COMPLEX ORGANIZATIONAL INTERVENTIONS

Rehabilitation after single-incident brain injury, including stroke, should be delivered through polymodal inpatient and community-based service delivery systems that differ by structure and process. A classification of these different service options, as a “rehabilitation typology,” remains to be agreed upon, at least cross-culturally, and would facilitate comparisons of different studies.¹⁶ The use of each component is determined by a variety of clinical and social factors, including length of time since injury, level of dependency, characteristics of the residual impairment, age of the patient, and resources available.¹⁷ Recommendations that different types of care after stroke are incorporated into systems of clinical care (e.g., Schwamm et al¹⁸) also emphasize the use of acute and rehabilitation interventions in parallel rather than in series (Fig. 48-1), so that strategies optimizing neural protection acutely are combined with key factors identified in postacute care. The benefits of organized complex polymodal interventions so far reported are likely to reflect mainly the effects of preventing the complications described later in this chapter and shown in Table 48-2, functional interventions that focus on teaching new skills, and the use of aids and appliances and environmental modifications to help patients adapt to their impairments, rather than the effects of neural reorganization.

Figure 48-1 Acute treatments focus on neural and systemic protection, whereas rehabilitation involves goal-directed restorative and compensatory training strategies. Both form the basis of practice in postacute stroke rehabilitation units. Optimal outcomes over time (→) are most likely to result from their use in parallel (A), not in series (B).

TABLE 48-2 Complications after Stroke during Inpatient Rehabilitation and Their Management

AED, antiepileptic drug; CBT, cognitive-behavior therapy; CCF, congestive cardiac failure; CPAP, continuous positive airway pressure; DVT, deep vein thrombosis; GI, gastrointestinal; NG/PEG, nasogastric/percutaneous endoscopic gastrostomy; NSAID, nonsteroidal anti-inflammatory drug; PE, pulmonary embolism.

Until about 1995, 20% to 50% of patients after stroke were managed acutely at home in some European countries,¹⁹ despite recommendations to the contrary.²⁰ The benefits of stroke unit care and the gradual introduction of thrombolysis and neuroprotective agents are, alone, likely to reduce this percentage dramatically. Furthermore, patients randomly assigned to receive acute care at home after moderately severe stroke (sufficient to cause persistent neurological deficit affecting continence, mobility, and self-care, and necessitating multidisciplinary treatment), rather than to receive care in an acute and rehabilitation stroke unit or a general ward with stroke team support, were significantly more likely at 1 year to be dead or institutionalized (24% versus 14%; P = 0.03) and less likely to be alive without severe disability (85% versus 71%; P = 0.002) than were patients admitted to the stroke unit.²¹

In their meta-analysis of 23 trials, the Stroke Unit Trialists’ Collaboration²² compared care in stroke units, nine of which were “comprehensive” and included both acute and rehabilitation components, with care given by alternative services, usually in a general medical or geriatric ward with or without a visiting stroke team. The meta-analysis showed that patients who receive unit-based inpatient care do not stay longer in hospital and are more likely to be alive, independent, and living at home 1 year after the stroke, regardless of gender, age, and stroke severity (Fig. 48-2). Thus, in comparison with alternative services, stroke unit care produced reductions in the odds of death at 1 year (odds ratio, 0.86; 95% confidence interval, 0.71 to 0.94; P = 0.005), of death or institutionalized care (odds ratio, 0.80; 95% confidence interval, 0.71 to 0.90; P = 0.0002), and of death or dependency (odds ratio, 0.78; 95% confidence interval, 0.68 to 0.89; P = 0.0003).

Figure 48-2 Care in stroke units or mixed rehabilitation units can be organized and is thus more effective than in general medical wards where it remains disorganized, with or without input from a peripatetic team. NS, nonsignificant.

(Adapted from Stroke Unit Trialists’ Collaboration: Organised inpatient [stroke unit] care for stroke. Cochrane Database Syst Rev 2001; [3]:CD000197.)

A subsequent systematic review of results from nine postacute units, which admitted patients more than 1 week after stroke, in comparison with an alternative service, revealed benefits similar to those resulting from combined acute and postacute units,²³ and in the United States, compliance with postacute but not acute stroke unit guidelines (Table 48-3) has been shown to be correlated with outcome.¹⁶ Secondary analysis has indicated that the reduction in stroke unit deaths probably results from a reduction in the complications of immobility rather than in neurological or cardiovascular complications. The increased number of patients discharged home from stroke units and the reduced requirement for institutional care were attributable largely to an increase in the number of patients returning home physically independent (Rankin score, 0 to 2), rather than dependent (Rankin score, 3 to 5).²⁴

TABLE 48-3 Dimensions of Acute Rehabilitation after Stroke

Compliance with these dimensions of process during postacute rehabilitation after stroke is correlated with outcome.

After Duncan PW, Horner RD, Reker DM, et al: Adherence to postacute guidelines is associated with functional recovery in stroke. Stroke 2002; 33:167-178.

Langhorne and Pollock²⁵ identified aspects of care common to 11 stroke units, 8 of which included acute care, reporting beneficial effects between 1985 and 2000 in the Stroke Unit Trialists’ Collaboration’s systematic review.²² These included comprehensive medical, nursing, and therapy assessments; integration of nursing care within the multidisciplinary team; early mobilization and treatment of hypoxia, hyperglycemia, and suspected infection; avoidance of urinary catheterization; and formalized goal-oriented multidisciplinary team care, with early discharge planning and education and involvement of caretakers.

Evans and associates²⁶ found that many of these aspects of care were not delivered by a peripatetic specialist stroke team (PSST). They compared care delivered to 304 patients randomly assigned to an acute and rehabilitation stroke unit (n = 152) or to general wards supported by a PSST (n = 152). Patients in the stroke unit were monitored more frequently, and more of those patients received oxygen, antipyretics, measures to reduce aspiration, and early nutrition than those in general wards. Many aspects of multidisciplinary care occurred more frequently in the stroke unit, in which complications including stroke progression, chest infections, and dehydration were less frequent. Early feeding, stroke unit management, frequency of complications, and measures to prevent aspiration independently affected outcome. In 1984, a PSST’s input to general ward care was shown to nonsignificantly reduce mortality rates and significantly improve functional recovery in men.²⁷ More recently, in a head-to-head comparison at 1 year of patients randomly assigned to receive care from the PSST with those receiving care in a combined acute and rehabilitation stroke unit, more patients in the PSST group were dead or institutionalized (30% versus 14%; P < 0.001) and fewer were alive without severe disability (66% versus 85%; P < 0.001).²¹ Retrospective data collected by van der Walt and colleagues²⁸ before and 2 years after the introduction of a PSST to a general ward showed significant improvements in prophylaxis for deep vein thrombosis, incontinence management, premorbid function documentation, frequent neurological observations, and early occupational therapy, with fewer severe complications (9% after versus 24% before; P = 0.004), reduced median length of stay (12.0 after versus 18.5 days before; P = 0.003), and more patients independent at discharge (32% after versus 9% before; P < 0.001). It is thus possible that practice developed in stroke units can now be delivered more effectively by a PSST, and further comparison of these two methods of service delivery remains legitimate.

More recently, high-dependency care has been introduced immediately after stroke, providing continuous rather than manual monitoring for hypoxia, hyperglycemia, hypotension, cardiac arrhythmias, and elevated body temperature during the first 48 to 72 hours after admission.^29,30 Currently, its effectiveness is relatively unexplored, but there is a suggestion that it may reduce mortality rates at 3 months and 1 year among patients with severe stroke, without increasing dependency.^31,32

Remarkably, the differences between stroke unit and alternative care persist for many years. Indredavik and colleagues³³ found that even 10 years after random assignment to an acute and rehabilitation stroke unit or a general ward, fewer patients in the stroke unit had died (75.5% versus 87.3%; P = 0.008), more were at home (19.1% versus 8.2%; P = 0.018), and more were at least partly independent with a Barthel Index score of 60 or higher (20.0% versus 8.2%; P = 0.012) or independent with a Barthel Index score of 95 or higher (12.7% versus 5.4%; P = 0.061). Increased survival times 5 years after random assignment to a stroke rehabilitation unit, versus a general medical or geriatric ward, were also reported by Lincoln and colleagues.³⁴

Early discharge, supported by a multidisciplinary outreach team, of medically stable patients after mild and moderate stroke, with an admission Barthel Index score of more than 9, supplements initial stroke unit gains. A meta-analysis of individual patient data from 11 trials of early supported discharge versus conventional care showed that the patients with early supported discharge had a reduced risk of death or dependency (odds ratio, 0.79; 95% confidence interval, 0.64 to 0.97; P = 0.02), a hospital stay shortened by 8 days (P < 0.0001), and significant improvement in extended activities of daily living (P = 0.05), although not in subjective health status or mood in either patients or caretakers. Patients were moderately disabled at discharge with a median discharge Barthel Index score of 15.³⁵ One study revealed that these gains can include better life quality, assessed by the Nottingham Health Profile, at 1 year,³⁶ and another study revealed that gains in domestic and extended activities of daily living are still evident after 5 years.³⁷

Once in the community, patients are known to be at risk of deteriorating as a result of multiple health problems, including falls, depression, and physical and social inactivity and isolation, in addition to age-related symptoms and comorbidity, and health-related quality of life has been shown to significantly decline in the 6 months after discharge.³⁸ Nursing home care does not substitute for stroke unit care,³⁹ but the place for and the optimal process in other service systems for patients later after stroke (e.g., nurse-led wards, nursing homes, or residential placements) remain to be examined.

A meta-analysis of trials of therapy-based outpatient or domiciliary rehabilitation, delivered by either a multidisciplinary team or by a physiotherapist or occupational therapist, with the goal of improving task-oriented behaviors, has shown that deterioration is prevented (odds ratio, 0.72; 95% confidence interval, 0.57 to 0.92; P = 0.009) and dependency in personal care reduced (95% confidence interval, 0.02 to 0.25; P = 0.02).⁴⁰ The effective components and best location for this type of service need further exploration, but benefit has been consistent in trials of community-based occupational therapy, which have provided sufficient data for a meta-analysis of eight trials that showed that intervention was associated with improved personal, extended, and leisure-based activities of daily living, depending on the intervention target.⁴¹ These findings were confirmed in a more wide-ranging systematic review of occupational therapy for stroke patients by Steultjens and associates,⁴² who also noted the need for further studies of the effectiveness of splinting.

Most studies of physiotherapy in patients in the community after stroke investigate the effect of a particular physiotherapy treatment on upper or lower limb function at the level of impairment and mobility, which may improve, rather than on limitations in activity and independence, which, if examined, may not improve. Domiciliary physiotherapy within 6 months after stroke has been shown to reduce probability of readmission after an average of only 2.9 (range, 1 to 8) visits⁴³ and to reduce dependency at less cost than for day hospital attendance.^44,45 In patients more than 1 year after stroke, only four to five physiotherapy sessions produced a clinically small but significant improvement in mobility.⁴⁶

Informal caretakers should be recognized as an important resource: They enable patients to remain in the community,⁴⁷ their support is likely to facilitate patient outcomes,⁴⁸ and depression is more severe in caretakers who feel poorly supported.⁴⁹ Formal support for caretakers is difficult to obtain in the United Kingdom. Trials of psychosocial interventions to support caretakers of patients with stroke, involving information packages, specialist nurses, a mental health worker, or family support workers, have failed to show functional or psychological benefit in patients and only modest psychosocial benefit for caretakers.^50–57 In contrast, there is evidence that caretaker adjustment is increased by education and counseling or by training in social problem-solving skills.⁵⁸ Kalra and colleagues^59,60 demonstrated that training informal caretakers in basic nursing skills and facilitation of personal care techniques reduced costs and caregiver burden and improved psychosocial outcomes for the caretaker and the patient, although there was still no change in patients’ rates of mortality, institutionalization, and disability.

MANAGEMENT OF NEUROLOGICAL IMPAIRMENTS

Evidence of dysphagia, with consequent risk of dehydration, further malnutrition, and chest infections resulting from aspiration, has been reported to occur clinically in about 50% of all patients with stroke admitted to a hospital, with videofluoroscopic evidence of a swallowing abnormality in up to 65% of patients and of aspiration in about 20%.^61,62 Videofluoroscopic and flexible endoscopic evaluation of swallowing increase the reliability with which aspiration can be identified^63,64 when clinicians select which patients need tube feeding. However, whether other aspects of dysphagia management, including dietary modification and compensatory swallowing techniques, reduce the need for tube feeding or the risk of aspiration pneumonia remains unclear despite assertion to the contrary.⁶⁵ The Feed or Ordinary Diet (FOOD) trials⁶⁶ revealed that in 859 patients randomly assigned to receive nasogastric feeding within a week, in contrast to more than a week, after stroke, absolute mortality rates were reduced by 5.8% (95% confidence interval, −0.8 to 12.5; P = 0.09) in the group fed early; in 321 patients randomly assigned to undergo percutaneous enteral gastrostomy (PEG) feeding or nasogastric tube feeding a median of 1 week after stroke, PEG feeding was associated with an absolute increase in risk of death of 1.0% (95% confidence interval, −10.0 to 11.9; P = 0.9) and an increased risk of death or other poor outcome of 7.8% (95% confidence interval, 0.0 to 15.5; P = 0.05). Thus, as is recommended current practice in hospitalized adult patients in general,⁶⁷ nasogastric feeding should be used for dysphagic patients soon after stroke, whereas PEG feeding is reserved for patients who do not tolerate nasogastric feeding and as required in the longer-term care of dysphagic patients.

After stroke, early incontinence is predictive of poor outcomes, including death, lengthy hospital stay, institutionalization, and severe disability.⁶⁸ New urinary incontinence occurs in 40% to 50% of patients during the first week after admission and in 10% to 20% by 6-month follow-up^68,69; it appears to remain at this level in the community over the long term.⁷⁰ The proportions of stroke patients with new fecal incontinence on admission and at 6-month follow-up are lower, at 30% to 40% and 5% to 10%, respectively.^68,69 Causes and predictors^68,71,72 of fecal and urinary incontinence include age; stroke type and severity; “functional incontinence” resulting from stroke-related cognitive and communication difficulties and mobility problems; autonomic neuropathies, usually diabetic in origin; medications causing bladder and bowel hyporeflexia; preexisting bladder outflow obstruction in men or stress incontinence in women; and perhaps depression and lowered self-esteem that result from the incontinence itself.^73,74

Current guidelines (e.g., Royal College of Physicians¹²) for the management of incontinence after stroke recommend initial definition of its cause and associations through a full clinical assessment. The neurogenic cause after stroke is usually disruption of suprapontine inhibition of bladder contractility resulting in detrusor hyperreflexia, but bladder hyporeflexia may occur after cerebellar stroke.⁷⁵ Evidence for the effectiveness of methods promoting continence after stroke is currently derived largely from trials in other (nonstroke) patient groups. Management of urinary incontinence⁷⁶ includes a combination of anticholinergic drug therapy, bladder training by prompted and/or timed voiding^77,78 or biofeedback,⁷⁹ and appropriate management of outflow obstruction and stress incontinence, for which one small study of 26 stroke patients randomly assigned to receive treatment or to control conditions showed no benefit from a 12-week pelvic floor muscle training program.⁸⁰ Similar general principles apply to the management of fecal incontinence,⁸¹ which in inpatients or in discharged patients is associated with needing help to use the toilet⁸² and thus might be addressed by improving functional independence, whereas constipation but not fecal incontinence after stroke is improved by a single clinical/educational nurse intervention up to 6 months later.⁸³

Impairment of trunk control and sitting balance after acute stroke is common, and recovery is a reliable predictor of functional outcome after stroke. Trunk control, measured by the trunk control test at rehabilitation admission a median of 5 weeks after stroke, accounted for 71% of the variance of the motor component of the Functional Independence Measure at discharge 10 weeks after stroke⁸⁴; as measured by the Postural Assessment Scale for Stroke Patients within 2 weeks after stroke, trunk control was strongly correlated with functional outcome, rated by a combined score of both personal (Barthel Index) and instrumental (Frenchay Activities Index) activities of daily living at 6 months after stroke.⁸⁵ A positive correlation has also been found between sitting balance and the Barthel Index score⁸⁶ and between sitting balance and gait at 6 months.⁸⁷ Mudie and associates⁸⁸ found that three specific treatment approaches implemented early after stroke improved sitting balance in the short term in comparison with a nonspecific control approach. Whether early targeting of sitting balance for treatment could improve mobility over the long term is an interesting question that remains to be explored.

Disordered motor control of the limbs after stroke is usually caused by the positive and negative effects of the upper motor neuron syndrome, in which the central abnormality is usually muscle weakness and co-contraction resulting from a failure of coordinated high-frequency motor neuron firing during movement, rather than high tone resulting from spasticity and contracture found at rest,⁸⁹ which is seen clinically in only 20% to 30% of hemiparetic patients 3 to 12 months after stroke.^90,91 Spasticity at rest, when it is a velocity-dependent increase in tonic stretch reflexes with exaggerated tendon jerks, is of diagnostic use but does not contribute significantly to impairment during active voluntary movement and function in either the leg or the arm^92,93 and is only one factor contributing to increased muscle tone in the upper motor neuron syndrome, in which co-contraction and biomechanical changes contribute significantly to the resistance to passive movement.⁹⁴ It has thus been easier to show that pharmacological treatments that reduce the hypertonus of spasticity and co-contraction benefit passive rather than active functional activities, including hand and perineal hygiene, dressing, pain, and limb position,^95–100 and even correction of equinovarus at the ankle to improve weight bearing by the affected leg is unlikely to translate into functional improvements in gait^101,102 unless it is followed by physical measures to reduce soft tissue shortening and a program of task-related training.^103,104

Thus, although treatments for spasticity after stroke achieve passive functional goals by preventing contracture or reduce pain and are of particular value in dependent patients unable to participate in skill learning, the results of treatments for spasticity in general and after stroke have emphasized the fundamental truth of Landau’s statement in 1974¹⁰⁵ that “If the major disability of the upper motor neuron syndrome is due to diminished neural input to segmental control of the final common path (negative symptoms), it follows that only some method of provoking [central nervous system] neuronal regeneration or of improving the potency of and control by surviving upper level neurons could provide a direct remedy.” Landau’s next sentence was “No such approach is on the horizon,” and at that time the focus in rehabilitation was on training in compensatory strategies and social and environmental modification. Since the mid-1990s, however, there have been investigations of neurobiological rehabilitation strategies designed to modify the neural processes and behaviors that occur during recovery (Fig. 48-3) (Dobkin, 2004).¹⁰⁶

Figure 48-3 Both patient A and patient B suffered a stroke affecting the right hand. The T1-weighted magnetic resonance imaging scans on the left (“infarct”) demonstrate that patient A suffered a left pontine infarction and patient B suffered a left middle cerebral artery territory infarction. The five scans to the right of each structural scan show patterns of brain activation, measured with functional magnetic resonance imaging, during repetitive gripping with the affected right hand at different time points over the first 3 months after stroke. (The brain is seen from above, with the anterior aspect at the top of the image.) Recovery of function in patient A is associated with early widespread activation within recognized motor related areas, followed by a gradual focusing toward a normal lateralized pattern; the 3-month (far right) scan shows activity confined largely to the left sensorimotor cortex, as in normal subjects. Recovery of function in patient B is also associated with focusing of the early pattern of widespread activation, but in this instance activity focuses toward the right sensorimotor cortex, ipsilateral to the affected hand. Ipsilateral sensorimotor cortex activation is rarely seen in normal subjects. This focusing is due to increasingly efficient neural circuitry, very similar to that seen during learning of a new complex motor task in the undamaged human brain. However, it is possible to focus only onto regions of the brain that can still generate a motor output signal. In the case of patient B, many of the left hemisphere cortical motor regions have been disconnected by the infarct. Thus, there appears to be a reliance on right motor cortex to generate right hand movements. This “secondary” system is necessarily less efficient, and this is reflected in less complete recovery of patient B in comparison with patient A (hand grip of 64% of the unaffected side for patient B at 3 months, in comparison with 89% for patient A). This finding suggests two important consequences: (1) It ought to be possible to use therapies that are designed to drive this process toward the optimally configured motor system and (2) it helps clinicians appreciate that damage to brain structures is likely to be a major constraint to this process and that reorganization can take place only in surviving brain regions; this has implications about how much recovery can be expected.

Meta-analyses of physiotherapy after stroke have revealed significant improvements in independence in activities of daily living and reduction in impairments with higher intensities of physiotherapy.^107–109 Traditional physiotherapeutic strategies emphasize cutaneous and proprioceptive stimulation, reduction of tone, and therapist-induced central facilitation of normal movement, rather than voluntary activation of proximal or distal musculature to relearn functional tasks. This type of task-related motor relearning was first shown to be effective in improving the ability to balance during seated reaching activities late after stroke.¹¹⁰ Although this approach has not yet been consistently shown to improve functional outcomes and resource use early after stroke,^111,112 its principles of repetition, task orientation, attention, and reward are similar to those that result in motor learning and cortical neural reorganization in nonhuman primates with and without ischemic cortical damage.¹¹³ It is thus a “neurobiological” approach to skill relearning in which repetition of graded and motivating task-related learning programs is envisaged to drive the acquisition of motor, linguistic, and cognitive skills and more complex behaviors through activity-dependent neural reorganization, regrowth, and replacement.¹¹⁴

Although these techniques can be shown to produce neuroplastic effects—in humans, usually by functional imaging techniques—these changes should not be confused with the potential clinical usefulness of these techniques, which requires a robust behavioral effect if measurable functional benefit is to occur. The effects of task-related training are seen most obviously in the function targeted for treatment; change has been demonstrated after stroke in upper and lower limb function and linguistic skills. Factors that optimize skill learning, including implicit rather than explicit feedback¹¹⁵ and a random rather than blocked learning schedule,¹¹⁶ are likely to affect the rate at which a task is learned and the extent of carryover into novel environments. In the arm, training is focused on practicing graded motivating functional tasks, including grasping, reaching, leaning, and manipulating clothing by the affected limb; in the trunk and leg, training is focused on the recovery of bed mobility, sitting and standing balance, transfers, gait stability and velocity, and stair climbing. Kwakkel and colleagues¹¹⁷ showed that leg training of this sort for an average of 30 minutes of therapy 5 days a week for 20 weeks after stroke onset, in contrast to immobilization of the limb in an inflatable cuff for a similar time, increased performance in activities of daily living (measured by the Barthel Index), walking ability, and also hand dexterity, whereas arm training increased arm dexterity alone.

What determines when, in whom, and how task-oriented training should be carried out after stroke, in an individual as well as in a population, remains open to investigation. The duration in training can be increased: for example, to an impractical 6 hours a day for 10 weekdays late after stroke. Training can be performed solely with a therapist or also automated with a robotic,^118,119 most commonly as a treadmill with or without body weight support for gait training¹²⁰ or in a virtual environment,¹²¹ or it can be encouraged, probably at both cortical¹²² and functional levels, and also in animals,¹²³ by constraint of the unaffected upper limb.^124–126 It can also be combined with a variety of impairment-based adjunctive therapies, including muscle-strengthening exercises,¹²⁷ an orthotic,¹²⁸ electrical stimulation of the arm or leg,^129–131 sensory stimulation,¹³² mental imagery,¹³³ and musical feedback.¹³⁴

The available evidence for the effectiveness of these and other physical therapies was systematically reviewed in 2004 by van Peppen and colleagues.¹³⁵ Which therapy modalities should be used, to what purpose, how they should be combined, at what stage after stroke, at what intensity and for how long, and in which patients are questions that are beginning to be answered clinically with an evidence base, rather than at best theory and at worst belief. The need to be able to do this is emphasized particularly by the evidence beginning to show long-term persistence of task-related training effects^136,137 and the useful effects of interventions late after stroke (e.g., Werner and Kessler¹³⁸ and van der Lee et al¹³⁹). It is accompanied by a need to explore how the new technologies, which may not always yield benefit (e.g., Richards et al¹⁴⁰), can best be put best to work.

Similar principles are beginning to drive trials of rehabilitation therapies of the focal, rather than global,¹⁴¹ cognitive disorders observed after stroke. According to results of clinical measures, language impairments affect about 40% of the stroke population at stroke onset¹⁴² and about 10% at 6 months after stroke.¹⁴³ Minimal spontaneous improvement in language functions occur more than 1 year after stroke. An early meta-analysis¹⁴⁴ and review¹⁴⁵ of trials of speech and language therapies concluded that there was evidence of effectiveness, but restriction of the evidence to randomized controlled trials, in comparison with no therapy, informal support, or another speech and language therapy, failed to yield clear evidence of effectiveness.¹⁴⁶ However, Bhogal and colleagues¹⁴⁷ found that increased intensity of therapy was associated with improvement, which occurred in four studies that provided an average of 9 hours of therapy per week for 11 weeks, in contrast to four studies with negative results that provided approximately 2 hours per week for 23 weeks. Although the type of therapy required remains unclear (e.g., Doesborgh et al¹⁴⁸), Pulvermüller and colleagues¹⁴⁹ suggested that the principles of (a) massed practice—for example, 30 to 35 hours of therapy over 10 days rather than the same therapy time over a longer period—and of (b) constraint (in this case of maladaptive methods of communication), and of (c) everyday behavioral relevance are important factors in generating neuroplastic and thus functional changes in language skills.

The presence of other cognitive problems in the acute phase are important independent predictors of long-term outcome, in addition to neurological deficits rated by standard measures, particularly unilateral neglect^150,151 and executive dysfunction.¹⁵² Neglect, as currently measured, is seen in more than 60% of patients assessed within 2 to 3 days of admission for stroke, more commonly after right than left hemisphere stroke.¹⁵³ Of patients with neglect 2 to 4 weeks after stroke, 80% still exhibit aspects of neglect on testing at 6 months,¹⁵⁴ and the poor functional outcome predicted by the presence of early neglect is still seen if neglect has resolved on testing,¹⁵¹ possibly as a result of residual problems with nonlateralized attention.¹⁵⁵ Neglect, like dysphasia, is heterogeneous; different combinations of lateralized and nonlateralized attentional and motor deficits are identifiable in each patient. A number of techniques have been shown to improve lateralized spatial representation, including caloric stimulation^156,157; contralesional neck muscle vibration, in one trial with scanning training¹⁵⁸; prism adaptation¹⁵⁹; and contralesional limb activation,¹⁶⁰ as well as alerting techniques to increase nonlateralized loss of attentional capacity.¹⁶¹ Some of these techniques have also been reported to improve activities of daily living, notably neck muscle vibration with scanning training¹⁵⁸ and prism adaptation,¹⁶² and contralesional limb activation in one study reduced length of stay by 24 days.¹⁵⁰ Whether a treatment battery tailored to the neglect subtypes present in each patient^163,164 will prove sufficiently practical and robust to produce significant functional benefits, not yet evident in the literature on systematic review,¹⁶⁵ remains to be explored.

In the context of stroke, use-dependent neural reorganization after damage is facilitated by other extrinsic and intrinsic inputs that modulate neural excitability and augment long-term synaptic potentiation. Use-dependent network remodeling at cellular and systems levels is also likely to be needed for the successful application of techniques to promote neural regrowth and/or replacement.^166,167 Adjunctive facilitators of this process include manipulation of attention¹⁶⁸; increase or decrease of somatosensory input, shown to improve swallowing¹⁶⁹ and leg¹⁷⁰ and arm¹²² function; neuromodulatory drugs¹⁷¹; electrical¹⁷² and magnetic¹⁷³ stimulation of the brain; and possibly exercise.¹⁷⁴

Pharmacological facilitation and inhibition of neurological recovery after stroke are also likely to be of clinical relevance. The use-dependent plasticity that underpins motor training and perceptual learning is dependent on N-methyl-D-aspartate receptor activation and γ-amino butyric acid–mediated (GABAergic) inhibition,^175,176 and these are in turn modulated by cholinergic, noradrenergic, serotonergic, and dopaminergic transmitter systems.¹⁷¹ Early experiments by Feeney and coworkers¹⁷⁷ in rats showed that motor recovery after suction ablation of the sensorimotor cortex was improved by amphetamine before practice on a beam-walking task; neither practice nor amphetamine alone had any benefit. Subsequently, α₂ antagonists, increasing noradrenergic effects, were found to facilitate motor recovery, and α₁ antagonists and α₂ agonists, which decrease norepinephrine release or block its postsynaptic effects, reinstate deficit in clinically recovered animals or slow their recovery.^178,179 There is also evidence that modulation of other neurotransmitter systems has similar effects.¹⁸⁰ The beneficial effect of these pharmacological interventions is dose dependent, and the timing of administration is likely to be important, the effect being dependent on close temporal linkage to behavioral experience. In animals, there is a strong correlation between the functional effects of altering levels of various neurotransmitters and their ability to facilitate increases in synaptic efficacy through long term potentiation.¹⁸⁰

Attempts have been made to translate these findings into treatments for stroke patients. In the first clinical trial of noradrenergic enhancement coupled with physical therapy, patients who had suffered from hemiplegic stroke were randomly assigned to receive either 10 mg of dextroamphetamine or placebo 45 minutes before physiotherapy.¹⁸¹ Follow-up assessment 24 hours after treatment revealed a 40% improvement from baseline scores with dextroamphetamine in comparison with placebo. Small numbers of patients in this study made interpretation difficult. A subsequent failure to replicate these findings by Borucki and associates¹⁸² was attributed to a different experimental design, particularly a failure to schedule physiotherapy immediately after amphetamine treatment. More recent studies with dextroamphetamine have had conflicting results,^183,184 but positive results have been published for daily doses of levodopa (100 mg) used in conjunction with physiotherapy.¹⁸⁵ Studies of the effects of drugs used in conjunction with speech therapy have yielded some success.¹⁸⁶ Although larger scale studies are under way, the use of pharmacological agents to facilitate the effects of physical and speech therapy is currently only experimental. Conversely, there is some evidence to suggest that patients receiving drugs that reduce the effects of norepinephrine or dopamine, or increase the effects of GABA, at the time of stroke or shortly afterward have poorer outcomes than do patients who do not receive any of these drugs¹⁸⁷; thus, care must be taken in their prescription.

Maladaptive or ineffective neural reorganization after stroke may contribute to emotionalism, central poststroke pain (CPSP), and involuntary movement disorders, which include hemiballismus, hemichoreoathetosis, distal resting and/or action tremor, and proximal postural tremor. CPSP, originally described in 1906 as a phenomenon after thalamic stroke by Dejerine and Roussy,¹⁸⁸ was found in 16 (8%) of 207 hospitalized stroke patients 6 months after stroke by Andersen and colleagues.^48,49 Pain may result from stroke involving spinothalamocortical afferent pathways in the medulla, pons, midbrain, thalamus, subcortical white matter, and cortex. Leijon and colleagues¹⁸⁹ found in 27 patients with CPSP that the thalamus was involved in only 9, lesions were in the lower brainstem in 8 and lateral and superior to the thalamus in 6, and onset of the pain was delayed by more than a month in 13 patients. The pain was described as “burning, aching, pricking, and lacerating”; the intensity was increased by external stimuli, either noxious (hyperalgesia) or not noxious (allodynia), the most common being joint movements, cold and light touch, and, in five patients, emotional stimuli. All patients had decreased temperature sensation. Hypersensitivity to cutaneous stimuli was found in 88% of the patients, and sensations of touch and vibration were abnormal in only 52% and 41%, respectively. There were also relatively low incidences of weakness and ataxia, which were present in 48% and 62%, respectively. Treatments in individual patients include all the drugs and physical methods used for neuropathic pain in general; descending modulation of pain pathways by electrical motor cortex stimulation,¹⁹⁰ rather than deep brain stimulation,¹⁹¹ and more recently repetitive TMS over the hand area of the motor cortex¹⁹² have shown some promise in the treatment of CPSP. Both deep brain stimulation and electrical motor cortex stimulation have been used with some success to treat poststroke movement disorders,¹⁹¹ whereas antidepressants are effective in treating emotionalism.^192–194

MANAGEMENT OF SYSTEMIC COMPLICATIONS

In a prospective study of 311 consecutive patients recruited within 7 days of stroke onset, Langhorne and colleagues¹⁹⁵ recorded complications in 85% of the patients during hospital admission (see Table 48-3); 6, 18, and 30 months after discharge, there was a high frequency of infections, falls, unexplained blackouts, pain, and symptoms of depression and anxiety. Seizures and chest infections tend to occur soon after admission; other common complications, including venous thromboembolism, skin breakdown, urinary tract infections, falls, and depression, are still occurring at 30 days.¹⁹⁶ More than 60% of patients have more than one complication during inpatient rehabilitation, and complications are more common with increasing age, stroke severity, premorbid disability, and low serum albumin level.^195–197 The occurrence of a complication is more likely to be associated with death,¹⁹⁶ and the causes of death most likely to be prevented by stroke unit care are those resulting from immobility, particularly as a result of infection and venous thromboembolism. More recently, emphasis on acute care after stroke has begun to define potentially avoidable complications that cause disordered physiological homeostasis acutely and may affect mortality rates, morbidity, length of stay, and longer-term outcome^31,32,198; these include airway patency and disordered breathing (lowering blood oxygenation), cardiac problems, blood pressure control, temperature regulation, and glycemic control.³⁰ The prevention and early treatment of the acute, early, and later systemic complications of stroke are thus an important aspect of organized multidisciplinary care and rehabilitation after stroke. They require at least easy access to acute care and diagnostic facilities,^196,197 if not high-dependency care in the acute phase for some patients, with continuous monitoring of cardiac, respiratory, metabolic, and neurological functions during the first 72 hours after stroke.

Venous thromboembolism was once common after stroke; symptomatic and asymptomatic deep vein thrombosis and pulmonary embolism were present in up to 70% of, respectively, all patients and those studied post mortem.²⁰⁰ In contrast, nearly 20 years later, Kelly and associates^200,201 found that in 102 unselected patients receiving stroke unit care, aspirin, and compressive therapy with stockings, symptomatic deep vein thrombosis and pulmonary embolism occurred in only 3% and 5% of patients, respectively; however, noninvasive magnetic resonance direct thrombus imaging revealed the prevalence of venous thromboembolism, deep vein thrombosis, and pulmonary embolism to be high (40%, 18%, and 12%, respectively, and increasing to 63%, 30%, and 20%) in nonambulatory patients with a Barthel Index score of 9 two days after stroke. In the context of early mobilization and stroke unit care, either aspirin or low-dose unfractionated heparin has been shown to reduce the frequency of symptomatic pulmonary embolism to below 1%.^203,204 Because there is no evidence that routine low-dose subcutaneous low-molecular-weight heparin is superior to aspirin in preventing pulmonary embolism or deep vein thrombosis,²⁰⁵ current guidelines for routine venous thromboprophylaxis (e.g., Royal College of Physicians¹²) recommend aspirin, early mobilization, and, for patients with leg weakness, graduated compression stockings, which prevented deep vein thrombosis in all hospitalized patients in one study²⁰⁵ but for which there is insufficient evidence at present to show that this is the case after stroke.²⁰⁶ Factors thought to confer higher risk, such as previous deep vein thrombosis or pulmonary embolism, necessitate individualized treatment decisions, and management in other high-risk groups, such as those initially nonambulatory after stroke, probably requires further definition (e.g., Kelly et al^201,202).

The detraining effects of immobility and reduced activity²⁰⁸ compound the increased difficulty in completing activities of daily living that results from comorbid musculoskeletal and cardiorespiratory dysfunction and from the increased aerobic requirements of walking that are secondary to neurological impairments. Even in normal people, bed rest for 4 to 6 weeks results in a loss in exercise capacity of 0.9% per day²⁰⁹ and a decrease in muscle strength of up to 40%.²¹⁰ After stroke, exercise capacity is reduced to about 60% of normal at 1 month,²¹¹ making the performance of activities of daily living effortful and fatiguing. Although activities of daily living exert a training effect, so that exercise capacity improves, it remains reduced at 6 months²¹² and probably in the long term. Activity intolerance is thus common among stroke survivors, who may work close to their individual maximal exercise capacity while engaged only in domestic chores, in comparison with age- and weight-matched controls.²¹³ Either aerobic training²¹⁴ or muscle strengthening^215,216 or both²¹⁷ usually improve targeted physiological outcomes and sometimes balance, walking speed, and distance, but not consistently activities of daily living^136,218; this finding is sufficient to contribute to recommendations for exercise after stroke from the American Heart Association.²¹⁹

Detraining is likely to be one of the factors contributing to poststroke fatigue, which occurs in 40% to 70% of patients late after stroke in community studies,^220–222 and patient-derived stroke-specific, health-related quality-of-life measures identify lack of energy as one of the four most consequential problems, together with those related to physical, psychosocial, and communication difficulties.^223,224 Despite its relevance to quality of life, the natural history, causes, and management of poststroke fatigue remain underinvestigated. It often occurs independently of prestroke fatigue and depression and poststroke depression^220,221,225 as “primary” poststroke fatigue, possibly resulting from stroke-disordered attention,²²⁶ or in association with physical impairment,²²¹ psychological factors, sleep-disordered breathing and obstructive sleep apnea,²²⁷ or incidental systemic comorbid conditions. Management clearly starts with accurate differential diagnosis, but there are no guidelines to direct management of primary poststroke fatigue or evidence that modafinil, which may²²⁸ or may not²²⁹ be of benefit in multiple sclerosis, is of benefit after stroke.

During hospital admission after stroke, patients suffering from undernutrition have a greater risk of pneumonia, other infections, and gastrointestinal bleeding than that in other patients.²³⁰ A higher risk of chest infection and poor nutritional status within the first month after stroke is predicted by abnormal swallowing, which is also associated with increased disability, length of hospital stay, and institutional care and is an independent predictor of mortality during the first 6 months after stroke^61,62; rates of death from pneumonia over the longer term are also higher if dysphagia persists.²³¹ Although the prevention of chest infections remains problematic, the prevention of undernutrition in hospital has been clarified. Routine oral supplementation versus no supplementation in admitted stroke patients could not be shown to reduce death or poor outcome when the rate of malnutrition on admission was only 8%²³²; whether it would with the higher rates of malnutrition on admission of about 40% reported before 1995²³³ has not been investigated. Nasogastric tube feeding within a week of stroke reduces the mortality rate,⁶⁶ probably by preventing malnutrition, as it does in other groups of vulnerable hospitalized patients,²³⁴ rather than by reducing the incidence of chest infections, which nonetheless occurred, for example, in 44% of nasogastrically fed patients reported by Dziewas and colleagues.²³⁵

Pain after stroke is most often musculoskeletal in origin, rather than CPSP, and the result of an exacerbation of preexisting osteoarthritis and/or reduced movement and mobility caused by the stroke. Shoulder pain occurs in more than 60% of patients during the 6 months after stroke and its incidence may increase after discharge.²³⁶ Its occurrence is correlated with prestroke shoulder pain, severe upper limb weakness, neglect, sensory loss, visual field defects, and glenohumeral subluxation, but their role in its pathogenesis remains uncertain, and shoulder subluxation is not clearly a cause. Vasomotor changes in the limb, shoulder-hand syndrome, and reflex sympathetic dystrophy may also be present but should not be confused with the painful shoulder, and other disorders of the shoulder including fractures should be ruled out. Shoulder pain causes considerable morbidity and, once established, is difficult to treat. Prevention, particularly in patients at high risk, should include support of the flaccid arm²³⁷ and appropriate handling techniques to avoid traction injury. Treatment is likely to be determined partly by the tone in shoulder muscles²³⁸: A flaccid shoulder requires support at all times and possibly benefits from functional electrical stimulation (e.g., Renzenbrink and IJzerman²³⁹), whereas a spastic shoulder requires maintenance of range of motion by physical techniques and possibly the addition of botulinum toxin. Local steroid injection should be avoided unless there is clear evidence of an inflammatory component to the pain.

There is a high cumulative incidence of psychiatric disorder after stroke. It can be difficult to diagnose in the presence of speech disturbance, anosognosia, and abulia, and it includes major and minor depression, agoraphobia, social withdrawal, apathy and self-neglect, irritability, and pathological emotionalism. However, in 128 community-dwelling patients during the first year after stroke, House and colleagues²⁴⁰ found that little of it persisted, and only two cases of major depression manifested during the whole 12 months. This frequency is lower than that found in other studies: for example, about 25% in a rehabilitation unit within 3 months after stroke²⁴¹ and 15% in community-dwelling patients 4 months after stroke in the study by Burvill and colleagues²⁴² in Perth, Australia. There remains a suggestion that left-sided anterior strokes affecting frontal-subcortical circuits predispose to depression,^243,244 but the majority of studies emphasize that the major risk factors for depression are nonspecific, of the sort seen in older people with other physical illness,²⁴⁵ and they include a history of previous depression, stroke severity, physical dependency, and social inactivity.^48,49,246 Over the longer term, any association between stroke and depressive symptoms is attributable largely to physical limitations.²⁴⁷ Although depression soon after stroke probably reduces later functional outcome²⁴⁸ and its presence 1 year later is clearly related to a reduced quality of life,⁶ two systematic reviews revealed that the evidence of benefit from early prevention or diagnosis and treatment was minimal.^249,250

FUTURE DEVELOPMENTS

Further research and development in rehabilitation after stroke should continue to focus on maximizing effective service delivery and investigating new organizational systems and restorative techniques and their transfer into routine clinical use.

Because of the evidence for the effectiveness of comprehensive acute and rehabilitation stroke unit care and subsequent supported discharge and intervention in the community, and the national recommendations and guidelines that this evidence makes possible, these service delivery systems of care and rehabilitation can be available to a greater proportion of patients admitted after stroke, particularly when linked to an explicit multisite audit of the organization and process of stroke care of the sort undertaken at intervals by the Royal College of Physicians of London.²⁵¹ Linkage of these stroke care systems into networks will be gradually developed to facilitate multicenter clinical research, hasten enlargement of the interventional evidence base,²⁵² and promote the delivery of coordinated rather than fragmented systems of care.¹⁸ Parallel networks aiming to facilitate the transfer of proof-of-principle studies in the translational restorative neurosciences, while the pharmaceutical industry continues to focus on new methods of neural protection, are likely to follow.

It is now established that the characteristic features of comprehensive stroke unit care include a standardized protocol for acute assessment, intermittent monitoring, medical treatments, and a strategy for the prevention of complications through the work of a multidisciplinary team integrating medical care, nursing, and rehabilitation. Clinical research networks will enable exploration of the effectiveness of further aspects of plant and process that contribute to outcome, including the usefulness of acute continuous physiological monitoring³²; other strategies to detect and prevent complications, such as very early mobilization²⁵³; and further definition of the aspects of acute and early goal-focused interdisciplinary rehabilitation that optimize functional outcome.¹⁷

Since the mid-1990s, the neural reorganization that takes place longitudinally in response to damage and its relation to functional recovery have become better understood.²⁵⁴ This understanding has, for example, provided a biological rationale for the efficacy of task-related training and the need for intensive therapy, and it has led to investigations of how and whether other drivers of neural reorganization can increase function. Although these techniques may reduce impairments in the laboratory and at the proof-of-principle level, the generation of robust functional gains remains elusive, and their introduction and relevance to the clinical domain remain unclear.²⁵⁵ Exploration of how restorative techniques can be applied clinically to generate functionally relevant neural reorganization, regrowth, and repair through novel technologies and drugs, alone or in combination, and in the context of task-related training is a journey of universal interest.