Interventions to Improve Recovery After Stroke

Key Points

Neural repair is a therapeutic strategy distinct from acute stroke strategies such as reperfusion: biologic targets are different. The goal is to boost function in surviving brain elements, not to salvage threatened tissue, and time windows are measured in days to weeks, not hours.
Many classes of therapy are under study in animals and in human trials to improve stroke recovery, including drugs, biologic agents, brain stimulation, activity-based therapies, cognitive-based therapies, and lesion bypass.
Some repair-based therapies are introduced days to weeks post-stroke, to amplify innate repair mechanisms; others are offered months to years post-stroke, to stimulate new brain plasticity.
Repair-based therapies improve behavioral outcomes on the basis of experience-dependent brain plasticity: a drug may galvanize the brain for repair, but behavioral reinforcement is also needed to achieve maximal gains. This is an important difference as compared to reperfusion and preventative stroke therapies, where patients need not perform any particular behavior for the drug to work.
Repair-based therapies are not a one-size-fits-all program—a single treatment is unlikely to improve outcomes across all infarct sizes or behavioral deficits. Research suggests a future where such therapies are individualized: based on measures of brain structure and function, genetics, and lifestyle factors, akin to what is currently done in many other fields of medicine.

Biology of Stroke Recovery Suggests Therapeutics Targets

A new stroke sets numerous biologic pathways into motion. These include the ischemic cascade acutely, followed by numerous immunologic events. Later, a sequence of cellular and molecular events emerges that supports spontaneous tissue repair.

Studies have provided insights into the mechanisms of post-stroke neural repair. In animals, an experimental stroke results in an ordered change in expression of numerous genes, including growth-related events such as growth factor release, increased levels of growth inhibitors such as Nogo and MAG, capillary growth, axonal sprouting, synaptogenesis, and glial cell activation. These changes are seen both near and distant from the injury, and generally peak during the initial weeks post-stroke. Human studies using noninvasive neuroimaging and neurophysiologic methods have identified stroke recovery mechanisms concordant with preclinical findings, with behavioral improvement accompanied by cortical map reorganization, regional changes in brain activity and global changes in brain networks both intra- and inter-hemispherically, and associated changes in brain structure. ^, Uninjured brain areas connected to an injured brain region may show depressed function, a process known as diaschisis, ^, resolution of which may be linked with behavioral improvement. Taken together, these restorative events represent potential therapeutic targets to promote neural repair.

Repair-Based Therapies are Distinct From Acute Therapies

A clear distinction must be made between repair versus acute stroke therapeutic strategies. Acute stroke treatments are based neuroprotection or reperfusion (tPA or thrombectomy), have an actionable timeframe measured in minutes-hours, the target is a clot, and the goal is to salvage threatened brain tissue. In contrast, repair-based treatments generally have a time window measured in days-weeks or longer, the target is the brain, and the goal is to promote favorable plasticity. Promoting plasticity in the brain is a more complex treatment goal than removing a fresh thrombus, and this affects clinical trial design (see below and Table 61.1 for a summary of key studies).

TABLE 61.1

Lessons From Key Repair-Related Stroke Trials.

Study	N	Time Post-Stroke When Treatment Started	Treatment Arms	Main Finding
Drugs
FLAME	118	5–10 days	[A] 20 mg/day fluoxetine [B] Placebo	[A] > [B] in Fugl-Meyer (FM) score change to day 90
1. Therapy outside of study from baseline to day 90 might influence outcomes but was not measured. 2. Randomization methods did not ensure groups matched at baseline for primary outcome measure; consequently, FM scores in fluoxetine group were higher at baseline, complicating results interpretation. 3. Note that this was a positive study that showed reducing depression is associated with reduced motor impairment.
FOCUS	3127	2–15 days	[A] 20 mg/day fluoxetine [B] Placebo	No difference in mRS at 6 months
1. Entry population very heterogeneous—treatments in stroke recovery do not benefit from a broad one-size-fits-all approach. 2. Primary outcome measure (mRS) lacks granularity, limiting interpretation of negative results. 3. Very broad population enrolled, baseline testing used dichotomous measures and so behavioral change over time could not be evaluated, all outcomes were patient-reported (no live exams done) and indeed were measured using questionnaires delivered by mail—pragmatic study design likely premature for current stage of drug development.
Effect of levodopa in combination with physiotherapy on functional motor recovery after stroke	53	3 weeks to 6 months	[A] Sinemet prior to physiotherapy [B] Placebo prior to physiotherapy	[A] > [B] for gains in arm and leg motor function
1. Strengths include daily therapy at a single inpatient facility, and physiotherapy carefully timed to match peak drug levels.
DARS	593	5–42 days	[A] Sinemet prior to routine therapy [B] Placebo prior to routine therapy	Ability to walk independently not different for [A] (41%) versus [B] (45%)
1. Inconsistent ingestion of drug and provision of therapy per protocol, diagnosis of stroke based on clinical findings—pragmatic study design likely premature for current stage of drug development.
Activity-Based Therapies
EXCITE	222	3–9 months	[A] Constraint induced motor therapy [B] Usual Care	[A] > [B] in Wolf Motor Function and Motor Activity Log at 12 months
1. A positive phase III study showing potential for a restorative therapy to reduce disability. 2. Groups not matched in intensity of therapy, making it difficult to conclude if between-group differences are due to differences in treatment content or intensity.
VECTORS	52	10 days (upon admission to inpatient rehab)	[A] High intensity constraint induced motor therapy [B] Standard intensity constraint induced motor therapy [C] Control (usual care occupational therapy)	[B] = [C] in ARAT scores at 90 days. [A] had significantly less improvement than [B] and [C]
1. Higher therapy doses are not always better. 2. Timing of intervention delivered is a complex modifier of treatment effect.
LEAPS	408	2 months	[A] Treadmill training 2 months post-stroke [B] Treadmill training 6 months post-stroke [C] Home exercise 2 months post-stroke	No difference between all groups
1. This is a positive study across all 3 treatment groups, with 52% of patients showing improved walking function; a pill started on day 1 post-stroke that improved function in 52% of patients would garner great praise—so should these interventions. 2. Intense training produces functional gains when initiated 2 or 6 months after stroke. 3. Use of a modality-specific endpoint (measure of gait) is a strength: provides granularity for measuring treatment effects and is directly aligned with content of treatment (gait therapy). 4. Use of an active control group matched in treatment intensity helped compare two reasonable treatment options. 5. Use of sliding dichotomous outcomes has strengths in stroke: the definition of success varies according to severity of baseline deficits. 6. Therapy provided outside of the study procedures was common and complicated interpretation of results.
The Queen Square Upper Limb Neurorehabilitation programme	224	Chronic stroke (median 18 months)	[A] 90 hours of therapy over 3 weeks of individualized occupational therapy and physical therapy	Statistically and clinically important gains in upper limb function sustained for at least 6 months post-therapy
1. High-dose therapy produced large gains, even in chronic stroke. 2. Large functional gains in upper limb, across a broad range of deficits, without adverse events or subject dropout, suggesting broad utility. 3. Global function at baseline was high (Barthel 19/20), creating interest in understanding whether study effects are also positive in more severely affected patients. 4. Lack of a control group, though substantial spontaneous recovery in chronic phase unlikely.
VA Robotics	127	>6 months post-stroke	[A] Robot therapy [B] Dose-matched, non-robot, upper extremity therapy [C] Usual Care	Gains were small and did not differ between groups.
1. Enrollees had severe baseline deficits; treatment gains are often limited in this population.
RATULS	770	1–260 weeks	[A] robot-assisted training [B] repetitive upper limb functional task therapy [C] Usual Care	Gains did not differ between groups
1. Use of a sliding responder analysis addresses inter-subject differences in baseline deficits. 2. Enrolling patients 1–260 weeks post-stroke means treating very heterogeneous brain states. 3. Pragmatic study design might have impeded hypothesis testing.
EXPLICIT	159	<14 days post-stroke	For patients with favorable prognosis: [A1] Modified constraint induced motor therapy [A2] Usual Care For patients with unfavorable prognosis: [B1] neuromuscular stimulation [B2] Usual Care	[A1] > [A2] in ARAT scores at 12, but not 26, weeks [B1] = [B2] in ARAT scores
1. Stratifying treatment choices based on prognosis reveals differences in response to repair-based therapies.
Telerehabilitation versus In-Clinic Therapy	124	3–36 weeks post-stroke	[A] Telerehabilitation therapy in the home [B] Therapy at an outpatient clinic	Statistically and clinically important gains in FM score that did not differ between groups, as hypothesized in this noninferiority trial.
1. Home-based telehealth is comparably efficacious for improving outcomes after stroke and could increase access to rehabilitation care. 2. Holistic approach integrated secondary stroke prevention with rehabilitation therapy. 3. Compliance rate of 98.3% in telerehabilitation group supports using games to drive therapy. 4. Results create interest in economic analysis.
Brain Stimulation
NICHE	167	3–12 months	[A] Therapy preceded by 1 Hz (low-frequency) transcranial magnetic stimulation (TMS) to contralesional motor cortex [B] Therapy preceded by sham TMS	Improved FM scores in both groups with no differences in responder rate between [A] and [B]
1. Optimal frequency of TMS stimulation, brain target, and time post-stroke may need to be individualized. 2. Most patients in both groups showed substantial functional gains following therapy.
EVEREST	164	≥4 months	[A] Therapy with electrical epidural stimulation of ipsilesional motor cortex [B] Therapy alone	Responder rates (4.5 point FM and 0.21 point Arm Motor Ability Test gain) not different between [A] (32%) and [B] (29%)
1. Translation was imprecise: positive preclinical studies required preserved motor evoked responses upon cortical stimulation but EVEREST did not; post hoc analysis found [A] had much higher responder rate (67%) when motor evoked response was preserved.
Transcranial Direct Current Stimulation to Treat Aphasia after Stroke	74	>6 months	[A] Speech therapy with tDCS to left temporal lobe [B] Speech therapy with sham stimulation	Change in correct object naming higher in [A] > [B]
1. Cortical stimulation target identified by a functional magnetic resonance imaging scan, insuring that stimulation was centered over functionally intact cortex, and underscoring the utility of using a measure of brain function to individualize restorative therapy details. 2. Employed a futility design, useful with heterogeneous population: the study did not provide evidence that this tDCS protocol is futile—this treatment is nonfutile and worthy of further study.

Repair-Based Therapies Under Study

Many classes of restorative therapy are under study to promote brain repair, using wide-ranging strategies. These are summarized in Box 61.1 , with most having reached human trials.

BOX 61.1

Repair-Based Therapies Under Study

BDNF , Brain-derived neurotrophic factor; B-FGF , basic fibroblast growth factor; G-CSF , granulocyte-colony stimulating factor; hCG , human chorionic gonadotropin; MAG , myelin-associated glycoprotein; OP-1 , osteogenic protein 1; SSRI , selective serotonin reuptake inhibitors.

Drugs

Including SSRIs, stimulants, l -Dopa, memantine, maraviroc

Biologic Agents

Growth factors: Including erythropoietin, hCG, BDNF, G-CSF, b-FGF, OP-1

Monoclonal antibodies: Including anti-MAG Ab, anti-Nogo A Ab

Stem cells: Including mesenchymal stromal cells and neural stem cells

Activity-Based Therapies

Including occupational, physical, and speech language therapy; constraint-induced movement therapy; robotics, telerehabilitation

Cognitive-Based Therapies

Including mental imagery, environmental enrichment, prism adaptation

Brain Stimulation

Including transcranial magnetic stimulation, transcranial DC stimulation, deep brain stimulation, vagal nerve stimulation

Lesion Bypass

Including brain-computer interface, nerve transfer surgery

Drugs

Many drugs have been studied to promote stroke recovery. Most are small molecules that target a specific brain neurotransmitter system. Many have the advantage of transport through the blood-brain barrier. The two most studied drug classes for stroke recovery target serotonin and dopamine.

Serotonergic

A number of small studies suggested benefits from selective serotonin reuptake inhibitors (SSRIs). The Fluoxetine for Motor Recovery After Acute Ischemic Stroke (FLAME) study was a double-blind, placebo-controlled trial that enrolled nondepressed hemiplegic/hemiparetic patients within 10 days of ischemic stroke onset. Patients were randomized to 3 months of oral fluoxetine (20 mg/day) or placebo. Patients randomized to fluoxetine showed significantly greater gains on the primary endpoint, change in the arm/leg Fugl-Meyer motor score to day 90 ( P = .003).

FOCUS (Effects of Fluoxetine on Functional Outcomes after Acute Stroke) was a pragmatic placebo-controlled trial that randomized 3127 patients with a clinical diagnosis of stroke 2–15 days prior to 6-months of either fluoxetine 20 mg/day or placebo and found no difference in the primary endpoint, distribution of the modified Rankin scale scores at 6 months. Several features of this study might limit its impact. First, an extremely heterogeneous population was enrolled—patients with any neurologic deficit and any degree of severity were eligible. The index infarct could be tiny or massive, located in any part of the brain, and produce any constellation of symptoms of any degree of severity. Repair-based therapies are not likely one-size-fits-all but instead are of maximal value in specific subpopulations aligned with treatment mechanism. Second, FOCUS used a pragmatic study design: baseline testing used dichotomous measures (e.g., able to lift both arms or not), which lack granularity and also prevent measurement of within-subject change over time, and study outcomes were patient-reported (no live exams done) and measured using questionnaires delivered by mail. Use of a pragmatic study design is likely premature for stroke recovery trials (see below). Third, the FOCUS’s primary outcome was the modified Rankin Scale, which, with only seven levels, captures global disability with limited granularity.

Other studies have described favorable effects of SSRIs on mood and anxiety. This is key, as post-stroke depression affects 31% of patients at any time point post-stroke, has a cumulative incidence of 55%, and has a presence strongly associated with higher risk for recurrent vascular events, poorer quality of life, poorer functional outcomes, poorer cognitive function, and higher mortality after stroke. Robinson et al. performed a multisite randomized controlled trial for prevention of depression among 176 nondepressed patients enrolled within 3 months of stroke onset. Patients randomized to the placebo arm were significantly ( P < .001) more likely to reach the primary outcome, development of major or minor depression, compared to patients in either of the two active study arms, which were the SSRI escitalopram or problem-solving therapy. Similarly, in the FOCUS study, the SSRI fluoxetine significantly (22%) reduced depression. In a subgroup analysis, cognitive outcomes at 12 months were significantly better among those randomized to SSRI. A separate subgroup analysis found a lower incidence of generalized anxiety disorder with SSRI or problem-solving therapy.

Together, these studies suggest that SSRIs might not improve functional outcomes across all patients with a stroke but might be useful in target subpopulations such as patients whose features are aligned with drug mechanisms of action. SSRIs might also be useful to improve mood and anxiety after stroke.

Dopaminergic

Dopamine is a monoaminergic catecholamine neurotransmitter important to key brain functions that include learning, plasticity, reward, motivation, and movement. Several smaller studies have suggested the potential to promote favorable brain plasticity. A double-blind, placebo-controlled study randomized 53 patients within 6 months of stroke onset to 3 weeks of daily l -Dopa (Sinemet) or placebo coupled with physiotherapy. The primary endpoint, improvement in trunk/leg and arm motor status by the Rivermead Motor Assessment 3 weeks after end of therapy, was significantly better in the Sinemet group.

In contrast, the DARS (Dopamine Augmented Rehabilitation in Stroke) trial was a multicenter, double-blind, placebo-controlled trial with some pragmatic design features that randomized 593 patients recruited 5–42 days post-stroke who could not walk independently to co-careldopa or placebo given prior to rehabilitation therapy. The study found no difference between groups in the primary endpoint, walking independently 8 weeks after randomization. This study had some pragmatic design features, and drug ingestion and rehabilitation therapy were not consistently provided per protocol. Pragmatic study design may be premature for the current stage of dopaminergic drug development.

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here