Although attention has recently been focused on the role of psychosocial factors in patients with cardiovascular disease (CVD), the factors that have the greatest influence on prognosis have not yet been elucidated. The aim of this study was to evaluate the effects of depression, anxiety, and anger on the prognosis of patients with CVD. Four hundred fourteen consecutive patients hospitalized with CVD were prospectively enrolled. Depression was evaluated using the Patient Health Questionnaire, anxiety using the Generalized Anxiety Disorder Questionnaire, and anger using the Spielberger Trait Anger Scale. Cox proportional-hazards regression was used to examine the individual effects of depression, anxiety, and anger on a combined primary end point of cardiac death or cardiac hospitalization and on a combined secondary end point of all-cause death or hospitalization during follow-up (median 14.2 months). Multivariate analysis showed that depression was a significant risk factor for cardiovascular hospitalization or death after adjusting for cardiac risk factors and other psychosocial factors (hazard ratio 2.62, p = 0.02), whereas anxiety was not significantly associated with cardiovascular hospitalization or death after adjustment (hazard ratio 2.35, p = 0.10). Anger was associated with a low rate of cardiovascular hospitalization or death (hazard ratio 0.34, p <0.01). In conclusion, depression in hospitalized patients with CVD is a stronger independent risk factor for adverse cardiac events than either anxiety or anger. Anger may help prevent adverse outcomes. Routine screening for depression should therefore be performed in patients with CVD, and the potential effects of anger in clinical practice should be reconsidered.
Attention has recently been focused on the role of psychosocial factors, including depression, anxiety, and anger, on the outcomes of patients with cardiovascular disease (CVD). Depression in particular has been shown to be a strong predictor of adverse short- and long-term outcomes in patients with CVD. Although anxiety and anger have also been shown to be strongly associated with adverse cardiac events in several studies, a 2008 American Heart Association science advisory recommended routine depression screening for all patients with coronary heart disease, but not anxiety or anger screening. Few studies have examined what psychosocial factors are most important in predicting outcomes for these patients. Therefore, the aims of the present study were to evaluate the incidence of depression, anxiety, and anger in patients with CVD; to investigate their relations with prognosis; and to compare their effects on various adverse patient outcomes.
Methods
For this prospective observational follow-up study, we recruited consecutive patients with CVD discharged from Nippon Medical School Hospital (Tokyo, Japan) from June 2009 to June 2011. Patients who were still enrolled in the trial were followed up until December 2011. We considered CVD to include acute myocardial infarction, unstable angina, stable angina, heart failure, arrhythmia, and peripheral artery disease. Patients who died before discharge, those who had dementia, and those who were not able to complete the questionnaires or give informed consent were excluded. A final sample size of 414 subjects was available for analysis. Clinical information and medical history, including blood examination and echocardiography, were obtained from medical records on a yearly basis during the follow-up period. Patients were also followed up by telephone interview to determine whether they had been hospitalized or if they had died during the past year. This study was approved by the research ethics committee of Nippon Medical School, and previous written informed consent was obtained from all participants.
The primary outcome was defined as the time from initial hospital discharge to cardiovascular hospitalization or cardiovascular death (whichever occurred first) within the follow-up period (6 to 30 months depending on enrollment date). Cardiovascular death was defined as death due to myocardial infarction, cerebral infarction, other vascular causes, heart failure, or documented sudden cardiac death. Cardiovascular hospitalizations included those related to myocardial infarction, heart failure, arrhythmia, ischemic stroke, transient ischemic attack, percutaneous coronary intervention, endovascular treatment, or cardiac surgery. To develop a fuller understanding of the relation between psychosocial factors and clinical outcomes, we also noted all-cause hospitalization or all-cause death as a secondary composite end point. The left ventricular ejection fraction (LVEF) was determined using 2-dimensional echocardiography. We defined a low LVEF as being lower than a cutoff point of 40%.
All patients completed 3 self-report questionnaires to assess depression, anxiety, and anger. Depression was determined using the Patient Health Questionnaire (PHQ-9; score range 0 to 27), anxiety using the Generalized Anxiety Disorder Questionnaire (GAD-7; score range 0 to 21), and anger using the Spielberger Trait Anger Scale (STAS; score range 10 to 40). The sensitivity and specificity of the Japanese version of the PHQ-9 are 84% and 95%, respectively, and those of the Japanese version of the GAD-7 are 88% and 82%, respectively. The Japanese version of the STAS is an acceptable scale for evaluating anger expression. We defined depression as a PHQ-9 score ≥10, anxiety disorder as a GAD-7 score ≥10, and anger as an STAS score ≥22. All patients received these psychological questionnaires a few days before hospital discharge to minimize the effects of their specific physical condition.
Continuous variables were tested for normal distribution using Shapiro-Wilk tests. Normally distributed continuous variables are expressed as mean ± SD, nonparametrically distributed variables as median (interquartile range), and categorical variables as frequency (percentage). Baseline clinical data were compared between patients with and without depression, anxiety, and anger using Student’s t tests and Mann-Whitney U tests. Categorical variables were analyzed using chi-square tests. Time-to-event outcomes using the Kaplan-Meier method were assessed using log-rank tests. Depression, anxiety, anger, a low LVEF (<40%), old age (≥75 years), low glomerular filtration rate (estimated glomerular filtration rate measured by the Modification of Diet in Renal Disease [MDRD] formula <60 ml/min/1.73 m 2 ), and a diagnosis of coronary artery disease, heart failure, arrhythmia, or other diseases had planned inclusion in the original models. To assess the robustness of the planned models, other potential factors (diagnosis at admission, including gender, body mass index, low-density lipoprotein cholesterol level, high-density lipoprotein cholesterol level, triglyceride level, glycosylated hemoglobin level, B-type natriuretic peptide level, previous myocardial infarction, hypertension, diabetes mellitus, dyslipidemia, and the use of β blockers, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, calcium channel blockers, statins, and diuretics) were eligible for entry into the models by stepwise selection (the significance level for entry into the model was 0.10). Finally, multivariate analysis using the Cox proportional-hazards model was performed to assess the relations between an adverse outcome (primary or secondary end point) and depression, anxiety, anger, old age, a low LVEF, low glomerular filtration rate, and high B-type natriuretic peptide level (>200 pg/ml) in the modified model. JMP version 9.0.3 (SAS Institute Inc., Cary, North Carolina) was used for all analyses.
Results
Demographic and clinical characteristics of the study sample are listed in Table 1 . The values of these characteristics and the assessments of 3 self-report questionnaires were collected at discharge. The mean patient age at baseline was 64.9 years (range 17 to 90). The gender ratio (men/women) was 2.4:1.
| Variable | All (n = 414) | Depression | Anxiety | Anger | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Yes (n = 46) | No (n = 368) | p Value | Yes (n = 25) | No (n = 385) | p Value | Yes (n = 128) | No (n = 280) | p Value | ||
| Age (yrs) | 64.9 ± 13.1 | 63.9 ± 13.2 | 65.0 ± 12.0 | 0.58 | 63.1 ± 13.0 | 65.1 ± 13.1 | 0.4 | 61.6 ± 12.4 | 66.4 ± 13.1 | <0.01 ∗ |
| Women (%) | 29.7 | 23.9 | 30.4 | 0.36 | 24.0 | 29.8 | 0.5 | 22.6 | 32.8 | 0.03 ∗ |
| Body mass index (kg/m 2 ) | 23.3 ± 4.3 | 21.5 ± 7.0 | 23.5 ± 3.9 | <0.01 ∗ | 23.6 ± 7.8 | 23.3 ± 4.1 | 0.7 | 23.8 ± 4.3 | 23.1 ± 4.4 | 0.20 |
| B-type natriuretic peptide (pg/dl) | 194 ± 297 | 189 ± 307 | 194 ± 296 | 0.91 | 163 ± 206 | 197 ± 302 | 0.60 | 165 ± 295 | 210 ± 299 | 0.17 |
| Glomerular filtration rate (ml/min/1.73 m 2 ) | 63.7 ± 23.1 | 59.6 ± 25.7 | 64.2 ± 22.8 | 0.20 | 62.9 ± 23.9 | 63.7 ± 23.2 | 0.87 | 65.9 ± 21.7 | 62.5 ± 23.8 | 0.16 |
| Low-density lipoprotein cholesterol (mg/dl) | 92.1 ± 28.3 | 83.4 ± 28.3 | 90.0 ± 28.3 | 0.06 | 98.4 ± 32.2 | 91.7 ± 28.2 | 0.26 | 91.1 ± 27.4 | 92.7 ± 29.0 | 0.59 |
| High-density lipoprotein cholesterol (mg/dl) | 45.3 ± 14.4 | 43.4 ± 17.1 | 45.2 ± 14.1 | 0.91 | 48.0 ± 18.0 | 45.1 ± 14.2 | 0.29 | 44.6 ± 14.2 | 45.7 ± 14.6 | 0.49 |
| Triglycerides (mg/dl) | 116 ± 59 | 123 ± 59 | 116 ± 59 | 0.72 | 120 ± 58 | 116 ± 60 | 0.74 | 114 ± 56 | 117 ± 60 | 0.69 |
| Glycosylated hemoglobin (%) | 5.7 ± 1.0 | 5.6 ± 0.7 | 5.7 ± 1.0 | 0.72 | 5.7 ± 1.4 | 5.7 ± 1.0 | 0.91 | 5.80 ± 1.0 | 5.69 ± 0.9 | 0.31 |
| LVEF (%) | 54.3 ± 17.5 | 52.6 ± 19.5 | 54.5 ± 17.3 | 0.49 | 53.3 ± 22.2 | 54.2 ± 17.2 | 0.82 | 54.2 ± 17.8 | 54.1 ± 17.5 | 0.97 |
| Previous myocardial infarction | 41.8% | 47.7% | 41.2% | 0.42 | 30.7% | 42.7% | 0.30 | 63.4% | 38.3% | 0.06 |
| Hypertension | 73.8% | 83.7% | 72.6% | 0.24 | 73.9% | 73.5% | 0.97 | 77.8% | 71.0% | 0.28 |
| Diabetes mellitus | 38.4% | 41.8% | 38.0% | 0.65 | 56.5% | 36.8% | 0.08 | 43.4% | 35.4% | 0.17 |
| Dyslipidemia (%) | 61.0 | 58.1 | 61.4 | 0.67 | 54.5 | 61.3 | 0.52 | 62.8 | 59.7 | 0.57 |
| β-Blockers (%) | 65.3 | 66.6 | 68.5 | 0.80 | 68.1 | 68.5 | 0.97 | 71.3 | 66.9 | 0.39 |
| Angiotensin-converting enzyme inhibitors or angiotensin receptor blockers (%) | 75.7 | 71.4 | 76.2 | 0.49 | 50.0 | 77.0 | <0.01 ∗ | 82.7 | 72.1 | 0.02 ∗ |
| Calcium channel blockers (%) | 29.8 | 33.3 | 29.3 | 0.59 | 40.9 | 28.6 | 0.22 | 27.0 | 30.6 | 0.47 |
| Statins (%) | 63.5 | 64.2 | 63.5 | 0.92 | 54.5 | 64.0 | 0.37 | 68.8 | 60.1 | 0.10 |
| Diuretics (%) | 41.5 | 42.8 | 41.4 | 0.85 | 36.3 | 42.3 | 0.58 | 34.4 | 45.6 | 0.03 |
Of the 414 patients, 54% (n = 226) were hospitalized for coronary artery disease, 29% (n = 121) for heart failure, and 10% (n = 44) for arrhythmia. Patients with arrhythmia experienced anxiety more often than patients with coronary artery disease and heart failure (9.1% vs 5.8% vs 4.3%, p <0.05), but those with coronary artery disease experienced anger more often than those with heart failure and arrhythmia (35.0% vs 21.4% vs 29.5%, p <0.05). The prevalence of depression in patients with coronary artery disease, heart failure, and arrhythmia was 11.1%, 10.7%, and 9.1%, respectively, and the differences were not statistically significant. At baseline, the mean PHQ-9 score was 3.8 ± 4.2 (range 0 to 22), the mean GAD-7 score was 3.0 ± 3.8 (range 0 to 27), and the mean STAS score was 19.3 ± 5.6 (range 10 to 39).
The mean follow-up time was 14.2 months, at which time 6 patients (1.4%) had died and 72 (17.3%) had been hospitalized at least once; of these 72 patients, 55 were hospitalized because of CVD. Multivariate Cox proportional-hazards regression analysis revealed that the LVEF, B-type natriuretic peptide level, and depression were all significantly associated with increased cardiovascular hospitalization or mortality ( Table 2 ). In this same model, anger was associated with a low rate of cardiovascular hospitalization or mortality. However, no significant association was observed between anxiety and this combined outcome. These analyses indicate that depression remained an independent predictor of incident CVD after adjusting for anxiety and anger. Anger, the LVEF, and diagnosis were significantly associated with the combined outcome of all-cause hospitalization or death ( Table 2 ), as was depression. Secondary analyses showed that no additional variables except B-type natriuretic peptide were added to the planned model through stepwise selection, confirming the robustness of the planned model. Figure demonstrates the rates of cardiovascular hospitalization or death according to each psychosocial symptom during 2-year follow-up ( Figure 1 ).
