Police work is considered a stressful occupation which not only involves danger and traumatic event exposure, but also organizational stressors such as lack of support, punishment-centered executive philosophies, and excessive paperwork (Violanti et al. 2006; Berg et al. 2005; Kop and Euwema 2001; Patterson 2003; Spielberger et al. 1982; Violanti and Aron 1994).

Depression may be one outcome of exposure to police work stressors and it has been associated with increased risk of cardiovascular disease (CVD) (Wang et al. 2010; Chen et al. 2006; Whipple et al. 2011). Barefoot and Schroll (1996) found that a high level of depressive symptoms predicted the subsequent occurrence of myocardial infarction and mortality in a 21-year follow-up study. Appels and Mulder (1988) found relationships between various negative psychological states and the occurrence of coronary heart disease (CHD). Depression has also been associated with biological outcomes which exacerbate the risk of CVD, including hyperactivity within the hypothalamic–pituitary–adrenal (HPA) axis, diminished heart rate variability, and ventricular instability (Musselman et al. 1998a, b).

Due to their occupational exposure and increased risk for depression, police officers may also be at higher risk for CVD (Violanti et al. 1998; Franke and Anderson 1994; Franke et al. 1998). Based on the previous research, we tested the hypothesis that depressive symptoms in police officers are associated with carotid artery intima-media thickness (CIMT) after adjustment for traditional CVD risk factors. Stewart et al. (2007) found that higher depressive symptoms at baseline were associated with greater 3-year change in carotid intima-media thickness even after taking into account demographic factors, cardiovascular risk factors, medication use, medical conditions, and other correlated negative emotions. Pizzi et al. (2010) found that IMT is higher in depressed subjects, indicating that atherosclerosis is accelerated in depressed patients. These authors contribute this to mechanisms which connect depression and coronary artery disease, such as inflammation and imbalance of the autonomic nervous system. Haas et al. (2005), adjusting for baseline cardiovascular risk factors, found participants with elevated depression scores at baseline were > 2 times as likely as those with no depressive symptoms to have carotid plaque. This study suggested that hypothalamic–pituitary–adrenal (HPA) dysregulation, diminished heart rate variability, altered blood platelet function, and noncompliance with medical treatments underlie the association between depression and cardiovascular disease. Conversely, Rice et al. (2009) in the Baltimore Longitudinal Study of Aging found no significant relationship between the trajectory of depressive symptoms and future carotid IMT.

CIMT progression rates have been associated with risk factors such as diabetes, smoking, hypercholesterolemia, and hypertension (Chambless et al. 2002; Ferrieres et al. 1999). CIMT has been used successfully to study arterial atherosclerotic disease in predicting cardiovascular disease outcomes (myocardial infarction, stroke), clinical coronary outcomes, determining the effectiveness of risk factor intervention, and in monitoring the progression of arterial wall thickness over time (Bots et al. 2003; Chambless et al. 1997; Grobbee and Bots 1994) In a meta-analysis of population-based CIMT studies, Lorenz et al. (2007) estimated the relative risk of myocardial infarction and stroke as 1.15 and 1.18, respectively (adjusted for age and gender), for each 0.10-mm thickness increase in common carotid IMT.

The mean age (±SD) of all officers (n = 412) was 41.0 ± 7.1 years (Table 1). The age ranges for men and women were 21–66 and 26–53 years, respectively. The majority of officers were Caucasian (76.8 %), 16.7 % were current smokers, and 19.9 % reported 8 drinks or more per week. Approximately 24 % of all officers had a history of hypertension and 2.7 % a history of diabetes. Women had slightly higher CES-D scores than men (8.7 ± 8.2 vs. 7.4 ± 6.6). In contrast, men had slightly higher mean CCA IMT (0.626 mm vs. 0.587 mm) and MMXIMT (0.801 mm vs. 0.729 mm) values than women.

Results for the association between demographic, lifestyle characteristics, and CES-D scores are presented in Table 2. Low-density lipoprotein (LDL) cholesterol and triglycerides generally increased with increasing quintiles of CES-D scores but the associations were not statistically significant. Race/ethnicity was associated with CES-D scores, P = 0.029. Use of antidepressant medications was also associated with CES-D scores, P = 0.004. Table 3 provides associations between selected demographics, traditional CVD risk factors, CCA IMT, and MMXIMT. Most results were in the expected direction. LDL cholesterol, glucose levels, diabetes, and hypertension were positively and significantly correlated with both CIMT measures, whereas high-density lipoprotein (HDL) was inversely and significantly associated with both. Mean levels of CCA IMT and MMXIMT differed significantly by race/ethnicity, with Caucasians having the lowest mean levels and Hispanic-American officers having the highest mean levels of both CCA IMT (P = 0.007) and MMXIMT (P = 0.007).

After adjustment for age, gender, race, education level, cigarette smoking status, alcohol intake, body mass index (BMI), HDL cholesterol, LDL cholesterol, triglycerides, glucose, and hypertension, officers with CES-D scores >16 had a higher mean CCA IMT value (0.628 ± 0.013) than officers with lower CES-D scores, but the difference was not statistically significant (P = 0.130) (data not shown). Similar results were observed with MMXIMT. The results in Table 4 show no significant association for CES-D scores with CCA IMT or MMXIMT, before or after adjustment for covariates.

To help clarify associations between depressive symptoms, CVD risk factors, and CIMT, we stratified by traditional CVD risk factors based on the cut-points recommended by the National Cholesterol Education Program Adult Treatment Panel III (NCEP ATP III), the American Heart Association and the National Heart, Lung, and Blood Institute, National Institutes of Health. With the exception of hypertension (SPB ≥ 140 mm Hg; DBP ≥ 90 mm Hg or medication use), no significant associations were observed between depressive symptoms and CIMT after stratification.

Hypertension status significantly modified the association between CES-D and CCA IMT (interaction P = 0.088) and between CES-D and MMXIMT (interaction P = 0.083). Table 5 displays mean values of CCA IMT and MMXIMT across quintiles of CES-D score, stratified by hypertension status. Among officers without hypertension, depressive symptoms were positively and significantly associated with CCA IMT in the unadjusted model (P = 0.041). This association remained statistically significant after inclusion of several confounders and CVD risk factors (P = 0.030). The associations between CES-D and MMXIMT were not statistically significant among the officers with or without hypertension.

Although sleep quality did not significantly modify the association between CES-D and CIMT, the associations among officers who reported poor sleep quality versus those who reported good sleep quality were substantially different (though not statistically significant) to warrant mention in this paper. Among officers who reported poor sleep quality, mean values of both CCA IMT and MMXIMT increased as depressive symptoms (CES-D scores) increased; adjusted p values for linear trends were =0.071 and 0.083, respectively. Among officers who reported good sleep quality, the associations of depressive symptoms with CCA IMT and MMXIMT did not show a similar positive relationship nor were they statistically significant.

The goal of this study was to investigate the cross-sectional association of depressive symptoms with subclinical cardiovascular markers CIMT and MMXIMT among police officers. As expected, traditional risk factors remained strong predictors of CIMT. Age, BMI, LDL cholesterol, and glucose levels were positively and significantly correlated with CIMT measures, whereas high-density lipoprotein (HDL) was inversely and significantly associated with both. We stratified by traditional CVD risk factors to reveal whether any of these risk factors might modify the association between depressive symptoms and CIMT. No significant differences were noted in the associations between depressive symptoms and CIMT after stratification with one exception: hypertension. A significant positive association between depressive symptoms and CIMT was observed in officers who were free of hypertension. Additionally, hypertension significantly modified the association between CES-D and CCAIMT. This result indicated that depressive symptoms may be independently associated with CIMT, yet only among participants free of hypertension. Recent work on depression and hypertension has provided evidence of increased sympathetic activity and increased blood pressure reactivity (Jonas and Lando 2000; Jonas et al. 1997; Bowman and Nicklin 1997; Veith et al. 1994; Waked and Jutai 1990). These studies suggest that depression may have an effect on the cardiovascular system that could lead to the development of hypertension and possible future CVD. Our results also showed that among officers who reported poor sleep quality (Table 6), mean levels of CIMT tended to increase as depressive symptoms increased. Even though sleep quality did not significantly modify the main association, our results suggest that poor sleep quality may act synergistically with depressive symptoms to increase CIMT. Violanti et al. (2009) found that police officers working midnight shifts combined with either shorter sleep duration or increased overtime may be at an increased risk for metabolic syndrome. The prevalence of metabolic syndrome among officers working the midnight shift was higher than that found in the National Health and Nutrition Examination Survey (NHANES III) 1988–1994 (Ford et al. 2002). Karlsson et al. (2001) found that obesity, high triglycerides, and low concentrations of HDL cholesterol seem to cluster together more often in shift workers than in day workers. Lasfargues et al. (1996) found night-shift workers have significantly higher levels of triglycerides, smoking, and obesity than controls. In a review of the literature, Wolk and Somers (2007) found that the weight of evidence suggested that sleep deprivation and shift work independently lead to the development of both insulin resistance and individual components of the metabolic syndrome. Other studies found the metabolic effects of shift work to include abdominal obesity, lower HDL cholesterol and higher triglycerides, and changes in glucose intolerance (Nagaya et al. 2002). Mean levels of CCA IMT and MMXIMT differed significantly by race/ethnicity, with Caucasians having the lowest mean levels and Hispanic-American officers having the highest mean levels of both CCA IMT (P = 0.007) and MMXIMT (P = 0.007). The small number of Hispanic Americans in our study prevented any meaningful conclusions concerning CIMT and depressive symptoms. African Americans had significantly greater mean CIMT than did Caucasians in our cohort; however, no significant interactions with depressive symptoms were found. This was not surprising, as African Americans are more likely to have a higher prevalence of CVD than whites and are also 1.5 times as likely as whites to have hypertension and 10 % less likely than whites to have their blood pressure under control (http://www.cdc.gov/nchs/data/series/sr_10/sr10_249.pdf).

In a recent study, higher levels of hopelessness (a diagnostic criterion of depression) were associated with increased CIMT (Whipple et al. 2011). The authors stated that hopelessness confers an independent risk of CVD. For minority officers, there may be social conditions or exposures within police work that increase the risk of both depressive symptoms and hopelessness.

Behaviors related to stress or depression may also increase the risk for CVD in officers. The low physical fitness level found among police officers may be an indicator. Williams (1987) found that a substantial number of officers in their sample were at elevated risk for atherosclerotic heart disease: 76 % had elevated cholesterol, 26 % had elevated triglycerides, and 60 % elevated body fat composition. Price et al. (1978) concluded that middle-aged police officers had CVD risk above that of the general population. Joseph et al. (2009) found that police officers have increased levels of atherosclerosis compared with a general population sample, which was not fully explained by elevated CVD risk factors, thereby potentially implicating other mechanisms whereby law enforcement work may increase CVD risk. Franke and Anderson (1994) found that public safety officers had a higher probability of developing CHD than did the Framingham Heart Study population. Steinhardt et al. (1991) found an inverse association between cardiovascular fitness and medical claims among police officers. Of interest is the fact that police officers either suffer from disease or die at a much earlier age than do reference groups such as municipal workers or the general US population (Violanti et al. 1998; Franke et al. 1998). Officers also died from CVD at earlier ages than the general population (Violanti et al. 1998). This suggests the possible influence of lifestyle risk factors for those diseases, including psychological stress and depression.

This study is cross-sectional and cannot determine the causal direction between depressive symptoms and CIMT. Future prospective investigations may help to clarify this issue. Other limitations of this study are the lack of complete recruitment of all members of the police force, residual confounding, and the possibility of selection bias. To clarify whether officers who were examined were representative of all officers, we compared those officers who participated versus those who did not; summary characteristics were as follows: 26 versus 23 % women; 30 versus 28 % above 45 years of age; 66 versus 71 % police officers (vs. higher rank); and 45 versus 39 % hired after 1990, respectively. The results suggest that findings are likely to be generalizable to the entire Buffalo Police Department. The potential for residual confounding in this analysis could remain due to incomplete adjustment, and there could be other unmeasured variables that could not be adjusted for in the analysis. Another limitation of the study is the limited period of time for which information on depressive symptoms is available (i.e., 7 days). Data on the history of depressive symptoms or depressive disorders were not available for this cohort. If the association was causal, it might take years of exposure to depressive symptoms before an adverse effect on carotid IMT would be evident.

This study has the advantage of being conducted in a controlled clinical setting, where the assessment of many CVD risk factors (such as blood pressure or cholesterol) is conducted by trained clinic personnel. Data collection for all studies was performed at the same site, using a standardized protocol during similar time frames by the same research team. As suggested by Joseph et al. (2009), subclinical CIMT is a predictive measure of future CVD disease. The use of subclinical markers prior to CVD event manifestation is beneficial, in that preventive efforts may be initiated well in advance of clinically apparent disease.

In sum, the present study suggests that, other than when stratified by hypertension, no significant associations were observed for depressive symptoms with CCA IMT and MMXIMT, after adjustment for demographics and traditional CVD risk factors. However, among officers without hypertension, mean levels of CCA IMT increased with increasing quintiles of CES-D scores. This association was statistically significant after adjustment for several confounders and CVD risk factors (BMI, HDL cholesterol, LDL cholesterol, triglycerides, and serum glucose levels). This result may reflect difficulty detecting an association of depressive symptoms with CIMT in individuals with hypertension due to the strong influence of hypertension on CIMT, whereas such an association may be easier to detect in those free of hypertension. If confirmed in other studies and in a prospective design, efforts to reduce depressive symptoms and to prevent its occurrence might prevent increased CIMT. There are some studies which do provide directional evidence that depressive symptoms may have an adverse effect on the cardiovascular system that could lead to the development of hypertension and CVD (Waked and Jutai 1990; Julius 1988). Future prospective work is planned with the present police cohort to help clarify such possible directional relationships between depressive symptoms and CVD. It will also be necessary to examine not only depressive symptoms, but other variables associated with police work such as hopelessness and behavioral lifestyles that may further exacerbate earlier signs of cardiovascular disease.