Acupuncture

Evidence Reviewed as of before: 11-08-2017
Author(s)*: Tatiana Ogourtsova, PhD(c) OT; Marc-André Roy, MSc; Nicol Korner-Bitensky, PhD; Robert Teasell, MD; Norine Foley, BASc; Sanjit Bhogal, MSc; Jamie Bitensky, MSc OT; Mark Speechley, MD; Annabel McDermott, OT
Patient/Family Information Table of contents

Introduction

Acupuncture is an ancient Chinese therapy involving the stimulation of specific trigger points along the body’s 18 meridian lines to help regulate the flow of Qi (energy). The meridian lines represent the normal flow of Qi through the body. It is believed that when this energy is disrupted, disease ensues. The use of thin metal needles or other acupuncture techniques is proposed to conduct Qi through its correct paths. The trigger points used are areas of the skin where Qi flows close to the surface and thus can be reached by the various acupuncture therapies.

While the exact mechanisms are not well defined in terms of Western medicine, there are biological responses that occur directly at the stimulus point and indirectly at other parts of the body. In addition to the use of fine needles, other methods of acupuncture include:

  • electro-acupuncture (current through the needles),
    L'électro-acupuncture
    Pictures courtesy of Ricardo Miranda,L.Ac
  • cupping (suction cups on trigger points),
    les ventouses
    Pictures courtesy of Ricardo Miranda,L.Ac
  • acupressure using trigger points (applying pressure with fingers or instruments),
  • reflexology (using pressure on the soles of the feet and inferior ankle to stimulate various parts of the body),
  • moxibustion (heat at trigger points, often combined with needles),la moxibustion
    la moxibustion
    Pictures courtesy of Ricardo Miranda,L.Ac
  • auriculotherapy (stimulating trigger points on the ear to affect other parts of the body),
  • laserpuncture and sonopuncture (using sound waves over trigger points).

Acupuncture has been used to treat many types of health problems and in the past decade has been advocated by some for the treatment of stroke. Recently, a number of studies have explored the use of acupuncture in stroke rehabilitation.

Patient/Family Information

Author: Tatiana Ogourtsova, PhD(c) OT, Marc-André Roy, MSc

What is acupuncture?

Acupuncture comes from ancient Chinese medicine. It has been used to treat pain in China for about 3000 years. The Chinese explanation involves Qi (pronounced Chee), an energy that flows through the body. The belief is that when this Qi is balanced (Yin and Yang), then the body is healthy. Qi flows through different lines within your body called “meridians”. With the most common form of acupuncture, an expert puts very small needles into specific areas of your body where Qi flows close to the surface of the skin.

There is some evidence that acupuncture works after operations to stop pain, after chemotherapy to stop feeling sick and vomiting, during pregnancy to stop feeling sick and after dental surgery for dental pain. It has also been used to treat headaches, tennis elbow, fibromyalgia (general muscle pain), low back pain, carpal tunnel syndrome and asthma.

While we are not sure exactly how it works, 3 possible explanations have been given:

  • Acupuncture blocks pain from traveling in your nerves
  • Acupuncture causes your body to make chemicals that prevent pain
  • Acupuncture opens or closes your veins and arteries in important areas of the body

Are there different kinds of acupuncture?

The most popular acupuncture is performed by putting thin metal needles into the skin. Other forms of acupuncture include:

  • electro-acupuncture, which again uses needles through which very small electrical currents are passed;L'électro-acupuncturePictures courtesy of Ricardo Miranda,L.Ac
  • auriculotherapy, which uses either needles or pressure on different spots of the ear which are trigger points for the entire body;
  • moxibustion, which uses heat at different spots on the body;moxibustion moxibustionPictures courtesy of Ricardo Miranda,L.Ac
  • sonopuncture, which uses sound waves at different spots on the body
  • cupping, which uses suction cups over areas such as the back or the legs to pull blood and other fluids in the area under the skin;cuppingPictures courtesy of Ricardo Miranda,L.Ac
  • acupressure, which uses pressure on different spots on the body;
  • reflexology, which uses pressure under the feet or the back part of the ankles.

Why use acupuncture after a stroke?

Acupuncture has been used after a stroke to treat spasticity (stiffness of muscles caused by the stroke), loss of function, loss of mobility, depression, aphasia (loss of speaking and writing skills), hemiplegia (loss of feeling and/or power to move one side of the body) and for pain reduction.

Does it work for stroke?

Experts have done some experiments to compare acupuncture with other treatments to see whether acupuncture helps people who have had a stroke.

In individuals with ACUTE stroke (< 4 weeks after stroke)
Thirteen high quality studies and 7 fair quality studies found that acupuncture:

  • Was not more helpful than other treatments for improving cognitive skills (e.g. memory, language); mood (e.g. depression); self-care skills (e.g. dressing, shopping); quality of life; physical skills (e.g. strength, range of motion, sensation, motor function of arms and legs); or mobility (e.g. balance, walking speed); but
  • Was more helpful than the usual treatment for improving swallowing skills and swallowing safety.

In individuals with SUBACUTE stroke (1 to 6 months after stroke)
One high quality study found that acupuncture:

  • Was not more helpful than pretend acupuncture for improving range of motion.

In individuals with CHRONIC stroke (> 6 months after stroke)
Three high quality studies and 1 low quality study found that acupuncture:

  • Was not more helpful than pretend acupuncture for improving mood (e.g. depression); self-care skills (e.g. dressing); mobility (e.g. walking endurance); physical skills (e.g. spasticity, range of motion, strength) or pain.

What can I expect?

Most people find that having acupuncture treatment causes very little pain, if any. In most cases you feel the needle going in, but it doesn’t hurt. Some people say they feel cramping, heaviness or tingling at the needle site or up the “meridian”.

The acupuncturist may use other treatments once the needles are in place. This depends on his/her training.

Side effects/risks?

As with any other use of needles, sanitation is very important to not spread germs. All acupuncturists should use new, individually packaged, disposable needles. If these are not used, don’t agree to treatment.

There is little risk related to acupuncture if done by a qualified professional. Side effects could include dizziness, feeling sick and feeling tired after treatment. There could also be a little bleeding at the needle site and some slight bruising. There is always a slight risk of infection when putting needles in the skin.

Who provides the treatment?

Acupuncture should be practiced by a trained health professional. For example, in Quebec (Canada) the practice of acupuncture is regulated by a professional Order and only members of the Order can practice it. Different health care professionals such as physicians and physiotherapists may use the trigger point needle technique as part of their treatment.

How many treatments?

This depends on the reason you are getting acupuncture. You should discuss the treatment plan with the acupuncturist before starting treatment. You might receive anywhere from one to 15 treatment sessions.

How much does it cost? Does insurance pay for It?

Acupuncture is not paid for by provincial insurance plans. However, it is covered by some private insurance plans. The cost for each session may vary from $40.00 to $90.00.

Is acupuncture for me?

Although the benefits of acupuncture have been talked about for hundreds of years, there is no strong scientific evidence that it works to reduce spasticity, loss of function, loss of mobility, depression, aphasia or pain. Yet, there are some people who say they have found it helpful.

Clinician Information

Note: When reviewing the findings, it is important to note that they are always made according to randomized clinical trial (RCT) criteria – specifically as compared to a control group. To clarify, if a treatment is “effective” it implies that it is more effective than the control treatment to which it was compared. Non-randomized studies are no longer included when there is sufficient research to indicate strong evidence (level 1a) for an outcome.

The current module includes 35 RCTs including 25 high quality RCTs, nine fair quality RCTs and one poor quality RCT. Numerous outcome measures were used throughout studies and outcomes include balance, cognitive function, dexterity, depression, functional independence, motor function, quality of life, swallowing function, etc. Studies conducted with patients in one phase of stroke recovery, be it the acute, subacute, or chronic phases of stroke recovery, predominantly reported that acupuncture was not more effective than comparison interventions in improving most outcomes (with the exception of dysphagia and swallowing function). By comparison, studies that included patients across stages of stroke recovery (e.g. patients in the acute or subacute phases of stroke recovery) generally reported that acupuncture was more effective than comparison interventions in improving outcomes (especially those related to cognitive function, health related quality of life, insomnia, mobility and swallowing function).

Results Table

View results table

Outcomes

Acute Phase

Balance
Not effective
1b

One high quality RCT (Hsieh et al., 2007) and one fair quality RCT (Johansson et al., 1993) investigated the effect of acupuncture on balance in patients with acute stroke.

The high quality RCT (Hsieh et al., 2007) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Balance was measured by the Fugl-Meyer Assessment (FMA – Balance) during treatment (2 weeks), at post-treatment (4 weeks), and follow-up (3 and 6 months post-stroke). No significant between-group differences were found at any time point.

The fair quality RCT (Johansson et al., 1993) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Balance was measured by the modified Chart for Motor Capacity Assessment – Balance at mid-treatment (1 month post-stroke), and follow-up (3 months post-stroke); measures were not taken at post-treatment (10 weeks). Significant between-group differences were found at both time points, favoring electroacupuncture vs. no acupuncture.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (conventional rehabilitation with no acupuncture) in improving balance in patients with acute stroke.
Note: 
However, one fair quality RCT found that acupuncture was more effective than no acupuncture in improving balance in patients with acute stroke; the studies differed in duration of the intervention (4 weeks vs. 10 weeks) and outcome measures used to assess balance.

Cognitive function
Not effective
1a

Two high quality RCTs (Rorsman & Johansson, 2006Chen et al., 2016) investigated the effect of acupuncture on cognitive function in patients with acute stroke.

The first high quality RCT (Rorsman & Johansson, 2006) randomized patients to receive acupuncture (including electroacupuncture), high intensity/low frequency transcutaneous electrical nerve stimulation TENS) or low intensity (subliminal)/high frequency TENS. Cognitive function was measured by the Mini-Mental State Examination (MMSE) at follow-up (3 and 12 months post-stroke); measures were not taken at post-treatment (10 weeks). No significant between-group differences were found at either time point.

The second high quality RCT (Chen et al., 2016) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Cognitive function was measured by the MMSE and the Montreal Cognitive Assessment (MOCA) at baseline, at post-treatment (3 weeks) and at follow-up (7 weeks). There were no significant between-group differences on either measure at post-treatment. There were significant differences in change scores on both measures from baseline to follow-up, favoring acupuncture vs. no acupuncture.

Conclusion: There is strong evidence (Level 1a) from 2 high quality RCTs that acupuncture is not more effective than comparison interventions (TENS, conventional rehabilitation with no acupuncture) for improving cognitive function in patients with acute stroke.
Note: 
However, one of the high quality RCTs reported gains in favour of acupuncture at follow-up.

Depression
Not effective
1b

One high quality RCT (Rorsman & Johansson, 2006) investigated the effect of acupuncture on depression in patients with acute stroke. The high quality RCT randomized patients to receive acupuncture (including electroacupuncture), high intensity/low frequency TENS or low intensity (subliminal)/high frequency TENS. Depression was measured at follow-up (3- and 12-months post-stroke) by the Hospital Anxiety and Depression Scale and the Comprehensive Psychiatric Rating Scale; measures were not taken at post-treatment (10 weeks). No significant between-group differences were found on either measure at either follow-up time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than comparison interventions (high intensity/low frequency TENS, low intensity/high frequency TENS) in improving depression in patients with acute stroke.

Dexterity
Not effective
1a

Two high quality RCTs (Johansson et al., 2001Park et al., 2005) investigated the effect of acupuncture on dexterity in patients with acute stroke.

The first high quality RCT (Johansson et al., 2001) randomized patients to receive electroacupuncture, high intensity/low frequency TENS or low intensity (subliminal)/high frequency TENS; all groups received conventional rehabilitation. Dexterity was measured by the Nine Hole Peg Test (NHPT) at follow-up (3 and 12 months post-stroke); measures were not taken at post-treatment (10 weeks). No significant between group differences were found at either follow-up time point.

The second high quality RCT (Park et al., 2005) randomized patients to receive manual acupuncture or sham acupuncture. Dexterity was measured by the NHPT at post-treatment (2 weeks). No significant between-group differences were found.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that acupuncture is not more effective than comparison interventions (TENS, sham acupuncture) in improving dexterity in patients with acute stroke.

Dysphagia
Effective
1b

One high quality RCT (Xia et al., 2016) investigated the effect of acupuncture on functional severity of dysphagia in patients with acute stroke and subsequent dysphagia. This high quality RCT randomized patients to receive acupuncture or no acupuncture; both groups received standard swallowing training. Functional severity of dysphagia was measured by the Dysphagia Outcome and Severity Scale at post-treatment (4 weeks). Significant between-group differences were found, favoring acupuncture vs. no acupuncture.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that swallowing training with acupuncture is more effective than a comparison intervention (swallowing training with no acupuncture) in improving functional severity of dysphagia in patients with acute stroke and subsequent dysphagia.

Functional independence
Not effective
1a

Ten high quality RCTs (Gosman-Hedstrom et al., 1998Johansson et al., 2001Sze et al., 2002Park et al., 2005Hsieh et al., 2007Hopwood et al., 2008Zhu et al., 2013Li et al., 2014Liu et al., 2016Xia et al., 2016) and six fair quality RCTs (Hu et al., 1993Johansson et al., 1993Wong et al., 1999Pei et al., 2001Min et al., 2008Wang et al., 2014) investigated the effect of acupuncture on functional independence in patients with acute stroke.

The first quality RCT(Gosman-Hedstrom et al., 1998) randomized patients to receive deep electroacupuncture, superficial acupuncture or no acupuncture; all groups received conventional rehabilitation. Functional independence was measured by the Barthel Index (BI) and Sunnaas Index at post-treatment (3 months) and at follow-up (12 months). No significant between-group differences were found on any measure at either time point.

The second high quality RCT(Johansson et al., 2001) randomized patients to receive electroacupuncture, high intensity/low frequency TENS or low intensity (subliminal)/high frequency TENS; all groups received conventional rehabilitation. Functional independence was measured by the BI at follow-up (3 and 12 months post-stroke); measures were not taken at post-treatment (10 weeks). No significant between group differences were found at either follow-up time point.

The third high quality RCT(Sze et al., 2002) randomized patients to receive manual acupuncture or no acupuncture; both groups received conventional rehabilitation. Functional independence was measured by the BI and the Functional Independence Measure (FIM) at post-treatment (10 weeks). No significant between-group differences were found on any measure.

The forth high quality RCT (Park et al., 2005) randomized patients to receive manual acupuncture or sham acupuncture. Functional independence was measured by the BI at post-treatment (2 weeks). No significant between-group differences were found.

The fifth high quality RCT (Hsieh et al., 2007) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Functional independence was measured by the FIM (total, self-care, social, mobility, locomotion, sphincter control, communication) during treatment (2 weeks), at post-treatment (4 weeks), and follow-up (3- and 6-months post-stroke). A significant between-group difference was found on only one score (FIM – social) during treatment (2 weeks), favoring electroacupuncture vs. no acupuncture. There were no other significant between-group differences on any measure, at any time point.

The sixth high quality RCT (Hopwood et al., 2008) randomized patients to receive electroacupuncture or placebo electroacupuncture. Functional independence was measured by the BI during treatment (3 weeks) and at several follow-up time points (6, 12, 25, and 52 weeks); measures were not taken at post-treatment (4 weeks). No significant between-group differences were found at any time point.

The seventh high quality RCT(Zhu et al., 2013) randomized patients to receive acupuncture or no acupuncture; both groups received conventional rehabilitation. Functional independence was measured by the BI at mid-treatment (1 month), post-treatment (3 months) and follow-up (6 months). No significant between-group differences were found at any time point.

The eighth high quality RCT (Li et al., 2014) randomized patients to receive verum acupuncture or sham acupuncture. Functional independence was measured by the modified BI and the modified Rankin Scale (mRS) at baseline, at mid-treatment (2 weeks), post-treatment (4 weeks), and follow-up (12 weeks). Significant between-group differences were found at post-treatment (both measures) and at follow-up (BI only), favoring verum acupuncture vs. sham acupuncture.
Note: Differences at post-treatment reflect change scores from baseline to post-treatment; differences at follow-up reflect scores at that time point as well as change scores from baseline to follow-up.

The ninth high quality RCT (Liu et al., 2016) randomized patients to receive manual acupuncture or no acupuncture. Functional independence was measured by the BI,the mRS and the FIM at post-treatment (2 weeks: FIM) and at follow-up (3 weeks: FIM; 1 month: FIM; 3 months: MRS, BI). No significant between-group differences were found on any measure at any time point.

The tenth high quality RCT (Xia et al., 2016) randomized patients to receive acupuncture or no acupuncture; both groups received standard swallowing training. Functional independence was measured by the modified BI at post-treatment (4 weeks). Significant between group differences were found, favoring acupuncture vs. no acupuncture.

The first fair quality RCT (Hu et al., 1993) randomized patients to receive acupuncture or no acupuncture; both groups received conventional rehabilitation. Functional independence was measured by the BI at post-treatment (4 weeks) and at follow-up (3 months). No significant between-group differences were found at either time point.

The second fair quality RCT (Johansson et al., 1993) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Functional independence was measured by the BI at mid-treatment (1 month post-stroke) and at two follow-up timepoints (3 and 12 months post-stroke); measures were not taken at post-treatment (10 weeks). Significant between-group differences were found at all time points, favoring electroacupuncture vs. no acupuncture.

The third fair quality RCT (Wong et al., 1999) randomized patients to receive electroacupuncture or no acupuncture. Functional independence was measured by the FIM (total, self-care, locomotion, sphincter control, transfers, communication, social interaction) at post-treatment (2 weeks). Significant between-group differences were found (FIM total, self-care, locomotion), favoring electroacupuncture vs. no acupuncture.

The forth fair quality RCT (Pei et al., 2001) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Functional independence was measured by the BI mid-treatment (1 and 2 weeks), at post-treatment (4 weeks) and at follow-up (3 months). Significant between-group differences were found at all time points, favoring electroacupuncture vs. no acupuncture.

The fifth fair quality RCT (Min et al., 2008) randomized patients to receive acupuncture or no acupuncture; both groups received conventional rehabilitation. Functional independence was measured by the modified BI at post-treatment (3 months). Significant between-group differences were found, favoring acupuncture vs. no acupuncture.

The sixth fair quality RCT (Wang et al., 2014) randomized patients to receive electroacupuncture or no electroacupuncture; both groups received conventional rehabilitation. Functional independence was measured by the BI at follow-up (3 and 6 months); measures were not taken at post-treatment (4 weeks). Significant between-group differences were found at 6-month follow-up only, favoring electroacupuncture vs. no electroacupuncture.

Conclusion: There is strong evidence (Level 1a) from eight high quality RCTs and one fair quality RCT that acupuncture is not more effective than comparison interventions (superficial acupuncture, no acupuncture, TENS, conventional rehabilitation, sham or placebo acupuncture) in improving functional independence in patients with acute stroke.
Note:
However, two high quality RCTs and five fair quality RCTs found that acupuncture was more effective than comparison interventions (sham acupuncture, standard swallowing training, no acupuncture, conventional rehabilitation) in improving functional independence in patients with acute stroke.

Health-related quality of life (HRQoL)
Not effective
1a

Five high quality RCTs (Gosman-Hedstrom et al., 1998; Johansson et al., 2001; Park et al., 2005; Hopwood et al., 2008Li et al., 2014) and one fair quality RCT (Johansson et al., 1993) investigated the effect of acupuncture on health-related quality of life (HRQoL) in patients with acute stroke.

The first high quality RCT (Gosman-Hedstrom et al., 1998) randomized patients to receive deep electroacupuncture, superficial acupuncture or no acupuncture; all groups received conventional rehabilitation. HRQoL was measured by the Nottingham Health Profile (NHP – energy level, pain, emotional reaction, sleep, social isolation, physical abilities) at post-treatment (3 months) and at follow-up (12 months). There were no significant between-group differences at post-treatment; there was a significant between-group difference in one component of HRQoL (physical abilities) at follow-up, favoring deep electroacupuncture vs. no acupuncture.

The second high quality RCT (Johansson et al., 2001) randomized patients to receive electroacupuncture, high intensity/low TENS or low intensity (subliminal)/high frequency TENS; all groups received conventional rehabilitation. HRQoL was measured by the NHP at follow-up (3 and 12 months post-stroke); measures were not taken at post-treatment (10 weeks). No significant between group differences were found at both follow-up time points.

The third high quality RCT (Park et al., 2005) randomized patients to receive manual acupuncture or sham acupuncture. HRQoL was measured by the EuroQoL (EuroQoL5 – Visual Analogue Scale) at post-treatment (2 weeks). No significant between-group differences were found.

The forth high quality RCT (Hopwood et al., 2008) randomized patients to receive electroacupuncture or placebo electroacupuncture. HRQoL was measured by the NHP during treatment (3 weeks) and at follow-up (6, 12, 25, and 52 weeks). There was a significant between-group difference in one score (NHP – Energy) during treatment and at all follow-up time points, favoring electroacupuncture vs. placebo acupuncture.

The fifth high quality RCT (Li et al., 2014) randomized patients to receive verum acupuncture or sham acupuncture. HRQoL was measured by the Stroke Specialization Quality of Life Scale (SS-QoL) at baseline, at mid-treatment (2 weeks), post-treatment (4 weeks), and at follow-up (12 weeks). Significant between-group differences were found at post-treatment and at follow-up, favoring verum acupuncture vs. sham acupuncture.
Note: Differences at post-treatment reflect change scores from baseline to post-treatment; differences at follow-up reflect scores at that time point as well as change scores from baseline to follow-up.

The fair quality RCT (Johansson et al., 1993) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. HRQoL was measured by the modified NHP at follow-up (3, 6 and 12 months post-stroke); measures were not taken at post-treatment (10 weeks). There were significant between-group differences in some components of HRQoL at 3 months post-stroke (energy, mobility, emotion, social isolation), at 6 months post-stroke (energy, mobility, emotion, social isolation, sleep), and at 12 months post-stroke (mobility, emotion), favoring electroacupuncture vs. no acupuncture.

Conclusion: There is strong evidence (Level 1a) from four high quality RCTs that acupuncture is not more effective than comparison interventions (superficial acupuncture, no acupuncture, TENS, sham or placebo acupuncture) in improving health-related quality of life in patients with acute stroke.
Note
: However, one high quality RCT found that acupuncture was more effective than a comparison intervention (sham acupuncture); this study used the SS-QoL to measure quality of life, rather than the NHP used by most other studies. In addition, one fair quality RCT found that acupuncture was more effective than no acupuncture in improving some components of the health-related quality of life.

Instrumental activities of daily living (IADLs)
Not effective
1b

One high quality RCT (Park et al., 2005) investigated the effect of acupuncture on IADLs in patients with acute stroke. This high quality RCT randomized patients to receive manual acupuncture or sham acupuncture. IADLs were measured by the Nottingham Extended ADL scale at post-treatment (2 weeks). No significant between-group differences were found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (sham acupuncture) in improving IADLs in patients with acute stroke.

Language function
Not effective
1b

One high quality RCT (Rorsman & Johansson, 2006) investigated the effect of acupuncture on language function with acute stroke. This high quality RCT randomized patients to receive acupuncture (including electroacupuncture), high intensity/low frequency TENS or low intensity (subliminal)/high frequency TENS. Language function was measured by the Token Test and FAS Word Fluency Test at follow-up (3 and 12 months post-stroke); measures were not taken at post-treatment (10 weeks). No significant between-group differences were found on any measure at either follow-up time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than comparison interventions (TENS) in improving language function in patients with acute stroke.

Memory
Not effective
1b

One high quality RCT (Rorsman & Johansson, 2006) investigated the effect of acupuncture on memory in patients with acute stroke. This high quality RCT randomized patients to receive acupuncture (including electroacupuncture), high intensity/low frequency TENS or low intensity (subliminal)/high frequency TENS. Memory was measured by the Rey Auditory Verbal Learning Test and Facial Recognition Memory Test at follow-up (3 and 12 months post-stroke); measures were not taken at post-treatment (10 weeks). No significant between-group differences were found on either measure of memory at either time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than comparison interventions (TENS) in improving memory in patients with acute stroke.

Mobility
Not effective
1b

One high quality RCT (Johansson et al., 2001) and one fair quality RCT (Johansson et al., 1993) investigated the effect of acupuncture on mobility in patients with acute stroke.

The high quality RCT (Johansson et al., 2001) randomized patients to receive electroacupuncture, high intensity/low TENS or low intensity (subliminal)/high frequency TENS; all groups received conventional rehabilitation. Mobility was measured by the Rivermead Mobility Index at follow-up (3 and 12 months post-stroke); measures were not taken at post-treatment (10 weeks). No significant between-group differences were found at either follow-up time point.

The fair quality RCT (Johansson et al., 1993) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Mobility was measured by the modified Chart for Motor Capacity Assessment (Walking) at mid-treatment (1 month post-stroke) and at follow-up (3 months post-stroke); measures were not taken at post-treatment (10 weeks). Significant between-group differences were found at both time points, favoring electroacupuncture vs. no acupuncture.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that electroacupuncture is not more effective than comparison interventions (TENS) in improving mobility in patients with acute stroke.
Note: 
However, one RCT found that acupuncture was more effective than no acupuncture in improving mobility in patients with acute stroke.

Motor function
Conflicting
4

Five high quality RCTs (Sze et al., 2002Hsieh et al., 2007Tan et al., 2013Li et al., 2014Liu et al., 2016) and three fair quality RCTs (Johansson et al., 1993Pei et al., 2001Min et al., 2008) investigated the effect of acupuncture on motor function in patients with acute stroke.

The first high quality RCT (Sze et al., 2002) randomized patients to receive manual acupuncture or no acupuncture; both groups received conventional rehabilitation. Motor function measured by the Fugl-Meyer Assessment (FMA) at post-treatment (10 weeks). No significant between-group differences were found.

The second high quality RCT (Hsieh et al., 2007) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Motor function was measured by the FMA (total score) at mid-treatment (2 weeks), post-treatment (4 weeks), and follow-up (3 and 6 months post-stroke). Significant between-group differences were found at mid-treatment, post-treatment and at 3 months post-stroke, favoring electroacupuncture vs. no acupuncture.

The third high quality RCT (Tan et al., 2013) randomized patients to receive electroacupuncture or no electroacupuncture. Motor function was measured by the FMA at post-treatment (14 days). Significant between-group differences were found at post-treatment, favoring electroacupuncture vs. no electroacupuncture.

The fourth high quality RCT (Li et al., 2014) randomized patients to receive verum acupuncture or sham acupuncture. Motor function was measured by the FMA – Upper and Lower Extremity scores combined at baseline, at mid-treatment (2 weeks), at post-treatment (4 weeks), and at follow-up (12 weeks). Significant between-group differences were found at post-treatment and at follow-up, favoring verum acupuncture vs. sham acupuncture.
Note: Differences at post-treatment reflect change scores from baseline to post-treatment; differences at follow-up reflect scores at that time point as well as change scores from baseline to follow-up.

The fifth high quality RCT (Liu et al., 2016) randomized patients to receive manual acupuncture or no acupuncture. Motor function was measured by the FMA at follow-up (1 month); measures were not taken at post-treatment (2 weeks). No significant between-group differences were found.

The first fair quality RCT (Johansson et al., 1993) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Motor function was measured by the modified Chart for Motor Capacity Assessment (motor function) at 1 and 3 months post-stroke (follow-up); measures were not taken at post-treatment (10 weeks). No significant between group differences were found at either time point.

The second fair quality RCT (Pei et al., 2001) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Motor function was measured by the FMA at mid-treatment (1 and 2 weeks), post-treatment (4 weeks) and at follow-up (3 months). Significant between-group differences were found at all time points, favoring electroacupuncture vs. no acupuncture.

The third fair quality RCT (Min et al., 2008) randomized patients to receive acupuncture or no acupuncture; both groups received conventional rehabilitation. Motor function was measured by the FMA at post-treatment (3 months). A significant between-group difference was found at post-treatment, favoring acupuncture vs. no acupuncture.

Conclusion: There is conflicting evidence (Level 4) regarding the effect of acupuncture on motor function. Two high quality RCTs and one fair quality RCT reported that acupuncture is not more effective than no acupuncture, whereas two other high quality RCTs and two fairquality RCTs found that acupuncture was more effective than comparison interventions (no/sham acupuncture) in improving motor function in patients with acute stroke. A fifth high quality RCT also reported of significant differences in change scores at post-treatment and follow-up.
Note:
There was significant variation between studies in type, frequency and duration of acupuncture.

Motor function - lower extremity
Not effective
1a

Three high quality RCTs (Hsieh et al., 2007Zhu et al., 2013Chen et al., 2016) and two fair quality RCTs (Wong et al., 1999Min et al., 2008) investigated the effect of acupuncture on lower extremity motor function in patients with acute stroke.

The first quality RCT (Hsieh et al., 2007) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Lower extremity motor function was measured by the Fugl Meyer Assessment (FMA – hip/knee/ankle motor function, lower extremity coordination and speed) at mid-treatment (2 weeks), post-treatment (4 weeks), and follow-up (3 and 6 months post-stroke). No significant between-group differences were found at any time point.

The second high quality RCT (Zhu et al., 2013) randomized patients to receive acupuncture or no acupuncture; both groups received conventional rehabilitation. Lower extremity motor function was measured by the Fugl-Meyer Assessment – Lower Extremity (FMA-LE) at mid-treatment (1 month), post-treatment (3 months), and at follow-up (6 months). No significant between-group differences were found at any time point.

The third high quality RCT (Chen et al., 2016) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Lower extremity motor function was measured by the FMA-LE at baseline, at post-treatment (3 weeks) and at follow-up (7 weeks). There were no significant differences at post-treatment; there were significant differences in change scores from baseline to follow-up, favoring acupuncture vs. no acupuncture.

The first fair quality RCT (Wong et al., 1999) randomized patients to receive electroacupuncture or no acupuncture. Lower extremity motor function was measured using Brunnstrom’s lower limb motor recovery at post-treatment (2 weeks). Significant between-group differences were found, favoring electroacupuncture vs. no acupuncture.

The second fair quality RCT (Min et al., 2008) randomized patients to receive acupuncture or no acupuncture; both groups received conventional rehabilitation. Lower extremity motor function was measured by the FMA–LE at post-treatment (3 months). Significant between-group difference were found, favoring acupuncture vs. no acupuncture.

Conclusion: There is strong evidence (level 1a) from 3 high quality RCTs that acupuncture is not more effective than a comparison intervention (no acupuncture) for improving lower extremity motor function in patients with acute stroke.
Note: 
One of the high quality RCTs reported a significant difference in change scores at follow-up, in favour of acupuncture vs. no acupuncture. Further, two fair quality RCTs reported that acupuncture was more effective than no acupuncture. There was significant variation in the frequency and duration of interventions.

Motor function - upper extremity
Not effective
1a

Three high quality RCTs (Hsieh et al., 2007Zhu et al., 2013Chen et al., 2016) and two fair quality RCTs (Wong et al., 1999Min et al., 2008) investigated the effect of acupuncture on upper extremity motor function in patients with acute stroke.

The first high quality RCT (Hsieh et al., 2007) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Upper extremity motor function was measured by the Fugl Meyer Assessment (FMA – shoulder / elbow / wrist / hand motor function, upper extremity coordination and speed) during treatment (2 weeks), at post-treatment (4 weeks), and follow-up (3 and 6 months post-stroke). Significant between-group differences were found during treatment (FMA – hand motor function, upper extremity coordination and speed), post-treatment (FMA – wrist motor function, hand motor function, upper extremity coordination and speed), and at both follow-up time points (FMA – wrist motor function, hand motor function, upper extremity coordination and speed), favoring electroacupuncture vs. no acupuncture.

The second high quality RCT (Zhu et al., 2013) randomized patients to receive acupuncture or no acupuncture; both groups received conventional rehabilitation. Upper extremity motor function was measured by the Fugl-Meyer Assessment – Upper Extremity scale (FMA-UE) at mid-treatment (1 month), post-treatment (3 months) and follow-up (6 months). No significant between-group differences were found at any time point.

The third high quality RCT (Chen et al., 2016) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Upper extremity motor function was measured by the FMA-UE at post-treatment (3 weeks) and follow-up (7 weeks). No significant between-group differences were found at either time point.

The first fair quality RCT (Wong et al., 1999) randomized patients to receive electroacupuncture or no acupuncture. Upper extremity motor function was measured by Brunnstrom’s upper limb motor recovery at post-treatment (2 weeks). Significant between-group differences were found, favoring electroacupuncture vs. no acupuncture.

The second fair quality RCT (Min et al., 2008) randomized patients to receive acupuncture or no acupuncture; both groups received conventional rehabilitation. Upper extremity motor function was measured by the FMA-UE at post-treatment (3 months). A significant between-group difference was found, favoring acupuncture vs. no acupuncture.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that acupuncture is not more effective than a comparison intervention (no acupuncture) in improving upper extremity motor function in patients with acute stroke.
Note: 
However; one high quality RCT and two fair quality RCTs found that acupuncture was more effective than a comparison intervention (no acupuncture) in improving upper extremity motor function in patients with acute stroke. Studies varied in terms of the intervention, frequency (2-6 times/week) and duration (2 weeks – 3 months) of the intervention, and outcome measures used.

Range of motion
No effective
1b

One high quality RCT (Hsieh et al., 2007) investigated the effect of acupuncture on range of motion in patients with acute stroke. This high quality RCT randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Range of motion was measured by the Fugl Meyer Assessment (FMA – range of motion) at mid-treatment (2 weeks), post-treatment (4 weeks), and follow-up (3 and 6 months post-stroke). There was a significant between-group difference in range of motion at 3 months post-stroke only, favoring electroacupuncture vs. no acupuncture.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that electroacupuncture is not more effective than a comparison intervention (no acupuncture) in improving range of motion in patients with acute stroke.

Sensation
Not effective
1b

One high quality RCT (Hsieh et al., 2007) investigated the effects of acupuncture on sensation in patients with acute stroke. The high quality RCT randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Sensation was measured by the Fugl Meyer Assessment (FMA – sensation) at mid-treatment (2 weeks), post-treatment (4 weeks), and follow-up (3 and 6 months post-stroke). No significant between-group differences were found at any time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (no acupuncture) in improving sensation in patients with acute stroke.

Spasticity
Conflicting
4

Two high quality RCTs (Park et al., 2005; Li et al., 2014) investigated the effect of acupuncture on spasticity in patients with acute stroke.

The first high quality RCT (Park et al., 2005) randomized patients to receive manual acupuncture or sham acupuncture. Spasticity was measured by the Modified Ashworth Scale (MAS) at post-treatment (2 weeks). No significant between-group differences were found.

The second high quality RCT (Li et al., 2014) randomized patients to receive verum acupuncture or sham acupuncture. Spasticity was measured by the MAS at baseline, at mid-treatment (2 weeks), post-treatment (4 weeks), and follow-up (12 weeks). Significant between-group differences in spasticity were found at post-treatment and follow-up, favoring verum acupuncture vs. sham acupuncture.
Note: Differences at post-treatment reflect change scores from baseline to post-treatment; differences at follow-up reflect scores at that time point as well as change scores from baseline to follow-up.

Conclusion: There is conflicting evidence (Level 4) regarding the effect of acupuncture on spasticity in patients with acute stroke. While one high quality RCT found manual acupuncture (2 weeks duration) was not more effective than sham acupuncture, a second high quality RCT reported a significant difference in change scores following verum acupuncture (4 weeks duration), in improving spasticity in patients with acute stroke.

Strength
Not effective
1a

Two high quality RCTs (Park et al., 2005; Hopwood et al., 2008) investigated the effect of acupuncture on strength in patients with acute stroke.

The first high quality RCT (Park et al., 2005) randomized patients to receive manual acupuncture or sham acupuncture. Strength was measured by the Motricity Index (MI) at post-treatment (2 weeks). No significant between-group differences were found.

The second quality RCT (Hopwood et al., 2008) randomized patients to receive electroacupuncture or placebo electroacupuncture. Strength was measured by the MI at mid-treatment (3 weeks) and at follow-up (6, 12, 25, and 52 weeks); measures were not taken at post-treatment (4 weeks). No significant between-group differences were found at any time point.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that acupuncture is not more effective than comparison interventions (sham acupuncture, placebo electroacupuncture) in improving strength in patients with acute stroke.

Stroke outcomes
Not effective
1a

Seven high quality RCTs (Gosman-Hedstrom et al., 1998; Park et al., 2005; Tan et al., 2013; Li et al., 2014; Zhang et al., 2015; Chen et al., 2016, Liu et al., 2016) and three fair quality RCTs (Si et al., 1998; Pei et al., 2001; Wang et al., 2014) investigated the effect of acupuncture on stroke outcomes in patients with acute stroke.

The first high quality RCT (Gosman-Hedstrom et al., 1998) randomized patients to receive deep electroacupuncture, superficial acupuncture or no acupuncture; all groups received conventional rehabilitation. Stroke outcomes were measured by the Scandinavian Stroke Study Group – Neurological score at post-treatment (3 months) and follow-up (12 months). No significant between-group differences were found at either time point.

The second high quality RCT (Park et al., 2005) randomized patients to receive manual acupuncture or sham acupuncture. Stroke outcomes were measured by the National Institutes of Health Stroke Scale (NIHSS) at post-treatment (2 weeks). No significant between-group differences were found.

The third high quality RCT (Tan et al., 2013) randomized patients to receive electroacupuncture or no electroacupuncture. Stroke outcomes were measured by the Modified Edinburg Scandinavian Stroke Scale and the NIHSS at post-treatment (14 days). Significant between-group differences were found on both measures at post-treatment, favoring electroacupuncture vs. no electroacupuncture.

The forth high quality RCT (Li et al., 2014) randomized patients to receive verum acupuncture or sham acupuncture. Stroke outcomes were measured by the NIHSS at mid-treatment (2 weeks), post-treatment (4 weeks), and follow-up (12 weeks). No significant between-group differences were found at any time point.

The fifth high quality RCT (Zhang et al., 2015) randomized patients to receive acupuncture or no acupuncture. Stroke outcomes were measured by the Scandinavian Stroke Scale at post-treatment (3 weeks). Significant between-group differences were found, favoring acupuncture vs. no acupuncture.
Note: Results were significant only for participants who had received 10 or more acupuncture sessions.

The sixth high quality RCT (Chen et al., 2016) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Stroke outcomes were measured by the NIHSS at baseline, during treatment (1 week), at post-treatment (3 weeks), and follow-up (7 weeks). There were no significant differences between groups during treatment or at post-treatment. There was a significant between-group difference in change scores from baseline to follow-up, favoring acupuncture vs. no acupuncture.

The seventh high quality RCT (Liu et al., 2016) randomized patients to receive manual acupuncture or no acupuncture. Stroke outcomes were measured by the NIHSS at post-treatment (2 weeks) and follow-up (3, 4, 12 weeks). No significant between-group differences were found at any time point.

The first fair quality RCT (Si et al., 1998) randomized patients to receive electroacupuncture or no acupuncture. Stroke outcomes were measured by the Chinese Stroke Scale (CSS – total score, motor shoulder/hand/leg, level of consciousness, extraocular movements, facial palsy, speech, walking capacity) at discharge from hospital (average of 37±12 days). Significant between group differences in some stroke outcomes (CSS – total, motor shoulder/hand/leg) were found at discharge, favoring electroacupuncture vs. no acupuncture.

The second fair quality RCT (Pei et al., 2001) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Stroke outcomes were measured by the CSS during treatment (1 and 2 weeks), at post-treatment (4 weeks) and at follow-up (3 months). Significant between-group differences in stroke outcomes were found at 2 weeks, 4 weeks and 3 months, favoring electroacupuncture vs. no acupuncture.

The third fair quality RCT (Wang et al., 2014) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Stroke outcomes were measured by the NIHSS at post-treatment (4 weeks) and at follow-up (3 months). Significant between-group differences were found at post-treatment, favoring electroacupuncture vs. no electroacupuncture. These differences were not maintained at follow-up.

Conclusion: There is strong evidence (Level 1a) from five high quality RCTs that acupuncture is not more effective than comparison interventions (superficial/no/sham acupuncture) in improving stroke outcomes in patients with acute stroke.
Note:
However, two high quality RCTs and three fair quality RCTs found that acupuncture is more effective than a comparison intervention (no acupuncture) in improving stroke outcomes in patients with acute stroke. Differences between studies, including variation in the type of acupuncture, treatment frequency/duration and outcome measures used may account for this discrepancy in findings.

Swallowing function
Effective
1a

Three high quality RCTs (Park et al., 2005; Chen et al., 2016; Xia et al., 2016) investigated the effect of acupuncture on swallowing function in patients with acute stroke.

The first high quality RCT (Park et al., 2005) randomized patients to receive manual acupuncture or sham acupuncture. Swallowing function was measured by the Bedside Swallowing Assessment (BSA) at post-treatment (2 weeks). Significant between group differences were found, favoring sham acupuncture vs. manual acupuncture (i.e. participants who received manual acupuncture presented with a higher incidence of unsafe swallow than participants who received sham acupuncture).

The second high quality RCT (Chen et al., 2016) randomized patients to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Swallowing function was measured by the BSA at post-treatment (3 weeks) and follow-up (7 weeks), and by Videofluoroscopic Swallowing Study (VFSS) at follow-up (7 weeks). Significant between-group differences were found at post-treatment (BSA) and at follow-up (BSA, VFDSS), favoring acupuncture vs. no acupuncture.

The third high quality RCT (Xia et al., 2016) randomized patients to receive acupuncture or no acupuncture; both groups received standard swallowing training. Swallowing function was measured by the Standardized Swallowing Assessment at post-treatment (4 weeks). Significant between-group differences were found, favoring acupuncture vs. no acupuncture.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that acupuncture is more effective than a comparison intervention (no acupuncture) in improving swallowing function in patients with acute stroke.
Note:
However, one high quality RCT found that acupuncture was LESS effective than a comparison intervention (sham acupuncture) in improving swallowing function in patients with acute stroke.

Swallowing-related quality of life
Effective
1b

One high quality RCT (Xia et al., 2016) investigated the effects of acupuncture on swallowing-related quality of life in patients with acute stroke and subsequent dysphagia. This high quality RCT randomized patients to receive acupuncture or no acupuncture; both groups received standard swallowing training. Swallowing-related quality of life was measured with the Swallowing Related Quality of Life scale at post-treatment (4 weeks). Significant between-group differences were found, favoring acupuncture vs. no acupuncture.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is more effective than a comparison intervention (no acupuncture with standard swallowing training) in improving swallowing related quality of life in patients with acute stroke and subsequent dysphagia.

Unilateral spatial neglect
Not effective
1b

One high quality RCT (Rorsman & Johansson, 2006) investigated the effect of acupuncture on unilateral spatial neglect in patients with acute stroke. This high quality RCT randomized patients to receive acupuncture (including electroacupuncture), high intensity/low frequency TENS or low intensity (subliminal)/high frequency TENS. Unilateral spatial neglect was measured by the Star Cancellation Test and Time Perception Test at follow-up (3 and 12 months post-stroke); measures were not taken at post-treatment (10 weeks). No significant between-group differences were found on any measure at either time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than comparison interventions (TENS) in improving unilateral spatial neglect in patients with acute stroke.

Walking speed
Not effective
1b

One high quality RCT (Park et al., 2005) investigated the effect of acupuncture on walking speed in patients with acute stroke. This high quality RCT randomized patients to receive manual acupuncture or sham acupuncture. Walking speed was measured by the 10 Meter Walk Test at post-treatment (2 weeks). No significant between-group differences were found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (sham acupuncture) in improving walking speed in patients with acute stroke.

Subacute phase

Range of motion
Not effective
1b

One high quality RCT (Naeser et al., 1992) investigated the effect of acupuncture on range of motion in patients with subacute stroke. This high quality RCT randomized patients to receive electroacupuncture or sham acupuncture. Isolated active range of motion was measured at post-treatment (4 weeks). No significant between-group differences were found.
Note: A subgroup analysis of patients with the lesion in half or less than half of the motor pathway areas revealed significant between-group differences, favoring electroacupuncture vs. sham acupuncture.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that electroacupuncture is not more effective than a comparison intervention (sham acupuncture) in improving isolated active range of motion in patients with subacute stroke.

Chronic phase

Depression
Not effective
1a

Two high quality RCTs (Fink et al., 2004; Wayne et al., 2005) investigated the effect of acupuncture on depression in patients with chronic stroke. This first high quality RCT (Fink et al., 2004) randomized patients to receive acupuncture or placebo acupuncture. Depression was measured by the von Zerssen Depression Scale at post-treatment (4 weeks) and follow-up (3 months). No significant between-group differences were found at either time point. 

The second high quality RCT (Wayne et al., 2005) randomized patients to receive acupuncture or sham acupuncture. Depression was measured by the Center for Epidemiological Surveys Depression at post-treatment (12 weeks). No significant between-group differences were found.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that acupuncture is not more effective than a comparison intervention (placebo/sham acupuncture) in improving depression in patients with chronic stroke.

Functional independence
Not effective
1b

One high quality RCT (Wayne et al, 2005) investigated the effect of acupuncture on functional independence in patients with chronic stroke. This high quality RCT randomized patients to receive acupuncture or sham acupuncture. Functional independence was measured by the Barthel Index at post-treatment (12 weeks). No significant between-group differences were found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (sham acupuncture) in improving functional independence in patients with chronic stroke.

Gait parameters
Not effective
1b

One high quality RCT (Fink et al., 2004) investigated the effect of acupuncture on gait parameters in patients with chronic stroke. This high quality RCT randomized patients to receive acupuncture or placebo acupuncture. Gait parameters (step length, cadence, mode of initial foot contact) were measured at first treatment, post-treatment (4 weeks), and follow-up (3 months). No significant between-group differences were found at any time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (placebo acupuncture) in improving gait parameters in patients with chronic stroke.

Grip strength
Not effective
1b

One high quality RCT (Wayne et al, 2005) investigated the effect of acupuncture on grip strength in patients with chronic stroke. This high quality RCT randomized patients to receive acupuncture or sham acupuncture. Grip strength was measured by Jamar dynamometer at post-treatment (12 weeks). No significant between-group differences were found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (sham acupuncture) in improving grip strength in patients with chronic stroke.

Health-related quality of life (HRQoL)
Not effective
1a

Two high quality RCTs (Fink et al., 2004; Wayne et al., 2005) investigated the effect of acupuncture on HRQoL in patients with chronic stroke.

This first high quality RCT (Fink et al., 2004) randomized patients to receive acupuncture or placebo acupuncture. HRQoL was measured by the Nottingham Health Profile and the Everyday Life Questionnaire at post-treatment (4 weeks) and follow-up (3 months). No significant between-group differences were found on either measure at either time point. 

The second high quality RCT (Wayne et al, 2005) randomized patients to receive acupuncture or sham acupuncture. HRQoL was measured by the Nottingham Health Profile at post-treatment (12 weeks). No significant between-group differences were found.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that acupuncture is not more effective than a comparison intervention (placebo/sham acupuncture) in improving health-related quality of life in patients with chronic stroke.

Impression of improvement
Not effective
1b

One high quality RCT (Fink et al., 2004) investigated the effect of acupuncture on impression of improvement in patients with chronic stroke. This high quality RCT randomized patients to receive acupuncture or placebo acupuncture. Impression of improvement was measured by the Clinical Global Impressions Scale at first treatment, post-treatment (4 weeks), and follow-up (3 months). Significant between-group differences in patients’ impression of improvement were found at post-treatment, favoring placebo acupuncture vs. acupuncture.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (placebo acupuncture) in increasing the impression of improvement in patients with chronic stroke. In fact, patients who received acupuncture showed lower impression of improvement as compared to those who received placebo acupuncture.

Mobility
Not effective
1b

One high quality RCT (Fink et al., 2004) investigated the effect of acupuncture on mobility in patients with chronic stroke. This high quality RCT randomized patients to receive acupuncture or placebo acupuncture. Mobility was measured by the Rivermead Mobility Index at first treatment, post-treatment (4 weeks), and follow-up (3 months). No significant between-group differences were found at any time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (placebo acupuncture) in improving mobility in patients with chronic stroke.

Motor function
Not effective
1a

Two high quality RCTs (Fink et al., 2004, Wayne et al., 2005) investigated the effect of acupuncture on motor function in patients with chronic stroke.

This first high quality RCT (Fink et al., 2004) randomized patients to receive acupuncture or placebo acupuncture. Motor function was measured by the Rivermead Motor Assessment at first treatment, post-treatment (4 weeks), and follow-up (3 months). No significant between-group differences were found at any time point.

The second high quality RCT (Wayne et al., 2005) randomized patients to receive acupuncture or sham acupuncture. Motor function was measured by the Fugl-Meyer Assessment at post-treatment (12 weeks). No significant between-group differences were found.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that acupuncture is not more effective than a comparison intervention (placebo/sham acupuncture) in improving motor function in patients with chronic stroke.

Pain
Not effective
1b

One high quality RCT (Fink et al., 2004) investigated the effect of acupuncture on pain in patients with chronic stroke. This high quality RCT randomized patients to receive acupuncture or placebo acupuncture. Pain was measured by Visual Analogue Scale at first treatment, post-treatment (4 weeks), and follow-up (3 months). No significant between-group differences were found at any time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (placebo acupuncture) in improving pain in patients with chronic stroke.

Range of motion - upper extremity
Not effective
1a

Two high quality RCTs (Wayne et al., 2005, Schaechter et al., 2007) investigated the effect of acupuncture on upper extremity range of motion in patients with chronic stroke.

The first high quality RCT (Wayne et al., 2005) randomized patients to receive acupuncture or sham acupuncture. Upper extremity range of motion (shoulder, elbow, forearm, wrist, thumb, digits) was measured at post-treatment (12 weeks). No significant between-group differences were found.

The second high quality RCT (Schaechter et al., 2007) randomized patients to receive acupuncture with electroacupuncture or sham acupuncture with sham electroacupuncture. Upper extremity active assisted range of motion was measured at 2 weeks post-treatment (12 weeks). No significant between-group differences were found.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that acupuncture is not more effective than comparison interventions (sham acupuncture, sham electroacupuncture) in improving upper extremity range of motion in patients with chronic stroke.

Spasticity - lower extermity
Not effective
1b

One high quality RCT (Fink et al., 2004) investigated the effect of acupuncture on lower extremity spasticity in patients with chronic stroke. This high quality RCT randomized patients to receive acupuncture or placebo acupuncture. Ankle spasticity was measured by the Modified Ashworth Scale and the Hoffman’s reflex (Hmax/Mmax ratio of the spastic leg) using the Nicolet Viking II device at first treatment, post-treatment (4 weeks), and follow-up (3 months). Significant between-group differences in spasticity (Hoffman’s reflex) were found at post-treatment, favoring placebo acupuncture vs. acupuncture. These differences were not maintained at follow-up.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (placebo acupuncture) in reducing ankle spasticity in patients with chronic stroke. In fact, patients who received acupuncture showed greater spasticity in their affected ankle as compared to those who received placebo acupuncture.

Spasticity - upper extermity
Not effective
1a

Two high quality RCTs (Wayne et al., 2005; Schaechter et al., 2007) and one poor quality crossover RCT (Mukherjee et al., 2007) investigated the effect of acupuncture on upper extremity spasticity in patients with chronic stroke.

The first high quality RCT (Wayne et al., 2005) randomized patients to receive acupuncture or sham acupuncture. Spasticity in the elbow and wrist was measured by the Modified Ashworth Scale at post-treatment (12 weeks). No significant between-group differences were found.

The second high quality RCT (Schaechter et al., 2007) randomized patients to receive acupuncture with electroacupuncture or sham acupuncture with sham electroacupuncture. Upper extremity spasticity was measured by the Modified Ashworth Scale at 2 weeks post-treatment (12 weeks). No significant between-group differences were found.

The poor quality crossover RCT (Mukherjee et al., 2007) randomized patients to receive electroacupuncture or no electroacupuncture; both groups received strengthening exercises. Spasticity of the wrist was measured at post-treatment (6 weeks). Significant between-group differences on one measure of wrist spasticity were found, favoring electroacupuncture vs. no electroacupuncture.
Note: Other measures of spasticity were taken, however between-group analyses were not performed.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that acupuncture is not more effective than comparison interventions (sham acupuncture, sham electroacupuncture) in reducing upper extremity spasticity in patients with chronic stroke.
Note
: However, a poor quality crossover RCT found a significant difference on one measure of wrist spasticity, in favour of electroacupuncture + strengthening exercises alone vs. strengthening exercises alone.

Walking endurance
Not effective
1b

One high quality RCT (Fink et al., 2004) investigated the effect of acupuncture on walking endurance in patients with chronic stroke. This high quality RCT randomized patients to receive acupuncture or placebo acupuncture. Walking endurance was measured by the 2-Minute Walk Test at first treatment, post-treatment (4 weeks), and follow-up (3 months). No significant between-group differences were found at any time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (placebo acupuncture) in improving walking endurance in patients with chronic stroke.

Phase not specific to one period

Balance
Not effective
1b

One high quality RCT (Alexander et al., 2004) investigated the effect of acupuncture on balance in patients with stroke. This high quality RCT randomized patients with acute/subacute stroke to receive acupuncture or no acupuncture for 2 weeks; both groups received conventional rehabilitation. Balance was measured by the Fugl-Meyer Assessment (FMA – Balance) at discharge from hospital. No significant between-group differences were found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (no acupuncture) in improving balance in patients with stroke.

Cognitive function
Effective
1b

One high quality RCT (Jiang et al., 2016) investigated the effect of acupuncture on cognitive function in patients with stroke. This high quality RCT randomized patients with acute/subacute stroke to receive acupuncture (AC) + conventional rehabilitation (CR), computerized cognitive rehabilitation (COG) + CR, combined AC+COG+CR, or CR alone. Cognitive function was measured by the Mini Mental State Examination and the Montreal Cognitive Assessment (MOCA) at baseline and at post-treatment (12 weeks). Significant between-group differences in change scores from baseline to post-treatment were found on both measures, favoring AC+CR vs. CR alone. There were no significant between-group differences between AC+CR vs. COG+CR.
Note: Significant between-group differences in change scores of both measures were also found in favour of COG+CR vs. CR alone; AC+COG+CR vs. CR alone; AC+COG+CR vs. AC+CR; and AC+COG+CR vs. COG+CR.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is more effective than a comparison intervention (conventional rehabilitation) in improving cognitive function in patients with stroke.
Note:
Combined acupuncture + computerized cognitive training was also found to be more effective than comparison interventions (acupuncture alone, computerized cognitive training alone, conventional rehabilitation) in improving cognitive function in patients with stroke.

Functional independence
Not effective
1a

Five high quality RCTs (Sallstrom et al., 1996 – and a follow-up by Kjendahl et al., 1997 –; Alexander et al., 2004; Schuler et al., 2005; Zhuang et al., 2012; Jiang et al., 2016) and one fair quality RCTs (Hegyi & Szigeti, 2012) investigated the effect of acupuncture on functional independence in patients with stroke.

The first high quality RCT (Sallstrom et al., 1996) randomized patients with acute/subacute stroke to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Functional independence was measured by the Sunnaas Index at post-treatment (6 weeks) and at 1 year post-discharge from hospital (Kjendahl et al., 1997, follow-up study). Significant between-group differences were found at post-treatment and at follow-up, favoring electroacupuncture vs. no acupuncture.

The second high quality RCT (Alexander et al., 2004) randomized patients with acute/subacute stroke to receive acupuncture or no acupuncture for 2 weeks; both groups received conventional rehabilitation. Functional independence was measured by the Functional Independence Measure (FIM) at discharge from hospital. A significant between-group difference was found in only one measure of functional independence (tub/shower transfer), favoring acupuncture vs. no acupuncture.

The third high quality RCT (Schuler et al., 2005) randomized patients with acute/subacute stroke to receive electroacupuncture or placebo acupuncture. Functional independence was measured by the Barthel Index at post-treatment (4 weeks) and at follow-up (6 months). No significant between-group differences were found at either time point.

The forth high quality RCT (Zhuang et al., 2012) randomized patients with acute/subacute stroke to receive acupuncture, conventional rehabilitation or combined acupuncture with conventional rehabilitation. Functional independence was measured by the modified Barthel Index at mid-treatment (2 weeks) and at post-treatment (4 weeks). No significant between-group differences were found at either time point.

The fifth high quality RCT (Jiang et al., 2016) randomized patients with acute/subacute stroke to receive acupuncture (AC) + conventional rehabilitation (CR), computerized cognitive rehabilitation (COG) + CR, combined AC+COG+CR, or CR alone. Functional independence was measured at baseline and at post-treatment (12 weeks) by the FIM. Significant between-group differences were found in FIM change scores from baseline to post-treatment, favoring AC+CR vs. CR alone. There were no significant differences between AC+CR vs. COG+CR.
Note: Significant differences in FIM change scores were also found in favour of COG+CR vs. CR alone; AC+COG+CR vs. CR alone; AC+COG+CR vs. AC+CR; and AC+COG+CR vs. COG+CR.

The fair quality RCT (Hegyi & Szigeti, 2012) randomized patients with acute/subacute stroke to receive acupuncture or no acupuncture for the time of hospitalization (duration not specified); both groups received conventional physical therapy. Functional independence was measured by the Barthel Index at 2 years post-stroke. Significant between-group differences were found, favoring acupuncture vs. no acupuncture.

Conclusion: There is strong evidence (Level 1a) from three high quality RCTs that acupuncture is not more effective than comparison interventions (no/placebo acupuncture, conventional rehabilitation) in improving functional independence in patients with stroke.
Note:
However, two high quality RCTs and one fair quality RCT found that acupuncture was more effective than a comparison intervention (no acupuncture, conventional rehabilitation alone) in improving functional independence in patients with stroke.

Health-related quality of life (HRQoL)
Effective
1b

One high quality RCT (Sallstrom et al., 1996; and Kjendahl et al., 1997 follow-up study) and one fair quality RCT (Hegyi & Szigeti, 2012) investigated the effect of acupuncture on HRQoL in patients with stroke.

The high quality RCT (Sallstrom et al., 1996) randomized patients with acute/subacute stroke to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. HRQoL was measured by the Nottingham Health Profile (NHP – Part I, Part II) at post-treatment (6 weeks) and at 1 year post-discharge from hospital (Kjendahl et al., 1997 follow-up study). Significant between-group differences were found at post-treatment (NHP Part I: sleep, energy) and at follow-up (NHP Part I: emotion, sleep, physical movement, energy; Part II), favoring electroacupuncture vs. no acupuncture.

The fair quality RCT (Hegyi & Szigeti, 2012) randomized patients with acute/subacute stroke to receive acupuncture or no acupuncture for the time of hospitalization (duration not specified); both groups received conventional physical therapy. HRQoL (general and physical statuses) was measured by Visual Analogue Scale at 2 years post-stroke. A significant between-group difference was found, favoring acupuncture vs. no acupuncture.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one fair quality RCT that electroacupuncture is more effective than a comparison intervention (no acupuncture) in improving health-related quality of life in patients with stroke.

Insomnia
Effective
1b

One high quality RCT (Kim et al., 2004) investigated the effect of acupuncture on insomnia in patients with stroke. This high quality RCT randomized patients with stroke (stage of recovery not specified) and insomnia to receive intradermal acupuncture or sham acupuncture. Symptoms of insomnia were measured by the Morning Questionnaire (MQ – sleep latency, sleep quality, condition upon awakening, ability to concentrate, ease of falling asleep, morning sleepiness), the Insomnia Severity Index (ISI) and the Athens Insomnia Scale (AIS) at mid-treatment (1 day) and post-treatment (2 days). Significant between-group differences were found at both time points (MQ – sleep quality, condition upon awakening, ability to concentrate, morning sleepiness; ISI; AIS), favoring intradermal acupuncture vs. sham acupuncture.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is more effective than a comparison intervention (sham acupuncture) in improving symptoms of insomnia in patients with stroke and insomnia.

Joint pain
Not effective
1b

One high quality RCT (Alexander et al., 2004) investigated the effect of acupuncture on joint pain in patients with stroke. This high quality RCT randomized patients with acute/subacute stroke to receive acupuncture for 2 weeks or no acupuncture; both groups received conventional rehabilitation. Joint pain was measured by the Fugl-Meyer Assessment (FMA – upper and lower extremity joint pain) at discharge from hospital. No significant between-group differences were found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (no acupuncture) in improving joint pain in patients with stroke.

Mobility
Effective
2a

One fair quality RCT (Hegyi & Szigeti, 2012) investigated the effect of acupuncture on mobility in patients with stroke. This fair quality RCT randomized patients with acute/subacute stroke to receive acupuncture or no acupuncture for the time of hospitalization (duration not specified); both groups received conventional physical therapy. Mobility was measured by the Rivermead Mobility Index at 2 years post-stroke. Significant between-group differences were found, favoring acupuncture vs. no acupuncture.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that acupuncture is more effective than a comparison intervention (no acupuncture) in improving mobility in patients with stroke.

Motor function
Not effective
1a

Three high quality RCTs (Sallstrom et al., 1996; and Kjendahl et al., 1997 follow-up study), Alexander et al., 2004, Zhuang et al., 2012) investigated the effect of acupuncture on motor function in patients with stroke.

The first high quality RCT (Sallstrom et al., 1996) randomized patients with acute/subacute stroke to receive electroacupuncture or no acupuncture; both groups received conventional rehabilitation. Motor function was measured by the Motor Assessment Scale at post-treatment (6 weeks) and at 1 year post-discharge from hospital (Kjendahl et al., 1997 follow-up study). Significant between-group differences were found, at both time points, favoring electroacupuncture vs. no acupuncture.

The second high quality RCT (Alexander et al., 2004) randomized patients with acute/subacute stroketo receive acupuncture for 2 weeks or no acupuncture; both groups received conventional rehabilitation. Motor function was measured by the Fugl-Meyer Assessment (FMA-total) at discharge from hospital. No significant between-group differences were found.

The third high quality RCT (Zhuang et al., 2012) randomized patients with acute/subacute stroke to receive acupuncture, conventional rehabilitation or combined acupuncture with conventional rehabilitation. Motor function was measured by the FMA at mid-treatment (2 weeks) and at post-treatment (4 weeks). No significant between-group differences were found at either time point.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that acupuncture is not more effective than a comparison intervention (no acupuncture, conventional rehabilitation) in improving motor function in patients with stroke.
Note:
However, one high quality RCT found that acupuncture was more effective than a comparison intervention (no acupuncture) in improving motor function in patients with stroke.

Motor function - lower extremity
Effective
1b

One high quality RCT (Alexander et al., 2004) investigated the effect of acupuncture on lower extremity motor function in patients with stroke. This high quality RCT randomized patients with acute/subacute stroke to receive acupuncture for 2 weeks or no acupuncture; both groups received conventional rehabilitation. Lower extremity motor function was measured by the Fugl-Meyer Assessment (FMA – lower extremity motor function) at discharge from hospital. Significant between-group differences were found, favoring acupuncture vs. no acupuncture.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is more effective than a comparison intervention (no acupuncture) in improving lower extremity motor function in patients with stroke.

Motor function - upper extremity
Not effective
1b

One high quality RCT (Alexander et al., 2004) investigated the effects of acupuncture on upper extremity motor function in patients with stroke. This high quality RCT randomized patients with acute/subacute stroke to receive acupuncture for 2 weeks or no acupuncture; both groups received conventional rehabilitation. Upper extremity motor function was measured by the Fugl-Meyer Assessment (FMA – Upper extremity motor function) at discharge from hospital. No significant between-group differences were found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (no acupuncture) in improving upper extremity motor function in patients with stroke.

Range of motion
Not effective
1b

One high quality RCT (Alexander et al., 2004) investigated the effect of acupuncture on range of motion in patients with stroke. This high quality RCT randomized patients with acute/subacute stroke to receive acupuncture for 2 weeks or no acupuncture; both groups received conventional rehabilitation. Joint motion was measured by the Fugl-Meyer Assessment (FMA – upper/lower extremity joint motion) at discharge from hospital. No significant between-group differences were found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than no acupuncture in improving upper and lower extremity range of motion in patients with stroke.

Sensation
Not effective
1b

One high quality RCT (Alexander et al., 2004) investigated the effect of acupuncture on sensation in patients with stroke. This high quality RCT randomized patients with acute/subacute stroke to receive acupuncture for 2 weeks or no acupuncture; both groups received conventional rehabilitation. Sensation was measured by the Fugl-Meyer Assessment (FMA – upper/lower extremity sensation) at discharge from hospital. No significant between-group differences were found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that acupuncture is not more effective than a comparison intervention (no acupuncture) in improving sensation in patients with stroke.

Stroke outcomes
Not effective
1a

Two high quality RCTs (Schuler et al., 2005; Zhuang et al., 2012) investigated the effect of acupuncture on stroke outcomes in patients with stroke.

The first high quality RCT (Schuler et al., 2005) randomized patients with acute/subacute stroke to receive electroacupuncture or placebo acupuncture. Stroke outcomes were measured by the European Stroke Scale at post-treatment (4 weeks) and at follow-up (6 months). No significant between-group differences were found at either time point.

The second high quality RCT (Zhuang et al., 2012) randomized patients with acute/subacute stroke to receive acupuncture, conventional rehabilitation or combined acupuncture with conventional rehabilitation. Stroke outcomes were measured by the Neurologic Defect Scale at mid-treatment (2 weeks) and at post-treatment (4 weeks). No significant between-group differences were found at either time point.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that acupuncture is not more effective than comparison interventions (placebo acupuncture, conventional rehabilitation) in improving stroke outcomes in patients with stroke.

Swallowing function
Effective
2b

One fair quality RCT (Mao et al., 2016) investigated the effect of acupuncture on swallowing function in patients with stroke. This fair quality RCT randomized patients with acute/subacute stroke and dysphagia to receive acupuncture + standard swallowing training or standard swallowing training alone. Swallowing function was measured by the Video Fluoroscopic Swallowing Study (VFSS), Standardized Swallowing Assessment (SSA) and the Royal Brisbane Hospital Outcome Measure for Swallowing (RBHOMS) at post-treatment (4 weeks). Significant between-group differences were found in two measures of swallowing function (VSFF, SSA), favoring acupuncture + standard swallowing training vs. standard swallowing training alone.

Conclusion: There is limited evidence (Level 2b) from one fair quality RCT that acupuncture with swallowing training is more effective than a comparison intervention (standard swallowing training alone) in improving swallowing function in patients with stroke.

References

Alexander D.N., Cen S., Sullivan K.J., Bhavnani G., Ma X., & Azen S.P. (2004). Effects of acupuncture treatment on poststroke motor recovery and physical function: a pilot study. Neurorehabilitation and Neural Repair, 18(4), 259-267.
https://www.ncbi.nlm.nih.gov/pubmed/15537996

Chen, L., Fang, J., Ma, R., Gu, X., Chen, L., Li, J., & Xu, S. (2016). Additional effects of acupuncture on early comprehensive rehabilitation in patients with mild to moderate acute ischemic stroke: a multicenter randomized controlled trial. BMC complementary and alternative medicine16(1), 226.
https://bmccomplementalternmed.biomedcentral.com/articles/10.1186/s12906-016-1193-y

Fink M., Rollnik J. D., Bijak M., Borstadt C., Dauper J., Guergueltcheva V., Dengler, Reinhard D., & Karst M. (2004). Needle acupuncture in chronic poststroke leg spasticity. Archives of Physical Medicine and Rehabilitation, 85(4), 667-672.
http://www.archives-pmr.org/article/S0003-9993(03)00941-9/abstract

Gosman – Hedstrom G., Claesson L., Klingenstierna U., Carlsson J., Olausson B., Frizell M., Fagerberg B., & Blomstrand C. (1998). Effects of acupuncture treatment on daily life activities and quality of life: a controlled, prospective, and randomized study of acute stroke patients. Stroke, 29, 2100-2108.
https://www.ncbi.nlm.nih.gov/pubmed/9756589

Hegyi, G., & Szigeti, G. P. (2012). Rehabilitation of stroke patients using yamamoto new scalp acupuncture: a pilot study. The Journal of Alternative and Complementary Medicine18(10), 971-977.
http://online.liebertpub.com/doi/abs/10.1089/acm.2011.0047

Hopwood V, Lewith G., Prescott P., & Campbell M.J. (2008). Evaluating the efficacy of acupuncture in defined aspects of stroke recovery: A randomized, placebo controlled single blind study, Journal of Neurology, 255, 858-866.
https://www.ncbi.nlm.nih.gov/pubmed/18465110

Hsieh R.L., Wang L.Y., & Lee W.C. (2007). Additional therapeutic effects of electro-acupuncture in conjunction with conventional rehabilitation for patients with first-ever ischaemic stroke. Journal of Rehabilitation Medicine, 39, 205-211.
https://www.medicaljournals.se/jrm/content/html/10.2340/16501977-0032

Hu H.H., Chung C., Li T.J., Chen R.C., Chen C.H., Chou P., Huang W.S., Lin J.C.T., & Tsuei J.J. (1993). A randomized controlled trial on the treatment for acute partial ischemic stroke with acupuncture. Neuroepidemiology, 12, 106-113.
https://www.ncbi.nlm.nih.gov/pubmed/8232703

Jiang, C., Yang, S., Tao, J., Huang, J., Li, Y., Ye, H., … & Chen, L. (2016). Clinical efficacy of acupuncture treatment in combination with rehacom cognitive training for improving cognitive function in stroke: a 2× 2 factorial design randomized controlled trial. Journal of the American Medical Directors Association17(12), 1114-1122.
http://www.sciencedirect.com/science/article/pii/S1525861016302997

Johansson K., Lindgren I., Widner H., Wiklund I., & Johansson B.B. (1993). Can sensory stimulation improve the functional outcome in stroke patients? Neurology, 43, 2189-2192.
https://www.ncbi.nlm.nih.gov/pubmed/8232927

Johansson B.B., Haker E., von Arbin M., Britton M., Langstrom G., Teréent A, Ursing D., & Asplund K. (2001). Acupuncture and transcutaneous nerve stimulation in stroke rehabilitation: a randomized, controlled trial. Stroke, 32, 707-713.
http://stroke.ahajournals.org/content/strokeaha/32/3/707.full.pdf

Kim Y.S., Lee S.H., Jung W.S., Park S.U., Moon S.K., Ko C.N., Cho K.H., Bae H.S. (2004). Intradermal acupuncture on shen-men and nei-kuan acupoints in patients with insomnia after stroke. American Journal of Chinese Medicine, 32, 771-778.
https://www.ncbi.nlm.nih.gov/pubmed/15633811

Kjendahl A., Sallstrom S., Egil Osten P., & Kvalvik Stanghelle J. (1997). A one year follow-up study on the effects of acupuncture in the treatment of stroke patients in the subacute stage: a randomized, controlled study. Clinical Rehabilitation, 11, 192-200.
https://www.ncbi.nlm.nih.gov/pubmed/9360031

Li, H., Liu, H., Liu, C., Shi, G., Zhou, W., Zhao, C., … & Sun, J. (2014). Effect of “Deqi” during the study of needling “Wang’s Jiaji” acupoints treating spasticity after stroke. Evidence-Based Complementary and Alternative Medicine2014.
https://www.hindawi.com/journals/ecam/2014/715351/abs/

Liu, C. H., Hsieh, Y. T., Tseng, H. P., Lin, H. C., Lin, C. L., Wu, T. Y., … & Zhang, H. (2016). Acupuncture for a first episode of acute ischaemic stroke: an observer-blinded randomised controlled pilot study. Acupuncture in Medicine34(5), 349-355.
http://aim.bmj.com/content/34/5/349

Mao, L. Y., Li, L. L., Mao, Z. N., Han, Y. P., Zhang, X. L., Yao, J. X., & Li, M. (2016). Therapeutic effect of acupuncture combining standard swallowing training for post-stroke dysphagia: A prospective cohort study. Chinese Journal of Integrative Medicine22(7), 525-531.
https://link.springer.com/article/10.1007/s11655-016-2457-6

Min, M., Xin, C., Yuefeng, C., Ping, R., & Jian, L. (2008). Stage-oriented comprehensive acupuncture treatment plus rehabilitation training for apoplectic hemiplegia. Journal of Traditional Chinese Medicine28(2), 90-93.
http://www.sciencedirect.com/science/article/pii/S0254627208600229

Mukherjee M., McPeak L.K., Redford J. B., Sun C., & Liu W. (2007). The effect of electro-acupuncture on spasticity of the wrist joint in chronic stroke survivors. Archives of Physical Medicine and Rehabilitation, 88, 159-166.
https://www.ncbi.nlm.nih.gov/pubmed/17270512

Naeser M.A., Alexander M.P., Stiassny-Eder D., Galler V., Hobbs J., & Bachman D. (1992). Real versus sham acupuncture in the treatment of paralysis in acute stroke patients: a CT scan lesion site study. Journal of Neuroengineering and Rehabilitation 6(4), 163-173.
http://www.bu.edu/naeser/acupuncture/publications/Naeser_RealvsSham_1992.pdf

Park J., White A.R., James M.A., Hemsley A.G., Johnson P., Chambers J., & Ernst E. (2005). Acupuncture for subacute stroke rehabilitation. A sham-controlled, subject and assessor-blind, randomized trial. Archives of Internal Medicine, 165, 2026-2031.
https://www.ncbi.nlm.nih.gov/pubmed/16186474

Pei, J., Sun, L., Chen, R., Zhu, T., Qian, Y., & Yuan, D. (2001). The effect of electro-acupuncture on motor function recovery in patients with acute cerebral infarction: a randomly controlled trial. Journal of Traditional Chinese Medicine21(4), 270-272.
http://europepmc.org/abstract/med/12014128

Rorsman I. & Johansson B. (2006). Can electro-acupuncture or transcutaneous nerve stimulation influence cognitive and emotional outcome after stroke? Journal of Rehabilitation Medicine, 38, 13-19.
https://www.ncbi.nlm.nih.gov/pubmed/16548081

Sallstrom S., Kjendahl A., Osten P.E., Stanghelle J.H., & Borchgrevink C.F. (1996). Acupuncture in the treatment of stroke patients in the subacute stage: a randomized, controlled study. Complementary Therapies in Medicine, 4, 193-197.
http://www.sciencedirect.com/science/article/pii/S0965229996800094

Schaechter J.D., Connell B.D., Stason W.B., Kaptchuk T.J., Krebs D.E., Macklin E.A., Schnyer R.N., Stein J., Scarborough D.M., Parker S.W., McGibbon C.A., & Wayne P.M. (2007). Correlated change in upper limb function and motor cortex activation after verum and sham acupuncture in patients with chronic stroke. Journal of Alternative and Complementary Medicine. 13, 527-532.
https://www.ncbi.nlm.nih.gov/pubmed/17604556

Schuler M.S., Durdak C., Hosl N.M., Klink A., Hauer K.A., & Oster P. (2005). Acupuncture treatment of geriatric patients with ischemic stroke: a randomized, double-controlled, single-blind study. Journal of the American Geriatrics Society, 53, 549-550.
https://www.ncbi.nlm.nih.gov/pubmed/15743311

Si Q.C., Wu G.C., & Cao X.D. (1998). Effects of electroacupuncture on acute cerebral infarction. Acupuncture and Electro-therapeutics Research, 23, 117-124.
https://www.ncbi.nlm.nih.gov/pubmed/9789586

Sze F.K., Wong E., Yi X., & Woo J. (2002). Does acupuncture have additional value to standard poststroke motor rehabilitation? Stroke, 33, 186-194.
https://www.ncbi.nlm.nih.gov/pubmed/11779909

Tan, F., Wang, X., Li, H. Q., Lu, L., Li, M., Li, J. H., … & Zheng, G. Q. (2013). A randomized controlled pilot study of the triple stimulation technique in the assessment of electroacupuncture for motor function recovery in patients with acute ischemic stroke. Evidence-Based Complementary and Alternative Medicine2013.
https://www.hindawi.com/journals/ecam/2013/431986/abs/

Wang, C., Wu, Z., Li, N., Zhao, Y., Tian, F., Zhou, X., & Wang, Z. (2014). Clinical curative effect of electric acupuncture on acute cerebral infarction: a randomized controlled multicenter trial. Journal of Traditional Chinese Medicine34(6), 635-640.
http://www.sciencedirect.com/science/article/pii/S0254627215300753

Wayne P. M., Krebs D. E., Macklin E. A., Schnyer R., Kaptchuk T. J., Parker S. W., Scarborough D.M., McGibbon C.A., Schaechter J.D., Stein J., & Stason W.B. (2005). Acupuncture for upper-extremity rehabilitation in chronic stroke: a randomized sham-controlled study. Archive of Physical Medicine Rehabilitation, 86, 2248-2255.
https://www.ncbi.nlm.nih.gov/pubmed/16344019

Wong A.M., Su T.Y., Tang F.T., Cheng P.T., & Liaw M.Y. (1999). Clinical trial of electrical acupuncture on hemiplegic stroke patients. American Journal of Physical Medicine and Rehabilitation, 78, 117-122.
https://www.ncbi.nlm.nih.gov/pubmed/10088585

Xia, W., Zheng, C., Zhu, S., & Tang, Z. (2016). Does the addition of specific acupuncture to standard swallowing training improve outcomes in patients with dysphagia after stroke? a randomized controlled trial. Clinical Rehabilitation30(3), 237-246.
http://journals.sagepub.com/doi/abs/10.1177/0269215515578698

Zhang, S., Wu, B., Liu, M., Li, N., Zeng, X., Liu, H., … & Wang, D. (2015). Acupuncture Efficacy on Ischemic Stroke Recovery. Stroke46(5), 1301-1306.
http://stroke.ahajournals.org/content/46/5/1301.short

Zhu, Y., Zhang, L., Ouyang, G., Meng, D., Qian, K., Ma, J., & Wang, T. (2013). Acupuncture in subacute stroke: no benefits detected. Physical Therapy93(11), 1447-1455.
https://academic.oup.com/ptj/article/93/11/1447/2735273/Acupuncture-in-Subacute-Stroke-No-Benefits

Zhuang, L. X., Xu, S. F., D’Adamo, C. R., Jia, C., He, J., Han, D. X., & Lao, L. X. (2012). An Effectiveness Study Comparing Acupuncture, Physiotherapy, and Their Combination in Poststroke Rehabifitation: A Multicentered, Randomized, Controlled Clinical Trial. Alternative Therapies in Health & Medicine18(3).
https://www.ncbi.nlm.nih.gov/pubmed/22875557

Aerobic Exercise

Evidence Reviewed as of before: 30-05-2021
Author(s)*: Tatiana Ogourtsova, PhD OT; Adam Kagan, B.Sc; Anita Petzold, BSc OT; Nathalie Serrat, BSC PT; Amanda Ischayek BSc PT; Sabrina Ianni, BSc, PT; Caroline Labelle, BSc PT; Sukhdeep Johal, Bsc PT; Monica Trozzo BSc. PT; Elissa Sitcoff, BA BSc; Annabel McDermott, OT; Nicol Korner-Bitensky, PhD OT
Expert Reviewer: Janice Eng, PhD PT; Pamela Duncan, PhD PT(C)
Patient/Family Information Table of contents

Introduction

It has been shown that patients with stroke have been shown to have low endurance during exercise, likely due to both the event and also as a secondary reaction to forced inactivity. It is also known that there is a positive connection between aerobic capacity and functional performance (Katz-Leurer et al. 2003).

Click here to view the AEROBICS 2019 Update Best Practice Recommendations.

Click here to access the Canadian Partnership for Stroke Recovery (CPSR) 2013 Clinicians’ guide.

Click here to access the CPSR 2013 Patients’ guide.

Patient/Family Information

Authors*: Erica Kader; Adam Kagan, B.Sc.; Nathalie Serrat, BSC PT; Amanda Ischayek BSc PT; Sabrina Ianni, BSc, PT; Caroline Labelle, BSc PT; Sukhdeep Johal, Bsc PT; Monica Trozzo BSc. PT; Elissa Sitcoff, BA BSc; Nicol Korner-Bitensky, PhD OT NOTE: *The authors have no direct financial interest in any tools, tests or interventions presented in StrokEngine.

What is aerobic exercise?

Aerobic exercise refers to physical activity that requires the body to use oxygen to generate energy. Participating in aerobic exercise is important to maintain a healthy body. A major benefit of aerobic exercise is that it conditions the heart and lungs. It does so by increasing the oxygen available to the body and enabling the heart to use oxygen more efficiently. In addition, aerobic exercise can also control body fat, increase energy, decrease tension, increase stamina, and improve mood. There are several different types of aerobic exercises that can be done at different levels of intensity for varying periods of time. Any activity that lasts longer than 3 minutes is considered aerobic (such as golf, biking, walking, and swimming).
Note: While other forms of exercises (such as those focused on flexibility and muscles training) are equally important, only those focusing on aerobic exercise will be addressed in this module.

Why is exercise important after I have had a stroke?

After a stroke, it is common to experience continued difficulties in mobility, for example in walking. It is important to continue to exercise despite these challenges to avoid a vicious cycle, where difficulty in mobility leads to lack of exercise, and lack of exercise leads to further muscle weakening and reduced fitness. Inactivity can contribute to physical complications, including osteoporosis and decreased circulation. It can also lead to loss of independence, depression, and social isolation. The more inactive you are, the harder it is to maintain cardiovascular, mental, and neurological health.

How do I begin to exercise after a stroke?

Before beginning an exercise program, it is recommended that you undergo a comprehensive medical evaluation to assess your specific needs. Your medical or rehabilitation team can work with you to develop an appropriate exercise regime (including types of activities, how often you should participate in activities and for how long) based on your individual needs and abilities.

What kind of activities should I do?

You should pick an activity that you will have fun doing. Examples of aerobic exercise activities include:

  • Golf
  • Walking
  • Dancing – With permission of Dr. Patricia McKinley, McGill School of Physical and Occupational Therapy
  • Swimming
  • Cycling
  • Tennis
  • Bowling

Gardening and housework are also great forms of aerobic exercise. Try adding exercise to your daily routine, for example, parking your car further away from your destination. Any form of physical activity can be beneficial as long as it is done regularly and consistently.

When it comes to bicycling, many people find it difficult or are afraid to fall. This problem can be solved by using a stationary bicycle. Stationary bicycles are a safe and effective means of low-impact, or light, aerobic exercise, so they are a good choice for people who have had a stroke. They can also be altered to fit your individual needs.

Treadmills are also helpful for walking, providing that there is a bar to hold on to, and a way to modify speed and intensity. A treadmill is especially useful to retrain people who have had a stroke to walk again.

Can I participate in the same exercise as before?

After a stroke, it may be difficult to resume the same activities that you enjoyed before. You may need to change your previous exercise regime, which may mean discovering new exercise activities that are perhaps less physically demanding. Things that you may need to modify are:

  • The level of difficulty of exercise
  • Length of time you exercise
  • How often you exercise

These will depend on your needs and abilities and should be assessed by a rehabilitation team. Certain equipment can also be used to facilitate exercising, such as handrails and assistive devices. For example, you may enjoy swimming but may need to find a pool that has special safety equipment and adaptations.

Who can help me resume my exercise activities?

While rehabilitation staff, such as occupational therapists, physiotherapists, social workers, recreation therapists, and psychologists will start you on your new exercise program, your family and friends are an excellent source of support to help you continue with success. Exercising with a friend or family member is motivating, encouraging, and of course more fun.

How much exercise should I do?

According to the American Heart Association, the recommended frequency of training is 3 to 7 days a week, with a duration of 20 to 60 minutes per day, depending on the patient’s level of fitness. ** Once again, however, it is very important that you seek medical advice before beginning an exercise program and get advice on how often and for how long you should be doing the activities.

Where can I participate in exercise?

While in the hospital or rehabilitation centre, you will participate in exercise programs developed and assisted by your rehabilitation team. When you are ready to go home, the team may show you how to continue with this exercise on your own, may recommend that you join an exercise program, or a combination of the two. Day centers, local community centers, and gyms in your area may be able to provide appropriate programs and support that you need.

Is it effective after stroke?

Experts have done some experiments to compare aerobic exercise with other treatments to see whether it helps people who have had a stroke.

In individuals with ACUTE stroke (< 4 weeks after stroke)

Studies found that aerobic exercise:

  • Was more helpful than the other treatments for improving awareness about stroke and walking endurance (i.e. your physical tolerance when walking).
  • Was as helpful as other treatments for improving cardiovascular fitness parameters (e.g. your blood pressure); quality of life; mood and affect (e.g. symptoms of depression and/or anxiety); and physical activity.

In individuals with CHRONIC stroke (> 6 months after stroke)

Studies found that aerobic exercise:

  • Was more helpful than the other treatments for improving cognitive function (e.g. memory, attention); grip strength; quality of life; walking endurance (i.e. physical tolerance when walking); and walking speed.
  • Was as helpful as other treatments for improving balance; cardiovascular fitness parameters (e.g. your blood pressure); executive functions (e.g. your ability to plan and sequence tasks); functional independence (i.e. your ability to perform tasks of daily life such as dressing and washing); mobility (walking, going up/down the stairs); mood and affect (e.g. symptoms of depression and/or anxiety); the strength of your leg muscles; and physical activity.

In individuals with stroke (unspecified time period post-stroke)

Studies found that aerobic exercise:

  • Was more helpful than the other treatments for improving balance; cardiovascular fitness parameters (e.g. your blood pressure); functional independence (your ability to perform tasks of everyday life such as dressing and washing); quality of life; the function of your legs and overall function of your body; spasticity (the tone of your muscles); walking endurance (your physical tolerance when walking); and walking speed.
  • Was as helpful as other treatments for improving cognitive abilities (e.g. memory, attention); dexterity (ability to manipulate small objects with your fingers); capacity to exercise; executive function (e.g. your ability to plan and sequence tasks); depression; fatigue; mobility (ability to move around); muscle strength; and the quality of sleep.

Are there any side effects or risks?

While exercise is mostly risk-free, it is important to stay within your own personal threshold. As mentioned before, it is best to consult with your doctor or therapist before beginning an exercise program. They will assist you in determining how often you should exercise, what activities you should participate in, and how intense they should be. If you were physically active before the stroke, you may or may not be able to continue with the same activities. You may simply need to modify those activities so they are easier for you. If you feel dizzy, have pain (especially in your chest) or have difficulty breathing, stop exercising immediately and tell your healthcare provider.

Clinician Information

Note: When reviewing the findings, it is important to note that they are always made according to randomized clinical trial (RCT) criteria – specifically as compared to a control group. To clarify, if a treatment is “effective” it implies that it is more effective than the control treatment to which it was compared. Non-randomized studies are no longer included when there is sufficient research to indicate strong evidence (level 1a) for an outcome.
Note: It is often difficult to say with absolute certainty whether a particular exercise intervention is “aerobic” in nature. In this module we include only those studies that had a clear aerobic exercise intervention. Specifically only those that included an outcome examining the effect of exercise on aerobic capacity (peak VO2, peak workload and peak heart rate during some sort of maximal aerobic test) were considered. Many of these studies also examined functional, physical and emotional outcomes and these results are included. As well, many studies to date that have examined the effect of aerobic exercise featured a “cocktail” of different types of treatment (e.g. strength training, flexibility training as well as a strong aerobic training component) so it is important to note that the effects of these interventions may be due in part to the combination of different treatments and not the aerobic component specifically.

This module focuses on aerobic exercise for people who have had a stroke. This module contains 16 studies, where 11 of them are of high quality. Three studies report effects of aerobic exercise for individuals early in their stroke recovery period (1 month or less after stroke). Nine studies report effects of aerobic exercise for individuals in their chronic stroke period (6 months or more after stroke). Four studies report effects of aerobic exercise for individuals after stroke with unspecified period.

Click here for the elearning course

Results Table

View results table

Outcomes

Acute phase

Cardiovascular fitness parameters
Not effective
1b

One high quality RCT (Wijkman et al., 2017) investigated the effect of aerobic exercise on cardiovascular fitness parameters in the acute phase of stroke recovery. This high quality RCT randomized patients to receive aerobic exercise or no scheduled physical exercise. Cardiovascular fitness parameters [Resting diastolic blood pressure, Resting systolic blood pressure (SBP), Peak SBP, Difference in SBP (peak – resting), Resting heart rate, Peak heart rate, Difference in heart rate (peak – resting), Aerobic capacity work rate] were measured by ergometer exercise test at post-treatment (12 weeks). No significant between-group differences were found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is not more effective than a comparison intervention (usual care without scheduled physical exercise) in improving cardiovascular fitness parameters in the acute phase of stroke recovery.

Health-related quality of life
Conflicting
4

Two high quality RCTs (Faulkner et al., 2015; Moren et al., 2016) investigated the effect of aerobic exercise on health-related quality of life (HRQoL) in the acute phase of stroke recovery.

The first high quality RCT (Faulkner et al., 2015) randomized patients to receive a resistance exercise and education program or written information. HRQoL was measured by the Short-Form 36 (SF-36: Physical component score, Mental component score, Mental health, Social functioning, Global health, Role physical, Role emotional, Vitality, Bodily pain, Physical functioning) at post-treatment (8 weeks) and follow-up (12 months). Significant between-group differences were found in change scores from baseline to post-treatment in some components of HRQoL (Physical component score, Global health, Role physical, Vitality, Physical Functioning), favouring exercise + education vs. written information. Differences did not remain significant at follow-up.

The second high quality RCT (Moren et al., 2016) randomized patients to receive physical activity or no treatment; both groups received usual care. HRQoL was measured by the EuroQoL 5 Dimension Visual Analogue Scale at follow-up (3 and 6 months). No significant between-group difference was found at either time point.

Conclusion: There is conflicting evidence (Level 4) regarding the effect of aerobic exercise on HRQoL in the acute phase of stroke recovery. While one high quality RCT found that an exercise + education program was more effective than written information alone, another high quality RCT found that physical activity was not more effective than no treatment.
Note:
Differences in outcome measures may explain the conflicting findings.

Mood and affect
Not effective
1b

One high quality RCT (Faulkner et al., 2015) investigated the effect of aerobic exercise on mood and affect in the acute phase of stroke recovery. This high quality RCT randomized patients to receive a resistance exercise and education program or written information. Mood and effect were measured by the Hospital Anxiety and Depression Scale (HADS: Anxiety, Depression) and the Profile and Mood States (PMS: Vigour, Depression, Confusion, Tension, Anger, Fatigue) at post-treatment (8 weeks) and follow-up (12 months). A significant between-group difference was found in change scores from post-treatment to follow-up in one measure (PMS: Fatigue), favouring exercise + education vs. written information. No other significant between-group differences were found at either time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is not more effective than a comparison intervention (written information) in improving mood and affect in the acute phase of stroke recovery.

Physical activity
Not Effective
1a

Two high quality RCTs (Faulkner et al., 2015; Moren et al., 2016) investigated the effect of aerobic exercise on physical activity in the acute phase of stroke recovery.

The first high quality RCT (Faulkner et al., 2015) randomized patients to receive a resistance exercise and education program or written information. Physical activity was measured by the International Physical Activity Questionnaire (IPAQ: Leisure time walk activity, Leisure time moderate activity, Leisure time vigorous activity, Total leisure time activity, Sitting time) at post-treatment (8 weeks) and follow-up (12 months). No significant between-group difference was found at either time point.

The second high quality RCT (Moren et al., 2016) randomized patients to receive physical activity or no treatment; both groups received usual care. Physical activity was measured by the Physical Activity of Moderate to Higher Intensity (MVPA) and number of steps per day at follow-up (3 and 6 months). No significant between-group differences were found at either time point.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that aerobic exercise is not more effective than a comparison intervention (written information) or no treatment in improving physical activity in the acute phase of stroke recovery.

Sroke awareness
Effective
1b

One high quality RCT (Faulkner et al., 2015) investigated the effect of aerobic exercise on stroke awareness in the acute phase of stroke recovery. This high quality RCT randomized patients to receive a resistance exercise and education program or written educational material. Stroke awareness was measured by the Stanford Medical Centre Stroke Awareness Questionnaire (SMCSAQ) at post-treatment (8 weeks) and follow-up (12 months). A significant between-group difference was found in change scores from baseline to post-treatment, favouring exercise + education vs. written educational material. Differences did not remain significant at follow-up.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is more effective than a comparison intervention (written educational material) in improving stroke awareness in the acute phase of stroke recovery.

Walking endurance
Effective
1b

One high quality RCT (Moren et al., 2016) investigated the effect of aerobic exercise on walking endurance in the acute phase of stroke recovery. This high quality RCT randomized patients to receive physical activity or no treatment; both groups received usual care. Walking endurance was measured by the 6 Minute Walk Test at follow-up (3 and 6 months). A significant between-group difference was found at 6-month follow-up, favouring physical activity vs. no treatment.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that physical activity is more effective than no treatment in improving long-term walking endurance of patients in the acute phase of stroke recovery.

Chronic phase

Balance
Conflicting
4

Four high quality RCTs (Pang et al., 2005; Liu-Ambrose & Eng, 2015; Lee et al., 2015; Moore et al., 2015) and one fair quality RCT (Lund et al., 2018) investigated the effect of aerobic exercise on balance in the chronic stage of stroke recovery.

The first high quality RCT (Pang et al., 2005) randomized patients to receive a community-based fitness and mobility exercise program (FAME) or a seated upper extremity program. Balance was measured by the Berg Balance Scale (BBS) at post-treatment (19 weeks). No significant between group difference was found.

The second high quality RCT (Liu-Ambrose & Eng, 2015) randomized patients to receive the FAME program or usual care. Balance was measured by the BBS at mid-treatment (3 months) and post-treatment (6 months). No significant between group difference was found at either time point.

The third high quality RCT (Lee et al., 2015) randomized patients to receive aerobic + resistance exercise training or light physical activity. Balance was measured by the Chair Sit and Reach Test and the Functional Reach Test at post-treatment (16 weeks). Significant between-group differences were found in both measures of balance, favouring aerobic + resistance exercise training vs. light physical activity.

The fourth high quality RCT (Moore et al., 2015) randomized patients to receive a fitness and mobility exercise program (adapted from the FAME program) or time-matched stretching. Balance was measured by the BBS at post-treatment (19 weeks). A significant between-group difference was found, favouring the fitness and mobility exercise program vs. stretching.

The fair quality RCT (Lund et al., 2018) randomized patients to receive aerobic training, resistance training or upper extremity training. Balance was measured by the BBS at post-treatment (12 weeks). No significant between-group differences were found.

Conclusion: There is conflicting evidence (Level 4) regarding the effectiveness of aerobic exercise in improving balance in the chronic stage of stroke recovery. While two high quality RCTs found that the FAME exercise program was not more effective than comparison interventions (seated upper extremity program, usual care), two high quality RCTs found that other aerobic exercise programs (aerobic + resistance exercise training, exercises adapted from the FAME program) were more effective than comparison interventions (light physical activity, stretching).

Cardiovascular fitness parameters
Conflicting
4

Five high quality RCTs (Pang et al., 2005; Gordon, Wilks & McCaw-Binns, 2013; Tang et al., 2014 and 2016; Lee et al., 2015; Moore et al., 2015) and two fair quality RCTs (Severinsen et al., 2014; Lund et al., 2018) investigated the effect of aerobic exercise on cardiovascular fitness parameters in the chronic stage of stroke recovery.

The first high quality RCT (Pang et al., 2005) randomized patients to receive a community-based fitness and mobility exercise program (FAME) or a seated upper extremity program. Maximal oxygen consumption was measured by the maximal exercise test on the Excalibur cycle ergometer at post-treatment (19 weeks). A significant between-group difference was found, favouring the FAME program vs. seated upper extremity exercises.

The second high quality RCT (Gordon, Wilks & McCaw-Binns, 2013) randomized patients to receive aerobic exercise or massage. Resting heart rate was measured by heart monitor at post-treatment (12 weeks) and follow-up (3 months). A significant between-group difference was found at follow-up only, favouring aerobic exercise vs. massage.

The third high quality RCT (Tang et al., 2014; 2016) randomized patients to receive aerobic training or balance/flexibility training. Peak oxygen consumption was measured by graded maximal exercise test using a leg cycle ergometer at post-treatment (6 months). No significant between-group difference was found.

The fourth high quality RCT (Lee et al., 2015) randomized patients to receive aerobic + resistance exercise training or light physical activity. Cardiovascular parameters (peripheral systolic blood pressure (SBP) / diastolic blood pressure (DBP), central SBP / DBP, Pulse Wave Velocity (PWV), Augmentation Index – AIx@75) were measured at post-treatment (16 weeks). Significant between-group differences were found for three cardiovascular fitness parameters (central DBP, PWV, AIx@75), favouring aerobic + resistance training vs. light physical activity.

The fifth high quality RCT (Moore et al., 2015) randomized patients to receive a fitness and mobility exercise program or time-matched stretching. Cardiovascular parameters (Peak oxygen consumption, Peak work rate, SBP, DBP) were measured by the maximal progressive recumbent bicycle exercise test and the semi-automated sphygmomanometer at post-treatment (19 weeks). Significant between-group differences were found for three cardiovascular fitness parameters (Peak oxygen consumption, Peak work rate, DBP), favouring the fitness and mobility exercise program vs. stretching.

The first fair quality RCT (Severinsen et al., 2014) randomized patients to receive aerobic training, resistance training or upper extremity training. Peak aerobic capacity (VO2 peak) was measured by the maximal progressive stepwise cycle ergometer test at post-treatment (12 weeks) and at follow-up (6 months). Significant between-group differences were found at post-treatment, favouring aerobic training vs. resistance training and aerobic training vs. upper extremity training; differences were not maintained at follow-up.

The second fair quality RCT (Lund et al., 2018) randomized patients to receive aerobic training, resistance training or upper extremity training. Cardiovascular fitness parameters (peak oxygen update, resting HR, maximal HR) were measured by the maximal progressive cycle ergometer test and a heart rate monitor at post-treatment (12 weeks). No significant between-group differences were found.

Conclusion: There is conflicting evidence (Level 4) regarding the effectiveness of aerobic exercise programs on cardiovascular fitness in the chronic phase of stroke recovery. While three high quality RCTs and one fair quality RCT found that aerobic exercise programs were more effective than comparison interventions (seated upper extremity exercises, light physical activity, stretching, resistance training, upper extremity training), two high quality RCTs* and one fair quality RCT found that aerobic exercise programs were not more effective than comparison interventions (massage, balance/flexibility training, upper extremity training).
*Note:
One of the high quality RCTs found that aerobic exercises were more effective than massage in the long-term.

Cognition
Effective
1b

One high quality RCT (Moore et al., 2015) investigated the effect of aerobic exercise on cognition in the chronic stage of stroke recovery. This high quality RCT randomized patients to receive a fitness and mobility exercise program or stretching. Cognition was measured by the Addenbrooke’s Cognitive Examination Revised at post-treatment (19 weeks). A significant between-group difference was found, favouring the fitness and mobility exercise program vs. stretching.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is more effective than a comparison intervention (stretching) in improving cognition in the chronic stage of stroke recovery.

Executive functions
Conflicting
4

Two high quality RCTs (Tang et al., 2014 & 2016; Liu-Ambrose & Eng, 2015) investigated the effect of aerobic exercise on executive functions in the chronic stage of stroke recovery.

The first high quality RCT (Tang et al., 2014 & 2016) randomized patients to receive aerobic training or balance/flexibility training. Executive functions were measured by the Verbal Digit Span Test Forward & Backward (working memory), Trail Making Test B (set shifting), and Colour-Word Stroop Test (selective attention and conflict resolution) at post-treatment (6 months). No significant between-group differences were found.

The second high quality RCT (Liu-Ambrose & Eng, 2015) randomized patients to receive a community-based Fitness and Mobility Exercise (FAME) program or usual care. Executive functions were measured by the Stroop Test (selective attention and conflict resolution), Trail Making Tests – Part A and B (set shifting) and verbal digit span forward/backward test (working memory) at mid-treatment (3 months) and post-treatment (6 months). At mid-treatment a significant between-group difference was found in one measure of executive functions (Trail Making Tests), favouring the FAME program vs. usual care. At post-treatment significant between-group differences were found in two measures of executive functions (Stroop Test; verbal digit span forward/backward test), favouring the FAME program vs. usual care.

Conclusion: There is conflicting evidence (Level 4) regarding the effectiveness of aerobic exercise in improving executive functions in the chronic stage of stroke recovery. One high quality RCT found that aerobic training was not more effective than a balance/flexibility program, whereas another high quality RCT found that aerobic exercise was more effective than usual care.

Functional independence
Not effective
1b

One high quality RCT (Gordon, Wilks & McCaw-Binns, 2013) investigated the effect of aerobic exercise on functional independence in the chronic stage of stroke recovery. This high quality RCT randomized patients to receive aerobic exercise or massage. Functional independence was measured by the Barthel Index at post-treatment (12 weeks) and follow-up (3 months). No significant between-group difference was found at either time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is not more effective than a comparison intervention (massage) in improving functional independence in the chronic stage of stroke recovery.

Functional status and service use
Not effective
1b

One high quality RCT (Gordon, Wilks & McCaw-Binns, 2013) investigated the effect of aerobic exercise on functional status and service use in the chronic stage of stroke recovery. This high quality RCT randomized patients to receive aerobic exercise or massage. Functional status and service use were measured by the Older Americans Resources and Services at post-treatment (12 weeks) and follow-up (3 months). No significant between-group difference was found at either time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is not more effective than a comparison intervention (massage) in improving functional status and service use in the chronic stage of stroke recovery.

Grip strength
Effective
1b

One high quality RCT (Lee et al., 2015) investigated the effect of aerobic exercise on grip strength in the chronic stage of stroke recovery. This high quality RCT randomized patients to receive aerobic + resistance exercise training or light physical activity. Grip strength of the unaffected hand was measured by handheld dynamometer at post-treatment (16 weeks). A significant between-group difference was found, favouring aerobic + resistance exercise training vs. light physical activity.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise with resistance training is more effective than a comparison intervention (light physical activity) in improving grip strength of the unaffected hand in the chronic stage of stroke recovery.

Health-related quality of life
Effective
1b

One high quality RCTs (Gordon, Wilks & McCaw-Binns, 2013) investigated the effect of aerobic exercise on health-related quality of life (HRQoL) in the chronic stage of stroke recovery. This high quality RCT randomized patients to receive aerobic exercise or massage. HRQoL was measured by the Short-Form-36 (SF-36: Physical health component, Mental health component) at post-treatment (12 weeks) and follow-up (3 months). A significant between-group difference was found at post-treatment (SF-36: Physical health component), favouring aerobic exercise vs. massage.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is more effective than a comparison intervention (massage) in improving health-related quality of life in the chronic stage of stroke recovery.
Note:
Between-group differences were only significant for one measure of health-related quality of life.

Mobility
Not effective
1b

One high quality RCT (Lee et al., 2015) investigated the effect of aerobic exercise on mobility in the chronic stage of stroke recovery. This high quality RCT randomized patients to receive aerobic and resistance exercise training or light physical activity. Mobility was measured by the Timed Up and Go Test at post-treatment (16 weeks). No significant between-group difference was found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is not more effective than a comparison intervention (light physical activity) in improving mobility in the chronic stage of stroke recovery.

Mood and affect
Not effective
1b

One high quality RCT (Liu-Ambrose & Eng, 2015) investigated the effect of aerobic exercise on mood and affect in the chronic stage of stroke recovery. This high quality RCT randomized patients to receive a community-based Fitness and Mobility Exercise (FAME) program or usual care. Mood and affect were measured by the 17-item Stroke Specific Geriatric Depression Scale at mid-treatment (3 months) and post-treatment (6 months). No significant between-group difference was found at either time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is not more effective than a comparison intervention (usual care) in improving mood and affect in the chronic stage of stroke recovery.

Muscle strength - lower extremities
Conflicting
4

Three high quality RCTs (Pang et al., 2005; Gordon, Wilks & McCaw-Binns, 2013; Lee et al., 2015) and two fair quality RCTs (Severinsen et al., 2014; Lund et al., 2018) investigated the effect of aerobic exercise on lower extremities muscle strength in the chronic stage of stroke recovery.

The first high quality RCT (Pang et al., 2005) randomized patients to receive a community-based fitness and mobility exercise program (FAME) or a seated upper extremity program. Isometric knee extension (paretic, non-paretic) was measured by dynamometer at post-treatment (19 weeks). A significant between group difference was found, favouring the FAME program vs. seated upper extremity exercises.

The second high quality RCT (Gordon, Wilks & McCaw-Binns, 2013) randomized patients to receive aerobic exercise or massage. Lower extremity strength (paretic, non-paretic) was measured by the Motricity Index at post-treatment (12 weeks) and follow-up (3 months). No significant between-group difference was found at either time point.

The third high quality RCT (Lee et al., 2015) randomized patients to receive aerobic + resistance exercise training or light physical activity. Lower extremity strength was measured by the 30-sec Chair Stand Test at post-treatment (16 weeks). A significant between-group difference was found, favouring aerobic + resistance exercise training vs. light physical activity.

The first fair quality RCT (Severinsen et al., 2014) randomized patients to receive aerobic training, resistance training or upper extremity training. Maximal isometric knee strength (paretic, non-paretic) was measured by dynamometer at post-treatment (12 weeks) and follow-up (6 weeks). A significant between-group difference in knee strength (paretic, non-paretic) was found at post-treatment, favouring resistance training vs. aerobic training; this between-group difference was maintained at follow-up. There was no significant difference in knee strength between aerobic training vs. upper extremity training.
Note: A significant between-group difference in knee strength (non-paretic only) was found at post-treatment, favouring resistance training vs. upper extremity exercises; this difference was maintained at follow-up.

The second fair quality RCT (Lund et al., 2018) randomized patients to receive aerobic training, resistance training or upper extremity training. Knee strength (paretic, non-paretic) was measured by dynamometer at post-treatment (12 weeks). No significant between-group differences were found.

Conclusion: There is conflicting evidence (Level 4) regarding the effectiveness of aerobic exercise on lower extremity strength in the chronic stage of stroke recovery. While two high quality RCTs found that aerobic exercise was more effective than comparison interventions (seated upper extremity program, light physical activity), one high quality RCT and two fair quality RCTs found that aerobic exercise was not more effective than comparison interventions (massage, upper extremity training, resistance training).
Note
: In fact, one of the fair quality RCTs found that resistance training was more effective than aerobic training for improving knee strength.

Physical activity
Not effective
1b

One high quality RCT (Pang et al., 2005) and one fair quality RCT (Shaughnessy, Michael & Resnick, 2012) investigated the effect of aerobic exercise on physical activity in the chronic stage of stroke recovery.

The high quality RCT (Pang et al., 2005) randomized patients to receive a community-based fitness and mobility exercise program (FAME) or a seated upper extremity program. Physical activity was measured by the Physical Activity Scale for Individuals with Physical Disabilities at post-treatment (19 weeks). No significant between group difference was found.

The fair quality RCT (Shaughnessy, Michael & Resnick, 2012) randomized patients to receive aerobic treadmill training or stretching. Physical activity was measured by the Yale Physical Activity Survey (YPAS: Housework, Yard work, Caretaking, Moderate physical activity, Recreational activities) at post-treatment (6 months). No significant between-group difference was found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one fair quality RCT that aerobic exercise is not more effective than comparison interventions (seated upper extremity program, stretching) in improving physical activity in the chronic stage of stroke recovery.

Self-efficacy and expectations
Not effective
2a

One fair quality RCT (Shaughnessy, Michael & Resnick, 2012) investigated the effect of aerobic exercise on self-efficacy & expectations in the chronic stage of stroke recovery. This fair quality RCT randomized patients to receive aerobic treadmill training or stretching. Self-efficacy & expectations were measured by Short Self-Efficacy and Outcomes Expectations for Exercises (Outcome expectations; Self-efficacy) at post-treatment (6 months). No significant between-group difference was found.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that aerobic exercise is not more effective than a comparison intervention (stretching) in improving self-efficacy & expectations in the chronic stage of stroke recovery.

Stroke outcomes
Not effective
1b

One high quality RCT (Moore et al., 2015) and one fair quality RCT (Shaughnessy, Michael & Resnick, 2012) investigated the effect of aerobic exercise on stroke outcomes in the chronic stage of stroke recovery.

The high quality RCT (Moore et al., 2015) randomized patients to receive a fitness and mobility exercise program or time-matched stretching. Stroke outcomes were measured by the Stroke Impact Scale (SIS: Stroke recovery, Mood, Strength, Memory, Communication, Activities of daily living, Community mobility, Hand function, Participation, Physical total) at post-treatment (19 weeks). A significant between-group difference was found in only one measure (SIS: Mood), favouring the fitness and mobility exercise program vs. stretching.

The fair quality RCT (Shaughnessy, Michael & Resnick, 2012) randomized patients to receive aerobic treadmill training or stretching. Stroke outcomes were measure by the SIS (Strength, Hand function, Activities of daily living, Mobility, Communication, Emotion, Memory and thinking, Participation, Overall sum, Recovery visual analogue scale) at post-treatment (6 months). No significant between-group difference was found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one fair quality RCT that aerobic exercise is not more effective than a comparison intervention (stretching) in reducing stroke outcomes in the chronic stage of stroke recovery.

Walking endurance
Effective
1a

Six high quality RCTs (Pang et al., 2005; Gordon, Wilks & McCaw-Binns, 2013; Tang et al., 2014 & 2016; Lee et al., 2015; Liu-Ambrose & Eng, 2015; Moore et al., 2015) and 2 fair quality RCTs (Severinsen et al., 2014; Lund et al., 2018) investigated the effect of aerobic exercise on walking endurance in the chronic stage of stroke recovery.

The first high quality RCT (Pang et al., 2005) randomized patients to receive a community-based fitness and mobility exercise program (FAME) or a seated upper extremity program. Walking endurance was measured by the 6-Minute Walk Test (6MWT) at post-treatment (19 weeks). A significant between-group difference was found, favouring the FAME program vs. seated upper extremity exercises.

The second high quality RCT (Gordon, Wilks & McCaw-Binns, 2013) randomized patients to receive aerobic exercise or massage. Walking endurance was measured by the 6MWT at post-treatment (12 weeks) and follow-up (3 months). No significant between-group differences were found at either time point.

The third high quality RCT (Tang et al., 2014; 2016) randomized patients to receive aerobic training or balance/flexibility training. Walking endurance was measured by the 6MWT at post-treatment (6 months). No significant between-group difference was found.

The fourth high quality RCT (Lee et al., 2015) randomized patients to receive aerobic + resistance exercise training or light physical activity. Walking endurance was measured by the 6MWT at post-treatment (16 weeks). A significant between-group difference was found, favouring aerobic + resistance exercise training vs. light physical activity.

The fifth high quality RCT (Liu-Ambrose & Eng, 2015) randomized patients to receive the FAME program or usual care. Walking endurance was measured by the 6MWT at mid-treatment (3 months) and post-treatment (6 months). A significant between-group difference was found at post-treatment, favouring the FAME program vs. usual care.

The sixth high quality RCT (Moore et al., 2015) randomized patients to receive a fitness and mobility exercise program or time-matched stretching. Walking endurance was measured by the 6MWT at post-treatment (19 weeks). A significant between-group difference was found, favouring the fitness and mobility exercise program vs. stretching.

The first fair quality RCT (Severinsen et al., 2014) randomized patients to receive aerobic training, resistance training or upper extremity training. Walking endurance was measured by the 6MWT at post-treatment (12 weeks) and at follow-up (6 months). No significant between-group differences were found.

The second fair quality RCT (Lund et al., 2018) randomized patients to receive aerobic training, resistance training or upper extremity training. Walking endurance was measured by the 6MWT at post-treatment (12 weeks). No significant between-group differences were found.

Conclusion: There is strong evidence (Level 1a) from four high quality RCTs that aerobic exercise is more effective than comparison interventions (seated upper extremity program, light physical activity, usual care, stretching) in improving walking endurance in the chronic stage of stroke recovery.
Note
: However, two high quality RCTs and two fair quality RCTs found that aerobic exercise was not more effective than comparison interventions (massage, balance/flexibility training, resistance training, upper extremity training).

Walking speed
Effective
1a

Two high quality RCTs (Lee et al., 2015; Moore et al., 2015) and two fair quality RCTs (Severinsen et al., 2014; Lund et al., 2018) investigated the effect of aerobic exercise on walking speed in the chronic stage of stroke recovery.

The first high quality RCT (Lee et al., 2015) randomized patients to receive aerobic + resistance exercise training or light physical activity. Walking speed was measured by the 10 Meter Walk Test (10mWT) at post-treatment (16 weeks). A significant between-group difference was found, favouring aerobic + resistance exercise training vs. light physical activity.

The second high quality RCT (Moore et al., 2015) randomized patients to receive a fitness and mobility exercise program or time-matched stretching. Walking speed was measured by the 10mWT at post-treatment (19 weeks). A significant between-group difference was found, favouring the fitness and mobility exercise program vs. stretching.

The first fair quality RCT (Severinsen et al., 2014) randomized patients to receive aerobic training, resistance training or upper extremity training. Walking speed was measured by the 10mWT at post-treatment (12 weeks) and at follow-up (6 months). No significant between-group differences were found at post-treatment. Significant between-group differences were found at follow-up, favouring resistance training vs. aerobic training, and upper extremity training vs. aerobic training.

The second fair quality RCT (Lund et al., 2018) randomized patients to receive aerobic training, resistance training or upper extremity training. Walking speed was measured by the 10mWT at post-treatment (12 weeks). No significant between-group differences were found.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs that aerobic exercise is more effective than comparison interventions (light physical activity, stretching) in improving walking speed in the chronic stage of stroke recovery.
Note: Two fair quality RCTs found that aerobic exercise was not more effective than comparison interventions (resistance training, upper extremity training). In fact, one fair quality RCT found that both resistance training and upper extremity training were more effective than aerobic exercise for improving walking speed.

Phase not specific to one period

Balance
Effective
1b

One high quality RCT (Sandberg et al., 2016) and one quasi-experimental design study (Marsden et al., 2016) investigated the effect of aerobic exercise on balance in patients with stroke.

The high quality RCT (Sandberg et al., 2016) randomized patients with acute/subacute stroke to receive an aerobic exercise program or no exercise program. Balance was measured by the Single Leg Stance Test (SLST: Right/left with eyes closed/open) at post-treatment (12 weeks). A significant between-group difference was found for three measures of balance (SLST: Right leg eyes open, Right leg eyes closed, Left leg eyes open), favouring aerobic exercise program vs. no exercise program.

The quasi-experimental design study (Marsden et al., 2016) assigned patients with acute, subacute or chronic stroke to receive a home- and community-based exercise program with aerobic content or usual care. Balance was measured by the Step Test (Right, Left) at post-treatment (12 weeks). A significant between-group difference was found on both measures of balance, favouring the aerobic exercise program vs. usual care.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one quasi-experimental study that aerobic exercise is more effective than comparison interventions (no exercise program, usual care) in improving balance in patients with stroke.

Cardiovascular fitness parameters
Effective
1a

Two high-quality RCTs (Wang et al., 2014; Sandberg et al., 2016) and one quasi-experimental design study (Marsden et al., 2016) investigated the effect of aerobic exercise on cardiovascular fitness parameters in patients with stroke.

The first high quality RCT (Wang et al., 2014) randomized patients with subacute/chronic stroke to receive low-intensity aerobic training or no training; both groups received conventional rehabilitation. Cardiovascular fitness parameters (Resting heart rate (RHR), Peak heart rate (PHR), Exercise test time) were measured at post-treatment (6 weeks). A significant between-group difference was found on one measure (exercise test time), in favour of aerobic training vs. no aerobic training.

The second high quality RCT (Sandberg et al., 2016) randomized patients with acute/subacute stroke to receive an aerobic exercise program or no exercise program. A cardiovascular fitness parameter (Peak work rate) was measured by the symptom-limited graded cycle ergometer test at post-treatment (12 weeks). A significant between-group difference was found, favouring aerobic exercise program vs. no exercise program.

The quasi-experimental design study (Marsden et al., 2016) assigned patients with acute, subacute or chronic stroke to receive a home- and community-based exercise program with aerobic content or usual care. Cardiovascular fitness parameters (Vo2peak absolute, Vo2peak relative, HR, R-value) were measured during the 6 Minute Walk Test (6MWT), Shuttle Walk Test and Cycle Progressive Exercise Test (also Duration, Workload measures) at post-treatment (12 weeks). A significant between-group difference was found in one parameter (6MWT: Vo2peak absolute, relative), favouring an aerobic exercise program vs. usual care.

Conclusion: There is strong evidence (Level 1a) from two high quality RCTs and one quasi-experimental study that aerobic exercise is more effective than no treatment for improving some cardiovascular fitness parameters (e.g. exercise test time, peak work rate, peak oxygen uptake) in patients with stroke.

Cognition
Not effective
2b

One fair quality RCT (Nave et al., 2019) investigated the effect of aerobic exercise on cognition in patients with stroke. This fair quality RCT randomized patients with acute/subacute stroke to receive aerobic physical fitness training using the PHYS-STROKE program or relaxation. Cognition was measured by the Montreal Cognitive Assessment at post-treatment (4 weeks) and follow-up (3, 6 months). No significant between group difference was found at any time point.

Conclusion: There is limited evidence (Level 2b) from one fair quality RCT that aerobic exercise is not more effective than a comparison intervention (relaxation) in improving cognition in patients with stroke.

Depression
Not effective
2b

One fair quality RCT (Nave et al., 2019) and one quasi-experimental design study (Marsden et al., 2016) investigated the effect of aerobic exercise on depression in patients with stroke.

The fair quality RCT (Nave et al., 2019) randomized patients with acute/subacute stroke to receive aerobic physical fitness training using the PHYS-STROKE program or relaxation. Depression was measured by the Centre for Epidemiological Studies Depression at post-treatment (4 weeks) and follow-up (3, 6 months). No significant between-group differences were found at any time points.

The quasi-experimental design study (Marsden et al., 2016) assigned patients with acute, subacute or chronic stroke to receive a home- and community-based exercise program with aerobic content or usual care. Depression was measured by the Patient Health Questionnaire-9 at post-treatment (12 weeks). No significant between-group difference was found.

Conclusion: There is limited evidence (Level 2b) from one fair quality RCT and one quasi-experimental study that aerobic exercise is not more effective than comparison interventions (relaxation, usual care) in reducing depression in patients with stroke.

Dexterity
Not effective
2b

One fair quality RCT (Nave et al., 2019) investigated the effect of aerobic exercise on dexterity in patients with stroke. This fair quality RCT randomized patients with acute/subacute stroke to receive aerobic physical fitness training using the PHYS-STROKE program or relaxation. Dexterity was measured by the Box and Block Test at post-treatment (4 weeks) and follow-up (3, 6 months). No significant between-group differences were found at any time points.

Conclusion: There is limited evidence (Level 2b) from one fair quality RCT that aerobic exercise is not more effective than a comparison intervention (relaxation) in improving dexterity in patients with stroke.

Executive functions
Not effective
2b

One fair quality RCT (Nave et al., 2019) investigated the effect of aerobic exercise on executive functions in patients with stroke. This fair quality RCT randomized patients with acute/subacute stroke to receive aerobic physical fitness training using the PHYS-STROKE program or relaxation. Executive functions were measured by the Trail Making Test (TMT – A, B) at post-treatment (4 weeks) and follow-up (3, 6 months) and the Regensburger Wort-Flüssigkeits-Test at follow-up (3 months). No significant between group differences were found on any of the measures at any time point.

Conclusion: There is limited evidence (Level 2b) from one fair quality RCT that aerobic exercise is not more effective than a comparison intervention (relaxation) in improving executive functions in patients with stroke.

Exercise capacity
Not effective
2b

One quasi-experimental design study (Marsden et al., 2016) investigated the effect of aerobic exercise on exercise capacity in patients with stroke. This quasi-experimental design study assigned patients with acute, subacute or chronic stroke to receive a home- and community-based exercise program with aerobic content or usual care. Exercise capacity was measured by the Shuttle Walk Test at post-treatment (12 weeks). No significant between-group difference was found.

Conclusion: There is limited evidence (Level 2b) from one quasi-experimental study that aerobic exercise is not more effective than a comparison intervention (usual care) in improving exercise capacity in patients with stroke.

Fatigue
Not effective
2b

One quasi-experimental design study (Marsden et al., 2016) investigated the effect of aerobic exercise on fatigue in patients with stroke. This quasi-experimental study assigned patients with acute, subacute or chronic stroke to receive a home- and community-based exercise program with aerobic content or usual care. Fatigue was measured by the Fatigue Assessment Scale at post-treatment (12 weeks). No significant between-group difference was found.

Conclusion: There is limited evidence (Level 2b) from one quasi-experimental study that aerobic exercise is not more effective than a comparison intervention (usual care) in reducing fatigue in patients with stroke.

Functional independence
Effective
1b

One high quality RCT (Wang et al., 2014) and one fair quality RCT (Nave et al., 2019) investigated the effect of aerobic exercise on functional independence in patients with stroke.

The high quality RCT (Wang et al., 2014) randomized patients with subacute/chronic stroke to receive low-intensity aerobic training or no training; both groups received conventional rehabilitation. Functional independence was measured by the Barthel Index at post-treatment (6 weeks). A significant between-group difference was found in favour of aerobic training vs. no training.

The fair quality RCT (Nave et al., 2019) randomized patients with acute/subacute stroke to receive aerobic physical fitness training using the PHYS-STROKE program or relaxation. Functional independence was measured by the Barthel Index (change scores) and the modified Rankin Scale at post-treatment (4 weeks) and follow-up (3, 6 months). No significant between-group differences were found at any time points.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is more effective than no training in improving functional independence in patients with stroke.

Gait parameters
Not effective
2a

One fair quality RCT (Nave et al., 2019) investigated the effects of aerobic exercise in improving gait parameters in patients with stroke. This fair quality RCT randomized patients with acute/subacute stroke to receive aerobic physical fitness training using the PHYS-STROKE program or relaxation. Gait parameters (Number of steps/day, Step length, Step Cadence, Gait Energy Cost) were measured at post-treatment (4 weeks) and follow-up (3, 6 months). No significant between-group differences were found at any time points.

Conclusion: There is limited evidence (Level 2a) from one fair quality RCT that aerobic exercise is not more effective than a comparison intervention (relaxation) in improving gait parameters in patients with stroke.

Health-related quality of life
Effective
1b

One high quality RCT (Sandberg et al., 2016), one fair quality RCT (Nave et al., 2019) and one quasi-experimental design study (Marsden et al., 2016) investigated the effect of aerobic exercise on health-related quality of life (HRQoL) in patients with stroke.

The high quality RCT (Sandberg et al., 2016) randomized patients with acute/subacute stroke to receive an aerobic exercise program or no exercise program. HRQoL was measured by the European Quality of Life Scale (EQ-5D: Total score; Visual analogue Scale) at post-treatment (12 weeks). A significant between-group difference was found for one measure of HRQoL (EQ-5D: Visual analogue scale), favouring aerobic exercise program vs. no exercise program.

The fair quality RCT (Nave et al., 2019) randomized patients with acute/subacute stroke to receive aerobic physical fitness training using the PHYS-STROKE program or relaxation. HRQoL was measured by the EuroQoL Quality of Life Questionnaire 5D-5L at post-treatment (4 weeks) and follow-up (3, 6 months). No significant between-group differences were found at any time points.

The quasi-experimental study (Marsden et al., 2016) assigned patients with acute, subacute or chronic stroke to receive a home- and community-based exercise program with aerobic content or usual care. HRQoL was measured by the Stroke and Aphasia Quality of Life-39 at post-treatment (12 weeks). No significant between-group difference was found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is more effective than a comparison intervention (no exercise program) in improving health-related quality of life in patients with stroke.

Mobility
Not effective
1b

One high quality RCT (Sandberg et al., 2016) and one fair quality RCT (Nave et al., 2019) investigated the effect of aerobic exercise on mobility in patients with stroke.

The high quality RCT (Sandberg et al., 2016) randomized patients with acute/subacute stroke to receive an aerobic exercise program or no exercise program. Mobility was measured by the Timed Up and Go test at post-treatment (12 weeks). A significant between-group difference was found, favouring aerobic exercise program vs. no exercise program.

The fair quality RCT (Nave et al., 2019) randomized patients with acute/subacute stroke to receive aerobic physical fitness training using the PHYS-STROKE program or relaxation. Mobility was measured by the Rivermead Mobility Index, Use of walking aids and Functional Ambulation Category, at post-treatment (4 weeks) and follow-up (3, 6 months). No significant between group differences were found on any of the measures at any time point.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is more effective than a comparison intervention (no exercise program) in improving mobility in patients with stroke.

Motor function
Effective
1b

One high quality RCT (Wang et al., 2014) investigated the effect of aerobic exercise on motor function in patients with stroke. This high quality RCT randomized patients with subacute/chronic stroke to receive low-intensity aerobic training or no training; both groups received conventional rehabilitation. Motor function was measured by the Fugl Meyer Assessment (FMA: Total motor score) at post-treatment (6 weeks). A significant between-group difference was found, favouring aerobic exercise vs. no training.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is more effective than no training (conventional rehabilitation alone) in improving motor function in patients with stroke.

Motor function - lower extremity
Effective
1b

One high quality RCT (Wang et al., 2014) investigated the effect of aerobic exercise on lower extremity motor function in patients with stroke. This high quality RCT randomized patients with subacute/chronic stroke to receive low-intensity aerobic training or no training; both groups received conventional rehabilitation. Lower extremity motor function was measured by the Fugl Meyer Assessment (FMA: Lower extremity score) at post-treatment (6 weeks). A significant between-group difference was found, favouring aerobic exercise vs. no training.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is more effective than no training in improving lower extremity motor function in patients with stroke.

Motor function - upper extremity
Not effective
1b

One high quality RCT (Wang et al., 2014) and one fair quality RCT (Nave et al., 2019) investigated the effect of aerobic exercise on upper extremity motor function in patients with stroke.

The high quality RCT (Wang et al., 2014) randomized patients with subacute/chronic stroke to receive low-intensity aerobic training or no training; both groups received conventional rehabilitation. Upper extremity motor function was measured by the Fugl Meyer Assessment (FMA: Upper extremity score) at post-treatment (6 weeks). No significant between-group difference was found.

The fair quality RCT (Nave et al., 2019) randomized patients with acute/subacute stroke to receive aerobic physical fitness training using the PHYS-STROKE program or relaxation. Upper extremity motor function was measured by the Rivermead Mobility Index (RMI – arm score) at post-treatment (4 weeks) and follow-up (3, 6 months). A significant between-group differences was found at 6-month follow up only, favouring aerobic physical fitness training vs. relaxation.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one fair quality RCT that aerobic exercise is not more effective than comparison interventions (no training, relaxation) in improving upper extremity motor function in patients with stroke.

Muscle strength
Not effective
2b

One fair quality RCT (Nave et al., 2019) investigated the effect of aerobic exercise on muscle strength in patients with stroke. This fair quality RCT randomized patients with acute/subacute stroke to receive aerobic physical fitness training using the PHYS-STROKE program or relaxation. Muscle strength was measured by the Medical Research Council (MRC) Scale at post-treatment (4 weeks) and follow-up (3, 6 months). No significant between-group difference was found at any time points.

Conclusion: There is limited evidence (Level 2b) from one fair quality RCT that aerobic exercise is not more effective than a comparison intervention (relaxation) in improving muscle strength in patients with stroke.

Sleep quality
Not effective
2b

One fair quality RCT (Nave et al., 2019) investigated the effect of aerobic exercise on sleep quality in patients with stroke. This fair quality RCT randomized patients with acute/subacute stroke to receive aerobic physical fitness training using the PHYS-STROKE program or relaxation. Sleep quality was measured by the Pittsburgh Sleep Quality Score at post-treatment (4 weeks) and follow-up (3, 6 months). No significant between-group difference was found at any time point.

Conclusion: There is limited evidence (Level 2b) from one fair quality RCT that aerobic exercise is not more effective than a comparison intervention (relaxation) in improving sleep quality in patients with stroke.

Spasticity
Effective*
2b

One fair quality RCT (Nave et al., 2019) investigated the effect of aerobic exercise on spasticity in patients with stroke. This fair quality RCT randomized patients with acute/subacute stroke to receive aerobic physical fitness training using the PHYS-STROKE program or relaxation. Spasticity was measured by the Resistance to Passive Movement Scale (REPAS) at post-treatment (4 weeks) and follow-up (3, 6 months). A significant between-group differences was found at follow up (6 months), favouring aerobic physical fitness training vs. relaxation.

Conclusion: There is limited evidence (Level 2b) from one fair quality RCT that aerobic exercise is more effective, in the long term*, than a comparison intervention (relaxation) in reducing spasticity in patients with stroke.

Stroke outcomes
Effective
1b

One high quality RCT (Sandberg et al., 2016) investigated the effect of aerobic exercise on stroke outcomes in patients with stroke. This high quality RCT randomized patients with acute/subacute stroke to receive an aerobic exercise program or no exercise program. Stroke outcomes were measured by the Stroke Impact Scale (SIS: Daily activities, Recovery) at post-treatment (12 weeks). A significant between-group difference was found in one measure (SIS: Recovery), favouring aerobic exercise program vs. no exercise.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is more effective than a comparison intervention (usual care) in reducing some stroke outcomes in patients with stroke.

Walking endurance
Effective
1b

One high-quality RCT (Sandberg et al., 2016), one fair quality RCT (Nave et al., 2019) and one quasi-experimental study (Marsden et al., 2016) investigated the effect of aerobic exercise on walking endurance in patients with stroke.

The high quality RCT (Sandberg et al., 2016) randomized patients with acute/subacute stroke to receive an aerobic exercise program or no exercise program. Walking endurance was measured by the  6 Minute Walk Test (6MWT) at post-treatment (12 weeks). A significant between-group difference was found, favouring aerobic exercise vs. no exercise.

The fair quality RCT (Nave et al., 2019) randomized patients with acute/subacute stroke to receive aerobic physical fitness training using the PHYS-STROKE program or relaxation. Walking endurance was measured by the 6MWT at post-treatment (4 weeks) and follow-up (3, 6 months). No significant between-group difference was found at any time points.

The quasi-experimental study design (Marsden et al., 2016) assigned patients with acute, subacute or chronic stroke to receive a home- and community-based exercise program with aerobic content or usual care. Walking endurance was measured by the 6MWT at post-treatment (12 weeks). A significant between-group difference was found, favouring aerobic exercise vs. usual care.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT and one quasi-experimental study design that aerobic exercise is more effective than comparison interventions (no exercise, usual care) in improving walking endurance in patients with stroke.

Walking speed
Effective
1b

One high quality RCT (Sandberg et al., 2016), one fair quality RCT (Nave et al., 2019) and one quasi-experimental study design (Marsden et al., 2016) investigated the effect of aerobic exercise on walking speed in patients with stroke.

The high quality RCT (Sandberg et al., 2016) randomized patients with acute/subacute stroke to receive an aerobic exercise program or no exercise program. Walking speed was measured by the  10 Minute Walk Test (10MWT) at post-treatment (12 weeks). A significant between-group difference was found, favouring aerobic exercise vs. no exercise.

The fair quality RCT (Nave et al., 2019) randomized patients with acute/subacute stroke to receive aerobic physical fitness training using the PHYS-STROKE program or relaxation. Walking speed was measured by the 10MWT post-treatment (4 weeks) and follow-up (3, 6 months). No significant between group difference was found at any time point.

The quasi-experimental study design (Marsden et al., 2016) assigned patients with acute, subacute or chronic stroke to receive a home- and community-based exercise program with aerobic content or usual care. Walking speed was measured by the 10MWT (Fast, Self-selected speed) at post-treatment (12 weeks). No significant between-group difference was found.

Conclusion: There is moderate evidence (Level 1b) from one high quality RCT that aerobic exercise is more effective than no exercise in improving walking speed in patients with stroke.

References

Faulkner, J., McGonigal, G., Woolley, B., Stoner, L., Wong, L., & Lambrick, D. (2015). A randomized controlled trial to assess the psychosocial effects of early exercise engagement in patients diagnosed with transient ischaemic attack and mild, non-disabling stroke. Clinical Rehabilitation29(8), 783-794.
https://journals.sagepub.com/doi/full/10.1177/0269215514555729?casa_token=R7BB6cYVToAAAAAA%3AOFN0VReAB8HomN2EmpuqI_RFSUW4EvcKLEpYsPlH-wmXYfvkRVv3fL-L7f-bbXGIlcrcGYQHdXrLmA

Gordon, C. D., Wilks, R., & McCaw-Binns, A. (2013). Effect of aerobic exercise (walking) training on functional status and health-related quality of life in chronic stroke survivors: a randomized controlled trial. Stroke, 44(4), 1179-1181.
https://www.ahajournals.org/doi/full/10.1161/STROKEAHA.111.000642

Lee, Y. H., Park, S. H., Yoon, E. S., Lee, C. D., Wee, S. O., Fernhall, B., & Jae, S. Y. (2015). Effects of combined aerobic and resistance exercise on central arterial stiffness and gait velocity in patients with chronic poststroke hemiparesis. American Journal of Physical Medicine & Rehabilitation94(9), 687-695.
https://journals.lww.com/ajpmr/fulltext/2015/09000/Effects_of_Combined_Aerobic_and_Resistance.3.aspx?casa_token=KwoozWjzPWwAAAAA:qkb9LlDGH-rMj5lOOEvpd7zIkx7zeRoZgmXHDRIACPyqBOgU4lzzJnbDaTd87F5Hgoc3Ukow6jCBAOUuUKoCau7PqcY

Liu-Ambrose, T., & Eng, J. J. (2015). Exercise training and recreational activities to promote executive functions in chronic stroke: a proof-of-concept study. Journal of Stroke and Cerebrovascular Diseases24(1), 130-137.
https://www.sciencedirect.com/science/article/abs/pii/S1052305714004169

Lund, C., Dalgas, U., Grønborg, T. K., Andersen, H., Severinsen, K., Riemenschneider, M., & Overgaard, K. (2018). Balance and walking performance are improved after resistance and aerobic training in persons with chronic stroke. Disability and rehabilitation, 40(20), 2408-2415.
https://www.tandfonline.com/doi/full/10.1080/09638288.2017.1336646?casa_token=-jl-IeFq0cwAAAAA%3AFp7OwEr3QabrfxNE9vWV4w6NJZFNrK0FFaoIFN6uRh9WuxC3_YIl2tvN7OYH7mid3DsUqUYcqVDZJQ

Marsden, D. L., Dunn, A., Callister, R., McElduff, P., Levi, C. R., & Spratt, N. J. (2016). A home-and community-based physical activity program can improve the cardiorespiratory fitness and walking capacity of stroke survivors. Journal of Stroke and Cerebrovascular Diseases, 25(10), 2386-2398.
https://www.sciencedirect.com/science/article/pii/S1052305716301240?casa_token=OTKqoIi5B6oAAAAA:wgMBHD1qRkcQTDIIAdW8kfzhih_3r343EUI_I3TE37wXRVSX_p3-kRPuNC3_Feh0vplp63m0bNg

Moore, S. A., Hallsworth, K., Jakovljevic, D. G., Blamire, A. M., He, J., Ford, G. A., … & Trenell, M. I. (2015). Effects of community exercise therapy on metabolic, brain, physical, and cognitive function following stroke: a randomized controlled pilot trial. Neurorehabilitation and neural repair, 29(7), 623-635.
https://journals.sagepub.com/doi/full/10.1177/1545968314562116

Morén, C., Welmer, A. K., Hagströmer, M., Karlsson, E., & Sommerfeld, D. K. (2016). The effects of “physical activity on prescription” in persons with transient ischemic attack: a randomized controlled study. Journal of Neurologic Physical Therapy40(3), 176-183.
https://www.ingentaconnect.com/content/wk/npt/2016/00000040/00000003/art00003

Nave, A. H., Rackoll, T., Grittner, U., Bläsing, H., Gorsler, A., Nabavi, D. G., … & Flöel, A. (2019). Physical Fitness Training in Patients with Subacute Stroke (PHYS-STROKE): multicentre, randomised controlled, endpoint blinded trial. BMJ, 366.
https://www.bmj.com/content/366/bmj.l5101.long

Pang, M. Y., Eng, J. J., Dawson, A. S., McKay, H. A., & Harris, J. E. (2005). A community‐based fitness and mobility exercise program for older adults with chronic stroke: A randomized, controlled trial. Journal of the American Geriatrics Society, 53(10), 1667-1674.
https://agsjournals.onlinelibrary.wiley.com/doi/full/10.1111/j.1532-5415.2005.53521.x?casa_token=DoqAsUl08_AAAAAA%3A0y0HmQGzsD6pcdOW8racpcePu-yJLUmwsp4FImdylTm-C8_pHOglUxgIDzz5tnUHOOlfi2m7OFAHCh8

Sandberg, K., Kleist, M., Falk, L., & Enthoven, P. (2016). Effects of twice-weekly intense aerobic exercise in early subacute stroke: a randomized controlled trial. Archives of physical medicine and rehabilitation, 97(8), 1244-1253.
https://www.sciencedirect.com/science/article/pii/S0003999316001052?casa_token=XJu7jzePsdMAAAAA:6pxMTWTNSxm5b6Grgh7QNcb8yB9hCKRQ1Y7zLW2nXFSYAeleqUShrz0WaADKN5Zt2Udm4rfI0O8

Severinsen, K., Jakobsen, J. K., Pedersen, A. R., Overgaard, K., & Andersen, H. (2014). Effects of resistance training and aerobic training on ambulation in chronic stroke. American journal of physical medicine & rehabilitation, 93(1), 29-42.
https://journals.lww.com/ajpmr/Fulltext/2014/01000/Effects_of_Resistance_Training_and_Aerobic.3.aspx?casa_token=mB1Gwq9fLDgAAAAA:pLqqIJFYJWzoq_bAd4de_t5rgAHqgYHuSe4BZCikntkwsUv9dggoHkMiquX7UiX-DuDiVa0ltkg3WSoaL3jKxJa1ah8

Shaughnessy, M., Michael, K., & Resnick, B. (2012). Impact of treadmill exercise on efficacy expectations, physical activity, and stroke recovery. The Journal of neuroscience nursing: journal of the American Association of Neuroscience Nurses, 44(1), 27.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3249605/

Tang, A., Eng, J. J., Krassioukov, A. V., Madden, K. M., Mohammadi, A., Tsang, M. Y., & Tsang, T. S. (2014). Exercise-induced changes in cardiovascular function after stroke: a randomized controlled trial. International Journal of Stroke, 9(7), 883-889.
https://journals.sagepub.com/doi/full/10.1111/ijs.12156?casa_token=BaEy8XeTLe8AAAAA%3AV1kHt11rKpBMnXOn1Vf4lap-4YXYAs8qECRbWkmOzU8eDZCSggtA4_buQjMTW3G_btKsnTGjZbkK_g

Tang, A., Eng, J. J., Tsang, T. S., & Liu-Ambrose, T. (2016). High-and low-intensity exercise do not improve cognitive function after stroke: A randomized controlled trial. Journal of rehabilitation medicine, 48(10), 841-846.
https://www.ingentaconnect.com/contentone/mjl/sreh/2016/00000048/00000010/art00002

Wijkman, M. O., Sandberg, K., Kleist, M., Falk, L., & Enthoven, P. (2018). The exaggerated blood pressure response to exercise in the sub‐acute phase after stroke is not affected by aerobic exercise. The Journal of Clinical Hypertension20(1), 56-64.
https://onlinelibrary.wiley.com/doi/full/10.1111/jch.13157

Wang, Z., Wang, L., Fan, H., Lu, X., & Wang, T. (2014). Effect of low-intensity ergometer aerobic training on glucose tolerance in severely impaired nondiabetic stroke patients. Journal of Stroke and Cerebrovascular Diseases, 23(3), e187-e193.
https://www.sciencedirect.com/science/article/pii/S1052305713003959?casa_token=GoBcYRstDJ8AAAAA:IJo9a1F9CwFK_YLGBHuYDi0q-ca7d6CUdht3OvsR6hI-MgDQIxyJV2khbzTwrKl5f4bSVQZR1N4

Excluded studies:

Bo, W., Lei, M., Tao, S., Jie, L. T., Qian, L., Lin, F. Q., & Ping, W. X. (2019). Effects of combined intervention of physical exercise and cognitive training on cognitive function in stroke survivors with vascular cognitive impairment: a randomized controlled trial. Clinical Rehabilitation, 33(1), 54-63.
Reason for exclusion: The main part of the treatment focused on endurance, strength and balance (30-35 minutes) and aerobic exercise was provided only for the first 5 minutes as warm-up (cycling, jogging).

Jin, H., Jiang, Y., Wei, Q., Chen, L., & Ma, G. (2013). Effects of aerobic cycling training on cardiovascular fitness and heart rate recovery in patients with chronic stroke. NeuroRehabilitation32(2), 327-335.
Reason for exclusion: Included in treadmill module.

Jin, H., Jiang, Y., Wei, Q., Wang, B., & Ma, G. (2012). Intensive aerobic cycling training with lower limb weights in Chinese patients with chronic stroke: discordance between improved cardiovascular fitness and walking ability. Disability and Rehabilitation34(19), 1665-1671.
Reason for exclusion: Included in treadmill module.

Aquatic interventions

Evidence Reviewed as of before: 26-02-2021
Author(s)*: Annabel McDermott, OT; Tatiana Ogourtsova, PhD OT
Patient/Family Information Table of contents

Introduction

Aquatic interventions are considered applicable to post-stroke rehabilitation as the properties of water support the effects of exercise on recovery.

An early Cochrane review (Merholz, Kugler & Pohl, 2011) that looked at the effect of water-based exercise on activities of daily living (ADLs) and other clinical outcomes (walking speed, postural control, muscle strength, aerobic fitness) after stroke included 4 studies (all studies are included in this review) and concluded that there was not enough evidence at that time to determine whether water-based exercise reduced disability after stroke.

Since that time there have been several systematic reviews and meta-analyses of aquatic interventions specific to the stroke population. Recent reviews have concluded that aquatic interventions are useful for improving balance (Iatridou et al., 2018; Xie et al., 2019; Nascimento et al., 2020) and walking skills (Xie et a., 2019; Nascimento et al., 2020). A positive effect has not been found in relation to ADLs (Xie et al., 2019).

Most recently, Veldema & Jansen’s (2020) analysis of 28 studies (all studies are included in this review) concluded that aquatic therapies are more effective than no treatment for improving walking, balance, emotional status/health-related quality of life, spasticity and physiological indicators; and are more effective than land-based therapies for improving walking, balance, muscular strength, proprioception, health-related quality of life, physiological indicators and cardiorespiratory fitness.

This Stroke Engine review includes 31 studies comprised of 11 high quality RCTs, 17 fair quality RCTs and 3 quasi-experimental studies. Most studies (N=26) were conducted with participants in the chronic phase of stroke recovery; all other studies were conducted with individuals in the subacute phase of recovery. For the purpose of this review, aquatic therapy interventions are defined as any stroke rehabilitation program conducted in controlled water environments. Aquatic programs encompass lower-extremity exercises, trunk exercises, balance activities, obstacle courses, dual-task training, hydrokinesitherapy, hydrotherapy, proprioceptive exercises, treadmill training, task-oriented training, and programs that draw on defined methods (e.g. Bad Ragaz Ring method and programs using Proprioceptive Neuromuscular Facilitation, Halliwick method, Ai Chi method). Control groups include no treatment and on-land programs (e.g. conventional rehabilitation, physical therapy, upper extremity function exercises, proprioceptive exercise, aerobic exercise, obstacle training, PNF lower extremity exercises, treadmill training/backward treadmill training, task-oriented training, trunk exercises, motor dual task training).

Overall, the results from this review found strong evidence (level 1a – from two or more high quality RCTs) to indicate that that aquatic interventions improve* lower extremity muscle strength and gait during the subacute phase of recovery; and balance, mobility and walking speed in the chronic phase of stroke recovery. Further, there was moderate evidence (level 1b – from at least one high quality RCT) that aquatic interventions improve* cardiovascular fitness parameters, gait parameters, muscle activity, pain and walking endurance in the chronic phase of stroke recovery.

* More than land-based interventions or no treatment.

Patient/Family Information

What are aquatic interventions?

 Aquatic interventions are exercise programs performed in a controlled water environment (e.g. in a pool).

Aquatic therapy is also referred to as:

  • water-based therapy
  • pool therapy
  • hydrotherapy
  • hydrokinesiotherapy

Why are aquatic interventions used for?

 It is common to experience physical difficulties after a stroke, such as difficulty with walking and balance. Exercise after a stroke is very important to recovery. It is necessary to continue to exercise after a stroke, to avoid further muscle weakness and reduced fitness. Exercise can also have a positive effect on mental health and neurological health.

Aquatic interventions can be suitable for different levels of ability and recovery after stroke. Aquatic therapy is used in stroke rehabilitation because water provides a safe and comfortable environment for exercise. There are several ways in which aquatic therapy assists recovery:

  • The density and viscosity (thickness) of water provides buoyancy to support body weight. This reduces the impact of movement on joints, allows for increased mobility, and reduces the risk of falls when exercising.
  • The hydrostatic pressure of water provides resistance for muscle strengthening. This pressure also provides increased sensory input to the muscles and joints.
  • Water can provide relief for muscles and joints.

Are there different types of aquatic interventions?

 Yes, there are different types of aquatic therapy used in stroke rehabilitation. Rehabilitation clinicians may choose a specific program because of:

  • the method (e.g. task-oriented training, Halliwick method, Ai Chi method)
  • the equipment (e.g. obstacle courses, treadmills)
  • the goal (e.g. improving upper body strength, lower body strength, balance or proprioception)

How do I do aquatic therapy?

 Your stroke rehabilitation team will talk with you to determine whether aquatic therapy is available, safe, and suitable for your recovery. Your rehabilitation clinician will develop a program that addresses your specific recovery needs and goals. Your clinician will supervise your session and will instruct you on the exercises and movements. They may assist you in the pool or direct you from beside the pool, depending on safety.

Do aquatic interventions work?

Researchers have done studies to see if aquatic therapy helps people who have had a stroke. There is good evidence that aquatic therapy can improve balance and walking skills.

There is strong evidence that aquatic therapy is helpful in the subacute phase of stroke recovery (1-6 months after the stroke) for improving:

  • lower extremity muscle strength
  • gait

There is moderate to strong evidence that aquatic therapy is helpful in the chronic phase of stroke recovery (more than 6 months after the stroke) for improving:

  • balance
  • mobility
  • walking speed
  • gait
  • walking endurance
  • cardiovascular fitness
  • muscle activity
  • pain

Other studies show that aquatic therapy is also helpful for improving emotional status and quality of life (related to health).

Note: These studies showed that aquatic interventions were more effective than land-based interventions or no treatment.

Are there any side effects or risks?

There are safety risks to consider when starting aquatic therapy after stroke. Risks include:

  • Slips and falls on wet surfaces around pools. The clinician will assess the individual’s safety and mobility before starting the program. The clinician will supervise or assist the individual to enter and exit the pool safely.
  • Drowning and heat exhaustion. The clinician will closely supervise the individual when doing exercises. The clinician will monitor the pool temperature and the individual’s wellness during the session.
  • Skin irritation and infection. Aquatic therapy should be done in a pool with controlled pH levels and an environment with frequent cleaning routines. The clinician should follow hygiene and infection control procedures.

Aquatic therapy should be done under supervision of a rehabilitation clinician. The clinician will choose a program that is safe and that suits the person’s recovery.

Who provides the treatment?

Aquatic interventions are performed under the supervision of a trained clinician. The clinician will choose specific exercises that use the physical properties of water to benefit the patient. The program and exercises will be selected according to each person’s recovery stage, abilities, needs and rehabilitation goals.

Who can help me?

 It is important to obtain medical clearance from your physician before starting an exercise program after stroke. Talk with your rehabilitation team if you are interested in aquatic therapy for stroke recovery.

Clinician Information

Note: When reviewing the findings, it is important to note that they are always made according to randomized clinical trial (RCT) criteria – specifically as compared to a control group. To clarify, if a treatment is “effective” it implies that it is more effective than the control treatment to which it was compared. Non-randomized studies are no longer included when there is sufficient research to indicate strong evidence (level 1a) for an outcome.

 

Results Table

View results table

Outcomes

Subacute phase

Activities of Daily Living
Conflicting
4

Three high quality RCTs (Zhang et al., 2016; Han & Im, 2018; Lee et al., 2018) investigated the effect of aquatic interventions on Activities of Daily Living (ADLs) in the subacute phase of stroke recovery.

The first high quality RCT (Zhang et al., 2016) randomized participants to receive aquatic therapy or land-based physiotherapy. ADLs were measured using the Barthel Index at post-treatment (8 weeks). A significant between-group difference was found, in favour of aquatic therapy vs. land-based physiotherapy.

The second high quality RCT (Han & Im, 2018) randomized participants to receive aquatic treadmill training or land-based aerobic exercise. ADLs were measured using the Korean modified Barthel Index at post-treatment (6 weeks). No significant between-group difference was found.

The third high quality RCT (Lee et al., 2018) randomized participants to receive aquatic treadmill training or on-land aerobic exercise. ADLs were measured using the Korean modified Barthel Index at post-treatment (4 weeks). No significant between-group difference was found.

Conclusion: There is conflicting evidence (level 4) regarding the effectiveness of aquatic interventions on Activities of Daily Living in the subacute phase of stroke recovery. While one high quality RCT found that aquatic interventions were more effective than on-land programs, two other high quality RCTs found that an aquatic therapy program was no more effective than a land-based aerobic program.

Arterial stiffness
Not effective
1b

One high quality RCT (Lee et al., 2018) investigated the effect of aquatic interventions on arterial stiffness in the subacute phase of stroke recovery. The high quality RCT randomized participants to receive aquatic treadmill training or on-land aerobic exercise. Arterial stiffness (paretic/non-paretic) was measured using an oscillometric method at post-treatment (4 weeks). No significant between-group difference was found.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that aquatic interventions are not more effective than on-land interventions for reducing arterial stiffness in the subacute phase of stroke recovery.

Balance
Conflicting
4

Two high quality RCTs (Tripp & Krakow, 2014; Lee et al., 2018) and one fair quality RCT (Chan et al., 2017) investigated the effect of aquatic interventions on balance in the subacute phase of stroke recovery.

The first high quality RCT (Tripp & Krakow, 2014) randomized participants to receive aquatic therapy using the Halliwick method, or time-matched conventional physiotherapy; both groups received additional physiotherapy. Balance was measured using the Berg Balance Scale (BBS) and the Functional Reach Test at post-treatment (2 weeks). A significant between-group difference was found on one measure (BBS), in favour of aquatic therapy vs. conventional physiotherapy.

The second high quality RCT (Lee et al., 2018) randomized participants to receive aquatic treadmill training or on-land aerobic exercise. Balance was measured using the BBS at post-treatment (4 weeks). No significant between-group difference was found.

The fair quality RCT (Chan et al., 2017) randomized participants to receive aquatic therapy or conventional rehabilitation; both groups received additional conventional rehabilitation. Balance was measured using the BBS at post-treatment (6 weeks). No significant between-group difference was found.

Conclusion: There is conflicting evidence (level 4) regarding the effectiveness of aquatic interventions on balance in the subacute phase of stroke recovery. While one high quality RCT found that a 2-week aquatic therapy program was more effective than physiotherapy, a second high quality RCT and a fair quality RCT found that aquatic interventions (4-week aquatic treadmill training, 6-week aquatic therapy) were not more effective than on-land rehabilitation programs.

Cardiorespiratory fitness parameters
Conflicting
4

Two high quality RCTs (Han & Im, 2018; Lee et al., 2018) investigated the effect of aquatic interventions on cardiorespiratory fitness parameters in the subacute phase of stroke recovery.

The first high quality RCT (Han & Im, 2018) randomized participants to receive aquatic treadmill training or land-based aerobic exercise. Cardiorespiratory fitness parameters (Peak oxygen uptake, Peak rate pressure product, Resting heart rate, Peak heart rate, Age-predicted maximum heart rate, Exercise tolerance test duration, Respiratory exchange ratio) were measured at post-treatment (6 weeks). Significant between-group differences were found on four of seven measures (Oxygen uptake, Peak heart rate, Age-predicted maximum heart rate, Exercise tolerance test duration), in favour of aquatic therapy vs. land-based aerobic exercise.

The second high quality RCT (Lee et al., 2018) randomized participants to receive aquatic treadmill training or on-land aerobic exercise. Cardiorespiratory fitness parameters (Resting heart rate, Resting systolic/diastolic blood pressure, Maximal heart rate, Maximal systolic/diastolic blood pressure, Maximal rate pressure product, Respiratory exchange ratio, Maximal oxygen consumption) were measured at post-treatment (4 weeks). No significant between-group differences were found.

Conclusion: There is conflicting evidence (level 4) regarding the effectiveness of aquatic interventions on cardiorespiratory fitness parameters in the subacute phase of stroke recovery. While both interventions compared aquatic treadmill training with on-land aerobic exercise, the 6-week program found between-group differences on some measures of cardiorespiratory fitness whereas the 4-week program found no differences between groups.

Gait
Effective
1a

Two high quality RCTs (Tripp & Krakow, 2014; Zhang et al., 2016) investigated the effect of aquatic interventions on gait in the subacute phase of stroke recovery.

The first high quality RCT (Tripp & Krakow, 2014) randomized participants to receive aquatic therapy using the Halliwick method, or time-matched conventional physiotherapy; both groups received additional physiotherapy. Gait was measured using the Functional Ambulation Categories at post-treatment (2 weeks). A significant between-group difference was found, in favour of aquatic therapy vs. conventional physiotherapy.

The second high quality RCT (Zhang et al., 2016) randomized participants to receive aquatic therapy or land-based physiotherapy. Gait was measured using the Functional Ambulation Categories at post-treatment (8 weeks). A significant between-group difference was found, in favour of aquatic therapy vs. land-based physiotherapy.

Conclusion: There is strong evidence (level 1a) from two high quality RCTs that aquatic interventions are more effective than on-land interventions for improving gait in the subacute phase of stroke recovery.

Mobility
Not effective
1b

One high quality RCT (Tripp & Krakow, 2014) and one fair quality RCT (Chan et al., 2017) investigated the effect of aquatic interventions on mobility in the subacute phase of stroke recovery.

The high quality RCT (Tripp & Krakow, 2014) randomized participants to receive aquatic therapy using the Halliwick method, or time-matched conventional physiotherapy; both groups received additional physiotherapy. Mobility was measured using the Rivermead Mobility Index at post-treatment (2 weeks). No significant between-group difference was found.

The fair quality RCT (Chan et al., 2017) randomized participants to receive aquatic therapy or conventional rehabilitation; both groups received additional conventional rehabilitation. Mobility was measured using the Timed Up and Go test and ambulatory skills were measured using the Community Balance and Mobility Test at post-treatment (6 weeks). No significant between-group differences were found.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT and one fair quality RCT that aquatic interventions are not more effective than on-land interventions for improving mobility in the subacute phase of stroke recovery.

Motor function - lower extremity
Not effective
1b

One high quality RCT (Lee et al., 2018) investigated the effect of aquatic interventions on motor function in the subacute phase of stroke recovery. The high quality RCT randomized participants to receive aquatic treadmill training or on-land aerobic exercise. Lower extremity motor function was measured using the Fugl-Meyer Assessment (FMA, FMA – Lower Limb score) at post-treatment (4 weeks). No significant between-group difference was found.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that aquatic interventions are not more effective than on-land interventions for improving lower extremity motor function in the subacute phase of stroke recovery.

Muscle strength - lower extremity
Effective
1a

Two high quality RCTs (Zhang et al., 2016; Lee et al., 2018) investigated the effect of aquatic interventions on lower extremity muscle strength in the subacute phase of stroke recovery.

The first high quality RCT (Zhang et al., 2016) randomized participants to receive aquatic therapy or land-based physiotherapy. Muscle activity (Knee extension/flexion torque/cocontraction ratio, Ankle dorsiflexion/plantarflextion torque/cocontraction ratio) was measured at post-treatment (8 weeks). Significant between-group differences were found in some measures (Knee extension torque, Knee extension cocontraction ratio, Ankle plantarflextion torque), in favour of aquatic therapy vs. land-based physiotherapy.

The second high quality RCT (Lee et al., 2018) randomized participants with subacute stroke to receive aquatic treadmill training or on-land aerobic exercise. Muscle strength was measured by dynamometer (isometric knee flexion/extension – paretic/non-paretic limb) at post-treatment (4 weeks). A significant between-group difference in muscle strength of the paretic limb (knee flexion, knee extension) was found, in favour of aquatic therapy vs. on-land aerobic exercise.

Conclusion: There is strong evidence (level 1a) from two high quality RCTs that aquatic interventions are more effective than on-land interventions for improving lower extremity muscle strength in the subacute phase of stroke recovery.

Quality of life
Not effective
1b

One high quality RCT (Lee et al., 2018) investigated the effect of aquatic interventions on quality of life in the subacute phase of stroke recovery. The high quality RCT randomized participants to receive aquatic treadmill training or on-land aerobic exercise. Health-related quality of life was measured using the EQ-5D-3L at post-treatment (4 weeks). No significant between-group difference was found.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that aquatic interventions are not more effective than on-land interventions for improving quality of life in the subacute phase of stroke recovery.

Spasticity
Not effective
1b

One high quality RCT (Zhang et al., 2016) investigated the effect of aquatic interventions on spasticity in the subacute phase of stroke recovery. The high quality RCT randomized participants to receive aquatic therapy or land-based physiotherapy. Spasticity was measured using the Modified Ashworth Scale (Knee flexion, Ankle dorsiflexion) at post-treatment (8 weeks). No significant between-group difference was found.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that aquatic interventions are not more effective than on-land interventions for reducing spasticity in the subacute phase of stroke recovery.

Walking endurance
Not effective
1b

One high quality RCT (Han & Im, 2018) and one fair quality RCT (Chan et al., 2017) investigated the effect of aquatic interventions on walking endurance in the subacute phase of stroke recovery.

The high quality RCT (Han & Im, 2018) randomized participants to receive aquatic treadmill training or land-based aerobic exercise. Walking endurance was measured using the Six Minute Walk Test at post-treatment (6 weeks). No significant between-group difference was found.

The fair quality RCT (Chan et al., 2017) randomized participants to receive aquatic therapy or conventional rehabilitation; both groups received additional conventional rehabilitation. Walking endurance was measured using the Two-Minute Walking Test at post-treatment (6 weeks). No significant between-group difference was found.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT and one fair quality RCT that aquatic interventions are not more effective than on-land interventions for improving walking endurance in the subacute phase of stroke recovery.

Chronic phase

Balance
Effective
1a

Six high quality RCTs (Chu et al., 2004; Zhu et al., 2016; Cha, Shin & Kim, 2017; Saleh, Rehab & Aly, 2019; Perez-de la Cruz, 2020; Perez-de la Cruz, 2021), ten fair quality RCTs (Noh et al., 2008; Lee, Ko & Cho, 2010; Park & Roh, 2011b; Park et al., 2014; Jung et al., 2014; Kim, Lee & Kim, 2015; Kim, Lee & Jung, 2015; Kim, Lee & Kim, 2016; Eyvaz, Dundar & Yesil, 2018; Aidar et al., 2018) and three quasi-experimental studies (Han, Kim & An, 2013; Montagna et al., 2014; Morer et al., 2020) investigated the effect of aquatic therapy interventions on balance in the chronic phase of stroke recovery.

The first high quality RCT (Chu et al., 2004) randomized participants to receive an aquatic lower extremity program or a land-based upper extremity training program. Balance was measured using the Berg Balance Scale (BBS) at post-treatment (8 weeks). No significant between-group difference was found.

The second high quality RCT (Zhu et al., 2016) randomized participants to receive hydrotherapy or land-based exercises. Balance was measured using the BBS and the Functional Reach Test (FRT) at post-treatment (4 weeks). A significant between-group difference was found on one measure (FRT), in favour of hydrotherapy vs. land-based exercise.

The third high quality RCT (Cha, Shin & Kim, 2017) randomized participants to receive aquatic therapy using the Bad Ragaz Ring method or time-matched conventional physical therapy; both groups received additional physical therapy. Balance was measured using the Biodex Balance Master (a balance measurement system) at post-treatment (6 weeks). A significant between-group difference was found, in favour of aquatic therapy vs. physical therapy.

The fourth high quality RCT (Saleh, Rehab & Aly, 2019) randomized participants to receive aquatic motor dual task training or land-based motor dual task training. Dynamic balance was measured using the Biodex Balance System (Overall Stability Index, Anteroposterior Stability Index, Mediolateral Stability Index) at post-treatment (6 weeks). A significant between-group difference was found on all measures of dynamic balance, in favour of aquatic therapy vs. land-based training.

The fifth high quality RCT (Perez-de la Cruz, 2020) randomized participants to receive Ai-Chi aquatic therapy, on-land exercises, or combined aquatic therapy + on-land exercises. Balance was measured using the Tinetti test (Total score), the 360 degree turn test, and single-leg stance balance tests (Right/Left leg) at post-treatment (12 weeks) and one-month follow-up. Significant between-group differences were found on two measures (Tinetti test, 360-degree turn test) at both timepoints, in favour of aquatic therapy vs. on-land therapy.
Note: Significant between-group differences were also found (Tinetti test, 360-degree turn test) at both timepoints, in favour of combined therapy vs. on-land exercises. There was a significant between-group difference (360 degree turn test) at both timepoints, in favour of combined therapy vs. aquatic therapy.

The sixth high quality RCT (Perez-de la Cruz, 2021) randomized participants to receive aquatic therapy, on-land exercises, or combined aquatic therapy + on-land exercises. Balance was measured using the BBS and tandem stance (eyes open) at post-treatment (12 weeks) and one-month follow-up. There were significant between-group differences on both measures at both timepoints, in favour of aquatic therapy vs. on-land exercises.
Note: There was a significant between-group difference in one measure (BBS) at both timepoints, in favour of aquatic therapy vs. combined therapy; in one measure (tandem stance) at follow-up only, in favour of combined therapy vs. aquatic therapy; and in one measure (tandem stance) at both timepoints, in favour of combined therapy vs. on-land exercises.

The first fair quality RCT (Noh et al., 2008) randomized participants to receive aquatic therapy or conventional rehabilitation. Balance was measured using the BBS at post-treatment (8 weeks). A significant difference in change scores from baseline to post-treatment was found, in favour of aquatic therapy vs. conventional rehabilitation.

The second fair quality RCT (Lee, Ko & Cho, 2010) randomized participants to receive aquatic task-oriented training or on-ground task-oriented training. Balance was measured using the Good Balance System to measure static balance (Anteroposterior/mediolateral sway velocity – eyes open, eyes closed) and dynamic balance (Time, Distance) at post-treatment (12 weeks). A significant between-group difference was found in dynamic balance only (time, distance), in favour of aquatic vs. on-ground task-oriented training.

The third fair quality RCT (Park & Roh, 2011b) randomized participants to receive aquatic exercises or land exercises. Static balance was measured using the Good Balance System (Mediolateral sway velocity – eyes open/eyes closed; Anteroposterior sway velocity – eyes open/eyes closed; Velocity movement – eyes open/eyes closed) at post-treatment (6 weeks). Significant between-group differences were seen in static balance measures (Mediolateral sway velocity – eyes open, Anteroposterior sway velocity – eyes open, Velocity movement – eyes open), in favour of aquatic exercises vs. land exercises.

The fourth fair quality RCT (Park et al., 2014) randomized participants to receive aquatic treadmill training or no additional training; both groups received conventional rehabilitation. Balance was measured using the Balance System SD (Static balance – anteroposterior sway, mediolateral sway, total; Dynamic balance) at post-treatment (4 weeks). No significant between-group differences were found.

The fifth fair quality RCT (Jung et al., 2014) randomized participants to receive aquatic obstacle training or land-based obstacle training. Static balance was measured using the Good Balance system (Mediolateral sway velocity – eyes closed, Anteroposterior sway velocity – eyes closed, Sway area) at post-treatment (12 weeks). Significant between-group differences were found on all measures, in favour of aquatic therapy vs. land-based obstacle training.

The sixth fair quality RCT (Kim, Lee & Kim, 2015) randomized participants to receive aquatic proprioceptive neuromuscular facilitation (PNF) lower extremity exercises or on-ground PNF lower extremity exercises. Balance was measured using the BBS, FRT and One Leg Stand Test at post-treatment (6 weeks). Significant between-group differences were found on all measures of balance, in favour of aquatic PNF exercises vs. on-ground PNF exercises.

The seventh fair quality RCT (Kim, Lee & Jung, 2015) randomized patients to receive aquatic coordination movement using Proprioceptive Neuromuscular Facilitation (PNF) and Neurodevelopmental Therapy (NDT) or NDT alone. Balance was measured using the BBS and the FRT at post-treatment (6 weeks). Significant between-group differences were found on both measures of balance, in favour of aquatic PNF vs. no aquatic therapy.

The eighth fair quality RCT (Kim, Lee & Kim, 2016) randomized participants to receive aquatic dual-task training or no aquatic therapy; both groups received neurodevelopmental therapy. Balance was measured using the BBS and FRT at post-treatment (6 weeks). A Significant between-group difference was found on both measures, in favour of aquatic therapy vs. no aquatic therapy.

The ninth fair quality RCT (Eyvaz, Dundar & Yesil, 2018) randomized participants to receive water-based exercises or land-based exercises; both groups received additional land-based exercises. Balance was measured using the BBS and the Sportak Balance Device (Static balance index, Dynamic balance index) at post-treatment (6 weeks). A significant difference was found on one measure (BBS), in favour of land-based exercise vs. water-based exercise.

The tenth fair quality RCT (Aidar et al., 2018) randomized participants to receive an aquatic exercise program or no treatment. Balance was measured using the BBS at post-treatment (12 weeks). A significant between-group difference was found, in favour of aquatic therapy vs. no treatment.

The first non-randomized study (Han, Kim & An, 2013) allocated participants to receive an aquatic proprioceptive exercise program or a land-based proprioceptive exercise program. Balance was measured using the BBS and sway area was measured using the Good Balance system (eyes open, eyes closed) at post-treatment (6 weeks). Significant between-group differences were found on all measures, in favour of aquatic therapy vs. land-based therapy.

The second non-randomized study (Montagna et al., 2014) assigned participants to receive aquatic physiotherapy using the Halliwick method. Balance was measured using the BBS at post-treatment (18 sessions). A significant improvement was found.

The third non-randomized study (Morer et al., 2020) assigned participants to receive aquatic therapy + thalassotherapy. Balance was measured using the BBS at post-treatment (2 weeks). A significant improvement was found.

Conclusion: There is strong evidence (level 1a) from five high quality RCTs, nine fair quality RCTs and one quasi-experimental study that aquatic therapy interventions are more effective than land-based interventions or no treatment for improving balance in the chronic phase of stroke recovery.
Note
: However, one high quality RCT found that an aquatic lower extremity intervention program was no more effective than a comparative land-based upper extremity program; one fair quality RCT found that aquatic treadmill training was no more effective than no treatment; and another fair quality RCT found that water-based exercises were less effective than land-based exercises for improving balance. In contrast, two other quasi-experimental studies noted a significant improvement in balance following aquatic interventions.

Cardiovascular fitness parameters
Effective
1b

One high quality RCT (Chu et al., 2004) investigated the effect of an aquatic intervention on cardiovascular fitness in the chronic phase of stroke recovery. The high quality RCT randomized participants to receive an aquatic lower extremity program or a land-based upper extremity training program. Maximal oxygen uptake (VO2max) and maximal workload (watts) were measured using a cycle ergometer test at post-treatment (8 weeks). A significant between-group difference was found on both measures, in favour of aquatic therapy lower extremity training vs. land-based upper extremity training.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that aquatic therapy intervention is more effective than a land-based intervention (upper extremity training) for improving cardiovascular fitness parameters in the chronic phase of stroke recovery.

Functional independence
Conflicting
4

Two fair quality RCTs (Kim, Lee & Kim, 2015; Eyvaz, Dundar & Yesil, 2018) have investigated the effect of aquatic therapy interventions on functional independence in the chronic phase of stroke recovery.

The first fair quality RCT (Kim, Lee & Kim, 2015) randomized participants to receive aquatic proprioceptive neuromuscular facilitation (PNF) lower extremity exercises or on-ground PNF lower extremity exercises. Functional independence was measured using the Functional Independence Measure (FIM) at post-treatment (6 weeks). A significant between-group difference was found, in favour of aquatic PNF exercises vs. on-ground PNF exercises.

The second fair quality RCT (Eyvaz, Dundar & Yesil, 2018) randomized participants to receive water-based exercises or land-based exercises; both groups received additional land-based exercises. Functional independence was measured using the FIM at post-treatment (6 weeks). No significant between-group difference was found.

Conclusion: There is conflicting evidence (level 4) regarding the effectiveness of aquatic therapy on functional independent in chronic phase of stroke recovery. While one fair quality RCT found that aquatic therapy was more effective than land-based exercises, another fair quality RCT found that it was not more effective.
Note:
Differences in outcomes may relate to the different forms of aquatic intervention and/or intervention intensity: aquatic exercises performed for 60 mins/session for 3 days/week were not more effective than land-based exercises; aquatic PNF performed for 30 mins/session, 5 sessions/week were more effective than land-based PNF exercises.

Gait ability
Not effective
2a

One fair quality RCT (Noh et al., 2008) has investigated the effect of aquatic interventions on gait ability in the chronic phase of stroke recovery. The fair quality RCT randomized participants to receive aquatic therapy or conventional rehabilitation. Gait ability was measured using the Modified Motor Assessment Scale at post-treatment (8 weeks). No significant difference was found.

Conclusion: There is limited evidence (level 2a) from one fair quality RCT that aquatic interventions are not more effective than land-based interventions for improving gait ability in the chronic phase of stroke recovery.

Gait parameters
Effective
1b

One high quality RCT (Saleh, Rehab & Aly, 2019) and three fair quality RCTs (Park et al., 2012; Furnari et al., 2014; Park et al., 2016) have investigated the effect of aquatic interventions on gait parameters in the chronic phase of stroke recovery.

The high quality RCT (Saleh, Rehab & Aly, 2019) randomized participants to receive aquatic motor dual task training or land-based motor dual task training. Gait parameters (Walking speed, Step length – paretic/non-paretic limb, Time of support on the paretic limb) were measured using the Biodex Gait Trainer at post-treatment (6 weeks). A significant between-group difference was found on all gait parameters, in favour of aquatic training vs. land-based training.

The first fair quality RCT (Park et al., 2012) randomized participants to receive aquatic treadmill training or land-based treadmill training. Gait parameters (Joint angles on heel contact and toe off the ground [hip flexion, knee extension, plantarflexion/dorsiflexion]) were measured at post-treatment (6 weeks). Significant between-group differences were found (hip flexion – heel contact, toe off; knee extension – heel contact, toe-off), in favour of aquatic treadmill training vs. land-based treadmill training.

The second fair quality RCT (Furnari et al., 2014) randomized participants to receive hydrokinesytherapy or conventional physical therapy; both groups received additional physical therapy. Gait parameters (gait speed, cadence, stance phase, swing phase, double support phase, semistep length) were measured using a Modular Clinical Electronic Baropodometer at post-treatment (8 weeks). Significant between-group differences were found on most measures (gait speed, cadence, stance phase, swing phase, double support phase), in favour of aquatic therapy vs. physical therapy.

The third fair quality RCT (Park et al., 2016) randomized participants to receive aquatic trunk exercises or land-based trunk exercises. Gait parameters were measured using the Gait trainer 2 analysis system (Walking speed, Walking cycle, Stance phase, Stride length, Symmetry index – stance phase/stride length) at post-treatment (4 weeks). Significant between-group differences were found on two gait parameters (Walking cycle, Stride length – paretic limb), in favour of land-based trunk exercises vs. aquatic trunk exercises.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT and two fair quality RCTs that aquatic interventions are more effective than land-based interventions for improving gait parameters in the chronic phase of stroke recovery.
Note
: However, one fair quality RCT found that land-based exercises were more effective than water-based trunk exercises for improving some gait parameters.

Mobility
Conflicting
4

Three high quality RCTs (Zhu et al., 2016; Cha, Shin & Kim, 2017; Perez-de la Cruz, 2021), six fair quality RCTs (Park et al., 2011a; Kim, Lee & Jung, 2015; Kim, Lee & Kim, 2015; Kim, Lee & Kim, 2016; Eyvaz, Dundar & Yesil, 2018; Aidar et al., 2018) and two quasi-experimental studies (Montagna et al., 2014; Morer et al., 2020) investigated the effect of aquatic interventions on mobility in the chronic phase of stroke recovery.

The first high quality RCT (Zhu et al., 2016) randomized participants to receive hydrotherapy or land-based exercises. Mobility was measured using the Timed Up and Go Test (TUG) at post-treatment (4 weeks). No significant between-group difference was found.

The second high quality RCT (Cha, Shin & Kim, 2017) randomized patients to receive aquatic therapy using the Bad Ragaz Ring method or time-matched conventional physical therapy; both groups received additional physical therapy. Mobility was measured using the TUG at post-treatment (6 weeks). No significant between-group difference was found.

The third high quality RCT (Perez-de la Cruz, 2021) randomized participants to receive aquatic therapy, on-land exercises, or combined aquatic therapy + on-land exercises. Mobility was measured using the TUG and Five Times Sit-to-Stand test (FTSTS) at post-treatment (12 weeks) and one-month follow-up. There were significant between-group differences in both measures at both timepoints, in favour of aquatic therapy vs. on-land exercises.
Note: There was a significant between-group difference in one measure (FTSTS) at both timepoints, in favour of combined therapy vs. aquatic therapy; there were significant between-group differences in both measures at both timepoints, in favour of combined therapy vs. on-land exercises.

The first fair quality RCT (Park et al., 2011a) randomized participants to receive aquatic exercises or land exercises. Mobility was measured using the Performance-Oriented Mobility Assessment at post-treatment (6 weeks). A significant between-group difference was found, in favour of aquatic exercise vs. land exercise.

The second fair quality RCT (Kim, Lee & Jung, 2015) randomized patients to receive aquatic coordination movement using Proprioceptive Neuromuscular Facilitation (PNF) and Neurodevelopmental Therapy (NDT) or NDT alone. Mobility was measured using the TUG at post-treatment (6 weeks). A significant between-group difference was found, in favour of aquatic PNF vs. no aquatic therapy.

The third fair quality RCT (Kim, Lee & Kim, 2015) randomized participants to receive aquatic proprioceptive neuromuscular facilitation (PNF) lower extremity exercises or on-ground PNF lower extremity exercises. Mobility was measured using the TUG at post-treatment (6 weeks). A significant between-group difference was found, in favour of aquatic PNF exercises vs. on-ground PNF exercises.

The fourth fair quality RCT (Kim, Lee & Kim, 2016) randomized participants to receive aquatic dual-task training or no aquatic therapy; both groups received neurodevelopmental therapy. Mobility was measured using the TUG and the Five Times Sit-to-Stand Test at post-treatment (6 weeks). A significant between-group difference was found on both measures, in favour of aquatic therapy vs. no therapy.

The fifth fair quality RCT (Eyvaz, Dundar & Yesil, 2018) randomized participants to receive water-based exercises or land-based exercises; both groups received additional land-based exercises. Mobility was measured using the TUG test at post-treatment (6 weeks). No significant between-group difference was found.

The sixth fair quality RCT (Aidar et al., 2018) randomized participants to receive an aquatic exercise program or no treatment. Mobility was measured using the TUG and a test of getting up from a sitting position at post-treatment (12 weeks). A significant between-group difference was found on both measures of mobility, in favour of aquatic therapy vs. no treatment.

The first non-randomized study (Montagna et al., 2014) assigned participants to receive aquatic physiotherapy using the Halliwick method. Mobility was measured using the TUG at post-treatment (18 sessions). A significant improvement was found.

The second non-randomized study (Morer et al., 2020) assigned participants to receive aquatic therapy + thalassotherapy. Mobility was measured using the TUG at post-treatment (2 weeks). A significant improvement was found.

Conclusion: There is conflicting evidence (level 4) regarding the effect of aquatic therapy on mobility in the chronic phase of stroke recovery. While one high quality RCT and five fair quality RCTs found that aquatic interventions were more effective than land-based interventions or no treatment, two high quality RCTs and one fair quality RCT found that aquatic interventions were not more effective than comparison interventions (land-based exercises, conventional physical therapy).

Mood
Effective
2a

Two fair quality RCTs (Aidar et al., 2013; Aidar et al., 2018) investigated the effect of aquatic interventions on mood in the chronic phase of stroke recovery.

The first fair quality RCT (Aidar et al., 2013) randomized participants to receive an aquatic exercise program or no treatment. Anxiety was measured using the State Trait Anxiety Inventory (IDATE – I Anxiety Trait; II Anxiety State) and depression was measured using the Beck Depression Inventory at post-treatment (12 weeks). A significant between-group difference was found on all measures of mood, in favour of aquatic therapy vs. no treatment.

The second fair quality RCT (Aidar et al., 2018) randomized participants to receive an aquatic exercise program or no treatment. Anxiety was measured using the State Trait Anxiety Inventory (IDATE – I Anxiety state, II Anxiety trait) and depression was measured using the BDI at post-treatment (12 weeks). A significant between-group difference was found on all measures of mood, in favour of aquatic therapy vs. no treatment.

Conclusion: There is limited evidence (level 2a) from two fair quality RCTs that aquatic interventions are more effective than no treatment in improving mood in the chronic phase of stroke recovery.

Muscle activity
Effective
1b

One high quality RCT (Cha, Shin & Kim, 2017) investigated the effect of aquatic interventions on muscle activity in the chronic phase of stroke recovery. The high quality RCT randomized patients to receive aquatic therapy using the Bad Ragaz Ring method or time-matched conventional physical therapy; both groups received additional physical therapy. Muscle activity was measured using electromyography (EMG – Tibialis anterior, Gastrocnemius) at post-treatment (6 weeks). A significant between-group difference was found in favour of aquatic therapy vs. physical therapy.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that aquatic interventions are more effective than land-based interventions for improving muscle activity in the chronic phase of stroke recovery.

Pain
Effective
1b

One high quality RCT (Perez-de la Cruz, 2020) and one quasi-experimental study (Morer et al., 2020) investigated the effect of aquatic interventions on pain in the chronic phase of stroke recovery.

The high quality RCT (Perez-de la Cruz, 2020) randomized participants to receive Ai-Chi aquatic therapy, on-land exercises, or combined aquatic therapy + on-land exercises. Pain was measured using a visual analogue scale at post-treatment (12 weeks) and one-month follow-up. A significant between-group difference was found at both timepoints, in favour of aquatic therapy vs. on-land exercises.
Note: A significant between-group difference in pain was found at both timepoints, in favour of combined therapy vs. on-land exercises.

The non-randomized study (Morer et al., 2020) assigned participants to receive aquatic therapy + thalassotherapy. Pain was measured using a visual analogue scale at post-treatment (2 weeks). A significant improvement was found.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that aquatic interventions are more effective than on-land interventions for reducing pain in the chronic phase of stroke recovery. A non-randomized study also reported reduced pain following aquatic intervention.

Postural stability - dynamic
Conflicting
4

Two fair quality RCTs (Kim, Lee & Kim, 2016; Kum & Shin, 2017) investigated the effect of aquatic interventions on dynamic postural stability in the chronic phase of stroke recovery.

The first fair quality RCT (Kim, Lee & Kim, 2016) randomized participants to receive aquatic dual-task training or no aquatic therapy; both groups received neurodevelopmental therapy. Postural stability when walking was measured using the Functional Gait Assessment at post-treatment (6 weeks). A significant between-group difference was found, in favour of aquatic therapy vs. no aquatic therapy.

The second fair quality RCT (Kum & Shin, 2017) randomized participants to receive underwater backward treadmill training or on-ground backward treadmill training. Postural stability when walking was measured using the Functional Gait Assessment at post-treatment (6 weeks). No significant between-group difference was found.

Conclusion: There is conflicting evidence (level 4) between two fair quality RCTs regarding the effectiveness of aquatic interventions on dynamic postural stability in the chronic phase of stroke recovery. One study found that aquatic dual-task training was more effective than no training whereas a second study found that underwater backward treadmill training was no more effective than on-ground training.

Postural stability - static
Not effective
2a

One fair quality RCT (Furnari et al., 2014) and one quasi-experimental study (Montagna et al., 2014) investigated the effect of aquatic interventions on static postural stability in the chronic phase of stroke recovery.

The fair quality RCT (Furnari et al., 2014) randomized participants to receive hydrokinesytherapy or conventional physical therapy; both groups received additional physical therapy. Static postural stability was measured using baropodometry (plantar surface, plantar load – paretic/non-paretic) and stabilometry (length of the ball – eyes open/closed) at post-treatment (8 weeks). A significant between-group difference was found on one measure (length of ball – eyes open/closed), in favour of aquatic therapy vs. conventional physical therapy.

The non-randomized study (Montagna et al., 2014) assigned participants to receive aquatic physiotherapy using the Halliwick method. Plantar pressure distribution was measured using baropodometry (anterioposterior/mediolateral – eyes open, eyes closed, sit-to-stand) at post-treatment (18 sessions). No significant improvement was found.

Conclusion: There is limited evidence (level 2a) from one fair quality RCT that aquatic interventions are not more effective than on-ground interventions for improving static postural stability in the chronic phase of stroke recovery. A quasi-experimental study also reported no significant improvement in static postural stability following aquatic physiotherapy.

Proprioception
Effective
2a

Two fair quality RCTs (Park et al., 2011a; Kum & Shin, 2017) and one quasi-experimental study (Han, Kim & An, 2013) investigated the effect of aquatic interventions on proprioception in the chronic phase of stroke recovery.

The first fair quality RCT (Park et al., 2011a) randomized participants to receive aquatic exercises or land exercises. Proprioception of knee movements was measured using the Biometrics Motion Analysis System at post-treatment (6 weeks). A significant between-group difference was found, in favour of aquatic exercise vs. land exercise.

The second fair quality RCT (Kum & Shin, 2017) randomized participants to receive underwater backward treadmill training or on-ground backward treadmill training. Proprioception was measured using the joint angle recurrence method by smartphone protractor application while the participant was in one-legged stance (paretic hip flexion/extension, paretic knee flexion/extension), at post-treatment (6 weeks). Significant between-group differences were found on all measures, in favour of underwater backward treadmill training vs. on-ground backward treadmill training.

The non-randomized study (Han, Kim & An, 2013) allocated participants to receive an aquatic proprioceptive exercise program or a land-based proprioceptive exercise program. Proprioception was measured using the Biometrics motion analysis system at post-treatment (6 weeks). A significant between-group difference was found, in favour of aquatic therapy vs. land-based therapy.

Conclusion: There is limited evidence (level 2a) from two fair quality RCTs and one quasi-experimental study that aquatic interventions are more effective than on-land interventions for improving proprioception in the chronic phase of stroke recovery.

Quality of life
Conflicting
4

One high quality RCT (Matsumoto et al., 2016), one fair quality RCT (Eyvaz, Dundar & Yesil, 2018) and two quasi-experimental studies (Montagna et al., 2014; Morer et al., 2020) investigated the effect of aquatic interventions on quality of life in the chronic phase of stroke recovery.

The high quality RCT (Matsumoto et al., 2016) randomized participants to receive aquatic therapy or no aquatic therapy; both groups received conventional rehabilitation. Quality of life was measured using the Short-Form 36 (SF-36: Physical functioning; Role physical; Bodily pain; General health; Vitality; Social functioning; Role-emotional; Mental health) at post-treatment (12 weeks). A significant between-group difference was found in change scores on all measures of quality of life in favour of aquatic therapy vs. no aquatic therapy.

The fair quality RCT (Eyvaz, Dundar & Yesil, 2018) randomized participants to receive water-based exercises or land-based exercises; both groups received additional land-based exercises. Quality of life was measured using the SF-36 (Vitality; Physical functioning; Role physical; Pain; General health; Social functioning; Role emotional; Mental health) at post-treatment (6 weeks). A significant between-group difference was found in one measure of quality of life (SF-36: Vitality), in favour of water-based exercises vs. land-based exercises. No other between-group differences were found.

The first non-randomized study (Montagna et al., 2014) assigned participants to receive aquatic physiotherapy using the Halliwick method. Quality of life was measured using the Stroke-Specific Quality of Life questionnaire (SS-QoL – Energy, Family roles, Language, Mobility, Mood, Personality, Self-care, Social roles, Thinking, Upper extremity function, Vision, Work/productivity, Total scores) at post-treatment (18 sessions). A significant improvement was found in one score only (Mobility).

The second non-randomized study (Morer et al., 2020) assigned participants to receive aquatic therapy + thalassotherapy. Health-related quality of life was measured using the EQ-5D (Mobility, Self-care, Usual activities, Pain/discomfort, Anxiety/depression scores) at post-treatment (2 weeks). A significant improvement was found on one measure (Mobility).

Conclusion: There is conflicting evidence (level 4) regarding the effectiveness of aquatic interventions on quality of life in the chronic phase of stroke recovery. One high quality RCT found that aquatic therapy was more effective than no aquatic therapy. One fair quality RCT found that water-based exercises were no more effective than comparable land-based exercises. Two quasi-experimental studies found an improvement in only one measure of quality of life (mobility) following aquatic interventions.

Self-perception of Health and Well-being
Effective
2b

One quasi-experimental study (Morer et al., 2020) investigated the effect of aquatic interventions on well-being in the chronic phase of stroke recovery. The non-randomized study assigned participants to receive aquatic therapy + thalassotherapy. Psychological well-being was measured using the WHO 5-item Well-Being Index and self-perception of health was measured using the EQ-VAS at post-treatment (2 weeks). A significant improvement was found on both measures.

Conclusion: There is limited evidence (level 2b) from one quasi-experimental study that aquatic interventions are effective for improving well-being following stroke.

Spasticity
Effective
1b

One high quality RCT (Matsumoto et al., 2016) investigated the effect of aquatic interventions on spasticity in the chronic phase of stroke recovery. This high quality RCT randomized participants to receive aquatic therapy or no aquatic therapy; both groups received conventional rehabilitation. Spasticity was measured using the Modified Ashworth Scale at post-treatment (12 weeks). A significant between-group difference was found, in favour of aquatic therapy vs. no aquatic therapy.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that aquatic therapy is more effective than no aquatic therapy for reducing spasticity in the chronic phase of stroke recovery.

Strength - lower extremity
Not effective
1a

One high quality RCT (Perez-de la Cruz, 2020) and four fair quality RCTs (Noh et al., 2008; Park et al., 2012; Kum & Shin, 2017; Eyvaz, Dundar & Yesil, 2018) investigated the effect of aquatic interventions on lower extremity strength in the chronic phase of stroke recovery.

The high quality RCT (Perez-de la Cruz, 2020) randomized participants to receive Ai-Chi aquatic therapy, on-land exercises, or combined aquatic therapy + on-land exercises. Lower extremity functional strength was measured using the 30-second chair stand test at post-treatment (12 weeks) and one-month follow-up. No significant between-group difference was found at either timepoint between aquatic therapy vs. on-land exercises.
Note: Significant between-group differences were found at both timepoints in favour of combined therapy vs. aquatic therapy, and in favour of combined therapy vs. on-land exercises.

The first fair quality RCT (Noh et al., 2008) randomized participants to receive aquatic therapy or conventional rehabilitation. Muscle strength was measured using an isokinetic device (knee flexors/extensors – paretic/nonparetic, lumbar flexors/extensors) at post-treatment (8 weeks). A significant difference was found on one measure of muscle strength (paretic knee flexor – change score from baseline to post-treatment), in favour of aquatic therapy vs. conventional rehabilitation.

The second fair quality RCT (Park et al., 2012) randomized participants to receive aquatic treadmill training or land-based treadmill training. Muscle strength was measured using the Short Physical Performance Battery at post-treatment. No significant between-group difference was found.

The third fair quality RCT (Kum & Shin, 2017) randomized participants to receive underwater backward treadmill training or on-ground backward treadmill training. Knee flexor and extensor Isokinetic strength (paretic, non-paretic) was measured by handheld dynamometer (maximal peak torque at 90-degrees/second, 120-degrees/second) at post-treatment (6 weeks). No significant between-group differences were found.

The fourth fair quality RCT (Eyvaz, Dundar & Yesil, 2018) randomized participants to receive water-based exercises or land-based exercises; both groups received additional land-based exercises. Lower extremity muscle strength (paretic, non-paretic sides) was measured at post-treatment (6 weeks). No significant between-group difference was found.

Conclusion: There is strong evidence (level 1a) from one high quality RCT and four fair quality RCTs that aquatic interventions are not more effective than on-land interventions for improving lower extremity strength in the chronic phase of stroke recovery.

Walking endurance
Effective
1b

One high quality RCT (Zhu et al., 2016) and one quasi-experimental study (Morer et al., 2020) investigated the effect of aquatic interventions on walking endurance in the chronic phase of stroke recovery.

The high quality RCT (Zhu et al., 2016) randomized participants to receive hydrotherapy or land-based exercises. Walking endurance was measured using the Two-Minute Walk Test at post-treatment (4 weeks). A significant between-group difference was found, in favour of hydrotherapy vs. land-based exercise.

The non-randomized study (Morer et al., 2020) assigned participants to receive aquatic therapy + thalassotherapy. Walking endurance was measured using the Six-Minute Walk Test at post-treatment (2 weeks). A significant improvement was found.

Conclusion: There is moderate evidence (level 1b) from one high quality RCT that aquatic interventions are more effective than on-land interventions for improving walking endurance in the chronic phase of stroke recovery. A quasi-experimental study also reported improved walking endurance following aquatic therapy.

Walking speed
Effective
1a

Two high quality RCTs (Chu et al., 2004; Matsumoto et al., 2016), three fair quality RCTs (Kim, Lee & Jung, 2015; Kim, Lee & Kim, 2016; Aidar et al., 2018) and one quasi-experimental study (Morer et al., 2020) investigated the effect of aquatic interventions on walking speed in the chronic phase of stroke recovery.

The first high quality RCT (Chu et al., 2004) randomized participants to receive an aquatic lower extremity program or a land-based upper extremity training program. Self-selected gait speed (m/sec) was measured over an 8-meter walking test at post-treatment (8 weeks). A significant between-group difference was found, in favour of aquatic therapy vs. upper extremity training.

The second high quality RCT (Matsumoto et al., 2016) randomized participants to receive aquatic therapy or no aquatic therapy; both groups received conventional rehabilitation. Walking speed was measured using the 10-Meter Walk Test (Speed, Cadence) at post-treatment (12 weeks). A significant between-group difference was found on both measures of walking speed, in favour of aquatic therapy vs. no aquatic therapy.

The first fair quality RCT (Kim, Lee & Jung, 2015) randomized patients to receive aquatic coordination movement using Proprioceptive Neuromuscular Facilitation (PNF) and Neurodevelopmental Therapy (NDT) or NDT alone. Walking speed was measured using the 10-Meter Walk Test at post-treatment (6 weeks). A significant between-group difference was found, in favour of aquatic PNF vs. no therapy.

The second fair quality RCT (Kim, Lee & Kim, 2016) randomized participants to receive aquatic dual-task training or no aquatic therapy; both groups received neurodevelopmental therapy. Walking speed was measured using the 10-Meter Walk Test at post-treatment (6 weeks). A significant between-group difference was found, in favour of aquatic therapy vs. no therapy.

The third fair quality RCT (Aidar et al., 2018) randomized participants to receive an aquatic exercise program or no treatment. Walking speed was measured using the Timed 7.62-Meter Walk test at post-treatment (12 weeks). A significant between-group difference was found, in favour of aquatic therapy vs. no treatment.

The non-randomized study (Morer et al., 2020) assigned participants with chronic stroke to receive aquatic therapy + thalassotherapy. Walking speed was measured using the 10-Meter Walk Test at post-treatment (2 weeks). No significant improvement was found.

Conclusion: There is strong evidence (level 1a) from two high quality RCTs and three fair quality RCTs that aquatic therapy is more effective than on-land interventions or no therapy for improving walking speed in the chronic phase of stroke recovery.

Walking skills
Not effective
2a

One fair quality RCT (Kum & Shin, 2017) investigated the effect of aquatic interventions on walking skills in the chronic phase of stroke recovery. The fair quality RCT randomized participants to receive underwater backward treadmill training or on-ground backward treadmill training. Walking skills were measured using the Figure-of-Eight Walk test at post-treatment (6 weeks). No significant between-group difference was found.

Conclusion: There is limited evidence (level 2a) from one fair quality RCT that aquatic interventions are not more effective than on-land interventions for improving walking skills in the chronic phase of stroke recovery.

Weight-bearing
Effective
2a

Two fair quality RCTs (Noh et al