Psychosocial distress and health status as risk factors for ten-year major adverse cardiac events and mortality in patients with non-obstructive coronary artery disease

Participants and procedure

The present study is part of the TweeSteden Mild Stenosis study (TWIST), an observational cohort to examine psychosocial risk factors for outcomes in patients with non-obstructive coronary artery disease ¹¹(ClinicalTrials.gov Id: NCT01788241). All participants who received a computed tomography (CT scan) or coronary angiography (CAG) were screened between January 2009 and March 2013, and eligible patients received information on the study. Inclusion criteria were having a CT scan with either a coronary artery calcium score ≥ lowest 10th percentile with mild atherosclerotic plaque, excluding those eligible for a subsequent CAG. Separately, all patients undergoing a CAG were examined for having visible wall irregularities without obstructive coronary occlusion (between 10-60% occlusion of coronary arteries). Exclusion criteria were having a history of obstructive coronary heart disease. In total 547 patients provided written informed consent. One person was additionally excluded. Coded paper questionnaires on sociodemographic, lifestyle, and psychological complaints were sent and received by postal mail at baseline, after 12, and 24 months. Electronic hospital record information was used for indicators of disease severity, medication, and comorbid conditions. The TWIST study received ethical approval in 2009 (METC Brabant, Protocol number: NL22258.008.08), with an update for screening for outcomes in 2021-2022 (ETZ: L1160.2020 > LP.2008.227 TWIST follow-up). Study data are stored in accordance with and conforming to good clinical practice and data protection guidelines.

Primary outcomes

Electronic hospital records were screened for outcomes by trained research assistants (LV, AB, ABE) between October 2021 and February 2022. Outcomes had previously been examined between October 2014 and April 2015¹³. New and previous events were double checked. Raw cardiovascular events with dates and information on the cause of the event at follow up were reported in the local Electronic Data Capture management system (ETZ Research Manager). Additional checks and interpretation of unclear outcomes were done. Coded variables were exported and categorized into primary and secondary outcomes.

The primary outcome was the occurrence of a major adverse cardiac event with mortality (referred to as MACE) during follow-up, including: cardiovascular mortality, cardiac events (PCI, CABG, (N)STEMI or cardiac arrest, heart failure), or non-cardiac mortality. The date of each event was coded. In case of multiple events, hierarchical choices were made; cardiovascular mortality was chosen over a cardiac event, and any cardiac event was chosen over non-cardiac mortality. For descriptive purpose, mortality causes were further subcategorized¹⁴. Time between the index coronary angiography or CT-scan at inclusion and MACE or censored date was calculated in years. In the absence of an event, patients were censored at the final screening date (February 3, 2022). The censoring date of patients lost to follow-up was set at 12 months after the last moment of contact¹⁵.

In the Clinicaltrials.gov preregistration done in 2009 we previously defined MACE as ‘the occurence of a recurrent coronary angiography, emergency hospitalization (for cardiac reasons), myocardial infarction, percutaneous coronary intervention (PCI) or coronary artery bypass graft surgery (CABG), mortality (cardiac/noncardiac)’. We choose to deviate from this preregistered definition due to having previously reported on ED visit or repeat testing¹³, to be able to distinguish between cardiac and non-cardiac mortality in the process of disease development, and the previous absence of heart failure as a valid disease progression indicator. In our opinion the current MACE definition better reflects cardiac disease progression, extended with non-cardiac mortality.

Psychosocial distress

Well validated baseline questionnaires were used¹². Type D personality was measured using the DS14⁵, measuring negative affectivity (NA) and social inhibition (SI), each with seven items on a five point Likert scale. Individuals were considered to have a Type D personality when they scored 10 or higher on both the NA and SI total score. For the analyses the NA and SI total scores were transformed into z-scores and multiplied to obtain their interaction Z(NA)*Z(SI). Both the z-scores and their interaction were included together as predictors in outcome analysis, with the interaction term testing the Type D personality effect⁶.

Depressive symptoms were assessed with the Beck Depression Inventory (BDI), using a sumscore of the 21 items on a 0-3 range, as well as with the cognitive-affective and somatic subscales (examined together). Separately, the 14 item, 0-3 range hospital anxiety and depression scale (HADS) was used to measure depressive symptoms (HADS-D) and anxiety (HADS-A).

Hostility was operationalized using the Cook-Medley Hostility scale (CMH), containing 27 items on a true/false scale¹⁶. The sumscore was used as a measure of hostility, with a higher score indicating a more hostile personality style. Positive affect was measured with 10 items on a 0-4 Likert scale from the Global Mood Scale [GMS-PA], with a higher sum score indicating a more positive mood.

Health status and quality of life

Fatigue was measured using the 10-item fatigue assessment scale [FAS], with a higher score indicating more fatigue. A modified version derived from the Seattle Angina Questionnaire was used to measure disease-specific heart-related complaints in five dimensions^11,17. A higher score indicates better health and fewer complaints, as well as the generic Short Form 12, derived from the SF36 was used to evaluate physical (physical component summary (PCS)), and mental health (mental component summary (MCS)).

Descriptive and potential confounding factors

Statistical analysis

Descriptive statistics stratified by presence or absence of MACE were reported and compared between these two groups using Chi-square test for categorical variables, and One-Way ANOVA for continuous scores or a non-parametric equivalent. The annualized event rate (AER) for MACE was calculated dividing the number of MACE events by the sum of total person-years of follow-up (either MACE or censored)*1000, expressed per 1,000 person-years. For comparison reasons, the AER was divided by the average follow-up time, named the ‘Annual MACE rate’¹⁸.

Cox proportional hazard analyses were performed using MACE as the primary outcome, and time since the baseline CAG or CT in years. An explorative model with sex and age was examined, with sex depicted in a Kaplan-Meier curve. The psychosocial and health status variables of interest were each entered separately in an analysis, adjusted for age and sex. In the complete adjusted model the CAG versus CT group and hypertension were added as indicators of disease severity to age and sex, and indicators of lifestyle (BMI and being physically active) since these tend to covary with psychosocial factors¹⁹.

Sensitivity analyses were done stratified by sex, as well as the interaction of sex-by-psychosocial factor adjusted for age²⁰. The Type D personality continuous score was explored, by adding the quadratic negative affectivity and social inhibition terms²¹. Sensitivity analyses for MACE were run with ANOCA and INOCA (reference ‘other’) group as an indicator of disease severity, adjusted for age and sex, as well as excluding non-cardiac mortality cases.

Comparability of MACE outcomes

Other studies which examined outcomes in patients with NOCAD or INOCA often included patients with a broader range of NOCAD. Huang and colleagues examined all-cause mortality after 9 years, and included both patients without any visible wall irregularities, as well as patients with either 1, 2, or 3 vessel NOCAD. The AER for all-cause mortality in their 1,2, or 3 vessel disease NOCAD group combined was 1.3% (113 cases/989 patients/9 years, Table 2 findings), with a higher risk for age, men, and 3-vessel disease²², which is in line with the findings of our study (1.1%). A higher risk of adverse outcomes for the CAG group compared to the CT group is in line with the meta-analysis of Wang and colleagues²³. Sedlak and colleagues observed that in the first year after inclusion, women had a higher risk of MACE compared to men, though no sex differences were present from year 1-3²⁴. Visual inspection of the Kaplan-Meier curve stratified for sex in the present study shows that sex differences became apparent after four years. Herscovici summarized annual MACE rates in studies, showing an annual MACE rate between 0.9%-2.4%¹⁸, comparable with the 2% observed in the present study. The WISE study in women with CAG-detected minimal CAD (≥20% but <50% stenosis) reported an all-cause mortality rate of 1.7%, and cardiac mortality was 1.1% AER over 10 years ²⁵. Event rates for serious adverse outcomes were recently estimated to be about 3% per year¹⁰. Notably, each study has slightly different inclusion criteria, as well as different criteria for MACE. A future update of a meta-analysis of risk of adverse outcomes in patients with NOCAD or INOCA including examining sex differences remains to be done.

Psychosocial factors and health status associated with MACE

Most pronounced findings of the present study were the finding that a lower positive mood, higher fatigue, and lower physical limitation were related to a higher risk of MACE in patients with NOCAD. Depressive symptoms were no longer associated with MACE after adjustment for confounders, which may be due to lack of sufficient statistical power to reliably detect an association. At the same time, there was an absence of anxiety, Type D personality, and hostility with MACE, nor were indicators of health status and angina associated with MACE. Fatigue is a core component of vital exhaustion, which is well observed risk factor for cardiovascular events⁸. In terms of mechanisms, an autonomic imbalance has been hypothesized for the association between exhaustion and cardiovascular risk²⁶. Likewise positive psychological wellbeing is related to reduced prognostic cardiovascular events in people with heart disease²⁷. The physical limitation scale was a modified version of the SAQ, which in our study represents limitations due to broader defined ‘cardiac problems’ rather than due to ‘chest pain, chest tightness, or angina’. For physical limitation, a poor (0-24 range), fair (25-49 range), and even good (50-74 range) score on the physical limitation scale of the SAQ is associated with mortality²⁸. In the present study, the health status subscale ‘physical limitations’ was on average good to excellent, and still predictive of a reduced risk for MACE. These factors show a protective association of experiencing less physical limitations, less fatigue, and better positive mood. This provides a potential starting point for interventions, e.g. aimed at enhancing positive mood²⁹, or aimed to develop a greater psychological flexibility rather than addressing negative mental health states, for example by using ‘acceptance and commitment therapy’ (ACT)³⁰.

Psychosocial factors, especially depressive symptoms, but also anxiety, Type D personality, and hostility have consistently been associated with MACE^3,31–33, which is acknowledged in current guidelines^1,2. A difference between these studies and the present cohort is that these studies predominantly followed patients with obstructive CAD, who had either acute coronary syndrome, a myocardial infarction, or who received invasive treatment such as PCI or CABG. In line with obstructive CAD, those studied populations usually have a majority (70%-80%) of men included. An exception is the WISE cohort study which observed somatic depressive symptoms, but not anxiety by itself to predicted adverse outcomes in women³⁴. However further stratification for obstructive versus nonobstructive origin was not examined, nor were other psychosocial factors such as hostility described for outcomes. In the present study, adverse mental health such as anxiety, Type D personality, and hostility were not predictive of MACE, and findings were unclear for depressive symptoms. The differences in findings may be due to several factors: differences in pathophysiological mechanisms in nonobstructive versus obstructive CAD¹⁰, which are also associated with psychosocial distress³⁵, differences in health behavior, age, and sex differences in people with obstructive versus nonobstructive CAD³⁶. It is relevant to examine the risk of psychosocial factors for outcomes further stratified for obstructive versus nonobstructive IHD.

Sex differences in psychosocial factors and outcomes

A consistent secondary finding, in addition to age being a risk for MACE, is that women presented with a lower risk for MACE compared to men. Though no significant sex interaction was observed, sex-stratified findings showed that, after adjustment, only in men, fatigue and physical limitations were associated with MACE. A notable limitation is that power is reduced in these secondary analyses. The absence of an interaction of psychosocial factors with sex, as well as the absence of an association of negative mental health states with MACE is relevant information in terms of gender bias. NOCAD remains underrecognized in general practice^37,38. Women with NOCAD, more often than men, are often dismissed from specialized care³⁷, their symptoms being ascribed to mental health state such as anxiety³⁹, and they experience not being taken seriously in their symptom presentation³⁸. One of the implications of the present findings is that though psychosocial care could and should be offered for patients to deal with their disease burden³⁸, an intervention aimed to reduce depressive symptoms or anxiety is unlikely to affect their prognosis.

Limitations

A limitation of the present study is that it is a single center cohort study, which could limit the generalizability of the findings towards patients not receiving appropriate care, other countries or ethnic subgroups which are currently underrepresented. Though psychosocial factors were measured at baseline, 12, and 24 months after inclusion, no potential changes over time for the psychosocial factors were taken into account in the present study. The follow-up time is a strength of the present study, though the power in secondary analysis was likely insufficient to reliably detect event risk factors within subgroups. The present study is a combination of patients being included via CAG (70%) as well as CT scan (30%), which created a more heterogeneous group of patients. Moreover, we did not solely focus on patients with ischemia or angina as an inclusion criteria, but included all consecutive patients who had nonobstructive coronary arteries, though no significant differences between these three groups emerged for MACE. A limitation is the absence of more robust data regarding microvascular dysfunction, which was not part of clinical practice between 2009-2013. A limited number of fractional flow reserve (FFR) measurements were performed during CAG, which was upcoming at that time, these were all within normal range.

In conclusion, in patients with CAG or CT scan detected nonobstructive CAD, after a follow-up of 10 years, in total 19% MACE events occurred, with a lower risk for women. A higher positive mood and experiencing less physical limitations were related to a lower risk for MACE, whereas more fatigue was related to a higher MACE risk, without significant sex interactions. No consistent significant risk for depressive symptoms, anxiety, Type D personality, hostility, mental and physical health status, or angina with MACE was observed. Positive mood, fatigue and physical limitations may benefit from specific interventions, and though reducing psychosocial distress is an intervention aim by itself, it is less likely to affect MACE in patients with NOCAD.