Effects of Heart Rate Variability-guided Training vs Standard Aerobic Prescription on Cardiovagal Modulation and Cardiorespiratory Fitness in Coronary Artery Disease

Background

Coronary artery disease (CAD) is characterized by the inability to adequately supply blood and oxygen to the heart, often resulting from the formation of atherosclerotic plaques within the arterial lumen. This inflammatory, multifactorial condition can impair the autonomic nervous system (ANS), disrupting the balance between the sympathetic and parasympathetic pathways and leading to autonomic dysfunction. Key indicators of this dysfunction include reduced heart rate variability (HRV) and delayed heart rate recovery, both of which signify heightened sympathetic activity and diminished parasympathetic activity. These changes predict worse cardiovascular outcomes in CAD patients, as the ANS compensates for myocardial ischemia and hypoxia by increasing sympathetic drive and reducing vagal activity to preserve cardiac contractility and output. This imbalance triggers a cascade of neurohumoral alterations affecting the cardiovascular, peripheral vascular, and renal systems, further exacerbating disease progression.

Regular exercise is known to play an important role in the prevention, management, and treatment of CAD, mainly through the preservation of endothelial function but also through the improvement of cardiovagal modulation and neurocardiovascular stress reactivity and by the increase in cardiorespiratory fitness. Patients with stable CAD should undergo a cardiac rehabilitation program to fulfill exercise guidelines. Protocols may vary in mode, intensity, frequency, and duration, whereas the most recommended types of exercise are walking and cycling, with intensities between 40-80% of VO2peak associated with increases in exercise capacity by 11-36%. Recent studies demonstrated that high-intensity interval training (HIIT) compared with moderate-intensity continuous training (MICT) promotes higher adaptation in VO2max and in post-exercise heart rate recovery, HIIT has also demonstrated to contribute to the restoration of endothelial function and autonomic balance, inducing reverse cardiac remodeling and increasing left ventricle morphology and function. However, MICT demonstrates higher adherence because it is a safer and more effective approach. Nevertheless, despite the known non-pharmacological benefits of exercise, recent studies demonstrate that more than 15% of the patients with CAD are exercise non-responders to increases in VO2max , i.e., an increase of 3.5 ml/kg/min in VO2max, which has been associated with a 10% decrease in all-cause of mortality and cardiovascular mortality, however the greater the increases in VO2max the greater the reduction in the risk.

Exercise individualization is increasing among apparently healthy populations due to the ease of the use of cellphone applications. Recently, HRV-guided training has been identified as an alternative prescription technique for cardiovascular endurance exercise prescription, this prescription technique demonstrated to contribute to greater improvements in endurance athletes. HRV-guided training concerns to exercise prescribed based on daily changes in HRV, depending on thresholds constructed at the individual level. After an initial characterization period, where the HRV after waking up is measured, the individual HRV profile is defined. Then, considering the resting values of HRV, exercise intensity will be determined for the training session, i.e., if LnRMSSD7day-roll-avg remained inside the smallest worthwhile change (SWC) (+), high-intensity or moderate-intensity training sessions is prescribed, and if LnRMSSD7day-roll-avg fell outside SWC (-), low intensity or rest is prescribed. This prescription technique leads to different workloads from the predefined training programs.

To the best of our knowledge, no study has yet compared the chronic effects of two different exercise protocols on cardiovagal modulation, and cardiorespiratory fitness in patients with CAD. Thus, this research aims to compare the chronic adaptations on cardiorespiratory fitness, cardiovagal modulation, baroreflex sensitivity (BRS), and arterial stiffness (AS) of two different techniques of exercise prescription (traditional vs individualized) and two training methods (HIIT vs MICT) in patients with CAD. We hypothesize that exercise individualization will promote higher adaptations in cardiovagal modulation, cardiorespiratory fitness, BRS and AS in patients with CAD, compared with traditional prescription.

Methods

Study design This longitudinal randomized trial will consist of comparing the effects of different exercise prescription techniques and different methods of exercise prescription (traditional MICT, traditional HIIT, and HRV-guided training) would exhibit differences in chronic adaptations in cardiorespiratory fitness, cardiovagal modulation, and BRS in patients with CAD. The study will consist of 12 weeks of supervised exercise with 3 sessions per week. This study protocol was submitted to the Faculty of Human Kinetics Institutional Review Board.

Sample size

Assuming a change in VO2max of 3.5 mL/kg/min to be clinically relevant, α = 0.05, 1-β = 0.80, an expected dropout of 15% for 12 weeks, and a chosen medium effect size of 0.25 due to gaps in the literature, the calculations yielded a total minimum of 48 participants (16 per group) (G*Power, version 3.1.9.4), however, the sample size will be updated, if necessary, after a pilot study consisting of 12 participants, 4 per group, (48x0.25=12).

Recruitment

Recruitment will be performed mainly at the Coronary Club of Lisbon (CORLIS) (phase III) at the Faculty of Human Kinetics and at the Hospital of Santa Cruz in Lisbon.

Measurements

All study measurements will be collected at the Faculty of Human Kinetics, University of Lisbon. Assessments of cardiopulmonary exercise test (CPET), cardiovagal modulation, body composition, and arterial stiffness, will be conducted both at baseline and after the 12-week intervention period. Participants will be requested to be fasted (> 4h), abstain from foods/drinks containing caffeine and alcohol (> 12h), and refrain from strenuous exercise (> 24h) before assessments, post-intervention assessments will be scheduled for the same time that baseline. The measures will follow the following order body composition, cardiovagal modulation (and baroreflex sensitivity), local and regional arterial stiffness, cardiopulmonary test, and post-maximal exercise cardiovagal modulation (and baroreflex sensitivity), and local and regional arterial stiffness.

Randomization

Baseline evaluation will be done before randomization, thus neither participants nor fieldwork staff will know the group assignment at this point. Due to the nature of the intervention, it will not be possible to blind participants after baseline assessments. After the baseline assessments, block randomization by group with balanced blocks of 12 participants will be performed by a researcher not involved in recruitment or data collection using REDCap (https://imm.medicina.ulisboa.pt/group/redcap/redcap_v7.6.7/). The participants will be randomly allocated to one of the three groups: HIIT, MICT, and HRV-guided training.

Exercise intervention

Exercise intervention will have a 12-week duration, with 3 sessions per week on non-consecutive days for all 3 groups and will be carried out at the Faculdade de Motricidade Humana. Certified exercise physiologists will supervise sessions. Heart rate will be monitored during all sessions with an HR monitor (S810i Polar, Polar Electro, Kempele, Finland; sampling rate 1000 Hz). All sessions will include a 10-minute warm-up, an aerobic component, a strength component, and a cool-down for both groups. The warm-up will focus on increasing HR, improving mobility, and performing preparatory stretches for the conditioning component. The strength component will consist of 6 arm and 2 leg machines with 2 sets of 8 to 12 reps at 50-60% 1RM per machine. The cool-down will transition from the conditioning component to the stretching phase, consisting of static and dynamic stretching exercises for all major muscle groups. Warm-up, strength component and coold-down will be standardized for the 3 groups. Sessions will be deemed completed when at least 83.3% (30 out of 36 sessions i.e. minimum adherence = 83.3%) of the prescribed exercise sessions have been successfully performed (reached the programmed intensity). The aerobic component in all groups will follow linear periodization. Volume and/or intensity will increase every 4 weeks.

Data analysis

All HRV analyses will be performed offline using the FisioSinal software built-in Matlab [25]. Time-domain and spectral power analyses will be conducted using 2-min time bins. Ectopic heartbeats (M = 1, SD = 4 b.min-1) will be excluded from the final analysis. The time-domain statistics used to characterize HRV will be SDNN and RMSSD. Non-linear time-domain parameters will be derived from the Poincaré plot including SD1 and SD2. The time-frequency domain analysis will be conducted using the Daubechy-12 discrete wavelet algorithm. Wavelet analysis will be chosen instead of fast Fourier transform as it is better suited to characterize the acute responses of the autonomic nervous system during exercise. The baroreflex effectiveness index (BEI) and BRS will be estimated using the spontaneous sequence method. Briefly, the method casts about adjacent oscillations (ramps) in systolic blood pressure (SBP; > 1 mmHg) and RR and ( > 4 ms). BRS will be defined as the average of the BRS slopes. BEI will be calculated as the total number of BRS events divided by the total number of BP ramps observed during the 2-minute time-bin.

Statistical methods

Statistical tests will be performed in R (R studio Version 4.2.2). Descriptive statistics will be employed to analyze demographic and clinical variables measured at the baseline for each group. Categorical variables will be presented using frequencies and percentages, while continuous variables will be expressed as mean ± standard deviation (or median and range).

For the primary and secondary outcomes, data will undergo analysis using linear mixed models in R, and lme4. The fixed effects will be defined as the moment of intervention, and group, and the random intercept will be defined as each participant. Parameter estimation will utilize restricted maximal likelihood and the Kenward-Roger approximation for degrees of freedom will be applied for evaluating significance, ensuring optimal type I error rates. Secondary analyses will explore adjustments for protocol deviations through a per-protocol analysis. Post a significant interaction, pairwise comparisons will be conducted with a Bonferroni adjustment. For main and interaction effects, partial eta squared (η2) will be computed as an effect size estimate. Effect sizes for pairwise comparisons will be reported as Cohen's d, complemented by 95% confidence intervals for mean differences. All statistical analyses will be conducted using R software, with a significant level of α = 0.05.