Abstract
Background
Heart failure (HF) is a common complication in individuals with coronary artery disease (CAD). Autonomic effects of psychological stress may play an important, under-recognized role in this relationship. We hypothesized that stress-induced autonomic dysfunction, measured by change in low-frequency heart rate variability (HRV) during mental stress challenge, associates with increased HF risk.
Methods
We examined 662 participants with stable CAD and no known diagnosis of HF who underwent mental stress challenge via a standardized speaking task in conjunction with Holter monitoring. We evaluated HRV in 5-minute windows and examined its change from rest to stress as our primary exposure. Repeated events Cox proportional hazard models were used to examine incident and recurrent acute HF in the outpatient and inpatient setting.
Results
The mean age was 58 years, 35 % were women, and 43 % self-identified as Black. In models adjusted for age, sex, race, comorbidities, ejection fraction, and resting low-frequency HRV, each standard deviation decrease (negative change) in low-frequency HRV change from rest to stress was associated with an increased risk of incident and recurrent acute HF (HR 1.39 [95 % CI 1.02–1.90], p = 0.035) over a median follow-up of 5.7 years. These estimates for HF risk were higher than those of resting HRV.
Conclusion
Greater decreases in low-frequency HRV change during acute mental stress challenge independently associate with higher risks of future HF development in individuals with stable CAD and had stronger effect sizes than resting HRV alone, highlighting an important role of stress autonomic pathways in the pathogenesis of HF.
Keywords: Psychological Stress, Autonomic Dysfunction, Heart Failure
1. Introduction
Heart failure (HF) is a common complication following acute myocardial infarction, with an estimated incidence between 10 % to 40 %, and leads to impaired quality of life and increased mortality risk [[1], [2], [3]]. Psychological conditions, such as depression, commonly accompany a diagnosis of HF [4] and are under-appreciated risk factors for HF progression despite being associated with poorer functional status and higher rates of HF hospitalizations [[5], [6], [7]]. The high burden of HF to society and its strong relationship with psychological factors underscore the need to study underlying mechanisms to aid in prevention efforts.
Prior studies have suggested that psychological stress precipitates HF, but these mechanisms have largely been unexplored [8]. Studies of psychological stress physiology in laboratory settings may help to better understand autonomic and other related mechanisms that may predispose individuals to HF. Laboratory-based public speaking tasks aim to stimulate the autonomic nervous system via cognitive challenge, critical evaluation, negative feedback for poor performance, and social pressure from an audience [9]. They have been shown to stimulate sympathetic activity, parasympathetic withdrawal, and activate the hypothalamus–pituitary–adrenal axis [[10], [11], [12]], which subsequently leads to cardiac remodeling, myocardial cell apoptosis, systemic inflammation, and ultimately the development and progression of HF [[13], [14], [15], [16], [17]].
Heart rate variability (HRV) is a digital biomarker of autonomic balance that can help evaluate stress-based mechanisms of HF risk. It measures the physiological variation between adjacent heart beats and can be used to quantify cardiac autonomic activity in response to various intrinsic and external stimuli [18]. Low-frequency HRV reflects both parasympathetic and sympathetic activity and may help quantify baroreceptor sensitivity, while high-frequency HRV is more strongly influenced by vagal activity during respiration [19]. Some studies suggests low-frequency HRV may have stronger prognostic value for mortality [20], as well as a stronger relationship with chronic stress conditions such as depression and PTSD, than high-frequency HRV [21,22].
Previous prognostic studies have been limited to HRV examination during rest periods, which is nonspecific to and provides limited insight on possible therapies. Previous studies have also suggested that studying HRV during laboratory stress challenge may yield important prognostic insights, although its role in HF prediction has largely been unexplored [23]. In this study, we examined the relationship between stress-induced autonomic dysfunction, as measured by change in low-frequency (primary exposure) and high-frequency (secondary exposure) HRV during mental stress challenge, with the incidence of acute HF in a pooled cohort of individuals with stable CAD and no prior history of clinical HF. We hypothesized that those with decreases in HRV with a laboratory-based stress challenge develop clinical HF at accelerated rates compared to those with preserved autonomic function in response to stress challenge.
2. Methods
2.1. Study participants
We examined a pooled cohort of participants with stable CAD from the Mental Stress Ischemia Prognosis Study (MIPS) [24] and the Myocardial Infarction and Mental Stress Study (MIMS2) [25]. We excluded participants with a history of clinical HF as well as those who were missing HRV data or loss to follow-up. Reasons for missing HRV data included specific scenarios within the Holter monitoring electrocardiogram (ECG) that precluded HRV analysis such as atrial fibrillation, paced-rhythm, and excessive noise artifact. MIPS and MIMS2 shared similar protocols and enrolled patients from Emory-affiliated hospitals between June 2011 and March 2016 [24,25]. The inclusion criteria for MIPS were patients ages 30 to 79 years of age with a documented history of CAD. The inclusion criteria for MIMS2 were patients ages 18 to 60 years of age who had been hospitalized with a myocardial infarction within the prior 8 months. The exclusion criteria for both studies were the same and included acute coronary syndrome or decompensated heart failure within a week of enrollment, severe comorbid medical or psychiatric disorders, use of psychotropic medications other than anti-depressants, immunosuppression, pregnancy or breastfeeding, body weight over 450 lb, language barriers, homelessness, and active illicit drug or alcohol abuse. At the baseline visit, all participants underwent clinical and psychosocial assessments, blood testing, and mental stress testing via a standardized public speaking task. Both studies were approved by the Institutional Review Board of Emory University and all participants provided written informed consent.
2.2. Mental stress protocol
Participants underwent mental stress testing in the morning following a 12-hour fast. Anti-anginal medications, such as beta-blockers, were held the day before the baseline visit. We evoked mental stress using a standardized public speaking task as previously described [24]. In brief, participants were given a 2-minute period to prepare a speech on how they would feel and react to a specific, distressing scenario, followed by a 3-minute period to deliver their speech in front of an audience of at least 4 people.
2.3. Heart rate variability, demographic, and clinical measures
Holter data were collected as part of an ancillary investigation that began after approximately 160 participants had already been enrolled in the first parent study, MIPS. Low-frequency HRV and high-frequency HRV during rest and mental stress were obtained from Holter monitoring ECG that were acquired with GE SEER light monitors (Waukesha, WI, USA). ECG data were edited for artifact and arrhythmia first, and then HRV was measured with commercially available software (GE MARS 8.0.2) during a 5-minute resting period and a 5-minute mental stress testing period that included both the 2-minute preparation and 3-minute speech phases. These were single five-minute intervals of time that did not require averaging. For both low-frequency and high-frequency HRV, we chose a five-minute sampling period because this is the duration of the stress challenge. This was also the default window size of our software, and previously published standards on short-term HRV measurement have also described 5 minute as an acceptable measurement duration [19,26]. Low-frequency HRV included frequencies in the 0.04 Hz to 0.15 Hz range and high-frequency HRV included frequencies in the 0.15 Hz to 0.40 Hz range. All HRV measurements were log-transformed and standardized via converting to z-scores.
Change in low-frequency HRV from rest to stress, the primary exposure of this study, was calculated by measuring the difference in log-transformed low-frequency HRV values between the two states. This involved subtracting the rest value from the stress value. We performed similar analyses for the change in high-frequency HRV from rest to stress as our secondary exposure. All exposure variables were treated as continuous variables in our models.
We chose to focus on low-frequency HRV because prior studies have shown that it is a stronger predictor for adverse cardiovascular outcomes than high-frequency HRV [20]. Other types of HRV, such as ultra-low frequency HRV and very-low frequency HRV, were not measured because they require longer recordings (typically 24-hours) [19], which would not be feasible during a laboratory-based acute mental stress testing. Ultra-low frequency HRV also corresponds more with very slow-acting biological processes, like the circadian rhythm, and thus would not be a suitable measure for our study [19]. Frequency-domain was chosen over time-domain HRV because certain frequency ranges can be associated with specific physiological processes. We also chose to focus on HRV, as oppose to heart rate, because HRV is a more reliable marker of autonomic regulation and is less likely to be confounded by factors such as aging, sinus node dysfunction, and residual beta-blockers effects.
Demographic data (such as race, sex, and age) and medical history (such as smoking status, body mass index, hypertension, hyperlipidemia, diabetes mellitus, myocardial infarction, cardiac revascularizations, and baseline ejection fraction) were obtained through standardized questionnaires and verified via review of electronic medical records. Resting heart rate and blood pressure were obtained at the initial baseline visit.
2.4. Evaluation of heart failure events
The primary outcome of this study was acute HF noted in either the inpatient or outpatient setting. The definitions used for HF were adapted from the Multi-Ethnic Study of Atherosclerosis (MESA) [27]. We included both probable and definite HF, where probable HF was defined as having a clinical diagnosis of acute HF failure by a clinician that required medical treatment, while definite HF required a clinical diagnosis of acute HF plus evidence of pulmonary congestion on chest x-ray and/or evidence of left ventricular dysfunction on imaging. Information on acute HF events was collected through review of medical records and telephone calls. Outside medical records were obtained if an event was reported to have occurred at another hospital outside of Emory Healthcare system. MIPS participants were contacted at 6-month intervals for the first 3 years and then at 5 years, while MIMS participants were contacted at 3-years and 5-years from their initial baseline visit. All-cause mortality was obtained through medical records, telephone calls, and social security death index. All HF events and deaths were adjudicated by a group of study cardiologists blinded from other study data. We did not include any composite outcomes, such as HF and cardiovascular mortality, because our goal was to specifically measure estimates of HF risk independently of atherosclerotic or fatal arrhythmic events in which previous research has already been performed [28].
2.5. Statistical analysis
For descriptive purposes, baseline characteristics were divided into quartiles for change in low-frequency HRV from rest to stress, with the lowest quartile (quartile 1) corresponding with the greatest decrease (most negative change) from rest to stress and the highest quartile (quartile 4) corresponding with the greatest increase (largest positive change) from rest to stress.
For our primary analysis, we examined the association of change in low-frequency HRV from rest to stress, as a continuous variable, with incident and recurrent acute HF (first and repeated events) using repeated events Cox proportional hazard models, specifically the Prentice-Williams-Peterson models with gap time [29]. The fully adjusted model for incident and recurrent acute HF had at least 5 outcomes (combined incident and recurrent) per covariate to limit bias from overfitting [30]. The advantage of using this model is that it accounts for disease severity with the assumption that more frequent admissions equates to greater disease severity [31]. This model also allows us to produce the most precise estimates by increasing the number of events, when compared to using first events alone. Additionally, recurrent HF events incorporate psychosocial and behavioral mediators, which may be advantageous when considering the potential impact of behavioral interventions. Using this cause-specific model may also be more suitable for our study, which aims to examine stress-related autonomic dysfunction as an etiologic mechanism of potential future HF, regardless of comorbid competing risk factors [32]. Of note, we did not examine inpatient and outpatient incident HF separately as potential sensitivity analyses of HF severity because most events (80 %) were inpatient and we already accounted for HF severity with our recurrent events model.
For our secondary analysis, we examined incident acute heart failure (first events only) using Fine and Gray’s sub-distribution hazard models and accounting for all-cause mortality as a competing risk in order to more specifically examine biological mechanisms for heart failure that could be applicable to individual risk-stratification [32]. This alternative model takes into account comorbidities not related to heart failure as alternative causes for censoring and may be useful for estimating individualized clinical prognosis. Because of fewer outcomes, we also examined for potential bias from overfitting by performing a sensitivity analysis of only select covariates that were imbalanced within the exposure subgroups. Overall, by evaluating two different types of outcomes models in parallel, we were able to evaluate the robustness of the hypothesized relationship between stress-induced autonomic dysfunction and HF [32].
Three sequential models were performed in the final analysis: model 1 adjusted for demographic factors (age, sex, and race); model 2 additionally adjusted for comorbidities (current smoking, hypertension, myocardial infarction, coronary artery bypass graft, percutaneous transluminal coronary angioplasty, hyperlipidemia, diabetes mellitus, and body mass index); model 3 additionally adjusted for baseline ejection fraction from echocardiogram or nuclear imaging. Models examining change in HRV from rest to stress were additionally adjusted for resting HRV because a low resting HRV may limit the potential additive impact of stress to further decrease it. We did not analyze the impact of obstructive versus non-obstructive myocardial infarction as possible confounders because only 13 % of participants with a history of myocardial infarction had non-obstructive disease. We also did not control for beta-blocker use because they were balanced between exposure groups. Similar analyses were performed for the change in high-frequency HRV from rest to stress as a secondary exposure (Supplemental Tables 1 and 2). A p-value of less than or equal to 0.05 was considered statistically significant for all analyses. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA).
3. Results
3.1. Participant characteristics
Of the 949 participants (MIPS n = 636, MIMS2 n = 313) in the pooled cohort, 185 and 11 participants respectively were excluded from analysis due to missing heart rate variability data. Of the 753 remaining participants, 85 were excluded for having a history of clinical HF and 6 were lost to follow-up, leaving 662 participants for the final analysis (Fig. 1).
Fig. 1.
Flow diagram of analysis cohort.
For the overall cohort, the mean age was 58 years (standard deviation (SD) ±10), 35 % were women, and 45 % self-identified as Black. The proportion of participants with diabetes mellitus and history of myocardial infarction were slightly higher in the those with greater decreases in low-frequency HRV from rest to stress (Table 1). The proportions of women, Black participants, other comorbid conditions, and beta-blocker use were similar across all quartile groups (Table 1). The mean baseline ejection fraction, mean heart rate at rest, and mean systolic and diastolic blood pressure at rest were also comparable across all groups (Table 1).
Table 1.
Baseline characteristics of participants in the overall cohort, stratified by quartiles for change in log-transformed low-frequency HRV from rest to stress.
| Participant characteristics | Total n = 662 |
Change in low-frequency HRV from rest to stress |
|||
|---|---|---|---|---|---|
| Quartile 1 (−15.04 to −1.33) n = 166 |
Quartile 2 (−1.34 to −0.21) n = 166 |
Quartile 3 (−0.22 to 0.88) n = 163 |
Quartile 4 (0.88 to 9.82) n = 167 |
||
| Age (mean years, SD) | 58 (10) | 58 (10) | 58 (10) | 58 (11) | 60 (10) |
| Female (n, %) | 232 (35 %) | 51 (31 %) | 67 (40 %) | 61 (38 %) | 53 (32 %) |
| Black (n, %) | 273 (41 %) | 65 (39 %) | 75 (45 %) | 67 (41 %) | 66 (40 %) |
| BMI (mean kg/m2, SD) | 30.2 (6.1) | 30.5 (6.3) | 30.4 (7.0) | 30.2 (5.8) | 29.6 (5.4) |
| Current Smoking (n, %) | 252 (38 %) | 65 (39 %) | 55 (34 %) | 64 (39 %) | 68 (41 %) |
| Hypertension (n, %) | 519 (78 %) | 134 (81 %) | 135 (81 %) | 125 (77 %) | 125 (75 %) |
| Hyperlipidemia (n, %) | 543 (82 %) | 136 (82 %) | 138 (83 %) | 135 (83 %) | 134 (80 %) |
| Diabetes mellitus (n, %) | 204 (31 %) | 61 (37 %) | 53 (32 %) | 44 (27 %) | 46 (28 %) |
| Myocardial infarction (n, %) | 390 (59 %) | 106 (64 %) | 102 (61 %) | 89 (55 %) | 93 (56 %) |
| CABG (n, %) | 180 (27 %) | 49 (30 %) | 44 (27 %) | 41 (25 %) | 46 (28 %) |
| PTCA (n, %) | 391 (59 %) | 94 (57 %) | 103 (62 %) | 96 (59 %) | 98 (59 %) |
| Ejection Fraction (mean %, SD) | 63 (14) | 62 (14) | 62 (15) | 63 (15) | 65 (13) |
|
Resting HR (mean bpm, SD) |
63 (11) | 64 (11) | 63 (10) | 63 (11) | 62 (10) |
|
Resting SBP (mean mmHg, SD) |
135 (19) | 137 (20) | 134 (19) | 134 (20) | 133 (18) |
|
Resting DBP (mean mmHg, SD) |
80 (11) | 82 (12) | 80 (10) | 79 (11) | 78 (11) |
| Beta-blocker use (n, %) | 514(78 %) | 130 (78 %) | 135 (81 %) | 124 (76 %) | 125 (75 %) |
Abbreviations: HRV.
heart rate variability, BMI; body mass index, CABG; coronary artery bypass graft, PTCA; percutaneous transluminal coronary angioplasty, HR; heart rate, BPM; beats per minute, SBP; systolic blood pressure, DBP; diastolic blood pressure.
3.2. Association of change in low-frequency from rest to stress with increased incident and recurrent acute heart failure
During a median follow-up time of 5.7 years, there were 93 acute HF events (51 first events and 42 repeated events). In the fully adjusted model (model 3), there was no association between resting low-frequency HRV with incident and recurrent acute HF (Table 2). In contrast, one SD decrease in low-frequency HRV change during mental stress challenge was associated with a 39 % increased risk of incident and recurrent acute HF (HR 1.39 [95 % CI 1.02–1.90] per SD decrease in ln(low-frequency HRV), p = 0.035, Table 2). There was no significant association between change in high-frequency HRV from rest to stress and incident and recurrent acute HF (HR 1.18 [0.86–1.61] per SD decrease in ln(high-frequency HRV), p = 0.306, Supplemental Table 1).
Table 2.
Repeated events Cox proportional hazard models showing the association of resting low-frequency heart rate variability and change in low-frequency heart rate variability from rest to stress with incident and recurrent heart failure.
| Exposures | Incident and recurrent heart failure |
||
|---|---|---|---|
| Model 1 |
Model 2 |
Model 3 |
|
| HR (95 % CI) | HR (95 % CI) | HR (95 % CI) | |
|
Resting Low-frequency HRV* |
1.07 (0.82–1.40) P = 0.597 |
1.02 (0.71–1.46) P = 0.934 |
1.12 (0.76–1.66) P = 0.568 |
|
Change in Low-frequency HRV from Rest to Stress*# |
1.37 (1.07–1.76) P = 0.012 |
1.42 (1.04–1.95) P = 0.028 |
1.39 (1.02–1.90) P = 0.035 |
*Per standard deviation decrease in ln(low-frequency HRV).
Model 1: Adjusted for age, sex, and race.
Model 2: Model 1 + adjusted for current smoking, myocardial infarction, coronary artery bypass graft, percutaneous transluminal coronary angioplasty, hypertension, hyperlipidemia, diabetes mellitus, and body mass index.
Model 3: Model 2 + adjusted for ejection fraction.
Abbreviations: HR; hazard ratio, CI; confidence intervals, HRV; heart rate variability.
Additionally adjusted for resting low-frequency HRV.
3.3. Association of change in low-frequency HRV from rest to stress with increased incident acute heart failure
In our competing risk models, there was no association between resting low-frequency HRV or high-frequency HRV with incident acute HF (Table 3 and Supplemental Table 2). In contrast, one SD decrease in low-frequency HRV change during mental stress challenge was associated with a 41 % increase in incident acute HF (HR 1.41 [95 % CI 1.00–1.99] per SD decrease in ln(low-frequency HRV), p = 0.047, Table 3) after multivariate adjustments. The effects were attenuated in the relationship between change in high-frequency HRV from rest to stress and incidence of acute HF (HR 1.34 [95 % CI 0.97–1.87] per SD decrease in ln(high-frequency HRV), p = 0.079, Supplemental Table 2). A sensitivity analysis of a reduced model which only featured imbalanced covariates (female sex, myocardial infarction, diabetes mellitus, and ejection fraction) showed similar results to Model 3 of Table 3 (Supplemental Table 3).
Table 3.
Fine and Gray’s sub-distribution hazard models showing the association of resting low-frequency heart rate variability and change in low-frequency heart rate variability from rest to stress with incident heart failure.
| Exposures | Incident heart failure |
||
|---|---|---|---|
| Model 1 |
Model 2 |
Model 3 |
|
| HR (95 % CI) | HR (95 % CI) | HR (95 % CI) | |
|
Resting Low-Frequency HRV* |
1.55 (1.11–2.16) P = 0.010 |
1.21 (0.89–1.66) P = 0.220 |
1.27 (0.92–1.74) P = 0.142 |
|
Change in Low-frequency HRV from Rest to Stress*# |
1.77 (1.27–2.46) P < 0.001 | 1.47 (1.04–2.09) P = 0.028 |
1.41 (1.00–1.99) P = 0.047 |
*Per standard deviation decrease in ln(low-frequency HRV).
Competing risk = all-cause mortality.
Model 1: Adjusted for age, sex, and race.
Model 2: Model 1 + adjusted for current smoking, myocardial infarction, coronary artery bypass graft, percutaneous transluminal coronary angioplasty, hypertension, hyperlipidemia, diabetes mellitus, and body mass index.
Model 3: Model 2 + adjusted for ejection fraction.
Abbreviations: HR; sub-distribution hazard ratio, CI; confidence intervals, HRV; heart rate variability.
Additionally adjusted for resting low-frequency HRV.
4. Discussion
In a large cohort of individuals with stable CAD, we found that greater decreases in low-frequency HRV during acute mental stress challenge independently predicts a higher risk of incident and recurrent acute HF. In other words, individuals who had excessive sympathetic activity following mental stress challenge were at greater risk for the development of heart failure later in life. This association persisted despite comprehensive multivariable adjustments for sociodemographic and cardiovascular factors, and were significant regardless of which survival model was chosen. The effect sizes for HRV changes with stress were stronger than rest HRV only, which emphasizes the importance of autonomic stress pathways specifically. Overall, our findings support stress-induced autonomic dysfunction as a novel mechanism for the development of clinical HF in individuals with stable CAD (Fig. 2).
Fig. 2.
In individuals with stable coronary artery disease, exaggerated reduction in low-frequency heart rate variability during mental stress challenge was associated with increased heart failure risk later in life.
To our knowledge, our study is the first to show that greater decreases in HRV in response to acute stress is associated with higher incidence of acute HF in individuals with stable CAD. Previous research has focused on resting or ambulatory HRV data and only in those without CAD. This builds upon a previous study from the Cardiovascular Health Study cohort, which also specifically examined HRV and HF risk in older community-based individuals [33]. They evaluated several 24-hour HRV parameters and found independent associations with future HF risk. Although they evaluated different HRV parameters in the setting of ambulatory monitoring, and not mental stress challenge specifically, the study supports the construct that digital biomarkers of autonomic function are relevant to future HF risk. Our study adds to the previous HF study by highlighting the role of psychosocial stimuli as important components of the autonomic risk predisposing certain individuals with stable CAD to HF. Another advantage to our design and use of short-term HRV data in the setting of stress provocation is that it may shed light on possible autonomic interventions that include stress pathways.
A possible mechanistic pathway underlying the relationship between mental stress and HF involves hemodynamic aberrancies caused by baroreceptor dysregulation. Reduced low-frequency HRV may indicate reduced baroreflex sensitivity [19,34], which may lead to more episodes of hypertension and greater blood pressure variability [35]. Reduced HRV and its exaggerated reductions during stressful challenges may indicate dysregulations in blood pressure, which induces pathological ventricular remodeling and subsequently HF [17,[36], [37], [38]]. Additionally, repeated episodes of acute stress may lead to bursts of excessive sympathetic nervous system activity that facilitates adverse cardiac remodeling and myocardial cell apoptosis overtime [16,39].
Our findings point to potentially novel approaches to the management of HF in individuals with stable CAD. Many therapies impact the autonomic nervous system and reduce stress like exercise training and neuromodulation. For example, exercise training has been shown to improve parasympathetic tone and reduce sympathetic activity [40], as well as reduce depressive symptoms. Baroreflex activation therapy has also been shown to improve quality of life and exercise capacity in patients with HF with reduced EF [41]. Guided breathing interventions may also impact autonomic and brain stress pathways that are more specific to the changes in HRV with stress [42].
Our findings also highlight the potentially important role stress pathways play in the large number of individuals with pre-clinical HF (defined as the asymptomatic presence of structural heart disease and/or elevated cardiac biomarkers) [43], such as those in our cohort. The changes in HRV with stress likely involved several brain regions that regulate the stress response, including the amygdala, hypothalamus, and prefrontal cortex. Emerging evidence suggests these regions play in important role in CVD risk [44,45]. Unfortunately, in many clinical settings, discussions of stress exposures and psychological symptoms are often overlooked. Pre-clinical HF is a public health concern, as approximately one-fourth develop symptomatic HF within 5 years [46]. Early identification of at-risk individuals is important because the five-year mortality increases approximately five-fold with advancement from pre-clinical to clinical HF [47]. The use of HRV with mental stress provocation in the clinical setting may help identify patients who are at highest risk of heart failure progression and who may warrant more aggressive lifestyle and pharmacological interventions. Disparities in HF outcomes have also been observed in individuals of low socioeconomic status and patients from minoritized racial and ethnic groups [[48], [49], [50]]. The role of stress in mediating increased HF outcomes in these groups, who have more frequent stress exposures, is another important avenue of future research.
The findings of this study are limited in terms of generalizability to those with CAD and in whom HRV data capture was possible. The findings of this study cannot be generalizable to patients without CAD because it is a pooled cohort from the Mental Stress Ischemia Prognosis Study (MIPS) [24] and the Myocardial Infarction and Mental Stress Study (MIMS2) [25], both of which only included high-risk individuals with stable CAD. Nonetheless, it is important to study this population because of their high morbidity and mortality rates, which necessitate more aggressive prevention efforts. Given HRV indices cannot be measured in participants with atrial fibrillation, paced rhythm, or excessive ectopy, individuals with these risks of developing HF were excluded from this study. The study cohort was also recruited from a single healthcare system in the southeastern United States, which may further limit generalizability to other groups that were not represented. Our survival models were subject to certain limitations as discussed in the methods sections, and as a result we needed to conduct two separate types of survival analyses which made different assumptions on the relationship between stress HRV and HF. Another potential limitation of this study is that our final model did not account for changes in respiratory frequency, which can occur during a public speaking task and subsequently affect HRV. Nonetheless, the likelihood of confounding was likely minimal because the participants were uniformly instructed to speak for the entire 3-minute stress period. Respiratory effects also primarily influence high-frequency HRV, not low-frequency HRV [19]. Another potential limitation to consider is that there may not have been complete wash-out of beta-blocker effects since these medications were only held for 24 hours prior to mental stress testing. Reassuringly, beta-blocker use was fairly uniform across HRV exposure groups and thus the likelihood of confounding was low. Participants with a history of atrial fibrillation were also not excluded from this study, although our findings were unlikely to be confounded by atrial fibrillation given the low (<1 %) prevalence of this arrhythmia in the study cohort and the inability to measure HRV when this rhythm occurred. We were also unable to discern the exact nature of the heart failure events due to limited medical documentation and clinical testing in many cases. This limited our ability to examine behavioral factors such as poor medical and dietary adherence. Lastly, because this is an observational study and residual confounding may exist, we cannot prove a causal relationship between reduced HRV following acute mental stress challenge and the risk of incident and recurrent heart failure.
In conclusion, we found that greater decreases in low-frequency HRV during acute mental stress challenge independently predicts a higher risk of incident and recurrent acute HF, suggesting that novel autonomic pathways involving the acute stress response play a role in future HF development in individuals with stable CAD. Our findings emphasize the importance of considering psychological stress physiology in the prevention of HF. However, further translational research is needed on possible interventions that improve autonomic stress physiology and possible risk-stratification tools that are practical in clinical settings. This may include ECG-based or wristband wearables that may quantify autonomic physiology during acute stress exposures in real-life situations. Overall, the improved understanding of heart-brain mechanisms gained by this study may be helpful in informing future preventive strategies in HF among individuals with stable CAD.
CRediT authorship contribution statement
Maggie Wang: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Chang Liu: Writing – review & editing, Methodology, Formal analysis. Anish Shah: Writing – review & editing, Data curation. Yi-An Ko: Writing – review & editing, Methodology. Rachel Lampert: Writing – review & editing. Yan V. Sun: Writing – review & editing. Kasra Moazzami: Writing – review & editing, Data curation. Mariana Garcia: Writing – review & editing, Data curation. Zakaria Almuwaqqat: Writing – review & editing, Data curation. Gabriel Najarro PA-C: Data curation. Samaah Sullivan: Writing – review & editing. Paolo Raggi: Writing – review & editing. J. Douglas Bremner: Writing – review & editing. Arshed A. Quyyumi: Writing – review & editing. Viola Vaccarino: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Alanna A. Morris: Writing – review & editing, Supervision. Amit J. Shah: Writing – review & editing, Supervision, Funding acquisition, Conceptualization.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This work was supported by grants P01 HL101398, R01 HL109413, R01 HL109413-02S1, R01 HL155711, K24 HL077506, K24 MH076955, UL1 TR000454, KL2 TR000455, K23 HL127251, and T32 HL130025 from the National Institutes of Health.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.ijcha.2025.101694.
Appendix A. Supplementary data
The following are the Supplementary data to this article:
References
- 1.Velagaleti R.S., Pencina M.J., Murabito J.M., et al. Long-term trends in the incidence of heart failure after myocardial infarction. Circulation. 2008;118(20):2057–2062. doi: 10.1161/CIRCULATIONAHA.108.784215. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Chen J., Hsieh A.F., Dharmarajan K., Masoudi F.A., Krumholz H.M. National trends in heart failure hospitalization after acute myocardial infarction for Medicare beneficiaries: 1998-2010. Circulation. 2013;128(24):2577–2584. doi: 10.1161/circulationaha.113.003668. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Gerber Y., Weston S.A., Enriquez-Sarano M., et al. Mortality associated with heart failure after myocardial infarction. Circ. Heart Fail. 2016;9(1) doi: 10.1161/CIRCHEARTFAILURE.115.002460. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Rutledge T., Reis V.A., Linke S.E., Greenberg B.H., Mills P.J. Depression in heart failure a meta-analytic review of prevalence, intervention effects, and associations with clinical outcomes. J Am Coll Cardiol. 2006;48(8):1527–1537. doi: 10.1016/j.jacc.2006.06.055. [DOI] [PubMed] [Google Scholar]
- 5.Endrighi R., Dimond A.J., Waters A.J., et al. Associations of perceived stress and state anger with symptom burden and functional status in patients with heart failure. Psychol Health. 2019;34(10):1250–1266. doi: 10.1080/08870446.2019.1609676. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Vaccarino V., Almuwaqqat Z., Kim J.H., et al. Association of mental stress–induced myocardial ischemia with cardiovascular events in patients with coronary heart disease. JAMA. 2021;326(18):1818–1828. doi: 10.1001/jama.2021.17649. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Almuwaqqat Z., Wittbrodt M., Young A., et al. Association of early-life trauma and risk of adverse cardiovascular outcomes in young and middle-aged individuals with a history of myocardial infarction. JAMA Cardiol. 2021;6(3):336–340. doi: 10.1001/jamacardio.2020.5749. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Harris K.M., Jacoby D.L., Lampert R., Soucier R.J., Burg M.M. Psychological stress in heart failure: a potentially actionable disease modifier. Heart Fail Rev. 2021;26(3):561–575. doi: 10.1007/s10741-020-10056-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Ketterer M.W., Freedland K.E., Krantz D.S., et al. Psychological correlates of mental stress-induced ischemia in the laboratory: the psychophysiological investigation of myocardial ischemia (PIMI) study. J. Health Psychol. 2000;5(1):75–85. doi: 10.1177/135910530000500112. [DOI] [PubMed] [Google Scholar]
- 10.Allen A.P., Kennedy P.J., Dockray S., Cryan J.F., Dinan T.G., Clarke G. The trier social stress test: principles and practice. Neurobiol Stress. 2017;6:113–126. doi: 10.1016/j.ynstr.2016.11.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hostinar C.E., McQuillan M.T., Mirous H.J., Grant K.E., Adam E.K. Cortisol responses to a group public speaking task for adolescents: variations by age, gender, and race. Psychoneuroendocrinology. 2014;50:155–166. doi: 10.1016/j.psyneuen.2014.08.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Kirschbaum C., Pirke K.M., Hellhammer D.H. The 'Trier Social Stress Test'–a tool for investigating psychobiological stress responses in a laboratory setting. Neuropsychobiology. 1993;28(1–2):76–81. doi: 10.1159/000119004. [DOI] [PubMed] [Google Scholar]
- 13.Brotman D.J., Golden S.H., Wittstein I.S. The cardiovascular toll of stress. Lancet. 2007;370(9592):1089–1100. doi: 10.1016/s0140-6736(07)61305-1. [DOI] [PubMed] [Google Scholar]
- 14.Steptoe A., Kivimäki M. Stress and cardiovascular disease. Nat. Rev. Cardiol. 2012;9(6):360–370. doi: 10.1038/nrcardio.2012.45. [DOI] [PubMed] [Google Scholar]
- 15.Osborne M.T., Shin L.M., Mehta N.N., Pitman R.K., Fayad Z.A., Tawakol A. Disentangling the links between psychosocial stress and cardiovascular disease. Circ. Cardiovasc. Imaging. 2020;13(8) doi: 10.1161/CIRCIMAGING.120.010931. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Amin P., Singh M., Singh K. β-adrenergic receptor-stimulated cardiac myocyte apoptosis: role of β1 integrins. J. Signal Transduct. 2011;2011 doi: 10.1155/2011/179057. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Mehra R., Tjurmina Olga A., Ajijola Olujimi A., et al. Research opportunities in autonomic neural mechanisms of cardiopulmonary regulation. JACC: Basic Transl. Sci. 2022;7(3_Part_1):265–293. doi: 10.1016/j.jacbts.2021.11.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Electrophysiology TFotESoCtNASoP. Heart Rate Variability. Circulation. 1996;93(5):1043-1065. DOI: 10.1161/01.CIR.93.5.1043. [PubMed]
- 19.Shaffer F., Ginsberg J.P. An overview of heart rate variability metrics and norms. Front. Public Health. 2017;5:258. doi: 10.3389/fpubh.2017.00258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Tsuji H., Venditti F.J., Jr., Manders E.S., et al. Reduced heart rate variability and mortality risk in an elderly cohort. the framingham heart study. Circulation. 1994;90(2):878–883. doi: 10.1161/01.cir.90.2.878. [DOI] [PubMed] [Google Scholar]
- 21.Shah A.J., Lampert R., Goldberg J., Veledar E., Bremner J.D., Vaccarino V. Posttraumatic stress disorder and impaired autonomic modulation in male twins. Biol. Psychiatry. 2013;73(11):1103–1110. doi: 10.1016/j.biopsych.2013.01.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Huang M., Shah A., Su S., et al. Association of depressive symptoms and heart rate variability in Vietnam war-era twins: a longitudinal twin difference study. JAMA Psychiat. 2018;75(7):705–712. doi: 10.1001/jamapsychiatry.2018.0747. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Hamilton J.L., Alloy L.B. Atypical reactivity of heart rate variability to stress and depression across development: systematic review of the literature and directions for future research. Clin. Psychol. Rev. 2016;50:67–79. doi: 10.1016/j.cpr.2016.09.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hammadah M., Al Mheid I., Wilmot K., et al. The mental stress ischemia prognosis study: objectives, study design, and prevalence of inducible ischemia. Psychosom. Med. 2017;79(3):311–317. doi: 10.1097/psy.0000000000000442. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Vaccarino V., Sullivan S., Hammadah M., et al. Mental stress–induced-myocardial ischemia in young patients with recent myocardial infarction. Circulation. 2018;137(8):794–805. doi: 10.1161/CIRCULATIONAHA.117.030849. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Malik M. Heart rate variability. Standards of measurement, physiological interpretation, and clinical use. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Eur. Heart J. 1996;17(3):354–381. [PubMed] [Google Scholar]
- 27.Polonsky T.S., Ning H., Daviglus M.L., et al. Association of cardiovascular health with subclinical disease and incident events: the multi‐ethnic study of atherosclerosis. J. Am. Heart Assoc. 2017;6(3) doi: 10.1161/JAHA.116.004894. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Shah A.S., Vaccarino V., Moazzami K., et al. Autonomic reactivity to mental stress is associated with cardiovascular mortality. Eur. Heart J. Open. 2024;4(6) doi: 10.1093/ehjopen/oeae086. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Prentice R.L., Williams B.J., Peterson A.V. On the regression analysis of multivariate failure time data. Biometrika. 1981;68(2):373–379. doi: 10.1093/biomet/68.2.373. [DOI] [Google Scholar]
- 30.Vittinghoff E., McCulloch C.E. Relaxing the rule of ten events per variable in logistic and cox regression. Am. J. Epidemiol. 2006;165(6):710–718. doi: 10.1093/aje/kwk052. [DOI] [PubMed] [Google Scholar]
- 31.Lahoz R., Fagan A., McSharry M., Proudfoot C., Corda S., Studer R. Recurrent heart failure hospitalizations increase the risk of mortality in heart failure patients with atrial fibrillation and type 2 diabetes mellitus in the United Kingdom: a retrospective analysis of Clinical Practice Research Datalink database. BMC Cardiovasc. Disord. 2022;22(1):234. doi: 10.1186/s12872-022-02665-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Austin P.C., Lee D.S., Fine J.P. Introduction to the analysis of survival data in the presence of competing risks. Circulation. 2016;133(6):601–609. doi: 10.1161/CIRCULATIONAHA.115.017719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Patel V.N., Pierce B.R., Bodapati R.K., Brown D.L., Ives D.G., Stein P.K. Association of Holter-derived heart rate variability parameters with the development of congestive heart failure in the cardiovascular health study. JACC Heart Failure. 2017;5(6):423–431. doi: 10.1016/j.jchf.2016.12.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Rahman F., Pechnik S., Gross D., Sewell L., Goldstein D.S. Low frequency power of heart rate variability reflects baroreflex function, not cardiac sympathetic innervation. Clin. Auton. Res. 2011;21(3):133–141. doi: 10.1007/s10286-010-0098-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Sugimoto Y., Kiyono K., Yoshino K. Analysis of the relationship between amplitude modulation of low frequency heart rate variability and blood pressure variability. Adv. Biomed. Eng. 2019;8:78–84. doi: 10.14326/abe.8.78. [DOI] [Google Scholar]
- 36.Florea V.G., Cohn J.N. The autonomic nervous system and heart failure. Circ. Res. 2014;114(11):1815–1826. doi: 10.1161/circresaha.114.302589. [DOI] [PubMed] [Google Scholar]
- 37.Mostarda C., Moraes-Silva I.C., Moreira E.D., et al. Baroreflex sensitivity impairment is associated with cardiac diastolic dysfunction in rats. J. Card. Fail. 2011;17(6):519–525. doi: 10.1016/j.cardfail.2011.02.007. [DOI] [PubMed] [Google Scholar]
- 38.Grassi G., Seravalle G., Quarti-Trevano F., et al. Sympathetic and baroreflex cardiovascular control in hypertension-related left ventricular dysfunction. Hypertension. 2009;53(2):205–209. doi: 10.1161/HYPERTENSIONAHA.108.121467. [DOI] [PubMed] [Google Scholar]
- 39.Gardner R.T., Ripplinger C.M., Myles R.C., Habecker B.A. Molecular mechanisms of sympathetic remodeling and arrhythmias. Circ. Arrhythm. Electrophysiol. 2016;9(2) doi: 10.1161/CIRCEP.115.001359. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Pearson M.J., Smart N.A. Exercise therapy and autonomic function in heart failure patients: a systematic review and meta-analysis. Heart Fail. Rev. 2018;23(1):91–108. doi: 10.1007/s10741-017-9662-z. [DOI] [PubMed] [Google Scholar]
- 41.Zile M.R., Lindenfeld J., Weaver F.A., et al. Baroreflex activation therapy in patients with heart failure with reduced ejection fraction. J. Am. Coll. Cardiol. 2020;76(1):1–13. doi: 10.1016/j.jacc.2020.05.015. [DOI] [PubMed] [Google Scholar]
- 42.Maniaci G., Daino M., Iapichino M., et al. Neurobiological and anti-inflammatory effects of a deep diaphragmatic breathing technique based on neofunctional psychotherapy: a pilot RCT. Stress Health. 2024;40(6):e3503. doi: 10.1002/smi.3503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Heidenreich P.A., Bozkurt B., Aguilar D., et al. 2022 AHA/ACC/HFSA guideline for the management of heart failure: a report of the american college of cardiology/american heart association joint committee on clinical practice guidelines. Circulation. 2022;145(18):e895–e1032. doi: 10.1161/CIR.0000000000001063. [DOI] [PubMed] [Google Scholar]
- 44.Moazzami K., Wittbrodt M.T., Lima B.B., et al. Higher activation of the rostromedial prefrontal cortex during mental stress predicts major cardiovascular disease events in individuals with coronary artery disease. Circulation. 2020;142(5):455–465. doi: 10.1161/circulationaha.119.044442. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Tawakol A., Ishai A., Takx R.A., et al. Relation between resting amygdalar activity and cardiovascular events: a longitudinal and cohort study. Lancet. 2017;389(10071):834–845. doi: 10.1016/s0140-6736(16)31714-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Goldberg L.R., Jessup M. Stage B heart failure. Circulation. 2006;113(24):2851–2860. doi: 10.1161/CIRCULATIONAHA.105.600437. [DOI] [PubMed] [Google Scholar]
- 47.Ammar K.A., Jacobsen S.J., Mahoney D.W., et al. Prevalence and prognostic significance of heart failure stages. Circulation. 2007;115(12):1563–1570. doi: 10.1161/CIRCULATIONAHA.106.666818. [DOI] [PubMed] [Google Scholar]
- 48.Schultz W.M., Kelli H.M., Lisko J.C., et al. Socioeconomic status and cardiovascular outcomes. Circulation. 2018;137(20):2166–2178. doi: 10.1161/CIRCULATIONAHA.117.029652. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Glynn P., Lloyd-Jones D.M., Feinstein M.J., Carnethon M., Khan S.S. Disparities in cardiovascular mortality related to heart failure in the United States. J. Am. Coll. Cardiol. 2019;73(18):2354–2355. doi: 10.1016/j.jacc.2019.02.042. [DOI] [PubMed] [Google Scholar]
- 50.Bahrami H., Kronmal R., Bluemke D.A., et al. Differences in the incidence of congestive heart failure by ethnicity: the multi-ethnic study of atherosclerosis. Arch. Intern. Med. 2008;168(19):2138–2145. doi: 10.1001/archinte.168.19.2138. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.