The remarkable decline in stroke mortality in

Comments, Opinions, and Reviews 1081 A Review of Stroke Rehabilitation and Physiotherapy Edzard Ernst, MD, PhD Most of the members of the therapeutic team in stroke rehabilitation take the effectiveness of physical treatments after stroke for granted. Yet, published data show that the evidence is not so straightforward or easy to evaluate. The majority of the hard evidence, however, does imply that stroke patients benefit from rehabilitation with physiotherapy. This benefit may be statistically small, but for a given individual, it could mean the difference between living at home or in an institution. Few studies address the question of the optimal physiotherapy in stroke rehabilitation. The evidence available today suggests that it does not matter which form of treatment is chosen and that any of the available approaches will improve the patient's functional status. In other words, if an optimal treatment exists, we have, so far, failed to identify it. Until further evidence emerges, we should therefore select therapies that are most cost-effective and that can be given to the largest number of patients. Well-planned clinical trials aimed at finding the best approach and discriminating potential responders from nonresponders are urgently needed. {Stroke 1990;21:1081-1085) The remarkable decline in stroke mortality in some countries during the last decades is predominantly due to a decrease in incidence and only to a lesser extent to an increase in survival rates. 1 Thus, the decline may be seen as a triumph of prevention. 2 At the same time, acute stroke therapy has failed to keep pace, 3 and the role of rehabilitation in reducing stroke fatality is not clear at all. There are many uncertainties surrounding stroke rehabilitation. In particular two key questions need to be asked: Is stroke rehabilitation effective? and Which is the optimal approach? Evidence Concerning the Effectiveness of Rehabilitation A number of uncontrolled studies have suggested that rehabilitation programs lead to an improvement in functional status that cannot be attributed merely to spontaneous recovery. 4-5 The introduction of stroke units brought about a significantly higher hospital discharge rate than historical controls. 67 Similarly, the number of secondary complications after stroke is reported to be lowered by admission to a specialized stroke unit. 8 Several groups have conducted controlled trials comparing the outcome of patients from stroke units From the Department of Physical Medicine and Rehabilitation, University of Vienna, AKH. Address for correspondence: E. Ernst, Department of Physical Medicine and Rehabilitation, University of Vienna, AKH, 1090 Vienna, Austria. Received October 23, 1989; accepted March 14, 1990. with those from medical wards. The first such study 9 reported no benefit for mildly and severely affected patients but a significantly better outcome for patients with moderate disabilities. Feigenson and colleagues 10 showed in a similar trial that patients admitted to a stroke unit were more likely to be discharged home and were able to walk better. Garraway et al 11-13 performed a randomized trial comparing 152 hemiplegic, conscious stroke patients from a medical ward to 155 similar patients from a stroke unit, who also received significantly more physical therapy. Sixty days after hospital admission, 50% of the stroke unit patients but only 32% of the medical ward patients were functionally independent (/j<0.01). n In the next year, however, this difference was lost. 12 Since the stay in the stroke unit averaged 55 days and that in the medical ward averaged 75 days, the authors conclude that longer rehabilitation is mandatory for sustained benefit. In a more detailed analysis of the same trial, 13 the authors also conclude that an early start of rehabilitation is more important than the absolute amount of physiotherapy administered. They have no doubt that the coordinated team approach used in a stroke unit is superior to that in a medical ward. Unfortunately, this study (like the majority of other trials) was not a blind one. Strikingly similar results were reported from Sweden. 14 In a nonrandomized trial, 60 stroke patients were submitted to individually adapted programs of early activation by a specially trained nursing staff; 52 stroke patients from other wards served as controls. According to an activity index of functional capacity,

1082 Stroke Vol 21, 7, July 1990 there was significant improvement only in the experimental group within 4 weeks after the acute event. After this period, both groups were treated conventionally. The effect of early activation was then lost during the following year. 15 Smith and coworkers 16 randomized 133 stroke patients who were fit enough to go through an intensive rehabilitation into three equally sized groups, an intensive rehabilitation group (4 entire days per week), a conventional rehabilitation group (3 half days per week), and a group that underwent no rehabilitation (patients were told only to carry on exercising as previously taught). In the first two groups, treatment was performed for up to 6 months. The functional outcome (index of daily life activities) was best in the first group. rehabilitation was less beneficial, and patients with no supervised rehabilitation deteriorated. Maximum benefit was obtained at 3 months, and continuation of the program merely maintained this success without extending it. This trial seems particularly relevant since it refutes Lind's view 17 that spontaneous recovery, rather than rehabilitation, accounts for most of the improvement in functional ability. Sivenius et al 18 compared 40 patients treated conventionally with 45 others who were rehabilitated intensively with physiotherapy twice per day until no more functional improvement was observed. The intensively treated patients showed significantly better functional recovery, but death and institutionalization rates were not different. As in the previous trial, recovery occurred predominantly during the first 3 months. Furthermore, there were several disadvantages: the sample size was small, the study was not a blind one, and there was no randomization. Strand et al 19 compared 110 acute stroke patients treated in a stroke unit with 183 patients from a medical ward. Mortality rates were not different, but significantly fewer stroke unit patients, both at 3 and 12 months, remained hospitalized, and more were able to perform important daily life activities. Later, the authors analyzed the same data and attempted to determine which patients would benefit and which would not benefit from a rehabilitation program. 20 They found that practically all stroke survivors were helped by such a program, but it is also possible that they were merely unable to identify which subgroup responded less. Other evidence has brought into question the effectiveness of rehabilitation programs. Three randomized controlled trials evaluating the usefulness of stroke rehabilitation were inconclusive, 21-23 mainly due to the small sample sizes. Another group 24 reported that physiotherapy did not lead to functional recovery nor prevent complications after stroke. A randomized controlled trial 25 showed that traditional care, consisting of physical therapy only on request of the physician, resulted in survival rates similar to those of team care based on routine physical, occupational, and speech therapy. For motor functions, males performed better when on team care, whereas females benefited more from traditional care. In terms of functional abilities, males receiving team care fared better, whereas women experienced no differences between the two treatments. Finally, one report 26 implied that a community-based program to coordinate inpatient and outpatient rehabilitation was not effective in terms of functional outcome (Barthel index). Unfortunately, controls were not matched, allowing for considerable bias in this particular study. It is difficult to come to firm conclusions based on these data 27 because study protocols differ widely among trials and stroke units vary from specialized acute care units 1920 to nonacute rehabilitation centers. 10 Moreover, selection criteria for patient admittance are not uniform, and duration, type, and onset of rehabilitation differ. Finally, measurements of outcome are not constant. 28 Yet, after critical evaluation, particularly of the randomized trials (Table 1), stroke rehabilitation does seem to have its value. Spontaneous recovery can be impressive, but rehabilitation-induced recovery seems to be greater on average. The benefit may be statistically small, but it is medically relevant since, for a given patient, it could mean the difference between living at home or being institutionalized. Rehabilitation should probably begin as soon as possible after the acute event. Whether highly intensive treatment is better than moderately intensive therapy is debatable. Even though the most marked improvement is achieved during the first 3 months, 29 rehabilitation should be continued for a longer period to prevent subsequent deterioration. Clearly, the evidence is not as convincing as one would wish, and further meticulously planned studies are badly needed. Evidence for Optimal Physiotherapy Medical doctors are often stunned by the range of choices of physiotherapeutical approaches; unacquainted with the techniques of researchers such as, Brunnstrom, Clayton, Coulter, Fay, Kabat, Knott, Rood, and Voss; and unfamiliar with the theories behind these techniques. 30 At closer inspection, the conclusions drawn by these sometimes competing schools of thought lack sound scientific proof, and published data that indicate optimal physiotherapy is limited. For the purpose of this communication, physiotherapy is defined as any physical treatment including therapeutic and ergotherapy. Stern et al 31 treated two groups of 31 stroke patients (partly randomized) with either proprioceptive neuromuscular facilitation or conventional physiotherapy. Patients were admitted into the trial up to 13 days after the acute event. There was no better recovery of mobility, strength, or function in the facilitation group than in the other populations. However, the sample size was small, the study was not blind, and rehabilitation started perhaps too late. Two earlier studies 32-33 had addressed similar questions but are not detailed here because of certain deficiencies. One study 32 used now-obsolete thera-

Ernst Physiotherapy After Stroke 1083 TABLE 1. Reference Garraway et al 11 ' 12 (1980) Randomized Trials on the Efficacy of Stroke Rehabilitation Therapy Mean age Blind Start of Duration of group n (yr) evaluation rehabilitation rehabilitation Outcome Stroke unit* 155 72.5 later than 3 55 days 60 days after stroke, 50% of days after stroke stroke unit patients and 32% of medical ward patients were independent. Difference was lost at 1-year follow-up. Medical ward 152 75 days Smith et al 16 (1981) Intensive caret care Self care 133 63 66 65 On discharge from acute care ADL improvement was greatest in intensive care group and intermediate in conventional group. Deterioration was observed in self-care group. Strand et al 1 '. 20 * (1985) Stroke unit Medical ward 110 183 72 73 later than 1 week after stroke 21 days 31 days 15% of stroke unit patients and 39% of medical ward patients remained hospitalized 3 months after stroke. Wood-Dauphinee et al 25 (1984) Traditional care 130 74.9 later than 1 week after stroke 5 weeks Team care 74.4 I 5 weeks Survival was similar in both groups. In motor performance, males were better with team care, and females were better with traditional care; in functional abilities, males were better with team care, and females fared the same in both groups. n is the number of patients in each therapy group. If a single value (or statement) is given for a particular study, this represents the therapy groups of that study combined. ADL, activities of daily living. 'Operated according to Isaacs. toutpatient treatments (essential details are missing). tthis study was only quasi-randomized (allocation by available space in stroke unit). Team work of specialists, early start, family education. Evaluation not strictly blind but by a noninvolved third party. pies, and the other study 3? used a facilitation group of only four patients, unmatched controls, and no randomization. Loggigian et al 34 performed a study in which 42 stroke patients (stroke history was <7 weeks) were randomly allocated to "traditional" or 's physiotherapy. Neither the Barthel index nor the Kendall muscle test showed group differences, but both groups improved in the same way. Again, the same drawbacks were present: the sample size was small, the start of rehabilitation was late, and the study was not a blind one. Lord and Hall 35 matched two groups of stroke patients (n=19 and 20) from two different centers. One center focused on "traditional" techniques (average stroke history was 9 days); the other focused on facilitation techniques (average stroke history at entry was 26 days). There was a considerable overlap concerning the treatments administered. Hospitalization was 29 days in one center and 68 days in the other. A validated telephone questionnaire on daily life activities 8 months after the acute event showed a slight, but clinically meaningless, advantage in self-feeding ability in the facilitation group. Other functions were not significantly different. This study suffers from the disadvantages of the above trials plus considerable intergroup inequality, as outlined above. Smedley et al 36 compared two groups of 25 patients each from two hospitals in a nonrandomized trial. Both groups were treated traditionally for 1 month, but the experimental group had to play with a modified "slot machine" to enhance motor function, strength, and coordination of the upper limb. All these variables improved significantly better in the experimental group. Yet this study is poorly controlled, and decisive details on treatment modalities and patient characteristics are missing. Dickstein and coworkers 37 performed a large quasirandomized (/i=131) study on three different physiotherapies: conventional, facilitation, and. The patients admitted to rehabilitation 16 days after the acute event were treated for. In the early phase, conventionally treated patients fared better in terms of ambulation. At the end of the trial, there were no significant differences among the groups with regard to functional recovery. In this study, rehabilitation was started late and was of short duration.

1084 Stroke Vol 21, 7, July 1990 TABLE 2. Controlled Trials of One Rehabilitation Approach Versus Another Reference Basmajian et al 38 (1987) Therapy group EMG feedback n 29 Mean age (yr) 62 Blind evaluation Randomized Start of rehabilitation later than 12 months after stroke Duration of rehabilitation Outcome 5 weeks Multiple upper limb functions showing no differences between groups 5 weeks Dickstein et al 37 PNF 131 70.5 16 days after stroke (on average) advantage of one group over another (Barthel index) Loggigian et al 34 (1983) Lord and Hall 35 Smedley et al 36 Stern et al 31 (1970) and Rood "Slot machine" plus therapeutic Therapeutic PNF 42 19 20 25 25 31 31 61.6 64.4 63.5 Within 7 weeks after stroke Within 8 months of stroke differences in improvement of Barthel index meaningful differences according to telephone questionnaire Motor function better with "slot machine" therapy differences in motor or functional improvement EMG, electromyographic; PNF, proprioceptive neuromuscular facilitation;, data not available. If a single value (or statement) is given for a particular study, this represents the therapy groups of that study combined. *Depending on administrative procedures. Basmajian et al 38 tested two physical therapies on hemiparetic stroke patients. Twenty-nine patients were randomly assigned to either treatment or a "behavioral cognitive physical therapy" consisting mainly of a specially designed electromyographic biofeedback program for the upper limb. The assessment of arm function was by blind evaluation. After 9 months, both therapies showed significant improvements, but there were no differences between the two groups. According to other reports 39 (most of them not randomized), biofeedback seems a promising approach in stroke rehabilitation. It is particularly favorable when initiated early in only mildly affected patients. 3840-42 This therapy is not costly or work intensive and can be performed on outpatients; it deserves further study to define its role in stroke rehabilitation. With the exception of one study, 36 which is even more problematic in design and execution than other investigations, all trials seem to convey the same message: the type of treatment does not matter as long as the stroke patient gets some sort of physiotherapy (Table 2). This must be sobering to the followers of one particular school, but it represents the state of the art today. In no way, however, is this conclusion nihilistic toward physical therapy and rehabilitation. On the contrary, it bears an optimistic prospect that stroke rehabilitation is preferable to spontaneous recovery, 43 regardless of which physiotherapy is chosen. One-sided approaches neglect the complexity of the problem and must be avoided. 4445 These data on effectiveness of stroke rehabilitation imply that early start is preferable. The trials comparing one approach to another are invariably burdened by the relatively late start of the particular intervention to be tested. Future studies aimed at clarifying the issue of optimal rehabilitation should begin therapy in the very early phase after the acute event. Existing data also suggest that if an optimal physical therapy exists, it has not yet been identified. Further investigation is also necessary to determine which patient benefits from a rehabilitation program and which does not. For the time being, the obvious conclusion is that stroke rehabilitation should be applied in forms that are most cost-effective and that can be administered to the largest number of patients. From our present knowledge, or rather from the lack of it, only one thing is undebatably and painfully clear: "Controlled clinical trials are essential if the role of rehabilitation, its indications, and its contraindications are to be more adequately understood." 46 References 1. Wolf PA: Cigarettes, alcohol and stroke. N Engl J Med 1986;315:1087-1089

Ernst Physiotherapy After Stroke 1085 2. Editorial: Why has stroke mortality declined? Lancet 1983; i:1195-1196 3. Grotta JC: Current medical and surgical therapy for cerebrovascular disease. N EnglJ Med 1987;317:1505-1516 4. Anderson AL, Hanvik LJ, Brown JR: A statistical analysis of rehabilitation in hemiplegia. Geriatrics 1950;5:204-218 5. Lehmann JF, Delateur BJ, Fowler RS, Warren CG, Arnold R, Scherzer G, Hurka R, Whitmore JJ, Masock AJ, Chambers KH: Stroke: Does rehabilitation affect outcome? Arch Phys Med Rehabil 1975;56:375-382 6. Dow RS, Dick HL, Crowell FA: Failures and successes in a stroke program. Stroke 1974;5:40-47 7. Adams GF: Prognosis and prospects of strokes, in Cerebrovascular Disease and the Ageing Brain. London, Churchill Livingstone, 1974, pp 141-145 8. Drake WF, Hamilton MJ, Blumenkrantz J: Acute stroke management and patient outcome: The value of neurovascular care units (NCU). Stroke 1973;4:933-945 9. McCann CT, Cuthbertson RA: Comparison of two systems of stroke rehabilitation in a general hospital. J Am Geriatr Soc 1976;24:211-216 10. Feigenson JS, Gitlow HS, Greenberg SD: The disability orientated rehabilitation unit: A major factor influencing stroke outcome. Stroke 1979;10:5-8 11. Garraway WM, Akktar AJ, Prescott RJ, Hockey L: Management of acute stroke in the elderly: Preliminary results of a controlled trial. Br Med J 1980;1:1040-1043 12. Garraway WM, Akktar AJ, Hockey L, Prescott RJ: Management of acute stroke in the elderly: Follow up of a controlled trial. BrMedJ 1980;2:827-829 13. Smith ME, Garraway WM, Smith DL, Akktar AJ: Therapy impact on functional outcome in a controlled trial of stroke rehabilitation. Arch Phys Med Rehabil 1982;63:21-24 14. Hamrin E: Early activation in stroke: Does it make a difference? Scand J Rehabil Med 1982;14:101-109 15. Hamrin E: One year after stroke: A follow up of an experimental study. Scand J Rehabil Med 1982;14:111-116 16. Smith DS, Goldenberg E, Ashburn A: Remedial therapy after stroke: A randomized controlled trial. Br Med J 1981; 282:517-520 17. Lind K: A synthesis of studies on stroke rehabilitation. / Chronic Dis 1982;35:133-149 18. Sivenius J, Pyorala K, Heinonen OP, Salonen J, Riekkinen P: The significance of intensity of rehabilitation of stroke: A controlled trial. Stroke 1985;16:928-931 19. Strand T, Asplund K, Eriksson S, Lithner F, Wester PO: A non-intensive stroke unit reduces functional disability and the need for long term hospitalization. Stroke 1985;16:29-34 20. Strand T, Asplund K, Eriksson S, Hagg E, Lithner F, Wester PO: Stroke unit care: Who benefits? Stroke 1985;17:377-381 21. Feldman DJ, Lee PR, Unterecker J, Lloyd K, Rusk HA, Toole A: A comparison of functionally orientated medical care and formal rehabilitation in the management of patients with hemiplegia due to cerebrovascular disease. / Chronic Dis 1962;15:197-310 22. Gordon EE, Kohn KH: Evaluation of rehabilitation methods in the hemiplegic patient. J Chronic Dis 1966;19:3-16 23. Peakock PB, Riley CP, Lampton TD, Raffael SS, Walker JS: The Birmingham Stroke Epidemiology and Rehabilitation Study, in Stewart GT (ed): Trends in Epidemiology. Springfield, 111, Charles C Thomas Publishing, 1972, pp 231-345 24. Brockelhurst JC, Andrews K, Richards B, Laycock PJ: How much physical therapy for patients with stroke? Br Med J 1978;l:1307-1311 25. Wood-Dauphinee S, Shapiro S, Bass E, Fletcher C, Georges P, Hensby V, Mendelson B: A randomized trial of team care following stroke. Stroke 1984;15:864-872 26. Dignan MB, Howard G, Toole JF, Becker C, McLeroy KR: Evaluation of the rth Carolina stroke care program. Arch Phys Med Rehabil 1987;67:382-386 27. Dombovy ML, Sandok BA, Basford JR: Rehabilitation for stroke: A review. Stroke 1986;17:363-369 28. rris JW, Hachinski VC: Stroke units or stroke centres? Stroke 1986;17:360-362 29. Lindmark B: Evaluation of functional capacity after stroke with special emphasis on motor function and activities of daily living. Scand J Rehabil Med 1988;21(suppl):l-40 30. De Lisa JA: Rehabilitation Medicine: Principles and Practice. Philadelphia, Lippincott, 1988 31. Stern PH, McDowell F, Miller JM, Robinson M: Effects of facilitation techniques in stroke rehabilitation. Arch Phys Med Rehabil 1970;71:526-531 32. Westcott EJ: Traditional regimens for hemiplegic patients. Am J Phys Med 1967;46:1012-1023 33. Zinn WM: Assessment, treatment and rehabilitation of adult patients with brain damage. Int Rehabil Med 1978;l:3-9 34. Loggigian MK, Samuels MA, Falconer J, Zagar R: Clinical trial for stroke patients. Arch Phys Med Rehabil 1983;64:364-367 35. Lord JP, Hall K: Neuromuscular reeducation versus traditional programs for stroke rehabilitation. Arch Phys Med Rehabil 1986;67:88-91 36. Smedley RR, Smedley WP, Aronica MJ, Fiorino AJ, Soucar E, Reynolds D: Slot machines: Their use in rehabilitation after stroke. Arch Phys Med Rehabil 1986;67:546-549 37. Dickstein R, Hockerman S, Pillar T, Shaham R: Stroke rehabilitation: Three therapy approaches. Phys Ther 1986;66:1233-1238 38. Basmajian JV, Gowland CG, Finlayson AJ: Stroke treatment: Comparison of integrated behavioral physical therapy vs. traditional physical therapy programs. Arch Phys Med Rehabil 1987;68:267-272 39. Health and Public Policy Committee, American College of Physicians: Biofeedback of neuromuscular disorders. Ann Intern Med 1985;102:854-858 40. Burnside G, Tobias HS, Bursill D: Electromyographic feedback in the rehabilitation of stroke patients. Arch Phys Med Rehabili 1982;63:217-222 41. Basmajian JV, Gowland CG, Branstater ME, Swanson L, Trotter J: EMG feedback treatment of upper limb in hemiplegic stroke patients: A pilot study. Arch Phys Med Rehabil 1982;63:613-616 42. Inglis J, Donald MW, Monga TN, Sprode M, Young MJ: Electromyographic biofeedback and physical therapy of the hemiplegic upper limb. Arch Phys Med Rehabil 1984; 65:755-759 43. Johnston MV, Keith RA: Cost benefits of medical rehabilitation: Review and critique. Arch Phys Med Rehabil 1983; 64:147-154 44. Paulsen HF, Ernst E (eds): Schlaganfall Therapie und Nachsorge. Miinchen, Zuckschwerdt, 1989, pp 1-3 45. Ernst E: Probleme der Schlanganfallrehabilitation. Perfusion 1990;3:62-65 46. World Health Organization Task Force on Stroke and Other Cerebrovascular Disorders: Recommendations on stroke prevention, diagnosis, and therapy. Stroke 1989;20:1407-1431 KEY WORDS rehabilitation cerebrovascular disorders physical therapy