In the 2011 “Expert Panel on Integrated Guidelines for Cardiovascular Health and Risk Reduction in Children and Adolescents,” several medical conditions among youth were identified that predispose to accelerated atherosclerosis and early cardiovascular disease (CVD), and risk stratification and management strategies for youth with these conditions were elaborated. Major depressive disorder (MDD) and bipolar disorder (BD) among youth satisfy the criteria set for, and therefore merit inclusion among, Expert Panel tier II moderate-risk conditions. The combined prevalence of MDD and BD among adolescents in the United States is ≈10%, at least 10 times greater than the prevalence of the existing moderate-risk conditions combined. The high prevalence of MDD and BD underscores the importance of positioning these diseases alongside other pediatric diseases previously identified as moderate risk for CVD. The overall objective of this statement is to increase awareness and recognition of MDD and BD among youth as moderate-risk conditions for early CVD. To achieve this objective, the primary specific aims of this statement are to (1) summarize evidence that MDD and BD are tier II moderate-risk conditions associated with accelerated atherosclerosis and early CVD and (2) position MDD and BD as tier II moderate-risk conditions that require the application of risk stratification and management strategies in accordance with Expert Panel recommendations. In this scientific statement, there is an integration of the various factors that putatively underlie the association of MDD and BD with CVD, including pathophysiological mechanisms, traditional CVD risk factors, behavioral and environmental factors, and psychiatric medications.
- AHA Scientific Statements
- bipolar disorder
- cardiovascular diseases
- coronary artery disease
- depressive disorder, major
- population at risk
In a 2006 American Heart Association scientific statement on cardiovascular risk reduction in high-risk pediatric patients, a list of 8 pediatric diagnoses associated with elevated cardiovascular risk were identified, and practical management recommendations were generated.1 In that scientific statement, tier II conditions required “pathophysiological evidence for arterial dysfunction indicative of accelerated atherosclerosis before 30 years of age.” In the 2011 “Expert Panel on Integrated Guidelines for Cardiovascular Health and Risk Reduction in Children and Adolescents” (henceforth, “Expert Panel”), special risk conditions were identified in accordance with the above definition for tier II (moderate risk). Specifically, moderate-risk conditions are those for which “the disease process has been shown to be associated with pathologic, physiologic, or subclinical evidence of accelerated atherosclerosis.”2 According to the Expert Panel, tier II included Kawasaki disease with regressed coronary aneurysms, chronic inflammatory disease (systemic lupus erythematosus, juvenile inflammatory arthritis), HIV infection, and nephrotic syndrome. Moreover, risk stratification in the Expert Panel indicated that children and adolescents with tier II conditions were to be moved to tier I (high risk) if they had 2 or more of 7 traditional cardiovascular risk factors or comorbidities (obesity, tobacco smoke exposure, hypertension, insulin resistance, dyslipidemia including high levels of low-density lipoprotein cholesterol, high levels of triglycerides, and low levels of high-density lipoproteins). Importantly, in the Expert Panel report, specific cut points and treatment goals were outlined for blood pressure, body mass index, glucose, and lipids, and recommendations were provided regarding lifestyle change and pharmacological treatment.
When the results of several studies to date are considered together, as reviewed in this statement, it can be concluded that the inclusion of adolescent mood disorders on the list of tier II moderate-risk pediatric diagnoses is warranted. Moreover, there is compelling evidence regarding excessive and premature cardiovascular disease (CVD)* among adults with major depressive disorder (MDD) and bipolar disorder (BD). The association between depression and CVD among adults is well known.3,4 In recent epidemiological studies in the United States, the prevalence of CVD among adults with MDD was nearly 3-fold greater than among adults without mood disorders, and adults with CVD and MDD were ≈7.5 years younger than adults with CVD who did not have mood disorders.5 Putative pathophysiological mechanisms of the increased risk of CVD among adults with mood disorders include hypothalamic-pituitary-adrenal axis and sympathomedullary hyperactivity, increased platelet reactivity, reduced heart rate variability, vascular inflammation, oxidative stress, and endothelial dysfunction (Figure 1). These processes are likely set into motion in part by adverse lifestyle behaviors that are disproportionate among people with mood disorders. The results of twin studies and molecular genetic studies also implicate shared genetic pathways between depression and CVD.6,7 There is also preliminary evidence based on adolescents regarding the familial nature of the CVD-depression link. For example, psychiatrically healthy adolescent offspring of parents with MDD have been shown to have increased aortic stiffness and blood pressure, along with decreased insulin sensitivity.8 Moreover, adolescents with a history of MDD have elevated rates of parental CVD.9
The association between BD and CVD appears to be at least as strong as the association with MDD, although this association is less commonly recognized. CVD is the leading cause of death in BD, with a standardized mortality ratio of 1.5 to 2.5.10,11 Among young adults specifically, the standardized mortality ratio may be as high as 8.12 By comparison, although malignancy is the second most common medical cause of death in BD, it does not appear to be more prevalent in BD than in the general population.10,13 Excess and premature CVD mortality in BD has been documented for >70 years, before the use of mood stabilizers and antipsychotic drugs.11,14 After control for multiple potential confounders, including medications, symptomatic severity in BD was independently associated with CVD mortality.15 Adults with BD in the US population have a 5-fold increased risk of CVD and manifest CVD 14 years earlier than adults without mood disorders, despite the fact that approximately half have never received any pharmacological treatment for BD and approximately three-fourths have not received antimanic medication.5,16,17 The prevalence of CVD among adults with BD is nearly 2-fold greater than among adults with MDD.5 On the basis of recent genome-wide association studies, calcium channels are implicated in the pathophysiology of BD,18,19 and these channels also appear to be salient to CVD.20,21
MDD and BD are the first and fourth most disabling conditions, respectively, among adolescents worldwide.22 BD involves repeated episodes of mania/hypomania (elated or irritable mood together with other symptoms) that generally, but not always, alternate with episodes of depression (sadness or lack of interest/pleasure along with other symptoms). MDD involves episodes of depression without mania/hypomania.23 The diagnostic criteria for mood episodes in MDD and BD are summarized in Table 1; more detailed descriptions and guidelines are available elsewhere.24–27 In the United States, the prevalence of mood disorders among adolescents is ≈10% (8.7% for MDD, 2.6% for BD).28 By comparison, the prevalence of the 4 existing Expert Panel tier II moderate-risk conditions ranges from ≈0.5% (chronic inflammatory disease) to <0.05% (HIV, Kawasaki disease, nephrotic syndrome). In other words, MDD and BD together are at least 10 times more prevalent than the 4 previously identified moderate-risk conditions combined. Because mood disorders are highly prevalent among adolescents and are generally amenable to treatment, there could be substantial cardiovascular benefits associated with improved identification, monitoring, and treatment of these conditions, with potential public health implications. Importantly, mood disorders that begin in childhood or adolescence are known to persist into adulthood and are known to be especially pernicious variants of these disorders, with a far greater burden of psychiatric symptoms than adult-onset mood disorders.29–36 This increased severity of child- and adolescent-onset mood disorders has been observed in clinical samples30,31,34 and representative epidemiological samples.29,32 The high prevalence of mood disorders in adolescents, the biological plausibility of an association between mood disorders and CVD, and the strength of association between mood disorders and CVD justify the need for a scientific statement to address CVD risk among adolescents with mood disorders.
The overall objective of the current statement is to increase awareness and recognition of mood disorders among young people as being moderate-risk conditions for early CVD. To achieve this objective, the primary specific aims of this statement are to (1) summarize evidence that mood disorders (MDD and BD) are Expert Panel tier II moderate-risk conditions associated with accelerated atherosclerosis and early CVD and (2) position MDD and BD as tier II moderate-risk conditions that require the application of risk stratification and management strategies in accordance with Expert Panel recommendations. In addition, in this statement, the contribution of traditional cardiovascular risk factors (eg, diabetes mellitus, obesity) and lifestyle behaviors (eg, sedentary lifestyle, tobacco smoking) to excessive CVD risk among adolescents and adults with mood disorders is examined. Finally, this statement summarizes the evidence regarding the association of pharmacological treatments commonly used in the treatment of adolescents and young adults with mood disorders (particularly second-generation antipsychotic drugs) with CVD and CVD risk factors. The success of this statement will rely on the extent to which it results in improved CVD risk factor screening, prevention, and intervention among adolescents and young adults with mood disorders and the extent to which it spurs increased collaboration between preventive cardiology, pediatrics, and psychiatry, together with consumers and other stakeholders, in achieving these goals.
Pathophysiological, Subclinical, and Clinical Evidence of Accelerated Atherosclerosis Among Adolescents and Young Adults With Mood Disorders
The articles reviewed in this section were identified with MEDLINE searches using the following medical subject headings (MeSH) terms and keywords: major depressive disorder or bipolar disorder, each cross-referenced with cardiovascular diseases or cardiovascular or heart disease or atherosclerosis or coronary disease or coronary vessels or coronary or endothelium or endothelial or arterial stiffness or vascular stiffness or arterial pressure, or flow mediated dilation, or reactive hyperemia. Articles were selected for inclusion if either all of the participants or a substantial proportion of participants were <30 years of age.
Evidence of Premature Cardiovascular Mortality
Two observational, population-based studies have linked MDD, attempted suicide, BD, and anxiety in children and young adults to increased risk of premature CVD or related deaths. The National Health and Nutritional Examination Survey included 7641 participants 17 to 39 years of age during 1988 to 1994 and performed follow-up of this cohort until 2006 (Table 2).37 After controlling for Framingham scores, diagnoses of clinical depression, as opposed to high levels of mood symptoms on a checklist, were associated with increased ischemic heart disease (IHD) mortality. The adjusted hazard ratio for IHD mortality was 3.70 (95% confidence interval [CI] 1.32–10.35) for depression (associated with an MDD or BD diagnosis) and 7.12 (95% CI 2.67–18.98) for history of attempted suicide; associations were stronger among women. Further controlling for education, income, body mass index, alcohol intake, and sedentary lifestyle did not affect these findings. The associations were stronger for IHD than for CVD (including endocarditis, myocarditis, heart failure, and cerebrovascular disease). The risk of IHD mortality attributable to MDD or attempted suicide or suicide was 13% among males (fourth strongest risk factor after obesity, smoking, and hypertension) and 65% among females (strongest risk factor).
In a Taiwanese national health insurance study involving >1 million participants of all age groups, including young adults, the association of MDD, BD, and anxiety disorder diagnoses with risk of IHD was examined. The relative risk for IHD among patients <20 years old was 2.19 for MDD, 2.11 for BD, and 9.88 for anxiety.38 Although the prevalence of IHD increased with increasing age in all diagnostic groups, the excessive relative risk for IHD among patients with MDD, BD, and anxiety was greatest among those <20 years old. Although these findings do not demonstrate causality, they are based on large samples and robust covariate modeling and provide a basis for hypothesizing that the spectrum of BD, MDD, suicide attempts, and even anxiety in adolescents and young adults may confer increased risk of subsequent CVD and cardiovascular mortality.
Increased Carotid Artery Intima-Media Thickness
There is evidence in epidemiological studies of a link between depressive symptoms (even when they are not severe or disabling enough to merit a full clinical diagnosis) and premature vascular aging as measured by carotid artery intima-media thickness (CIMT). Findings from the Cardiovascular Risk in Young Finns study, a population-based longitudinal follow-up study, linked early adult depressive symptoms to increased CIMT in 410 young men.39 Depressive symptoms in earlier years among men did not predict CIMT after controlling for depressive symptoms concurrent with CIMT assessment. There was no significant association between depressive symptoms and CIMT in women. In a cross-sectional study of 157 black and white 16- to 21-year-olds, depressive symptoms were not correlated with right and left common CIMT.40 However, higher levels of depressive symptoms were associated with higher pulse-wave velocity, indicative of increased vascular stiffness, independent of covariates (including body mass index, systolic blood pressure, and social class; lipid profiles were not measured). These studies together indicate a plausibly adverse but as yet inconsistent relationship between depressive symptoms and CIMT among adolescents and young adults. Importantly, all of these studies focused on self-reported depressive symptoms. This approach is likely to capture both patients with clinically significant depression and those who are transiently distressed or do not have the degree of functional impairment needed to merit a diagnosis.
One of the earliest signs of risk for CVD is impaired endothelial function, measured by ultrasound-determined brachial artery flow-mediated dilation or by digital pulse-wave amplitude, in which higher values are considered better. There are 6 studies in young adults and adolescents with diagnoses of MDD or depressive symptoms in relation to endothelial function. In 15 young adults with MDD and 15 healthy control subjects, percent change in brachial artery diameter was smaller among those with MDD.41 Participants with MDD also had elevated monocyte chemotactic protein-1, soluble intercellular adhesion molecule-1, and E-selectin, markers of risk for atherosclerosis. No study participants were taking antidepressant medications or had elevated cardiovascular risk factors. By contrast, in another study of 50 untreated depressed young adults matched to 50 healthy control subjects, there were no group differences in flow-mediated dilation42; however, plasma nitrite/nitrate concentrations were lower among the depressed participants, which suggests decreased nitric oxide production. In a small controlled study, participants with BD (n=27), matched with healthy control subjects with regard to tobacco use, age, and sex, did not differ from the control subjects with regard to flow-mediated dilation.43 Among 248 teenagers, symptom scores for depression, anger, and anxiety were associated with lower pulse-wave amplitude scores among girls only.44 In a single longitudinal study, 135 female teenagers at high risk for MDD were followed up over a period of 2.5 years, and pulse-wave amplitude was assessed at 12-month intervals.45 Depressive symptoms were negatively associated with endothelial function at each time period but did not predict endothelial function at later time points. In this sample of adolescents, endothelial dysfunction and depression symptoms travelled together. Taken together, depressive symptoms may be associated with impaired endothelial function in adolescent and young adult females; the association of MDD and BD with endothelial function is not consistent.
There is evidence from several studies of early CVD mortality, CVD, and imaging proxies for atherosclerosis; however, results are mixed. In most studies, depressive symptoms were determined via self-report; however, the strongest evidence emerged from studies in which mood disturbance was ascertained by use of semistructured research interviews that yielded psychiatric diagnoses. Indeed, the study that yielded the strongest data used the most rigorous assessment of mood (ie, diagnostic interviews) and the most valid cardiovascular outcome (ie, CVD mortality).37 In a previous meta-analysis, the relative risk of incident CVD in adults conferred by diagnoses of MDD (2.69; 95% CI 1.63–4.43) was meaningfully stronger than the relative risk conferred by symptoms of depression (1.49; 95% CI 1.16–1.92).46 These differences between diagnoses and symptoms are likely mediated by more severe, persistent, or recurrent elevations in pathophysiological processes discussed in the section “Pathophysiological Processes Bridging Mood Disorders and CVD.” Other methodological factors may also have contributed to the inconsistency of findings, including small sample sizes and model overfitting (insufficient power). Sensitivity and reliability are important considerations when using noninvasive vascular imaging47; however, most studies reviewed in this section were conducted by groups with expertise in these methods. Future studies with larger sample sizes and research-grade assessment of mood symptoms, as well as psychiatric diagnoses, are warranted to resolve some of the inconsistency in these findings.
Traditional Cardiovascular Risk Factors in Relation to MDD and BD
The metabolic syndrome (MetS) is a clustering of clinical and biochemical risk factors for CVD and type 2 diabetes mellitus (T2DM). Several definitions for metabolic syndrome have been proposed. The most often cited are from the National Cholesterol Education Program, Adult Treatment Panel II, World Health Organization, and International Diabetes Federation.48 MetS does not have a consensus definition in pediatric populations, although various definitions define risk according to pediatric cut points, most commonly the International Diabetes Federation definition.49 There is evidence from several studies that the prevalence of MetS among adults with MDD or depressive symptoms is elevated.50 Similarly, an increased hazard for MetS among people with BD has been documented across continents, countries, and cultures.51,52 Each MetS component has been documented to be present at a higher rate than with matched people in the general population, with abdominal obesity being the most replicated finding.
Risk of MetS components among adolescents with BD is a major concern that has been the subject of comparatively little research to date.53–57 No published study could be identified that primarily sought to determine the prevalence of MetS in a pediatric MDD or BD population. Notwithstanding this, data from the South Carolina Medicaid program, covering all medical services and medication prescriptions between January 1996 and December 2005, indicated that children and adolescents with BD are differentially affected by several medical disorders, including but not limited to obesity and T2DM.58 Moreover, these conditions, including obesity and T2DM, were found to precede diagnoses of BD.54 Similarly, research based on medical claims data indicated that children and adolescents diagnosed with BD were more likely to access several medical services, including cardiology services.55 Finnish population-based prospective data indicated that there was a bidirectional association between depressive symptoms and MetS between childhood (mean age, 12 years) and early adulthood (mean age, 33 years).59
Results from longitudinal studies indicated that pediatric obesity increased the risk for MDD and vice versa.60,61 Merikangas et al62 examined the association between MDD and obesity among adolescents 12 to 19 years old in the 2001 to 2004 National Health and Nutritional Examination Survey. MDD was associated with a significantly increased prevalence of obesity among males and among non-Hispanic blacks, whereas in the overall sample, the association of MDD with increased prevalence of obesity was no longer significant after controlling for demographic variables (adjusted odds ratio 1.6, 95% CI 0.9–2.9). There was prior evidence from adults that certain depressive subtypes may confer differential risk of obesity and that these subtypes may have genetic underpinnings63; however, this has yet to be examined in youth.†
There has been cross-sectional and longitudinal evidence of a bidirectional association between BD and overweight/obesity.64 In a large cohort of children and adolescents with BD, 42% of participants were overweight/obese.53 Because the vast majority of this sample had received prior treatment, analyses could not specifically examine untreated participants. As expected, second-generation (or atypical) antipsychotic drugs were significantly associated with overweight/obesity; however, other variables were also significantly and independently associated with overweight/obesity, including nonwhite race, earlier age of BD onset, and history of physical abuse and psychiatric hospitalization. In South Carolina Medicaid data, adolescent BD was associated with an increased risk of obesity (odds ratio 1.92, 95% CI 1.53–2.40).54
Insulin Resistance and Diabetes Mellitus (T2DM)
There is evidence of bidirectional associations between depressive symptoms and T2DM among adults.65 Moreover, among young adult males in 2 Finnish population samples, severity of depressive symptoms was associated with insulin resistance.66,67 In meta-analytic findings, depressive symptoms along with other psychological symptoms were more common among children with versus without T2DM,68 and there was preliminary evidence that the association between depression and T2DM was especially strong in youth.69 There is pharmacoepidemiological evidence of elevated rates of hyperglycemia and T2DM in pediatric BD treated with psychotropic medication, although diagnoses of T2DM frequently precede treatment.54,70
In pharmacoepidemiological studies, the rate of dyslipidemia in younger populations treated for BD is increased relative to the general population. For example, a retrospective cohort study based on a US health insurance organization examined claims for 17 884 adolescents (13–17 years of age) with BD (or schizophrenia) receiving care from 1997 to 2006.71 The incidence rate per 100 000 person-years of dyslipidemia was 346.4 (95% CI 274.9–431.0) in this group compared with 86.6 (95% CI 76.4–97.7) in the general population. The adjusted hazard ratio for developing dyslipidemia for the psychiatric cohort versus the general population was 1.66 (95% CI 1.22–2.28).71 A separate study reported increased prevalence of dyslipidemia (ie, 51% and 48% had elevated triglycerides and low high-density lipoprotein levels, respectively) in a mixed sample (n=95) that included 49 youth with BD.72
Hypertension is significantly more prevalent among adults with MDD and BD than in the general population and occurs at younger ages among adults with MDD and BD.5 No original article was identified that primarily sought to report on the rate of hypertension among youth with MDD or BD. However, South Carolina Medicaid data indicated that hypertension was prospectively associated with an increased risk of incident BD among adolescents.54
Findings to date have demonstrated a significantly increased prevalence of traditional CVD risk factors among adults with MDD and BD. Although limited in scope, available findings from youth with MDD and BD are largely convergent with adult findings. The data regarding MDD among youth have included representative population samples, many of whom were medication-naïve. In contrast, most data regarding youth with BD are based on clinical or pharmacoepidemiological data. Future studies based on representative population samples of youth with BD are needed.28 Prospective studies incorporating repeated measures of mood and traditional CVD risk factors are needed to better understand the bidirectional relationships between these variables among youth with MDD and BD.
Pathophysiological Processes Bridging Mood Disorders and CVD
As described in several comprehensive review articles, multiple systemic processes have been implicated in the association between mood disorders and CVD.3,4,73,74 Although other processes have also been examined, the sections below highlight inflammation, oxidative stress, and autonomic dysfunction in particular because these arguably have the strongest data thus far.
Over the past 20 years, a rapidly growing body of research supports the role of inflammation in MDD and BD among adults.75–78 Meta-analytic findings demonstrated increased inflammation in MDD.79 Similarly, 3 meta-analyses in 2013 yielded generally convergent findings in BD, namely, that there is evidence of a proinflammatory imbalance during symptomatic episodes.80–82 Preliminary genetic, epigenetic, and neurobiological findings suggest that this process may be part of the pathogenesis of MDD and BD rather than being a secondary epiphenomenon.83–87
Given the adult literature on this topic, excessive inflammation could underlie some of the association of depression with cardiovascular risk in youth. Indeed, in a handful of small case-control studies, clinically depressed adolescents have shown higher serum levels of some proinflammatory cytokines (eg, interleukin [IL]-1β and IL-6) than healthy control participants.88–90 The results of larger studies of community youth have been mixed, with most but not all studies demonstrating graded associations between depressive symptoms and inflammatory biomarkers such as IL-6 and C-reactive protein (CRP).88,91–93 Findings from adults suggest that inflammation may be particularly salient in early-onset BD,94,95 and there are preliminary findings from youth with BD. In a high-risk offspring study, an aberrant inflammatory gene expression signature was observed among 88% of adolescent and young adult BD offspring affected with mood disorders versus 19% of control subjects.96 Finally, in a study of 30 adolescents with BD, manic symptom severity was significantly associated with high-sensitivity CRP (r=0.37, P=0.04).97 High-risk levels of high-sensitivity CRP (≥2 μg/mL)98 were observed in 40% of the sample, and mean high-sensitivity CRP (3.1±4.6 μg/mL) was 3-fold higher than normal and nearly as high as in acute juvenile inflammatory arthritis.99 Because both CRP and IL-6 forecast later CVD risk in apparently healthy people,100–102 these findings suggest the possibility that low-grade inflammation is part of what underlies the increased risk of CVD in MDD and BD.
The question arises as to the direction of the association between depression and inflammation. From the available cross-sectional data in youth, it is not possible to discern how or why depression and inflammation come to be associated; however, it appears from both animal studies and human studies in adults that a bidirectional relationship exists.77 Rodents treated with inflammatory cytokines develop “sickness behaviors” that resemble depressive symptoms and can include irritability, anhedonia, loss of slow-wave sleep, reduced food intake, and circadian disruptions (particularly in cortisol release).103 Parallel symptom profiles often arise in patients treated with adjuvant cytokines for malignant melanoma.104,105 Consistent with the hypothesis that inflammation may contribute to the pathogenesis of some depressive episodes, a recent randomized controlled trial found that infusion of the tumor necrosis factor-α receptor antagonist infliximab can ameliorate depressive symptoms in patients with low-grade chronic inflammation.106 Moving in the other direction, depressed patients are prone to smoking, obesity, a sedentary lifestyle, and poor sleep quality,107,108 all of which foster inflammation.
Two multiwave prospective studies of adolescents have yielded results that are consistent with this bidirectional view. In 1 study, 147 teenaged females at high risk for an initial depressive episode were assessed every 6 months for 2.5 years.109 The transition to depression was accompanied by increases in levels of IL-6 and CRP, particularly for participants who had experienced previous childhood adversity. In time-lagged models among participants with previous adversity, high CRP levels persisted even after the depressive syndrome had resolved. These lingering effects were bidirectional. Among the participants with previous adversity, high IL-6 forecast risk of depression 6 months later, independent of inflammation at that time. Another study conducted follow-up of 1420 children from the Great Smoky Mountains Study who were assessed up to 9 times between the ages of 9 and 21 years.110 In time-lagged models, CRP did not forecast later depression risk; however, over time, CRP levels rose in concert with the number of depressive episodes. Of participants who experienced ≥2 depressive episodes, 42% subsequently exceeded the Centers for Disease Control and Prevention’s high-risk cutoff for CRP, which is 3.0 mg/L.111 In contrast, 9.8% of participants with 0 depressive episodes and 11.4% of participants with 1 depressive episode exceeded the high-risk cutoff.
Oxidative stress comprises a disturbance of increased pro-oxidant to antioxidant ratio, which leads to potential damage in numerous systems.112 Increased oxidative stress has been linked to endothelial dysfunction.113 Studies have demonstrated that oxidative stress and endothelial dysfunction contribute to cardiovascular risk factors including hypertension,114 diabetes mellitus,115 and dyslipidemia.116 The confluence of endothelial dysfunction and oxidative stress is implicated in the development of atherosclerosis.113,117
Oxidative stress markers are commonly increased in people with MDD and BD.118–124 The most recent meta-analytic data in MDD, based on 23 studies and 4980 patients, evidenced large elevations of oxidative stress during depression, as well as small to medium elevations in antioxidant markers.125 In addition, in several prospective studies, reductions in oxidative stress markers corresponded with improvement in psychiatric symptoms.126–129 Levels of thiobarbituric acid reactive substances and nitric oxide activity may be especially increased in patients with BD.119,129 Taken together, findings from these studies lend support to the hypothesis that oxidative damage may be involved in the pathophysiology of MDD and BD among adults.
In longitudinal observational epidemiological studies, impaired autonomic nervous system function (defined noninvasively by measures such as heart rate variability) is associated with incident hypertension130,131 and diabetes mellitus.132–135 Sympathetic activation plays a critical role in the persistence of hypertension through altered arterial baroreflex receptivity and by the promotion of renal dysfunction.136 Lower resting heart rate variability is also associated with incident clinical cardiovascular events among adults with and without diabetes mellitus.137–139 In most studies, there is a stronger association with fatal CHD and sudden cardiac death, which suggests an arrhythmic pathogenesis.140
Autonomic dysfunction is a primary pathological factor that relates psychological and neurological functioning to cardiovascular events.141–148 Under normal conditions, the parasympathetic division of the autonomic nervous system is responsible for neurovegetative functions (eg, appetite, sleep, energy), whereas the sympathetic division prepares the body to respond to a challenge by promoting coagulation and platelet activation, constricting arteries and vessels, and increasing hepatic production of glucose to transport to the muscles for immediate energy. In population and clinical studies, autonomic dysfunction is typically estimated by use of noninvasive estimates of the beat-to-beat variability in heart rate. Through statistical transformations of heart rate variability, individual contributions from the parasympathetic and sympathetic divisions can be estimated. However, in the absence of a stimulus to elicit sympathetic functioning, parasympathetic inputs are most commonly captured. Consequently, the majority of studies reviewed in this section are based on estimates of low parasympathetic input. Relatively fewer studies have investigated BD in relation to autonomic dysfunction, and those studies that have been conducted have relied on very small sample sizes. The following summary of findings therefore includes both MDD and BD.
Autonomic nervous system dysfunction is common in MDD, having been observed in adults147,149 and adolescents.141 In a 2010 meta-analysis of 18 articles,150 patients who were clinically depressed had significantly less favorable measures of autonomic dysfunction reflecting parasympathetic function than nondepressed patients; moreover, depression severity was correlated with autonomic dysfunction. Some90,151,152 but not all153 studies have suggested that autonomic nervous system functioning, most commonly estimated as parasympathetic withdrawal, is worse in patients with BD than in those without. Although autonomic dysfunction has been identified in the pathway between depressive disorders and CVD in adults,154 autonomic dysfunction is less likely to lead to end-organ damage in adolescents and young adults because of the decades-long process by which CVD develops over time. Rather, the pathway by which CVD risk is enhanced in depression and BD is more plausibly related to the development of CVD risk factors such as hypertension and diabetes mellitus. CVD is a distal consequence that is mechanistically plausible given the association of depressive disorders with numerous pathophysiological changes in the vasculature that result from hypertension and diabetes mellitus. The potential impact of psychotropic medications on autonomic and other findings is discussed in “Psychotropic Medications and Cardiovascular Risk Factors in Youth.”
Behavioral and Environmental Factors Contributing to CVD Risk
A number of behavioral and psychosocial characteristics that are disproportionately prevalent among adolescents and young adults with mood disorders are associated with increased CVD risk. Such characteristics include early maltreatment, sleep disturbance, sedentary lifestyle, suboptimal nutrition, and tobacco smoking and substance abuse.
This section focuses on maltreatment specifically, rather than negative life events in general, because the literature is most robust on this topic. Nonetheless, other stressful life events, particularly if severe, recurrent, or persistent, may also contribute to CVD risk. Some portion of the association of depression with cardiovascular risk may reflect a common influence of childhood maltreatment on the pathogenesis of these conditions. A recent study estimated that 13.7% of children in the United States experience parental maltreatment annually.155 Maltreatment can take the form of neglect or emotional, physical, or sexual abuse.155,156
With regard to depression, children exposed to maltreatment are at increased risk of developing an episode of major depression in their lifetime,157–160 are prone to residual symptoms of depression once the episode subsides, have frequent depressive recurrences, and are resistant to a variety of front-line treatments.161 With regard to biology, relative to nonexposed children, maltreated youth demonstrate dysregulation of the hypothalamic-pituitary-adrenal axis,162,163 higher levels of CRP,164 and more signs of cardiometabolic risk through early and middle adulthood.165–168 In samples of clinically depressed adults, the prevalence of MetS and low-grade inflammation is significantly higher among those who were exposed to childhood adversities, including maltreatment, than among those who were not.109,169,170 With regard to vascular outcomes, lifetime rates of stroke, myocardial infarction, and other forms of CVD are elevated among people who report having been maltreated as children.109,171,172 However, many of the studies on this topic have used retrospective and unconfirmed reports of maltreatment, which complicates interpretation of their results. Given the above evidence that early maltreatment has been associated with mood disorders and with CVD, continued research bridging these topics, particularly prospective research and research that includes confirmed reports of maltreatment, is warranted.
Sleep disturbances are included among the diagnostic criteria for both MDD and BD.23 A variety of sleep disturbances have been associated with symptoms of mood disorders among youth.152,173,174 One study showed that depressed adolescent boys had short REM (rapid eye movement) latency and more frequent nighttime arousals, although depressed adolescent girls showed the same sleep patterns as their healthy counterparts.175 Circadian phase shifts can occur before the onset of and during a depressive episode in adults and adolescents. Among children and adolescents, disrupted circadian rhythms also occur among those with mood disorders, resulting in delayed and often shortened sleep. Both short and long durations of sleep can precede mood disorders in adolescents and young adults.176 Interestingly, short and long sleep duration have also been associated with increased CIMT, increased inflammatory markers, and poor cardiovascular outcomes in adults177; short sleep duration has also been associated with obesity and insulin resistance among children and adolescents.178,179 Although circadian therapies can alleviate depressive symptoms and mood disturbances in adolescents and young adults,176,180,181 the effect of these therapies on cardiovascular risk factors is not known.
Sedentary behavior increases cardiovascular risk factors and disease among adolescents, young adults, and older adults, and physical activity decreases cardiovascular risk through multiple pathways, including reduced body weight, improved endothelial and immune function, and improved blood pressure.182 People with MDD or depressive symptoms are more likely to be sedentary relative to those without mood disorders.183–185 The association of physical activity with BD is inconsistent. Some studies have demonstrated reduced physical activity and reduced exercise tolerance in BD,186,187 whereas other studies have demonstrated no significant differences.188 Although many studies have been cross-sectional, and thus, the relationship may not be causal, prospective data from the National Health and Nutritional Examination Survey I demonstrated that physical inactivity was a significant predictor of the development of depressive symptoms in white women.189 Similarly, in a prospective study of a Swedish cohort of young males, low fitness related to sedentary behavior was a significant risk factor for later depression.190 In contrast, in a German population study, regular physical activity among adolescents and young adults predicted an increased incidence of BD at prospective follow-up.191 However, the extent to which sedentary behavior influences the relationship between depression and CVD is not yet clear.
Nutrition could play a role in the association of depression with cardiovascular risk; however, relatively little attention has been devoted to exploring this hypothesis. A small study reported higher glycemic load and higher scores on Western and modern dietary patterns among women with BD.192 In a handful of cross-sectional studies with adults, depression has been associated with reduced levels of micronutrients and their metabolites, some of which may be relevant to CVD.193,194 Disparities in zinc, folate, and vitamins D and E have been reported between depressed and nondepressed adults,195,196 although as a whole, the findings have been inconsistent and constrained by methodological limitations.197 In a recent meta-analysis, there were significantly lower blood zinc concentrations among depressed adults (n=1642) than among control subjects (n=804).198 Greater depression severity was associated with greater zinc deficiency. In 1 prospective study, older adults with low serum vitamin D were at increased risk for developing a depressive episode over the next 3 to 6 years.199 Low vitamin D has also been reported in depressed adolescents, and with open-label supplementation, these patients’ mood symptoms improved significantly.200 Of note, vitamin D is the only supplement recommended in the recent Expert Panel.2 Thus far, it is unclear whether micronutrient disparities account for the association of depression with cardiovascular risk.
Fish oils, specifically omega-3 fatty acids, have also received much attention in relation to the link between CVD and mood disorders.201,202 Observational data suggest that greater seafood consumption is associated with lower rates of BD.203 Omega-3 deficits have been reported among adults with MDD and BD.204,205 A small study compared youth with versus without BD and found large between-group differences in red blood cell membrane concentrations of omega-3 levels; however, these were not significant after controlling for dietary intake.206 However, greater docosahexaenoic acid levels were associated with significantly lower depressive symptoms in the BD group. The Ryukyus Child Health Study (n=3067 boys and 3450 girls, 12–15 years old) examined fish intake and omega-3 levels in relation to depressive symptoms.207 Higher eicosapentaenoic intake was associated with significantly lower depressive symptoms, and a similar trend was observed for docosahexaenoic intake. Clinical trials of omega-3 supplementation, spurred by these findings, have yielded positive findings for the treatment of depression among adults with MDD and BD.208,209 Preliminary studies also suggest potential benefits of omega-3 treatment among youth with BD and MDD.210–212 Recent studies have drawn into question the role of omega-3 treatment for secondary prevention of CVD in the general population,213 as well as for adjunctive treatment of depression among adults with CVD.214 However, given the above findings, future studies examining the impact of omega-3 treatment on CVD-relevant outcomes (eg, endothelial function, inflammation) and mood symptoms among youth with MDD and BD are warranted.
Tobacco Smoking and Substance Use
Cigarette smoking is the most significant behavioral risk factor associated with CVD risk,215 and the severity of CVD increases as a function of the number of pack-years.216 Adults with MDD and BD are 2 to 3 times more likely to be smokers and less likely to successfully quit smoking.217–220 Adolescents with MDD and BD are more likely to begin smoking, and begin smoking earlier, than their nondepressed counterparts,221–224 a significant association because nearly all adult smokers begin smoking as adolescents. There is some evidence that smoking exacerbates the association between CVD and mood disorders. For example, a recent cross-sectional study found that smoking, in combination with elevated body mass index, increases the risk of CVD among adults with depression, BD, and other mood disorders.225
Similar to the data on smoking, adolescents with MDD and BD are at increased risk of alcohol and substance abuse.226–229 However, little research has examined the contribution of alcohol and substance abuse to the elevated CVD risk among those with mood disorders.
In summary, a variety of behavioral and psychosocial factors may contribute to the increased risk of CVD among youth and adults with MDD and BD, potentially acting through the aforementioned pathophysiological mechanisms (eg, inflammation, oxidative stress) and cardiovascular risk factors (eg, increased blood pressure). Primary among these are childhood maltreatment, sleep disorders, physical inactivity, and smoking. These linkages are in need of further study, particularly among adolescents and young adults with MDD and BD. With that said, many of the studies we reviewed statistically controlled for these lifestyle variables and still found significant associations between mood disorders and cardiovascular risk. As such, lifestyle variables may contribute to but do not fully explain the link between mood disorders and cardiovascular risk.
Psychotropic Medications and Cardiovascular Risk Factors in Youth
We recognize that although there are many unanswered questions regarding the role of psychiatric medications in the association of mood disorders with CVD, this topic is nonetheless timely and salient to this statement. For this reason, and although a systematic review of this topic is beyond the scope of this statement, we offer a brief overview and several tentative conclusions on this topic.
There is some evidence to suggest that autonomic dysfunction is attributable to medications used to manage depression147,230; however, a meta-analysis150 concluded that the adverse effect of antidepressant drugs was restricted to adults taking tricyclic antidepressants, and similar findings were recently reported among adults participating in randomized controlled clinical trials.231 Tricyclic antidepressant drugs are not indicated for the treatment of mood disorders among youth and are not commonly used in this population, whereas selective serotonin reuptake inhibitors (SSRIs) are the pharmacological treatment of choice for depression among youth.232 The topic of SSRIs and CVD has been examined in great detail among adults, and the current literature suggests that SSRIs may have a salutary effect on CVD (particularly citalopram and sertraline) and are unlikely to have a deleterious effect231,233; however, there is evidence that SSRIs may confer a risk of obesity and possibly glycemic control problems.234,235 This topic has received limited attention among youth to date. In a large study of adolescents with treatment-resistant depression, 12 weeks of treatment with venlafaxine (a serotonin and norepinephrine reuptake inhibitor; n=166) was associated with greater increases in diastolic blood pressure than treatment with SSRIs (n=168).234 The same study found an ≈2-kg increase in weight for venlafaxine-treated youth and ≈3-kg increase in weight in SSRI-treated youth. Although significant weight gain is not a frequently reported side effect in placebo-controlled clinical trials, most trials have not systematically reported changes in weight/obesity. Moreover, anorexia and weight loss are a common symptom of depression, such that weight gain in itself may reflect symptom improvement and may not be a negative outcome. The association between antidepressant drugs and glycemic control is complex. It remains unclear whether antidepressant drugs have a salutary, neutral, or deleterious effect on glycemic control, because different studies have reported each of these associations.235–237
Lithium and Anticonvulsant Drugs
Lithium, carbamazepine, and divalproex are associated with significant weight gain among youth with BD.56,238,239 The magnitude of this association is significantly smaller than second-generation antipsychotic drug (SGA)–associated weight gain among youth.240 There is also pharmacoepidemiological evidence that exposure to these medications among youth is associated with hypertension.241 A recent 8-week study found that in contrast to treatment with risperidone, neither lithium (n=62) nor divalproex (n=78) yielded significant changes in blood glucose.239 Divalproex does not appear to increase risk of dyslipidemia in clinical studies of youth with BD and may in fact improve lipid profiles to some degree.239,242 Despite its association with weight gain, observational data in adults with BD suggest that long-term treatment with lithium may attenuate the risk of CVD.243
Second-Generation Antipsychotic Drugs
SGAs are highly efficacious for mania, and several have received US Food and Drug Administration approval for the treatment of adolescent BD and are considered first-line treatments.244,245 Unfortunately, antimanic SGAs (with the exception of ziprasidone) have been shown to negatively impact each of the MetS components in pharmacoepidemiological studies and in controlled trials.56,240,246,247 Despite metabolic monitoring guidelines (ie, glucose, lipids, weight, blood pressure) for patients treated with SGAs from the American Diabetes Association/American Psychiatric Association248 and the International Society for Bipolar Disorder,249 adherence to these monitoring guidelines is poor, particularly among youth.250,251 Patient-, clinician-, and system-level strategies are warranted to increase the proportion of youth achieving guideline-concordant monitoring benchmarks.
Antidepressant medications (including SSRIs) and mood-stabilizing medications (particularly SGAs) can cause weight gain and may also affect other metabolic parameters among adults and youth. There is no doubt these are undesirable side effects in a population with increased risk of CVD; however, it is important to acknowledge that despite the clear evidence regarding CVD risk factors, evidence that these medications cause increases in CVD or CVD mortality is lacking to date.231,252 Reasons for this discrepancy are uncertain; however, compelling contradictory findings regarding high-potency statins (decreased risk of CVD despite increased risk of T2DM) suggest that this discrepancy is not unique to psychiatric medications.253 Pending definitive studies on this topic, one might speculate that the anti-inflammatory effects of these medications, or other yet unknown pharmacological properties, may partially mitigate their adverse effects on traditional CVD risk factors.96,254–257
Although it remains possible that psychotropic medications contribute in part to CVD risk in MDD and BD, there are 3 primary reasons to conclude that MDD and BD are moderate-risk conditions independent of the effect of psychotropic medications and that medications should not be the only focus. First, the strongest evidence regarding these medications relates to increased CVD risk factors, and the best available evidence suggests that the association between mood disorders and CVD is independent of CVD risk factors.37,46,258 Second, the association between mood disorders and excessive CVD risk was described decades before the advent of these medications.11,14,259 Third, many if not most people in population-based studies on this topic had not received pharmacological treatment for their mood disorders.5,16,260 This is likely to be particularly true for adolescents, at least 60% of whom do not receive any treatment for their MDD or BD, let alone pharmacological treatment.261
Overall Summary and Future Directions
The central goal of this statement is to position MDD and BD alongside other pediatric diseases previously identified as moderate risk for CVD. Despite some inconsistencies, MDD and BD clearly satisfy the criteria that have been set for tier II moderate-risk conditions and equal or exceed the evidence for other conditions in this category. Taken together, the evidence reviewed in this statement demonstrates that MDD and BD among youth are tier II moderate-risk conditions that predispose to accelerated atherosclerosis and early CVD. Cardiovascular risk among youth with MDD and BD should therefore be managed in accordance with recent Expert Panel integrated guidelines, applying the same recommendations made for other moderate-risk conditions (Figure 2).2 The magnitude of increased risk for CVD in adulthood is substantial, most likely because of a combination of factors that include direct effects of MDD and BD through shared pathophysiological processes (eg, inflammation), direct effects of mood disorder symptoms (eg, sleep disruption), indirect effects of mood disorder symptoms (eg, smoking, suboptimal nutrition and physical activity), and accumulation of excessive traditional CVD risk factors. Importantly, there is evidence of clustering of risk factors (eg, smoking, hypertension, obesity, suboptimal nutrition or physical activity) among people, including youth, with MDD and BD. As a result, although MDD and BD warrant inclusion among tier II moderate-risk conditions overall, careful assessment is warranted to additionally identify those youth with risk factor clustering (≥2 risk factors), for whom a tier I high-risk designation would apply.
Future Research Directions
Similar to the association between depression and CVD among adults, future studies are needed to determine whether any potential impact of other psychiatric disorders, such as anxiety disorders, on CVD is independent, cumulative, or synergistic.262 Several studies discussed in this scientific statement demonstrated sex differences. Therefore, future studies are also needed to achieve a better understanding of sex differences as they relate to the risk of CVD among youth with mood disorders. In addition, although many studies examined mood symptoms rather than diagnoses, the strength of the observed associations was in some instances greater when mood diagnoses were examined. Future studies should therefore ascertain both diagnoses and symptoms and should examine whether the severity, persistence, and recurrence of symptoms impacts CVD risk. There is an unrealized opportunity to increase our knowledge on this topic by incorporating CVD-related measures in studies of youth with MDD and BD. Traditional CVD risk factors and high-sensitivity CRP, as well as ascertainment of CVD-related family medical history, can be readily and inexpensively measured and should be incorporated into large cohort studies of youth with MDD and BD, as well as youth at familial risk for MDD and BD. The converse also holds true: mood symptoms and diagnoses, as well as family psychiatric history, should be incorporated into large cohort studies focused on CVD risk among youth. Approaching the mood-CVD link from these complementary perspectives offers the chance to maximize progress in understanding the epidemiology and familial nature of this link. Finally, no studies have systematically examined integrated treatment strategies among youth, although preliminary approaches to mitigating the effect of mood-stabilizing medications on obesity and T2DM have been described.263–266 It is therefore unclear whether treatment of the mood symptoms of MDD and BD among youth can reduce the risk of accelerated atherosclerosis and premature CVD, and this is an important question for future research to address. In particular, research regarding the dual mood-related and CVD-related benefits of complementary and alternative interventions such as exercise, mindfulness meditation, antioxidants, and fish oils is warranted. Indeed, future clinical trials would benefit from the incorporation of CVD-related measures and mood-related measures to generate more integrated data regarding the global benefits and risks of interventions.
Future Clinical Directions
Medications used to treat mood disorders among youth, particularly those used to treat BD, confer additional risk of weight gain and other metabolic disturbances. Although the magnitude of the impact of these medications on CVD and CVD mortality has yet to be elucidated, improved metabolic monitoring is warranted to mitigate the accumulation of traditional CVD risk factors. Current guidelines for the treatment of BD among youth acknowledge the importance of metabolic monitoring of those taking mood-stabilizing medications but do not address the importance of metabolic monitoring of youth with BD irrespective of treatment.26,267 Similarly, current guidelines for the treatment of MDD among youth do not adequately incorporate cardiovascular risk factors.24,232,268 Future guidelines should integrate cardiovascular risk factor monitoring irrespective of treatment.
A transformational change is required in the management of MDD and BD among youth to meaningfully integrate cardiovascular risk assessment and management into the day-to-day treatment of these conditions. Related changes have begun in the treatment of adults with MDD and BD. Indeed, there is preliminary evidence that integrating cardiovascular risk assessment and management in the treatment of adults with MDD and BD may have salutary effects both on CVD risk and on psychiatric outcomes.269–271 Although unfortunately, CVD is already highly prevalent by middle age among adults with MDD and BD, there remains a substantial window of opportunity in which to intervene to prevent these outcomes among youth.
To meaningfully change the cardiovascular risk associated with MDD and BD among youth, a concerted effort across stakeholder groups will be required, including pediatricians and other primary care providers, psychiatrists, patients and their families, research funding agencies, and policy makers. Pending the development of evidence-based guidelines that are specific to youth with MDD and BD, it is important that the above stakeholders collaboratively endeavor to ensure that previously described Expert Panel integrated guidelines, with consideration of MDD and BD as additional moderate risk conditions, are consistently applied to these youth.
Disclaimer: The views expressed in this paper are those of the authors and do not necessarily reflect those of the National Heart, Lung, and Blood Institute; the National Institutes of Health; or the US federal government.
The American Heart Association makes every effort to avoid any actual or potential conflicts of interest that may arise as a result of an outside relationship or a personal, professional, or business interest of a member of the writing panel. Specifically, all members of the writing group are required to complete and submit a Disclosure Questionnaire showing all such relationships that might be perceived as real or potential conflicts of interest.
This statement was approved by the American Heart Association Science Advisory and Coordinating Committee on February 10, 2015, and the American Heart Association Executive Committee on March 22, 2015. A copy of the document is available at http://my.americanheart.org/statements by selecting either the “By Topic” link or the “By Publication Date” link. To purchase additional reprints, call 843-216-2533 or e-mail .
The American Heart Association requests that this document be cited as follows: Goldstein BI, Carnethon MR, Matthews KA, McIntyre RS, Miller GE, Raghuveer G, Stoney CM, Wasiak H, McCrindle BW; on behalf of the American Heart Association Atherosclerosis, Hypertension and Obesity in Youth Committee of the Council on Cardiovascular Disease in the Young. Major depressive disorder and bipolar disorder predispose youth to accelerated atherosclerosis and early cardiovascular disease: a scientific statement from the American Heart Association. Circulation. 2015;132:965–986.
Expert peer review of AHA Scientific Statements is conducted by the AHA Office of Science Operations. For more on AHA statements and guidelines development, visit http://my.americanheart.org/statements and select the “Policies and Development” link.
Permissions: Multiple copies, modification, alteration, enhancement, and/or distribution of this document are not permitted without the express permission of the American Heart Association. Instructions for obtaining permission are located at http://www.heart.org/HEARTORG/General/Copyright-Permission-Guidelines_UCM_300404_Article.jsp. A link to the “Copyright Permissions Request Form” appears on the right side of the page.
* CVD is used in this statement to denote atherosclerotic/coronary/ischemic heart disease; however, when referencing previous studies that included structural or infectious heart disease within CVD, this statement specifies the term that best reflects the above definition of CVD.
† “Youth” is used in this statement as an umbrella term for children, adolescents, and/or young adults (defined here as mean sample age within 1 standard deviation of 30 years).
- © 2015 American Heart Association, Inc.