Author + information
- Received April 9, 1999
- Revision received January 18, 2000
- Accepted March 6, 2000
- Published online July 1, 2000.
- Wybe Nieuwland, MD∗,
- Marike A Berkhuysen, PhD†,‡,
- Dirk J van Veldhuisen, MD, PhD, FACC∗,* (, )
- Johan Brügemann, MD, PhD∗,
- Martin L.J Landsman, MD, PhD∗,
- Eric van Sonderen, PhD§,
- K.I Lie, MD, PhD∥,
- Harry J.G.M Crijns, MD, PhD∗ and
- Piet Rispens, MD, PhD†
- ↵*Reprint requests and correspondence: Dr. D. J. van Veldhuisen, Department of Cardiology/Thoraxcenter, University Hospital Groningen, P.O. Box 30001, 9700 RB Groningen, the Netherlands
We sought to study the influence of frequency of exercise training during cardiac rehabilitation on functional capacity (i.e., peak oxygen consumption [VO2] and ventilatory anaerobic threshold [VAT]) and quality of life (QoL).
Although the value of cardiac rehabilitation is now well established, the influence of the different program characteristics on outcome has received little attention, and the effect of frequency of exercise training is unclear. Functional capacity is regularly evaluated by peak VO2 but parameters of submaximal exercise capacity such as VAT should also be considered because submaximal exercise capacity is especially important in daily living.
Patients with coronary artery disease (n = 130, 114 men; mean age 52 ± 9 years) were randomized to either a high- or low-frequency program of six weeks (10 or 2 exercise sessions per week of 2 h, respectively). Functional capacity and QoL were assessed before and after cardiac rehabilitation. Global costs were also compared.
Compared with baseline, mean exercise capacity increased in both programs: for high- and low-frequency, respectively: peak VO2 = 15% and 12%, Wmax = 18% and 12%, VAT = 35% and 12% (all p < 0.001). However, when the programs were compared, only VAT increased significantly more during the high-frequency program (p = 0.002). During the high-frequency program, QoL increased slightly more, and more individuals improved in subjective physical functioning (p = 0.014). We observed superiority of the high-frequency program, especially in younger patients. Mean costs were estimated at 4,455 and 2,273 Euro, respectively, for the high- and low-frequency programs.
High-frequency exercise training is more effective in terms of VAT and QoL, but peak VO2 improves equally in both programs. Younger patients seem to benefit more from the high-frequency training.
Cardiac rehabilitation is a well-established treatment in patients with coronary artery disease. Beneficial effects on exercise capacity, quality of life (QoL) and probability of recurrent events and hospitalization are reported in various studies using different programs (1–3). Although there are some data available regarding different levels of exercise intensity (4–7), little or no data exist on the optimal frequency of exercise.
Outcome assessment of cardiac rehabilitation programs and of medication trials is usually focused on peak exercise capacity, especially peak oxygen consumption (peak VO2). However, the value of peak VO2 has been recently questioned as an outcome parameter of studies, especially in patients with chronic heart failure. Indeed, it has been suggested that submaximal endurance exercise capacity should also be evaluated, for example, the anaerobic threshold, because this might better reflect exercise in daily life (8).
Therefore, this study compared the effects of high-frequency with low-frequency exercise training on both peak VO2 and ventilatory anaerobic threshold (VAT) during outpatient cardiac rehabilitation.
Patients who had been hospitalized with manifestations of documented coronary artery disease (myocardial infarction, angina pectoris, bypass surgery or angioplasty) were referred to our cardiac rehabilitation center by the departments of cardiology of a general and a university hospital. They were eligible for the study if their ages were between 30 and 70 years. Exclusion criteria were unstable angina, clinically unstable heart failure, unstable arrhythmias (e.g., sustained ventricular tachycardias or exercise induced polymorf ventricular tachycardias), contraindications for exercise training (e.g., endocarditis or other systemic infectious diseases), other exercise limiting concurrent condition (e.g., chronic obstructive pulmonary disease, skeletal or muscular disorders) or a psychosocial indication for inpatient cardiac rehabilitation (severe depression or panic disorder). Patients were randomized to either a high- or low-frequency exercise training during a six-week outpatient (phase II) cardiac rehabilitation program. Randomization was executed externally after assessment of baseline data and obtaining written informed consent. The study-protocol was approved by the institutional review board and was in accordance with the Helsinki Declaration. Baseline left ventricular function, ejection fraction and wall motion score index (9) were evaluated by echocardiography (Vingmed CFM 800; Vingmed Sounds, Norway).
Outline of the training programs
Duration of the rehabilitation stage was six weeks. The high-frequency program consisted of two training sessions each day, five days a week. The low-frequency program consisted of one training session a day, twice a week, without advice or prescription of additional exercise outside the program. Each training session consisted of cycling on an ergometer (6 min warm-up, 20 min endurance training with heart rate (HR) maintained on 60% to 70% of HR Reserve, 4 min cool-down) and 45 to 60 min sports activities (swimming, walking or jogging, ballsports, calisthenics). All patients joined the education program and participated in relaxation therapy and breathing technique instructions once a week (teaching awareness of respiration and of bodily tension). Spouses were also invited to join two exercise-sessions and an education program. A dietician, social worker and/or psychologist counselled patients individually.
Exercise capacity was measured both before and at the end of the exercise training program. All patients were familiarized with the exercise testing protocol by a preliminary exercise test with respiratory gas exchange measurement one to three days before the baseline exercise test. Graded symptom-limited exercise tests were performed on an electromagnetically braked cycle ergometer (Excalibur, Lode, Netherlands) as previously described in detail (10). The protocol consisted of a 3-min warm-up period at a workload of 20 W. At the next stage, the workload was increased to 50 W and then by 10 W every subsequent minute. Patients were instructed to maintain a speed of 60 to 70 rotations per min and were encouraged to perform maximally to symptoms of dyspnea or general fatigue to a level of perceived exertion of 19 to 20 according to the Borg scale (11). A complete 12-lead electrocardiogram was monitored continuously. Blood pressure was measured by cuff sphygmanometry before exercise, every 3 min during exercise and during the 6 min after exercise. A capillary blood sample was obtained within 45 s after peak exercise to measure blood-lactate concentration. During the final exercise test, we also obtained a blood sample at a submaximal exercise stage, that is, the same exercise stage as peak workload during the baseline exercise test. Patients breathed through a mask with a turbine volume transducer, measuring the volume of inspired and expired air. Respired gases were withdrawn from the mask for determination of O2 and CO2 and were analyzed breath by breath (Oxycon Champion, Jaeger, Netherlands). The gas analyzers as well as the volume transducer were calibrated before each test. Peak VO2 was defined as the mean VO2 of the last minute of the exercise test. Age- and gender-adjusted peak VO2 was calculated as a fraction of predicted peak VO2(12). Respiratory exchange ratio (RER) was calculated on line (VO2/VCO2). Ventilatory anaerobic threshold was determined using the RER = 1 method (13) and was reported in (percentage of) peak workload (Wmax).
Subjective improvement on health was assessed with the RAND-36, which is a Dutch version of the MOS SF-36 (Medical Outcomes Survey 36-item Short Form health domains) (14,15). This questionnaire is widely accepted and well validated and was completed before and immediately after the six-week program. We used seven subscales (perception of general health, vitality, physical functioning, mental health, health change, social functioning and bodily pain) to evaluate changes in QoL.
Calculation of costs
Costs of both programs were calculated on the basis of the rates of the actual system of calculation. In this system the actual duration of treatment by the various professionals directly determined the calculated costs. The program was subdivided into intake procedure, actual treatment (exercise training, individual counselling and education program) and exit-procedure, and the approximate mean time of treatment was calculated.
Statistics were obtained using SPSS (PC+, version 5.01, 1992; SPSS Inc., Chicago, Illinois). Differences between groups (high- vs. low-frequency and improvement vs. no improvement) were analyzed using unpaired t test; differences of program-effect were analyzed with multivariate analysis of variance for repeated measures. The individual effect of training programs was defined as an improvement if there was an increase of >50% of SD of the mean baseline parameter, and no improvement was defined as an increase <50% (16). Differences in individual improvement between both programs were tested with chi-square test and reported in risk-ratio. Correlation between QoL and exercise capacity was reported in Pearson product moment correlation coefficient (R). Significance was expected to occur when (two-tailed) p values were below 0.05. Group data for each variable are expressed as mean value ± SD.
We randomized 130 of a total of 186 patients who met inclusion and exclusion criteria. Reasons for nonrandomization were: 1) there was no reliable measurement of baseline parameters (i.e., exercise tests, echocardiography or questionnaires) (11 patients) and 2) a refusal to participate in one or another program (36 patients preferred one of the two programs, 9 refused for other reasons). There were no significant differences between both groups in baseline characteristics, that is, demographic, left ventricular function, exercise capacity and most parameters of health-related QoL (Table 1). However, on three parameters of QoL we observed a slightly significant difference at baseline, that is, mental health, vitality and social functioning (p of difference, respectively: p = 0.03, p = 0.04, p = 0.05). During the program five patients dropped out (high-frequency—one and low-frequency-four; p = ns). For one patient (high-frequency program) the drop-out was caused by occurrence of unstable angina pectoris, treated by coronary bypass surgery; four patients (low-frequency program) stopped attending exercise sessions for lack of motivation (3×) or because of resumption of work (1×). Other patients attended all the exercise sessions. If patients were not able to join a session, the session was rescheduled. All exercise tests were stopped if symptoms appeared, that is, general fatigue and/or dyspnea with or without some leg discomfort. In none of the patients was angina the reason for termination of the test.
Increase of exercise capacity was highly significant during both programs (Table 2). Mean increase of peak VO2 was comparable between both programs, while VAT increased more during high-frequency training (p = 0.002). Also Wmax and exercise duration increased significantly more during high-frequency training. Heart rate, systolic blood pressure, RER and blood lactate concentration during peak exercise did not change during either program (neither were there significant differences between both programs). Most individuals showed an improvement in both Wmax and peak VO2 during both programs (Fig. 1). Peak VO2 improved in half of the patients during both programs (high- and low-frequency: 30/60 and 30/62, respectively; p = ns); by contrast, significantly more individuals improved their VAT during the high-frequency program (high- and low-frequency response: 35/48 and 19/49; p = 0.002) (Table 3). Patients who improved their VAT during high-frequency training were significantly younger than patients who did not improve (50 ± 9 vs. 56 ± 8 years; p < 0.05), but no significant difference in age was observed during low-frequency training between both categories (52 ± 10 vs. 53 ± 9 years). The “improvement” group tended to have a better baseline exercise capacity than the “no-improvement” group during high-frequency training (Wmax 151 ± 40 vs. 126 ± 40, peak VO2 23.8 ± 6.8 vs. 21.3 ± 6.3; response p = 0.065 and p = 0.241). In contrast, during the low-frequency program, the improvement group tended to have a lower baseline exercise capacity (Wmax 146 ± 37 vs. 147 ± 36, peak VO2 22.6 ± 5.1 vs. 24.2 ± 6.1; response p = 0.937 and p = 0.320). This difference between the “improvement” and “no-improvement” groups was also observed concerning peak VO2 and QoL; however, this was not significant. Mean endurance exercise capacity increased significantly during both programs. This was reflected not only by a higher VAT but also by a lower HR, RER and serum lactate concentration at submaximal exercise (respectively 124 vs. 128/min, 1.04 vs. 1.07 and 4.7 vs. 5.4 mmol/l; program × time interaction effect, respectively, p = 0.047, p < 0.001; p = 0.134).
Almost all measures of the RAND-36 improved significantly in both treatment groups; this tended to be greater during high-frequency exercise (Table 2). This difference between programs was statistically significant on two subscales (mental health and health change). Also, during the high-frequency program, significantly more individuals reported improvement in subjective physical functioning (Fig. 1). Mean improvement in subjective physical functioning was significantly correlated with mean improvement of VAT (r = 0.178, p = 0.035) but not with improvement on peak VO2 (r = 0.090, p = 0.169).
Mean costs of intake and exit procedure were 591 and 318 Euro, respectively; individual counselling, approximately 682 Euro; and education program, 182 Euro. The only difference in costs between both programs was the exercise training; exercise training in the high-frequency program was five times higher than it was in the low-frequency program; respectively, 273 and 545 Euro. Total costs of both programs were 4,455 and 2,273 Euro.
Cardiac rehabilitation has become a well-established treatment modality in patients with heart disease. Effects have been demonstrated on functional and psychosocial recovery and on cardiac morbidity and (in meta-analysis) mortality (1–3). Studies also demonstrated beneficial effects in various categories of cardiac patients. In addition to postmyocardial infarction patients, patients with chronic heart failure or after heart transplantation were shown to benefit from such programs (17–19). However, most of the programs applied in these studies were not compared in a randomized way, and critical program evaluation, therefore, is crucial because development of programs tailored to the individual is demanded nowadays (20).
The present randomized study focused on program evaluation and demonstrated a beneficial effect in favor of high-frequency exercise training on QoL and (submaximal) exercise capacity. Both qualities bear on important issues that are related to the goal of rehabilitation, namely restoration of normal daily functioning. This superiority of high-frequency training was not only clear in mean improvement but also on the number of individuals who experienced significant improvement (of VAT: high- and low-frequency response 35/14 and 19/49). These better results, however, were obtained with almost twice as much money, and the question therefore arises whether high-frequency training should always be preferred. Although high-frequency training is common in some centers, low-frequency training is used the most and is known to be effective. This study confirmed the efficacy of low-frequency training, demonstrating a mean increase of at least 10% on most parameters. In most patients this increase would be enough to reach their individual rehabilitation targets (e.g., restoration of recreational activities, resumption of work). An additional benefit elicited by high-frequency training might be indicated only in specific patients, such as those with a severely decreased exercise capacity or with high physical demands in daily life. It might be speculated that only in these patients will this additional effect be sustained long-term, while in other patients this program effect would disappear. It is, however, of clinical relevance to evaluate whether the effects of this high-frequency program are maintained over a longer period of time. Our results also suggest, particularly, that younger patients might benefit.
In contrast to peak VO2, VAT improved significantly more during the high-frequency program. Additionally, a differential higher improvement of Wmax (compared with peak VO2) was measured during the high-frequency program. This discrepancy might be explained by a higher increase of VAT, by which patients might sustain exercise testing longer because of a later increase of blood lactate concentration. Other potential explanations might be a greater increase in muscular strength and the improvement of motor skills, improving the energetic efficiency of movement. A higher motivation is unlikely to explain the differential higher increase in Wmax. If patients have a higher motivation, they might sustain a higher level of anaerobiosis and symptoms of fatigue. However, blood lactate concentration, RER and peak HR were comparable between both groups.
The differential effects on peak and endurance exercise capacity raise the question of which parameter to prefer in the assessment of an intervention. Several trials in chronic heart failure have also shown disappointing effects on peak VO2, while parameters reflecting submaximal exercise capacity, such as QoL or 6-min walk distance, were favorably affected (21,22). Considering that most physical activities are on a submaximal level, this might explain the observed relation between VAT and QoL (which relates to subjective physical functioning) observed in this study.
The present study population cannot be automatically considered representative for each patient after a recent coronary event, because of referral bias. In general, only a proportion of patients are referred for cardiac rehabilitation after a recent coronary event and often only those with a decreased functional capacity or psychosocial problems. Generally, those referred are highly motivated, and this might explain the relatively low dropout rate and low initial exercise capacity in this study. In addition, we included mainly men. This is in line with the population generally referred for cardiac rehabilitation. A low number of female patients were observed by others as well and can be partly explained by a lower incidence of coronary artery disease among women. Also, women are less likely to be referred, for several reasons (23). Furthermore, some ethnic groups might be referred less in some societies. This is of particular interest because ethnicity might influence the clinical profile and predict the outcome of rehabilitation (24). This last issue did not, however, apply to our study. We included only Caucasians, and in our region, ethnicity has hardly any impact on referring patients for cardiac rehabilitation.
Physical activity outside the program might have disturbed the trial. These activities were not controlled for. Physical activity in the low-frequency program might have elicited an extra training stimulus, diminishing the difference between both programs. By contrast, extra physical activity during the high-frequency program might have had an adverse effect by overtraining. If this is true, it may have decreased the mean physiological benefits of the high-frequency program. Symptoms of overtraining were not assessed systematically in our study. In addition, a high-frequency program might not be feasible for all patients, because of other scheduled activities, and it might be speculated that a somewhat shorter program would already generate a more pronounced physiological effect.
A relatively short program of six weeks was applied in this study, this being a common length for programs used in Europe (16). Whether the results of this study are applicable to programs with a longer length (for example, three months) is unknown. However, a longer program with this high-frequency exercise training will not be easily applied in clinical practice, because of the high costs.
High-frequency exercise training in cardiac rehabilitation is more effective than low-frequency training, as VAT and QoL increased more during high-frequency exercise training, while, by contrast, there was no program effect on peak VO2. This differential effect stresses the importance of assessing VAT, especially when therapy is directed at improving functional capacity in daily life.
☆ This study was supported by a grant of the Netherlands Heart Foundation (nr. 92-354). Dr. D. J. van Veldhuisen is a Clinical Scientific Investigator of the Netherlands Heart Foundation.
- heart rate
- peak VO2
- peak oxygen consumption
- quality of life
- respiratory exchange ratio
- ventilatory anaerobic threshold
- peak workload
- Received April 9, 1999.
- Revision received January 18, 2000.
- Accepted March 6, 2000.
- American College of Cardiology
- O’Connor G.T.,
- Buring J.E.,
- Yusuf S.,
- et al.
- Hedbäck B.,
- Perk J.,
- Wodlin P.
- Worcester M.C.,
- Hare D.L.,
- Oliver R.G.,
- et al.
- Wilson J.R.,
- Rayos G.,
- Yeoh T.K.,
- et al.
- Schiller N.B.,
- Shah P.M.,
- Crawford M.,
- et al.
- Nieuwland W.,
- Berkhuysen M.A.,
- van Veldhuisen D.J.,
- et al.
- Jones N.L.,
- Campbell E.J.M.
- Denollet J.,
- Brutsaert D.L.
- ↵European Heart Failure Training Group. Experience from controlled trials of physical training in chronic heart failure. Protocol and patient factors in effectiveness in the improvement in exercise tolerance. Eur Heart J 1998;19:466–75.
- Thompson D.R.,
- Bowman G.S.,
- Kitson A.L.,
- et al.
- Krum H.,
- Sackner-Bernstein J.D.,
- Goldsmith R.L.,
- et al.
- O’Callahan W.G.,
- Teo K.K.,
- O’Riordan J.,
- et al.