Physiological Outcomes of Aerobic Exercise Training in Hemiparetic Stroke Patients
Background and Purpose In hemiparetic individuals, low endurance to exercise may compound the increased energy cost of movement and contribute to poor rehabilitation outcomes. The purpose of this investigation was to describe how hemiparetic stroke patients responded to intense exercise and aerobic training.
Methods Forty-two subjects were randomly assigned to an exercise training group or to a control group. Treatments were given three times per week for 10 weeks in similar laboratory settings. Baseline and posttest measurements were made of maximal oxygen consumption, heart rate, workload, exercise time, resting and submaximal blood pressures, and sensorimotor function.
Results Only experimental subjects showed significant improvement in maximal oxygen consumption, workload, and exercise time. Improvement in sensorimotor function was significantly related to the improvement in aerobic capacity. After treatment, experimental subjects showed significantly lower systolic blood pressure at submaximal workloads during the graded exercise test.
Conclusions We conclude that hemiparetic stroke patients may improve their aerobic capacity and submaximal exercise systolic blood pressure response with training. Sensorimotor improvement is related to the improvement in aerobic capacity.
Approximately 500 000 documented cerebrovascular accidents, or strokes, occur in the United States each year.1 At least 60% cause persistent neurological impairments that affect functional abilities.2 Traditional emphasis in stroke rehabilitation has been on improving self-care abilities through muscle strength and coordination training.3 However, stroke patients are known to have low endurance to exercise, which may decline further after discharge from formal rehabilitation.4 5 Low endurance may compound the increased energy cost of movement associated with residual hemiparesis and may contribute to poor rehabilitation outcomes.3 Yet the role of aerobic exercise training in the functional recovery of patients after stroke has not been well documented. Little information is available on the responses to aerobic exercise and effects of aerobic exercise training in hemiparetic stroke patients.6 7
The purposes of this investigation were to (1) describe the responses of hemiparetic stroke patients to intense exercise and (2) determine the effect of aerobic training on cardiovascular and functional outcome measures.
Subjects and Methods
The study included male and female subjects aged 21 to 77 years who had had a hemispheric stroke at least 6 months before study entry, were medically stable, and had completed a formal rehabilitation program after the stroke. Subjects with brain stem lesions were not included in the study. Cerebral vascular events and lesion location were documented by a computed tomographic scan and medical diagnosis. Lesions were classified by hemisphere. Subjects had mild-to-moderate hemiparesis including both an upper and lower limb, which was documented on physical examination and on the Fugl-Meyer Index.8 Exclusion criteria related to disorders that would preclude maximal exercise testing or confound the measurement of maximal exercise parameters. Specifically, exclusion criteria included clinical evidence of unstable cardiac disease, uncontrolled hypertension, peripheral vascular disease, pulmonary disease such as asthma or other chronic obstructive pulmonary diseases, renal or hepatic failure, alcohol or drug abuse, anemia, diabetes mellitus requiring insulin therapy, and other disorders potentially associated with medical instability, such as metastatic cancer, systemic infection, hypoglycemia, or hyperthyroidism.
After a thorough explanation of the study, all subjects signed informed consent forms that were approved by the institutional human studies committee. Subjects underwent a complete history and physical examination, chest radiograph, electrocardiogram, and laboratory blood tests to screen for comorbid diseases. Subjects who met entry criteria were randomly assigned to either the experimental group (a 10-week aerobic exercise training program) or to the control group (a 10-week program of passive range-of-motion exercise). Subjects received their respective treatments three times per week, for 30 minutes each session, for 10 weeks in similar laboratory environments. Baseline and posttest measurements included resting heart rate, blood pressure, body weight, maximal heart rate, oxygen consumption (V̇o2), expiration per minute (V̇e), carbon dioxide production (V̇co2), workload, exercise time, and exercise blood pressure at submaximal workloads, as well as an assessment of sensorimotor function using the Fugl-Meyer Index.8
Graded Exercise Test Protocol
Exercise tests were performed in a humidity- and temperature-controlled laboratory and at the same time of day for each subject. Tests were conducted according to standard criteria of the American College of Sports Medicine.9 Subjects were acclimated to the testing procedures. The exercise protocol began with seated rest on the bicycle ergometer for 2 minutes. Dynamic exercise began at 10 W, and workloads increased by 10 W each minute until maximal effort was achieved. Criteria for a maximal effort included voluntary exhaustion (subjects could no longer continue pedaling) and a respiratory exchange ratio (RER) greater than 1.15. An exercise test was terminated and subjects were excluded from the study if there were any untoward cardiac signs or symptoms that warranted stopping the test for patient safety.9 At the point of voluntary exhaustion, final measurements were made, and the subjects were given cool-down and recovery periods. To document reliability of the exercise measurements, a randomly selected subset of subjects was given two pretreatment exercise tests separated by 48 hours. Subjects were maintained on all drug prescriptions, including dosage, throughout the study.
Exercise Training Protocol
Subjects in the experimental group were exercised on an adapted cycle ergometer for 30 minutes a session, three times per week, for 10 weeks. During the first 4 weeks of the program, the training load was gradually increased from a workload representing 30% to 50% of maximal effort to the highest level attainable by the subject. The highest training load was then maintained for the final 6 weeks of training.
Passive Range-of-Motion Protocol
Subjects in the control group were given passive exercise for range of motion to body joints in a systematic procedure for 30 minutes, three times a week, for 10 weeks. Heart rate and blood pressure were monitored as in the experimental aerobic training group.
Heart rate was obtained from the RR interval on the electrocardiogram. Resting heart rate was determined after the subject sat for 20 minutes at rest in the laboratory before getting on the ergometer for a graded exercise test. Maximal heart rate during the graded exercise test was the average heart rate during the last 30 seconds of exercise.
Brachial artery blood pressure in the subject’s unaffected arm was measured using a calibrated mercury sphygmomanometer. Procedures followed the American Heart Association Recommendations for Human Blood Pressure Determination.10 The systolic blood pressure (SBP) was determined at the first Korotkoff sound and the diastolic blood pressure (DBP) at the fifth Korotkoff sound. Resting blood pressure was determined after the subject sat for 20 minutes at rest in the laboratory before getting on the ergometer for a graded exercise test. Blood pressure during exercise was measured every 2 minutes during the graded exercise test and at specified intervals during training sessions.
Exercise Metabolic Parameters
V̇o2, V̇co2, V̇e, and RER were continuously determined during graded exercise tests using a breath-by-breath respiratory gas analysis assembly and appropriate computerized software (Medical Graphics Cardiopulmonary Exercise Stress Testing System, model 2001). Maximal values for exercise parameters were the average values during the last 30 seconds of exercise.
The electronically braked ergometer (Mijnhardt, model KEM-3) features an automatic resistance adjustment to maintain the specified workload within a wide range of cadence (50 to 70 rpm). Subjects achieved the specified workload by maintaining the cadence between 50 to 70 rpm. The maximal workload for each graded exercise test was the highest workload maintained for at least 30 seconds.
Exercise time was the total amount of time in seconds of the graded exercise test counted from the first downward stroke of the ergometer pedal until the subject stopped pedaling.
Functional improvement was assessed with the Fugl-Meyer Index,8 a 113-item observer-rated measure of sensorimotor function. Total and subscale scores were used to determine functional improvement with training. Previous work in our laboratory established the α internal consistency reliability of the instrument as .96 for the total scale and .75, .56, .51, and .96 for the sensation, proprioception, balance, and motor subscales, respectively. The intraclass correlation of interrater reliability was .99 for the total instrument and .97, .94, .63, and .99 for the sensation, proprioception, balance, and motor subscales, respectively.11 The Fugl-Meyer Index was administered by three registered physical and occupational therapists who were trained in the procedure.
This was a randomized, controlled clinical trial with a 2×2 experimental design to determine the impact of aerobic training on physiological and functional outcomes in hemiparetic stroke patients. A power analysis using data from our preliminary case work showed that a sample size of 21 per cell will produce a power of .80 with an effect size of .884 (standardized units) with an α of .05. The preliminary case work also showed that we could expect this magnitude of effect size for the principle variables in the study. Descriptive statistics were used to characterize the sample and to display baseline exercise responses. All data are presented as mean±SEM. Data were subjected to a series of statistical analyses. First, treatment effects were determined by a repeated-measures ANOVA of baseline and posttreatment scores by treatment group. The differences between baseline and posttest measurements were then computed as percentage changes to describe how exercise training changed maximal V̇o2, workload, exercise time, and Fugl-Meyer Index. These percentage changes were used in subsequent ANOVA and ANCOVA of treatment outcomes. Two-way ANOVA was used to determine the influence of lesion location (right or left hemisphere) on treatment outcomes. A series of covariate analyses were performed to determine the influence of age, pretreatment fitness level, training workload, and pretreatment functional status on treatment outcomes. Pearson correlations were performed to determine the magnitude and significance of the relationship between improvement in aerobic fitness and improvement in functional outcome.
To determine the effect of training on exercise blood pressure responses, SBP and DBP were measured at three submaximal workloads during exercise (0, 20, and 40 W). These three blood pressure readings were examined in a subset of subjects (n=25) who completed at least 40 W of exercise. For each of the three workloads, we measured the magnitude of change with treatment as the difference between pretest and posttest measures. These derived-change scores were subjected to repeated-measures ANOVA to determine differences in treatment effects by group.
Finally, a secondary analysis was performed to determine the relationships between the magnitude of pretreatment SBP and DBP response and the average percentage of drop in exercise blood pressure from before to after treatment. The pretreatment magnitude of blood pressure rise, or blood pressure response during exercise, was determined by the difference between blood pressure readings at 40 W (4 minutes) and blood pressure at the resting workload (0 minute). Pearson correlations were used to determine the relationship between the variables.
Descriptive data are displayed in Table 1⇓. Fifty-five subjects met medical eligibility criteria. Of these, 10 subjects were excluded after the first exercise test for medical reasons, and 3 subjects declined to participate for personal reasons. The sample consisted of 23 men and 19 women, aged 43 to 72 years. All subjects had completed a stroke rehabilitation program for an average of 216±40.3 days (range, 14 to 1098 days). The mean score on the Fugl-Meyer Index was 177±6, which indicates moderate sensorimotor impairment. There were no statistically significant baseline differences between demographic or clinical characteristics of treatment groups.
Reliability of Exercise Responses
Twenty-five subjects were given two baseline exercise tests from which intraclass correlation coefficients and coefficients of determination were derived.12 The intraclass correlation coefficients for maximal V̇o2, heart rate, workload, SBP, and DBP measured at 40 W of exercise were .94, .97, .99, .83, and .72, respectively. The coefficients of determination for these variables were .94, .97, .99, .85, and .74, respectively.
Exercise Responses and Aerobic Training Effects
The baseline and posttreatment measurements are displayed in Table 2⇓. The improvement in maximal V̇o2 for exercise subjects ranged from 0% to 35.7%. Changes in resting measurements were not significant for either group.
The two-way ANOVAs for treatment group and lesion location were not significant for V̇o2, workload, or exercise time expressed as percent improvement. In the covariate analyses, where percent improvement in V̇o2 was the dependent variable, the training workload was the only significant covariate (F=5.97, df=2, P=.022).
Mean values for baseline and posttreatment SBP and DBP were measured at submaximal points during the graded exercise tests and are shown in the Figure⇓. These blood pressure values include data on 25 subjects who exercised through at least 40 W on the graded exercise test. Seventeen of the 25 subjects in this subgroup had diagnosed hypertension, whereas the remaining 8 subjects were without hypertension and were equally distributed between the experimental and the control group. The ANOVA was significant for the SBP treatment group by workload interaction (F=3.27, df=2, P=.047). A similar pattern of change with treatment was observed for DBP, but the DBP treatment group by workload interaction was not quite significant (F=2.21, df=2, P=.12).
SBP response, measured in the baseline exercise test, was moderately correlated with the average drop in SBP from the baseline to the posttreatment exercise test across all three workloads measured (r=−.62, P=.04). The correlation for DBP was not significant.
Relationship Between Improvement in Aerobic Fitness and Improvement in Sensorimotor Function
The ANOVA for between-group changes in Fugl-Meyer Index scores was nonsignificant. The Pearson correlation of the improvement in V̇o2 and the improvement in the total score of the Fugl-Meyer Index in the aerobic exercise treatment group was significant (r=.56, P=.012). However, the correlations for the four subscales of the Fugl-Meyer Index were not significant.
Subjects studied were moderately impaired hemiparetic stroke patients who were medically stable and well into the long-term recovery phase after stroke. The baseline aerobic capacity of subjects was less than that observed in otherwise normal hypertensive subjects of similar ages.13 This finding in hemiparetic patients is likely related to low endurance, the reduction in number of motor units capable of being recruited during dynamic exercise,14 and the reduced oxidative capacity of paretic muscle.15 The maximal heart rate responses were also influenced by the presence of β-adrenergic receptor antagonists in some subjects. However, these subjects’ average maximal workload (61.6 W) and maximal heart rate (134.1 beats per minute [bpm]) are similar to the maximal workload and heart rate during aerobic testing reported in the three previous investigations of responses to intense exercise in hemiparetic patients. King et al,5 in a study of 70 hemiparetic stroke patients who used a Schwinn Air-Dyne ergometer adapted for use in a wheelchair, reported an average workload of 336 Kpm/min (or 56 W) and an average peak heart rate of 121 bpm. Bjure et al16 and Bachynski-Cole and Cumming17 studied 10 and 8 hemiparetic stroke patients, respectively, using an ergometer and reported average peak heart rates of 124 bpm and 126 bpm, respectively. Comparable data on maximal V̇o2, V̇co2, and RER, as observed in the present study, are not available in previous reports of intense exercise in hemiparetic stroke patients. However, the absolute amount of oxygen consumed per submaximal workload in these hemiparetic subjects is greater than that observed in normal subjects of similar age and body size.18 This oxygen increase confirms the observations of Hoskins,4 who showed the increased energy cost of submaximal aerobic exercise in hemiparetic subjects. The increased energy cost of exercise may be related to the reduced efficiency of motion and to the presence of spasticity.18 19 The increased energy cost of spastic muscle has been reported in subjects with spasticity due to multiple sclerosis.19
Exercise training significantly increased mean maximal V̇o2, workload, and exercise time. However, the improvement across subjects was not uniform. The only variable among training workload, age, baseline level of fitness, lesion location, and sensorimotor function that significantly predicted treatment response was training workload. The percent improvements in peak workload and exercise time are greater than those expected for the percent improvement in maximal V̇o2. This suggests that efficiency of motion is improving to a greater extent than is aerobic capacity.
In a study of one-leg and two-leg exercise in hemiparetic subjects, Landin and colleagues15 showed a reduced blood flow, an augmented lactic acid production, and a decreased capacity to oxidize free fatty acids in paretic muscle. Given that the purpose of our investigation was to demonstrate the overall trainability of hemiparetic subjects using a two-leg exercise protocol, our findings do not explicate the physiological mechanisms that influence fatigue, endurance, or trainability. We cannot be sure whether the increment in V̇o2 occurred as a result of central or peripheral adaptation, or both. Furthermore, the ability of paretic muscle to increase oxidative metabolism with exercise training is unknown. This study does suggest, however, that the overall improvement in aerobic capacity has implications for the submaximal effort that hemiparetic patients engage in each day; these patients can carry out their daily activities at a lower percentage of maximal V̇o2.
Exercise training significantly improved the exercise SBP across the three workloads examined. Although others have shown the value of exercise training in reducing resting blood pressure in hypertensive subjects,20 21 the results of this study demonstrate that exercise training attenuates SBP during moderate exercise in hemiparetic subjects. A pattern of attenuation was also observed for DBP at two of the three workloads; however, these changes were not statistically significant. This may be attributable to insufficient power of the study to detect small effects. The pretreatment SBP response correlated significantly with the percent improvement in the average exercise blood pressure; this correlation suggests that the improvement in SBP is greatest in subjects who had the highest baseline responses to exercise workload. This suggestion may be especially important because, for a given absolute workload, the magnitude of baseline blood pressure response in these hemiparetic subjects was greater than that observed in hypertensive subjects of similar ages.22 The elevations in blood pressure during exercise observed in this study may be comparable to those that occur during activities of daily life, and the influence of elevated blood pressure on the risk of future stroke is well documented.23 Many of the hemiparetic subjects studied were hypertensive and taking a variety of antihypertensive medications. The observed reductions in blood pressure with training are in addition to attenuations achieved with drug treatment alone, as observed in the baseline exercise test.
Our inability to detect changes in resting blood pressure may be related to the maximum effect of the antihypertensive drug treatment in the resting condition. Although others have observed a synergistic effect of exercise training and drug treatment on resting blood pressure in hypertensive subjects, this has not been a consistent finding.20 The reduction in resting blood pressure in hypertensive subjects after training has also been associated with low-intensity, rather than moderate- or high-intensity, training workloads.20 The relatively high training load in the present study may have contributed to the lack of change in resting blood pressure.
The ANOVA for between-group changes in sensorimotor function was nonsignificant. However, because the improvement in V̇o2 was not uniform among subjects, it was hypothesized that improvement in sensorimotor function might be related to the magnitude of improvement in aerobic capacity. To test this hypothesis, the data from the treatment group was subjected to a Pearson correlation of the improvement in V̇o2 and the improvement in the total score of the Fugl-Meyer Index; the result was significant. The moderate and significant correlation between the Fugl-Meyer Index and the improvement in aerobic capacity suggests that sensorimotor improvement with training is related to the percent improvement in aerobic capacity, or exercise intensity, rather than bicycle exercise per se.
In summary, the results of this study demonstrate that moderately disabled, chronically hemiparetic stroke patients may improve their aerobic capacity with adequate exercise training. The subjects undergoing aerobic exercise also demonstrated improvement in functional workload and exercise time to a greater extent than expected for the increase in aerobic capacity. A subset of subjects in the aerobic exercise group also showed significant attenuation of exercise SBP with training. This result may have important implications for reduction of cardiovascular risk in patients who have clinically significant elevations of blood pressure during exercise. Finally, the improvement in aerobic capacity was significantly related to improvement in sensorimotor function, which indicates that exercise training functionally benefits those subjects able to train at an intensity high enough to increase aerobic capacity.
This study was funded by Department of Health and Human Services National Institutes of Health grant NRO209603.
- Received June 24, 1994.
- Revision received September 28, 1994.
- Accepted October 14, 1994.
- Copyright © 1995 by American Heart Association
American Heart Association. Heart and Stroke Facts. Dallas, Tex: American Heart Association; 1993.
Duncan PW, Badke MB, eds. Stroke Rehabilitation: The Recovery of Motor Control. Chicago, Ill: Year Book Medical Publishers; 1987.
Hoskins TA. Physiologic responses to known exercise loads in hemiparetic patients. Arch Phys Med Rehabil. 1975;56:544.
King JL, Guarracini M, Lenihan L, Freeman D, Gagas B, Boston A, Bates E, Nori S. Adaptive exercise testing for patients with hemiparesis. J Cardiopulmonary Rehabil. 1989;9:237-242.
Fugl-Meyer AR. Post-stroke hemiplegia assessment of physical properties. Scand J Rehabil Med. 1980;7(suppl):85-93.
American College of Sports Medicine. Guidelines for Exercise Testing and Prescription. Philadelphia, Pa: Lea & Febiger; 1991.
Frohlich ED, Grim C, Labarthe DR, Maxwell MH, Perloff D, Weidman WH. Recommendations for human blood pressure determination by sphygmomanometers: steering committee of the American Heart Association. Circulation. 1988;77(suppl):502A-514A.
Hubalik CA. The Effect of Bicycle Exercise on Balance and Sensorimotor Function in Hemiparetic Stroke Patients. Chicago, Ill: University of Illinois at Chicago; 1991. Thesis.
Kirk RE. Experimental Design. Pacific Grove, Calif: Brooks/Cole Publishing Co; 1993:162.
Heath GW, Hagberg JM, Ehsani AA, Holloszy JO. A physiological comparison of young and older endurance athletes. J Appl Physiol. 1983;51:634-640.
Ragnarsson KT. Physiologic effects of functional electrical stimulation-induced exercises in spinal cord-injured individuals. Clin Orthop. 1988;126:53-63.
Bachynski-Cole M, Cumming G. The cardiovascular fitness of disabled patients attending occupational therapy. Occup Ther J Res. 1985;5:233-242.
Brinkmann JR, Hoskins TA. Physical conditioning and altered self-concept in rehabilitated hemiplegic patients. Phys Ther. 1979;59:859-865.
Paterson DH, Cunningham DA, Himann JE, Rechnitzer PA. Long term effects of exercise training on VO2max in older men. Can J Sport Sci. 1988;13:124P-125P.
Hagberg JM, Seals DR. Exercise training and hypertension. Acta Med Scand Suppl. 1986;711(suppl):131-136.
Marceau M, Kouame N, Lacourciere Y, Cleroux J. Effects of different training intensities on 24-hour blood pressure in hypertensive subjects. Circulation. 1993;88:2803-2811.
Keli S, Bloemberg B, Kromhout D. Predictive value of repeated systolic blood pressure measurements for stroke risk. Stroke. 1992;23:347-351.