Persistent cognitive depressive symptoms are associated with coronary artery calcification.

OBJECTIVES: The association between depression and sub-clinical atherosclerosis remains unclear. By assessing depressive symptoms only at one point in time, most previous studies have failed to ascertain long-term exposure. We examined the association of long-term depressive symptoms assessed at three time points (over 10 yrs) with a marker of sub-clinical atherosclerosis.
METHODS: Participants included 454 healthy, non-medicated men and women from the Whitehall II epidemiological cohort without known cardiovascular disease (CVD). Depressive symptoms were assessed at three time points (over 10 yrs) and coronary atherosclerosis was assessed at follow-up in terms of coronary artery calcification (CAC).
RESULTS: 18.9% of the sample reported depressive symptoms at least once during follow-up. Participants that were persistently depressed had over a two-fold increased risk of detectable CAC (Agatston score>0) (odds ratio [OR]=2.56, 95% CI, 1.14-5.78) and high CAC (Agatston score > or = 100) (OR=2.36, 1.04-5.35) compared with never depressed after adjustment for age, sex, and a range of conventional cardiac risk factors. These associations were more robust in men. Participants who were depressed on only one occasion were not at elevated risk of CAC.
CONCLUSIONS: Persistent cognitive symptoms of depression assessed over several time points, but not on a single occasion, are related to sub-clinical coronary atherosclerosis in men free of known CVD and diabetes. Copyright 2010 Elsevier Ireland Ltd. All rights reserved.

Introduction

Depressive symptoms have been consistently associated with greater risk of coronary heart disease (CHD) [1], although it is unclear whether depression precedes atherosclerosis. Several non-invasive imaging techniques can be employed to examine the extent of sub-clinical atherosclerosis, which begins early in life and precedes CHD [2]. Assessment of sub-clinical atherosclerosis before clinical disease is manifest helps delineate the temporal relationship between depression and CHD. Experimental studies in primates have demonstrated an association between depressive behaviour and accelerated atherogenesis [3], but evidence on humans is mixed.

The existing studies have generally utilised measures of carotid intima-media thickness (IMT), and carotid or coronary plaque burden (calcified or non-calcified) as markers of sub-clinical atherosclerosis. In some studies depression was associated with greater IMT [4,5], greater aortic and coronary calcifications [5,6] and greater progression of carotid IMT [7,8] and plaques [9]. In other studies concurrent depression has not been associated with carotid IMT or carotid plaques [10], nor with coronary calcification [11,12]. Furthermore, in a recent longitudinal study changes in depressive symptoms assessed over 15 years and history of major depression were unrelated to carotid IMT [13].

Since atherosclerosis develops over a long period of time, chronic or recurrent depression rather than transient symptoms may be associated more consistently with sub-clinical atherosclerosis. By assessing depressive symptoms at one point in time only, most previous studies may have failed to ascertain such long-term exposure.

In the present study, depressive symptoms were recorded at three time points over the 10 years before the measurement of sub-clinical atherosclerosis. It was therefore possible to contrast the relationship of point estimates and persistent depressive symptoms with sub-clinical disease.

Methods

Participants

A sub-sample of participants from the Whitehall II epidemiological cohort [14] was recruited for the assessment of coronary artery calcification (CAC) in 2006 to 2008. Enrolment was restricted to persons who were of white European origin, had no history or objective signs of CHD and no previous diagnosis or treatment for hypertension, diabetes, inflammatory diseases, or allergies. In addition, participants with major depression or taking anti-depressant medication in the 12 months prior to the heart scan assessment were not eligible. From the initially invited 1169 Whitehall II participants, 27.6% were not eligible (mainly because of prescribed medications) and 25.9% declined to take part (38.6% from lower work grades and 20.3% from higher grades). Thus, the study population included 543 participants, 294 men and 249 men, aged 53–76 years. They were prohibited from using any anti-histamine or anti-inflammatory medication 7 days before testing and were rescheduled if they reported colds or other infections on the day of testing. At the time of CAC assessment 8.7% of the sample were using lipid-lowering medication (statins). All participants gave full informed consent to participate in the study and ethical approval was obtained from the University College London committee on the Ethics of Human Research.

Depressive symptoms and health assessment

Depressive symptoms were assessed from a four-item scale derived from the 30-item General Health Questionnaire (GHQ-30), which has been validated by clinical interview [15]. The items assessed cognitive symptoms of depression only, and caseness was defined as a sum score of four or more (range 0 to 12), based on receiver operator curve analyses [15]. Depressive symptoms were assessed at three time points before the assessment of sub-clinical atherosclerosis (1997, 2001, 2002/4).

A full health assessment according to standardised protocols at the 2002/4 visit included anthropometric measures to determine body mass index (BMI) and a venous blood sample after at least 5 hours of fasting. Blood was analysed for C-reactive protein (CRP), which was performed using high-sensitivity ELISA (R & D Systems, Oxford, UK) and total and high-density-lipoprotein (HDL) cholesterol (serum for lipid analyses was refrigerated at −4 °C and assayed within 72 hours). LDL-cholesterol concentration was calculated using the Friedewald formula. In addition, self reported information about smoking and alcohol was recorded.

Assessment of coronary artery calcium

The assessment of CAC was performed in 2006–2008 using electron beam computed tomography (GE Imatron C-150, San Francisco, CA) as previously described [16]. In brief, 40 contiguous 3 mm slices were obtained during a single breath-hold starting at the carina and proceeding to the level of the diaphragm. Scan time was 100 ms/slice, synchronised to 40% of the R-R interval. Agatston and volumetric calcium scores were calculated to quantify the extent of CAC by a single experienced investigator blinded to the clinical data on an Aquarius workstation (TeraRecon Inc., San Mateo, CA). Since calcified volume was very highly correlated with Agatston score (Spearman's r = 0.99), we present data for Agatston score only.

Statistical analysis

Participants were categorised as having no depressive symptoms at any time point, one episode, and two or more episodes of depression. Because of the skewed distribution of CAC values, multivariate logistic regression analyses were employed to examine the association of depressive symptoms episodes with the presence of detectable CAC (Agatston score > 0) and high CAC (Agatston score ≥ 100). This threshold was based on the St. Francis Heart Study that demonstrated maximum sensitivity and specificity for detecting cardiovascular events at a threshold calcium score of greater than or equal to 100 [17]. We calculated odds ratios (OR) and 95% confidence intervals (CI) for the risk of CAC according to depressive symptoms category, adjusting sequentially for age, sex, education, smoking, statin medication use, alcohol, HDL- and LDL-cholesterol, CRP, and BMI. This modelling strategy was employed to examine if the associations between depression and atherosclerosis were independent of behavioural and cardiovascular risk factors. All analyses were conducted using SPSS version 15.

Results

Of the 543 participants, 60 individuals had missing data on depressive symptoms at one or more time points and a further 29 participants had missing clinical data. They were excluded from the analysis. There were no differences in demographic characteristics of those participants included and excluded from the present analyses. In the remaining sample of 454 participants with complete data on CAC, 18.9% reported depressive symptoms at least once and the remainder did not meet the threshold for depression on any of the three occasions. There were no notable differences in characteristics of the participants in relation to depressive symptoms, apart from the younger age of depressed (Table 1).

Table 1

Characteristics of the study population according to episodes of depressive symptoms assessed in 1997, 2001, 2002/4.

Variable	None (N = 368)	Once (N = 50)	Two+ (N = 36)
Age at baseline (yrs)	54.6 ± 5.7	52.1 ± 4.5*	51.8 ± 3.9*
Men (%)	54.9	48.0	58.3
Education (yrs)	14.7 ± 3.8	15.0 ± 4.1	14.7 ± 3.4
Current smokers (%)	9.8	10	13.9
Alcohol (units/wk)	11.4 ± 12.0	9.3 ± 9.6	11.4 ± 11.6
Systolic BP (mmHg)	123.6 ± 13.9	122.8 ± 15.2	123.4 ± 16.1
Body mass index (kg/m2)	25.4 ± 3.7	25.3 ± 4.4	25.9 ± 3.9
Total cholesterol (mmol/L)	5.85 ± 0.95	6.17 ± 1.10	5.63 ± 0.93**
HDL cholesterol (mmol/L)	1.71 ± 0.47	1.77 ± 0.50	1.63 ± 0.42
LDL cholesterol (mmol/L)	3.59 ± 0.90	3.82 ± 0.97	3.44 ± 0.84
Triglycerides (mmol/L)	1.19 ± 0.59	1.27 ± 0.66	1.19 ± 0.57
C-reactive protein (mg/L)	1.52 ± 2.01	1.68 ± 2.43	2.01 ± 2.36
Statin medication use (%)	8.4	10.0	11.1

p < 0.05 compared with never depressed.

p < 0.05 compared with once depressed.

Clinically relevant levels of CAC (Agatston score ≥ 100) were detected in 25.3% of the sample (36.4% men, 12.1% women), and CAC was not detectable in 43% of participants (29.1% men, 59.4% women). In Table 2a, persistent depression was associated with higher risk of detectable CAC (Agatston score > 0) in age and sex adjusted analyses (OR = 2.40, 1.09–5.29), which remained significant after further adjustments for education, smoking, alcohol, statin medication use and traditional cardiovascular risk factors (OR = 2.56, 1.14–5.78). This pattern of results was observed in both men and women although the point estimates did not reach conventional levels of statistical significance in the sex stratified analyses (see Table 2a).

Table 2a

Persistent depressive symptoms and risk of detectable coronary artery calcification (Agatston score > 0).

Depression incidence	Cases/N	Age adjusted OR (95% CI)	Model 1 OR (95% CI)	Model 2 OR (95% CI)
Alla
Never	204/368	1.00	1.00	1.00
Once	30/50	1.70 (0.89–3.25)	1.65 (0.86–3.18)	1.61 (0.83–3.12)
Twice or more	25/36	2.40 (1.09–5.29)	2.37 (1.07–5.26)	2.56 (1.14–5.78)
p-trend		0.038	0.047	0.040
Women
Never	66/166	1.00	1.00	1.00
Once	10/26	1.12 (0.47–2.69)	1.17 (0.48–2.83)	1.05 (0.42–2.63)
Twice or more	8/15	2.22 (0.75–6.61)	2.40 (0.79–7.28)	2.54 (0.83–7.82)
p-trend		0.355	0.301	0.263
Men
Never	138/202	1.00	1.00	1.00
Once	20/24	3.14 (1.01–9.78)	3.01 (0.96–9.43)	3.09 (0.97–9.83)
Twice or more	17/21	2.59 (0.82–8.16)	2.54 (0.80–8.07)	2.55 (0.76–8.59)
p-trend		0.050	0.060	0.068

Model 1: with further adjustment for education, smoking, alcohol.

Model 2: with further adjustment for systolic blood pressure, HDL and LDL cholesterol, CRP, BMI, statins medication.

All analyses contain additional adjustment for sex.

In Table 2b, we present results for depression and clinically relevant levels of CAC (Agatston score ≥ 100). Participants who were persistently depressed had over a two-fold increased risk of clinically relevant CAC (age and sex adjusted OR = 2.46, 95% CI, 1.12–5.42) compared with never depressed, whereas transient depressive symptoms were unrelated to CAC (OR = 1.12, 0.51–2.45). The associations between depressive symptoms and CAC persisted in the fully adjusted model (OR = 2.36, 1.04–5.35). In further analyses stratified by sex, the association was only observed in men but not in women for whom the numbers of significant CAC cases were smaller (p [for sex interaction] = 0.01, Table 2b). In persistently depressed men there was over a three-fold increased risk of CAC (fully adjusted OR = 3.87, 95% CI, 1.41–10.61) compared with men without any depression.

Table 2b

Persistent depressive symptoms and risk of clinically relevant coronary artery calcification (Agatston score ≥ 100).

Depression incidence	Cases/N	Age adjusted OR (95% CI)	Model 1 OR (95% CI)	Model 2 OR (95% CI)
Alla
Never	92/368	1.00	1.00	1.00
Once	10/50	1.12 (0.51–2.45)	1.11 (0.50–2.45)	1.02 (0.47–2.25)
Twice or more	13/36	2.46 (1.12–5.42)	2.44 (1.10–5.41)	2.36 (1.04–5.35)
p-trend		0.082	0.089	0.122
Women
Never	22/166	1.00	1.00	1.00
Once	2/26	0.65 (0.14–3.01)	0.53 (0.11–2.56)	0.53 (0.10–2.73)
Twice or more	1/15	0.60 (0.07–4.90)	0.54 (0.06–4.55)	0.49 (0.05–4.61)
p-trend		0.782	0.647	0.464
Men
Never	70/202	1.00	1.00	1.00
Once	8/24	1.42 (0.55–3.67)	1.46 (0.56–3.81)	1.34 (0.50–3.61)
Twice or more	12/21	3.90 (1.50–10.19)	3.94 (1.48–10.45)	3.87 (1.41–10.61)
p-trend		0.020	0.021	0.031

Model 1: with further adjustment for education, smoking, alcohol.

Model 2: with further adjustment for systolic blood pressure, HDL and LDL cholesterol, CRP, BMI, statins medication.

All analyses contain additional adjustment for sex.

Discussion

The present study examined the association between long-term depressive symptoms assessed at three time points with measures of coronary atherosclerosis. The results show that persistent depressive symptoms were associated with CAC in men although participants reporting depressive symptoms only at one time point were not at elevated risk of CAC. The findings therefore suggest it is the persistence of symptoms over time that is important.

Only one previous study [13] has examined long-term depressive symptoms assessed at different time points and they observed no association between persistent depression and carotid IMT. We did, however, observe associations for CAC. The reason why persistent depression is related to CAC but not IMT remains unclear. Some evidence has suggested that carotid IMT reflects an adaptive response of the vessel wall to changes in shear and tensile stress rather than atherosclerosis [18]. Studies that have examined associations between CAC and IMT have reported mixed findings, with some showing strong-graded associations [19] while others have demonstrated that there is a potential for misclassification from using different measures [20,21]. In addition, it is thought that CAC is a more reliable predictor of subsequent CHD events than carotid IMT [22,23]. Thus, the lack of associations between depressive symptoms and carotid IMT might be because measures of extracoronary atherosclerosis do not accurately reflect coronary atherosclerosis. It is also possible that the underlying pathology linking depression with cardiovascular disease (CVD) is specific to the coronary vessels. However, arguing against this interpretation is evidence linking depression with higher incidence of stroke [24,25]. The present study therefore demonstrates the importance of taking repeated measures of depression over time and suggests that different assessments of atherosclerosis might partly account for inconsistencies in the existing literature.

We observed considerable sex differences in the association between depression and CAC. Our sex-stratified analysis was, however, limited by the lower prevalence of CAC in women and small numbers of participants in depressed groups. These findings should therefore be replicated in further studies. Some previous evidence has suggested that CAC is a stronger predictor for CHD events than conventional risk factors in men whereas this relation is less pronounced in women [26]. Thus if CAC does not reflect risk to the same extent in women, this might partly explain why depression was associated with CAC only in men.

Several pathophysiological processes, such as inflammation, haemostasis, altered metabolic and cardiac autonomic control, might play an intermediate role in the causal pathway linking depression and coronary atherosclerosis [27]. In a recent study, adiposity appeared to partly mediate the association between depression and CAC in women [28] and other work suggests that health behaviours play a role in the depression-CVD risk link [29]. In the present sample, however, there were no significant differences in health behaviours, blood pressure, lipids, adiposity, or CRP between depressed and non-depressed participants. This might be because we specifically recruited participants that were healthy and without history of major depression or objective signs of CHD and no previous diagnosis or treatment for hypertension, diabetes, inflammatory diseases. In addition, the present sample of participants was originally recruited from the British civil service and may not be representative of the general population.

Several limitations should be addressed. Our measure of depression assessed cognitive symptoms only. Given that previous studies suggest that somatic symptoms of depression might better predict CHD events [30,31], the assessment of different dimensions of depression in the present study may have strengthened our findings. We did not obtain a measure of CAC at baseline, thus we cannot infer a longitudinal relationship between depressive symptoms and CAC with certainty, since this association might merely reflect differences already present at baseline. Therefore, we cannot rule out the possibility of reverse causality that presence of CAC might partly drive depressive symptoms. However, given that the participants were healthy at the time of the heart scan it is unlikely that significant CAC was present at baseline (10 yrs earlier). Lastly, the study was limited to white European participants without history of known CVD and diabetes, thus the results may not therefore generalise to the wider population since the presence and extent of CAC varies among different racial groups [32].

In conclusion, persistent cognitive symptoms of depression assessed over several time points, but not on a single occasion, are related to sub-clinical coronary atherosclerosis in men free of known CVD and diabetes.

Conflict of interest statement

The authors have not declared any conflict of interest.