With over 17 million deaths per year, cardiovascular disease is the leading cause of mortality worldwide [1]. For heart disease survivors, rehabilitation is recommended to reduce cardiac mortality risk. Phase two cardiac rehabilitation (CR) is a comprehensive outpatient therapy focused heavily on aerobic exercise. Recent evidence has demonstrated that participation in exercise-based CR lowered mortality risk by 41% among a large cohort (N=12,265) of coronary artery disease (CAD) patients [2]. Mortality risk reduction is largely associated with improvement in functional capacity (FC), which is the primary exercise-related outcome of patients enrolled in CR. A linear relationship exists between improvement of FC and mortality risk during CR [3]. Functional capacity is defined as the ability to perform activities of daily living such as walking or climbing stairs and is an assessment of central hemodynamics (cardiac output) and peripheral skeletal muscle function (arterial-venous oxygen difference) [4]. Functional capacity is measured as maximal or peak oxygen consumption (VO₂max or VO₂peak reported in ml.kg.min), or reported in metabolic equivalents (METs) achieved on a treadmill exercise test; whereas 1 MET is equal to 3.5 ml.kg.min of oxygen consumption. A 1 MET improvement in FC as a result of CR has been associated with a 30% reduction in cardiac mortality, along with improved quality of life, lower incidences of cardiac events, and reduced hospitalizations [5-8]. Therefore, increasing FC is an important outcome measure among CAD patients who complete CR.

To achieve improvement in FC, patients receive an exercise prescription which is typically 36 exercise sessions completed over a 12 to 18-week time frame (targeting a frequency of 3 sessions per week) [9]. Guidelines suggest that the duration of each aerobic exercise session should last between 20-60 minutes and be performed at an intensity range of 40-80% FC (using peak heart rate, VO₂, or MET values from a graded exercise test) [10]. A patient’s exercise workload throughout CR is considered the most important component of the exercise program. Brawner et al. [11] indicated that a decrease in mortality risk is achieved when patients exercise at intensities above 3.5 METs after 4 weeks of CR. In addition, recent meta-analytical data has suggested that high workload interval exercise is superior to continuous moderate workload exercise in improving FC among cardiac patients [12]. However, it is important to evaluate the impact of other components of the exercise prescription such as duration and frequency on FC gains after CR participation. Evidence has suggested that patients who exercise more frequently each week of CR (5 sessions/wk) achieved greater improvements in FC than those who exercised less (3 sessions/wk) [13,14]. However, limited research regarding the influence of frequency on the outcome of CR exists. Regarding exercise session duration, similar changes in FC have been reported among patients that exercise between 30 to 60 minutes; however, greater improvement would be expected among patients who exercise longer [15,16]. This study was performed to define those parameters for CR which would lead to the most benefit, and therefore by implication, have the greatest effect on improving mortality. The purpose of the study was to retrospectively investigate the effects of three exercise prescription components (duration, workload and frequency) on the FC outcome among cardiac patients who completed 36 sessions of treadmill exercise in CR.

Study protocol

The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the University of New Mexico Institutional Review Board for Human Subject Research. Patient data were extracted from an outpatient cardiac rehabilitation facility in Albuquerque, NM. Outcome data, measured as FC (ml.kg.min), was extracted from the American Association of Cardiovascular and Pulmonary Rehabilitation (AACVPR) registry and exercise prescription data was extracted from the facility’s database between January 2019 and March 2019. Inclusion criteria included, recent CAD diagnosis (including myocardial infarction and coronary revascularization diagnosis within past 6 months), completion of 36 sessions of CR between 2013 and 2018, no previous participation in CR, no missing exercise testing and program data for the patient. Heart failure diagnosis and multiple cardiovascular incidences and/or diseases were exclusion criteria. Determination of CAD was based on recorded disease in the database.

Data from 180 patients, including sex, age, diagnosis, pre- and post-exercise test results measured in peak METs, and exercise program data, including exercise duration, treadmill speed and incline, and exercise frequency per week were extracted in an Excel file (Microsoft, Redmond, WA) which was transferred into IBM SPSS (IBM Corp. Released 2017. IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY: IBM Corp) for data analysis.

Exercise testing and programming

Each patient completed a pre- and post- FC assessment, which consisted of a symptom-limited submaximal graded exercise stress test. After a 30 second warm up at 1 mph with no incline, speed was gradually increased by 0.5 mph every minute until a speed of 3.0 mph was reached. Workload was then further increased every minute by either a further speed increase of 0.2/0.3 mph or a 2% grade increase. Test termination criteria included an RPE of 5 (hard) on the 0-10 Borg scale, the patient requested to stop, evidence of cardiovascular decompensation or musculoskeletal or neurological complications. Functional capacity was estimated based on final walking speed and grade and reported in METs.

Based on the pre-CR exercise test, an individual progressive exercise prescription was developed for each patient by the clinic staff. There were no set goals for each session; rather over the course of 36 CR sessions patients were encouraged to increase exercise duration and workload for improvement in FC. For most patients, a final total duration of 60 minutes was set as goal over the 36 CR sessions. Various modes of aerobic exercise (e.g., treadmill, cycle ergometer, Nustep®) were available for patients. Duration, workload, and mode of exercise were documented. However, only treadmill exercise was used for data analyses, due to missing workload data for other modes of exercise. Patients did not engage in strength training and no data about exercise outside of the rehabilitation setting was recorded.

Data extraction

The results of the submaximal exercise test performed before (pre) and after (post) CR were extracted in METs. The recorded treadmill speed and incline from each exercise session were recorded to calculate workload by using the established American College of Sports Medicine (ACSM) formula (VO₂ = [0.1 * speed] + [1.8 * speed * grade] + 3.5). To allow for a better picture of exercise completed every 12 sessions, the workload data and duration was averaged for sessions 11,12, and 13, representing session 12; 23,24 and 25 were averaged and reported as session 24; as well as 34-36 were averaged and reported as session 36. The reason for this approach was to eliminate possible variability caused by using only sessions 12, 24, and 36. Averaging the sessions prior to- and after sessions 12, 24, and sessions prior to session 36 created a more realistic view of the patients exercise workloads and progression. The data for day one was not averaged and represents the actual exercise prescription completed in the first CR session, or the starting point for each patient’s exercise program.

Treadmill walking was included as the sole mode of exercise. Intensity was calculated as peak METs for each exercise session. The average exercise intensity over 36 sessions was determined based on the 4 time points established (days 1, 12, 24, 36). A percentage of the average exercise intensity over 36 sessions was calculated based on the pre-CR peak MET value determined by the baseline treadmill test. Only treadmill duration was recorded for the duration of exercise. Total exercise duration was not extracted due to unknown workload data for modalities other than treadmill walking. Treadmill duration was averaged in minutes for sessions 1, 12 (average of 11,12, 13), 24 (average of 23, 24, 24), and 36 (average of 34, 35, 36). Frequency was determined based on weekly sessions attended to achieve 36 sessions.

Statistical analysis

A multiple regression analysis was performed to establish a regression model for predicting peak MET changes after 36 sessions of CR. The dependent variable for the regression model was determined as ‘change in peak METs’ which represents the change (either up or down) between pre-CR peak METs to post-CR peak METs. To identify independent correlates of the changes in METs a backward stepwise method with probability criteria range of p=0.05 to p=0.10 was applied. The initial independent variables entered into the analysis were treadmill walking workload (as an average percentage based on the FC achieved in the pre-CR exercise test); duration of treadmill walking in minutes; frequency (number per sessions attended weekly); age, measured in years; and sex (male or female). Before any data were analyzed, it was checked and met the criteria for normal distribution, homoscedasticity, independence, and multi-collinearity. Statistical outliers were identified during diagnostic evaluation, and analyses were conducted with and without the outliers to gauge the influence of the values on the calculated regression weights. Pre- and post- outcome data (change in peak METs) were compared using a student t-test. Statistical significance was set as p<0.05.

Patient characteristics

A total of 180 patient records were extracted. Patients with missing peak metabolic equivalents (METs) pre- or post-CR data; who did not walk on the treadmill; or participated in CR previously were excluded from the study (n=29). The characteristics of 151 patients (103 males; 48 female) are located in table one. The average FC at pre-CR was 4.4 ± 1.9 METs. At post-CR the average value was 6.7 ± 1.7 METs with a mean increase of 2.25 ± 1.7 METs for all patients (p<0.01). Patients attended on average 2.7 ± 0.49 exercise sessions per week, with most patients completing three sessions weekly (n=102). Over the 36 sessions of CR, patients averaged 10.92 ± 3.58 minutes of treadmill exercise. On average the total exercise duration including all exercise modalities, was 45.5 ± 9.0 minutes per session among all patients. The maximal exercise duration was 75 minutes, and each patient completed a minimal duration of 22 minutes per session.

Data used for multiple linear regression model

A multiple regression model was constructed to estimate change in peak METs from pre-CR to post-CR using the independent variables age, sex, frequency, average treadmill duration and average treadmill workload. Sex was eliminated in the final model. Both models are represented in table three.

The outliers did not drive the change in peak METs. The resultant regression model, calculated with the potential outliers included, can be found in table two. The results of the multiple linear regression suggest that a significant proportion of the total variation in change in peak METs was predicted by age, frequency, average treadmill duration and average treadmill workload, F(4,151)=14.164, p<0.001. Multiple R² indicates that approximately 26% of the variation in change in peak METs was predicted by the model (Tables 1-3).

Interpretation of multiple linear regression model

Treadmill exercise duration was a significant predictor of change in FC, with a 1 MET increase in FC for every additional 10 minute of treadmill exercise (p<0.001). The average treadmill exercise duration per session for the present study was similar for male and female patients (10.9 and 11.0 minutes, respectively). The patient with the longest average treadmill walking duration (25.1 minutes), improved FC by 6.2 METs. Workload was the strongest single predictor in the model. For each percentage increase in workload, the change in peak METs increased by 0.029 (p<0.001). Workload increased on average by 12 percent between each twelve sessions period (sessions 1-12, sessions 12-24, sessions 24-36). On days one, twelve, twenty-four, and thirty-six, patients exercised on average at 74, 86, 98 and 104% of pre-CR peak METs, respectively. The average workload for all 36 sessions among men and women was 89.6 ± 20.6 and 95.4 ± 27.2% of peak METs, respectively. Our model suggests that an average increase in exercise workload of 35% across 36 sessions yields a 1 MET increase in FC. Frequency was a not a significant predictor in the model, but trended (p=0.05), and suggests that each additional day per week of exercise as part of CR predicted an improvement in peak METs by 0.4. Age trended toward significance (p=0.06) and was a negative predictor of change in peak METs in the current model. With each additional year in age, the total change in peak METs was -0.26.

The primary aim of this retrospective study was to determine the role of exercise prescription variables on FC change among CAD patients who completed 36 sessions of treadmill exercise in CR. According to our model, average treadmill workload (reported as a percentage of pre peak METs test) and treadmill walking duration were significant predictors of FC improvement when exercise frequency, and patient age were held constant.

Workload was the strongest single predictor of change in FC which is in accordance with previous meta-analytical findings [17,18]. Each percent increase in intensity resulted in a 0.029 increase in peak METs. This corresponds to a 1 MET increase in peak METs for every 35% increase in exercise intensity over the course of a 36 session CR program. The training workload of patients throughout the study was above the 40-80% VO₂peak recommended by the AACVPR [10]. In previous treadmill studies in which patients exercised between 50-60% of VO₂peak, FC improved by 0.8 METs [15]; and, by 1.7 METs when patients exercised above 80% VO₂peak [19]. These results indicate that the ability to achieve higher exercise intensity in CR yields larger gains in FC among CAD patients [15]. Similarly, patients who are unable to progress exercise workloads in CR may be at higher risk for cardiac mortality [11]. As previously mentioned, patients who exercise at an exercise workload at, or above 3.5 METs by CR session 12 have a lower risk of cardiac mortality [11]. Further evidence has demonstrated that continued progression of exercise workload throughout 36 CR sessions is important for improving FC. Specifically, increasing treadmill exercise workloads between CR sessions 12 and 36 were significant predictors of FC changes among CAD patients [20]. The results of this study add to the evidence that exercise workload (i.e., intensity) is an integral component to exercise programming for CAD patients enrolled in CR. The results should be interpreted with caution as workload in the current study was based on the initial symptom limited submaximal exercise test and determined based on only treadmill exercise during each session, and, therefore does not represent the workload for the total exercise session.

Average treadmill walking duration was a significant predictor of change in peak METs (p<0.001). No difference in walking duration was noted between men and women in the current study. However, the longer a patient walked on the treadmill yielded a larger improvement in peak METs. The treadmill duration often reported in CR studies ranged from 30 to 45 minutes, and it must be noted that these previous studies all examined walking as the sole mode of exercise in a CR program [15,16,21,22]. Blumenthal et al. [16] showed that CAD patients who walked for 30 to 45 minutes, three times weekly for 12 weeks achieved an average improvement of 0.94 METs. Similarly, in other studies, patients who walked for either 35 or 46 minutes in each session, improved FC by 0.7 METs regardless of the minutes walked [23,24]. Further, Rognmo et al. [15] reported a 0.77 increase in peak METs among CAD patients who walked for an average of 41 minutes over 30 sessions. Reasons for the larger FC change in the current study (2.2 METs) compared to previous studies (0.7 METs), could be the higher reported exercise workload (i.e., intensity) in the present study (90% peak METs) vs. 50-60% reported on average in previous studies. In addition, only patients who completed the full 36 sessions were extracted for analyses while previous studies did not report adherence outcomes [25-27].

There is commonly an inverse relationship between exercise workload and duration, where the higher the workload, the lower duration required to gain exercise benefits. The faster a patient walks and the higher the treadmill incline, the shorter the exercising bout. Accordingly, walking duration should be prolonged at lower treadmill workloads. The current data shows that every additional ten minutes of walking predicted a 1 MET increase as does a 35% increase in workload over 36 CR sessions. Based on these findings, if a CAD patient, no matter the age or sex, walks once weekly for 20 minutes and increases workload by 35% of peak METs over 36 CR-sessions, a 3.4 MET improvement in FC should be achieved.

Frequency was not an independent predictor in the current model (p=0.05), but the model suggested that frequency was related to change in peak METs by 0.4 for each additional day per week on which a patient exercised. This demonstrates the clinical importance of exercise frequency in CR. Le Bris et al. [13] reported that CAD patients who exercised five days weekly had a greater increase in VO₂peak than patients who exercised on three days. This corresponds with the current findings. According to the model, for a patient who exercises three times weekly, peak METs would increase by 1.2 over 36-sessions; further, a patient would improve FC by 2.0 METs when exercising five days weekly. The average frequency among patients was 2.66 ± 0.48 days per week. However, an increase of peak METs by 0.4 for each day of training per week is clinically meaningful.

Age was a negative non-significant predictor (-0.26, p>0.05), but indicates that peak METs were lowered by -0.26 for every year of age. However, the model suggested a relationship of age to changes in peak METs. A study measuring the effect of cycling exercise on CAD patients in the age ranges of 45-65, 65-75, and >75 years of age, measured smaller increases in total work capacity with increasing in age [28]. In addition, age has been identified as a clinical negative determinant for peak METs, which shows that the increase would be lower in elderly individuals [29]. The increase in peak METs was greater in men than in women in this study. Cannistra and colleagues [30] reported that women and men with pre-CR peak MET values of 4.1 and 5.5, respectively, increased to 4.9 METs in women and 6.4 METs in men post-CR. However, the relative increase was higher in women (20%) compared to men (14%), but men had an overall higher absolute change [30].

Limitations to the study include the determination of influence of exercise modalities other than treadmill walking done by patients. Total exercise duration could not be considered in the study due to incomplete and inconsistent data recording, only treadmill exercise could be included in this study. Further limitations include no progression of duration or intensity of treadmill walking for some patients. While sample size reflects CAD patients who completed 36 sessions of CR, additional research is needed to further evaluate components of exercise prescriptions among the whole cardiovascular disease population.

Exercising as part of CR is important to improve FC and leads to a decreased mortality risk. Treadmill walking has a large impact on the change in FC. The results of this study demonstrated especially the importance of treadmill walking workload and duration for improving FC throughout CR. Therefore, increasing the average treadmill duration and workload across 36 sessions of CR is very important. This study demonstrated the importance of treadmill walking in achieving the primary goal of CR, which is an improvement in FC.

There are no conflicts of interest.

This research received no external funding.

Many thanks to New Heart Fitness and Health in Albuquerque, NM and especially Omar Negrete and Dr. Barry Ramo for their help and willingness to share their outcome and exercise data.

All authors contributed to this study.