Adherence to High-Intensity Interval Training in Cardiac Rehabilitation

A REVIEW AND RECOMMENDATIONS

Abstract

Purpose: 

High-intensity interval training (HIIT) is gaining popularity as a training approach for patients attending cardiac rehabilitation (CR). While the literature has focused on the efficacy of HIIT for improving cardiorespiratory fitness (CRF), particularly when compared with moderate intensity exercise, less emphasis has been placed on adherence to HIIT. The aim of this review was to summarize the current literature regarding adherence to HIIT in CR patients with coronary artery disease.

Review Methods: 

A review identified 36 studies investigating HIIT in CR patients with coronary artery disease. Methods and data were extracted for exercise or training adherence (to attendance, intensity, and duration), feasibility of protocols, and CRF. The review summarizes reporting of adherence; adherence to HIIT and comparator/s; the influence of adherence on changes in CRF; and feasibility of HIIT.

Summary: 

Adherence to the attendance of HIIT sessions was high and comparable with moderate-intensity exercise. However, adherence to the intensity and duration of HIIT was variable and underreported, which has implications for determining the treatment effect of the exercise interventions being compared. Furthermore, additional research is needed to investigate the utility of home-based HIIT and long-term adherence to HIIT following supervised programs. This review provides recommendations for researchers in the measurement and reporting of adherence to HIIT and other exercise interventions to facilitate a sufficient and consistent approach for future studies. This article also highlights strategies for clinicians to improve adherence, feasibility, and enjoyment of HIIT for their patients.

In Brief

This review summarizes the current literature regarding adherence to high-intensity interval training (HIIT) in cardiac rehabilitation patients. Adherence to attendance of HIIT is high and comparable to moderateintensity exercise; however, adherence to intensity and duration is underreported. This review provides recommendations for researchers and clinicians to improve reporting and adherence to HIIT.

High-intensity interval training (HIIT) involves alternating periods of high-intensity intervals (eg, 85-95% maximal heart rate [HR_max] or rating of perceived exertion [RPE]: 15-18) with periods of lower intensity or no exercise.¹ As HIIT involves submaximal or near-maximal efforts (≤100% maximal oxygen uptake or ≤95% HR_max), it is more commonly applied to cardiac populations rather than sprint interval training, which involves periods of supramaximal or all-out sprint efforts.¹ There is interest in the use of HIIT as a training modality for patients attending cardiac rehabilitation (CR), given the efficacy for increasing cardiorespiratory fitness (CRF; peak oxygen uptake [V˙o_2peak]). Improvement in V˙o_2peak is doubled with HIIT compared with traditional moderate-intensity continuous training (MICT) in patients with cardiometabolic disease, while limiting the time commitment to training.¹ Furthermore, the reported difference between HIIT and MICT on V˙o_2peak improvement in patients with coronary artery disease (1.4-1.8 mL/kg/min)^2–4 is clinically meaningful, with each 1 mL/kg/min improvement in V˙o_2peak associated with a 6% reduction in hospital readmissions and a 13% decrease in all-cause mortality.⁵

International guidelines have called for further evidence on the feasibility, safety, and adherence to HIIT, before it is recognized as a standard option for patients attending CR.⁶ There are limited collective data on the adherence to, and feasibility of, HIIT in patients with cardiovascular disease. Quindry et al⁷ reported that adherence to supervised HIIT appeared similar to MICT based on study dropout rates. However, additional research is needed to fully understand the dynamics of adherence within supervised and home-based HIIT interventions, as this is fundamental to long-term outcomes.⁷ Exercise intensity does not appear to influence the degree of adherence,⁸ which is more affected by training mode (running/walking more than cycling), training duration, and overall time commitment.⁸

The Consensus of Exercise Reporting Template (CERT) was developed by a team of international experts to improve the reporting of exercise interventions in exercise research across all study designs⁹ as an adjunct to other reporting frameworks (eg, CONSORT, EQUATOR). The CERT was developed to increase clinical uptake of exercise interventions, enable research replication, reduce research waste, and improve patient outcomes.⁹ Three components in CERT that specifically relate to describing adherence to exercise interventions include (1) how adherence to exercise is measured and reported, (2) how adherence or fidelity to the exercise is assessed/measured, and (3) the extent to which the intervention was delivered as planned.⁹ As well as providing an explanation for the effect or lack of effect of an intervention, adherence data provide valuable information to inform future studies and clinical translation.⁹

The aim of this review is to summarize the current literature regarding adherence to HIIT in CR. Specifically, we review how adherence is reported in the literature, adherence results, the influence of adherence on changes in CRF, and feasibility of HIIT, and provide recommendations for future studies.

REVIEW OF THE LITERATURE

A review of the literature was conducted, including full-text studies published in English until September 2020. We searched electronic databases PubMed, MEDLINE, ProQuest, Scopus, and Web of Science using key words “high-intensity interval training,” “interval training,” “interval exercise,” “cardiac rehabilitation,” “coronary artery disease,” “coronary heart disease,” “acute coronary syndrome,” and “myocardial infarction.” We included studies involving exercise interventions with intervals of high-intensity exercise separated with active or passive recovery. High-intensity exercise was defined as ≥85% peak heart rate (HR_peak) or HR_max; ≥85% V˙o_2peak or maximum oxygen uptake (V˙o_2max); ≥80% heart rate reserve or V˙o₂ reserve; ≥15 on the Borg rating of perceived exertion scale¹⁰ (RPE); or ≥100% peak power output. Studies specifically involving patients with heart failure were beyond the scope of this review. A total of 36 distinct studies (from 52 publications) met our criteria, including 28 randomized controlled trials, one observational study, three retrospective studies, and four long-term follow-up studies. Study characteristics are outlined in Supplemental Digital Content 1, available at: https://links.lww.com/JCRP/A217. Exercise adherence methodology and results are outlined in Table 1 and Supplemental Digital Content 2, available at: https://links.lww.com/JCRP/A218. Female participants were underrepresented, averaging 20 ± 16% of recruited subjects. For studies with multiple publications, all related articles have been cited together in Supplemental Digital Content 1, available at: https://links.lww.com/JCRP/A217, with the original publication. Since adherence to an exercise intervention is related to whether the proposed exercise prescription is achieved, adherence data were reviewed and extracted for exercise prescription components “frequency (or attendance),” “intensity,” and “time (or duration)” of the prescribed type of exercise (ie, HIIT).

Table 1 - Exercise Adherence Methodology and Results From Randomized Controlled Trials^a

Study (Author, Year, Country)	Exercise Prescriptionb	Adherence Methodologyb	Adherence Resultsb	Improvement in Exercise Capacity (%, Mean Change)
Aamot et al (2014)14, Norway	HIIT: 85–95%HRpeak MICT: Not applicable	Adherence to attendance: Number of sessions completed—70% criteria (≥17/24 sessions) Adherence to intensity and/or duration: monitored by HR monitors and assessed as (1) achieving target HR (2) mean HR in final 2 min of work intervals; and (3) minutes spent in target HR	Criteria for attendance: Patients meeting attendance criteria (70%): Treadmill group (TE): 100%, group exercise (GE): 100%, home-based group (HE): 87% (26/30) The main reason for nonattendance was lack of time. All participants achieved the target HR. Mean training HR: TE = 90% HRpeak, GE = 89% HRpeak, HE = 90% HRpeak Minutes spent in target HR: TE = 10.3±2.8 min, GE = 10.2±4.2 min, and HE = 12.3±4.5 min	Assessed as V˙o 2peak 12 wk TE: 12% (+4.3 mL/kg/min) GE: 10% (+3.3 mL/kg/min) HE: 8% (+2.8 mL/kg/min)
Abdelhalem et al (2018)35, Egypt	HIIT: 85-95% HRR MICT: 60% HRR	Adherence to attendance: Number of sessions completed Adherence to intensity and/or duration: no methods reported	Attendance: All participants completed the program with no missing sessions. Training intensity and/or duration: data not reported.	Assessed as treadmill METs 12 wk HIIT: 53% (+4.0 METs) MICT: 30% (+2.5 METs)
Boidin et al (2019)15, Canada	HIIT: 100% PPO MICT: 60% PPO	Adherence to attendance: Number of sessions completed—66% criteria applied for per-protocol analysis. Adherence to intensity and/or duration: no methods reported	Attendance: Completion of sessions: HIIT = 99%, and MICT = 100% Study dropout was 25% (6/24) for HIIT and 0/19 for MICT. Training intensity and/or duration: data not reported.	Not reported
Cardozo et al (2015)16, Brazil	HIIT: 90% HRpeak MICT: 75%HRpeak	Adherence to attendance: Number of sessions completed—75% criteria applied for analysis exclusion Adherence to intensity and/or duration: no methods reported	Attendance: All participants completed the program. Training intensity and/or duration: data not reported.	Assessed as V˙o 2peak 16 wk HIIT: 18% (+3.8 mL/kg/min) MICT: 0% (+0.1 mL/kg/min)
Conraads et al (2015)22, Belgium (SAINTEX-CAD Study)	HIIT: 90-95% HRpeak MICT: 65-75% HRpeak	Adherence to attendance: Number of sessions completed—32/36 criteria applied for per-protocol analysis or ≥10 training sessions in final 4 wk. Adherence to intensity and/or duration: Mean training HRs/workloads calculated by averaging 4 × HRs/workloads of each training session (HIIT: measured at end of work interval; MICT: measured every 10 min). Mean HRs/workloads expressed as %HRpeak/workload from most recent CPX. RPE was measured at the end of the training session.	Attendance: Sessions completed: HIIT = 99% (35.7 ± 1.1), MICT = 99% (35.6 ± 1.5) Mean training HR: HIIT = 88% HRpeak, MICT = 80% HRpeak (P = .001) Mean Training RPE: HIIT = 13.5 ± 1.6, MICT = 12.5 ± 1.5 (P = .001) Mean training workload: HIIT = 86% PWC, MICT = 63% PWC (P = .001)	Assessed as V˙o 2peak 6 wk HIIT: 15% (+3.2 mL/kg/min) MICT: 13% (+2.8 mL/kg/min) 12 wk HIIT: 23% (+5.1 mL/kg/min) MICT: 20% (+4.4 mL/kg/min) Per-protocol analyses were not different
Currie et al (2013)26, Canada	HIIT: 89-110% PPO MICT: 51-65% PPO	Adherence to attendance: Number of sessions completed—no criteria specified Adherence to intensity and/or duration: no methods reported	Attendance: sessions completed: Supervised training: HIIT = 80% (19 ± 4), MICT completed 92% (22 ± 3) (P >.05). Home-based training: HIIT = 11 ± 10, MICT = 14 ±1 4 Mean training duration: HIIT = 40 ± 17 min/session, MICT = 45 ± 3 min/session. Mean training HR: During supervised training, HIIT = 73 ± 10% of age-predicted HRmax, and MICT = 65 ± 4% of age-predicted HRmax (P < .005) For home-based training HIIT = 68 ± 5% of age-predicted HRmax, and MICT = 60 ± 7% of age-predicted HRmax (P < .005)	Assessed as V˙o 2peak 12 wk HIIT: 24% (+4.7 mL/kg/min) MICT: 19% (+3.6 mL/kg/min)
Ha-Yoon et al (2018)17, Korea	HIIT: 85-100% HRmax MICT: 60-70% HRmax	Adherence to attendance: Number of sessions completed—18/18 criteria applied for study exclusion. Adherence to intensity and/or duration: no methods reported	Attendance: patients completing 100% sessions: HIIT = 96% (23/24) and 95% (21/22) for MICT. Dropout rates = 1/24 for HIIT (ankle injury) and 1/22 for MICT (occupational reason). Training intensity and/or duration: data not reported. Investigators reported that exercise was performed with sufficient intensity to reach the prescribed HR value.	Assessed as V˙o 2peak 9-10 wk HIIT: 24% (+7.6 mL/kg/min) MICT: 9% (+2.4 mL/kg/min)
Heber et al (2020)39 Austria	HIIT: 100% PPO MICT: 60% PPO	Adherence to attendance: Number of sessions completed—no criteria specified. Adherence to intensity and/or duration: no methods reported	Attendance: Completion of training protocol according to randomization: HIIT = 68% (27/40) and 79% (33/42) for MICT (P = .26). Protocol deviation: 13% (5/40) patients unable or unwilling to perform HIIT (swapped to MICT). Reasons given were knee pain, discomfort, and lack of motivation. 1/42 patients unwilling to perform MICT (swapped to HIIT) Dropout rates = 8% (11/40) in HIIT, 21% (9/42) in MICT. Percentage of time spent in target HR: HIIT = 94% (85-100%) for wk 1-6 and 79% (61-88%) for wk 7-12 MICT = 93% (84-100%) for wk 1-6 and 76% (67-91%) for wk 7-12. Mean training HR: HIIT = 92 ± 7% HRmax, MICT not reported. Home-based training not evaluated because of lack of compliance with self-report documentation.	Assessed as V˙o 2peak Data not reported, only displayed graphically. Improvement in V˙o 2peak was not statistically different between groups.
Jayo-Montoya et al (2020)12, Spain (INTERFARCT Project)	HIIT: 85-95% HRpeak MICT: Not specified	Adherence to attendance: Number of sessions completed—80% criteria applied to define adherence. Adherence to intensity and/or duration: Exercise intensity controlled and monitored by HR monitors and RPE.	Attendance: Not reported. Dropout rates: HIIT = 21% (12/56), control = 21% (3/14) Training intensity and/or duration: Data not reported.	Assessed as V˙o 2peak 16 wk Control: −2% (−0.5 mL/kg/min) LV HIIT: 15% (+3.5 mL/kg/min) HV HIIT: 22% (+5.0 mL/kg/min)
Karlsen et al (2008)29, Norway	HIIT: 85-95% HRpeak MICT: Not applicable	Adherence to attendance: Number of sessions completed—no criteria specified. Adherence to intensity and/or duration: HR and RPE monitored and recorded. RPE after the training session.	Attendance: Participants attended 97% (29/30) of sessions. Mean training RPE: 14.4 ± 1 (after training session)	Assessed as V˙o 2peak 10 wk HIIT: 16% (+4.5 mL/kg/min)
Keteyian et al (2014)28, United States	HIIT: 80-90% HRR MICT: 60-80% HRR	Adherence to attendance: Number of sessions completed—no criteria specified. Adherence to intensity and/or duration: HR monitored continuously by HR monitor and used by staff/patients to adjust workload. RPE was reported during the session (during work and recovery intervals for HIIT).	Attendance: Number of sessions completed: HIIT = 29 ± 1 and MICT = 29 ± 2. Mean training RPE: HIIT = 15 ± 1 for work intervals and 12 ± 1 for recovery MICT = 12 ± 1 during session Mean training HR: Not reported	Assessed as V˙o 2peak 10 wk HIIT: 16% (+3.6 mL/kg/min) MICT: 7% (+1.7 mL/kg/min)
Kim et al (2015)40, South Korea	HIIT: 85-95% HRR MICT: 70-85% HRR	Adherence to attendance: Number of sessions completed—no criteria specified. Adherence to intensity and/or duration: All training sessions monitored by HR. RPE was assessed during and after each training session. Workload was adjusted continuously to maintain HR target.	Attendance: Patients completing 100% of sessions: HIIT = 88% (14/16) and MICT = 100% (16/16) Training HR: All HIIT subjects achieved 85% HRR within 1 min after recovery interval. The target HR was achieved during 86% of the high-intensity interval duration. Training HR data were not reported for MICT. Training RPE: No differences reported between HIIT and MICT (RPE data not provided).	Assessed as V˙o 2peak 6 wk HIIT: 22% (+6.4 mL/kg/min) MICT: 8% (+2.5 mL/kg/min)
Kim and Choi (2020)36, South Korea	HIIT: 85% HRR Maximal HIIT: 95-100% of HRR	Adherence to attendance: Number of sessions completed—no criteria specified. Adherence to intensity and/or duration: All training sessions were monitored by HR and RPE. Workload was adjusted continuously to maintain HR target.	Adherence: Patients who attended all training sessions: HIIT = 80% (24/30) and maximal HIIT = 77% (23/30) Training intensity and/or duration: Data not reported.	Assessed as V˙o 2peak 4 wk HIIT: 17% (+3.7 mL/kg/min) Maximum HIIT: 31% (+8.3 mL/kg/min)
Lee et al (2019)25, Canada	HIIT: 90-95% HRpeak, and/or RPE ≥17 MICT: 60-80% V˙o 2peak.	Adherence to attendance: Number of sessions completed—no criteria specified. Frequency and duration recorded in a self-report weekly exercise diary. Adherence to intensity and/or duration: During supervised sessions, HR and speed data were monitored using a GPS watch monitor. During home-based training, HR and RPE were recorded in a self-report weekly exercise diary.	Adherence: Proportion of exercise sessions completed: For all sessions: HIIT = 76 ± 14% and MICT 72 ± 15% (P >.05) For supervised sessions: HIIT = 84% (21/25) and MICT = 80% (20/25). Mean training HR: HIIT = 89 ± 3% HRpeak and MICT = 68 ± 7% HRpeak Mean training RPE: HIIT = RPE 16.7 ± 0.6, and MICT = 11.2 ± 1.3 (P < .01).	Assessed as V˙o 2peak 26 wk HIIT: 6% (+1.3 mL/kg/min) MICT: 2% (+0.4 mL/kg/min)
Madssen et al (2014)38, Norway	HIIT: 85-95% HRpeak MICT: 70% at HRpeak	Adherence to attendance: Number of sessions completed—no criteria specified. Adherence to intensity and/or duration: HR monitors were used to help achieve target intensity	Adherence: All patients attended > 90% (32/36) of sessions. Training intensity and/or duration: Data not reported.	Assessed as V˙o 2peak 12 wk HIIT: 11% (+3.3 mL/kg/min) MICT: 7% (+2.0 mL/kg/min)
Madssen et al (2014)41, Norway	HIIT: 85-95% HRpeak MICT: Not applicable	Adherence to attendance: Number of sessions/wk—criteria not specified. Adherence to intensity and/or duration: Assessed as physical activity level via self-report questionnaires asking how often they exercised/wk (<1, 2, 3, and >4 sessions/wk), the mean duration of each exercise session (30, 30-45, 45-60, and >60 min/session). Self-report questionnaires also asked the mean intensity level during sessions (light, medium, or hard intensity). Time spent in moderate or vigorous physical activity was objectively measured by accelerometry in a subgroup of n = 18 (37%).	Attendance: Completion of sessions was 98% (7.8/8) in maintenance and control groups. Self-reported exercise frequency (maintenance vs control): ≥4 times/wk: 19% (4/21) vs 4% (1/22) 2-3 times/wk = 76% (16/21) vs 77% (17/22) <1 time/wk = 0% vs 22% (5/23) Self-reported exercise duration (maintenance vs control): >60 min: 29% (6/21) vs 32% (7/22) 45-60 min: 38% (8/21) vs 23% (5/22) 30-45 min: 29% (6/21) vs 23% (5/22) 0-30 min: 5% (1/21) vs 23% (5/22) Self-reported exercise intensity (maintenance vs control): at high intensity = 43% (9/21) vs 35% (6/22) at moderate intensity = 52% (11/21) vs 59% (10/22) Time spent at moderate intensity: Maintenance = 58 min/d, control = 111 min/d Time spent at high intensity: Maintenance = 3 min/d, control = 7 min/d	Assessed as V˙o 2peak 12 mo Maintenance group: 3% (+0.9 mL/kg/min) Control group: 3% (+0.8 mL/kg/min)
Moholdt et al (2009)20, Norway	HIIT: 90% HRpeak MICT: 70% HRpeak	Adherence to attendance: Number of sessions completed—70% criteria Frequency of sessions was recorded by self-report exercise diary. Adherence to intensity and/or duration: Assessed by HR monitors and RPE on 0-10 Borg scale. HR and RPE were assessed during the work intervals.	Attendance: Completion of supervised sessions: HIIT = 82% (16.4/20), and MICT = 84% (16.7/20) In addition, all subjects completed an additional 16.5 (range: 12-19) other training sessions during 4 wk. Self-reported exercise sessions at 6-mo: ≥3 times/wk HIIT = 74% (17/23), MICT = 68% (17/24) 1-2 times/wk: HIIT = 13% (3/23), MICT = 20% (5/24) <1 times/wk: HIIT = 13% (3/23), MICT = 8% (2/24) Mean training HR: HIIT = 92 ± 5% HRpeak and MICT = 74 ± 4% HRpeak Mean training RPE: HIIT = 6.1 ± 0.9 (during work intervals), MICT = 2.9 ± 0.9. In the other training sessions at the center, the patients exercised with high intensity in 31% and moderate intensity in 69% of the registered sessions, respectively. Adherence to randomized training at 6 mo: ≥3 times/wk: HIIT = 52% (12/23), MICT = 64% (16/25) Protocol deviation at 6 mo: HIIT = 22% (5/23) doing moderate intensity exercise MICT: 4% (1/25) doing higher intensity exercise	Assessed as V˙o 2peak 4 wk HIIT: 12% (+3.3 mL/kg/min) MICT: 9% (+2.3 mL/kg/min) 6 mo HIIT: 19% (+5.1 mL/kg/min) MICT: 13% (+3.3 mL/kg/min)
Moholdt et al (2012)24, Norway	HIIT: 90% HRpeak Usual care: Not specified	Adherence to attendance: Number of sessions completed—no criteria specified. Adherence to intensity and/or duration: All patients wore HR monitors. Training HR was calculated from HR in 8 min with the highest intensity throughout each session. For HIIT, this was the last 2 min of each work interval. For the usual care exercise, this was selected by visual inspection.	Attendance: Completion of sessions: HIIT = 83% (20.4 ± 5.0), usual care = 79% (19.1 ± 4.0) Mean training HR: HIIT = 87 ± 4% HRpeak, usual care = 78 ± 7% HRpeak (P < .001)	Assessed as V˙o 2peak 12 wk HIIT: 15% (+4.6 mL/kg/min) Usual care: 8% (+2.5 mL/kg/min)
Moholdt et al (2012)30, Norway	HIIT: 90% HRpeak Residential group: Light, moderate, high (RPE: 11-17)	Adherence to attendance: Number of sessions completed—only in HIIT group. Assessed with self-report exercise diary. Adherence to intensity and/or duration: No methods reported	Attendance: Number of weekly sessions: HIIT group reported in 1.6 ± 1.6 HIIT sessions/wk and 2.4 ± 1.9 moderate intensity sessions/wk Exercise adherence not reported for residential group. Self-reported exercise sessions at 6 mo: 42% (5/12) achieved HIIT 3 times/wk 17% (2/12) reported a combination of HIIT and MICT 33% (4/12) reported doing MICT instead of HIIT 1/12 did no exercise.	Assessed as V˙o 2peak 4 wk Usual care: 17% (+3.9 mL/kg/min) 26 wk HIIT: 19% (+4.6 mL/kg/min)
Munk et al (2009)37, Norway	HIIT: 80-90% HRmax Usual care: Not applicable	Adherence to attendance: Number of sessions completed—no criteria specified. Adherence to intensity and/or duration: Training sessions were monitored with HR watches allowing the patient to achieve the target HR.	Attendance: Patients attending >90% sessions = 95% (19/20). Training intensity and/or duration: Data not reported.	Assessed as V˙o 2peak 26 wk HIIT: 17% (+3.9 mL/kg/min) Control: 8% (+1.5 mL/kg/min)
Pedersen et al (2015)13, Norway (The CUT-IT Trial)	HIIT: 85-90% HRpeak MICT: Not applicable	Adherence to attendance: Number of sessions completed—≥60% criteria of attendance used for per-protocol analysis. Adherence to intensity and/or duration: monitored using HR monitors, lactate measurements, and the RPE Borg Scale.	Attendance: Patients who completed the study per protocol with valid measurements were 74% (26/35). Training intensity and/or duration: Data not reported.	Assessed as V˙o 2peak (relative to fat-free mass) 12 wk HIIT: 11% (+12.9 mL/kg/min need)
Rognmo et al (2004)21, Norway	HIIT: 80-90% V˙o 2peak 85-95% HRpeak MICT: 50-60% V˙o 2peak 65-75% HRpeak	Adherence to attendance: Number of sessions completed—70% criteria applied for adherence Adherence to intensity and/or duration: All subjects used HR monitors during every training session and were encouraged to exercise as close to the upper intensity border as possible. Workload was continually adjusted to maintain desired HR target. Borg RPE (6-20 scale) was recorded at the end of the training session.	Attendance: Completion of sessions: HIIT = 94% (28.3, range: 25-30), MICT = 85% (25.4, range: 21-30) (P = .074) Mean training RPE (at the end of the session) HIIT = 14.4 and MICT = 13.5 (P = .093). Mean training HR: data not reported.	Assessed as V˙o 2peak 10 wk HIIT: 18% (+6.0 mL/kg/min) MICT: 8% (+2.7 mL/kg/min)
Taylor et al (2020)11, Australia (The FITR-Heart Study)	HIIT: RPE 15-17 MICT: RPE 11-13	Adherence to attendance: Assessed as the number of randomized sessions completed with sufficient duration—70% criteria (adherent if 70% of sessions completed with 70% duration). During home-based training, frequency and duration were recorded in self-report weekly exercise diaries. Adherence to intensity and/or duration: Assessed as the number of sessions adherent to prescribed intensity. Deemed adherent if average training intensity was within RPE ≥15 (or ≥85% HRpeak) for HIIT during work intervals and 11-13 for MICT during session. During the supervised training, HR was measured by ECG, pulse oximetry, or HR monitor. For MICT, HR was measured >10 min into session and RPE at the end of each exercise modality. For HIIT, the highest HR and RPE were recorded for each high-intensity interval. Exercise intensity was then assessed as average training RPE, peak training RPE, average training %HRpeak, and peak training %HRpeak. Average RPE and HR were calculated by averaging the measurements taken during each exercise modality (eg, treadmill, bike) for MICT, or each high-intensity interval for HIIT. Training HR was not measured continuously throughout the entire exercise session. During home-based training, RPE was recorded by self-report exercise diary. The HIIT group used HR monitors for the initial 3 mo of training. They were instructed to record HR and RPE as per the supervised sessions.	Criteria for attendance: Patients meeting adherence to attendance criteria: 4 wk: HIIT = 98% (43/44), MICT = 91% (39/43), P >.05 3 mo: HIIT = 68% (25/37), MICT = 78% (31/41), P >.05 6 mo: HIIT = 65% (24/37), MICT = 67% (26/39), P >.05 12 mo: HIIT = 56% (19/34), MICT = 70% (26/37), P >.05 Criteria for attendance and intensity: Patients meeting adherence and intensity criteria combined: 4 wk: HIIT = 91% (39/44), MICT = 91% (29/43), P >.05. 3 mo: HIIT = 68% (25/37), MICT = 75% (30/41), P >.05 6 mo: HIIT = 57% (21/37), MICT = 39% (15/39), P >.05 12 mo: HIIT = 53% (18/34), MICT = 41% (15/39), P >.05 Criteria for intensity: Patients meeting adherence to intensity criteria: 4 wk: HIIT = 91% (39/44), MICT = 98% (42/43), P >.05 3 mo: HIIT = 84% (31/37), MICT = 88% (35/41), P >.05 6 mo: HIIT = 70% (26/37), MICT = 54% (21/39), P >.05 12 mo: HIIT = 68% (23/34), MICT=49% (18/41), P >.05 Mean training HR: Supervised training: HIIT = 87 ± 6% HRpeak, MICT = 70 ± 8% HRpeak (P < .001) Home-based training at 3 mo: HIIT = 88 ± 7% HRpeak Mean training RPE: Supervised training: HIIT = 16.0 ± 1.1 and MICT = 12.4 ± 0.6 (P < .001) Home-based training at 3 mo, HIIT = 16.5 ± 1.2, MICT = 12.4 ± 0.5 for MICT (P < .001) Protocol deviation: During home-based training at 6 mo, 24% of HIIT participants started exercising at a lower intensity on their own accord, and 38% of MICT participants were exercising at higher intensities. At 12 mo, 20% of HIIT participants started exercising at a lower intensity on their own accord, and 33% of MICT participants were exercising at higher intensities.	Assessed as V˙o 2peak 4 wk (intention-to-treat) HIIT: 10% (+2.9 mL/kg/min) MICT: 4% (+1.2 mL/kg/min) 4 wk (per protocol) HIIT: 18% (+6.0 mL/kg/min) MICT: 8% (+2.7 mL/kg/min) 12 wk (intention-to-treat) HIIT: 9% (+2.6 mL/kg/min) MICT: 8% (+2.2 mL/kg/min) 26 wk (intention-to-treat) HIIT: 11% (+3.1 mL/kg/min) MICT: 6% (+1.7 mL/kg/min) 52 wk (intention-to-treat) HIIT: 10% (+2.9 mL/kg/min) MICT: 7% (+1.8 mL/kg/min) 52 wk (per protocol) HIIT: 18% (+5.2 mL/kg/min) MICT: 8% (+2.2 mL/kg/min)
Trachsel et al (2019)43, Canada	HIIT: 100% PPO (or 15 RPE) MICT: RPE 12-14.	Adherence to intensity and/or duration: Methods not reported Adherence to intensity and duration: No methods reported	Attendance: Data not reported. Training intensity and/or duration: Data not reported.	Assessed as V˙o 2peak (relative to fat-free mass) 12 wk HIIT: 11% (+3.1 mL/kg/min) Usual care: 0% (+0.1 mL/kg/min)
Tschentscher et al (2016)27, Austria	HIIT: 85-95% HRpeak MICT: 65-85% HRpeak Other (pyramid): 65-95% HRpeak	Adherence to attendance: Number of sessions completed—no criteria specified. Adherence to intensity and/or duration: No methods reported	Attendance: To exercise sessions for all groups was 99.2%. Mean training HR: HIIT = 76 ± 6% HRpeak, MICT = 74 ± 8% HRpeak, and pyramid = 72 ± 9% HRpeak (P >.05) Mean training workload: HIIT = 58 ± 8% PWC, MICT = 60 ± 8% PWC, and pyramid = 56 ± 9% PWC (P >.05)	Assessed as peak exercise capacity (W) 6 wk HIIT: 23% (30W) MICT: 21% (27W) Pyramid: 25% (30W)
Villelabeitia-Jaureguizar et al (2016)34, Spain	HIIT: 50% of PPO from steep ramp test MICT: VT1 + 10%	Adherence to attendance: Number of sessions completed—no criteria specified. Adherence to intensity and/or duration: Patients reported peak RPE (Borg 6-20) during the exercise session. The exercise prescription for HIIT was prescribed on the basis of %PPO reached in the steep ramp test, whereas the training workload is reported as %PPO reached in CPX. Methods for measurement of HR not reported.	Attendance: Completion of exercise sessions: HIIT = 92% and MICT = 88% (P > 0.05). Mean training RPE: Not different between HIIT and MICT during the first month of training (RPE 11-13) and second month (RPE, 14-16). Mean training workload: HIIT = 105 ± 22% PPO (wk 1-4), 135 ± 30% PPO (wk 5-8). Training workload data for MICT were not reported. Mean training V˙o 2peak: MICT = 64 ± 9% V˙o 2peak (wk 1-4), corresponding to VT1, and 70 ± 9% V˙o 2peak, corresponding to VT1 + 10%.	Assessed as V˙o 2peak 8 wk HIIT: 23% (+4.5 mL/kg/min) MICT: 12% (+2.5 mL/kg/min)
Warburton et al (2005)31, Canada	HIIT: 85-95% of HRR MICT: 65% of HRR.	Adherence to attendance: Number of supervised sessions completed—no criteria specified. Adherence to intensity and/or duration: HR monitors used during supervised sessions. Individual workloads adjusted daily to maintain desired HR range.	Attendance: Supervised training attendance data not reported. For home-based training, completion rates were 99 ± 2% for both HIIT and MICT. Training intensity: Data not reported.	Assessed as V˙o 2peak Data not reported, only displayed graphically. Improvement in V˙o 2peak was not statistically different between groups after 16 wk.
Wehmeier et al (2020)32, Germany	HIIT: 85-95% HRmax MICT: 50-75% HRmax	Adherence to attendance: Number of supervised sessions completed—no criteria specified. Adherence to intensity and/or duration: All sessions monitored by ECG.	Attendance: Both HIIT and MICT completed an average of 11.5/12 training sessions and 11.5 hr of cycle ergometry. Mean training RPE (at the end of the session): HIIT = 13.1 and MICT = 12.9	Assessed as V˙o 2peak 3 wk HIIT: 18% (+3.4 mL/kg/min) MICT: 4% (+0.9 mL/kg/min)

Abbreviations: CPX, maximal cardiopulmonary exercise test; HIIT, high-intensity interval training; HR, heart rate; HR_max, percentage of maximum heart rate from maximal exercise test; METs, metabolic equivalents; MICT, moderate-intensity continuous training; %HR_peak, percentage of peak heart rate from maximal exercise test; %HRR, percentage of heart rate reserve; %PPO, percentage of peak power output achieved during a maximal exercise test; PWC, peak work capacity; RPE, rating of perceived exertion (on the 6-20 Borg Scale unless otherwise specified); V˙o_2peak, peak oxygen uptake measured during a maximal exercise test.

^aExercise adherence methodology and results for Observational, Retrospective, and Follow-Up Studies is available in Supplemental Digital Content 2 (https://links.lww.com/JCRP/A218).

^bHR_peak and HR_max are often used interchangeably and have been reported within the table as consistent with how the metric was reported by the study.

REPORTING OF ADHERENCE

A criterion for adherence was defined in 13 of 36 (36%) studies^11–23 (Figure 1). For the majority, this was reported to be based on the total number of sessions attended or completed^12–23 and ranged from 66-100% of sessions completed (Table 1). One study¹¹ reported accounting for training intensity and duration in addition to attendance. The criterion for adherence was used to exclude participants from the primary analysis in 6 of 13 studies^16–21 and applied for a secondary treatment (per-protocol) analysis in 5 of 13 studies.^{11,13–15,22} One study²³ applied the criterion solely to determine adherence, and for one study,¹² the application of the criterion was not defined.

Figure 1.: Graphical representation of studies reporting on aspects of exercise training adherence. Abbreviation: RPE, rating of perceived exertion.

For the randomized, observational, and retrospective studies, attendance was mainly reported as the number/proportion of sessions completed,^{20–22,24–34} or the number/proportion of participants who completed 70-100% of sessions^{11,13–19,27,35–41} (Table 1, Figure 1). One study reported attendance using both methods,²³ and three studies did not report on attendance to the exercise sessions.^12,42,43 For the follow-up studies,^19,20,44,45 adherence to the protocol was not reported but rather the effect of group allocation on exercise adherence. This was reported as the frequency or minutes of exercise/wk, either as self-reported sessions/wk^19,46 or objectively measured time spent at moderate or vigorous intensities.^44,45

Reporting of exercise intensity during training was highly variable among studies (Table 1, Figure 1). For randomized, observational, and retrospective studies, 12 of 32 studies reported training HR.^{11,14,18,20,22,24–27,33,39,40} One study reported only training HR for HIIT (but not MICT),³⁹ one study reported only absolute HR values (rather than %HR_peak),¹⁸ and one study reported relative HR in relation to age-predicted HR_max (rather than HR_peak from a maximal exercise test).²⁶ Six studies reported that training HR was measured during the high-intensity intervals,^{11,14,22,24,33,40} while six studies did not specify.^{18,20,25–27,39} One study also reported %HR_max during the recovery intervals.³³ There were 11 of 32 studies that reported intensity as training RPE, either in relation to the effort during the high-intensity intervals^{11,20,23,25,28,33,34} or at the end of the session.^21,22,29,32 All studies that reported RPE used the 6-20 Borg scale, except for one study²⁰ that used the 0-10 Borg scale. Two studies also reported RPE during the recovery intervals.^23,28 Three studies reported intensity as a percentage of peak power output³⁴ or peak work capacity.^22,27 Two studies with longer interventions included self-report exercise intensity.^30,41 Follow-up studies also used self-report exercise intensity^19,46 or accelerometry-based intensity data.^44,45 Ten studies did not report any training intensity data.^{12,13,15,17,31,35–38,43}

Few studies reported on whether participants achieved the prescribed duration of exercise sessions (Table 1 and Figure 1). Three studies reported the mean duration of sessions^18,26,41 and one study included duration in the criteria for adherence.¹¹ Overall, it was generally unclear whether a session deemed “completed” involved assessment of session duration in addition to session attendance.

ADHERENCE TO ATTENDANCE

The majority of studies evaluated adherence to HIIT in comparison to a form of MICT.^{11,15–18,20–22,24,26–28,31,32,34,35,38–40,42,45,46} For supervised programs, 10 studies^{20–22,24–28,34,35} compared attendance to sessions between HIIT and MICT, with all studies reporting similar attendance to sessions between HIIT (mean = 92%, range: 82-100%) and MICT (mean = 90%, range: 80-100%) (mean difference = 3.6%, Figure 2A). Eight studies^{11,15–17,35,38–40} compared the proportion of patients attending all sessions or meeting a prespecified criteria for supervised programs, with all studies reporting no differences between HIIT (mean = 93%, range: 68-100%) and MICT (96%, range: 79-100%) (mean difference = 3.9%, Figure 2B). Using linear regression, we found no influence of sex (as % of females recruited) on (1) proportion of HIIT sessions attended; (2) proportion of HIIT participants meeting study adherence criteria; or (3) difference between HIIT and MICT for adherence to attendance.

Figure 2.: (A) Session attendance. Proportion of HIIT or MICT sessions attended and the difference in proportions between the two exercise groups. Solid squares represent differences in adherence between HIIT and MICT. The solid diamond represents the mean difference ± SD. Image created using Comprehensive Meta-Analysis Software (V3.3.070, Biostat). (B) Participant attendance. Proportion of participants adherent to all sessions, or a priori adherence criteria within HIIT or MICT, and the difference in proportions between the two exercise groups. Solid squares represent differences in adherence between HIIT and MICT. The solid diamond represents the mean difference ± SD. Abbreviations: HIIT, high-intensity interval training; MICT, moderate-intensity continuous training. Image created using Comprehensive Meta-Analysis Software V3.3.070 (Biostat).

ADHERENCE TO INTENSITY

Of the studies that reported training intensity as %HR_peak for HIIT during supervised programs (n = 10), seven studies reported average training intensities >85%HR_peak (range: 87-92%),^{11,14,20,22,24,25,39} one study reported average training intensity of 84%HR_max,³³ and two studies reported average intensity <80%HR_peak.^26,27 Currie et al²⁶ reported mean training intensities of 73%HR_peak and 68%HR_peak for supervised and home-based training, respectively, and Tschentscher et al²⁷ reported a mean training intensity of 76%HR_peak for supervised training. Neither study reported methods for measuring training intensity, and, therefore, training HR may not be reflective of the time spent at high intensity. Furthermore Currie at al²⁶ reported training HR relative to age-predicted maximum, rather than HR_max achieved during a maximal exercise test.

In comparison to the training HR reported for HIIT, five studies reported mean training HR for MICT <75%HR_peak (range: 65-74%HR_peak).^{11,20,25–27} Two studies^22,24 reported training intensities for MICT of 80% and 79% HR_peak, respectively; however, neither study restrained participants to exercise at a moderate intensity.

Of the studies that reported training as RPE (n = 10) during supervised programs, four studies reported average training RPE ≥15/20 (range: 15.0-16.7)^11,25,28 or equivalent to a hard RPE on the 0-10 Borg scale (average training RPE of 6.1),²⁰ and six studies^{21–23,29,32,33} reported average RPE <15 (range: 13.1-14.4). However, four of the six studies reporting an average RPE <15 or hard,^21,22,29,32 measured RPE at the end of the session, which may be reflective of the effort for the session overall rather than the peak effort during the high-intensity intervals. Of the remaining studies,³³ prescribed HIIT at a lower RPE range of 14-16 (compared with RPE 15-17) prescribed in other trials.^11,14 The other study reported that 58% of HIIT participants trained below the target RPE of 15 during the high-intensity intervals.²³ In comparison to HIIT, seven studies reported mean training RPE for MICT <RPE 14/20 (range: 11.2-13.5)^{11,21,22,25,28,32} or equivalent to a somewhat hard RPE on the 0-10 Borg scale (average training RPE of 2.9).²⁰ Because of heterogeneity of study reporting for adherence to intensity, and lack of reporting for adherence to duration, we were unable to investigate the influence of sex on these aspects of adherence.

LONG-TERM ADHERENCE

Only two studies have assessed long-term adherence to HIIT compared with MICT following supervised CR.^11,20 Moholdt et al²⁰ found that after 5 mo of home-based training, 52% of HIIT participants continued HIIT ≥3 times/wk and 64% of MICT participants continued MICT ≥3 times/wk. The proportion of HIIT participants that stopped HIIT in favor of moderate-intensity training was 35%, compared to 4% MICT participants who started higher-intensity exercise. In contrast, The FITR-Heart Study,¹¹ found that a higher proportion of MICT participants (38%) started high-intensity exercise after 5-mo of home-based training, and a lower proportion of HIIT participants who swapped to a moderate-intensity exercise (24%). However, the percentage of HIIT participants who continued HIIT ≥3 times/wk in The FITR-Heart Study at 6 mo (57%) and 12 mo (53%)¹¹ was similar to Moholdt et al.²⁰ Furthermore, total exercise adherence (reported as the average number of weekly sessions for any exercise) was similar to MICT at 6 mo (3.5 ± 1.5 vs 3.7 ± 1.6) and 12 mo (3.1 ± 1.8 vs 3.5 ± 2.1).¹¹

ADHERENCE TO HOME-BASED HIIT

Only two studies have investigated home-based HIIT compared with center-based CR. Moholdt et al³⁰ compared a 6-mo home-based HIIT program with a 4-wk residential CR program. While Aamot and colleagues¹⁴ compared a 12-wk home-based HIIT with 12-wk center-based HIIT programs and then completed a follow-up at 12 mo,⁴⁴ Moholdt et al³⁰ provided verbal instructions but no supervised HIIT sessions prior to home-based HIIT. The authors found a similar improvement in V˙o_2peak with the 6-mo home-based HIIT program compared with an intensive 4-wk CR program (17 vs 19%, respectively), with 42% of participants maintaining three HIIT sessions/wk at 6 mo and 17% doing a combination of HIIT and MICT. The home-based HIIT group reported an average of 1.6 ± 1.6 HIIT sessions/wk and 2.4 ± 1.9 moderate-intensity sessions/wk. Aamot et al¹⁴ provided participants with two supervised instructional HIIT sessions prior to home-based training. There were no significant differences in V·o_2peak improvement between the 12-wk home-based HIIT program (8%) and 12-wk hospital-based HIIT programs (10-12%). Regarding adherence, the mean number of HIIT sessions for the home-based program was 24 (10-24), similar to the hospital-based program of 24 (17-24). However, 13% of participants in home-based HIIT did not meet the 70% adherence criteria, which was significantly higher than the hospital-based programs (0%). Few studies have reported training intensity data during home-based HIIT. Two studies^11,14 have reported an average home-based HIIT training intensity of 89% HR_peak, compared with center-based HIIT training intensities of 87% and 90% HR_peak, respectively. One study¹¹ reported an average home-based HIIT RPE of 16.3, which was the same as center-based HIIT.

INFLUENCE OF ADHERENCE ON CARDIORESPIRATORY FITNESS

Fourteen studies comparing HIIT with MICT included assessment of V˙o_2peak directly following the supervised exercise intervention.^{11,16,17,20–22,24–26,28,32,34,38,40} The average improvement in V˙o_2peak for HIIT was 17% (range: 6-24%) compared with 8% (range: 0-20%) for MICT. Ten studies reported that the improvement in V˙o_2peak with HIIT was significantly greater compared with MICT.^{11,16,17,21,24,28,32,34,38,40} One study²⁵ reported that the improvement in V˙o_2peak was greater for HIIT only when accounting for adherence using per-protocol analysis. Three studies^20,22,26 reported no significant differences for improvement in V˙o_2peak between HIIT and MICT, even when accounting for per-protocol analysis based on adherence to attendance. However, in these studies, the training intensities of participants likely contributed to the similarities in V˙o_2peak improvement for HIIT and MICT. In the study by Currie et al,²⁶ where HIIT was prescribed as 89-100% peak power output, the average training intensity reported for HIIT was 73 ± 10%HR_max, which is lower than the criteria usually prescribed for HIIT.¹ In the SAINTEX-CAD study,²² despite the HIIT group achieving an appropriate mean training intensity of 88%HR_peak, the authors acknowledged that MICT participants trained at higher intensities (mean training intensity = 80%HR_peak) than prescribed for MICT (65-75%HR_peak). Finally, in the study by Moholdt et al,²⁰ mean training HR during the randomized training sessions for both HIIT (92%HR_peak) and MICT (74%HR_peak) was within prescribed targets; however, both HIIT and MICT participated in 3-5 additional training sessions/wk for which 31% were of a high intensity.

Three studies have investigated the long-term effects of HIIT compared with MICT on CRF and exercise adherence at 9-mo,⁴⁵ 11-mo,¹¹ and 27-mo⁴⁶ following supervised CR. Moholdt et al⁴⁶ found less deterioration in V˙o_2peak for HIIT participants compared with MICT after a 27-mo follow-up period. This is likely a result of the HIIT group maintaining a greater frequency of total exercise than MICT, with 82% of the group HIIT engaging in any exercise ≥2 times/wk compared with 52% in the MICT group. The SAINTEX-CAD study⁴⁵ found that both HIIT and MICT maintained V˙o_2peak equally after a 9-mo follow-up period. The authors did not report on adherence to the randomized training, instead reporting exercise adherence as meeting physical activity guidelines and time spent in moderate and vigorous activities, which was not different between HIIT and MICT.⁴⁵ The FITR Heart Study¹¹ also found that both HIIT and MICT had similar improvements in V˙o_2peak at 12 mo for intention-to-treat analyses. However, when adherence to the recommended frequency (>3 times/wk) and intensity of HIIT and MICT were accounted for in the prespecified per-protocol analysis, the improvement in V˙o_2peak was superior for HIIT compared with MICT with a clinically relevant mean difference of 3.0 mL/kg/min.¹¹

FEASIBILITY OF HIIT

Eleven studies commented on the feasibility of HIIT compared with MICT. Feasibility referred to various aspects including adherence to attendance,^11,23,26,33 adherence to intensity,^22,23,25 tolerability or enjoyment of the protocol,^{11,23,25,28,32,33,39} safety,^23,33 and implementation of the protocol into the CR program.^11,28 In some studies, the definition of feasibility was not specified.^15,27 Eight studies determined HIIT to be feasible^{11,15,23,26–28,32,33} and three studies reported that HIIT was less feasible than MICT.^22,25,39 Two of these studies reported that HIIT was not feasible in relation to maintaining ≥ 90%HR_peak for the entire 4-min duration.^22,25 Two studies also reported that HIIT was less feasible based on patient perspectives.^25,39 Heber et al³⁹ reported that HIIT was not feasible in five patients (13%), with patients reporting knee pain, discomfort, and lack of motivation, and that it was too demanding. All patients instead completed MICT for the remainder of the intervention. Lee et al²⁵ studied HIIT compared with MICT in a female-only cohort. Although some women expressed a sense of enjoyment and accomplishment associated with performing HIIT, the majority of women described HIIT to be a daunting undertaking and expressed preference for performing MICT over HIIT.²⁵ Specifically, women reported physical discomfort related to biomechanical limitations to jogging or running, back or hip pain, shortness of breath, and hyperventilation.⁴⁷ In contrast, Way et al²³ found that all male and female participants expressed satisfaction with HIIT participation, and Taylor et al¹¹ found similar rates of enjoyment between HIIT and MICT. Furthermore, Taylor et al¹¹ found that the frequency and reasons stated for being unable to complete the exercise protocol were similar between HIIT and MICT, as well as unpleasant symptoms and injuries.¹¹ Both studies^11,23 employed a variety of modes for HIIT, including cycling, treadmill, elliptical, and dance/movement-based routines.

Supplemental Digital Content 3, available at: https://links.lww.com/JCRP/A219, presents qualitative data from The FITR-Heart Study,⁴⁸ relating to why participants were or were not continuing with HIIT following supervised training and home-based training (male: 28; female: 4). For comparison, Supplemental Digital Content 4, available at: https://links.lww.com/JCRP/A220, presents the same qualitative data for MICT participants (male: 25; female: 6). Twelve participants for HIIT (male: 11, female: 1) and 13 participants for MICT (male: 13; female: 0) did not provide any qualitative comments. One of 43 HIIT participants at 4 wk and 6 of 34 (18%) HIIT participants at 12 mo (all male) reported not continuing with HIIT. Four of the six participants discontinued HIIT to exercise at a lighter intensity. For MICT, one of 44 participants at 4 wk and 10 of 37 (27%) participants at 12 mo (male: 9; female: 1) reported not continuing with MICT. Six of the 10 (all male) instead chose to exercise at a higher intensity. However, many of the comments from patients participating in HIIT and MICT reported feelings of motivation, confidence, enjoyment, and beneficial effects for their physical and mental well-being. Way et al²³ found similar common themes within their qualitative analysis of HIIT.

DISCUSSION

A review of the current literature found that adherence to the attendance of HIIT sessions in CR patients with coronary artery disease is high and comparable with MICT. In contrast, adherence to the intensity and duration of HIIT is underreported. Only 36% of studies defined an adherence criterion for the exercise interventions, and for the majority of studies, this was reported to be based on sessions attended or completed. Whether adherence to intensity and duration was assumed or incorporated into determining whether a session was attended or completed was not reported. Based on the available literature, adherence to HIIT (encompassing all necessary components of attendance, intensity, and duration) could not be determined. This review illustrates that adherence to intensity influences the improvement in CRF which can have important implications for determining the treatment effect of exercise interventions being compared. Previous studies have observed that participants will tend to lower the intensity from high to moderate as time goes on.⁴⁹ Therefore, including the reporting of adherence to intensity and duration is important for determining whether the correct dose of exercise was delivered.

With important questions still to be answered relating to HIIT in CR, including the optimal protocol, reporting of adherence to HIIT protocols and its comparators is vital for determining effectiveness. For example, for studies designed to determine whether HIIT is superior to MICT for improving V·o_2peak and other outcomes, reporting and justifying whether the comparison is truly moderate-intensity exercise or a continuous training of a moderate to high intensity. Furthermore, adherence to the interventions should be reported and can be considered in sensitivity or explanatory analyses (such as treatment or per-protocol analysis, or used as a covariate) to illustrate the treatment effect in addition to the intention-to-treat effect.

Conducting exercise training trials is challenging, especially when sample sizes are calculated according to a presumed level of adherence to the intervention. Treatment or per-protocol analysis, which excludes nonadherent participants from the analysis, is a form of explanatory analysis that describes the treatment effect or efficacy of an intervention for those who adhered to it. While intention-to-treat analysis should be the primary analysis for randomized controlled trials, given it limits bias in estimating the efficacy of an intervention,⁵⁰ it can underestimate the magnitude of the treatment effect if there is substantial nonadherence to one or all of the interventions being compared.⁵¹ Calculation of required sample sizes should be performed accordingly. With appropriate measuring, reporting, and application of adherence data, treatment analysis is a valuable secondary analysis to explore the treatment effect of an intervention.

Long-term follow-up studies consistently report the influence of HIIT on overall exercise adherence, rather than whether participants want to, are able to, and/or actually do continue HIIT long-term. Therefore, it remains unclear whether adhering to HIIT offers superior benefit for CRF, and whether HIIT can be performed effectively and safely by patients with coronary artery disease in home-based settings. This is even more important, given the recent closures of center-based CR facilities during a global pandemic. Therefore, we recommend that follow-up studies also report ongoing adherence to HIIT, in addition to the effect of the interventions on overall exercise adherence.

This review was limited to patients with coronary artery disease. Therefore, our findings cannot be generalized to other patients attending CR, including those with heart failure, implanted devices, valve surgery, congenital heart disease, and spontaneous coronary artery dissection. Furthermore, to facilitate a timely review of the literature, we used a rapid review process, which is less rigorous than a systematic review. Study selection and data extraction were conducted by a single author, and quality of the included studies was not assessed. This process may increase the risk of bias, data extraction errors, and study omission, compared with a comprehensive systematic review process.

APPLICATION TO PRACTICE

Recommendations for Clinicians

As with any exercise prescription, HIIT should be prescribed with consideration of patient goals and preferences. A recent statement from the American Association of Cardiovascular and Pulmonary Rehabilitation has advocated the importance of progressing the intensity of exercise for patients in CR programs including considering HIIT.⁵² It provides an additional exercise option for patients to be used in conjunction with other forms of exercise (such as MICT) to achieve the recommended exercise volumes for health and/or weight loss.^53,54 To improve adherence to and enjoyment of HIIT, realistic expectations regarding target intensities should be set. For example, the 4 × 4 min HIIT protocol commonly prescribed on the basis of the Norwegian model⁵⁴ (ie, four bouts of 4-min high-intensity intervals, interspersed with 3-min active recovery intervals) was not designed to elicit a training intensity of 85-95%HR_peak for the entire 4-min work interval. Instead, the goal is to reach the target intensity (85-95%HR_peak) within 2-min and maintain target intensity for the final 2-min of each interval.^55,56 The RPE scale can also be used when prescribing HIIT due to the potential limitations for using HR in isolation.⁵⁶ Alternatively, RPE can be used in isolation for prescribing HIIT if accurate HR_peak data are not available or patients are taking HR-modulating therapy.⁵⁶ To increase enjoyment and reduce the impact/discomfort on musculoskeletal joints, a variety of modalities should be encouraged. In addition to walking or jogging, HIIT can be applied to cycling, rowing ergometer, elliptical, aerobics/dance classes, or swimming/water aerobics. Furthermore, progressive HIIT protocols may be better tolerated,⁴² allowing patients time to adapt and manage longer intervals of high-intensity exercise over a period of time. Music can also provide further enjoyment. The Norwegian Ullevaal model¹⁹ provides an example of how the traditional Norwegian 4 × 4 min HIIT protocol can be applied to aerobic exercise classes. Intensity is adjusted using individual %HR_peak, RPE, and the speed (beats/min) of the music, with faster-paced songs used for the aerobic-style exercises during the high-intensity intervals and slower-pace songs used for resistance and flexibility exercises during the recovery intervals. While studies in patients with coronary artery disease have consistently demonstrated a favorable safety profile,^{11,22,24,28,57} it should be acknowledged that the majority of studies to date have employed maximal cardiopulmonary exercise testing (CPX) to medically evaluate patients prior to HIIT participation. To maximize safety in patients with cardiometabolic disease, guidelines have been developed to assist clinicians with appropriate screening and monitoring for HIIT.⁵⁶ Initially, there may be additional time commitment regarding education and monitoring of HIIT. It may also be challenging for patients with cognitive impairment or multiple comorbidities to manage HIIT independently.

Recommendations for Researchers

Table 2 and Supplemental Digital Content 5, available at: https://links.lww.com/JCRP/A221, outline a set of recommendations for measuring and reporting adherence to HIIT interventions. In this framework, adherence to the intervention encompasses components from the American College of Sports Medicine (ACSM) exercise prescription principles (type, frequency, intensity, and duration).⁵³ First, we outline 10 recommendations to provide a consistent approach for measuring and reporting adherence to HIIT interventions. Second, for each recommendation, detailed descriptions are provided with examples to report methods and calculate the variables. A third component is an example table for how the data may be presented. Finally, additional outcomes are suggested to inform feasibility or tolerability of the intervention. These recommendations were formulated on the basis of the ACSM exercise prescription principles, the literature examined as part of this review, the challenges identified when comparing studies, practical experience of the authors, and methods that will allow for comparison of studies in relation to adherence data in the future.

Table 2 - Recommendations for Measuring and Reporting Adherence to HIIT and Other Exercise Interventions^a

Recommendations

Define criteria for global adherence in the methods to assess whether adherence to the intervention was achieved, and whether this was accounted for in analysis. Describe in the methods how adherence was measured and recorded for allocated training sessions. Describe in the methods how nonprotocol exercise/physical activity was measured and reported. Report the adherence to attendance of prescribed sessions for all exercise groups (as % of sessions attended compared with sessions prescribed, and the number and % of participants who meet the attendance criteria for adherence). Report the mean training intensity data for all exercise groups (subjective and objective). Report the adherence to intensity of the attended sessions for all exercise groups (as % of sessions meeting the intensity criteria compared with sessions attended). Report the mean training duration data of sessions attended for all exercise groups. Report the adherence to duration of the attended sessions for all exercise groups (as % of sessions meeting the duration criteria compared with sessions attended). Report nonprotocol exercise/physical activity. Report the global adherence of prescribed sessions for all exercise groups (as the % of sessions adherent compared with sessions prescribed, and the number and % of participants who meet the global criteria for adherence). If per-protocol analysis was performed, report whether study results were different for those who adhered to the intervention (treatment effect).

Detailed description of recommendations and examples

1. Define criteria for global adherence in the methods to assess whether adherence to the intervention was achieved, and whether this was accounted for in the analysis.

Describe criteria for global adherence in the methods to assess whether adherence to the intervention was achieved. Global criteria for adherence should encompass components of attendance, intensity, and duration. Consider alternatives to %HRpeak as intensity criteria for short intervals (eg, % peak work capacity) or with patients on rate-limiting therapy (eg, RPE). Example: Global adherence to the intervention was assessed as ≥70% attendance at sessions where training was at the prescribed intensity and ≥75% of the prescribed duration. For the HIIT group, adherence to intensity was assessed as an average training intensity ≥85% HRpeak and/or RPE ≥ 15, and the MICT group was assessed as an average training intensity between 65% and 75% HRpeak and/or RPE between 11 and 13. Adherence to duration was assessed as completing ≥75% of session time (23 min out of 30 min) and for the HIIT group ≥3 of the 4 × 4 min high-intensity intervals. Describe whether intervention adherence was accounted for in the analysis. Example: An explanatory per-protocol analysis was performed, including only participants meeting the global criteria for adherence to the intervention.

2. Describe in the methods how adherence was measured and recorded for allocated training sessions.

Describe how intensity was measured for the allocated training. Parameters of training intensity should include subjective measures of intensity (RPE) and objective measures of intensity (training HR or workload). Training workload may be measured using treadmill speed and incline, or watts, and/or metabolic equivalents (METs). To convert treadmill or cycle workloads to METs, the use of prediction equations from the FRIEND registry58,59 or American College of Sports Medicine (ACSM)54 is recommended. For high-intensity intervals ≥ 2 min, we recommend measuring HR in the final minute or at the end of the high-intensity interval. For high-intensity intervals of short duration (<2 min), we recommend using training workload as the objective measure, as HR may underestimate the training stimulus if there is not sufficient time to increase. Details that should be included: How often and when was intensity measured? Was intensity measured continuously (eg, HR) and/or at certain time points during the session (eg, RPE, workload). Was intensity measured and recorded during each high-intensity interval? By whom was the intensity measured and recorded? (eg, did study staff measure and record or did patient self-report and self-record?). Example 1: Training HR was measured by pulse oximetry and RPE was measured using the 6-20 Borg scale. For HIIT, HR and RPE were measured during the last minute of each high-intensity interval. For MICT, HR and RPE were measured halfway through and at the end of training prior to cooldown. For each session, both HR and RPE were recorded as peak intensity using the highest HR and RPE recorded. Average intensity was determined from all HR and RPE data recorded during the high-intensity intervals. Example 2: Training HR was measured continuously by an HR monitor, and RPE was measured using the 6-20 Borg scale. For HIIT, HR was averaged over the final 2 min of each high-intensity interval and RPE was measured at the end of each high-intensity interval. For MICT, HR was averaged for the entire exercise time (excluding warm-up and cooldown) and RPE was measured at the end of the session prior to cooldown. Example 3: Training workload was measured by treadmill speed and incline or cycling watts, and RPE was measured using the 6-20 Borg scale. For HIIT, training workload and RPE were measured during the final minute of each high-intensity interval and for MICT, workload and RPE were measured at the end of each exercise bout prior to cooldown. Describe how intensity was recorded for the allocated training. Example 1: For each session, both HR and RPE were recorded as peak intensity using the highest HR and RPE recorded, and average intensity using an average of all HR and RPE data recorded. Example 2: For each session, both workload and RPE were recorded as peak intensity using the highest workload and RPE recorded, and average intensity using an average of all workload and RPE data recorded. Workload from each exercise machine was converted to metabolic equivalents using standardized equations to allow for a standardized metric to be recorded and averaged. Describe how duration was measured and recorded for allocated training. Parameters of training duration may include total session duration, minutes spent in high-intensity intervals, or number of intervals. Example 1: For both exercise groups, session duration was measured and recorded as the total session duration. For HIIT, the number of high-intensity intervals was also measured and recorded. Example 2: For both HIIT groups, session duration was measured and recorded as the time spent in the high-intensity intervals by measuring the number of high-intensity intervals completed and the duration of each high-intensity interval (eg, if patient managed 3 × 4 min high-intensity intervals but only 2 min for the last high-intensity interval, time spent in high-intensity intervals = 14 min.

3. Describe in the methods how nonprotocol exercise/physical activity was measured and recorded.

Describe how nonprotocol exercise or physical activity was measured and recorded. Nonprotocol exercise can be measured as the additional exercise sessions and exercise minutes, and whether exercise intensity is different to allocated training. Can use self-report exercise diaries and/or fitness tracking devices with automatic upload to capture frequency, intensity, and duration of additional exercise. Overall physical activity can be measured by accelerometry or self-report physical activity questionnaires (eg, International Physical Activity Questionnaire). Example: Participants were provided with an exercise diary to record the type, frequency, and duration of their exercise each week, as well as a rating of intensity using the 6-20 Borg scale. These exercise diaries were checked by study staff for nonprotocol exercise, which was recorded as the number of additional exercise sessions and exercise minutes, and whether the intensity was different to the allocated training. Accelerometers were also used to measure physical activity for a 7-d period at each time point.

4. Report the adherence to attendance of prescribed sessions for all exercise groups as the % of sessions attended compared with sessions prescribed, and the number and % of participants who meet the attendance criteria for adherence.

For each participant, determine the adherence to attendance as the % of sessions attended compared with the number prescribed. Example: If a participant attends 10 sessions compared with 12 prescribed, adherence to attendance for that participant will be 10/12 = 83%. For each exercise group, average the adherence to attendance from each participant to calculate and report the group adherence to attendance (as % of sessions). For each exercise group, determine and report the number and % of participants who meet the attendance criteria for adherence. Example: Based on the example criteria (≥70% attendance at sessions), any participant with an adherence to attendance ≥70% would meet the criteria and any participant with adherence to attendance <70% would not meet the criteria. The example participant above with an adherence to attendance of 83% would meet the adherence criteria.

5. Report the mean training intensity data for all exercise groups (subjective and objective).

For each participant, calculate the mean training intensity across all training sessions. For HR or workload intensities, calculate the relative %HRpeak or % peak work capacity, based on the peak values from the maximal exercise test. For %HRpeak, use the maximal exercise test with the highest HRpeak. For % peak work capacity, use the peak workload from the initial maximal exercise test. Example 1: If a participant attends three sessions with a training RPE of 17.0, 16.0, and 16.0, the average training RPE will be the average of the three sessions = RPE 16.3. Example 2: If a participant attends three sessions with a training HR of 145 bpm, 148 bpm, and 150 bpm, the average training HR will be the average of the three sessions = 148 bpm. If the participant's HRpeak was 155 bpm during the baseline test and 160 bpm during the follow-up test, the average training %HRpeak should be calculated as 148/160 = 93% HRpeak based on the follow-up test (highest HRpeak). Example 3: If a participant attends three sessions with a training workload of 7.1 METs, 7.3 METs, and 7.8 METS, the average training workload will be the average of the three sessions = 7.4 METs. If the participant's peak workload was 7.1 METs in the baseline test and 8.0 METs during the follow-up test, the average training intensity should be calculated as 7.4/7.1 = 104% peak work capacity based on the baseline test. For each exercise group, calculate and report the mean training intensity from the average training intensities from each participant.

6. Report the adherence to intensity of the attended sessions for all exercise groups (as the % of sessions meeting the intensity criteria compared with sessions attended).

For each participant, determine the adherence to intensity as the % of sessions meeting the intensity criteria compared with the number of sessions attended. This may be determined separately for supervised sessions, home-based sessions, or as a combination. Example: If a participant meets the intensity criteria of ≥85% HRpeak for eight sessions compared with 10 sessions attended, the adherence to intensity for that participant will be 8/10 = 80%. For each exercise group, average the adherence to intensity for each participant to calculate and report the group adherence to intensity (as % of sessions).

7. Report the mean training duration data for all exercise groups.

For each participant, calculate the mean training duration. Example 1: If a participant attends three sessions with training durations of 30 min, 30 min, and 28 min, the average training duration will be the average of the three sessions = 29 min. Example 2: If a participant attends three sessions with time spent in high-intensity intervals of 16 min, 12 min, and 16 min, the average time spent in high-intensity intervals will be the average for the three sessions = 15 min. Example 3: If a participant attends three sessions and completes four intervals, three intervals, and four intervals, respectively, the average number of intervals will be the average for the three sessions = 3.7 intervals. For each exercise group, calculate and report the mean training duration from the average training intensities for each participant.

8. Report the adherence to duration of the attended sessions for all exercise groups (as the % of sessions meeting the duration criteria compared with sessions attended).

For each participant, determine the adherence to duration as the % sessions meeting the duration criteria compared with the number of sessions attended. Example 1: If a participant completes at least three out of four high-intensity intervals for nine sessions compared with the 10 sessions he or she attended, the adherence to duration for that participant will be 9/10 = 90%. Example 2: If a participant completes a session time of ≥ 27 min out of 30 min for nine sessions compared with the 10 sessions he or she attended, the adherence to duration for that participant will be 9/10 = 90%. For each exercise group, average the adherence to duration for each participant to calculate and report the group adherence to duration (as % of sessions).

9. Report non-protocol exercise/physical activity.

For each participant, calculate the number of additional exercise sessions and exercise minutes and the number of additional sessions reported to be at a different intensity to the allocated training. For each exercise group, calculate and report the mean additional exercise sessions and exercise minutes from each participant. For each exercise group, determine the number of participants who trained at a different exercise intensity to the allocated training.

10. Report the global adherence of prescribed sessions for all exercise groups (as the % of sessions adherent compared with sessions prescribed, and the number and % of participants who meet the global criteria for adherence). If per-protocol analysis was performed, report whether study results were different for those who adhered to the intervention (treatment effect).

For each participant, determine global adherence as the % of sessions meeting both intensity and duration criteria compared with sessions prescribed. Example: Using the example outlined in Recommendation no. 1 as the global criteria for adherence (≥70% attendance at sessions where training is at the prescribed intensity and ≥75% of the prescribed duration). If a participant attends 10 sessions out of the prescribed 12 sessions, however, he or she meets only the prescribed intensity for eight sessions (Recommendation no. 6 example), and meets the duration criteria for nine sessions (Recommendation no. 8 example), the number of sessions meeting both criteria is eight sessions, and therefore global adherence for that participant will be 8/12 = 66%. For each exercise group, average the global adherence from each participant to calculate and report the group global adherence (as % of sessions). For each exercise group, determine and report the number and % of participants who meet the global criteria for adherence. Example: Based on the example criteria, any participant with a global adherence ≥70% would meet the criteria and any participant with a global adherence <70% would not meet the criteria. The aforementioned example participant with a global adherence of 66% would not meet the global criteria for adherence. For secondary per-protocol analysis, exclude participants who do not meet the global criteria for adherence and rerun the analyses. Report whether the study results were different for per-protocol analyses compared with intention-to-treat analyses.

Additional data collection to inform feasibility (or tolerability) of the intervention

Number of participants who dropped out of the study (and reasons why) Number of participants who discontinued the intervention but remained in the study Number of participants with an interruption to the intervention (eg, missed ≥3 consecutive sessions or ≥1 wk of training), and whether sessions were rescheduled and completed accordingly. Reason for nonattendance to training session and/or interruption or discontinuation of training intervention Advised by study staff or clinicians due to medical or safety reason Patient decision (eg, due to lack of time, lack of motivation, symptoms, injury, fatigue, or recent medical history) Reason for nonadherence to the prescribed intensity: Clinician reduced intensity (eg, due to signs, symptoms, injury, fatigue, or recent medical history) Patient unable to exercise at prescribed intensity (eg, due to symptoms, fatigue, injury, physiological limitation) Patient unwilling to exercise at prescribed intensity (eg, due to lack of motivation, fear, chose to work at a lower or higher intensity) Reason for nonadherence to the prescribed duration Clinician stopped session early (eg, due to patient signs, symptoms, injury, or fatigue) Patient stopped exercising early (eg,. due to lack of time, lack of motivation, symptoms, injury, or fatigue) Training intensity parameters (RPE, HR, or workload) at the end of the recovery intervals

Abbreviations: HIIT, high-intensity interval training; HR, heart rate; HR_peak, peak heart rate (obtained from maximal exercise test); METs, metabolic equivalents; MICT, moderate-intensity continuous training; RPE, rating of perceived exertion (based on 6-20 or 1-10 Borg scale).

^aSee Supplemental Digital Content 5, available at: https://links.lww.com/JCRP/A221, for Example Results Reporting template.

For measurement and reporting of intensity, we recommend both subjective (eg, RPE) and objective measures (eg, %HR_peak and/or % peak work capacity). Objective measures of intensity should be reported relative to peak values to allow for comparison between interventions and other studies. For shorter high-intensity intervals (eg, <2 min), using % peak work capacity is recommended as %HR_peak may underestimate the training stimulus when there is insufficient time for HR to rise. For %HR_peak, an accurate HR_peak obtained during a maximal exercise test should be used (while taking any prescribed HR-modulating therapy), rather than age-predicted equations that can show significant variation in populations with⁵⁸ and without rate-limiting medical therapy (ie, β-blockade).⁵⁹ To avoid overestimating training intensity, we recommend using the HR_peak from the CPX with highest HR, unless there has been an adjustment in HR-modulating therapy during the intervention.⁵⁵ Using HR_peak from the baseline CPX instead of the CPX with the highest HR can overestimate training intensity by 4-5%HR_peak.¹¹ We recommend using subjective measures in addition to objective measures as there may be limitations in solely using %HR_peak.⁵⁶ These include an inaccurate HR_peak if a maximal effort is not achieved during the maximal test (eg, due to peripheral fatigue); if HR-modulating therapy is up-titrated throughout the intervention period; and when exercise testing and training occur at a different time of day, the HR-modulating effect of therapy may be different.⁶⁰ As an alternative to reporting %HR_peak, relative intensity can be reported as the training workload relative to peak work capacity during the initial maximal exercise test. Although this may be difficult when a variety of exercise modalities are utilized, converting workload to metabolic equivalents using the FRIEND (Fitness Registry and the Importance of Exercise: a National Database)^61,62 or ACSM⁵³ prediction equations can provide a standardized workload metric. To determine whether a significant difference in training intensity exists between training interventions, we recommend that HIIT intensity is recorded toward the end of the high-intensity interval. Additional measurement of intensity parameters at the end of the recovery intervals may help explain adherence to intensity during the high-intensity intervals. If physiological recovery is inadequate, this can lead to premature fatigue whereby patients may be unable to tolerate or provide sufficient exertion for subsequent intervals.⁶² Finally, measuring and reporting the duration of the intervention (in terms of total session duration, number of intervals, and time spent at high intensity) in addition to intensity allows for the determination of exercise volume, which may help identify the existence of dose-response effects.

As a starting point for the design and reporting of exercise studies, including HIIT interventions, we recommend using the CERT⁹ as a framework to ensure a comprehensive description of the exercise intervention/s. However, full reporting of exercise prescription methods may be futile without parallel precise reporting of the adherence to the intervention.⁶³ Accurate reporting of adherence for the comparator group (eg, MICT or control) is equally important and recommended in addition to the HIIT intervention. Furthermore, reporting of adherence, nonprotocol exercise/physical activity, and other novel adherence outcomes (such as discontinuation or interruption of interventions, early session termination, or intensity modification of sessions) not only reveals patterns of adherence and/or explanation of findings but also provides important data related to the feasibility or tolerability of the exercise protocol.⁶⁴ Moreover, applying an adherence criterion for secondary treatment or per-protocol analyses can be used to determine whether nonadherence underestimated the efficacy of the intervention for the intention-to-treat analysis. However, for reasons stated above, this should not be used as the primary analysis but rather an explanatory or sensitivity analysis. Furthermore, effort should be directed to designing studies with adequate sample sizes, which may reduce the influence of nonadherence on study results.

For home-based training, collection of adherence data can be challenging particularly when relying on self-report exercise diaries. With increasing access to technology, fitness-tracking devices with automatic data upload (that record session frequency, intensity, and duration) can allow for more accurate data collection. If fitness-tracking or HR-monitoring devices are not available, we recommend including self-report measures of exercise intensity (eg, RPE) in addition to duration and exercise type (eg, randomized vs nonprotocol exercise). For long-term follow-up studies, whereby objective exercise training data are difficult to continuously collect, we recommended using self-report exercise training questionnaires. Participants should be asked to document the same adherence components relating to type (eg, interval or continuous), frequency, intensity, and duration of sessions.

SUMMARY

This review highlights that adherence to the attendance of HIIT sessions in patients with coronary artery disease is high and comparable with MICT. However, adherence to the intensity and duration of HIIT in patients with coronary artery disease is underreported, which has implications for determining the treatment effect of exercise interventions being compared. Moreover, there remains limited evidence regarding long-term adherence to HIIT. This review provides recommendations for researchers in the measurement and reporting of adherence to HIIT and other exercise interventions to facilitate a sufficient and consistent approach for future studies. Furthermore, we have provided recommendations for clinicians to improve adherence, feasibility, and enjoyment of HIIT for their patients.