After stroke, the early 1 and persistent 2 decline in aerobic

Transcription

1 Effects of an Aerobic Exercise Program on Aerobic Capacity, Spatiotemporal Gait Parameters, and Functional Capacity in Subacute Stroke Neurorehabilitation and Neural Repair Volume 23 Number 4 May The Author(s) / Ada Tang, MSc, Kathryn M. Sibley, MSc, Scott G. Thomas, PhD, Mark T. Bayley, MD, Denyse Richardson, MD, William E. McIlroy, PhD, and Dina Brooks, PhD Background and objective. In spite of the challenges, engaging in exercise programs very early after stroke may positively influence aerobic capacity and stroke-related outcomes, including walking ability. The objective of this study was to evaluate the feasibility of adding aerobic cycle ergometer training to conventional rehabilitation early after stroke and to determine effects on aerobic capacity, walking ability, and health-related quality of life. Methods. A prospective matched control design was used. All participants performed a graded maximal exercise test on a semi-recumbent cycle ergometer, spatiotemporal gait assessments, 6-Minute Walk Test, and Stroke Impact Scale. The Exercise group added 30 minutes of aerobic cycle ergometry to conventional inpatient rehabilitation 3 days/week until discharge; the Control group received conventional rehabilitation only. Results. All Exercise participants (n = 23) completed the training without adverse effects. In the 18 matched pairs, both groups demonstrated improvements over time with a trend toward greater aerobic benefit in the Exercise group with 13% and 23% increases in peak VO 2 and work rate respectively, compared to 8% and 16% in the Control group (group-time interaction P =.71 and.62). A similar trend toward improved 6-Minute Walk Test distance (Exercise 53% vs Controls 23%, P =.23) was observed. Conclusion. Early aerobic training can be safely implemented without deleterious effects on stroke rehabilitation. A trend toward greater improvement in aerobic capacity and walking capacity suggests that such training may have an early beneficial effect and should be considered for inclusion in rehabilitation programs. Keywords: Stroke; Rehabilitation; Exercise; Walking After stroke, the early 1 and persistent 2 decline in aerobic capacity leads to diminished physiologic fitness reserve. 3 Lowered peak exercise responses result from reductions in the number of recruited motor units, as well as oxidative capacity and endurance of the paretic muscles. 4 These significant cardiovascular and neuromuscular changes, combined with the high energy costs of walking 5,6 and the presence of comorbidities, result in considerable functional limitations for this population. A meta-analysis of aerobic training trials in the chronic phase (>3 months) poststroke revealed a significant homogeneous standard effect size in improved aerobic capacity, walking speed, and walking endurance. 7 Recent studies have also investigated the effects of aerobic exercise training early after stroke (approximately 2 weeks). In a trial of 14 participants, supported treadmill training added to inpatient rehabilitation resulted in greater improvement in peak aerobic capacity (VO 2 peak) compared to a control group. 8 A larger trial (n = 92) using cycle ergometry exercise found lower resting heart rate (HR), higher peak work rate (WR) and longer exercise time post-training (VO 2 peak was not directly measured). 9 Despite these reports, traditional stroke rehabilitation models in the subacute phase are typically focused on reducing the impact of sensorimotor deficits, sequelae more commonly associated with stroke, without a significant aerobic training component. 10 Impairments in cardiorespiratory function and neuromotor control are independent, yet can mutually reinforce one another: poor cardiorespiratory fitness may restrict the recovery of appropriate motor patterns resulting from increased metabolic demands, whereas neurological dyscontrol may sufficiently impede movement required to the challenge the cardiorespiratory system. The interaction between sensorimotor and cardiorespiratory impairments culminates in the task of walking, which presents an interactive challenge to rehabilitation professionals. Few studies have examined the effects of aerobic exercise on gait, and the parameters measured have been limited with From the Department of Physical Therapy (AT, KMS, DB), Institute of Medical Science (AT, KMS, SGT, DB, WEM), and Faculty of Physical Education and Health (SGT), University of Toronto, Ontario; Toronto Rehabilitation Institute, Ontario (AT, KMS, SGT, MTB, DR, DB, WEM); Heart and Stroke Foundation of Ontario Centre for Stroke Recovery, Sunnybrook Health Sciences Centre, Toronto, Ontario (WEM, DB); and the Department of Kinesiology, Faculty of Applied Health Sciences, University of Waterloo, Ontario (WEM), Canada. Address correspondence to: Dina Brooks, PhD, Department of Physical Therapy, University of Toronto, 500 University Avenue, Room 160, Toronto Ontario Canada M5G 1V7. dina.brooks@utoronto.ca. 398

2 Tang et al / Effects of an Aerobic Exercise Program in Subacute Stroke 399 inconsistent findings. Two aerobic intervention trials used cycle ergometry training within 3 months after stroke. One study observed no post-training changes in 10-meter walk time. 9 The other trial used electromyography to measure muscle coordination during pedaling and did not observe corresponding changes in lower extremity neuromotor control in a secondary analysis, 11 despite improvements in 10-meter walk time. 12 The inclusion of more detailed gait assessments in such intervention trials may provide useful insights regarding the potential effectiveness of aerobic training on walking competency and may be important to establish improvements in efficiency during lower limb function. This has yet to be explored in the literature. Issues related to the selection of appropriate training modalities are challenging with respect to stroke. When considering paradigms that would appropriately target both the cardiorespiratory system and neuromuscular control of gait, the obvious option would be to choose a walking program (overground or treadmill). However, the use of such modalities may exclude a significant proportion of the population for whom aerobic training is desired. Particularly early after stroke, at least 50% of survivors may be nonambulatory 13 and would not be able to participate in an aerobic walking program. Indeed, in a trial conducted by Macko and colleagues, 3 24% of individuals screened were excluded based on specific issues related to the ability to perform the walking intervention. In addition, even among those who are capable of independent ambulation, a significant proportion is unable to walk a speed or distance necessary to achieve aerobic benefits. As an alternative to walking programs, seated ergometry paradigms improve aerobic capacity, and neuromuscular networks for pedaling and gait share common neural resources. 14 Such a model offers additional trunk support for individuals with poor postural control (such as stroke), 15 and we propose may be suitable for targeting both cardiorespiratory fitness and improving neuromuscular control of gait in a wide range of patients after stroke. 16 As an addi tional benefit, considering feasibility in subacute rehabilitation faculties, this training has the potential benefits of lower demands on real-time clinical supervision. The purpose of this study was (1) to establish the feasibility of structured aerobic cycle ergometry training for individuals with a broad range of motor involvement in the subacute phase (< 3 months) after stroke, and (2) to compare the short-term effects of adding such training to conventional stroke rehabilitation on aerobic capacity and spatiotemporal gait parameters, functional ambulation, and health-related quality of life. We hypothesized that semi-recumbent cycle training, chosen as the modality that is most applicable to the broadest range of participant abilities, could be feasibly added to stroke rehabilitation. Furthermore, we hypothesized that aerobic training would increase aerobic capacity and gait speed, improve spatial characteristics and temporal symmetry of gait at preferred and maximal paces, and improve functional ambulation and health-related quality of life. Method The study procedures were followed in accordance to institutional guidelines and were approved by the local university and hospital research ethics committees. Informed written consent was obtained from all study participants. Participants Individuals less than 3 months poststroke were recruited from the Toronto Rehabilitation Institute if they were able to provide informed consent, understand the evaluation procedures, walk at least 5 meters independently, and have a Chedoke- McMaster Stroke Assessment (CMSA) 17 leg impairment score of at least 3 (where spasticity and weakness are marked) but less than 7 (normal limb function including complex movement patterns with appropriate muscle timing and coordination). Participants were excluded if they exhibited any contraindications to maximal exercise testing as outlined by the American College of Sports Medicine, 18 or musculoskeletal impairments or pain that would limit pedaling ability. All participants in the Exercise group were included in the assessment to establish feasibility of early aerobic training poststroke. Because of established age-related changes and sex differences in aerobic capacity, 19 a matched control design was used to determine training effects, where the Control group participants were matched to the Exercise group in terms of age (± 5 years) and sex and did not receive the intervention. Four hundred forty patient admissions to the inpatient stroke rehabilitation unit were screened from September 2003 to April Screening occurred within the first week of inpatient rehabilitation; participants were entered into the study 8.7 ± 0.7 days after admission. Participant demographic information, stroke lesion type and location, time poststroke, degree of neurologic deficit using the National Institutes of Health Stroke Scale, 20 functional ability using the Functional Independence Measure, 21 and level of leg impairment using the CMSA 17 were collected. Measures All measures were performed upon study entry and before discharge from inpatient rehabilitation. Assessors were not blinded to group allocation. Aerobic capacity. All participants completed a graded maximal exercise test using a semi-recumbent cycle ergometer (Biodex Medical Systems, Shirley, NY) and measurement of respiratory gas exchange (AEI Technologies, Pittsburgh, PA). The test protocol, described in detail elsewhere, 15 provided important information for evaluating peak aerobic capacity and thus prescribing appropriate training loads. Gait assessment. Participants who were capable of independent ambulation walked across a 5-meter-long pressure-sensitive

3 400 Neurorehabilitation and Neural Repair mat (CIR Systems, Clifton, NJ). Three trials were performed at the participant s preferred pace and 3 trials at their maximally comfortable speed (fast pace). Participants used walking aids as prescribed by the treating physical therapist. A research assistant provided safety supervision during each trial. Primary outcome measures from the preferred and fast-paced gait assessments were gait speed (cm/s) and a calculated between-limb temporal symmetry ratio. Temporal symmetry ratio was determined by first calculating the ratio of swing time/ stance time for each limb. For description purposes, the nonparetic swing time/stance time ratio was then divided by the paretic swing time/stance time ratio. The resulting ratio indicated that a ratio greater than 1.0 represented increased time spent on the nonparetic limb, and a score greater than 1.1 was considered outside the normal range. For statistical purposes, the ratio was calculated by dividing the limb with the larger ratio by the limb with the smaller ratio. This generated an absolute ratio of symmetry with perfect symmetry equaling a ratio of 1.0 and any asymmetry increasing from 1.0, irrespective to the direction of the asymmetry. Although this calculation could not indicate whether more time was spent on the nonparetic or paretic limb, these rectified symmetry scores ensured that statistical comparisons would have increased validity. Six-Minute Walk Test (6MWT). Standardized instructions 22 were given to walk as far as possible over a 30-meter course in 6 minutes. The distance covered was the primary outcome of this test. This was compared to reference equations for the 6MWT and also reported as a percentage of predicted distance walked. 23 When necessary, participants used gait aids as prescribed by the treating physical therapist. No encouragement was provided during the test. Because of the known practice effect in other populations, 24 at least 1 practice trial was performed. Stroke Impact Scale (SIS). The SIS 25 is an interviewadministered instrument combining disability and healthrelated quality of life dimensions and has been validated on individuals with ischemic and hemorrhagic stroke ranging from mild to severe disability. 26 The strength, activities of daily living (ADL)/instrumental ADL, mobility, and hand function domains were combined to form a Physical Function subscale, 27 which was used for analysis. Intervention All participants received conventional inpatient rehabilitation provided by the facility, including physical, occupational, and/or speech and language therapy. Therapy was available 5 days per week for 4 to 5 weeks where an individualized program was provided, focused on regaining maximal independence in mobility and daily activities. In general, rehabilitation consisted of 1 to 1.5 hours of physical therapy, 1 hour of occupational therapy, and 30 minutes of speech and language therapy. Physical therapy did not usually include a structured, progressive aerobic training program. The treating physical therapists were instructed not to alter their treatment plan for study participants. Control group participants received usual individualized treatment only. The Exercise group, matched to the Control group by age and sex, was provided with an additional program of individualized aerobic training on a semirecumbent cycle ergometer, 3 days per week for up to 30 minutes per session, led by a member of the research staff. Initial training intensity for this group was set by: (1) systolic blood pressure less than 220 mm Hg; (2) 50% to 75% of WR achieved at VO 2 peak; (3) rating of perceived exertion (RPE) between 4 and 6 out of 10. Although HR was recorded, it was not used to define training intensity because several participants were taking medications that affected HR response to exercise. The program was progressed by increasing training duration until 30 minutes was achieved with concurrent increases in WR within the participants tolerance, monitored by RPE. Statistical Analysis Although the 6MWT and the SIS were primary outcome measures in the study, sample size calculation was based on the SIS, given that some of our participants were unable to walk upon study entry. Using 2-tailed test, with type I error of 0.05 and power of 90%, a clinically significant difference in the SIS (ie, 15 points) would be detected with a minimum sample of 15 subjects. 26 To account for potential withdrawals, we recruited 18 participants into each group. All participants in the Exercise group were included in the feasibility analysis. Descriptive statistics were performed on all measures at both time points. Paired t tests were used to compare baseline participant demographics between groups. The 3 walking trials across the pressure-sensitive mat at each pace were averaged and used for analysis. A 2-factor repeated-measures analysis of variance was used to determine differences between groups and over time in the matched pairs groups using specific a priori contrasts. Percent change for continuous variables was calculated using the change score divided by the baseline score. Post hoc analyses of variance were conducted that included all participants in the study to confirm the findings of the original matched pairs design. Statistical Analysis Software Version 8 was used with a significance level of P <.05. Results Figure 1 depicts the flow of participants through the study. Of the 440 patient admissions screened for the study, 292 were ineligible for the study based on the study inclusion criteria. Most common reasons for exclusion included comprehension difficulties from language, cognitive, or behavioral issues (n = 57; 19.6%); cardiovascular (n = 34; 12.8%) or musculoskeletal (n = 25; 8.4%) comorbidity restricting exercise or pedaling ability; and very severe or mild leg impairment level (<3 or >6 CMSA) (n = 34; 12.8%). Of the remaining 148 who were eligible, 29 declined participation and from October 2005 to April 2006, 62 were eligible but not entered because recruitment during this period was limited to matching participant pairs within the groups. Thus, 57 were entered into the study with 32 allocated into the Exercise group and 25 into the Control group.

4 Tang et al / Effects of an Aerobic Exercise Program in Subacute Stroke 401 Figure 1 Flow of Participants Through the Study (September 2003 to April 2006) Screened for study n = 440 Eligible for study n =148 Entered into study n = 57 Ineligible for study n = 292 Declined participation or not matched for group allocation n = 91 Group Allocation Exercise Group n = 32 Control Group n = 25 Completed study n = 23 Discontinued n = 9 Reasons: acute low back pain, rapid recovery of leg function, abnormal blood pressure response during testing, withdrew self from study (n = 3), uncontrolled resting blood pressure, unable to maintain target pedaling cadence for exercise testing, anticipated short length of rehabilitation stay Completed study n = 22 Discontinued n = 3 Reasons: leg and back discomfort on cycle ergometer, abnormal electrocardiogram noted on testing, poor comprehension Feasibility of Exercise Training in the Subacute Poststroke Phase Figure 1 shows the allocation of participants. Of the 32 participants allocated to the Exercise group, 9 withdrew or were withdrawn before commencing, leaving 23 who participated in the training program. Note that no individuals withdrew from the study once the training commenced. Characteristics are presented in Table 1. At baseline, 8 (35%) participants did not require any aids for ambulation, 1 (4%) used a cane, 9 (39%) used a walker, and 4 (17%) were nonambulatory. The Exercise group participants completed 9.2 ± 0.7 (range, 5 13) training sessions, representing 90.5 ± 1.5% of scheduled sessions. The number of training sessions is primarily associated with the duration of stay as inpatients. No adverse effects were encountered. By discharge, participants increased from a mean training duration of 16.6 ± 6.3 minutes (range, 5 30 minutes) to 27.1 ± 0.9 minutes (range, minutes) and from an initial training intensity of 57.5 ± 11.7% (range, 43 80) of peak WR from the baseline exercise tests to 83.1 ± 5.0% (range, ).

5 402 Neurorehabilitation and Neural Repair Table 1 Characteristics of Participants in Exercise Group (n = 23) Effects of Exercise Training on Aerobic Capacity, Spatiotemporal Gait Characteristics, Functional Ambulation, and Health-Related Quality of Life To determine the program s effects, 18 Exercise group participants were matched to 18 Controls (11 men, 7 women), based on age and sex. Baseline characteristics are presented in Table 2, and Table 3 summarizes the assessment results. Due to equipment malfunction, gait assessments were not performed on 8 individuals at baseline and 4 participants at discharge. Both groups demonstrated improvements over time in most of the aerobic outcomes (time effect P <.05). There were no group time interaction effects but there were trends toward greater improvement in the Exercise compared with the Control group. Of note, peak VO 2 and WR among Exercisers increased 13% and 23%, respectively, compared with the Control group, who demonstrated improvements of 8% and 16%, respectively. There was a modest (5%) increase in peak HR in the Exercise group versus a larger 11% increase in the Control group. Eight individuals (22%, 4 in the Exercise group and 4 in the Control group) were nonambulatory at study entry and 2 individuals (6%, both in the Control group) were still unable to walk at discharge. For individuals who were able to perform the gait assessment, preferred walking speed increased on average 0.18 ± 0.17 m/s at the discharge assessment (time effect P =.0003), with no additional improvement in the Exercise group. Similarly, fast-paced gait speed improved on average 0.13 ± n Mean ± SE (range) Men/Women 12/11 Ischemic/Hemorrhagic/ Unknown stroke type 17/5/1 Right/Left hemisphere affected 12/11 Comorbidities Hypertension 16 Hyperlipidemia 7 Diabetes mellitus 2 Coronary artery disease 1 Chronic obstructive pulmonary disease 2 History of smoking 7 Anti-hypertensive medication use None/β-blockers/ACE inhibitors/both 7/6/9/1 Age, years 64.7 ± 3.6 (19 90) Time poststroke, days 17.8 ± 3.1 (6 62) Body Mass Index 26.3 ± 1.1 ( ) National Institutes of Health Stroke Scale score 4.7 ± 0.4 (2 11) Chedoke-McMaster Stroke Assessment Leg score 4.4 ± 0.2 (3 6) Functional Independence Measure score 84.9 ± 3.4 (57 113) Abbreviation: ACE, angiotensin-converting enzyme. Table 2 Baseline Participant Characteristics for 18 Matched Pairs a P Value Exercise Control for Group Group Group Differences Age, years 64.7 ± ± Days poststroke 19.1 ± ± Days in study 23.9 ± ± National Institutes of Health 4.9 ± ± Stroke Scale score CMSA Leg score 4.2 ± ± Functional Independence 84.0 ± ± Measure score a Values are mean ± SE. Abbreviation: CMSA, Chedoke-McMaster Stroke Assessment m/s in both groups by discharge (time effect P =.0002). As the symmetry data were not normally distributed, analysis was performed on the transformed natural logarithm. The Exercise group demonstrated a trend toward greater improvement in gait symmetry than Controls (8% reduction in symmetry ratio compared to 0.1%, group-time interaction P =.08). At the fast pace, the Exercise group was more asymmetric than the Control group (group effect P =.04), and both groups demonstrated a trend of improved fast-paced symmetry at discharge (time effect P =.09). On inspection of the individual data, it was determined that a high proportion (45%) of individuals fell within normal symmetry ranges on admission, and accordingly we would not have expected them to demonstrate improvements in symmetry after the intervention. To specifically determine the effects of cycle ergometry exercise on participants who initially presented with asymmetric gait, a post hoc analysis was conducted with the subgroup of Exercise (n = 5) and Control (n = 6) group participants with baseline rectified symmetry ratios greater than 1.1. The subanalysis confirmed the greater temporal asymmetry among the Exercise group participants compared to the Controls at both speeds (group effect preferred pace P =.01, fast pace P =.03) and also revealed that both subgroups improved over time (time effect preferred pace P =.04, fast pace P =.02). There was a trend toward a greater improvement in gait symmetry at the preferred pace in the asymmetric subgroup of Exercise participants (13.9%) compared with the Control group subgroup (3.6%) (interaction effect, P =.14). There was no difference in the degree of improvement between the subgroups at fast pace (interaction effect, P =.82). A trend toward greater benefit in functional ambulation was also observed in the Exercise group compared with the Controls. Distance walked on the 6MWT increased by 53% among the Exercisers, versus 23% improvement in the Control group (group-time interaction P =.23). The SIS Physical subscale score improved similarly in both groups by discharge. Subsequent post hoc analyses of variance conducted that included all participants who completed the study (n = 23 in

6 Tang et al / Effects of an Aerobic Exercise Program in Subacute Stroke 403 Table 3 Pretreatment and Posttreatment Scores for 18 Matched Pairs a Exercise Group Control Group P Value Pretreatment Posttreatment % Change Pretreatment Posttreatment % Change Time Group b Time Aerobic capacity VO 2 peak, ml kg 1 min ± ± ± ± ± ± Peak WR, watts 46.4 ± ± ± ± ± ± 7.1 < Peak HR, beats/min ± ± ± ± ± ± VeT, ml kg 1 min ± ± ± ± ± ± Gait assessment Preferred pace Gait speed, meters/second 0.68 ± ± ± ± ± ± 12.1 < Gait symmetry (n = 20) 1.31 ± ± ± ± ± ± Gait symmetry (n = 11) b 1.53 ± ± ± ± ± ± Fast pace Gait speed, meters/second 0.99 ± ± ± ± ± ± 5.3 < Gait symmetry (n = 20) 1.24 ± ± ± ± ± ± Gait symmetry (n = 11) b 1.35 ± ± ± ± ± ± Functional ambulation 6MWT distance, meters ± ± ± ± ± ± 5.6 < (% predicted) (50.0%) (65.3%) (52.7%) (63.9%) Health-related quality of life SIS Physical subscale 47.3 ± ± ± ± ± ± 6.5 < Abbreviations: WR, work rate; HR, heart rate; VeT, ventilatory threshold; 6MWT, 6-Minute Walk Test; SIS, Stroke Impact Scale. a Values are mean ± SE. b Post hoc gait symmetry analysis using participants with asymmetric gait at study entry only. the Exercise group, n = 22 in the Control group) confirmed the findings of the original matched pairs design. There were no baseline differences between the groups when all participants were included, nor were there any changes to the training effects on the measures of interest. Discussion The results from this study contribute to the small body of research on aerobic training in the subacute phase poststroke. The work further reinforces the feasibility of implementing such a program in an inpatient stroke rehabilitation setting. In addition to examining the addition of an exercise program to inpatient rehabilitation and investigating the training effects, 7,8 a more detailed evaluation of gait outcomes to infer neuromotor control of walking was performed. Aerobic capacity, preferred and fast walking speed, functional ambulation, and health-related quality of life improved in both groups. There was, however, a trend toward greater gains in aerobic capacity and 6MWT distance in the Exercise group. Feasibility of Early Aerobic Training After Stroke The early and significant deconditioning that occurs after stroke 1,2 underscores the importance of establishing effective interventions to minimize the impact on long-term stroke recovery. Further to previous reports that support adding aerobic training to inpatient rehabilitation soon after stroke, 7,8 the current study expanded the inclusion criteria to include individuals admitted to inpatient stroke rehabilitation who also represent a broader range of abilities. Semirecumbent cycle ergometry was selected as the training modality due to anticipated poststroke postural and lower extremity dyscontrol that would limit treadmill or overground walking previously studied, 8 thus allowing those with moderate disability to participate. Indeed, 20% of the Exercise group participants were nonambulatory at study entry and even those capable of ambulation may have required safety supervision. While body weight supported treadmill training has been used effectively among individuals with subacute stroke, 8 there is limited access to such equipment in most clinical settings and ergometry has the added benefit of also allowing the feet to be affixed to the pedals, accommodating for impairments in leg motor control. Because the mechanical coupling of a traditional cycle ergometer does permit greater participation of the nonparetic leg during pedaling, ergometry training paradigms designed to encourage the use of the paretic leg may be of benefit to overcome this compensation. 16 Ongoing work is focused on examining other encouraged-use paradigms using a variety of training modalities. All Exercise participants who performed initial maximal exercise testing and started aerobic training were able to complete an average of an additional 27 minutes of exercise training 3 times per week beyond conventional stroke rehabilitation without adverse effects. Over 90% of scheduled training sessions were attended. While the influence of these additional sessions on the usual care was not specifically evaluated, there were no accounts of negative impact on therapy reported by the participants or treating therapists.

7 404 Neurorehabilitation and Neural Repair With respect to generalizability of the results, it is important to note that 148 of the 440 admissions screened for study entry were deemed eligible. Many were excluded based on comprehension difficulties and cardiovascular or musculoskeletal comorbidities. We acknowledge that such issues are common among stroke survivors and may limit the study s external validity. To increase the applicability to the broader range of stroke survivors, program adaptations may be implemented, such as vigorous screening and monitoring for complex cardiovascular comorbidities, using interpreter services, or incorporating supported communication or cognitive behavioral strategies. Furthermore, because the exercise intervention was limited to length of stay on the stroke rehabilitation unit, those with anticipated short stays (2 weeks or less) but would otherwise have been eligible were not recruited into the study. Finally, because we recruited from a rehabilitation facility, the results are most applicable to this cohort of stroke survivors and not necessarily to those that do not receive rehabilitation. Effects of Training Improvements in aerobic capacity were observed in both groups, demonstrating that aerobic fitness is modifiable early after stroke. Previous studies that have implemented aerobic training in the subacute stroke phase have reported trainingrelated benefits including increased VO 2 peak with supported treadmill ambulatory training, 8 and higher peak WR and exercise time and lower resting HR with cycle ergometry training. 9 Of note, however, is the potentially more impaired cohort of stroke survivors studied in the current trial evidenced by the proportion of nonambulatory Exercise participants. In comparison, the work of Katz-Leurer and colleagues 8 appeared to include participants at higher levels of walking ability with 40% capable of outdoor ambulation. Despite the stroke-related limitations in functional mobility in our sample, we found a trend toward greater peak VO 2 and WR in the Exercise group compared to the Controls, differences that suggest the potential effectiveness of early exercise intervention in reducing impact of compromised fitness after stroke, even among those with moderate impairment. While the improvements in aerobic capacity were not significantly different between the groups, this may be due in part to the short training duration afforded by the length of stay in inpatient rehabilitation. Implementing a longer period of training by continuing beyond inpatient rehabilitation would likely contribute to greater benefit among participants who receive the additional exercise training. Comparable improvements were observed in both groups in preferred- and fast-paced gait speed, with no added effect of the exercise training. Of interest, however, was the trend toward greater improvement in symmetry ratio at the preferred walking pace among the Exercise participants, an outcome afforded through the detailed assessment of walking competency. One confound to the present group differences was the baseline differences in gait symmetry. One could expect a more modest improvement in gait symmetry among the control group due to lower average severity of gait symmetry. Even with the partici pants with symmetric gait removed from the pool used for analysis, there was a trend toward greater improvement posttraining in preferred pace symmetry in the Exercise group. Nonetheless, with the Exercise group being asymmetric at baseline compared to the generally symmetric Control group, the results may provide preliminary support that an additional early benefit to short-term aerobic ergometry training may be improved lower limb dys-coordination during walking in individuals with already compromised neuromotor control. Energy costs of hemiparetic gait have been reported to be 1.5 to 2 times higher than that of normal walking. 6 As such, improve ments in gait symmetry may reduce these inefficiencies and lower the related physiological costs. Additionally, the trend toward greater improvement in 6MWT performance among the Exercise group is noteworthy. With a mean improvement of 127 meters walked, this distance was 40 meters more than the Control group and greater than the 54 meters cited as the minimum change in walking distance to be considered clinically important. 28 While this value is cited from the chronic respiratory disease literature and should be interpreted with caution for stroke, it nevertheless represents a greater relative change for a population with more significant limitations in walking ability. A limitation to the study was the lack of control of therapy time and aerobic exercise provided through conventional rehabilitation. The extent of aerobic exercise that was incorporated into the participants regular stroke rehabilitation program is not known; the gains observed in our Control group may be attributed to this. The modest effect size is most likely a product of the relatively short training duration. It is our view that such gains seen with early exercise, when augmented with a necessary continuation of training after discharge, will translate to important and significant changes in aerobic and functional capacity poststroke. As institution-based rehabilitation programs move toward shorter lengths of stay, the potential benefits of aerobic training programs will be limited if they are restricted to inpatient rehabilitation duration only. The subacute phase poststroke is considered important not only to minimize the extent of cardiorespiratory deconditioning, but also to exploit the window of opportunity for neuromotor recovery and early behavioral change. Starting training as early as possible is an essential element to maximizing stroke recovery and helping to better ensure involvement in outpatient and/or community programs. Continued training through structured outpatient or communitybased programs that parallel opportunities provided for individuals with cardiac disease 29 must be considered. Furthermore, exploring novel strategies using a seated model to accommodate for a broader range of abilities that simultaneously address aerobic training and encourage paretic limb 16 use may lead to additional improvements in gait outcomes and cardiorespiratory fitness, potentially translating to greater functional benefit. Implications Results from this study have important implications for shaping practice to aid in stroke recovery. Low baseline VO 2 peak

8 Tang et al / Effects of an Aerobic Exercise Program in Subacute Stroke 405 values are comparable to earlier reports 1 and emphasize the important concern that the exercise capacity among stroke survivors is severely compromised even in the very early phase. Despite the observed improvements in aerobic capacity and gait symmetry observed in both groups, there appears to be additional cardiorespiratory and neuromotor benefit with early cycle ergometer training poststroke. The results from this study may be used to inform care provided through conventional stroke rehabilitation programs, supporting the routine implementation of aerobic exercise to improve functional ambulation. We anticipate that engaging in early exercise training after stroke and continuing this for longer duration beyond inpatient rehabilitation not only has further potential benefits for improving hypertension and cholesterol levels thereby lowering recurrent stroke risk, 30 but ultimately may also translate to improved health-related quality of life. In summary, early aerobic training program can be safely implemented among moderately impaired individuals in the subacute phase after stroke without negative effects on participation in conventional stroke rehabilitation. Medical prescreening and exercise testing are recommended for safety and for determining appropriate training intensities. Despite limitations in the training duration, the trend toward greater improvement observed in the Exercise group in aerobic capacity, 6MWT distance, and some modest evidence of potential benefits to gait symmetry are important findings suggesting that early aerobic training can begin to have a beneficial effect on aerobic capacity, neuromotor control, and functional ambulation. Acknowledgments The study was conducted at the Toronto Rehabilitation Institute, Toronto, Ontario, Canada. Preliminary results from this work have been presented at the following conferences: (1) World Congress of Physical Therapy, Vancouver, May 2007 Early exercise intervention after stroke: influence on aerobic and functional capacity A Tang, KM Sibley, SG Thomas, MT Bayley, WE McIlroy, D Brooks; (2) 6th World Congress on Aging and Physical Activity, August 2004 Early Exercise Intervention After Stroke: Influence on Aerobic and Functional Capacity A Pilot Study A Tang, KM Sibley, D Brooks, SG Thomas, WE McIlroy; and (3) 6th World Congress on Aging and Physical Activity, August 2004 Early Exercise Intervention After Stroke: Influences on Muscle Activation Patterns and Implications for Gait KM Sibley, A Tang, D Brooks, SG Thomas, and WE McIlroy. None of the authors have a conflict of interest related to the publication of this manuscript. The study was supported by the Canadian Institutes of Health Research (CIHR) Grant # MOP WEM was a Canada Research Chair, DB held a CIHR New Investigator Award, AT was supported by the Government of Ontario / Heart and Stroke Foundation of Ontario, the Toronto Rehabilitation Institute and the Physiotherapy Foundation of Canada, KMS is supported by the Natural Sciences and Engineering Research Council of Canada. We acknowledge the support of Toronto Rehabilitation Institute who receives funding under the Provincial Rehabilitation Research Program from the Ministry of Health and Long Term Care in Ontario. We thank the following people for their help and support: A. Cheng, MD; L. Biasin, PT; J. Komar, PT; and J. Lymburner, PT. References 1. MacKay-Lyons MJ, Makrides L. Exercise capacity early after stroke. Arch Phys Med Rehabil. 2002;83: MacKay-Lyons MJ, Makrides L. Longitudinal changes in exercise capacity after stroke. Arch Phys Med Rehabil. 2004;85: Macko RF, Smith GV, Dobrovolny CL, Sorkin JD, Goldberg AP, Silver KH. Treadmill training improves fitness reserve in chronic stroke patients. Arch Phys Med Rehabil. 2001;82: Potempa K, Braun LT, Tinknell T, Popovich J. Benefits of aerobic exercise after stroke. Sports Med. 1996;21: Gerston J, Orr W. External work of walking in hemiparetic patients. Scand J Rehabil Med. 1971;3: Fredrickson E, Ruff R, Daly JJ. Physiological cost index as a proxy measure for the oxygen cost of gait in stroke patients. Neurorehabil Neural Repair. 2007;21: Pang MY, Eng JJ, Dawson AS, Gylfadottir S. The use of aerobic exercise training in improving aerobic capacity in individuals with stroke: a metaanalysis. Clin Rehabil. 2006;20: da Cunha-Filho IT, Lim PA, Quereshy H, Henson H, Monga TN, Protas EJ. A comparison of regular rehabilitation and regular rehabilitation with supported treadmill ambulation training for acute stroke patients. J Rehabil Res Dev. 2001;38: Katz-Leurer M, Shochina M, Carmeli E, Friedlander Y. The influence of early aerobic training on the functional capacity in patients with cerebrovascular accident at the subacute stage. Arch Phys Med Rehabil. 2003;84: MacKay-Lyons MJ, Makrides L. Cardiovascular stress during a contemporary stroke rehabilitation program: is the intensity adequate to induce a training effect? Arch Phys Med Rehabil. 2002;83: Kautz SA, Duncan PW, Perera S, Neptune RR, Studenski SA. Coordination of hemiparetic locomotion after stroke rehabilitation. Neurorehabil Neural Repair. 2005;19: Duncan PW, Studenski S, Richards L, et al. Randomized clinical trial of therapeutic exercise in subacute stroke. Stroke. 2003;34: Wade DT, Wood VA, Heller A, Maggs J, Langton Hewer R. Walking after stroke: measurement and recovery over the first 3 months. Scand J Rehabil Med. 1987;19: Raasch CC, Zajac FE. Locomotor strategy for pedaling: muscle groups and biomechanical functions. J Neurophysiol. 1999;82: Tang A, Sibley KM, Thomas SG, McIlroy WE, Brooks D. Maximal exercise test results in sub-acute stroke. Arch Phys Med Rehabil. 2006;87: Sibley KM, Tang A, Brown DA, Brooks D, McIlroy WE. Feasibility of adapted aerobic cycle ergometry tasks to encourage paretic limb use after stroke: a case series. J Neurol Phys Ther. 2008;32: Gowland C, Van Hullenaar S, Torresin WD, et al. Chedoke-McMaster Stroke Assessment: Development, Validation, and Administration Manual. Hamilton: Chedoke-McMaster Hospitals and McMaster University; American College of Sports Medicine. ACSM s Guidelines for Exercise Testing and Prescription. 7th ed. Philadelphia: Lippincott Williams & Wilkins; Bruce R, Kusumi F, Hosmer D. Maximal oxygen intake and nomographic assessment of functional aerobic impairment in cardiovascular disease. Am Heart J. 1973;85: Brott T, Adams HP, Olinger CP, et al. Measurements of acute cerebral infarction: a clinical examination scale. Stroke. 1989;20: Keith RA, Granger CV, Hamilton BB, Sherwin FS. The Functional Independence Measure: A New Tool for Rehabilitation. Vol. 1. New York: Springer-Verlag; 1987.

9 406 Neurorehabilitation and Neural Repair 22. American Thoracic Society. ATS statement: guidelines for the six-minute walk test. Am J Respir Crit Care Med. 2002;166: Enright PL, Sherrill DL. Reference equations for the six-minute walk in healthy adults. Am J Respir Crit Care Med. 1998;158: Solway S, Brooks D, Lacasse Y, Thomas SG. A qualitative systematic review of the measurement properties of functional walk tests used in the cardiorespiratory domain. Chest. 2001;119: Lai SM, Studenski S, Duncan PW, Perera S. Persisting consequences of stroke measured by the Stroke Impact Scale. Stroke. 2002;33: Finch LE, Brooks D, Stratford P, Mayo NE. Physical Rehabilitation Outcome Measures. 2nd ed. Hamilton: Canadian Physiotherapy Association; Duncan PW, Wallace D, Lai SM, Johnson D, Embretson S, Laster LJ. The Stroke Impact Scale Version 2.0: evaluation of reliability, validity and sensitivity to change. Stroke. 1999;30: Redelmeier DA, Bayoumi AM, Goldstein RS, Guyatt GH. Interpreting small differences in functional status: the Six Minute Walk test in chronic lung disease patients. Am J Respir Crit Care Med. 1997;155: Canadian Association of Cardiac Rehabilitation. Canadian Guidelines for Cardiac Rehabilitation and Cardiovascular Disease Prevention. Winnipeg: Canadian Association of Cardiac Rehabilitation; Lee CD, Folsom AR, Blair SN. Physical activity and stroke risk. Stroke. 2003;34: For reprints and permission queries, please visit SAGE s Web site at