Association Between Anger and Mental Stress-induced MI
The mean and median age was 50 years, half of patients were women, and 60% were non-white (Table I). Almost half of the patients had an ST-elevation MI (45%); and after the index MI, 75% underwent percutaneous coronary interventions and 11% coronary artery bypass surgery. Psychosocial factors were common, with almost one-third of patients (32%) reporting income below poverty and 37% having significant depressive symptoms (BDI-II score >13).
Myocardial perfusion could not be quantified in 5 subjects due to poor image quality. For mental stress, the mean and SD for the SDS for ischemia quantification was 2.31 ± 2.69 (range 0–13). For physical stress, the mean SDS was 2.74 ± 3.24 (range 0–13). Based on a predefined cut-off point of SDS ≥3 for mental stress and ≥4 for exercise/pharmacologic stress, 36 patients (39%) had mental stress ischemia, 32 (34%) had physical stress ischemia, and 18 (20%) had both mental and physical stress ischemia. Seven (9%) subjects had angina during exercise, and 18 (22%) had ST-segment depression with exercise stress. Descriptive characteristics of anger subscales are shown in Table II.
Age was inversely associated with state and trait anger, indicating that the older the age, the less severe the anger scores (online Appendix Supplementary Table I). There were no significant differences based on other demographic factors. Current smokers, hypertensive subjects, and subjects reporting ≥1 angina episode per month tended to have higher anger-out scores, but the levels of other anger dimensions were similar. All anger subscales were positively associated with higher depressive and anxiety symptoms.
Heart rate, systolic and diastolic blood pressure, and rate-pressure product (heart rate times systolic blood pressure) significantly increased with mental stress (online Appendix Supplementary Table II) and with exercise/pharmacologic stress. None of these changes was significantly associated with anger dimensions (data not shown). Changes in subjective ratings of distress, nervousness, fearfulness, and anxiety with mental stress were also not related to anger, but the change in subjective ratings of anger was weakly but significantly associated with trait anger (regression coefficient 0.05; 95% CI 0.01–0.09). None of the above subjective ratings was found to be associated with ischemia during mental stress.
State anger, trait anger, and anger expression out were all significantly associated with the SDS with mental stress in unadjusted analysis, denoting more ischemia (Table III). After adjustment for age, sex, race, smoking status, Gensini score, and depression and anxiety symptoms, both state and trait anger remained significantly associated with the SDS with mental stress. Each incremental IQR increase in state-anger score (corresponding to 3 score points) was associated with 0.36 U–adjusted increase in SDS (95% CI 0.14–0.59); the corresponding association for trait anger was 0.95 (95% CI 0.21–1.69) per IQR increase (corresponding to 6 points). These associations translated into 2.1% increased myocardium ischemic involvement with each IQR progressively higher state-anger score (95% CI 1.0%-3.2%) and 5.4% increased myocardial ischemic involvement with each IQR higher trait-anger score (95% CI 1.8%-8.9%). For both state and trait anger, the SDS was higher for higher levels of scale item means (Figure)). For state anger, being on average moderately angry or very angry was associated with approximately 4 times higher SDS compared with the 2 lower categories. For trait anger, there was a gradual increase in SDS going from 1 (almost never angry), to 4 (almost always angry). Anger out was no longer significantly associated with the SDS after adjustment for depression and anxiety symptoms. No association was found for the other anger dimensions. No significant sex or age interactions were found.
(Enlarge Image)
Figure.
Mental stress SDSs according to state and trait anger levels. Shown are mental-stress mean SDS and SEs according to state and trait anger score item mean levels. Numbers on bars indicate number of patients. For state anger, categories 3 and 4 were collapsed due to limited sample size.
Using a similar analytic strategy, we found that none of the anger dimensions was significantly associated with the SDS during exercise or pharmacologic stress (online Appendix Supplementary Table III).
To rule out the possibility that SPECT imaging artifacts may have influenced our results, we performed a sensitivity analysis after excluding 9 subjects with significant artifacts identified through a systematic review of all SPECT scans by an experienced cardiologist. Such exclusion did not change the association between anger subscales and both mental and physical stress SDS (online Appendix Supplementary Table IV).
Results
Study Sample
The mean and median age was 50 years, half of patients were women, and 60% were non-white (Table I). Almost half of the patients had an ST-elevation MI (45%); and after the index MI, 75% underwent percutaneous coronary interventions and 11% coronary artery bypass surgery. Psychosocial factors were common, with almost one-third of patients (32%) reporting income below poverty and 37% having significant depressive symptoms (BDI-II score >13).
Myocardial perfusion could not be quantified in 5 subjects due to poor image quality. For mental stress, the mean and SD for the SDS for ischemia quantification was 2.31 ± 2.69 (range 0–13). For physical stress, the mean SDS was 2.74 ± 3.24 (range 0–13). Based on a predefined cut-off point of SDS ≥3 for mental stress and ≥4 for exercise/pharmacologic stress, 36 patients (39%) had mental stress ischemia, 32 (34%) had physical stress ischemia, and 18 (20%) had both mental and physical stress ischemia. Seven (9%) subjects had angina during exercise, and 18 (22%) had ST-segment depression with exercise stress. Descriptive characteristics of anger subscales are shown in Table II.
Correlates of Anger Subscales
Age was inversely associated with state and trait anger, indicating that the older the age, the less severe the anger scores (online Appendix Supplementary Table I). There were no significant differences based on other demographic factors. Current smokers, hypertensive subjects, and subjects reporting ≥1 angina episode per month tended to have higher anger-out scores, but the levels of other anger dimensions were similar. All anger subscales were positively associated with higher depressive and anxiety symptoms.
Changes in Hemodynamic Measures and Subjective Distress With Mental Stress
Heart rate, systolic and diastolic blood pressure, and rate-pressure product (heart rate times systolic blood pressure) significantly increased with mental stress (online Appendix Supplementary Table II) and with exercise/pharmacologic stress. None of these changes was significantly associated with anger dimensions (data not shown). Changes in subjective ratings of distress, nervousness, fearfulness, and anxiety with mental stress were also not related to anger, but the change in subjective ratings of anger was weakly but significantly associated with trait anger (regression coefficient 0.05; 95% CI 0.01–0.09). None of the above subjective ratings was found to be associated with ischemia during mental stress.
Mental Stress–Induced Myocardial Ischemia and Anger
State anger, trait anger, and anger expression out were all significantly associated with the SDS with mental stress in unadjusted analysis, denoting more ischemia (Table III). After adjustment for age, sex, race, smoking status, Gensini score, and depression and anxiety symptoms, both state and trait anger remained significantly associated with the SDS with mental stress. Each incremental IQR increase in state-anger score (corresponding to 3 score points) was associated with 0.36 U–adjusted increase in SDS (95% CI 0.14–0.59); the corresponding association for trait anger was 0.95 (95% CI 0.21–1.69) per IQR increase (corresponding to 6 points). These associations translated into 2.1% increased myocardium ischemic involvement with each IQR progressively higher state-anger score (95% CI 1.0%-3.2%) and 5.4% increased myocardial ischemic involvement with each IQR higher trait-anger score (95% CI 1.8%-8.9%). For both state and trait anger, the SDS was higher for higher levels of scale item means (Figure)). For state anger, being on average moderately angry or very angry was associated with approximately 4 times higher SDS compared with the 2 lower categories. For trait anger, there was a gradual increase in SDS going from 1 (almost never angry), to 4 (almost always angry). Anger out was no longer significantly associated with the SDS after adjustment for depression and anxiety symptoms. No association was found for the other anger dimensions. No significant sex or age interactions were found.
(Enlarge Image)
Figure.
Mental stress SDSs according to state and trait anger levels. Shown are mental-stress mean SDS and SEs according to state and trait anger score item mean levels. Numbers on bars indicate number of patients. For state anger, categories 3 and 4 were collapsed due to limited sample size.
Using a similar analytic strategy, we found that none of the anger dimensions was significantly associated with the SDS during exercise or pharmacologic stress (online Appendix Supplementary Table III).
To rule out the possibility that SPECT imaging artifacts may have influenced our results, we performed a sensitivity analysis after excluding 9 subjects with significant artifacts identified through a systematic review of all SPECT scans by an experienced cardiologist. Such exclusion did not change the association between anger subscales and both mental and physical stress SDS (online Appendix Supplementary Table IV).