Introduction
Meta-analyses have shown that depressed, but otherwise healthy, individuals have an increased risk of developing coronary artery disease (CAD) (Van der Kooy et al. Reference Van der Kooy, van Hout, Marwijk, Marten, Stehouwer and Beekman2007) and that depressed cardiac patients are at an increased risk of subsequent cardiovascular events (Carney et al. Reference Carney, Rich, Freedland, Saini, teVelde, Simeone and Clark1988; Barefoot et al. Reference Barefoot, Helms, Mark, Blumenthal, Califf, Haney, O'Connor, Siegler and Williams1996; Wouts et al. Reference Wouts, Oude Voshaar, Bremmer, Buitelaar, Penninx and Beekman2008). Large randomized controlled trials aimed at improving cardiovascular prognosis by treatment of co-morbid depression, however, did not affect vascular outcome and had only modest effects on depression (Glassman et al. Reference Glassman, O'Connor, Califf, Swedberg, Schwartz, Bigger, Krishnan, van Zyl, Swenson, Finkel, Landau, Shapiro, Pepine, Mardekian, Harrison, Barton and McIvor2002; Lesperance et al. Reference Lesperance, Frasure-Smith, Koszycki, Laliberte, van Zyl, Baker, Swenson, Ghatavi, Abramson, Dorian and Guertin2007; van Melle et al. Reference van Melle, de Jonge, Honig, Schene, Kuyper, Crijns, Schins, Tulner, van den Berg and Ormel2007).
Generalized atherosclerotic disease, the underlying process that contributes to vascular events, can be reliably measured by the intima–media thickness (IMT) of the carotid artery and has been associated consistently with a negative cardiovascular outcome (Bots et al. Reference Bots, Hofman, de Jong and Grobbee1996, Reference Bots, Hoes, Koudstaal, Hofman and Grobbee1997). Epidemiological studies have shown up to a twofold increased risk for generalized atherosclerosis in depressed versus never depressed patients (Jones et al. Reference Jones, Bromberger, Sutton-Tyrrell and Matthews2003; Tiemeier et al. Reference Tiemeier, Breteler, van Popele, Hofman and Witteman2003, Reference Tiemeier, van Dijck, Hofman, Witteman, Stijnen and Breteler2004; Elovainio et al. Reference Elovainio, Keltikangas-Jarvinen, Kivimaki, Pulkki, Puttonen, Heponiemi, Juonala, Viikari and Raitakari2005; Chen et al. Reference Chen, Chen, Kuo, Chiang, Ko and Lin2006; Pizzi et al. Reference Pizzi, Manzoli, Mancini and Costa2008; Spitzer et al. Reference Spitzer, Volzke, Barnow, Krohn, Wallaschofski, Ludemann, John, Freyberger, Kerner and Grabe2008).
An important aspect in examining co-morbidity between depression and CAD is the definition of depression. According to DSM-IV-TR, depressive disorder is a syndrome including both cognitive and somatic symptoms (de Jonge et al. Reference de Jonge, Ormel, van den Brink, van Melle, Spijkerman, Kuijper, van Veldhuisen, van den Berg, Honig, Crijns and Schene2006a). These symptoms are reflected in most self-report questionnaires measuring the severity of depression. This may result in inflated associations between CAD and depression. For example, symptoms of fatigue, sleep disturbances and loss of libido can be related to CAD, to depression or to both (Lindgren et al. Reference Lindgren, Fukuoka, Rankin, Cooper, Carroll and Munn2008). Individual symptoms from both the cognitive and somatic symptom cluster have been associated with an increased risk of cardiovascular morbidity (Pedersen et al. Reference Pedersen, Denollet, Daemen, van de Sande, de Jaegere, Serruys, Erdman and van Domburg2007; Davidson et al. Reference Davidson, Burg, Kronish, Shimbo, Dettenborn, Mehran, Vorchheimer, Clemow, Schwartz, Lesperance and Rieckmann2010; Doyle et al. Reference Doyle, Conroy, McGee and Delaney2010), although these results remain to be confirmed and contradicting results have been found for fatigue and anhedonia (Davidson et al. Reference Davidson, Burg, Kronish, Shimbo, Dettenborn, Mehran, Vorchheimer, Clemow, Schwartz, Lesperance and Rieckmann2010; Doyle et al. Reference Doyle, Conroy, McGee and Delaney2010). Of note, more consistent results are found on the symptom cluster level. In depressed cardiac patients, only the somatic-affective symptom cluster of depression, and not the cognitive-affective symptom cluster, was related to subsequent vascular events (de Jonge et al. Reference de Jonge, Ormel, van den Brink, van Melle, Spijkerman, Kuijper, van Veldhuisen, van den Berg, Honig, Crijns and Schene2006; Martens et al. Reference Martens, Hoen, Mittelhaeuser, de Jonge and Denollet2010).
The current study was conducted to examine the relationship between depressive symptom clusters and atherosclerosis in a community-based sample of people aged between 50 and 70 years. We hypothesized that the association between depression and atherosclerosis would be driven primarily by the somatic-affective symptom cluster within the depressive syndrome and that this association would not be affected by the clinical vascular disease status, i.e. the presence of CAD.
Method
Study population
A total of 1517 subjects from the Nijmegen Biomedical Study (NBS) were included. The NBS is a population-based survey described in detail elsewhere (Hoogendoorn et al. Reference Hoogendoorn, Hermus, de Vegt, Ross, Verbeek, Kiemeney, Swinkels, Sweep and den Heijer2006). We invited participants, aged 50–70 years, to the hospital to participate in a detailed assessment of atherosclerotic disease, its risk factors and consequences (Holewijn et al. Reference Holewijn, den Heijer, Swinkels, Stalenhoef and de Graaf2009). Exclusion criteria were a diagnosis of dementia or a history of stroke, as these conditions might directly affect the neurobiological brain circuits involved in depression (Alexopoulos, Reference Alexopoulos2005). The Medical Ethics Committee of the Radboud University Nijmegen Medical Centre approved the study protocol (which is in accordance with the Declaration of Helsinki), and all participants provided written informed consent.
Primary outcome measures
Carotid IMT
Carotid IMT was determined using semi-automatic edge-detection software (M'AthSTD, version 2.0, Metris, France). IMT was defined as the mean IMT of four measured segments of the distal common carotid artery: the far wall left, near wall left, far wall right and near wall right. Longitudinal images of the most distal 10 mm of both the far wall and the near wall of both common carotid arteries were obtained in the optimal projection (anterolateral, lateral or posterolateral). IMT was measured in an area free of plaque, which was defined as an area with an IMT ⩾1.5 times the surrounding IMT. All measurements were carried out in end-diastole using the R-wave of a simultaneously recorded electrocardiogram (ECG) as a reference frame. From each frame the mean IMT was calculated over at least 7.5 mm of the above-mentioned 10 mm segment (yielding a quality index of at least 75%). The outcome variable was defined as the mean IMT of the near and far wall of both common carotid arteries. IMT measurements were performed after an overnight fast or in the afternoon 6 h after a standardized breakfast. Participants were asked to abstain from caffeinated products for at least 12 h and not to smoke for 12 h before the visit. After standardizing the measurement conditions, there were no significant differences between the measurements performed in the morning and those performed in the afternoon (details available in ter Avest et al. Reference ter Avest, Holewijn, Stalenhoef and de Graaf2005). Participants were measured in the supine position in a temperature-controlled (23–24°C) room. The equipment used was a 7.5 MHz transducer of an AU5 ultrasound system (Esaote Biomedical, Italy), connected to a computer with a data acquisition board. All measurements were highly standardized and performed by well-trained and certified sonographers. The reproducibility of our IMT measurements has been reported previously (Holewijn et al. Reference Holewijn, den Heijer, Swinkels, Stalenhoef and de Graaf2009).
Depressive symptoms
Depressive symptoms were measured by the Beck Depression Inventory (BDI; Beck & Steer, Reference Beck and Steer1984), which is a 21-item self-report questionnaire with excellent psychometric characteristics. Each item is rated on a scale of 0–3, with 0 representing ‘absence’ and 1–3 representing increasing levels of severity of the symptom. The BDI yields a total score ranging from 0 to 63. Information on history of depression was collected on the basis of an interview.
Principal components analysis (PCA) with oblimin rotation was conducted on the 21 individual BDI items to obtain fewer factors/components while retaining the original item information. We selected PCA rather than factor analysis for two reasons: (1) its ultimate aim is to reduce data to components useful for other purposes (in this case to examine associations of IMT and aggregate types of depressive symptoms rather than individual depressive symptoms), as opposed to the primary aim of factor analysis, which is to reveal underlying variables that cause manifest variables to co-vary; and (2) its superior ability to remedy multi-collinearity among factors. Factor scores were calculated on unstandardized item factor loadings and transformed into standardized z scores (using the Anderson–Rubin method) to increase their interpretability (Costello & Osborne, Reference Costello and Osborne2009).
Three criteria were used to select the best overall solution, namely factors with eigenvalues >1, the scree plot of eigenvalues and the number of complex items. Although five factors emerged with an eigenvalue >1, the combination of the above-mentioned criteria indicated that a two-factor solution was the optimal solution. The PCA specified to extract only two factors revealed a solution comparable to the traditional two-factor structure of the BDI (i.e. cognitive-affective versus somatic-affective symptoms); see Table 1 for comparison [Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy: 0.898; Bartlett's test of sphericity: χ2=5074, df=210, p<0.001, explained variance factor 1: 24.3%; factor 2: 7.6%].
Table 1. Factor loadings of depressive symptom dimensions and relationship to Beck Depression Inventory (BDI) items and previous dimensional constructs
a Rotate component. Extraction method: principal component analysis. Rotation method: oblimin with Kaiser normalization.
Assessment of covariates
Variables that were examined as potential confounders were age, sex, metabolic syndrome (MS), smoking status, physical activity, alcohol use, cardiovascular medication and cardiovascular disease status. Individual components of the MS were measured; mean arterial pressure was measured using an oscillometric sphygmanometer (Criticon model no. 1846, Criticon Inc., USA). Waist circumference was measured at the level of the umbilicus. Triglycerides and glucose concentrations were determined using commercially available enzymatic reagents (AEROSET® System, Abbott, USA). MS was defined according to the International Diabetes Federation (IDF, 2009). Smoking status, alcohol use, physical activity and cardiovascular disease status were assessed during a short interview. Smoking behaviour was categorized as current, former or never. Physical activity was based on the number of exercise sessions per week of more than 30 min moderate to vigorous activity (Stampfer et al. Reference Stampfer, Hu, Manson, Rimm and Willett2000). Because of a skewed distribution (that could not be normalized by transformation), a median split was used (0 or 1 v. 2 or more exercise moments/week). CAD was assessed by a trained interviewer and defined as a history of treated angina pectoris, myocardial infarction, a history of percutaneous transluminal coronary angioplasty or coronary artery bypass grafting (Kriegsman et. al. Reference Kriegsman, Penninx, van Eijk, Boeke and Deeg1996). Alcohol intake was dichotomized into severe use (>21 for males and >14 for females) and non-severe use.
Medication use was defined by the use of at least half of the defined daily dose and based on brought-along medication containers. Eight different classes of medication were selected based on their potential influence on atherosclerotic disease and/or its association with depression and entered as dichotomies (yes/no). These classes were antidepressants (ATC N06AXXX), statins (ATC C10AXXX), angiotensin-converting enzyme (ACE) inhibitors (ATC C09AXXX), angiotensin II antagonists (C09CATC), Ca+-channel blockers (C08CXXX, C08DXXX, C08EXXX), beta-blockers (ATC C07.AXXX), diuretics (C03XXXX), and analgesics (N02BAXX).
Data analysis
In the case of one or two missing items on the BDI we imputed the series mean, which is a reliable procedure in the case of less than 10% missing values (Shrive et al. Reference Shrive, Stuart, Quan and Ghali2006). Only 71 subjects had one or two missing items (percentage missing: 86 out of 26.481 items≈0.3%). Subjects with more than two missing items were excluded. As BDI scores showed a skewed distribution we used the log-transformed BDI scores in all analyses.
IMT values were normally distributed in our population. To evaluate associations of the IMT with BDI scores, we used multiple linear regression analysis with the IMT as the dependent variable. Multiple regression models were conducted, first corrected only for age and sex (block 1) and subsequently also for the individual measurements that make up for the MS [mean arterial blood pressure (mmHg), triglycerides (mmol/l), high density lipoprotein (mmol/l), glucose (mmol/l) and waist circumference (cm)] (block 2) and lifestyle factors such as smoking status (former/current/never), alcohol use (dichotomized), physical activity (dichotomized), history of depressive disorder (dichotomized) and prevalent CAD (block 3); and finally also medication use [antidepressants (yes/no), statins (yes/no), ACE inhibitors (yes/no), angiotensin II antagonists (yes/no), Ca+-channel blockers (yes/no) and beta blockers (yes/no), diuretics (yes/no), (analgesics yes/no), other medication (yes/no)] was entered (block 4).
In a second multiple regression model we replaced the BDI sum score by the factor scores (somatic-affective or cognitive-affective symptoms) as a measure for depression. To exclude the possibility that the main effects are caused by CAD as a stressful vascular event, we subsequently stratified the analyses by the presence of CAD by adding an interaction term between BDI sum score and CAD. All statistical analyses were carried out using SPSS version 16.0 (SPPS Inc., USA). All analyses were tested two-sided; p values <0.05 were considered statistically significant.
Results
Of the 1517 subjects who consented to participation in the study of non-invasive measurements of atherosclerosis, 29 subjects were excluded because of a history of stroke. None of the participants were diagnosed with dementia. In addition, 227 (15.3%) subjects were excluded because of missing data caused by: not responding to the postal questionnaire containing the BDI (n=181); having three or more missing items on the BDI (n=36); or violating the rules for a reliable measurement of atherosclerotic disease or its risk factors (i.e. having smoked before coming to the hospital, n=3; not obeying the pretest fasting protocol, n=2; and not stopping their lipid-lowering medication, n=5).
Subjects with missing data (227/1488, 15.3%) differed from subjects included (n=1261) with respect to mean age (62.1 v. 60.0 years, t 1486=2.49, p=0.013), severe alcohol use (6% v. 12%, t 1486=−2.70, p=0.007), physical activity (30% v. 37%, t 1486=−2.13, p=0.033) and current smoking (25% v. 17%, t 1486=2.82, p=0.005). The IMT, however, did not differ between these groups (0.86 v. 0.84 mm, t 1456=1.53, p=0.13).
The final study population consisted of 1261 persons with a mean age of 61.0 years (s.d.=5.9 years) and 642 (51%) were females. The median score on the BDI was 4 (interquartile range 2–8, range 0–30) and the mean IMT was 0.84 mm (s.d.=0.12 mm, range 0.57–2.54 mm). Table 2 presents the characteristics of all participants stratified by CAD status.
Table 2. Differences between CAD patients and healthy subjects
CAD, Coronary artery disease; IMT, intima–media thickness; s.d., standard deviation; BDI, Beck Depression Inventory; IQR, interquartile range; ACE, angiotensin-converting enzyme.
* Significant difference at a p<0.01 level.
As shown in Table 3, we found a significant correlation between the BDI sum score and the IMT. After we divided the depressive symptoms into a cognitive-affective symptoms cluster and a somatic-affective symptom cluster, the IMT proved to be associated negatively with the former (β=−0.06, p=0.02) and positively with the latter (β=0.11, p<0.001).
Table 3. Association of depressive symptoms with intima–media thickness (IMT), corrected for different sets of covariates
CAD, Coronary artery disease; BDI, Beck Depression Inventory; MS, metabolic syndrome.
a Linear regression analysis adjusted for sex and age.
b Previous model plus glucose, mean arterial pressure, waist circumference, high density lipoprotein and triglycerides. Similar results were obtained when MS was entered as a dichotomous variable (yes/no) rather than its separate components.
c Previous model plus smoking (current/former/never), alcohol consumption (yes/no), physical activity (yes/no), history of depression (yes/no), in case of all participants also prevalent cardiovascular disease (yes/no).
d Previous model plus medication antidepressants (yes/no), statins (yes/no), Ca+-channel blockers (yes/no), angiotensin II antagonists (yes/no), beta blockers (yes/no), angiotensin-converting enzyme (ACE) inhibitors (yes/no), diuretics (yes/no), analgesics (yes/no), other cardiovascular medication (yes/no).
To examine whether the relationship between IMT and depressive symptoms was explained primarily by CAD, we added an interaction term of CAD and depressive symptoms. While controlling for the effects of other covariates, this interaction term was significant in model 1 (β=0.22, p=0.011; model statistics: R 2=0.24; F 26,1234=18.46, p<0.001). However, in model 2, only the interaction between the somatic-affective symptom cluster and CAD remained significant (β=0.15, p<0.001; model statistics: R 2=0.25; F 28,1232=18.42, p<0.001), and not the interaction between the cognitive-affective symptom cluster and CAD (β=−0.023, p=0.38). The somatic-affective symptom cluster was associated significantly with the IMT in patients with and without CAD, whereas the cognitive-affective symptom cluster was associated negatively with IMT in the group without CAD.
Discussion
Main findings
We have shown that the positive association between depressive symptoms and subclinical atherosclerosis, as measured by the IMT, is driven primarily by the somatic-affective symptom cluster of depression. Furthermore, as the association was present in patients with and without CAD, the relationship seems to be specific for atherosclerotic disease independent of vascular events, such as a stressful life event.
Relationship between depressive symptoms and IMT
Previous studies have reported a positive relationship between depressive symptoms and IMT (Jones et al. Reference Jones, Bromberger, Sutton-Tyrrell and Matthews2003; Tiemeier et al. Reference Tiemeier, van Dijck, Hofman, Witteman, Stijnen and Breteler2004; Chen et al. Reference Chen, Chen, Kuo, Chiang, Ko and Lin2006), although negative findings have been reported as well (Jones et al. Reference Jones, Bromberger, Sutton-Tyrrell and Matthews2003; Spitzer et al. Reference Spitzer, Volzke, Barnow, Krohn, Wallaschofski, Ludemann, John, Freyberger, Kerner and Grabe2008). As our effect-size was relatively small, these negative results can be explained by the low patient numbers (Scarisbrick et al. Reference Scarisbrick, Jones and Isackson1993; Spitzer et al. Reference Spitzer, Volzke, Barnow, Krohn, Wallaschofski, Ludemann, John, Freyberger, Kerner and Grabe2008), and also by the lower age band of the population, resulting in significantly less severe atherosclerotic disease (Jones et al. Reference Jones, Bromberger, Sutton-Tyrrell and Matthews2003).
Different hypotheses have been postulated to explain the direction of association between depressive symptoms and atherosclerosis, including lifestyle factors, such as increased smoking and decreased physical activity in depressed patients, and pathophysiological disturbances associated with depression, such as hypercortisolaemia, low-grade inflammation and autonomic arousal (Raison et al. Reference Raison, Capuron and Miller2006; Vreeburg et al. Reference Vreeburg, Hoogendijk, van Pelt, Derijk, Verhagen, van Dyck, Smit, Zitman and Penninx2009; Carney & Freedland, Reference Carney and Freedland2009a). These mechanisms all have a final common pathway by promoting atherosclerotic disease. Two studies have identified a common genetic pathway for CAD and depression (Scherrer et al. Reference Scherrer, Xian, Bucholz, Eisen, Lyons, Goldberg, Tsuang and True2003; Kendler et al. Reference Kendler, Gardner, Fiske and Gatz2009). However, cytokines have been shown to cause depressive symptoms by acting on the hypothalamus (Nijm et al. Reference Nijm, Kristenson, Olsson and Jonasson2007). As atherosclerotic disease is currently regarded as a low-grade inflammatory process (Schuett et al. Reference Schuett, Luchtefeld, Grothusen, Grote and Schieffer2009), depression may thus be an epiphenomenon. Which of these mechanisms prevail may be determined by longitudinal studies with a vascular end-point that include measures of depression in addition to measures of subclinical atherosclerosis such as IMT and inflammatory markers. To date, such studies are lacking.
Depressive symptom clusters and IMT
We have shown that the somatic symptoms, but not the cognitive symptoms, of depression contribute to the positive relationship between depression and the IMT. This finding suggests that depression might be an epiphenomenon of atherosclerotic disease; symptoms originating from subclinical atherosclerotic disease inflate the depressive symptom score resulting in too broad a definition of depression (see Fig. 1). In other words, subjects with symptoms originating from atherosclerotic disease will score higher on the somatic-affective symptoms of the BDI, in which case only a small additional score on cognitive items is needed to meet the current definition of clinical depression. This hypothesis is supported by studies that have demonstrated that somatic-affective, but not cognitive-affective, symptoms of depression predict cardiovascular morbidity and mortality (de Jonge et al. Reference de Jonge, Ormel, van den Brink, van Melle, Spijkerman, Kuijper, van Veldhuisen, van den Berg, Honig, Crijns and Schene2006; Linke et al. Reference Linke, Rutledge, Johnson, Vaccarino, Bittner, Cornell, Eteiba, Sheps, Krantz, Parashar and Bairey Merz2009). Furthermore, cluster analyses of symptoms of cardiovascular disease in elderly patients with ischaemic heart disease yielded a cluster with patients particularly suffering from fatigue and sleep disturbances, and, as expected, the highest depressive symptom scores were found in this patient cluster (Lindgren et al. Reference Lindgren, Fukuoka, Rankin, Cooper, Carroll and Munn2008).
Fig. 1. Paper hypothesis.
The inverse association between IMT and the cognitive cluster seems puzzling at a first glance. In our opinion, it is most probably explained by an artefact due to a combination of low scores and negative factor loadings on the cognitive items. Moreover, previous research has shown that single cognitive symptoms are associated with an increased risk for cardiovascular morbidity (Pedersen et al. Reference Pedersen, Denollet, Daemen, van de Sande, de Jaegere, Serruys, Erdman and van Domburg2007; Davidson et al. Reference Davidson, Burg, Kronish, Shimbo, Dettenborn, Mehran, Vorchheimer, Clemow, Schwartz, Lesperance and Rieckmann2010) and no studies have shown a decreased risk for cardiovascular disease in subjects with high cognitive scores (de Jonge et al. Reference de Jonge, Ormel, van den Brink, van Melle, Spijkerman, Kuijper, van Veldhuisen, van den Berg, Honig, Crijns and Schene2006).
Of note, within a group of patients suffering from depression, higher IMT values were associated with a later age of onset (Smith et al. Reference Smith, Blumenthal, Babyak, Doraiswamy, Hinderliter, Hoffman, Waugh and Sherwood2009). The effect of atherosclerosis on the development of late-onset depression was, in that study, attributed to cerebrovascular damage, whereas it may also be explained by overlapping symptoms between atherosclerosis and (somatic-affective symptoms of) depression.
The association between the somatic-affective factor and IMT is further extended by the fact that we also found a positive association in subjects without a history of CAD, although it was significantly weaker than in subjects with CAD. Two explanations can be put forward. First, CAD may lead to a cardiotoxic process accompanied by high depressive symptoms scores (de Jonge, Reference de Jonge2009). As this process is thought to be time limited and the majority of our population suffer from chronic CAD, this is only a partial explanation. Second, the (very) small IMT values in non-CAD participants may hardly give rise to (somatic) symptoms. This explanation, however, suggests a non-linear relationship between increasing atherosclerosis and depressive symptoms (which is also indicated by the significant interaction term in our analyses). Nevertheless, the significant association in the group without CAD implies that the association between depression and CAD is driven not only by psychological stress associated with a life-threatening event but also by an acute biological process associated with a myocardial infarction. Because of this low level of somatic-affective symptoms and the smaller patient numbers, the overall BDI sum score was no longer related to the IMT. This has also been reported in a study of middle-aged women with low IMT values (Jones et al. Reference Jones, Bromberger, Sutton-Tyrrell and Matthews2003).
Methodological considerations
The strengths of the present study are the large sample size and reliable measures of depressive symptom clusters and subclinical atherosclerotic disease in middle-aged and older people at risk of cerebrovascular disease but prior to clinical stroke. For proper interpretation of the results, some limitations should be addressed. First, because vascular depression is known to have a specific somatic-affective symptom profile (Naarding et al. Reference Naarding, Tiemeier, Breteler, Schoevers, Jonker, Koudstaal and Beekman2007), this type of depression might influence our findings. For this reason we excluded all evident cases of cerebrovascular disease by excluding subjects with a history of stroke (Krishnan et al. Reference Krishnan, Taylor, McQuoid, MacFall, Payne, Provenzale and Steffens2004; Steffens, Reference Steffens2004). Nevertheless, many researchers consider magnetic resonance imaging (MRI) of the cerebrum as the gold standard to identify vascular depression. Second, the cross-sectional design of this study does not allow our results to be interpreted causally. To date, no prospective studies are available relating depressive symptom clusters to subclinical vascular disease. Third, we used a population sample, so bias towards the most healthy people may have occurred, resulting in less (advanced) atherosclerosis and depressive symptoms. This may have resulted in less overall variance of the atherosclerosis and depression measures and also lowered the explained variance of the factors found.
Conclusions
In addition to replicating a significant association between depressive symptoms and subclinical atherosclerotic disease, we are the first to show that this correlation could be explained by the somatic-affective symptom cluster within the depression symptomatology. This has major conceptual and clinical implications. From a conceptual point of view, subclinical atherosclerotic disease may inflate depressive symptom scores, possibly with large effects in CAD patients. This could argue for adaptations of the criteria for depressive disorder in patients with atherosclerosis, with less emphasis on somatic-affective symptoms parallel to the degree of atherosclerotic disease. From a clinical point of view, these results may suggest misdiagnosis of depression in some cardiac patients due to the presence of somatic-affective symptoms that reflect the severity of the atherosclerotic disease but are not part of a formal depressive syndrome. This hypothesis is supported by previous findings showing a worse cardiac outcome in CAD patients with treatment-resistant depression only (de Jonge et al. Reference de Jonge, van den Brink, Spijkerman and Ormel2006b; Carney & Freedland, Reference Carney and Freedland2009b), as those somatic-affective symptoms due to atherosclerosis may not be affected by antidepressants but are related to cardiac outcome.
Acknowledgements
This paper was supported by an NWO (Dutch Scientific Organization) Clinical Fellowship (grant no. 40-00703-97-231) awarded to R. C. Oude Voshaar.
Declaration of Interest
None.
