Skip to content

Rehabilitation and recovery

TL;DR — Organized multidisciplinary stroke-unit care improves survival and independence, and coordinated early supported discharge shortens hospital stay and reduces dependency for selected patients with mild-to-moderate disability (Langhorne 2020, PMID 32324916; Langhorne 2017, PMID 28703869). Rehabilitation should begin after physiological stabilization, but AVERT showed that a higher-dose, very early mobilization protocol begun within 24 hours worsened the odds of a favorable three-month outcome compared with usual stroke-unit care (AVERT Trial Collaboration 2015, PMID 25892679). Recovery is domain-specific: task-specific repetitive practice has modest motor benefit; speech-language therapy improves functional communication; evidence for attention and memory rehabilitation is thinner and often does not translate to daily activity (French 2016, PMID 27841442; Brady 2016, PMID 27245310; Loetscher 2019, PMID 31706263). Technology amplifies practice but rarely replaces skilled therapy: robot-assisted arm training did not improve three-month upper-limb function in RATULS, whereas paired vagus-nerve stimulation plus rehabilitation improved chronic arm impairment in a selected population (Rodgers 2019, PMID 31128926; Dawson 2021, PMID 33894832). Routine fluoxetine does not improve global functional recovery and increases fractures, falls, seizures, or hyponatremia across large pragmatic trials (FOCUS Collaboration 2019, PMID 30528472; AFFINITY Trial Collaboration 2020, PMID 32702334; EFFECTS Trial Collaboration 2020, PMID 32702335).

Recovery is not one endpoint

The Stroke Recovery and Rehabilitation Roundtable distinguishes restitution of the original neural function, compensation through alternative behavior, and adaptation to persistent disability. Mixing these processes makes mechanistic trials uninterpretable (Bernhardt 2017, PMID 28697708).

Domain Impairment-level measure Activity/participation measure Common interpretive error
Motor Fugl-Meyer Assessment, strength, synergies Action Research Arm Test, gait speed, six-minute walk, real-world use Calling compensation “neurological recovery”
Language Naming, comprehension, repetition Functional communication and conversation Treating test-item gain as restored participation
Cognition Attention, memory, executive tests Medication, finance, navigation, return to work Inferring daily independence from a trained cognitive task
Swallowing Instrumental physiology and aspiration Safe oral intake, nutrition, pneumonia Equating diet advancement with normal physiology
Global disability NIHSS early; mRS later Living situation, participation, caregiver dependence Hiding domain-specific gains inside an ordinal global score

Standardized time windows—hyperacute 0–24 hours, acute 1–7 days, early subacute 7 days–3 months, late subacute 3–6 months, chronic >6 months—reduce comparison error across recovery studies (Bernhardt 2017, PMID 28697708).

Organized care is itself an intervention

The 2020 Cochrane network meta-analysis found that organized inpatient stroke-unit care reduces death, dependency, and need for institutional care compared with general-ward care; the benefit is attributable to coordinated specialist processes rather than one device or drug (Langhorne 2020, PMID 32324916).

Core work includes early complication prevention, swallowing and nutrition management, positioning and mobility, continence, communication access, mood/cognition screening, caregiver training, discharge planning, and secondary prevention. The 2026 AHA/ASA adult rehabilitation guideline supersedes the 2016 document and extends current guidance across comorbidities, falls and fracture prevention, technology, caregiver support, participation, work, and driving (Richards 2026, PMID 42657476; Winstein 2016, PMID 27145936).

Early supported discharge (ESD) transfers rehabilitation home through a coordinated multidisciplinary team. Across 17 trials and 2,422 participants, ESD reduced the composite of death or long-term dependency (OR 0.80, 95% CI 0.67–0.95), shortened hospital stay by about six days, and produced the clearest benefit in mild-to-moderate disability when delivered by a coordinated team (Langhorne 2017, PMID 28703869).

Model Evidence-supported use Boundary
Dedicated stroke unit Broad acute-stroke population Requires coordinated multidisciplinary protocols, not merely geographic cohorting (PMID 32324916)
Inpatient rehabilitation Medically stable patients needing intensive interdisciplinary therapy Intensity must remain tolerable and goal-linked (PMID 27145936)
Early supported discharge Selected mild-to-moderate disability with safe home setting Uncoordinated early discharge is not ESD (PMID 28703869)
Telerehabilitation Access bridge or hybrid follow-up Evidence does not establish superiority to matched in-person therapy (Laver 2020, PMID 32002991)

Timing and dose: more is not always better

AVERT randomized 2,104 patients within 24 hours of ischemic or hemorrhagic stroke to frequent, higher-dose very early mobilization plus usual care or usual stroke-unit care. Favorable mRS outcome at three months was less likely with the intensive protocol (46% vs 50%; adjusted OR 0.73, 95% CI 0.59–0.90), despite fewer immobility complications being the motivating hypothesis (AVERT Trial Collaboration 2015, PMID 25892679).

The lesson is not bed rest. It is that treatment dose has timing-dependent biology: early upright activity, session frequency, total minutes, stroke severity, and physiological tolerance interact. “Early mobilization” without a protocol description is not reproducible.

Repetitive task training pooled 33 trials with 1,853 participants and produced modest improvements in arm function, hand function, walking distance, and functional ambulation; effects were sustained up to six months for some outcomes, but trial quality and intervention heterogeneity limited certainty (French 2016, PMID 27841442).

Fitness training evidence supports cardiorespiratory and mixed training for walking speed, walking capacity, balance, and fitness, but evidence for mortality, dependence, and durable participation is insufficient; adherence and adverse-event reporting remain inconsistent (Saunders 2020, PMID 32196635).

Upper-limb recovery

EXCITE randomized 222 people three to nine months after first stroke who retained some wrist and finger extension. Two weeks of constraint-induced movement therapy improved timed task performance and self-reported use versus usual care, with gains persisting to 12 months; the eligibility requirement means this is not evidence for a plegic hand (Wolf 2006, PMID 17077374).

Intervention Population/dose Quantitative finding Interpretation
Constraint-induced movement therapy 222 selected patients, 3–9 mo, residual hand movement Significant WMFT and Motor Activity Log improvement maintained 12 mo Effective for a selected motor-capable subgroup (Wolf 2006, PMID 17077374)
Repetitive task training 33 trials, 1,853 participants Small-to-moderate domain effects; heterogeneity substantial Task practice is a principle, not one protocol (French 2016, PMID 27841442)
Robot-assisted training RATULS, 770 participants, 36 sessions/12 wk No significant improvement in prespecified ARAT success at 3 mo vs usual care More repetitions did not yield superior activity outcome (Rodgers 2019, PMID 31128926)
Mirror therapy 62 studies, 1,982 participants Moderate-quality evidence for motor-function and ADL improvement; pain evidence low quality Cheap adjunct; blinding and small-study effects remain concerns (Thieme 2018, PMID 29993119)
Paired vagus-nerve stimulation 108 chronic ischemic-stroke participants Fugl-Meyer upper-extremity change 5.0 vs 2.4 points after in-clinic therapy; difference 2.6 (95% CI 1.0–4.2) Device augments, rather than replaces, structured practice (Dawson 2021, PMID 33894832)

RATULS is a useful negative result: robot training and enhanced repetitive therapy delivered many sessions, but neither improved the primary three-month ARAT success threshold versus usual care, although secondary impairment measures changed (Rodgers 2019, PMID 31128926).

VNS-REHAB enrolled people at least nine months after ischemic stroke with moderate-to-severe arm weakness and preserved minimum movement. Active paired stimulation improved impairment and the 90-day clinically meaningful response rate after therapy relative to sham; surgery, selection, durability, and scalability constrain generalization (Dawson 2021, PMID 33894832).

Walking, balance, and aerobic capacity

LEAPS randomized 408 people two months after stroke with walking speed <0.8 m/s to early or late body-weight-supported treadmill training or a progressive home exercise program. At one year, 52% improved functional walking ability, with no significant difference among groups; early locomotor training produced more dizziness/faintness and multiple falls in severely impaired participants (Duncan 2011, PMID 21612471).

Electromechanical/robot-assisted gait training plus physiotherapy increased the odds of independent walking compared with physiotherapy alone, particularly in people who were nonambulatory and treated early, but did not clearly increase walking velocity or capacity; devices should target the subgroup and outcome they can plausibly change (Mehrholz 2020, PMID 33091160).

Fitness training should be separated from gait-task training. Aerobic conditioning targets cardiorespiratory reserve, mixed training can improve mobility, and resistance training targets strength; trial comparisons are weakened when all are called “exercise” (Saunders 2020, PMID 32196635).

Aphasia and communication

The Cochrane aphasia review included 57 randomized comparisons and 3,002 participants. Speech-language therapy improved functional communication, reading, writing, and expressive language compared with no therapy, but high-intensity/high-dose therapy also had higher dropout; delivery must be tolerable and communication goals meaningful (Brady 2016, PMID 27245310).

Communication access is a safety intervention: aphasia can prevent valid consent, pain reporting, mood assessment, medication teaching, and participation in decisions. Supported conversation, trained communication partners, and accessible written/visual materials are part of rehabilitation rather than optional courtesy (Winstein 2016, PMID 27145936).

Recovery continues in the chronic phase for some people, but spontaneous change, test–retest learning, therapy specificity, and compensatory strategies must be separated in trial design (Bernhardt 2017, PMID 28697708).

Cognition, neglect, mood, and fatigue

Memory rehabilitation studies show short-term improvement in subjective memory reports but insufficient evidence for objective memory, independence, mood, or quality of life, and no established long-term benefit (das Nair 2016, PMID 27581994).

Attention rehabilitation may improve divided attention immediately after treatment, but evidence is insufficient for global attention, functional independence, mood, or quality of life, and durability is uncertain (Loetscher 2019, PMID 31706263).

Post-stroke fatigue is common and disabling, yet intervention trials remain small and heterogeneous. It must be distinguished from sleep disorder, depression, medication effects, infection, anemia, cardiopulmonary limitation, and effort associated with cognitive or motor impairment (Kennedy 2018, PMID 29525448).

Depression treatment and recovery enhancement are different indications. An antidepressant may be appropriate for diagnosed depression; this does not justify routine pharmacological “neurorecovery” treatment in patients without depression.

Fluoxetine: reversal of an early signal

FLAME, a small trial of 118 patients with ischemic stroke and motor deficit, reported greater 90-day Fugl-Meyer motor improvement with fluoxetine 20 mg/day than placebo, generating a plausible neuroplasticity signal (Chollet 2011, PMID 21216670).

Three large pragmatic trials then tested six months of routine fluoxetine 20 mg begun 2–15 days after stroke:

Trial n/setting Functional result Benefits Harms
FOCUS 3,127, UK No mRS improvement at 6 mo; common OR 0.951 (95% CI 0.839–1.079) Less new depression More fractures (FOCUS Collaboration 2019, PMID 30528472)
AFFINITY 1,280, Australia/NZ/Vietnam No mRS improvement No recovery benefit More falls, fractures, epileptic seizures (AFFINITY Trial Collaboration 2020, PMID 32702334)
EFFECTS 1,500, Sweden No mRS improvement Less new depression More fractures and hyponatremia (EFFECTS Trial Collaboration 2020, PMID 32702335)

The updated Cochrane review concluded that low-risk-of-bias SSRI trials show no reduction in disability or dependency and increased seizures and fractures; routine SSRI prescription to enhance recovery is unsupported (Legg 2021, PMID 34780067).

Prediction should guide planning, not ration therapy

Early motor impairment strongly predicts later motor status, while corticospinal-tract integrity measured by motor-evoked potentials or diffusion imaging can refine prediction in more impaired patients. Clinical assessment alone has limited individual-level accuracy, and biomarker algorithms require calibration across centers and populations (Stinear 2014, PMID 25200094).

Prediction can support goal setting, equipment planning, discharge destination, and trial stratification. It should not be used as a self-fulfilling reason to withhold practice, communication access, prevention of complications, or compensatory training.

Prognostic layer Example What it can contribute Failure mode
Clinical Initial strength, NIHSS, sitting balance Fast baseline stratification Ceiling/floor effects and sedation
Neurophysiology TMS motor-evoked potential Corticospinal functional integrity Availability, contraindications, operator dependence
Imaging Lesion load/corticospinal tract injury Structural pathway integrity Scanner/protocol variability; association mistaken for destiny
Context Cognition, mood, social support, housing Real-world capacity to practice and participate Social disadvantage misclassified as biological non-recovery

Telerehabilitation and access

The 2020 Cochrane review found insufficient evidence that telerehabilitation improves activities of daily living compared with in-person rehabilitation or usual care; studies were small and heterogeneous, with limited evidence on cost and adverse events (Laver 2020, PMID 32002991).

Telerehabilitation can reduce travel and expand specialist reach, but excludes people without connectivity, devices, caregiver support, safe space, or adequate communication interfaces. Hybrid service trials should measure reach, adherence, caregiver time, and inequity—not only impairment scores.

Open questions

  • What is the dose–timing response surface for mobility and task practice in the first week after stroke? AVERT shows that one high-dose protocol within 24 hours can harm, not where the optimum lies (PMID 25892679).
  • Which biological and behavioral markers predict response to a specific therapy rather than merely spontaneous recovery (Stinear 2014, PMID 25200094; Bernhardt 2017, PMID 28697708)?
  • Can adaptive dosing based on fatigue and within-session performance outperform fixed hours without reducing total meaningful practice (French 2016, PMID 27841442)?
  • Does paired VNS produce durable participation gains, and which lesion/pathway phenotypes account for response heterogeneity (Dawson 2021, PMID 33894832)?
  • How can aphasia, cognition, neglect, and caregiver burden be integrated into trials whose primary endpoint is motor impairment (Brady 2016, PMID 27245310; Loetscher 2019, PMID 31706263)?
  • Which telerehabilitation designs expand access rather than transferring work and cost to patients and families (Laver 2020, PMID 32002991)?

References

  1. Langhorne P, Ramachandra S. Organised inpatient (stroke unit) care for stroke: network meta-analysis. Cochrane Database Syst Rev. 2020;4:CD000197. PMID 32324916
  2. Winstein CJ, Stein J, Arena R, et al. Guidelines for adult stroke rehabilitation and recovery. Stroke. 2016;47:e98-e169. PMID 27145936 2a. Richards LG, et al. 2026 Guideline for Adult Stroke Rehabilitation and Recovery: A Guideline From the American Heart Association and American Stroke Association. Stroke. 2026. PMID 42657476
  3. Langhorne P, Baylan S. Early supported discharge services for people with acute stroke. Cochrane Database Syst Rev. 2017;7:CD000443. PMID 28703869
  4. AVERT Trial Collaboration. Efficacy and safety of very early mobilisation within 24 h of stroke onset (AVERT). Lancet. 2015;386:46-55. PMID 25892679
  5. Bernhardt J, Hayward KS, Kwakkel G, et al. Agreed definitions and a shared vision for new standards in stroke recovery research. Int J Stroke. 2017;12:444-450. PMID 28697708
  6. French B, Thomas LH, Coupe J, et al. Repetitive task training for improving functional ability after stroke. Cochrane Database Syst Rev. 2016;11:CD006073. PMID 27841442
  7. Saunders DH, Sanderson M, Hayes S, et al. Physical fitness training for stroke patients. Cochrane Database Syst Rev. 2020;3:CD003316. PMID 32196635
  8. Wolf SL, Winstein CJ, Miller JP, et al. Effect of constraint-induced movement therapy on upper extremity function 3 to 9 months after stroke: EXCITE. JAMA. 2006;296:2095-2104. PMID 17077374
  9. Rodgers H, Bosomworth H, Krebs HI, et al. Robot assisted training for the upper limb after stroke (RATULS). Lancet. 2019;394:51-62. PMID 31128926
  10. Thieme H, Morkisch N, Mehrholz J, et al. Mirror therapy for improving motor function after stroke. Cochrane Database Syst Rev. 2018;7:CD008449. PMID 29993119
  11. Dawson J, Liu CY, Francisco GE, et al. Vagus nerve stimulation paired with rehabilitation for upper limb motor function after ischaemic stroke. Lancet. 2021;397:1545-1553. PMID 33894832
  12. Duncan PW, Sullivan KJ, Behrman AL, et al. Body-weight-supported treadmill rehabilitation after stroke. N Engl J Med. 2011;364:2026-2036. PMID 21612471
  13. Mehrholz J, Thomas S, Kugler J, et al. Electromechanical-assisted training for walking after stroke. Cochrane Database Syst Rev. 2020;10:CD006185. PMID 33091160
  14. Brady MC, Kelly H, Godwin J, et al. Speech and language therapy for aphasia following stroke. Cochrane Database Syst Rev. 2016;2016:CD000425. PMID 27245310
  15. das Nair R, Cogger H, Worthington E, et al. Cognitive rehabilitation for memory deficits after stroke. Cochrane Database Syst Rev. 2016;9:CD002293. PMID 27581994
  16. Loetscher T, Potter KJ, Wong D, et al. Cognitive rehabilitation for attention deficits following stroke. Cochrane Database Syst Rev. 2019;2019:CD002842. PMID 31706263
  17. Kennedy C, Kidd L. Interventions for post-stroke fatigue: a Cochrane review summary. Int J Nurs Stud. 2018;85:136-137. PMID 29525448
  18. Chollet F, Tardy J, Albucher JF, et al. Fluoxetine for motor recovery after acute ischaemic stroke (FLAME). Lancet Neurol. 2011;10:123-130. PMID 21216670
  19. FOCUS Trial Collaboration. Effects of fluoxetine on functional outcomes after acute stroke (FOCUS). Lancet. 2019;393:265-274. PMID 30528472
  20. AFFINITY Trial Collaboration. Safety and efficacy of fluoxetine on functional outcome after acute stroke (AFFINITY). Lancet Neurol. 2020;19:651-660. PMID 32702334
  21. EFFECTS Trial Collaboration. Safety and efficacy of fluoxetine on functional recovery after acute stroke (EFFECTS). Lancet Neurol. 2020;19:661-669. PMID 32702335
  22. Legg LA, Rudberg AS, Hua X, et al. Selective serotonin reuptake inhibitors (SSRIs) for stroke recovery. Cochrane Database Syst Rev. 2021;11:CD009286. PMID 34780067
  23. Stinear CM, Byblow WD, Ward SH. An update on predicting motor recovery after stroke. Ann Phys Rehabil Med. 2014;57:489-498. PMID 25200094
  24. Laver KE, Adey-Wakeling Z, Crotty M, et al. Telerehabilitation services for stroke. Cochrane Database Syst Rev. 2020;1:CD010255. PMID 32002991