Screening for Psychological Distress and Risk of Cardiovascular Disease and Related Mortality: A Systematized Review, Meta-analysis, and Case for Prevention

Abstract

Background:

Psychological distress – elevated symptoms of depression, anxiety, posttraumatic stress disorder (PTSD), or psychosocial stress – has been associated with cardiovascular disease (CVD) risk. Despite increasing attention to the importance of these factors for CVD prevention, the state of this science requires updated synthesis to enable practice recommendations. Moreover, it is unknown if psychological distress based on screeners, validated self-report instruments that efficiently identify those who may require mental health services or additional support, is associated with incident CVD.

Methods:

MEDLINE, EMBASE, and PsychInfo were searched for studies including adults without a past psychiatric diagnosis, who were screened at baseline for depression, anxiety, PTSD, stress, or general mental health symptoms, and followed for >6 mo to determine their risk for incident CVD (i.e., atrial fibrillation, acute coronary syndrome, coronary heart disease, peripheral vascular disease, heart failure, or a composite). A meta-analysis was used to aggregate results to determine if clinically significant levels of psychological distress were associated with CVD onset.

Results:

The search identified 28 investigations that represented 658,331 participants (58% women). Fifteen studies had adequate data for the meta-analysis, which indicated that those reporting high psychological distress showed a 28% greater risk of incident CVD compared to those with low or no distress.

Conclusions:

Rapid screening for psychological distress is a helpful and efficient approach to understanding an individual’s CVD risk profile. Additional investigations are needed to improve prospective evidence concerning psychosocial stress. Conducting analyses by sex may better elucidate the benefits of psychological distress screening and encourage more widespread adoption in CVD prevention.

Condensed Abstract

Over 600,000 patients across 28 studies were evaluated by meta-analysis to determine if psychological distress assessed with brief, screeners was prospectively associated with cardiovascular disease (CVD). Psychological distress was associated with a 28% greater risk of incident CVD, providing context for the importance and feasibility of psychosocial assessment for better primary prevention.

Primary prevention initiatives are gaining in prevalence to forestall the onset of cardiovascular disease (CVD).^1,2 In the United States (U.S.), patients and their healthcare providers are currently afforded the most advanced biobehavioral toolkit for cardiovascular risk reduction.³ Still, with an obesity epidemic, sedentary lifestyles, an aging population, and broadening social health disparities, CVD prevalence and associated costs of healthcare and human life continue to rise.⁴

Psychological health is an important dimension of cardiovascular health and wellbeing.^5,6 Significant evidence from epidemiology, psychology, cardiology, and public health shows that psychological distress – i.e., elevated symptoms of depression, anxiety, posttraumatic stress disorder (PTSD), or perceived psychosocial stress – is associated with earlier CVD onset, more rapid CVD progression, poorer prognosis, and an increased risk of related death.^7–14 INTERHEART, a global, case-control study of first myocardial infarction, was one of the largest investigations to demonstrate the importance of psychological health: the population attributable risk of psychological stress and depression was 33%, which exceeded the risk associated with some traditional factors for CVD (e.g., hypertension, physical inactivity).¹⁵ Additional, prospective cohort studies and meta-analytic summaries have since reinforced these findings, offering compelling evidence that psychological distress – based on clinical diagnoses, diagnostic interviews, or self-reported symptoms – is involved in the risk for, and burden of, CVD.^{9,12–14,16–18}

With accumulating evidence linking psychological health and risk for CVD,¹⁹ leaders in cardiovascular medicine increasingly acknowledge psychological health’s importance in cardiac and vascular risk.^5,20 Contradictory to this acknowledgement, and despite cardiovascular rehabilitation’s longstanding empirically-based focus on psychological health and stance that psychological health is as equally important as traditional factors, professional associations in cardiovascular medicine continue to be cautious about translating evidence concerning psychological health and cardiovascular risk into clinical guidelines for distress screening and management. Most prominently, the American Heart Association (AHA) recently updated their cardiovascular health metric to “The Essential 8” to include sleep, and suggested that psychological health serves as a context for other health factors (e.g., sleep, weight), but that greater evidence is needed to guide the implementation of psychological screening and management.¹⁹ Historically, this tempered enthusiasm may have translated into a greater emphasis on psychological distress surveillance in patients with established CVD – e.g., AHA Recommendations for Screening, Referral, and Treatment for Depression focused on patients with coronary heart disease – rather than identifying opportunities for primary prevention.²¹ As the cardiovascular risk that is associated with psychological distress likely begins well before CVD onset, and managing such symptoms appears strategic for reducing cardiovascular risk, patients and providers alike would benefit from the earlier identification of psychological distress.

Across medicine, policymakers, thought leaders, and professional organizations are advocating for more widespread surveillance of psychological distress in routine care settings. For example, since 2009 the U.S. Preventive Services Task Force has recommended annual depression screening among adults aged ≥18 yr and highlights increased depression risk among those with CVD,²² and the Centers for Medicare & Medicaid Services have covered annual depression screening for adults since 2011,²³ but changes to U.S. healthcare policy develop slowly. As the current healthcare delivery landscape demands efficient evaluation of patients’ distress,²⁴ efforts to monitor psychological health using brief, validated self-report measures rather than more burdensome, comprehensive, psychological evaluations or psychiatric interviews continue to gain momentum. Several recent meta-analyses have summarized literature concerning psychological distress and initial CVD risk.^7,9–11,14 Yet, no such investigation has specifically focused on the use of brief, self-report screening measures to evaluate distress. This effort is integral to creating screening guidelines for psychological distress in the service of CVD prevention, efforts which will particularly benefit populations with a high risk of CVD.

The primary objective of this investigation was to update and extend literature concerning psychological health and the incidence of new-onset CVD or related mortality by conducting a systematized review of all recent studies in which psychological distress was identified with brief, self-report screening measures only. A complementary objective was to quantify the strength of recent evidence via meta-analysis. Given the breadth of this literature, and past meta-analyses of subdomains of psychological health, a final objective was to identify opportunities for further research concerning psychological health screening and CVD and to develop recommendations for applying these results to improve the implementation of such measures in CVD primary prevention.

METHODS

Search Procedures

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement²⁵ was used to guide the literature search and associated statistical analyses. Studies were included in the systematized review if they were 1) published in the last 5 yr (2017 – 2022), to provide estimates based on the most recent state of this science, 2) designed to evaluate the prospective association of psychological health (i.e., symptoms of depression, anxiety, PTSD, psychosocial stress, or general mental health-related quality of life) that was assessed at baseline with a validated self-report measure, and risk for incident CVD events, diagnoses, hospitalization, or related mortality, 3) participants were aged ≥18 yr and free from a diagnosed psychiatric disorder at baseline, 4) participants did not have CVD at baseline or CVD was controlled in analyses of more heterogeneous samples, and 5) the follow-up duration was ≥ 6 mo. Studies were excluded if psychological health was assessed with a formal psychiatric interview or psychological evaluation, or if the predictor was based on a psychiatric diagnosis derived from the medical record. Due to the different pathophysiology of stroke compared to other types of CVD and recent findings highlighting the complexity of psychological distress as a predictor of fatal and nonfatal ischemic stroke,^26,27 investigations that included stroke within a composite endpoint were excluded. Despite being a common endpoint, studies that focused on hypertension were also excluded due to being a risk factor for CVD rather than established CVD; and further, since first diagnosis of hypertension typically predates CVD by many years, prevention processes and timelines are likely very different. We excluded records from non-peer reviewed sources such as theses, dissertations, and conference abstracts, due to different standards of peer review. We also excluded studies not available in English. The primary meta-analysis was limited to studies reporting adjusted hazard ratios (HR) and associated 95% CI (or other information allowing for their calculation) for an association between psychological distress and a relevant CVD outcome. Studies that only reported continuous HR were excluded because our primary interest was in determining if high distress (relative to no or low distress) was associated with CVD risk.

Information Sources and Search Strategy

A medical librarian (LMF) searched the MEDLINE (Ovid), EMBASE (Embase.com), and American Psychological Association (APA) PsychInfo (EBSCO) databases for relevant journal articles. The final search was conducted on May 16, 2022. The following search terms were among those used to identify the appropriate articles: anxiety; depression; depressive disorder; stress disorders, posttraumatic; posttraumatic stress disorder; perceived stress; psychosocial stress; stress; GAD-7; PHQ-9; Patient Health or Penn State Worry or General Health Questionnaire; DBI or STAI or BAI or BDI-II; or state-trait or Spielberger or mental health) W3 (inventory HDRS or Ham-D or HADS or CESD or CES-D or HDI or SDS or GDS or HADS-D; Hamilton or hospital or center for epidemiologic studies or self-rating or geriatric or Zung self-rating) W3 (depression scale); CVD; coronary disease; myocardial infarction; myocardial ischemia; death, sudden, cardiac; coronary artery disease; atrial fibrillation; heart failure; peripheral artery disease; peripheral vascular diseases; risk; incidence; or new or onset or risk or predict; and all the possible combinations of these terms.

Duplicate records were removed via auto-deduplication in EndNote 20 and manual deduplication. Title and abstract screening were completed in Covidence according to inclusion and exclusion criteria outlined above, and irrelevant articles were excluded. Next, the full texts of the remaining articles were thoroughly examined according to the criteria, and unrelated studies were again excluded in EndNote. To maintain objectivity, screening and data extraction activities were performed by two independent reviewers (AEG and E.C.G.). To ensure a comprehensive search, the reference lists used within all the collected articles were manually reviewed. Interrater reliability was assessed for full text review, with raters agreeing on 77% of the records. Discrepancies were resolved through discussion until consensus was reached.

Study Selection and Data Extraction

A total of 3,983 records were initially identified through the database searches: MEDLINE (n=1616), EMBASE (n=1733), and APA PsycInfo (n=594). Of the initial search, 498 duplicate records were removed, and 3,485 records were eligible for title and abstract screening.

Data from the final studies were extracted using detailed coding forms. Items included: article title, first author’s name, year of publication, place of study, sample size, assessment method, demographics of the sample (average age, sex, race and ethnicity, education, marital status employment), type of study, the prevalence of depression, anxiety, and stress, method of assessing psychological distress, CVD outcomes, and statistical details (i.e., maximally adjusted HR and 95% CI, covariates). For inclusion in the meta-analysis, when studies with overlapping samples were identified, we prioritized the study with the largest sample which reported an association between baseline psychological distress and a relevant outcome. When studies with overlapping samples reported different cardiovascular outcomes, we prioritized the study reporting the most relevant cardiovascular outcome. If studies reported HR associated with tertiles or quantiles, the highest category was included. Fixed effects methods were used to combine subgroups and derive a study-level effect. For studies reporting outcomes associated with more than one metric of psychological distress (i.e., depression and anxiety), we prioritized the non-depression psychological construct, given the more limited number of such studies. We requested additional information from authors of 5 additional studies, but none responded in the allotted time (i.e., 4 mo).

Statistical Analysis

For the primary analysis, adjusted HR reflecting the association between psychological distress and CVD morbidity risk were used as the measure of effect; other measures of relative risk were considered equivalent for the purposes of this analysis. Hazard ratios > 1 indicated that psychological distress (e.g., high symptoms of depression) was associated with a greater incidence of CVD. As part of a sensitivity analysis, a broader definition of relative risk was used to pool adjusted HRs and unadjusted or adjusted ORs of associations between psychological distress and risk of CVD morbidity or mortality.

Random effects procedures with a restricted information maximum likelihood approach were used to aggregate effect sizes (i.e., logarithmic adjusted HR) and corresponding 95% CI to estimate the overall effect, which was converted back to a HR.^28,29 The Q statistic was computed to assess heterogeneity (i.e., inferred from a significant value). Outcome consistency across studies was estimated based on the I² index and its corresponding 95% CI.^30,31 I² values of 25%, 50%, and 75% are interpreted as low, medium, or high heterogeneity.³² Analyses were conducted using the Stata meta package Version 16 (Stata).³³

We assessed publication bias for analyses with outcomes that were reported in ≥10 studies.³⁴ Visual inspection of funnel plots and Egger’s test were used to evaluate the possibility of publication bias and small study effects, respectively.^35–37

RESULTS

Of the 3,445 records from the original search, 3,251 were excluded in the title and abstract review, and 166 records were excluded in the full-text review. Ultimately, 28 investigations met the inclusion criteria and had an endpoint of CVD morbidity or mortality, of which 18 were included in our analyses (Figure 1 includes additional details of the study selection process).

Figure 1.

PRISMA 2020 flow diagram of screening and selection procedures.

Sample Characteristics

Altogether, the studies represented 658,331 participants (58.4% women; 66.9% White [n=13 reported data on race]; 67.0% married or cohabiting [n=11 reported relevant data]). Characteristics of the included prospective cohort studies are depicted in Table 1, separated by the type of psychological distress. Most investigations included samples of healthy adults (n=13), although others were limited to adults middle-aged and older (n=10), members of specific racial/ethnic groups (n=3), men (n=2), post-9/11 emergency workers (n=1), or U.S. military Veterans (n=1). Most participants were self-selected from the broader population (n=21), but some were recruited through a clinical contact (n=4), electronic health records (n=1), or other means (n=4; e.g., employed in the British Civil Service). In terms of risk factors for CVD, 21.9% of participants smoked (n=24), 11.2% had diabetes (n=19), 41.7% had hypertension (n=17), and the average body mass index was 26.3 ± 4.4 (n=11).

Table 1.

Characteristics of the included studies, separated by area of psychological distress.

Study	Region	Baseline, yr	Sample, n	Follow-up duration, mo	Age, mean or median yr	Women, %	Measure of psychological distress	CVD outcomes	Covariates
Depression
Deschenes et al., 2019	Quebec, Canada	2009–2010	33455	84	53	57	PHQ	CAD, HF, MI	Age, sex, ethnicity, education, smoking, alcohol, PA, cholesterol, DM, HTN
Dixon et al., 2021	12 Southeastern U.S. states	2002–2009	23937	132	53 (median)	64	CES-D	HF	Age, sex, race, HTN, HLD, DM, BMI, smoking, income, education, employment, marital status, alcohol, PA, number of close friends, depression, antidepressant use, cerebrovascular disease
Feng et al., 2020	Nord-Trøndelag County, Norway	2006–2008	37402	96	53	57	HADS-Depression	AF	Age, sex, weight, height, smoking, occupation, marital status, PA, alcohol use, chronic disorders, blood glucose, BP, triglycerides, HDL, CRP
Gaffey et al., 2022	Jackson, Mississippi, US	2000–2004	2651	120	53 (median)	64	CESD	HF	Age, education, income, HTN, DM, CHD, eGFR, total cholesterol, LVEF, alcohol abuse, smoking, obesity, PA, HR, SBP
Garg et al., 2019	6 U.S. sites (Baltimore, MD; Chicago, IL; Forsyth County, NC; Los Angeles, CA; New York City, NY; St. Paul, MN)	2000–2002	6664	156	62	53	CESD	AF	Age, sex, race, education, income, clinic site, cigarette smoking, BMI, height, DM, glucose, SBP, PA, statin use, antihypertensive use, alcohol use
Han et al., 2022	28 provinces in China	2011–2012	8621	84	58	48	CESD	Composite	Age, sex, zip code, education, smoking, alcohol use, BMI, SBP, antidepressant use, medical history
Harshfield et al., 2020	Consortium of 21 studies from Europe, North America, and Australia	ERFC: 1974–2010	162036	114	63	73	CESD	CAD	Age, sex (stratified), smoking, DM
Karlsen et al., 2020	Six U.S. sites (Birmingham, AL; Minneapolis, MN; Palo Alto, CA; Pittsburg, PA; Portland, OR, San Diego, CA)	2003–2005	3095	144	76	0	GADS	MI, HF, Other	Age, anxiety, education, ethnicity, DM, antidepressant use, smoking, alcohol use, BP, cholesterol, BMI, PA
Ladwig et al., 2017	Germany	1984–1985	3428	120	58	0	von Zerssen symptom checklist: Depression/Exhaustion subscale	CAD mortality	Age, HCL, obesity, HTN, smoking, DM.
Lemogne et at., 2017	France	1993	10541	252	48	26	CESD	Composite	Age, sex, occupational grade, parental history of CVD, alcohol use, smoking, PA, BMI, HTN, dyslipidemia, DM, sleep complaints
Li et al., 2019	28 provinces in China	2011–2012	12417	48	58	51	CESD	Composite	Age, sex, residence, marital status, education, smoking, alcohol use, SBP, BMI, history of diabetes, HTN, DLP, CKD, antihypertensives, DM medication, lipid lowering therapy
Li et al., 2020	28 provinces in China	2011–2012	6810	24	58	53	CESD	Composite	Age, sex, residence, marital status, baseline CES-D, education, smoking, alcohol use, obesity, HTN, DM, DLP, chronic kidney disease, inflammation
Piantella et al., 2021	London, England, U.K.	1997–1999	7610	132	56	3	General Health Questionnaire - Depression	CAD	Age, gender, smoking, BMI
Poole and Steptoe, 2018	England	2004	2472	120	63	51	CESD	CAD	Age, sex, ethnicity, cohabitation, wealth, smoking, SMI, alcohol use, regular physical activity, cognitive function, HTN
Poole and Jackowska., 2019	England	2014–2015	5034	72	66	55	CESD	MI, Other	Age, sex, relationship status, income, BMI, smoker, alcohol use, PA, HTN, sleep problems
Rantanen et al. 2020	“Harjavalta and Kokemäki, Finland	2005–2007	2522	96	58	56	Beck Depression Inventory	MI, PAD, angina, CAD	Age, gender, education, smoking, alcohol use, PA, HTN, DLP
Remch et al., 2018	New York City, USA	2012–2013	5971	48	51	17	PHQ	MI, CV events	Age, sex, BP, total cholesterol, BMI, tobacco use, respirator use
Vu et al., 2021	4 U.S. sites: Washington County, MD, Forsyth County, NC, Suburbs of Minneapolis, MN, and Jackson MS	2011–2013	6025	66	75	59	CES-D	HF	Age, sex, race, education, income, smoking, alcohol use, PA, DM, BMI, HR, eGFR
Yu et al., 2022	Guizhou province, China	2010–2012	7735	84	44	52	PHQ	MI	Age, sex, ethnicity, education, marriage, occupation, smoking, alcohol use, PA, history of T2DM, HTN, DLP, BMI
Zhu et al., 2022	28 provinces in China	2013–2014	9595	60	58	52	CES-D	CAD, HF, MI angina, Other	Age, gender, marital status, education, residency, smoking, alcohol use, HTN, DM, DLP, sleep duration
Anxiety
Feng et al., 2020	Nord-Trøndelag County, Norway	2006–2008	37402	96	53	57	HADS-Anxiety	AF	Age, sex, weight, height, smoking, occupation, marital status, physical activity, alcohol use, chronic disorders, metabolic components (glucose, BP, triglycerides, HDL, CRP)
Karlsen et al. 2020	Six U.S. sites (Birmingham, AL; Minneapolis, MN; Palo Alto, CA; Pittsburg, PA; Portland, OR, San Diego, CA	2000–2002	3095	180	76	0	GADS	MI, HF, Other	Age, education, ethnicity, smoking, BMI, PA, alcohol use, DM, BP, cholesterol, antidepressant use, anxiety
PTSD
Remch et al., 2018	New York City, U.S.	2012–2013	5971	48	51	17	PCL-C (civilian)	MI, Composite	Age, sex, BMI, use of a respirator, BP, total cholesterol, tobacco use
Scherrer et al., 2020	U.S.	2008–2012	1079	84	49	17	PCL	Composite	Age, race, gender, marital status, health insurance, depression, anxiety, sleep disorder, substance use, smoking, DM, HTN, HLD, obesity, duration of PTSD psychotherapy, antidepressant use
Psychosocial Stress
Graff et al., 2017	Denmark	2010	114337	48	Largest age group: 55–64 (22.6%)	54	PSS	AF	Physical and psychiatric comorbid conditions, SES, lifestyle factors
Santosa et al., 2021	21 countries	2001–2003	118706	122	50	59	Composite: PSS, Recent Adverse Life Events, Financial Stress	CAD	Age, sex, education, marital status, location, obesity, HTN, smoking, DM, family history of CVD, center
General Mental Health/Health-Related Quality of Life
Bonaccio et al., 2018	Molise region, Italy	2005	17102	60	53	53	Mental HRQoL	CAD	Age, sex, education, household income, occupational class, marital status, cancer, DM, HTN, HCL, psychological assessment, PA, BMI, diet, smoking, physical/metal health, CRP
Nilsson et al. 2020	County of Östergötland, Sweden	2003–2004	1001	156	57	50	SF-36 Mental Health Subscale	MI, Composite	Age, sex, and reporting at least one disease and/or neck/back pain
Phyo et al., 2021	Australia, U.S.	2010–2014	19106	56	74 (median)	56	Medical Outcomes Study, SF-12 Mental Component Scale	Composite	Age, sex, race and ethnicity, education, living situation, country, smoking, alcohol use, PA
Pinheiro et al., 2019	U.S.	2003–2007	22229	101	64	58	HRQoL, SF-12 Mental Component Summary score	MI, Composite	Age, sex, race, education, relationship status, access to care, income, health insurance, residence, DM, HTN, AF, medication use, LVH, BMI, cholesterol, hsCRP, eGFR, CKD
Wimmelman et al., 2021	Copenhagen, Denmark	2009–2011	6750	72	54	31	Satisfaction with Life Scale, SF-36 Vitality subscale	Composite	Age, sex, education, BMI, smoking, alcohol use, coronary calcium index, employment, social support

Notes. Duration of follow-up is reported as average rather than total. Abbreviations: AF, atrial fibrillation; BMI, body mass index; CAD, coronary artery disease; CES-D, Center for Epidemiologic Studies Depression Scale; CKD, chronic kidney disease; CVD, cardiovascular disease; DM, diabetes mellitus; DLP, dyslipidemia; eGFR, estimated glomerular filtration rate; GADS, Goldberg Anxiety and Depression Scales; HADS, Hospital Anxiety and Depression Scales; HF, heart failure; HCL, hypercholesterolemia; hsCRP, high sensitivity c-reactive protein; HRQoL, health-related quality of life; HTN, hypertension; HLD, hyperlipidemia; LVH, left ventricular hypertrophy; MI, myocardial infarction; PA, physical activity; PCL, PTSD Checklist; PHQ, Patient Health Questionnaire; PTSD, post-traumatic stress disorder; PSS, Perceived Stress Scale; SBP, systolic blood pressure; SES, socioeconomic status; SF-36, Short Form-36; SF-12, Short Form-12

The period of data collection ranged considerably, with the earliest baseline in 1974 and the most recent baseline in 2015. The average follow-up period was 98.3 ± 45.5 mo. Most investigations examined depressive symptoms (n=20) or general mental health symptoms (n=5), and a few examined symptoms of anxiety (n=2), PTSD (n=2), and stress (n=2). Of the studies that focused on depression, the Center for Epidemiological Studies Depression Scale was particularly common (n=12), and other measures consisted of the Patient Health Questionnaire (n=3), the Beck Depression Inventory (n=1), the Goldberg Anxiety and Depression Scales (n=1), the Hospital Anxiety and Depression Scales-Depression subscale (n=1), and the Depression and Exhaustion subscale of the von Zerssen symptom checklist. Anxiety was assessed with the Goldberg Anxiety and Depression Scales (n=1), and the Hospital Anxiety and Depression Scales-Depression subscale (n=1). Assessment of PTSD was with the PTSD Checklist-Civilian version (n=2). Stress was assessed with the Perceived Stress Scale (n=1) or a composite of that scale, major adverse life events, and financial stress in the past year (n=1). Finally, other general mental health screening was conducted with the SF-36 Mental Health Subscale (n=2), Mental Health Related Quality of Life Scale (n=2), SF-12 Mental Component Summary score (n=2), and the Satisfaction with Life Scale (n=1).

Outcomes consisted of a composite endpoint (n = 10), coronary artery disease (n = 9), myocardial infarction (n = 8), atrial fibrillation (n = 3), and heart failure (n = 6), although many studies included distinct tests of more than one outcome. Documentation of outcomes was most often based on a diagnosis or an event (eg, hospitalization; n = 15), and data were collected via self-report (n = 13), from a national register/database (n = 9), or by medical chart review (n = 6).

Meta-analytic Results

Out of the studies that met criteria and were reviewed, 15 studies that reported associations between baseline psychological distress and subsequent risk of CVD had sufficient data to include in the primary analysis. Overall, high symptoms of distress were associated with a 28% greater risk of CVD morbidity (95% CI, 1.18 – 1.39; Figure 2). Analyses of heterogeneity among the studies showed that effects were moderately heterogenous (Q[14]=28.99, P=.010, I²=48%). Study results were also summarized by distinct dimensions of psychological health (Figure 3). Evidence of publication bias was not observed, and relevant plots are depicted in SDC 1.

Figure 2.

Forest plot depicting analyses of psychological distress identified via brief screening and CVD morbidity and mortality, published 2017–2022 (n=15).

Figure 3.

Psychological distress and risk of incident CVD and related mortality, by type of measure.

A sensitivity analysis was conducted treating odds ratios (ORs) as equivalent to HR and including studies with CVD mortality as the outcome (overall, n=18; Figure 4). Like results from the primary analyses, high distress was associated with a 28% greater risk of CVD morbidity and mortality. Analyses of the studies showed that effects were heterogenous (Q[17]=54.54, P<.001, I²=63%).

Figure 4.

Forest plot depicting sensitivity analysis of psychological distress and CVD morbidity and mortality (n=18).

DISCUSSION

This systematized review consisted of 28 studies published from 2017 to 2022, which focused on self-reported psychological distress identified by brief screeners and incident CVD morbidity and mortality. Variability in methodology and statistical reporting resulted in 15 investigations available for the primary systematized meta-analysis. Overall, psychological distress (i.e., elevated symptoms of depression, anxiety, PTSD, stress, or worse mental health-related quality of life) was associated with a 28% increased risk of CVD morbidity or mortality. Although multiple dimensions of psychological health were combined for the primary analysis, the observed effects were in a similar direction and range as previous meta-analyses of distinct psychological dimensions, and which included psychiatric diagnoses (e.g., relative risks of 1.21–1.90^{9–11,13,16,38}). Our investigation re-affirms the importance of psychological distress in cardiovascular risk, and more prominently, suggests that psychological health captured with “brief and valid methods…that can then be documented within the medical records”¹⁹ may be sufficient to approximate the associated CVD risk and guide intervention.

Past reviews and meta-analyses of psychological health and CVD risk have targeted unidimensional psychological health, with the most recent versions published in 2016–2019.^7,9–11,14 However, as recently stated by Lloyd-Jones and colleagues, “Psychological health is multidimensional…a contextual driver of cardiovascular health.”¹⁹ Psychological disorders and different dimensions of psychological symptoms are also highly comorbid.³⁹ Understanding associations between one aspect of psychological health and one CVD outcome provides foundational evidence, but may complicate potential assessment processes, leading clinicians to assume that all dimensions must be assessed and making the evidence less digestible or actionable. For example, if there are separate, significant findings concerning the associations of depression, anxiety, and psychosocial stress in relation heart failure, there may be uncertainty about what to target if a patient’s symptoms are clinically elevated on all measures and with limited access to mental health providers. Given the complexity of lifestyle modification to improve CVD risk and the importance of psychological distress for adherence and maintenance of such changes,¹⁹ targeting multiple aspects of risk concordantly with psychological health using tailored strategies may be most successful.⁴⁰ In other words, meet the patient where they are starting CVD prevention from, and apply knowledge about their psychological health to tailor their prevention moving forward.

A clear gap from existing research is the limited evaluation of sex and gender differences. Compared to men, women have distinct experiences of trauma throughout the lifespan, unique sources of psychosocial stress, increased perceived stress during adulthood, and show different presentations of psychological distress, including a 2-fold greater lifetime prevalence of depression and some anxiety disorders.^41–44 These distinct vulnerabilities and presentations likely translate to sex differences in CVD risk, physiological mechanisms, CVD onset and presentation, all of which are integral to CVD primary prevention.^45–51 In a follow-up to the primary INTERHEART analyses, compared to men, women with moderate or high psychosocial distress showed a greater risk of MI (OR: 2.58 vs. 3.49).⁵² Additionally, women may be un- or undertreated for depression: among 1075 women participating in primary prevention and cardiac rehabilitation, 39% still reported depressive symptoms meeting the threshold for depression.⁵³ Despite these well-recognized distinctions and gaps in care, few cohort studies or meta-analyses of psychological health and CVD risk have reported effects separately by sex.^54,55 This delineation is required to understand potential subgroup differences in psychological risk and to mount a better clinical offense for men and women, respectively. A pioneer in this endeavor has been the European Society of Cardiology, which has already made significant efforts towards sex-specific CVD risk stratification and that which includes mental health.⁵⁶

The present results provide additional data substantiating the utility of integrated psychological care within CVD primary prevention and supporting research and clinical recommendations to extend this work. To begin, there is considerable scientific heterogeneity in evaluating and treating psychological contributions to CVD risk, including the variables used for statistical adjustment. Next, many studies do not statistically control for the use of psychotropic medications; this oversight is problematic as certain medication classes (i.e., selective serotonin reuptake inhibitors) may contribute to vascular risk⁵⁷ and have also been associated with a lower risk of recurrent acute coronary syndrome (ACS) in patients with post-ACS depression.⁵⁸ Related, open questions are which domains of psychological health are most important when evaluating CVD risk, which screening measures are preferable, and how to best combine brief assessments.¹⁹ To help bridge this evidentiary gap, more rigorous tests of the prospective associations between psychosocial stress, particularly assessed with the Perceived Stress Scale or validated measures of specific types of stress (e.g., financial, caregiving), and CVD risk are needed. Moreover, self-reported psychological distress may vary depending on one’s life circumstances, yet the reviewed studies only provide a one-time snapshot of distress and generally do not account for changes in distress over time or due to psychological treatment. While yearly screening for depression has been recommended for primary care,²² the frequency of psychological screening needed to monitor cardiovascular risk is unknown. As another area of focus, for decades there has been a call for more experimental evidence concerning behavioral and biological underpinnings of psychological-CVD associations and clinical trials of stepped care approaches (e.g., using the PHQ-9 to screen followed by a more comprehensive evaluation and treatment).¹⁰ This research is important, yet one should consider if future mechanistic studies will actually help individuals significantly more than receiving available, validated, self-report screeners and gold standard behavioral and pharmacological treatments.

There is a continued cost to delaying implementation of psychological distress screening in primary prevention, particularly given the close associations between psychological health and other lifestyle factors implicated in cardiovascular health, and more generally, a patient’s motivation, engagement, and adherence to treatment. Even if associations between psychological distress and incident CVD are not causal, meta-analytic evidence supports the utility of assessing psychological health, particularly given the high comorbidity between psychological disorders that have been linked to cardiovascular risk – anxiety, depression, and PTSD.³⁹ Screening psychological distress may be especially advantageous to mediate social determinants of health that systematically increase CVD risk among certain racial, ethnic, and socioeconomic groups; this approach can be readily employed in clinical-, community-, and population-based healthcare settings.¹⁹ Interestingly, the present effect of psychological distress was observed despite variability in study locales, demographics, and different approaches to psychological and cardiovascular health management. Possible mechanisms of action for the observed effect of psychological distress could be those associated with self-care. For example, individuals who are distressed may be less likely to care for themselves, adhere to medical advice or medications, and engage in riskier health behaviors, offering an added reason to screen for distress in prevention settings.

Alleviating psychological distress should be a healthcare priority at any stage in patient-centered cardiovascular care – whether primordial prevention, primary prevention among those with a family history of CVD or who are otherwise high risk, prehabilitation to prepare for non-urgent cardiac surgery, or secondary prevention after a major surgical intervention or cardiovascular event. Integrated primary care offers one promising clinical service delivery model, which includes psychological distress assessment and management, and can be adapted for CVD prevention.^59–62 Besides benefiting patients, this model could be financially advantageous for single-payer and managed care systems alike.^63,64 In the U.S., health policy should be informed by robust randomized clinical trials and links to meaningful benchmarks including quality of life and reduced healthcare expenditures; those trials to date concerning psychological risk and CVD lag millions of dollars behind investments in biomedical factors – and cannot catch up. To create better policy, preventive cardiology should draw from the excellent data supporting psychosocial care in cardiovascular rehabilitation and from other models of integrated care for noncommunicable chronic diseases, to invest more readily in interprofessional training and team-based care.^63,65 Thus, we are left to reflect on whether the current evidence is sufficient and if there is an ethical imperative to begin implementing that evidence and refining approaches over time.

Strengths and Limitations

Our study summarizes the most recent evidence concerning psychological distress assessed by self-reported screening and cardiovascular risk among adults. Still, there were limitations to this investigation. First, studies that assessed different psychological health dimensions were combined in the primary analysis, which is an uncommon statistical approach. However, we prioritized this approach given the high degree of comorbidity between the psychological disorders of focus, and a goal of broadly understanding the effect of psychological distress. Given the small number of studies assessing dimensions of psychological health besides depression, conclusions should be interpreted cautiously. Similarly, the duration of follow-up and outcomes were varied; additional research is needed to clarify whether the strength of the association between elevated psychological distress and CVD onset differs by cardiovascular outcome (e.g., acute coronary syndrome, coronary artery disease, peripheral vascular disease, versus heart failure). Second, there was considerable heterogeneity in the methods and statistical analyses from the included studies. For example, most studies did not report unadjusted effects, and many did not screen for pharmacological treatments of psychological health or sleep, which have long-term implications for cardiovascular health.^57,58 We only included studies that provided dichotomous data in the meta-analysis in order to prioritize clinical interpretability, but excluding continuous data may over or underestimate prognostic effects.⁶⁶ Third, there are recognized subgroup differences in psychological distress prevalence and CVD risk but most included studies did not present data by sex or race. Future investigations should do so to better understand associations at the population level and among distinct demographic groups; these data may encourage scientific replication and inform more tailored primary prevention initiatives. Fourth, although we excluded or statistically accounted for known CVD, it is possible that individuals with elevated psychological distress had subclinical or undiagnosed, and therefore, uncontrolled CVD. Fifth, studies with stroke or hypertension as primary outcomes or included in a composite outcome were excluded due to the differing etiology or clinical timelines involved with such vascular events or conditions, although other meta-analyses have shown similar associations between psychological distress and risk for incident stroke.⁶⁷ Sixth, with a spotlight on sleep in cardiovascular risk, and well-described comorbidity between psychological disorders and deficient sleep (e.g., insomnia),³⁹ sleep may be a more or less critical point of management than psychological distress. Seventh, methodologically, we did not include the Newcastle-Ottawa scale, which is often used to evaluate risk of bias in non-randomized studies. Risk of publication bias was otherwise formally assessed, but further investigation should be considered. Finally, although every effort was made to identify relevant studies, some records may have been inadvertently missed in the systematized review processes.

CONCLUSIONS

This meta-analysis included investigations of the prospective effect of psychological distress on CVD morbidity and mortality in which psychological health was assessed with brief, self-report measures only. Our results showed meaningful effects, reaffirming and extending evidence to account for psychological distress in CVD primary prevention, particularly using validated screeners. In short, psychological health screening can be useful for providers to elucidate CVD risk more comprehensively, and management of psychological distress may reduce CVD risk and increase quality of life. Unfortunately, screening for psychological health continues to be the exception rather than the norm in CVD prevention. More research concerning measures of stress and sex-specific associations will help, and best practices for the implementation of such screening and clinical decision support in cardiovascular medicine should be tested and delineated. Still, when merging the current data concerning psychological health and CVD risk with distress screening approaches from other clinical specialties, there may already be sufficient evidence to implement distress screening for primary cardiovascular prevention and doing so may be more beneficial and ethical than not screening.

Key Perspective.

What is novel?

• This meta-analysis of research published in the last 5 yr demonstrates that psychological distress evaluated with brief screening measures only is associated with a 28% greater risk of first-onset cardiovascular disease (CVD).

• These results are comparable to those in previous meta-analyses including data from formal psychological evaluations or medical record diagnoses, suggesting that screeners alone are sufficient for capturing the CVD risk associated with psychological distress.

What are the clinical and/or research implications?

• Using brief psychological screeners in clinical or community settings is feasible and helpful for early CVD risk stratification.

• Even without meeting criteria high psychological distress, patients may benefit from gold standard evidence-based interventions or additional supportive resources to aid CVD primary prevention.