Author + information
- Received October 4, 1996
- Revision received February 4, 1997
- Accepted February 20, 1997
- Published online June 1, 1997.
- Paul Dubach, MDA,*,
- Jonathan Myers, PhD, FACCB,
- Gerald Dziekan, MDA,
- Ute Goebbels, MDA,
- Walter Reinhart, MDA,
- Peter Muller, MDA,
- Peter Buser, MDC,
- Peter Stulz, MDC,
- Paul Vogt, MDD and
- Reto Ratti, MDA
- ↵*Dr. Paul Dubach, Cardiology Department, Kantonsspital, CH-7000 Chur, Switzerland.
Objectives. The aim of this study was to evaluate the effects of high intensity exercise training on left ventricular function and hemodynamic responses to exercise in patients with reduced ventricular function.
Background. Results of studies on central hemodynamic adaptations to exercise training in patients with chronic heart failure have been contradictory, and some research has suggested that training causes further myocardial damage in these patients after a myocardial infarction.
Methods. Twenty-five men with left ventricular dysfunction after a myocardial infarction or coronary artery bypass graft surgery were randomized to an exercise training group (mean age ± SD 56 ± 5 years, mean ejection fraction [EF] 32 ± 7%, n = 12) or a control group (mean age 55 ± 7 years, mean EF 33 ± 6%, n = 13). Patients in the exercise group performed 2 h of walking daily and four weekly sessions of high intensity monitored stationary cycling (40 min at 70% to 80% peak capacity) at a residential rehabilitation center for a period of 2 months. Ventilatory gas exchange and upright hemodynamic measurements (rest and peak exercise cardiac output; pulmonary artery, wedge and mean arterial pressures; and systemic vascular resistance) were performed before and after the study period.
Results. Maximal oxygen uptake (Vo2max) increased by 23% after 1 month of training, and by an additional 6% after month 2. The increase in Vo2max in the trained group paralleled an increase in maximal cardiac output (12.0 ± 1.8 liters/min before training vs. 13.7 ± 2.5 liters/min after training, p < 0.05), but maximal cardiac output did not change in the control group. Neither stroke volume nor hemodynamic pressures at rest or during exercise differed within or between groups. Rest left ventricular mass, volumes and EF determined by magnetic resonance imaging were unchanged in both groups.
Conclusions. High intensity exercise training in patients with reduced left ventricular function results in substantial increases in Vo2max by way of an increase in maximal cardiac output combined with a widening of maximal arteriovenous oxygen difference, but not changes in contractility. Training did not worsen hemodynamic status or cause further myocardial damage.
(J Am Coll Cardiol 1997;29:1591–8)
Central cardiac and pulmonary hemodynamic abnormalities are major pathophysiologic features of chronic heart failure (CHF). These abnormalities include elevations in pulmonary artery and pulmonary capillary wedge pressure, increases in left ventricular volume and reduced stroke volume and cardiac output at rest and during exercise. These changes and the symptoms that accompany them underlie reduced exercise tolerance, which is the major cause of morbidity associated with CHF. Moreover, left ventricular volume, ejection fraction and filling pressures are all important prognostic markers in this condition ([1–3]).
In recent years, exercise training has become an accepted therapeutic modality in CHF. Significant adaptations in the peripheral skeletal muscle after training, leading to a widening of the arteriovenous oxygen (a-Vo2) difference, have been documented in the last several years ([4–7]). Other favorable adaptations due to training have included increases in mitochondrial volume and density (), enhanced vasodilator capacity () and reductions in systemic vascular resistance during exercise ([4, 6, 10]). However, results of studies on central hemodynamic adaptations to programs of regular exercise training have been mixed. Some investigators ([4–7]) have reported significant increases in maximal cardiac output after training in patients with CHF, whereas others ([7–11]) have not. Left ventricular filling pressure during exercise has been reported to increase ([7, 11]), decrease ([6, 11]) or remain unchanged ([5, 7]) in response to programs of exercise training. It has also been reported () that exercise training may lead to a deterioration of global and regional left ventricular performance among patients who begin the program with significant ventricular asynergy.
We performed the present study to clarify the influence of exercise training on central hemodynamic variables at rest and during exercise among patients with left ventricular dysfunction who were in the early phase of recovery after a myocardial infarction or coronary artery bypass surgery. To clearly separate the exercise intervention from the spontaneous healing process, a high intensity training stimulus was employed in a randomized, controlled design.
Twelve male patients (mean age 56 ± 5 years) participated in the exercise group and 13 male patients (mean age 55 ± 7 years) participated in the control group. Clinical characteristics of the two groups are outlined in Table 1. All patients had sustained a myocardial infarction, and some had undergone bypass surgery, and their hospital course included the diagnosis of heart failure. Before hospital admission, none had had a previous history of heart failure. Patients exhibiting angina or electrocardiographic (ECG) signs of ischemia during exercise were excluded from the study. The presence of heart failure was documented by signs, symptoms and angiographic evidence of reduced left ventricular function (ejection fraction [EF] <40%) due to coronary artery disease. All were limited by fatigue or dyspnea on baseline exercise testing, and none had clinical evidence of pulmonary disease. All patients were New York Heart Association functional class II or III. Six patients in the exercise group and five in the control group had mild to moderate mitral regurgitation. The study was approved by the investigational review board at the Kantonsspital, Chur, Switzerland. Written informed consent was obtained, and all patients had stable symptoms for at least 1 month before randomization.
1.2 Study Design.
Group assignment was randomized, with the study period lasting 8 weeks. Patients in both groups underwent invasive hemodynamic exercise testing initially and again after 2 months. Each patient’s pharmacologic regimen remained stable throughout the study period.
1.3 Exercise Training.
After stabilization and initial testing, patients in the exercise group resided in a rehabilitation center in Seewis, Switzerland for 8 weeks. Seewis is a small village in the mountains with an elevation of 3,500 ft. (1,067 m). The center has its own staff of physicians, consisting of a medical director and three interns/residents. Program components included education, exercise and low fat meals prepared three times daily by the center’s cook. Two outdoor walking sessions (duration ∼1 h each) were performed daily, once in the morning and once in the afternoon. Walking intensity was stratified into four levels based on clinical status, exercise capacity and performance on a 500-m walking test (50-m increase in altitude) on a nearby hill. The patients were accompanied by a physician during these walking sessions. Exercise leaders carried a two-way radio for communication with the center in case of emergency. A van equipped with emergency equipment followed the group.
In addition to these walking periods, the patients performed four 45-min periods/week of monitored stationary cycling. The cycling sessions were designed to elicit an intensity equal to roughly 70% to 80% of the patient’s heart rate reserve, and the work load was increased progressively over the 2 months as tolerated. Control patients remained at home to convalesce, received usual clinical follow-up and were encouraged not to exercise beyond a level associated with normal activities of daily living.
1.4 Exercise Testing.
On the day of testing, patients were requested to abstain from food, coffee and cigarettes for 3 h before the test. Standard pulmonary function tests were performed. Maximal exercise testing was performed on an electrically braked cycle ergometer using an individualized ramp protocol. Briefly, this test entails choosing an individualized ramp rate to yield a test duration of ∼10 min (). Arterial blood lactate samples were drawn every minute throughout the test. A 12-lead ECG was monitored continuously, and blood pressure was measured manually every minute during exercise and throughout the recovery period. The patient’s subjective level of exertion was quantified every minute with use of the Borg 6-20 scale (). All tests were continued to volitional fatigue or dyspnea; no patients were limited by angina.
Respiratory gas exchange variables were acquired continuously throughout exercise by using the Schiller CS-100 metabolic system. Gas exchange variables analyzed were oxygen uptake, carbon dioxide production, minute ventilation, respiratory rate, tidal volume, oxygen pulse and respiratory exchange ratio. The lactate threshold was determined visually by consensus among two experienced reviewers using a computerized plot of the oxygen uptake versus lactate relation.
1.5 Hemodynamic Measurements.
Invasive testing was performed by using a sequential, two-step procedure. After a 12-h fast, the patient was brought to the cardiac catheterization laboratory where he underwent standard sterile preparation procedures on arrival. A Sentron thermodilution/pressure-tipped sensor catheter (Sentron Inc., Basel, Switzerland) was then inserted into the brachial vein and positioned into the pulmonary artery under fluoroscopic guidance. The catheter contains several lumens: One has a transducer tip used for measuring pulmonary artery/capillary wedge pressure; another, which terminates 2.5 cm from the tip, was used to sample mixed venous blood; and a third, which terminates 29 cm from the tip, was used to measure right atrial pressure. A 5F introducer sheath was then inserted into the radial artery. This sheath was connected to a third pressure transducer for monitoring arterial pressure and sampling of arterial blood.
After both sheaths were secured, the patient was helped off the catheterization table and onto an upright cycle ergometer placed in the catheterization laboratory. After a 5-min rest period, right atrial, pulmonary artery and radial artery calibration procedures were performed. Baseline upright values were then obtained. These included right atrial pressure, pulmonary artery pressure, pulmonary capillary wedge pressure, systolic and diastolic arterial pressure, mean arterial pressure, radial artery oxygen saturation, mixed venous (pulmonary artery) oxygen saturation, hemoglobin concentration, heart rate, basal oxygen consumption, Fick cardiac output, systemic vascular resistance and pulmonary vascular resistance.
Cardiac output was measured with use of the Fick equation as [Oxygen uptake/a-Vo2difference]. Mean systemic vascular resistance (dynes·s·cm−5) was calculated as (Mean systemic arterial pressure− Mean right atrial pressure)/Cardiac output. Mean pulmonary resistance was determined as Mean pulmonary artery pressure/Cardiac output. Ramp exercise was performed as described before, and hemodynamic variables were measured at upright rest and every 2 min during exercise. At peak exercise, all hemodynamic measurements were repeated in an orderly fashion so as to obtain them as accurately and efficiently as possible.
1.6 Left Ventricular Volumes and Mass.
Rest left ventricular volumes and mass were determined by using a 1.5-T magnetic resonance imaging (MRI) scanner with the patient supine. With use of T1-weighted spin-echo sequences, the angulation of the left ventricular long-axis view was defined in a transverse and in a parasagittal scout image. The left ventricular short-axis view was defined as imaging planes perpendicular to the left ventricular long axis. The four-chamber view was defined as an imaging plane encompassing the insertion of the anterior mitral leaflet and the apex in an oblique coronal view parallel to the interventricular septum. Cine MRI was performed by using ECG-gated gradient refocused echo sequences with a flip angle of 30° and an echo time of 6 ms. Slice thickness was 8 mm, and 14 to 16 frames/RR interval were acquired in a single imaging plane. The whole heart was continuously encompassed from the base to the apex in a short axis, and four additional cine MRI sequences were performed in the four-chamber view, thus allowing for correction of partial volume effects and for regional wall motion analysis of the apex. Total imaging time was 20 min. Left ventricular volumes were calculated as the sum of the measured cavity area times slice thickness of all slices covering the left ventricle. Left ventricular mass was obtained by the sum of the ventricular cavity area times the slice thickness multiplied by the specific myocardial gravity (1.05).
Statistical Graphics Corporation software was used to perform analysis of variance comparing hemodynamic and gas exchange responses between groups. A one-way, repeated measures analysis was used, and post-hoc multiple comparison procedures were performed by using the Scheffé method. At randomization, group differences were compared by using unpaired ttests or chi-square analysis where appropriate.
No initial differences were observed between groups in clinical or demographic data, including age, height, weight, blood pressure at rest, pulmonary function, EF or maximal oxygen uptake (Vo2max) (Table 1). No untoward events occurred during any of the exercise testing or training procedures. Three patients in the exercise group and one in the control group did not undergo invasive hemodynamic exercise testing during either the initial or the follow-up test because of either clinical reasons or technical difficulties.
2.1 Maximal Exercise Testing.
Both groups achieved a maximal respiratory exchange ratio >1.20 and had perceived exertion levels of ∼19 on all three tests, suggesting that maximal effort was generally achieved (Table 2). No patient in either group was limited by angina, and none exhibited ECG evidence of ischemia during baseline exercise testing. No differences were observed within or between groups in maximal heart rate or blood pressure. The exercise group demonstrated a 23% increase in Vo2max from test 1 to test 2 (19.3 ± 3.0 to 23.9 ± 4.8 ml/kg per min, p < 0.01), and a further 5% increase from test 2 to test 3 (Fig. 1). Concomitant increases in maximal minute ventilation, CO2production, exercise time and watts achieved were observed in the exercise group. No differences between tests were observed among control patients in Vo2max, exercise time or watts achieved.
Oxygen uptake at the lactate threshold increased significantly among all three tests in the exercise group, whereas small but insignificant decreases were observed in control patients (Fig. 1). Similar increases in exercise time and watts achieved at the lactate threshold were observed among patients in the exercise group, whereas the control group demonstrated small decreases in these variables. No differences were observed within or between groups for heart rate, systolic or diastolic blood pressure, minute ventilation, CO2production, respiratory exchange ratio, lactate or perceived exertion at this point.
2.2 Hemodynamic Exercise Testing.
In the exercise group, a significant reduction in rest heart rate was observed (99 ± 16 before vs. 83 ± 18 after training, p < 0.05). No other notable differences were observed within or between groups at rest (Table 3). Maximal cardiac output increased in the exercise group from 12.0 liters/min before training to 13.7 liters/min after training (p < 0.05). This increase was accompanied by a widening of the maximal a-Vo2difference from 13.1 ± 1.3 to 14.8 ± 1.6 ml O2/100 ml (p < 0.05). No differences were observed within either group before or after the study period in rest or maximal exercise mean arterial, wedge, right atrial or pulmonary artery pressures, stroke volume, pulmonary vascular resistance or systemic vascular resistance (Fig. 2).
2.3 Left Ventricular Volume and Mass.
Left ventricular volume, mass and EF in the two groups are presented in Table 4. No differences in left ventricular mass or end-systolic or end-diastolic volume were observed within or between groups. Changes in EF were similar between groups (39.0 ± 9 before vs. 38.2 ± 10% after in the exercise group and 37.0 ± 10 before vs. 38.3 ± 13% after in the control group).
3.1 Exercise Training in Patients With Reduced Ventricular Function.
Until recent years, it was considered prudent to exclude most patients with reduced left ventricular function from participation in programs of exercise rehabilitation. This view was based on the general perception that patients with myocardial damage were unlikely to benefit from training and the belief that the potential risks of exercise in this group outweighed any benefits. In recent years, however, several trials have demonstrated that these patients may benefit considerably from training. In these studies ([4–8, 15]), training benefits were attributed mainly to adaptations in the periphery, including an enhanced a-Vo2difference through improvements in the oxidative capacity of the skeletal muscle, in vasodilator capacity and in autonomic function. Changes in central cardiac and pulmonary hemodynamic status at rest and during exercise, such as reductions in pulmonary artery and pulmonary capillary wedge pressures, improvements in stroke volume, EF or cardiac output and reductions in systemic vascular resistance, would also be of benefit to these patients, but few studies have addressed these responses. To our knowledge, no previous studies have evaluated in such patients the effect of training on hemodynamic responses to exercise in the upright position.
Central hemodynamic abnormalities are the primary pathophysiologic feature in patients with CHF. The major goal of any therapy in this condition is to reduce left ventricular filling pressures, increase cardiac output or reduce afterload, or all three. Any intervention that does not accomplish one or more of these objectives would have an adverse effect on the patient. For example, an increase in pulmonary artery pressure at a matched level of exercise would suggest a worsening of ventricular function. Maximal cardiac output and rest filling pressures correlate with exercise capacity in these patients (), and it is well documented that these three variables are strongly associated with outcome ([17, 18]). Although exercise training improves exercise capacity and appears to lead to modest improvements in cardiac output during exercise ([4, 5]), the effect of training on other measures of central hemodynamics at rest and during exercise is unclear.
3.2 Present Results.
The major observation in this study was the absence of differences in central cardiac and pulmonary pressures within and between groups despite a considerable training response (29% increase in Vo2max). These data suggest that the benefits of training in patients with heart failure are due mainly to the combination of an increase in maximal cardiac output and a widening of the a-Vo2difference. The benefits of exercise training on the peripheral skeletal muscle of patients with reduced ventricular function have been well documented in recent years by invasive () and noninvasive ([8, 15]) techniques. Maximal cardiac output could increase through one or a combination of several mechanisms, including an increase in maximal heart rate, enhanced left ventricular end-diastolic volume, intrinsic improvement in myocardial contractility or a reduction in systemic vascular resistance during exercise. We did not observe a significant reduction in systemic vascular resistance, and the volume response to upright exercise is difficult to measure technically (). The significant (1.7 liters/min) increase in maximal cardiac output was an important finding that could be attributed to the combination of the small change in maximal heart rate, small increase in maximal stroke volume and small reduction in maximal systemic vascular resistance.
3.3 Previous Studies.
Several studies have assessed hemodynamic responses to supine exercise before and after training among patients with reduced ventricular function. Sullivan and associates () trained 12 patients for 4 to 6 months and observed no differences in pulmonary artery, pulmonary wedge or systemic arterial pressures at rest or during exercise. Although maximal cardiac output increased by 1.0 liters/min, this difference was not significant. Using a 1-month residential training program in Germany, Jette et al. () reported no differences between trained and control groups in maximal cardiac output, but EF increased in both groups. A subgroup of patients with EF <30% had a marked increase in pulmonary wedge pressure at maximal exercise. Scalvini et al. () reported a significant increase in mean maximal pulmonary artery pressure after training only among patients with severely impaired left ventricular function. These patients did not have improved Vo2max with 5 weeks of training, whereas mildly impaired patients had increased Vo2max and demonstrated no changes in central hemodynamic variables at rest or during exercise in response to training.
Differences in the findings among these studies are probably due to differences in degree of left ventricular impairment, in training programs, in timing of rehabilitation relative to the coronary event and in exercise position (supine vs. upright) when the measurements were performed. The present study is the only one of which we are aware to evaluate hemodynamic responses to training by using upright exercise. In contrast to patients in some previous studies, the patients in our exercise group had only mild impairment (mean EF 31.5 ± 7%, Vo2max 19.4 ± 3 ml/kg per min), and their heart failure was less chronic and did not exhibit severely abnormal hemodynamic pressures initially. In an effort to clearly separate the effects of training from the spontaneous healing known to occur in the weeks after a myocardial event (), we trained our patients at a comparatively high level of intensity and discouraged control patients from engaging in exercise beyond normal activities of daily living.
Exercise training among patients who have reduced ventricular function after a recent myocardial infarction has been the topic of recent debate ([7, 12, 21–23]). It has been suggested () that patients with anterior infarction who exhibit marked ventricular asynergy on entry to an exercise program are prone to further global and regional ventricular asynergy and to a reduction in contractility. Although subsequent studies have not confirmed this view, the effect of exercise training on myocardial remodeling in these patients remains in dispute ([12, 21–23]). Our study patients differed somewhat from those in the aforementioned studies (all of our patients had newly diagnosed reduced left ventricular function after anterior or inferior myocardial infarction), but we saw no evidence of further myocardial damage, observing instead increased cardiac output and no change in rest EF, left ventricular mass or end-diastolic volumes. The small (5%) increase in end-diastolic volume in the exercise group, although not statistically significant, is noteworthy in that it suggests that training may tend to increase end-diastolic volume and underscores the need for a larger trial with a longer follow-up.
Concentrated, high intensity exercise training in patients with reduced left ventricular function after a recent myocardial infarction resulted in a significant increase in exercise capacity. These increases were achieved through the combination of an increase in maximal cardiac output and a widening of the maximal a-Vo2difference, but no changes were observed in contractility or systemic vascular resistance. Training had no effect on rest ventricular mass, volumes or function or on central hemodynamic pressures at rest or during exercise. These findings have important implications for policy guidelines in cardiac rehabilitation. Although data are contradictory, the perception remains pervasive that exercise training in patients with left ventricular dysfunction after a myocardial infarction may be dangerous or lead to further myocardial damage. Our patients with mildly reduced left ventricular function had no evidence of worsening or improvement in ventricular function as determined by detailed invasive hemodynamic measurements at rest and during exercise. It appears from the present data that patients with reduced left ventricular function can participate in an exercise program safely, and benefit in a fashion similar to that of patients with normal left ventricular function after a myocardial infarction.
☆ This study was supported in part by a grant from Schweizerische Herzstiftung, Bern, Switzerland.
- arteriovenous-oxygen difference
- chronic heart failure
- electrocardiogram, electrocardiographic
- ejection fraction
- magnetic resonance imaging
- maximal oxygen uptake
- Received October 4, 1996.
- Revision received February 4, 1997.
- Accepted February 20, 1997.
- The American College of Cardiology
- Cintron G,
- Johnson G,
- Francis G,
- Cobb F,
- Cohn J
- White HD,
- Norris RM,
- Brown MA,
- Brandt PWT,
- Witlock RML,
- Wild CJ
- Coats AJS,
- Stamatis A,
- Radaelli A,
- et al.
- Sullivan MJ,
- Higginbotham MB,
- Cobb FR
- Tavazzi L,
- Ignone G
- Jette M,
- Heller R,
- Landry F,
- Blumchen G
- Hambrecht R,
- Niebauer J,
- Fiehn E,
- et al.
- Myers J,
- Dubach P,
- Goebbels U,
- et al.
- Jugdutt BI,
- Michorowski BL,
- Kappagoda CT
- Myers J,
- Buchanan N,
- Walsh D,
- et al.
- Adamopolous S,
- Coats AJS,
- Brunette F,
- et al.
- Cohn JN
- Froelicher VF,
- Myers J,
- Follansbee W,
- Labovitz A
- Dubach P, Myers J, Dziekan G, et al. Effect of exercise training on myocardial remodeling in patients with reduced ventricular function following myocardial infarction: application of MRI. Circulation. In press.
- Iskandrian AS
- Giannuzzi P,
- Tavazzi L,
- Temporelli PL,
- et al.