High trait anger is linked to adverse cardiovascular outcomes. A potential antidote to the cardiotoxic influence of anger is trait forgiveness (TF), as it has shown associations with improved blood pressure (BP) and cardiovagal tone regulation in cardiac patients. However, it has yet to be determined if anger and forgiveness independently predict cardiovascular parameters. Trait anger (State-Trait Anger Expression Inventory-2) and TF (Tendency to Forgive Scale) were evaluated in 308 (M = 21.11years ± SD = 2.52) healthy female volunteers allocated to 3 related, yet distinct, studies. Hierarchical multiple regressions tested the incremental contribution of TF after accounting for anger. Study 1 assessed autonomic modulation through beat-to-beat BP and spectral analysis to examine sympathovagal balance and baroreflex functioning. Study 2 used tonometry and pulse wave analysis for aortic hemodynamics. Study 3 assessed 24-hour ambulatory BP and ambulatory arterial stiffness index. Hierarchical models demonstrated that anger was significantly associated with increased sympathovagal tone, increased hemodynamic indices, high ambulatory BPs, and attenuated BP variability and baroreflex. In contrast, TF was associated with more favorable hemodynamic effects (i.e., decreased ventricular work and myocardial oxygen consumption). In conclusion, these results demonstrate divergent cardiovascular effects of anger and forgiveness, such that anger is associated with a more cardiotoxic autonomic and hemodynamic profile, whereas TF is associated with a more cardioprotective profile. These findings suggest that interventions aimed at decreasing anger while increasing forgiveness may be clinically relevant.
Considerable attention has been given to the relation between anger and increased cardiac risk, which is relevant to both healthy and cardiac patients. For example, research indicates anger to increase the risk of coronary heart disease among initially healthy patients and to lead to poorer prognosis for patients with heart disease. Although the mechanism linking anger to increased cardiovascular risk is not well understood, impaired cardiovascular autonomic modulation and increased ventricular workload may be implicated. A potential antidote to the cardiotoxic influence of anger and hostility may be the cardioprotective properties provided by trait forgiveness (TF). TF has been shown to lower blood pressure (BP) and improve heart rate (HR) variability. There is even some evidence that forgiveness predicts mortality, suggesting that failure to forgive unconditionally may be life threatening.
We sought to investigate anger and TF and their potentially divergent relations with cardiovascular risk factors. We employed markers of cardiovascular functioning and tested the relation among these psychological constructs, sympathetic nervous system (SNS) activity, BP control, and noninvasive aortic hemodynamics. We tested the overall hypothesis that anger would predict markers of cardiotoxicity and that TF would be associated with cardioprotection. To this end, we carried out 3 related, yet distinct, studies to test the incremental and unique contribution of TF in comparison with anger in examining the functioning of BP, cardiac autonomic modulation, and aortic hemodynamics. Study 1 assessed autonomic modulation through beat-to-beat BP and power spectral analysis to examine the differential contribution of SNS and parasympathetic nervous system activation on baroreflex sensitivity (BRS) and HR modulation. Study 2 assessed aortic hemodynamics through applanation tonometry and pulse wave analysis (PWA) to allow the measurement of noninvasive surrogates of aortic hemodynamics. Study 3 assessed 24 hour ambulatory BP and ambulatory arterial stiffness.
Methods
A total of 308 healthy young women (M = 21.11 years ± SD = 2.52) participated in this research as approved by the University’s intuitional review board. Subjects were allocated to one of the following studies: Study 1—cardiovascular autonomic modulation and baroreflex function; Study 2—aortic hemodynamics; and Study 3—24-hour ambulatory BP. To minimize potential cardiovascular risk confounders, participants were excluded from study participation through an online health screening assessment if they smoked, exercised regularly as defined as >120 minutes per week in the previous 6 months, were hypertensive as defined as BP ≥140/90 mm Hg, had major chronic diseases, or were taking β blockers, antidepressants, or stimulants. Participants were asked to abstain from caffeine, alcohol, and strenuous physical activity for at least 24 hours before testing and were asked to not eat for 4 hours before testing. Participants were tested in the early follicular phase of the menstrual cycle to avoid potential variations in pressure wave morphology and cardiovascular functioning.
In Study 1, after laboratory familiarization, anthropometrics were measured. Participants then completed a health questionnaire that included a health history form and an anger and TF scale. All data collection was conducted in the afternoon in a quiet, dimly lit, temperature-controlled room (23 ± 1°C) at the same time of the day (±2 hours) to minimize potential diurnal variations in cardiovascular reactivity. After instrument calibration and a 10-minute resting period in a seated position, beat-to-beat finger BP was recorded for 5 minutes.
In Study 2, participants were first introduced to the laboratory setting and familiarized with the study procedures. Body measurements (i.e., height and weight) were taken followed by the completion of a health questionnaire that included a health history form and an anger and TF scale. Data collection was conducted in the afternoon in a quiet, dimly lit, temperature-controlled room (23 ± 1°C) at the same time of the day (±2 hours). Participants were seated and given a 10-minute rest before any measurements were performed. Within 5 minutes after the rest period, measurements for peripheral brachial BP and applanation tonometry of the radial artery for central aortic hemodynamics were taken.
In Study 3, after completing an online health questionnaire, eligible participants were scheduled for a laboratory appointment to complete a 24-hour ambulatory BP assessment. Upon arrival, participants completed an anger scale, a TF scale, health characteristics (height and weight) were measured, and participants were fitted with an ambulatory BP device, which began from 08:00 to 11:00 hours and concluded when the recorder was returned to the laboratory the following day.
The trait subscale of the State-Trait Anger Expression Inventory-2 was used to measure trait anger. Reliability for the sample was α = 0.87. TF was measured using the 4-item Tendency to Forgive Scale. Responses were summed into 1 overall score, with a possible range of 4 to 20. Reliability for the sample was α = 0.81.
Beat-to-beat BP, HR, systolic BP, and diastolic BP were recorded through finger plethysmography (NIBP-100 Biopac Inc., Goleta, California). This method has been shown to provide accurate measurement of BP changes compared with intra-arterial BP. Mean BP was calculated as systolic BP and diastolic BP, where (1/3) systolic BP + (2/3) diastolic BP = mean BP. The BP peaks were used to calculate the time duration of intervals among heartbeats (R wave to R wave interval, RRI) and were automatically detected using commercially available software (WinCPRS, Turku, Finland). The RRIs were inspected for artifacts, premature beats, and ectopic episodes to calculate heart rate variability (HRV) parameters. The main spectral components of the HRV that we calculated, by means of Fast Fourier transformation, were the low frequency (LF; 0.04 to 0.15 Hz) and the high frequency (HF; 0.15 to 0.4 Hz). The use of absolute units (ms 2 ) for HF and LF may be obtained in proportion to the total power, which is expressed in normalized units (nu). Normalization is used to exclude the influence of other fractal components. Because there is structural algebraic redundancy inherent in the normalized spectral HRV measures with respect to each other (normalized low frequency [LFnu] = 1 − normalized high frequency [HFnu]) and also with respect to the LF/HF ratio, here we report LFnu to denote cardiac sympathovagal tone.
Baroreflex functioning was evaluated through BRS calculated from the electrocardiogram and beat-by-beat BP files with the use of the cross-correlation method, which is a time-domain sequential method for baroreflex function based on spontaneous systolic BP and R-R variability changes.
Indices of vascular function and aortic hemodynamics were obtained using brachial BP and applanation tonometry through PWA, which is defined as the examination of the functioning of the arterial (central) pulse wave, allowing for accurate assessment of central hemodynamic functioning. Brachial BP was recorded using an automated oscillometric device (HEM-705CP; Omron Healthcare, Vernon Hill, Illinois). Brachial systolic and diastolic BP was used to calibrate radial waveforms obtained from a 10-second epoch using a high-fidelity tonometer (SPT-301B; Millar Instruments, Houston). PWA provides a more sensitive marker of cardiovascular function than brachial BP. We measured brachial mean BP, aortic mean BP, systolic pressure time interval (STI; indicator of left ventricular work), diastolic pressure time interval (DTI; coronary perfusion), the ratio of DTI to STI expressed as a percentage (subendocardial viability index [SVI]; surrogate of myocardial perfusion and coronary flow reserve), and rate pressure product (RPP = systolic BP × HR; myocardial oxygen consumption). All measurements were obtained in duplicate and averaged. Aortic BP waveforms were derived using a generalized transfer function (SphygmoCor, AtCor Medical, Sydney, Australia). Only high-quality measurements (>80% operator index) were considered for analysis.
Ambulatory BP measurements were collected using validated oscillometric 90217A SpaceLabs (Spacelabs; Wokingham, Berkshire, UK) recorders and calibrated to take 4 measurements per hour for 24 hours. To calculate the ambulatory arterial stiffness index (AASI), the regression slope of ambulatory diastolic BP on ambulatory systolic BP from unedited 24-hour recordings, taken at a rate of 4 per hour, was computed for each participant. The 24-hour mean BP was calculated from the recordings. AASI was defined as 1 minus the regression slope. The stiffer the arterial tree, the closer the regression slope and AASI are to 0 and 1, respectively. The BP dipping was defined as the degree of fall (%) in nocturnal mean arterial pressure relative to the diurnal mean BP: 100 × (1 − [nighttime mean BP ÷ daytime mean BP]).
Pearson correlation coefficients evaluated univariate associations. Hierarchical multiple regression (HMR) analyses were conducted to test the relation between anger and TF with cardiovascular parameters and to demonstrate the incremental contribution of TF from anger in accounting for variance in cardiovascular parameters. A priori alpha level of p <0.05 was considered to be significant. SPSS version 18.0 (SPSS Inc., Chicago, Illinois) was used for all analyses.
Results
Figure 1 shows how the participants were allocated across the 3 studies. Table 1 lists summary statistics for all continuous variables, including demographics, anger, TF, and cardiovascular parameters for Studies 1, 2, and 3.
| Study | 1 (n = 134) | 2 (n = 80) | 3 (n = 94) |
|---|---|---|---|
| Variable (M ± SD) | |||
| Age (years) | 21.28 ± 2.61 | 21.01 ± 2.44 | 21.02 ± 2.57 |
| Height (m) | 1.67 ± 0.08 | 1.65 ± 0.07 | 1.63 ± 0.08 |
| Weight (kg) | 68.72 ± 15.85 | 68.41 ± 9.95 | 64.00 ± 9.17 |
| Body mass index (kg/m 2 ) | 24.35 ± 4.26 | 25.10 ± 4.22 | 24.09 ± 4.17 |
| State-trait anger | 15.40 ± 1.59 | 15.90 ± 1.70 | 16.68 ± 1.05 |
| Tendency to forgive | 12.78 ± 1.52 | 14.60 ± 2.67 | 13.77 ± 2.03 |
| Heart rate (bpm) | 78.41 ± 9.95 | 74.53 ± 8.17 | — |
| Mean blood pressure (mm Hg) | 92.56 ± 7.28 | 83.82 ± 6.83 | — |
| Normalized low frequency | 0.65 ± 0.09 | — | — |
| Baroflex sensitivity (ms/mm Hg) | 16.81 ± 10.94 | — | — |
| Brachial systolic blood pressure (mm Hg) | — | 114.38 ± 7.58 | — |
| Brachial diastolic blood pressure (mm Hg) | — | 68.86 ± 7.51 | — |
| Aortic systolic blood pressure (mm Hg) | — | 97.43 ± 6.40 | — |
| Aortic diastolic blood pressure (mm Hg) | — | 69.78 ± 7.62 | — |
| Aortic mean blood pressure (mm Hg) | — | 82.29 ± 6.68 | — |
| Systolic time interval (mm Hg/s.min −1 ) | — | 1546.90 ± 215.47 | — |
| Diastolic time interval (mm Hg/s.min −1 ) | — | 3392.82 ± 288.47 | — |
| Subendocardial viability index (%) | — | 223.33 ± 33.45 | — |
| Rate pressure product (bpm × mm Hg × 100) | — | 59.93 ± 9.62 | — |
| Ambulatory arterial stiffness index | — | — | 0.28 ± 0.16 |
| Ambulatory 24-hour (bpm) | — | — | 78.66 ± 8.69 |
| Ambulatory 24-mean BP (mm Hg) | — | — | 85.42 ± 6.08 |
| Mean BP dp (%) | — | — | 4.94 ± 7.37 |
In Study 1, 134 participants (M age = 21.28 years, SD = 2.61) qualified for study inclusion. Pearson correlations indicated no statistically significant associations (p <0.05) between anger and TF scores with demographic or anthropometric characteristics. HMR analyses were performed to examine the unique relation TF had with each of the beat-to-beat cardiovascular parameters while controlling for the influence of anger. HMR provided an analysis of the incremental contribution of TF scores above and beyond anger scores in accounting for variance in cardiovascular values. For each cardiovascular parameter serving as an outcome, Model 1 of the HMR contained anger as a predictor, whereas Model 2 added TF as a predictor. Model 2 of the HMR analyses (see Table 2 ) indicated that anger, while controlling for TF, had significant relations with all measured autonomic and cardiovascular outcomes with higher anger scores associated with higher HR, mean BP, LFnu, and lower BRS. Model 2 of the HMR also indicated a significant relation between TF scores and LFnu (but not for HR, mean BP, or BRS) while controlling for anger. The addition of TF in Model 2 of the HMR analyses indicated that TF was able to uniquely predict 7% of the variance in LFnu values; higher TF scores were associated with lower LFnu.
| Variable | Model | Predictors | β | sr | p | Model R 2 | Model ΔR 2 | Model F |
|---|---|---|---|---|---|---|---|---|
| HR (bpm) | 1 | Anger | 0.31 | 0.31 | <0.001 | 0.10 | F(1, 132) = 13.36, p <0.001 | |
| 2 | Anger | 0.33 | 0.32 | <0.001 | 0.11 | 0.01 | ΔF(1, 131) = 1.53, p = 0.219 | |
| Forgive | −0.11 | −0.10 | 0.219 | |||||
| Mean BP (mm Hg) | 1 | Anger | 0.23 | 0.23 | 0.007 | 0.05 | F(1, 132) = 7.43, p = 0.007 | |
| 2 | Anger | 0.22 | 0.21 | 0.015 | 0.06 | 0.00 | ΔF(1, 131) = 0.54, p = 0.464 | |
| Forgive | −0.06 | −0.06 | 0.464 | |||||
| LFnu | 1 | Anger | 0.58 | 0.58 | <0.001 | 0.34 | F(1, 132) = 33.89, p <0.001 | |
| 2 | Anger | 0.56 | 0.58 | <0.001 | 0.41 | 0.07 | ΔF(1, 131) = 22.12, p <0.001 | |
| Forgive | −0.26 | −0.31 | 0.009 | |||||
| BRS (ms/mm Hg) | 1 | Anger | −0.18 | −0.18 | 0.041 | 0.03 | F(1, 132) = 4.26, p = 0.041 | |
| 2 | Anger | −0.18 | −0.18 | 0.045 | 0.03 | 0.00 | ΔF(1, 131) = 0.01, p = 0.917 | |
| Forgive | 0.01 | −0.01 | 0.917 |
In Study 2, 80 participants (M age = 21.01 years, SD = 2.44) qualified for study inclusion. Inclusion criteria were identical to Study 1. Pearson correlations indicated no statistically significant associations (p <0.05) between anger and TF scores with demographic or anthropometric characteristics. Model 1 of the HMR contained anger as a predictor, whereas Model 2 added TF as a predictor. Model 2 of the HMR analyses (see Table 3 ) indicated that anger, while controlling for TF, had significant relations with both brachial and aortic hemodynamic pressures, STI, DTI, and RPP, with higher anger scores associated with higher pressures, STI, DTI, and RPP. Model 2 of the HMR also indicated significant relation (while controlling for anger) between TF scores and HR, STI, SVI, and RPP, with higher TF scores associated with less HR, STI, and RPP but greater SVI. The addition of TF in the Model 2 of the HMR analyses indicated that TF was able to uniquely predict 10% of the variance in HR values, 14.2% of STI values, 12.6% in SVI values, and 9.8% in RPP values.
| Variable | Model | Predictors | β | sr | p | Model R 2 | Model ΔR 2 | Model F |
|---|---|---|---|---|---|---|---|---|
| HR (bpm) | 1 | Anger | 0.148 | 0.148 | 0.229 | 0.022 | F(1, 78) = 1.47, p = 0.229 | |
| 2 | Anger | 0.173 | 0.173 | 0.142 | 0.122 | 0.100 | ΔF(1, 77) = 7.41, p = 0.008 | |
| Forgive | −0.317 | −0.316 | 0.008 | |||||
| BSBP (mm Hg) | 1 | Anger | 0.488 | 0.488 | <0.001 | 0.238 | F(1, 78) = 20.63, p <0.001 | |
| 2 | Anger | 0.503 | 0.501 | <0.001 | 0.271 | 0.033 | ΔF(1, 77) = 2.94, p = 0.091 | |
| Forgive | −0.182 | −0.182 | 0.091 | |||||
| BDBP (mm Hg) | 1 | Anger | 0.693 | 0.693 | <0.001 | 0.480 | F(1, 78) = 61.04, p <0.001 | |
| 2 | Anger | 0.704 | 0.701 | <0.001 | 0.497 | 0.017 | ΔF(1, 77) = 2.19, p = 0.144 | |
| Forgive | −0.131 | −0.130 | 0.144 | |||||
| BMAP (mm Hg) | 1 | Anger | 0.690 | 0.690 | <0.001 | 0.476 | F(1, 78) = 59.87, p <0.001 | |
| 2 | Anger | 0.701 | 0.702 | <0.001 | 0.497 | 0.021 | ΔF(1, 77) = 2.72, p = 0.104 | |
| Forgive | −0.145 | −0.200 | 0.104 | |||||
| ASBP (mm Hg) | 1 | Anger | 0.613 | 0.613 | <0.001 | 0.376 | F(1, 78) = 39.76, p <0.001 | |
| 2 | Anger | 0.620 | 0.618 | <0.001 | 0.384 | 0.008 | ΔF(1, 77) = 0.80, p = 0.373 | |
| Forgive | −0.088 | −0.087 | 0.373 | |||||
| ADBP (mm Hg) | 1 | Anger | 3.18 | 0.694 | <0.001 | 0.482 | F(1, 78) = 61.37, p <0.001 | |
| 2 | Anger | 3.23 | 0.702 | <0.001 | 0.497 | 0.015 | ΔF(1, 77) = 1.91, p = 0.171 | |
| Forgive | −0.635 | −0.122 | 0.171 | |||||
| AMAP (mm Hg) | 1 | Anger | 0.690 | 0.690 | <0.001 | 0.476 | F(1, 78) = 59.87, p <0.001 | |
| 2 | Anger | 0.701 | 0.699 | <0.001 | 0.497 | 0.021 | ΔF(1, 77) = 2.72, p = 0.104 | |
| Forgive | −0.145 | −0.145 | 0.104 | |||||
| STI (mm Hg/s.min −1 ) | 1 | Anger | 0.513 | 0.513 | <0.001 | 0.263 | F(1, 78) = 23.56, p <0.001 | |
| 2 | Anger | 0.543 | 0.542 | <0.001 | 0.405 | 0.142 | ΔF(1, 77) = 15.54, p <0.001 | |
| Forgive | −0.378 | −0.377 | <0.001 | |||||
| DTI (mm Hg/s.min −1 ) | 1 | Anger | 0.575 | 0.575 | <0.001 | 0.330 | F(1, 78) = 32.56, p <0.001 | |
| 2 | Anger | 0.569 | 0.567 | <0.001 | 0.335 | 0.005 | ΔF(1, 77) = 0.497, p = 0.483 | |
| Forgive | 0.072 | 0.071 | 0.483 | |||||
| SVI (%) | 1 | Anger | −0.155 | −0.155 | 0.206 | 0.024 | F(1, 78) = 1.63, p = 0.206 | |
| 2 | Anger | −0.184 | −0.183 | 0.114 | 0.150 | 0.126 | ΔF(1, 77) = 9.64, p = 0.003 | |
| Forgive | 0.356 | 0.355 | 0.003 | |||||
| RPP (bpm × mm Hg) | 1 | Anger | 0.379 | 0.379 | 0.001 | 0.144 | F(1, 78) = 11.10, p = 0.001 | |
| 2 | Anger | 0.405 | 0.403 | <0.001 | 0.242 | 0.098 | ΔF(1, 77) = 8.42, p = 0.005 | |
| Forgive | −0.314 | −0.313 | 0.005 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
