Cardiorespiratory Responses During High-Intensity Interval Training Prescribed by Rating of Perceived Exertion in Patients After Myocardial Infarction Enrolled in Early Outpatient Cardiac Rehabilitation

Objective: We aimed to determine the cardiorespiratory responses during, and adaptations to, high-intensity interval training (HIIT) prescribed using ratings of perceived exertion (RPE) in patients after myocardial infarction (MI) during early outpatient cardiac rehabilitation (CR). Methods: We prospectively recruited 29 MI patients after percutaneous coronary intervention who began CR within 2 weeks after hospital discharge. Eleven patients (seven men; four women; age: 61 ± 11 yrs) who completed ≥24 supervised HIIT sessions with metabolic gas exchange measured during HIIT once weekly for 8 weeks and performed pre- and post- CR cardiopulmonary exercise tests were included in the study. Each HIIT session consisted of 5–8 high-intensity intervals [HIIs, 1-min RPE 14–17 (Borg 6–20 scale)] and low-intensity intervals (LIIs, 4-min RPE < 12). Metabolic gas exchange, heart rate (HR), and blood pressure during HIIT were measured. Results: The mean oxygen uptake (V˙O2) during HIIs across 88 sessions of HIITs [91 (14)% of V˙O2peak, median (interquartile range, IQR)] was significantly higher than the lower limit of target V˙O2 zone (75% of V˙O2peak) recommended for the HII (p < 0.001). Exercise intensity during RPE-prescribed HIITs, determined as %V˙O2peak, was highly repeatable with intra-class correlations of 0.95 (95% CI 0.86– 0.99, p < 0.001). For cardiorespiratory adaptations from the first to the last session of HIIT, treadmill speed, treadmill grade, treadmill power, V˙O2HII, %V˙O2peak, and VE during HIIs were increased (all p < 0.05), while no difference was found for HR, %HRpeak and systolic blood pressure (all p > 0.05). V˙O2peak increased by an average of 9% from pre-CR to post-CR. No adverse events occurred. Conclusion: Our results demonstrate that HIIT can be effectively prescribed using RPE in MI patients during early outpatient CR. Participation in RPE-prescribed HIIT increases exercise workload and V˙O2 during exercise training without increased perception of effort or excessive increases in heart rate or blood pressure.


INTRODUCTION
Exercise-based cardiac rehabilitation (CR) is a secondary prevention tool used worldwide to improve physical function and prognosis in patients after myocardial infarction (MI) (1,2). High-intensity interval training (HIIT) has recently emerged as an alternative or adjunct strategy to traditional moderateintensity continuous training (3). HIIT involves alternating periods ranging from a few seconds to 4 min of higher intensity exercise [high-intensity intervals, HIIs: 85 to 95% of peak heart rate (HR) corresponding to 75 to 85% of peak oxygen uptake (VO 2 )] with 1 to 4 min of lower intensity exercise (low-intensity intervals, LIIs: <60% of peak HR) during an exercise session (4). HIIT has been shown to result in similar or greater improvements in aerobic capacity and other health outcomes compared to moderate-intensity continuous training (4). However, the relationships between patient safety, perception of effort, and cardiorespiratory responses and adaptations during HIIT sessions in patients after MI have not been reported. Gaps in our understanding of the relationship between effort perception and cardiorespiratory responses limit our ability to provide optimal guidance for prescription, implementation, and safety of HIIT in CR.
The most common metrics to prescribe aerobic exercise intensity during CR includeVO 2 , HR, and their derivative indicators such as percentages of predicted/peak HR andVO 2 ; reserves of HR andVO 2 ; and metabolic equivalents (METs) (5). During outpatient CR, continuous monitoring ofVO 2 is impractical and, while continuous HR monitoring is feasible, the high number of MI patients prescribed rate modulating pharmacotherapy (e.g., beta-blockers) makes HR a highly variable metric for exercise prescription (6). Furthermore, many patients who begin CR have not performed a graded exercise test, and peak HR has not been determined (7). For these patients, prescribing exercise intensity using predicted peak HR as a guide is imprecise.
Ratings of perceived exertion (RPE) are a practical alternative for prescribing exercise intensity and facilitates relative patient autonomy and progression of exercise intensity during CR (4,8). Our CR program has used RPE, accompanied by continuous HR and periodic blood pressure monitoring, to prescribe exercise intensity for several decades (9, 10). Our CR staff are experienced in instructing patients on the proper use of the 6-20 Borg RPE scale. Patients are carefully instructed on the use of RPE as part of their baseline graded exercise test and during their first supervised exercise session in CR (10). We have previously demonstrated that RPE-prescribed HIIT improves body composition, characteristics of metabolic syndrome, and cardiorespiratory fitness in patients after MI (11,12). However, the cardiorespiratory responses directly measured with metabolic gas exchange during, and adaptations to HIIT across several exercise sessions in MI patients have not been previously reported.
Therefore, this study aimed to determine the cardiorespiratory responses and adaptations during HIIT exercise sessions prescribed using RPE in patients with MI who participate in early outpatient CR. We hypothesized that: (1) Using RPE to prescribe exercise intensity will effectively elicit a desired HIIT cardiorespiratory response, and (2) RPE-based HIIT will result in an increasingVO 2 during exercise training across exercise sessions without increased perception of effort or excessive increases in heart rate and blood pressure.

Participants and Study Design
This prospective observational study initially recruited 29 consecutive MI patients with percutaneous coronary intervention who were referred to outpatient CR within 2 weeks of discharge from inpatient care (our traditional time to begin CR) at Mayo Clinic, Rochester, MN, USA from February 1 st , 2017, to September 30 th, 2018. Thirteen patients who did not perform a post-CR cardiopulmonary exercise test (CPET), two who refused to wear a metabolic gas collector/mask during exercise training, and three who changed their exercise type from treadmill to recumbent stationary cycle were excluded. Eleven patients [seven men, four women; age: 62 [11] yrs, median (interquartile range, IQR); BMI: 33.0 (7.2) kg/m 2 ; the interval between hospital dismissal and the start of CR: 14 [4] days] who completed ≥24 sessions of supervised HIIT on a treadmill with metabolic gas exchange measured during HIIT once per week for eight consecutive weeks and who performed pre and post CR CPET were included. Cardiovascular medications were unchanged during the study period. Participants were free of angina at low exercise intensities, symptomatic arrhythmias, symptomatic heart valve disease, musculoskeletal limitations to exercise training, and significant frailty or weakness (i.e., inability to engage in HIIT). Study procedures were approved by the Institutional Review Board for Research at Mayo Clinic (Rochester, MN, USA; 15-007977) and conformed to the standards set forth by the Declaration of Helsinki. Patients were informed regarding testing procedures and potential risks of participation before providing written, informed consent.

High-Intensity Interval Training Intervention
Our protocol for HIIT has been routinely applied in CR for more than a decade and was described previously (4,13). It was designed for routine use by patients who begin CR within 2 weeks of hospital dismissal and start HIIT after 1 week of moderate intensity aerobic exercise. Components of the exercise prescription were Frequency: set as three sessions per week for eight consecutive weeks. Intensity: HIIs lasting 1 min at an RPE 14-17 [Borg 6-20 RPE scale] followed by 4-min LIIs at RPE < 12, treadmill speed and grade were self-selected by patients themselves to achieve the target RPEs, and the RPE scores were obtained at the end of each interval. Type: a treadmill was used under continuous observation by clinical exercise physiologists. Time: initial time started at 30 min and gradually progressed to 40 min per session; Volume: at least 24 sessions of HIIT completed; Progression: the number of HIIs was gradually increased from 5 to 8 throughout the study according to the patients' expectations and the judgment of clinical exercise physiologists.
Patients were carefully instructed in the proper use of the Borg RPE scale during their first supervised exercise session in outpatient CR. All patients performed adaptive exercise training during the 1st week of CR (three sessions) using RPE ratings of 11-13 to facilitate a gradual accommodation to exercise training, ensure the ability to engage in sustained exercise for a minimum of 20 min, and to become accustomed to the use of the RPE scale. Following the gradual adaptation phase (week-1), the patients started HIIT. Each exercise session began with 5-10 min of lowintensity warm-up (RPE 8-10) and ended with a 5-min lowintensity cool down.

Cardiopulmonary Exercise Testing
CPETs were conducted by clinical exercise physiologists with cardiologist oversight at the beginning and end of CR. The exercise modality and end-test criteria were consistent between pre-and post-CR tests for all patients. Our operation and interpretation procedures for CPET have been described previously (13).

Metabolic Gas Exchange Measurements During HIIT
Breath-by-breathVO 2, carbon dioxide production (VCO 2 ), breathing frequency (f B ), and tidal volume (V T ) were measured continuously using a standard cardiorespiratory diagnostic system (Ultima Series 6 CPX TM , MGC Diagnostics Corporation, Minnesota, USA) during an RPE-prescribed HIIT sessions once each week. Continuous cardiorespiratory measurements were performed during a total of 88 HIIT exercise sessions. The cardiorespiratory diagnostic system was calibrated for flow and gas concentrations before each session according to the manufacturer's recommendations using a 3-liter syringe and calibration gases of known concentration. To minimize the influence of subsidiary work, and thereforeVO 2 and VCO 2 , patients were instructed to refrain from excessive stabilization (i.e., using handrails) during all exercise sessions. Minute V E was calculated as the product of V T and breathing frequency (f B ). The respiratory exchange ratio (RER) was calculated as the ratio of VCO 2 toVO 2 .VO 2HII andVO 2LII were calculated as the highest averageVO 2 of three consecutive breaths during the HIIs and LIIs (i.e., the highest single-breathVO 2 value and the preceding and following breaths), respectively. Other metabolic gas exchange values (i.e., RER, VCO 2 , V E , V E /VCO 2 ) were determined by averaging values of the final 15-s of HII and LII, respectively, of each HIIT session.

Heart Rate and Blood Pressure Measurements
During each HIIT session, HR and rhythm were continuously measured via electrocardiogram (ECG) telemetry (Q-Tel RMS, Welch Allyn, New York, USA). The HRs at the end of HIIs and LIIs were recorded. Systolic (SBP) and diastolic (DBP) blood pressures were measured via manual sphygmomanometer by clinical exercise physiologists at rest and during the final 15-s of HII and LII, respectively, of each HIIT session.

Sample Size Calculation
As the primary hypothesis of this study is that RPE prescribed exercise can effectively elicit the desired exercise intensity of HIIT (more than 75%VO 2peak ) in patients after MI, the %VO 2peak during HIIs (VO 2HII /VO 2peak × 100), the gold standard of exercise intensity, was set as the primary endpoint. %VO 2peak during HIIs was used to calculate the sample size. We applied the repeated measures analysis model of the Power Analysis & Sample Size software, version 15.0 (NCSS, LLC, USA) to calculate the sample size. The main parameters are as follows: mean %VO 2 peak during HIIs was 75%, eight sessions of HIITs data from consecutive 8 weeks were collected, the mean increase of %VO 2peak during HIIs was 10% across sessions, the standard deviation was 6%, autocorrelation was between 0.2 and 0.4, the dropout rate was estimated 20%. To achieve a power (1-β) of 90% with an α of 0.05, 11 participants were required.

Statistical Analysis
A minimum of five HIIs was performed by all patients during all HIIT sessions. Therefore, regardless of the number of HIIs performed (ranging from 5 to 8), the final five intervals were used to make comparisons among HIIT sessions. A familiarization HIIT session was used to ensure physiological responses were accurately characterized. The familiarization session was excluded from data analyses, and the second HIIT session was categorized as the first HIIT session. A total of 88 sessions of HIIT data (8 × 11) with gas exchange measurements were included in the analysis.
Data are presented as median [IQR] for continuous variables and frequency and percentage for categorical variables. Repeatability analysis of exercise intensity in terms of %VO 2peak was performed with intra-class correlation (ICC) (14) using a random-effects model. Exercise workload (ie., treadmill speed, grade and power in watts) and cardiorespiratory variables measured during the HIIT training sessions (i.e., RPE,VO 2 , %VO 2peak , V E , HR, %HR, BP, energy expenditure [EE] per min and per session) were compared within (HIIs vs. LIIs) and between sessions (first vs. last) using repeated-measures analysis of variance (ANOVA). Pre-and post-CR CPET measurements Exercise workload and cardiorespiratory responses to RPE prescribed HIIT are presented in Figure 1 and Table 2. The highest meanVO 2 during HIIs of 88 sessions of HIITs [91 (14)% ofVO 2peak , median (IQR)] was significantly higher than for the targetVO 2 (75% ofVO 2peak ) recommended for HIIs (p < 0.001). The ICC of exercise intensity, %VO 2peak , between the RPE-prescribed HIIT sessions was 0.95 (95% CI, 0.86 to 0.99, p < 0.001). The values of treadmill speed, treadmill grade, power,VO 2 , %VO 2peak , HR, %HR peak , SBP, V E , V T and f B in the HIIs were significantly greater than those in the LIIs during each exercise session (all p < 0.01), which was consistent with the values of RPE during the HIIs vs. during the LIIs [15 (2) vs. 11 (2), median (IQR), p < 0.001]. A difference of 9-11% between %HR peak and %VO 2peak was present for the LIIs and is consistent with conventional wisdom as reported in the literature (16) that %HR peak is greater than %VO 2peak at a constant workrate. However, for the HIIs, %HR peak was not higher than %VO 2peak . For the HIIs of the first HIIT session, median %HR peak and %VO 2peak were identical (88%), while for the final session %VO 2peak was greater than %HR peak (97% vs. 90%) during HIIs. No differences were found in DBP and RER between HIIs and LIIs (all p > 0.05).
Comparisons of the first vs. the last exercise sessions to assess cardiorespiratory adaptations during RPE-prescribed HIIT are presented in Table 2 and Figure 2. No differences were found for RPE, HR, %HR peak , V T , f B , V E /VCO 2 , SBP and DBP (all p > 0.05) between the first and last session for both HIIs and LIIs. However, treadmill speed, treadmill grade, power,VO 2 , %VO 2peak , EE per minute and per session and V E increased significantly from the first to the last session for the HIIs (all p < 0.05), while no changes were detected for the LIIs (all p > 0.05). No adverse events related to exercise training occurred during the study.  Repeated-measures ANOVA was used for all assessments. Data were expressed as mean + up limit of 95% confidence interval for high-intensity intervals and mean-low limit of 95% confidence interval for low-intensity intervals in (B-G). *Significantly higher than low-intensity interval, p < 0.001.

DISCUSSION
Our study provided unique metabolic gas exchange data obtained during HIIT sessions and established the efficacy of using selfselected exercise intensity based on a target RPE range as a safe and practical method of prescribing HIIT for patients after MI during early outpatient CR. Our method of prescribing 5-8 one minute HIIs with RPE 14-17 interspersed with 4-min LIIs (RPE <12) during a 40-min session of training was effective in eliciting aVO 2 of >95% of pre-trainingVO 2peak during the final HIIT session. We demonstrated that over 8 weeks of HIIT, patients were able to exercise at a higherVO 2 without a concurrent increase in RPE or excessive HR and blood pressure response. Patients were able to increase the rate of energy expenditure during the HIIT sessions without an increased perception of effort which is potentially clinically important for decreasing body fat stores with the attendant metabolic health benefits.
Aamot et al. (17) reported that using RPE to prescribe exercise intensity during HIIT resulted in a lower than expected intensity based on %HR peak , (detected 82% HR peak vs expected 85% HR peak ) during HIIs (18) in patients with coronary artery disease in which 80% patients regularly used beta-blockers. In our study, we utilized both %VO 2peak (gold standard to reflect exercise intensity) and %HR peak to assess the validity of RPE-prescribed HIIT. Both %VO 2peak and %HR peak achieved during HIIs were more than the required levels to meet the designation of high-intensity (75%VO 2peak and 85% HR peak ) for all sessions.VO 2HII increased from 88% oḟ VO 2peak in the first HIIT session to 97% ofVO 2peak in the last session.
Patient progression in exercise training dose without the sacrifice of safety is a core tenet of cardiac rehabilitation. We observed that despite an increasedVO 2HII across exercise sessions, no significant increases in HR, DBP, SBP, or perception of effort were found during RPE-prescribed HIIT. The underlying reasons for this remain unclear.VO 2 is an integrated indicator of the systems that transport and utilize oxygen, including the respiratory (oxygen uptake from the atmosphere), heart (oxygen transport), peripheral vasculature (oxygen transport, tissue perfusion, tissue diffusion), and skeletal muscle (oxygen extraction and utilization) (19,20). In the present study, HR, O 2 pulse [a surrogate for stroke volume (21)], (G,F) Present the changes in the relationship of %VO 2peak and %HR peak between the first and last HIIT sessions during high-and low-intensity intervals, respectively. HII is high-intensity interval. LII is low-intensity interval. Repeated-measures ANOVA was used for all assessments. was used for all assessments. Data were expressed as mean difference [95% confidence interval]. V E /VCO 2 [an indicator of ventilatory efficiency (22)], RER (a variable to reflect degree of exertion) were not significantly changed during 8 weeks of RPE-prescribed HIIT sessions, which suggests that peripheral vasculature and skeletal muscle adaptations may have contributed to the increase inVO 2HII across sessions. This hypothesis is supported by our pre-clinical studies in mice that demonstrated regular exercise improved the structure and function of the aortic endothelium (23) and mitochondria in skeletal muscle (24,25). However, additional research on the mechanisms responsible for these observations is needed.
In order to evaluate the cardiopulmonary adaptations during 8 weeks of HIIT sessions., we studied the relationship between %VO 2peak and %HR peak during the HIIT sessions. During LIIs, the values for %VO 2peak were 61% for both first and last sessions, and the corresponding %HR peak values were 70 and 72%, respectively. This is consistent with previous reports, where %HR peak was ∼10% higher than %VO 2peak (26). However, during HIIs, the %VO 2peak was 88% for the first session and 97% for the last session, while the corresponding %HR peak remained ∼90% for both sessions. This is a clear disconnect from the assumed relationship of %VO 2peak and %HR peak . Though it has been assumed that the %HR peak -%VO 2peak relationship holds during HIIT, the expected linear relationship between %VO 2peak and %HR peak was established during graded exercise testing with cardiopulmonary measurements and may differ during HIIT. Further studies are needed to elucidate the mechanisms responsible for these observations.
The effect of HIIT on cardiorespiratory fitness in patients with coronary artery disease has been reported, with meanVO 2peak increases ranging from 11 to 25% (4). A recent study from our group demonstrated that RPE-prescribed HIIT during early outpatient CR significantly improvedVO 2peak by 18% (pre-CR vs. post-CR, 23.0 ± 6.3 vs. 28.0 ± 5.9; mean change 5.0 ± 2.5 ml·kg −1 ·min −1 ) in 42 MI patients (12). In the present study, HIIT improved cardiorespiratory fitness (i.e.,VO 2peak ) in 9 of 11 (82%) patients, with a mean improvement of only 9%. Possibly related to the small sample size in the current study, statistically significant changes were found inVO 2peak related to body mass and % predictedVO 2peak , but not inVO 2peak independent of body mass. The percentage of non-improvement (non-responder: failure to improveVO 2peak ) in CR was 18% in the present study, which is consistent with the data reported in the study by Savage et al. (27) in which 81 out of 385 patients (21%) failed to improveVO 2peak during outpatient CR using moderate-intensity continuous training. Our finding of non-improvement inVO 2peak with HIIT in some patients is a novel finding. Non-improvement in CR may be associated with exercise intensity, comorbidity score, self-reported physical function, diabetes, and baselineVO 2peak (27). In the present study, mean baselineVO 2peak was normal and may be a factor in our findings of a less than typical increase inVO 2peak and identification of non-responders with HIIT.
The present study did not assess change in left ventricular systolic or diastolic function resulting from HIIT. The literature suggested that HIIT is an effective strategy to attenuate left ventricular remodeling in clinically stable heart failure patients with reduced ejection fraction (28). The effect of HIIT on left ventricular function in heart failure with preserved ejection fraction is controversial (29), while the positive effects of HIIT on exercise capacity and quality of life in patients with MI and heart failure have been reported (4). The patients in the present study were not diagnosed with HFpEF. Further studies are warranted to investigate the effects of HIIT on cardiac function in patients after MI and heart failure with preserved ejection fraction.

LIMITATIONS
Our study examined a single, unique HIIT protocol in MI patients and may not be generalizable to other methods of prescribing HIIT or to other clinical populations. Because measuring metabolic gas exchange data during multiple 30-40 min CR exercise sessions is technically and logistically challenging, we studied only a limited number of patients. Our patients' average baselineVO 2peak was in the normal range for healthy individuals and our subjects are not representative of typical post-MI patients. In addition, we did not compare RPE vs. HR-based prescriptions for HIIT. While RPE certainly appears to be an effective prescriptive tool for HIIT, it remains unknown if RPE is the optimal prescription method despite its previously discussed advantages. Additionally, the cardiorespiratory assessments made during HIIT did not include direct measures of cardiac function (e.g., echocardiogram) and relied on an indirect method for cardiac adaptations (i.e., VO 2 and V E /VCO 2 ). As such, future studies should consider performing more comprehensive and direct measurements to identify the specific central and peripheral mechanisms responsible for the cardiorespiratory adaptations to RPEprescribed HIIT in patients after MI.

CONCLUSIONS
RPE is an effective and safe method for prescribing HIIT for patients enrolled in early outpatient CR after uncomplicated MI. Using RPE eliminates reliance on heart rate for exercise intensity prescription and may be advantageous for patients who do not perform a pre-CR exercise test and for individuals receiving heart rate modulating medications. Using an RPE target of 14-17 during 1-min of high-intensity exercise elicits a robustVO 2HII of >90% ofVO 2peak. The expected relationship between %HR peak and %VO 2peak (%HR peak > %VO 2peak ) is not present during the HIIs of the HIIT. Patients are comfortable performing 5-8 oneminute intervals at >90% ofVO 2peak during a 40-minute aerobic exercise session. Over the course of eight weeks of HIIT-based CR, patients increased treadmill speed and grade, andVO 2HII without an increase in perception of effort or excessive increases in heart rate and blood pressure.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Mayo Clinic Institutional Review Board. The patients/participants provided their written informed consent to participate in this study.