| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Research Reports |
JR Carey, PT, PhD, is Professor, Program in Physical Therapy, University of Minnesota, 420 Church St SE, Minneapolis, MN 55455 (USA)
Address all correspondence to Dr Carey at: carey007{at}umn.edu
I am pleased to offer comments on the article by Sullivan et al,1 which represents an important contribution to the scientific literature addressing recovery of walking ability in people with chronic stroke. Its particular value is its scientific approach to a practical problem—how to maximize recovery of walking ability while being cost conscious. The authors explored 2 questions: (1) whether a resisted cycling program that simulates some of the motoric demands of walking can be as effective as the more costly body-weight–supported treadmill training (BWSTT) and (2) whether 2 combined interventions can result in higher outcomes than 1 intervention combined with a sham treatment.
Commendably, the authors created 4 treatment groups that represent common physical therapist practice and used a power analysis to derive an appropriate number of subjects for each group. Because of the notoriously difficult task of recruiting a sufficient number of people with stroke who meet designated criteria, they used 3 different clinics. This not only expanded their subject recruitment base, it also contributed to higher generalizability of their findings. They used a blocked randomization procedure to ensure an equal distribution of subjects to each group and, more importantly, to promote homogeneity of subjects across groups with respect to severity of walking impairment. Furthermore, because differences in severity strongly influence treatment outcomes, they used an analysis of covariance with severity of walking impairment as the covariate to account for any remaining differences in this factor across groups. Although the therapists administering the treatments varied across these clinics, as did the blinded testers, consistency among clinics was reasonably ensured through standardized training and peer review.
As with most research pursuing practical questions, there were unanticipated variations in the protocol, and—to the authors' credit—these cases were analyzed by an independent group to determine whether the involved subjects should be retained in the full analysis. Also common to research involving intensive interventions in patient populations, there were adverse events. The authors listed these events and reported the rulings from an independent group as to whether it was the research that caused the event. In as comprehensive a manner as possible, the authors analyzed the results for all subjects who completed the study, and, additionally, they conducted an intention-to-treat analysis of all randomized subjects.
The chief finding was that the resisted cycling training, despite its demands on muscles needed for walking, was not as effective as BWSTT in improving walking speed. The authors explained this outcome as evidence of the importance of combining specificity of training with intensity of training. However, a confounding factor exists in that subjects who received BWSTT also received gait instruction, presumably after each BWSTT treatment, in an over-ground setting over a 15-m distance to reinforce the training presented on the treadmill. As the cycling group did not receive such instruction, it is not entirely certain whether the resultant differences in walking speed among groups were due to the treadmill training or to the over-ground instruction or to the combination of the 2 interventions. The authors acknowledged this concern but contended that such "low-intensity activity" could not likely account for the observed magnitude of differences among groups. Perhaps so, but this remains unproven.
More importantly, this study stimulates questions for future studies on the mechanisms for observed behavioral improvements. Is the biologic substrate supporting improved walking speed at the level of the muscle with treatment-induced conversion of fiber type or enlargement of cross-sectional area?2 Or is it at the level of the spinal cord with facilitation of alpha-motoneurons and tuning of a pattern generator?3 Or is it at the level of the motor cortex with heightened excitability of surviving neurons that prior to training were too down-regulated to be recruited voluntarily?4,5 And how do any such neural changes as these correlate to behavioral changes? These are important questions to answer so that the most effective treatment regimens can be used. Fortunately, technology is advancing rapidly so that pursuing these questions is within grasp. For example, whereas locomotion studies have not been amenable to probing of brain organization with functional magnetic resonance imaging, studies with brain probing in people with stroke are now accumulating with the use of near-infrared spectroscopy.6–9
As a provocation related to mechanism in the present study, the possibility exists that the improvement in walking speed had nothing to do with the task-specific nature of the treadmill training. Instead, it could have been due to exercise-induced trophic factors leading to improved excitability in neurons that have survived the stroke but fallen dormant from disuse. It is known that neurotrophins, including brain-derived neurotrophic factor (BDNF), have a potent effect on synaptic activity and neural excitability (for a review, see Vaynman and Gomez-Pinilla10). Kim et al11 showed in rats with an induced stroke that 12 days of treadmill training led to significant increases in BDNF and to improvements in both motor and sensory skills not specific to the treadmill task. In humans who are healthy, Rojas Vega et al12 and Ferris et al13 demonstrated that intense lower-extremity ergometry led to significant increases in serum BDNF concentrations. Winter et al14 found the same increases with intense sprinting. The provocative point is that intense, repetitive physical activity mobilizes BDNF and that it may be this mobilization, and not so much task-specific versus task-nonspecific training, that is the critical factor leading to neuroplastic changes subserving improved function.
It is for this reason that I must question the use of upper-extremity ergometry by Sullivan et al as a sham treatment. Their sham treatment involved 10 sets of 20 revolutions of a hand cycle with the resistance of the crank adjusted so that subjects could not complete any more than 20 revolutions per set. Whereas they identified other studies15,16 that used upper-extremity exercises as a sham treatment for walking outcomes, the exercises described in these studies (eg, writing, manipulating cards, grasp and release of objects) seem to be much less exertive than the ergometry used in the present study.
Related to their second question, the authors found that BWSTT combined with their "sham" was more effective in increasing lower-extremity strength than either BWSTT combined with cycling or BWSTT combined with lower-extremity resistive exercises. They explained the difference in strength outcomes as a deleterious effect in the 2 latter groups consequent to overtraining of muscles, and they presented a reasoned defense of this supposition. Still, given the effect of intensive, repetitive physical activity on neurotrophins, and, in turn, the effect of neurotrophins on central activation, the possibility of an additive effect in the group that received BWSTT combined with upper-extremity ergometry also should be considered. Admittedly, this argument is weakened by the commonality of a resistance element in all 3 training regimens such that all groups could be expected to show additive effects and not just the upper-extremity ergometry group. Perhaps the rate at which this group performed their resistive exercise was more demanding compared with the other groups and this higher intensity resulted in greater BDNF mobilization, as was found by Winter et al14 in comparing high-impact anaerobic sprinting with low-impact aerobic running in athletes.
The article by Sullivan et al sets a high model for other investigators to follow in the unending pursuit of the most effective combinations of treatment to maximize recovery following stroke. As a final remark, the uncommon combination of physical training with cognitive training deserves special study. Studies17–19 have shown that spatial learning and complex learning in rats, as adjuncts to physical training, have resulted in higher levels of neurotrophins in the brain compared with rats receiving an equal amount of physical training alone without learning. Thus, simultaneous with BWSTT, the new challenge would be to create task variations on such factors as speed, friction, compressibility, perturbations, and so forth of the treadmill belt. The resultant combination would involve not only "forced use" of paretic muscles but also "forced processing" of the cognitive system to manage the variations. Collaboration with equipment manufacturers would be needed to generate some of these changes. Similarly, variations with other systems (visual, auditory, vestibular) also could be manipulated to actually achieve a desired but reasonable "contextual interference"20 that may actually hinder performance in the short run but optimize recovery in the long run as the brain reorganizes.
 | References |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
