depression is also associated with excessive sympathetic and diminished parasympathetic nervous system activity, which may influence the progression of heart failure by impaired left ventricular function and renal sodium retention [38, 39]. Autonomic dysfunction may also lead to decreased heart rate variability [40], which in turn is associated to arrhythmia and sudden cardiac death [41]. Some mechanisms or factors such as dietary Ω-3 fatty acid intake [36] or inflammation [42] are also associated with an increased risk of depression, indicating that they might be common causes for depression as well as CVD. In the present study, we were not able to discern these potential mechanisms. Our rigorous exclusion of men with prevalent CVD and diabetes at baseline and our extensive adjustment for classical cardiovascular and lifestyle risk factors indicate, however, that the relation between depression and CVD mortality cannot be explained by unhealthy behavior or severity of disease. The relatively high risk we observed for mortality from stroke associated with depressive symptoms may be supportive for the hypothesis that vascular pathology of the brain precedes depression [43].

Fig. 1

Hazard curves for cardiovascular mortality according to country-specific tertiles of depressive symptoms, adjusted for age, country (Finland, The Netherlands, Italy), education (years), body mass index (continuously), smoking (never/past/current), alcohol intake (0/1–29/ ≤30 g/day), systolic blood pressure (mmHg), total and high-density lipoprotein cholesterol levels (mmol/l) and physical activity (min/week).

Table 3

Adjusted hazard ratios for five-point increase and country-specific tertiles of depressive symptoms and cardiovascular mortality, after exclusion of men who died from cardiovascular disease in the 5 years following baseline

	Years of exclusion after baseline
	1	2	3	4	5
Number excluded	41	64	85	112	144
Depressive symptoms
Five-points	1.16 (1.08–1.25)	1.16 (1.07–1.26)	1.17 (1.08–1.28)	1.14 (1.04–1.26)	1.16 (1.03–1.30)
Low	Reference	Reference	Reference	Reference	Reference
Middle	1.28 (0.85–1.92)	1.24 (0.80–1.92)	1.24 (0.78–2.06)	1.30 (0.78–2.17)	1.43 (0.77–2.66)
High	2.18 (1.48–3.20)	2.22 (1.47–3.26)	2.21 (1.46–3.55)	2.06 (1.25–3.40)	2.52 (1.38–4.60)
P for trend	< 0.0001	< 0.0001	0.0002	0.004	0.002

Hazard ratios are adjusted for age, country (Finland, The Netherlands, Italy), education (years), body mass index (continuously), smoking (never/past/current), alcohol intake (0/1–29/ ≤30 g/day), systolic blood pressure (mmHg), total and high-density lipoprotein cholesterol levels (mmol/l) and physical activity (min/week).

Finally, some methodological issues of the current study must be considered. First, selective participation of healthier respondents and exclusion of men with missing values who had a worse health profile may have led to a dilution of the observed hazard ratios. Second, the use of death certificates may have resulted in misclassification of some of the mortality endpoints. To minimize misclassification, however, in The Netherlands and Italy the death certificates were verified with clinical records and in Finland they were linked with the death registers. Furthermore, because the classification of mortality outcome was blinded for baseline status, possible misclassification of the outcome was not related with depression, and will therefore only have resulted in dilution of the associations. Finally, in an observational study the possibility of residual confounding cannot be excluded. We adjusted for many known classical cardiovascular risk factors, however, to minimize this possibility.

In conclusion, the results of the present longitudinal study among elderly men provide strong support for the hypothesis that depressive symptoms are a true risk factor for the occurrence of cardiovascular disease. Future studies should investigate what the underlying mechanisms are and which factors may be intermediates or common causes. They may suggest what preventive or therapeutic strategies may be effective to reduce depression in the elderly, who are often undertreated, and consequently the risk of cardiovascular disease.

Footnotes

Acknowledgements

The authors would like to thank Ingeborg Van der Tweel for her statistical advice, Carolien van den Brink and Boukje van Gelder for their database management.

References

1 British Heart Foundation. European cardiovascular disease statistics. London: British Heart Foundation; 2000.

Hemingway

Marmot

Evidence based cardiology: psychosocial factors in the aetiology and prognosis of coronary heart disease. Systematic review of prospective cohort studies. BMJ 1999; 318:1460–1467.

3 Schleifer

Macari-Hinson

Coyle

Slater

Kahn

Gorlin

. The nature and course of depression following myocardial infarction. Arch Intern Med 1989; 149:1785–1789.

4 van Melle

de Jonge

Spijkerman

Tijssen

Ormel

van Veldhuisen

. Prognostic association of depression following myocardial infarction with mortality and cardiovascular events: a meta-analysis. Psychosom Med 2004; 66:814–822.

Barth

Schumacher

Herrmann-Lingen

Depression as a risk factor for mortality in patients with coronary heart disease: a meta-analysis. Psychosom Med 2004; 66:802–813.

Carney

Freedland

Miller

Jaffe

AS.

Depression as a risk factor for cardiac mortality and morbidity: a review of potential mechanisms. J Psychosom Res 2002; 53:897–902.

Frasure-Smith

Lesperance

Reflections on depression as a cardiac risk factor. Psychosom Med 2005; 67(Suppl 1):S19–S25.

8 Schoevers

Geerlings

Beekman

Penninx

Deeg

Jonker

. Association of depression and gender with mortality in old age. Results from the Amsterdam Study of the Elderly (AMSTEL). Br J Psychiatry 2000; 177:336–342.

9 Penninx

Guralnik

Mendes de Leon

Pahor

Visser

Corti

. Cardiovascular events and mortality in newly and chronically depressed persons >70 years of age. Am J Cardiol 1998; 81:988–994.

10.

Whooley

Browner

WS.

Association between depressive symptoms and mortality in older women. Study of Osteoporotic Fractures Research Group. Arch Intern Med 1998; 158:2129–2135.

11.

11 Marzari

Maggi

Manzato

Destro

Noale

Bianchi

. Depressive symptoms and development of coronary heart disease events: the Italian longitudinal study on aging. J Gerontol A Biol Sci Med Sci 2005; 60:85–92.

12.

12 Mendes de Leon

Krumholz

Seeman

Vaccarino

Williams

Kasl

. Depression and risk of coronary heart disease in elderly men and women: New Haven EPESE, 1982–1991. Established Populations for the Epidemiologic Studies of the Elderly. Arch Intern Med 1998; 158:2341–2348.

13.

13 Wassertheil-Smoller

Applegate

Berge

Chang

Davis

Grimm

Jr . Change in depression as a precursor of cardiovascular events. SHEP Cooperative Research Group (Systolic Hypertension in the Elderly). Arch Intern Med 1996; 156:553–561.

14.

Abramson

Berger

Krumholz

Vaccarino

Depression and risk of heart failure among older persons with isolated systolic hypertension. Arch Intern Med 2001; 161:1725–1730.

15.

Williams

Kasl

Heiat

Abramson

Krumholz

Vaccarino

Depression and risk of heart failure among the elderly: a prospective community-based study. Psychosom Med 2002; 64:6–12.

16.

16 Stewart

North

West

Sharples

Simes

Colquhoun

. Depression and cardiovascular morbidity and mortality: cause or consequence? Eur Heart J 2003; 24:2027–2037.

17.

Rugulies

Depression as a predictor for coronary heart disease. A review and meta-analysis. Am J Prev Med 2002; 23:51–61.

18.

Everson

Roberts

Goldberg

Kaplan

GA.

Depressive symptoms and increased risk of stroke mortality over a 29-year period. Arch Intern Med 1998; 158:1133–1138.

19.

Jonas

Mussolino

ME.

Symptoms of depression as a prospective risk factor for stroke. Psychosom Med 2000; 62:463–471.

20.

20 Bijnen

Feskens

Caspersen

Giampaoli

Nissinen

Menotti

. Physical activity and cardiovascular risk factors among elderly men in Finland, Italy, and the Netherlands. Am J Epidemiol 1996; 143:553–561.

21.

Keys

Seven countries: a multivariate analysis of death and coronary heart disease. London: Harvard University Press; 1980.

22.

Zung

WW.

A self-rating depression scale. Arch Gen Psychiatry 1965; 12:63–70.

23.

Zung

Richards

Short

MJ.

Self-rating depression scale in an outpatient clinic. Further validation of the SDS. Arch Gen Psychiatry 1965; 13:508–515.

24.

Zung

WW.

A cross-cultural survey of symptoms in depression. Am J Psychiatry 1969; 126:116–121.

25.

Kivela

Pahkala

Sex and age differences of factor pattern and reliability of the Zung Self-rating Depression Scale in a Finnish elderly population. Psychol Rep 1986; 59(2 Pt 1):587–597.

26.

DeForge

Sobal

Self-report depression scales in the elderly: the relationship between the CES-D and ZUNG. Int J Psychiatry Med 1988; 18:325–338.

27.

Dunn

Sacco

WP.

Psychometric evaluation of the Geriatric Depression Scale and the Zung Self-Rating Depression Scale using an elderly community sample. Psychol Aging 1989; 4:125–126.

28.

Biggs

Wylie

Ziegler

VE.

Validity of the Zung Self-rating Depression Scale. Br J Psychiatry 1978; 132:381–385.

29.

Kromhout

Nissinen

Menotti

Bloemberg

Pekkanen

Giampaoli

Total and HDL cholesterol and their correlates in elderly men in Finland, Italy, and The Netherlands. Am J Epidemiol 1990; 131:855–863.

30.

Caspersen

Bloemberg

Saris

Merritt

Kromhout

The prevalence of selected physical activities and their relation with coronary heart disease risk factors in elderly men: the Zutphen Study, 1985. Am J Epidemiol 1991; 133:1078–1092.

31.

Hoeymans

Feskens

van den Bos

Kromhout

Measuring functional status: cross-sectional and longitudinal associations between performance and self-report (Zutphen Elderly Study 1990–1993). J Clin Epidemiol 1996; 49:1103–1110.

32.

Rose

Blackburn

Cardiovascular survey methods. Monogr Ser World Health Organ 1968; 56:1–188.

33.

Commenges

Letenneur

Joly

Alioum

Dartigues

JF.

Modelling age-specific risk: application to dementia. Stat Med 1998; 17:1973–1988.

34.

34 SAS Institute I. Statistical analysis system, version 8.2. Cary: SAS Institute, Inc; 1997.

35.

Wulsin

Singal

BM.

Do depressive symptoms increase the risk for the onset of coronary disease? A systematic quantitative review. Psychosom Med 2003; 65:201–210.

36.

Severus

Littman

Stoll

AL.

Omega-3 fatty acids, homocysteine, and the increased risk of cardiovascular mortality in major depressive disorder. Harv Rev Psychiatry 2001; 9:280–293.

37.

Musselman

Evans

Nemeroff

CB.

The relationship of depression to cardiovascular disease: epidemiology, biology, and treatment. Arch Gen Psychiatry 1998; 55:580–592.

38.

38 Veith

Lewis

Linares

Barnes

Raskind

Villacres

. Sympathetic nervous system activity in major depression. Basal and desipramine-induced alterations in plasma norepinephrine kinetics. Arch Gen Psychiatry 1994; 51:411–422.

39.

Packer

New concepts in the pathophysiology of heart failure: beneficial and deleterious interaction of endogenous haemodynamic and neurohormonal mechanisms. J Intern Med 1996; 239:327–333.

40.

40 Carney

Freedland

Veith

Cryer

Skala

Lynch

. Major depression, heart rate, and plasma norepinephrine in patients with coronary heart disease. Biol Psychiatry 1999; 45:458–463.

41.

Luukinen

Laippala

Huikuri

HV.

Depressive symptoms and the risk of sudden cardiac death among the elderly. Eur Heart J 2003; 24:2021–2026.

42.

Tiemeier

van Dijck

Hofman

Witteman

Stijnen

Breteler

MM.

Relationship between atherosclerosis and late-life depression: the Rotterdam Study. Arch Gen Psychiatry 2004; 61:369–376.

43.

43 Alexopoulos

Meyers

Young

Campbell

Silbersweig

Charlson

‘Vascular depression’ hypothesis. Arch Gen Psychiatry 1997; 54:915–922.