Effects of traditional Chinese exercises on the rehabilitation of limb function among stroke patients: A systematic review and meta-analysis

Accepted Manuscript Effects of traditional Chinese exercises on the rehabilitation of limb function among stroke patients: A systematic review and meta-analysis Li Ge, Qing-Xiang Zheng, Yan-Tan Liao, Jing-Yu Tan, Qiu-Lin Xie, Mikael Rask PII:

S1744-3881(17)30287-6

DOI:

10.1016/j.ctcp.2017.08.005

Reference:

CTCP 757

To appear in:

Complementary Therapies in Clinical Practice

Received Date: 11 July 2017 Revised Date:

20 August 2017

Accepted Date: 22 August 2017

Please cite this article as: Ge L, Zheng Q-X, Liao Y-T, Tan J-Y, Xie Q-L, Rask M, Effects of traditional Chinese exercises on the rehabilitation of limb function among stroke patients: A systematic review and meta-analysis, Complementary Therapies in Clinical Practice (2017), doi: 10.1016/j.ctcp.2017.08.005. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Effects of traditional Chinese exercises on the rehabilitation of limb function among stroke patients: a systematic review and

RI PT

meta-analysis

Li Ge1,*, Qing-Xiang Zheng1, Yan-Tan Liao2，Jing-Yu Tan1, Qiu-Lin Xie1，Mikael Rask3

Nursing College, Fujian University of Traditional Chinese Medicine, Fuzhou, China

2

Rehabilitative Hospital Affiliated to Fujian University of Traditional Chinese

SC

1

Medicine, Fuzhou, China

Department of Health and Caring Sciences, Linnaeus University, S-35195 Växjö,

M AN U

3

Sweden *

Corresponding author. Nursing College, Fujian University of Traditional Chinese

Medicine, No.1 Qiu Yang Road, Shangjie, Minhou, Fuzhou, Fujian Province, China. Tel: +86 0591 22861365; Fax: +86 0591 22861029.

AC C

EP

TE D

E-mail address: [email protected] (Li Ge).

ACCEPTED MANUSCRIPT

Effects of traditional Chinese exercises on the rehabilitation of limb function among stroke patients: a systematic review and

RI PT

meta-analysis

Key words: Traditional Chinese exercises; Stroke; Limb function; Systematic review; Meta-analysis; Randomized controlled trial

SC

ABSTRACT

Objective: To systematically review literature about the rehabilitative effects of

M AN U

traditional Chinese exercises (TCEs) on limb function among patients with stroke. Methods: Systematic review and meta-analysis of randomized controlled trials (RCTs). Twelve electronic databases were searched from their inceptions to February 2017, including PudMed, The Cochrane Library, Web of Science, EMBase, Science Direct, PsycINFO, Cumulative Index to Nursing and Allied Health Literature, Allied

TE D

and Complementary Medicine, Chinese Scientific Journal Database, China National Knowledge Infrastructure, Chinese Biomedical Literature Database and WanFang Data. RCTs were located to examine the rehabilitative effects of TCEs on limb

EP

function among stroke patients. Two authors independently screened the literature, extracted data and assessed the risk bias of the included studies. Methodological

AC C

quality evaluation and meta-analysis of included studies was performed by using Cochrane Collaboration's tool (RevMan 5.3). Results: A total of 31 RCTs with 2349 participants were included. Results of meta-analysis showed that TCEs produced positive effects on limb motor function (random effects model, standardized mean difference [SMD] = 1.21, 95% confidence interval [CI] = 0.66 to 1.77, P＜0.01), balance function{Berg balance scale: (random effects model, SMD = 2.07, 95%CI = 1.52 to 2.62, P＜0.01), timed-up-and-go test: (fixed effects model, mean difference [MD] = -1.77, 95%CI = -2.87 to -0.67, P＜ 0.01)}, activities of daily living (ADL) ability {Barthel Index scale: (random effects 1

ACCEPTED MANUSCRIPT model, MD = 15.60, 95%CI = 7.57 to 23.63, P＜0.01), Modified Barthel Index scale: (random effects model, MD = 12.30, 95%CI = 7.48 to 17.12, P＜0.01)}, and neurological impairment (fixed effects model, MD = -2.57, 95%CI = -3.14 to -2.00, P ＜0.01). After subgroup analysis and sensitivity analysis, the positive effects did not

RI PT

be affected by different types of TCEs and different lengths of intervention time. However, TCEs were no benefit to physical function on Short Physical Performance Battery and 2-minute Step Test among stroke patients.

Conclusion: Current evidence showed that TCEs produced positive effects on limb

SC

motor function, balance function, ADL ability and neurological impairment among stroke patients. More large-scale, high-quality, multiple center RCTs are required to

M AN U

further verify above conclusions in the future.

1. Introduction

As a worldwide disease with high morbidity, high mortality and high disability rate, stroke is the second leading cause of death in the world [1], and is also a leading

TE D

cause of functional impairments and disabilities [2]. Approximately 80% of stroke survivors suffer different degrees of motor functional impairments [3]. Limb dysfunction is one of the most common functional impairments among stroke

EP

survivors, which changes survivors’ lifestyle and lowers the daily activities ability and quality of life as well. It not only brings stroke survivors personal suffering, but also

AC C

brings a heavy burden to family members and society [4]. It is thus important to effectively improve the limb dysfunction of stroke patients by technologies of treatment and care.

Traditional Chinese exercises (TCEs) are one type of low-level aerobic [5],

gentle, healthcare, and self-healing exercises originating from ancient China [6], which encompass a dual nature of physical exercise and medical therapy [7]. TCEs combine body posture adjustment, breathing exercise and mindfulness practice in order to activate muscles and tendons and thus dredge meridians and collaterals, promote blood circulation, and regulate internal organs [7-9]. The effects of a 2

ACCEPTED MANUSCRIPT coordinative development on physical and mental health as well as preventing and curing diseases are achieved. Due to the effects, the low cost, and being easy to learn without any equipment, TCEs are attracting an increasing interest in the rehabilitation field of chronic diseases.

RI PT

TCEs majorly include Qigong, Tai Chi, Baduanjin, Yijinjing, Wuqinxi, Liuzijue, and Daoyin. These types of TCEs are different and each of them has their own characteristics (Box 1). However, So far a few studies have indicated that TCEs not only improved cardiovascular outcomes [24], quality of life and depression for

SC

chronic diseases [25], but also reduced the risk of stroke and made the elderly relax their body as well [26, 27]. A recent meta-analysis showed that TCEs had beneficial

M AN U

effects on the balance ability of stroke patients in short term and might be positive on their gait [28]. However, to our knowledge, there is no consensus regarding the benefits of TCEs on the rehabilitation of limb function among stroke patients. Therefore, the purpose of this systematic review was to summarize the effects of TCEs on the rehabilitation of limb function among stroke patients in order to provide

TE D

a theoretical basis for future research and practice. Box 1 The characteristics and advantages of each traditional Chinese exercises Characteristics

Advantage

Qigong

Qigong is one kind of Kung Fu,

Qigong can regulate the circulation of Qi

which includes Baduanjin, Yijinjing, Wuqinxi and Liuzijue. It emphasizes

and blood, dredge meridians and collaterals [10], awaken and improve self-healing

the practice and regulation of Qi

ability, inspire the motive power of human

based on the adjustment of breathing,

life [11], and thus prevent and treat diseases.

AC C

EP

Name

mind and physical activities, and aims to help people keep healthy [10].

Tai Chi

Tai

Chi

is

a

form

of

Tai Chi not only can activate muscles and

mind-body-breath exercise or martial

tendons and regulate the circulation of Qi

art characterized by slow, rhythmic,

and blood, but also dredge meridians and

circular movements [12].During Tai

collaterals [12]. Tai Chi has been widely

Chi exercise, body’s center of gravity

used as the method of treatment and

is constantly changed between two

rehabilitation in many kinds of diseases,

legs and both knees are kept slightly

such as stroke, rheumatoid arthritis [12],

flexing.

heart failure and chronic pulmonary diseases [13]. 3

obstructive

ACCEPTED MANUSCRIPT Baduanjin is a set of Qigong

Baduanjin can regulate the circulation of Qi

jin

exercise,

eight

and blood and dredge meridians and

separate and smooth movements,

collaterals [14]. It can improve middle-aged

breathing exercise, mind regulation

and elderly people’s physical function,

and relaxation [14]. Each of the eight

cardiovascular function, joint flexibility,

movements can benefit particular

balance ability and sleep quality [14], and

organs or different parts of body [15].

reduce the risk factors of cerebrovascular

which

includes

diseases [15]. Yijinjing is a set of Qigong exercise,

Yijinjing

g

which includes 12 separate movements. It is a mind-body

function, mental health and movement disorders. It also can enhance the body's

exercise

breathing

organs function and immunity ability to

exercise, the adjustment of physical

make people strong, delay the aging process

postures and movements, and mind

[17], keep a balance of Yin and Yang in the

regulation [16].

body, and prevent diseases [16].

Wuqinx

Wuqinxi is a set of Qigong exercise,

Wuqinxi not only regulates the circulation of

i

which imitates five kinds of animals’ Qi and blood, but also dredges meridians

physiological

M AN U

movement,

includes

improve

SC

Yijinjin

that

can

RI PT

Baduan

including

tigers’ and collaterals. Meanwhile, it can enhance

movement, deer’s movement, bears’ muscle strength and improve cardiovascular movement, monkeys’ movement and birds’ movement [18].

outcomes [19].

Liuzijue is a set of Qigong exercise,

Liuzijue can regulate the activities of Qi and

e

which includes six separate, slow and

promote Qi to activate blood. Therefore, it

smooth movements. It is practiced by

has good regulative effects on respiratory

means

the

function, especially for chronic obstructive

pronunciation of six Chinese words ( "xu", "he", "hu", “si”, “chui”, “xi” )

pulmonary diseases [21]. Meanwhile, it can improve heart function and inspire the

and coordinated with six physical

motive power of human life [20].

TE D

Liuziju

reading

aloud

EP

of

postures and movements to keep fit [20].

Daoyin refers to extend body and

AC C

Daoyin

practice

tendons

bones

dredge meridians and collaterals, promote

accompanied by breathing exercise

blood circulation and regulate internal

under the guidance of consciousness.

organs, and thus prevent and treat diseases

It

[23].

includes

breathing

and

Daoyin can activate muscles and tendons,

posture

adjustments,

exercise,

mind-body

relaxation, mind concentration and use, and rhythmic movements and self-massage [22].

2. Methods 2.1. Search strategy 4

ACCEPTED MANUSCRIPT Relevant studies were searched for final included studies from two sources: electronic databases and reference lists. The following 12 electronic databases were searched from every electronic database inceptions to February 2017: PudMed, The Cochrane Library, Web of Science, EMBase, Science Direct, PsycINFO, Cumulative

RI PT

Index to Nursing and Allied Health Literature (CINAHL), Allied and Complementary Medicine (AMED), Chinese Scientific Journal Database (VIP), China National Knowledge Infrastructure (CNKI), Chinese Biomedical Literature Database (CBM), and WanFang Data. The Mesh terms and free words were combined to construct the

SC

search strategies, and the retrieval way was adjusted according to the different features of database. In order to search for all relevant studies, we referred to the

M AN U

reference lists of all retrieved articles to expand the search scope. Articles published were restricted to Chinese and English. The Mesh terms and keywords were used for this study included: “Qigong”, “Tai Chi”, “Baduanjin”, “Yijinjing”, “Wuqinxi”, “Liuzijue”, “Daoyin”, “Stroke”, “Apoplexy”, “cerebral infarction” , “cerebral

TE D

hemorrhage”, and“Cerebral vascular accident”, etc.

2.2. Inclusion and exclusion criteria

The following inclusion criteria of all studies included in the present review were:

EP

1) randomized controlled trials (RCTs); 2) all the participants were confirmed as stroke by the standards of stroke diagnosis, brain computed tomography (CT) or

AC C

magnetic resonance imaging (MRI) examination; 3) The forms of TCEs included Qigong, Tai Chi, Baduanjin, Yijinjing, Wuqinxi, Liuzijue, Daoyin; Participants in the intervention group accepted one or more forms of regular TCEs either combined with or without other treatments, and participants in the control group received routine rehabilitation therapy or exercise, in which including routine rehabilitation exercise, acupuncture, phone calls, routine health guidance, routine therapy and so on; 4) primary outcome measures included Fugl-Meyer assessment (FMA), Berg balance scale (BBS), Fug1-Meyer balance (FM-B) assessment, Timed-up-and-go test, Short Physical Performance Battery (SPPB) test, 2-minute step test; secondary outcome 5

ACCEPTED MANUSCRIPT measures included activities of daily living (ADL) assessment and NIH stroke scale (NIHSS). The studies were excluded if they were repeated publication or the original data could not be extracted from the literature. The participants would be excluded when they were with cognitive impairment, heart failure, renal failure, severe stroke,

2.3. Selection of studies and data extraction

RI PT

and exercise contraindication.

Two authors independently screened literature, extracted data and cross-checked.

SC

In case of disagreement, the two authors discussed together or discussed with another author. The inclusion criteria were strictly followed to screen titles and abstracts of

M AN U

retrieved studies and excluded the irrelevant studies. Potential full-text studies were reviewed to determine if eligible. Data extraction’s information mainly included: 1) the basic information of the literature; 2) the characteristics of participants; 3) the interventions; 4) frequency and duration of interventions; 5) outcome indicators.

TE D

2.4. Quality assessment

The methodological quality and the bias risk of each included RCT were evaluated by using Cochrane Collaboration's tool, i.e. Review Manager 5.3 (RevMan

EP

5.3), according to the Cochrane Handbook [29]. The evaluation included: 1) random sequence generation (checking for potential selection bias); 2) allocation concealment

AC C

(checking for potential selection bias); 3) blinding of participants and personnel (checking for potential performance bias); 4) blinding of outcome assessment (checking for potential detection bias); 5) incomplete outcome data (checking for potential attrition bias); 6) selective outcome reporting (checking for reporting bias); 7) other bias and size of study. The risk of bias were evaluated by two authors independently and cross-checked. Discrepancies were resolved through discussion in the author group.

2.5. Data synthesis 6

ACCEPTED MANUSCRIPT All data were analyzed by RevMan 5.3 and input by cross-checking method to ensure accuracy. The mean difference (MD) was used when the same instruments were conducted to measure the trials’ outcomes; otherwise, the standardized mean difference (SMD) was applied. All statistical analysis was calculated by 95%

RI PT

confidence intervals (95%CIs). Statistical heterogeneity between trials was analyzed by Chi-squared test and I2 test. If the statistical heterogeneity of Chi-squared test P > 0.01 and I2 < 50%, we considered the studies as homogeneity, and then the fixed effect model was used for meta-analysis; otherwise, when statistical heterogeneity of

SC

Chi-squared test P < 0.01 and I2 > 50%, considering significant heterogeneity, we conducted subgroup analysis (according to the type of experimental interventions,

M AN U

frequency and duration of interventions, etc.) or sensitivity analysis to investigate the potential sources of clinical heterogeneity and the reliability of the results by converting the random-effects model into the fixed-effects model. The funnel plot analyses would be performed to determine potential publication bias if every comparison included more than 10 studies. The inspection level of meta-analysis is α

3. Results

TE D

= 0.05, and the value of P < 0.05 indicates statistical significance [29].

EP

A total of 1209 articles were identified from the database searches, and 403 articles of which were excluded after duplicate check. After browsing each title and

AC C

abstract of the remaining 806 articles, 740 articles were excluded because the participants, interventions, and types of studies did not meet the inclusion criteria. Among the remained 66 articles, 35 full-text articles were excluded because they were not RCTs (n=9), they were review studies (n=5) or clinical case reports (n=1), they were repeated publication (n=10), the outcome measures did not meet the inclusion criteria (n=10). Finally, 31 RCTs articles were obtained in the study including 4 articles [12,30-32] in English and 27 articles in Chinese [33-59]. The flow chart for study selection is shown in Figure 1.

7

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Fig. 1 Flow diagram of study selection.

3.1. Study characteristics In total, 2349 participants were included in the present study. The average number of participants in each study was 75.8, ranging from 16 to 158. Trials originated from South Korea [12], the United States [30,31], and China [32-59]. Four types of interventions were used: Tai Chi was used in 20 studies [12,30-32,36-51]; 8

ACCEPTED MANUSCRIPT Daoyin was used in three studies [33-35]; Yijinjing was used in two studies [52,53]; and Baduanjin was used in six studies [54-59]. The intervention duration ranged from two to 52 weeks, of which 12 weeks were accounted for the majority. The intervention frequencies were in great difference, of which 5-7 times per week were

RI PT

accounted for the majority. The types of limb function included: 12 studies for the limb motor function [33-35,37,38,48,49,51-54,56], 21 studies for balance function [12,31,37-48,50-52,54-57],

11

studies

for

ADL

ability

[33,34,36-38,40,42,43,49,51,59], two studies [30,31] for physical function and two

SC

studies [36,58] for neurological impairment. Characteristics of the analyzed trials are

M AN U

shown in Table 1.

3.2. Methodological quality and risk of bias for the included studies Methodological quality and risk of bias for the included studies were shown in Table 2. Randomization sequence generation methods were reported in 14 studies [30-32,34,37-41,44,49,54,55,59], and allocation concealment methods were reported

TE D

in five studies [31,32,38,40,41]. Only one study [38] reported a reasonable dropout rate and 11 studies [30-32,36-38,42,48,53,57,58] used intention-to-treat analysis. Ten studies reported blinding of outcome assessment [31,38,40,45,46,48,49,52,57,58].

EP

The quality of all included studies was not optimum. Firstly, the sample size of 22 studies (70.96%) was less than 50 subjects. Secondly, the “random sequence

AC C

generation”, “allocation concealment”, and “size of study”were considered to contribute to high risk of bias of the included studies. However, all of the included studies were judged as “low risk of bias” in terms of blinding method because it was very difficult about exercise therapy to take blinding of participants and personnel and the outcomes were unlikely influenced by blinding. 3.3. Effects of TCEs on the rehabilitation of limb function In the present study, the types of limb function were included limb motor function, balance function, ADL ability, physical function, and neurological impairment. Balance function, ADL ability and physical function were measured using various 9

ACCEPTED MANUSCRIPT scales in the included studies. The results were reported as change scores from baseline to final values. The intervention duration were departed into three phases: short-term intervention (less than 8 weeks), mid-term intervention (8-16 weeks) and long-term intervention (more than 16 weeks). A summary of the meta-analyses is

AC C

EP

TE D

M AN U

SC

RI PT

shown in Table 3.

10

ACCEPTED MANUSCRIPT

Intervention （Intervention E/Control C） ）

Patients Characteristic, Sample Size（ （Intervention E/Control C） ） E（n=11）Mean age(SD)：53.45 ± 11.54 y；

E: Tai Chi;

C（n=11）Mean age(SD)：55.18 ± 10.20 y；

C: General physical therapy

Taylor-Piliae et al

E (n=53)

E: Tai Chi;

2014[30]

C (n=48) Mean age(SD)：68.2 ± 10.3 y；

C: weekly phone calls

Taylor-Piliae et al

E (n=13)

E: Tai Chi;

2012[31]

C（n=12）Mean age(SD)：64.5 ± 10.9y；

Au-Yeung

et

al

2009[32]

weeks.

step test

60-minute class, 3 times per week, for 12

SPPB;2-minute

C: Usual care + weekly phone calls

weeks.

step test

E（n=59）Mean age(SD)：61.7 ± 10.5 y；

E: Tai Chi;

Each week, 1 hour of group practice was

TUGT

C（n=55）Mean age(SD)：65.9 ± 10.7 y；

C: General physical therapy

Mean age(SD)：72.8 ± 10.1y；

supplemented by 3 hours of self-practice, for 12 weeks.

E: Daoyin + Rehabilitation therapy

20 min each time, 2 times per day, 6 times per

2013[33]

C（n=20）Mean age(SD)：58.60±12.45 y；

C: Rehabilitation therapy

week, for 2 weeks.

Jiao JK 2011[34]

E（n=45）Mean age(SD)：51.18±7.60 y；

E: Daoyin

30~45 min/session, for 6 weeks.

FMA-U; MBI

C（n=45）Mean age(SD)：50.69±6.99 y；

C: Bobath

40 min each time, 2 times per day, for 2 weeks.

FMA-U

NIHSS; BI

F

et

al

SPPB;2-minute

E（n=20）Mean age(SD)：62.45±13.04 y；

Wang

et

60-minute class, 3 times per week, for 12

Mean age(SD)：71.5 ± 10.3 y；

al

2014[35]

E（n=60）Mean age(SD)：55.2±2.5 y； C（n=60）Mean age(SD)：52.2±2.9 y；

TE D

DC

Outcomes TUGT

EP

Zheng

Duration & Frequency of Trials Period

60 min/session, 2 times per week, for 6 weeks.

M AN U

Kim et al 2015[12]

SC

Article,Year

RI PT

Table 1 Characteristics of the analyzed trials.

E: Daoyin + Constraint-induced movement therapy +

FMA; MBI

Routine rehabilitation therapy

AC C

C: Routine rehabilitation therapy

Zheng WC et al

E（n=51）Mean age(SD)：59±13 y；

E: Tai Chi;

30 min each time, 2 times per day, for 50 ~ 52

2015[36]

C（n=55）Mean age(SD)：60±12 y；

C: Routine rehabilitation exercise

weeks.

Zhou L 2015[37]

E（n=21）Mean age(SD)：55.33±9.47 y；

E: Tai Chi + Routine rehabilitation exercise

60 min per day, 5 times per week, for 4 weeks.

C（n=19）Mean age(SD)：54.74±7.02 y；

C: Routine rehabilitation exercise

11

FMA;FMA-U; FMA-L; BBS; BI

ACCEPTED MANUSCRIPT

E（n=34）Mean age(SD)：62.6±5.7 y；

E: Tai Chi gait intensive + Acupuncture + Routine

20~30-minute class, 2 times per day, 10 times

FMA-L;BBS;

C（n=34）Mean age(SD)：63.3±6.0 y；

rehabilitation exercise

per week, for 6 weeks.

MBI

C: et

al

E（n=14）Mean age(SD)：60.71±7.32 y；

E: Tai Chi

2016[39]

C（n=16）Mean age(SD)：58.56±8.52 y；

C: Routine balance rehabilitation exercise

Xie F 2008[40]

E（n=24）Mean age：35-66 y；

E: Tai Chi

C（n=24）Mean age：8 岁-62 y；

C: Routine balance rehabilitation exercise

Huang YD 2016[41]

E（n=8）Mean age(SD)：65.00±6.16 y；

E: Tai Chi

C（n=8）Mean age(SD)：63.63±7.37 y；

C: Routine exercise at home + phone calls or home follow-up

E（n=25）Mean age(SD)：60.48±8.29 y；

E: Tai Chi

C（n=25）Mean age(SD)：60.92±10.07 y；

C: Routine rehabilitation therapy

E（n=50）Mean age(SD)：54.3±13.8 y

E: Tai Chi

2013[43]

C（n=50）Mean age(SD)：55.2±14.6 y；

C: Routine rehabilitation exercise

Li HY 2012[44]

E（n=20）Mean age(SD)：57.8±7.35 y；

E: Tai Chi + ProKin biofeedback balance training

C（n=20）Mean age(SD)：58.7±6.45 y；

system + Routine rehabilitation therapy

Yang

ZB

et

al

TE D

Wang XY 2016[42]

SC

XB

M AN U

Wang

Acupuncture + Routine rehabilitation

RI PT

Zhou ZG 2013[38]

60 min per day, 5 times per week, for 12 weeks.

BBS

Week 1-2：10 min each time, 3 times per day;

BBS; MBI

Week 3-10: 30 min each time, 3-4 times per day. 60 min per day, 2 times per week, for 24 weeks.

BBS; TUGT

30 min per day, 5 times per week, for 8 weeks.

BBS;MBI; TUGT

45 min per day, 6 times per week, for 4 weeks.

BBS; BI

40 min per day, for 6 weeks.

BBS; FM-B

2 times per week, for 12 weeks.

BBS

15 min per day, 6 times per week, for 8 weeks.

BBS

BBS

QA

et

al

2010[45] Fu

C（n=34）Mean age(SD)：50.4±8.8 y；

2016[46]

C（n=30）Mean age(SD)：60.3±8.4 y；

C: Routine rehabilitation exercise

Liu TJ et al 2009[47]

E（n=24）Mean age(SD)：52.13±14.13 y；

E: Tai Chi + Routine rehabilitation exercise

30 minutes per day, 7 times per week, for 12

C（n=24）Mean age(SD)：53.51±12.63 y；

C: Routine rehabilitation exercise

weeks.

E（n=28）Mean age(SD)：51.43±15.63 y；

E: Tai Chi + Routine rehabilitation therapy

40-minutes class, 5 times per week, for 8

et

al

C: Routine rehabilitation therapy E: Tai Chi + Routine rehabilitation exercise

HX

et

E: Tai Chi

E（n=30）Mean age(SD)：59.7±7.6 y；

Yang

CX

E（n=34）Mean age(SD)：52.5±10.8 y；

AC C

Zhou

EP

C: Routine rehabilitation therapy

al

12

FMA-U; FMA-L;

ACCEPTED MANUSCRIPT

C（n=21）Mean age(SD)：54.85±11.85 y；

C: Routine rehabilitation therapy

weeks.

BBS

Li X et al 2011[49]

E（n=35）Mean age(SD)：56±5.58 y；

E: Tai Chi + Routine rehabilitation therapy

30~35 min per day, 5 times per week, for 6

FMA-U; MBI

C（n=32）Mean age(SD)：54±6.23 y；

C: Routine rehabilitation therapy

E（n=40）Mean age(SD)：60.14±10.25 y；

E: Tai Chi + Routine rehabilitation therapy

C（n=40）Mean age(SD)：48.23±12.32 y；

C: Routine rehabilitation therapy

E（n=76）Mean age(SD)：65.3 士 12.2 y；

E: Sitting Tai Chi

2012[51]

C（n=82）Mean age(SD)：63.5 士 11.3 y；

C: Rehabilitation exercise

Jia WZ 2008 [52]

E（n=18）Mean age(SD)：47.70±15.5 y；

E: Yijinjing + Routine rehabilitation exercise

C（n=16）Mean age(SD)：51.5±16.37 y；

C: Routine rehabilitation exercise

E（n=55）Mean age：40 - 80 y；

E: Yijinjing + Acupuncture + Routine rehabilitation

C（n=56）Mean age: 40 - 80 y；

exercise

et

al

2014[50] Gao

Zhang

JB

GD

et

et

al

al

2016[53]

20 min each time, 2 times per day, 7 times per

BBS

week, for 12 weeks.

SC

XD

weeks.

M AN U

Xu

RI PT

2016[48]

60 min each time, 2 times per day, for 4 weeks.

FMA;BBS;BI

30 min per day, 5 times per week, for 12 weeks.

FMA-U; FMA-L; BBS; FM-B

30 min per day, 5 times per week, for 12 weeks.

FMA

FMA-L; BBS

C: Routine rehabilitation exercise al

2016[54]

Bai

YJ

et

al

2011[55] Zhang

B

et

al

2015[56] Zhang

E（n=31）Mean age(SD)：55.07±4.81 y；

E: Baduanjin + Balance training

20 min each time, 2 times per day, 5 days as

C（n=31）Mean age(SD)：56.72±3.57 y；

C: Balance training

one therapeutic course and 1 day for rest,

E（n=30）Mean age(SD)：7±4.5 y；

E: Baduanjin + Balancing function training

C（n=30）Mean age(SD)：51.3±7.5 y；

C: Balancing function training

E（n=20）Mean age(SD)：49.36±10.32 y；

E: Baduanjin + Routine rehabilitation exercise

C（n=20）Mean age(SD)：50.49±14.16 y； M

et

al

E（n=115）:Mean age:33-82 y；

2013[57]

C（n=106）:Mean age:33-82 y；

Guo J et al 2013[58]

E（n=115）:Mean age:33-82 y； C（n=106）:Mean age:33-82 y；

Cai W et al 2011[59]

TE D

et

totally 8 therapeutic courses (6 weeks). 20 min each time, 2 times per day, for 6 weeks.

BBS

20~40 min each time, 7 times per week, for 10

FMA-U; FMA-L;

C: Routine rehabilitation exercise

weeks.

BBS; FM-B

E: Baduanjin +Routine rehabilitation therapy

20 min each time, 2 times per day, for 6 weeks.

BBS

20 min each time, 2 times per day, for 6 weeks.

NIHSS

BI

EP

Y

AC C

Zhang

C: Routine rehabilitation therapy E: Baduanjin + Routine rehabilitation therapy C : Routine rehabilitation therapy

E（n=30）Mean age(SD)：60.27±10.48 y；

E: Sitting Baduanjin + Routine health guidance

30 min per day, 4~5 times per week, for 12

C（n=30）Mean age(SD)：61.27±7.42 y；

C: Routine health guidance

weeks.

13

ACCEPTED MANUSCRIPT

RI PT

Abbreviations SPPB: Short Physical Performance Battery; TUGT: The Timed-up-and-go Test; FMA: Fugl-Meyer Assessment; FMA-U: Fugl-Meyer Assessment of Upper Limb; FMA-L: Fugl-Meyer Assessment of Lower Limb; MBI: Modified Barthel Index; BI: Barthel Index; NIHSS: NIH Stroke Scale; BBS: Berg Balance Scale;

AC C

EP

TE D

M AN U

SC

FM-B: Fugl-Meyer Balance Scale ;

14

ACCEPTED MANUSCRIPT Table 2 Risk of bias assessment of the included trials. Trials

Random

Allocation

Blinding of

Blinding of

Incomplete

Selective

sequence

concealment

Participants

outcome

outcome

outcome

and

assessment

data

reporting

generation

Other bias

Size of study

Unclear Risk

Unclear Risk

Low Risk

Unclear Risk

Unclear Risk

Low Risk

Unclear Risk

High Risk

Taylor-Piliae et al 2014[30]

Low Risk

Unclear Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Taylor-Piliae et al 2012[31]

Low Risk

Low Risk

Low Risk

Low Risk

Low Risk

Low Risk

Low Risk

High Risk

Au-Yeung et al 2009 [32]

Low Risk

Low Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Low Risk

High Risk

Zheng DC et al 2013 [33]

Unclear Risk

Unclear Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Jiao JK 2011 [34]

Low Risk

Unclear Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Wang F et al 2014 [35]

Unclear Risk

Unclear Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Zheng WC et al 2015 [36]

Unclear Risk

Unclear Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Zhou L 2015 [37]

Low Risk

Unclear Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Low Risk

High Risk

Zhou ZG 2013 [38]

Low Risk

Low Risk

Low Risk

Low Risk

Low Risk

Low Risk

Low Risk

High Risk

Wang XB et al 2016 [39]

High Risk

High Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Xie F 2008 [40]

Low Risk

Unclear Risk

Low Risk

Low Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Huang YD 2016 [41]

Low Risk

Low Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Low Risk

High Risk

Wang XY 2016 [42]

Low Risk

Low Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Yang ZB et al 2013 [43]

Unclear Risk

Unclear Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Li HY 2012 [44]

Low Risk

Unclear Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Zhou QA et al 2010 [45]

Unclear Risk

Unclear Risk

Low Risk

Low Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Fu CX et al 2016 [46]

High Risk

Liu TJ et al 2009 [47]

High Risk

Yang HX et al 2016 [48]

High Risk

Li X et al 2011 [49]

Low Risk

Xu XD et al 2014 [50]

Unclear Risk

Gao JB et al 2012 [51]

High Risk

Jia WZ 2008 [52] Zhang GD et al 2016 [53]

M AN U

SC

RI PT

Kim et al 2015 [12]

TE D

personnel

Low Risk

Low Risk

Low Risk

Low Risk

Unclear Risk

High Risk

High Risk

Low Risk

High Risk

Low Risk

Low Risk

Unclear Risk

High Risk

High Risk

Low Risk

Low Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Unclear Risk

Low Risk

Low Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Unclear Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Unclear Risk

High Risk

High Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Unclear Risk

Unclear Risk

Low Risk

Low Risk

Low Risk

Low Risk

Unclear Risk

High Risk

High Risk

High Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Unclear Risk

High Risk

AC C

EP

High Risk

Zhang Y et al 2016 [54]

Low Risk

Unclear Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Bai YJ et al 2011 [55]

Low Risk

Unclear Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Zhang B et al 2015 [56]

Unclear Risk

Unclear Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Unclear Risk

High Risk

Zhang M et al 2013 [57]

High Risk

High Risk

Low Risk

Low Risk

Low Risk

Unclear Risk

Unclear Risk

Unclear Risk

Guo J et al 2013 [58]

High Risk

High Risk

Low Risk

Low Risk

Low Risk

Unclear Risk

Unclear Risk

Unclear Risk

Cai W et al 2011 [59]

Low Risk

Unclear Risk

Low Risk

Unclear Risk

Low Risk

Unclear Risk

Unclear Risk

High Risk

15

ACCEPTED MANUSCRIPT Table 3 Summary of meta-analysis.

Study outcomes

No. Of

No.of

Statistical

trails

participants

method

Effect

Test for overall

Heterogeneit

effect

y

estimate Z

(I2)

P

Traditional Chinese Exercise for The Limb Motor Function with Fugl-Meyer Assessment Scores Fugl-Meyer Assessment Scores

12

1002

Std.Mean Difference

1.21 [0.66 to 1.77]

＜0.0001

RI PT

(IV, Random ,95%CI)

4.29

＞50%

Traditional Chinese Exercise for The Limb Motor Function with Different Part of Fugl-Meyer Assessment Scores Fugl-Meyer Assessment of Limb

4

349

Mean Difference

16.57 [9.87 to 23.26]

4.85

＜0.00001

＞50%

7.02 [3.45 to 10.60]

3.85

0.0001

＞50%

2.80

0.005

＞50%

1.19 [0.52 to 1.86]

3.47

0.0005

＞50%

0.80[0.22 to 1.38]

2.82

0.006

＞50%

1.17 [0.35 to 1.99]

2.80

0.005

＞50%

1.69 [0.36 to 3.02]

2.49

0.01

＞50%

1.48[0.93 to 2.04]

5.23

＜0.00001

＞50%

0.77 [0.40 to 1.14]

4.06

＜0.0001

＞50%

2.07[1.52 to 2.62]

7.36

＜0.00001

＞50%

0.83[-0.10 to 1.77]

1.74

0.08

＞50%

-1.77 [-2.87 to -0.67]

3.16

0.002

48%

2.29[1.61 to 2.98]

6.56

＜0.00001

＞50%

1.92[1.03 to 2.81]

4.25

＜0.0001

＞50%

NR

NR

NR

(IV, Random ,95%CI) Fugl-Meyer Assessment of Upper

7

440

(IV, Random ,95%CI)

Fugl-Meyer Assessment of Lower

6

293

Limb

Mean Difference (IV, Random ,95%CI)

SC

Limb

Mean Difference

2.36 [0.71 to 4.02]

Tai Chi

8

511

M AN U

Different Types of Traditional Chinese Exercise for The Limb Motor Function with Fugl-Meyer Assessment Scores Std.Mean Difference

(IV, Random ,95%CI)

Baduanjin

3

191

Std.Mean Difference

(IV, Random ,95%CI)

Yijinjing

3

179

Std.Mean Difference

(IV, Random ,95%CI)

Daoyin

3

250

Std.Mean Difference

TE D

(IV, Random ,95%CI)

Different Length of Intervention for The Limb Motor Function with Fugl-Meyer Assessment Scores Short-Term Intervention

10

(less than 8 weeks)

( 8-16 weeks)

Std.Mean Difference

(IV, Random ,95%CI)

7

447

EP

Mid-Term Intervention

635

Std.Mean Difference

(IV, Random ,95%CI)

Traditional Chinese Exercise for Different Types of Balance Scale 19

AC C

Berg Balance Scale

Fugl-Meyer Balance Scale

Timed-up-and-go Test

3

4

1272

Std.Mean Difference (IV, Random ,95%CI)

114

Mean Difference (IV, Random ,95%CI)

202

Mean Difference (IV, Fixed ,95%CI)

Different Types of Traditional Chinese Exercise for Balance with Berg Balance Scale Scores Tai Chi

14

855

Std.Mean Difference (IV, Random ,95%CI)

Baduanjin

4

383

Std.Mean Difference (IV, Random ,95%CI)

Yijinjing

1

34

NR

16

NA

ACCEPTED MANUSCRIPT Different Length of Intervention for Balance with Berg Balance Scale Scores Short-Term Intervention

8

749

(less than 8 weeks)

Std.Mean Difference

2.22[1.56 to 2.88]

6.59

＜0.00001

＞50%

2.04[1.08 to 3.00]

4.17

＜0.0001

＞50%

NR

NR

NR

(IV, Random ,95%CI)

Mid-Term Intervention

10

507

( 8-16 weeks)

Std.Mean Difference (IV, Random ,95%CI)

Long-Term Intervention

1

16

NR

RI PT

( more than 16 weeks) Traditional Chinese Exercise for Different Types of Physical Function Scale Short Physical Performance Battery

2

126

(SPPB)

Mean Difference

-0.46[-1.28 to 0.36]

1.09

0.27

35%

-7.02[-20.78 to 6.75]

1.00

0.32

53%

(IV, Fixed ,95%CI)

2-minute Step Test

2

126

Mean Difference

NA

SC

(IV, Random ,95%CI) Traditional Chinese Exercise for Different Types of Activities of Daily Living Ability Assessment scores Barthel Index Scores(BI)

5

464

Mean Difference

Modified Barthel Index Scores(MBI)

6

363

3.81

0.0001

＞50%

12.30 [7.48 to 17.12]

5.00

＜0.00001

＞50%

-2.57[-3.14 to -2.00]

8.88

＜0.00001

M AN U

(IV, Random ,95%CI)

15.60 [7.57 to 23.63]

Mean Difference

(IV, Random ,95%CI)

Traditional Chinese Exercise for NIH stroke scale(NIHSS) scores NIH stroke scale(NIHSS) scores

2

327

Mean Difference

(IV, Fixed ,95%CI)

TE D

IV: Inverse Variance, CI: Confidence Interval, NR: not reported , NA: Not applicable.

3.3.1. Effects of TCEs on the rehabilitation of Limb motor function 879 participants in 12 studies [33-35,37,38,48,49,51-54,56] were measured limb

EP

motor function by using Fugl-Meyer assessment (FMA). The pooled results from these studies showed that the FMA scores in the intervention group were significantly

AC C

higher than that in the control group (random effects model, SMD = 1.21, 95%CI = 0.66 to 1.77, P＜0.01). Because of the FMA in the trials included Fugl-Meyer assessment of limb (FMA), Fugl-Meyer assessment of upper limb (FMA-U), Fugl-Meyer assessment of lower limb (FMA-L), we performed a subgroup analysis for that. Results of the subgroup analysis showed that FMA scores [33,37,51,53] (random effects model, MD = 16.57, 95%CI = 9.87 to 23.26, P＜0.01), FMA-U scores [34,35,37,48,49,52,56] (random effects model, MD= 7.02, 95%CI = 3.45 to 10.60, P＜0.01), FMA-L scores [37,38,48,52,54,56] (random effects model, MD= 2.36, 95%CI = 0.71 to 4.02, P＜0.01) were significantly higher in the intervention 17

0

ACCEPTED MANUSCRIPT group compared with the control group. Moreover, there were another two subgroups analysis according to different types of TCEs and different lengths of intervention time. Regarding to different types of TCEs, results of the subgroup meta-analysis showed that Tai Chi exercise group [37,38,48,49,51] (random effects model, SMD =

RI PT

1.19, 95%CI = 0.52 to 1.86, P＜0.01), Baduanjin exercise group [54,56] (random effects model, SMD = 0.80, 95%CI = 0.22 to 1.38, P＜0.01), Yijinjing exercise group [52,53] (random effects model, SMD = 1.17, 95%CI = 0.35 to 1.99, P＜0.01) and Daoyin exercise group [33-35] (random effects model, SMD = 1.69, 95%CI = 0.36 to

SC

3.02, P= 0.01) were remarkably improved the limb function than that in the control group. On the part of different lengths of intervention time, the limb motor function in

M AN U

the intervention group were significantly improved at the short-term intervention [33-35,37,38,49,51,54] (random effects model, SMD = 1.48, 95%CI = 0.93 to 2.04, P ＜0.01) and mid-term intervention [48,52,53,56] (random effects model, SMD = 0.77, 95%CI = 0.40 to 1.14, P＜0.01).

TE D

3.3.2. Effects of TCEs on the rehabilitation of balance function 1272 participants in 19 studies [37-48,50-52,54-57] were measured balance function by Berg Balance scale(BBS). The results of meta-analysis showed that the

EP

BBS scores in intervention group were significantly higher than that in the control group (random effects model, SMD= 2.07, 95%CI = 1.52 to 2.62, P ＜ 0.01).

AC C

Regarding to different types of TCEs and different lengths of intervention time, a subgroup meta-analysis was conducted. Results of the meta-analysis showed that Tai Chi [37-48,50,51] exercise group (random effects model, SMD = 2.29, 95%CI = 1.61 to 2.98, P＜0.01) and Baduanjin [54-57] exercise group (random effects model, SMD = 1.92, 95%CI = 1.03 to 2.81, P＜0.01) were remarkably improved the balance function compared with the control group. However, only one study [52] reported that the BBS scores in Yijinjing exercise group increased compared with the control group. Besides, the BBS scores in the intervention group were significantly higher at the short-term intervention [37,38,43,44,51,54,55,57] (random effects model, SMD= 2.22, 18

ACCEPTED MANUSCRIPT 95%CI = 1.56 to 2.88, P＞0.05) and mid-term intervention [39,40,42,45-48,50,52,56] (random effects model, SMD= 2.04, 95%CI = 1.08 to 3.00, P＞0.05) compared with the control group. One study indicated that the BBS scores were increased at the long-term intervention [41].

RI PT

Moreover, Fugl-Meyer Balance scale (FM-B) was used to measure balance function in three studies [44,52,56] involving 114 participants. The pooled results from these studies indicated that the FM-B scores were not significantly different between the intervention group and the control group. (random effects model, MD=

SC

0.83, 95%CI = -0.10 to 1.77, P＞0.05). In addition, Timed-up-and-go test (TUGT) was conducted to measure balance function in four studies [12,31,41,42] involving

M AN U

202 participants. The pooled results from these studies showed that the TUGT scores in the intervention group were significantly higher than that in the control group (fixed effects model, MD= -1.77, 95%CI = -2.87 to -0.67, P＜0.01).

3.3.3. Effects of TCEs on the rehabilitation of ADL ability

TE D

827 participants in 11 studies [33,34,36-38,40,42,43,49,51,59] were measured ADL ability by Barthel Index scale (BI) and Modified Barthel Index scale (MBI). The pooled results from these studies indicated that the BI scores [36,37,43,51,59]

EP

(random effects model, MD= 15.60, 95%CI = 7.57 to 23.63, P＜0.01) and the MBI scores [33,34,38,40,42,49] (random effects model, MD= 12.30, 95%CI = 7.48 to

AC C

17.12, P＜0.01) in the intervention group were significantly higher than that in the control group.

3.3.4. Effects of TCEs on the rehabilitation of physical function 126 participants in two studies [30,31] were measured physical function by Short Physical Performance Battery(SPPB) and 2-minute step test. The pooled results from these studies indicated that the SPPB scores (fixed effects model, MD= -0.46, 95%CI = -1.28 to 0.36, P＞0.05) and 2-minute step test scores (random effects model, MD= -7.02, 95%CI = -20.78 to 6.75, P＞0.05) were not significantly different between the 19

ACCEPTED MANUSCRIPT intervention group and the control group.

3.3.5. Effects of TCEs on the rehabilitation of neurological impairment 327 participants in two studies [36,58] were measured neurological impairment

RI PT

by NIH stroke scale(NIHSS). The pooled results from the two studies showed that TCEs significantly improved participants’ neurological impairment (fixed effects

3.4. Sensitivity and publication bias analysis

SC

model, MD= -2.57, 95%CI = -3.14 to -2.00, P＜0.01).

Although the subgroup analysis has been carried out, the results still had obvious

M AN U

statistical heterogeneity. If there were more than two studies with highly heterogeneity (Chi-squared test P < 0.01 and I2 > 50%), a sensitivity analysis would be performed to identify the possible source of heterogeneity and to investigate the reliability of the results. A summary of the sensitivity analyses is shown in Table 4. After the sensitivity analysis, we found that the I2 values were obviously decreased

TE D

after removing the dubious studies, and the study results were relatively stable. The two dubious studies [38,51] were the most frequently identified sources for contributing the heterogeneity. On the aspect of the outcomes of “TCEs for FMA-U

EP

scores”, “TCEs for balance with BBS scores”, “Baduanjin for balance with BBS scores” , “short-term intervention for balance with BBS scores”, “mid-term

AC C

intervention for balance with BBS scores” , “TCEs for MBI scores”, the decrease of I2 values was inconspicuous after the sensitivity analysis.

3.5. Publication bias

The funnel plot analyses were conducted to evaluate potential publication bias in FMA scores (include 12 studies) and BBS scores (include 19 studies), the results indicated that there might be publication biases because the graphs were not symmetrical. The funnel plot analyses were shown in Figure 2 and Figure 3.

20

ACCEPTED MANUSCRIPT Table 4 Summary of sensitivity analysis. Outcomes No. trails

of

Before sensitivity analysis MD/SMD(Model

I2

P

Method for sensitivity

Outcomes

analysis

No.

) 95%CI

of

After

sensitivity analysis

MD/SMD(Model

trails

) 95%CI

6

0.45 [0.17 to 0.73]

P

I2

0.002

51%

Traditional Chinese Exercise for The Limb Motor Function with Fugl-Meyer Assessment Scores 12

1.21 [0.66, 1.77]

＜0.0001

94%

Removing Gao JB 2012 &

& Wang F 2014 & Zhang Y 2016 & Zhou ZG 2013 Traditional Chinese Exercise for The Limb Motor Function with Fugl-Meyer Assessment Scores 4

16.57 [9.87 to 23.26]

＜0.00001

74%

Removing Gao JB 2012

3

RI PT

Jiao JK 2011 & Li X 2011

13.35 [8.33 to 18.37]

＜0.00001

21%

10.63 [9.44 to 11.82]

＜0.00001

73%

4

0.59 [0.10 to 1.07]

0.02

30%

6

0.81 [0.56 to 1.05]

＜0.00001

42%

2

0.54 [0.09 to 0.99]

0.02

39%

2

0.75 [0.41 to 1.09]

＜0.0001

0

2

1.07 [0.74 to 1.41]

＜0.00001

48%

8

1.14[0.94 to 1.33]

＜0.00001

39%

5

0.61[0.38 to 0.85]

＜0.00001

0

10

1.24[0.95 to 1.52]

＜0.00001

57%

2

0.36[-0.24 to 0.96]

0.24

0

Traditional Chinese Exercise for The Limb Motor Function with Fugl-Meyer Assessment of Upper Limb Scores 7.02 [3.45 to 10.60]

0.0001

96%

Removing Jia WZ 2008 & Yang HX 2016 & Zhang B 2015

4

SC

7

6

2.36 [0.71 to 4.02]

0.005

90%

M AN U

Traditional Chinese Exercise for The Limb Motor Function with Fugl-Meyer Assessment of Lower Limb Scores Removing Zhang Y 2016 & Zhou ZG 2013 Tai Chi for The Limb Motor Function with Fugl-Meyer Assessment Scores 8

1.19 [0.52 to 1.86]

0.0005

91%

Removing Gao Juanbi 2012 & Zhou Zugang 2013

Baduanjin for The Limb Motor Function with Fugl-Meyer Assessment Scores 0.80[0.22 to 1.38]

0.006

63%

Removing Zhang Y 2016

TE D

3

Yijinjing for The Limb Motor Function with Fugl-Meyer Assessment Scores 3

1.17 [0.35 to 1.99]

0.005

80%

Removing Jia WZ 2008

Daoyin for The Limb Motor Function with Fugl-Meyer Assessment Scores 3

1.69 [0.36 to 3.02]

0.01

94%

Removing Jiao JK 2011

10

EP

Short-Term Intervention for The Limb Motor Function with Fugl-Meyer Assessment Scores 1.48[0.93 to 2.04]

＜0.00001

90%

Removing Gao JB 2012 & Jiao JK 2011

7

AC C

Mid-Term Intervention for The Limb Motor Function with Fugl-Meyer Assessment Scores 0.77 [0.40 to 1.14]

＜0.0001

62%

Removing Jia WZ 2008 & Jia WZ 2008

Traditional Chinese Exercise for Berg Balance Scale Scores 19

2.07[1.52 to 2.62]

＜0.00001

93%

Removing Bai YJ 2011 & Jia WZ 2008 & Li HY 2012 & Xie F 2008 & Xu XDg 2014 & Yang HX 2016 & Zhang M 2013 & Zhou QA 2010 & Zhou ZG 2013

Traditional Chinese Exercise for Fugl-Meyer Balance Scale 3

0.83[-0.10 to 1.77]

0.08

78%

Removing Li HY 2012

21

ACCEPTED MANUSCRIPT Tai Chi for Balance with Berg Balance Scale Scores 14

2.29[1.61 to 2.98]

＜0.00001

93%

Removing Gao JB 2012 &

8

1.02[0.81 to 1.23]

＜0.00001

0

2

2.10[1.39 to 2.81]

＜0.00001

60%

Li HY 2012 & Xie F 2008 & Xu XD 2014 & Zhou QA 2010 & Zhou ZG 2013

4

1.92[1.03 to 2.81]

＜0.0001

90%

Removing Zhang B 2015 & Zhang M 2013

Short-Term Intervention for Balance with Berg Balance Scale Scores 8

2.22[1.56 to 2.88]

＜0.00001

92%

Removing Gao JB 2012 &

4

＜0.00001

66%

0.87[0.49 to 1.25]

＜0.00001

61%

3

11.34[8.55 to 14.14]

＜0.00001

49%

4

7.82[5.60 to 10.04]

＜0.00001

60%

Li HY 2012 & Yang ZB 2013 & Zhou L 2015

2.04[1.08 to 3.00]

＜0.0001

95%

Removing Xie F 2008 & Xu XD 2014 & Zhou QA 2010

5

15.60 [7.57 to 23.63]

0.0001

95%

7

M AN U

Traditional Chinese Exercise for Barthel Index Scores (BI) Removing Cai W 2011 &

2.48[1.99 to 2.97]

SC

Mid-Term Intervention for Balance with Berg Balance Scale Scores 10

RI PT

Baduanjin for Balance with Berg Balance Scale Scores

Gao JB 2012 Traditional Chinese Exercise for Modified Barthel Index Scores (MBI) 12.30 [7.48 to 17.12]

＜0.00001

94%

Removing Li X 2011 & Xie

EP

TE D

F 2008.

AC C

6

Fig. 2 Funnel plot about Traditional Chinese exercise on the limb motor function with Fugl-Meyer assessment. 22

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Fig. 3 Funnel plot about Traditional Chinese exercises on balance function with Berg balance scale.

4. Discussion

TE D

In the present systematic review, 31 studies about effects of TCEs on the rehabilitation of limb function involving 2349 stroke participants were identified. The intervention duration ranged from two to 52 weeks, of which 12 weeks were

EP

accounted for the majority. The intervention frequency was in big difference, of which 5-7 times per week accounted for the majority. The overall effect analysis indicated

AC C

that TCEs had positive effects on the rehabilitation of limb function among stroke patients including the positive improvements on limb motor function, balance function, activities of daily living ability and neurological impairment. A recent systematic review indicated that TCEs have beneficial effects on the

balance function among stroke patients, which was same to our meta-analysis [28]. However, in the review article, subgroup analysis was not conducted to compare the effects on balance function by different types of TCEs; Timed-up-and-go test (TUGT) and Fugl-Meyer balance (FM-B) scores were only performed descriptive analysis; and a meta-analysis on BBS score was only used to indicate the effect on short-term 23

ACCEPTED MANUSCRIPT intervention. The results of our meta-analysis indicated that BBS scores were not only significantly improved at short-term intervention (less than eight weeks) but also the mid-term intervention (8-16 weeks). The different types of TCEs’ subgroup analysis showed that Tai Chi and Baduanjin had beneficial effects on improving the BBS

RI PT

scores. Furthermore, the TCEs could also improve the TUGT scores. However, due to the significantly clinical heterogeneity and different types of TCEs, the FM-B scores were insignificantly different between the intervention group and the control group.

According to traditional Chinese medicine theories (TCM), “Qi deficiency, blood

SC

stasis and blockage of meridians” are considered as one of important mechanisms causing limb dysfunction for stroke patients [60,61]. Therefore, “relax and activate

M AN U

muscles, promote blood circulation, activate meridians and collaterals” are regarded as the TCM therapeutic principles on stroke patients with limb dysfunction [62]. Based on TCM theories, TCEs are believed that they can activate muscles and tendons, dredge meridians and collaterals, promote blood circulation, and regulate internal organs [7-9]. Thus, the effects of improving limb function and a coordinative

TE D

development on physical and mental health are achieved. Our findings indicated that TCEs could be a promising intervention for the rehabilitation of limb function among stroke patients.

EP

Currently, there are many kinds of rehabilitation training for stroke patients with limb dysfunction, such as acupuncture therapy [63], robot-assisted lower-limb

AC C

rehabilitation [64], etc. However, due to the expensive medicine expenses and the lack of rehabilitative training professionals and equipment, stroke patients could not obtain timely, effective and adequate treatment, which affected the rehabilitative effects. Meanwhile, the majority of patients with hemiplegia longs for regaining independent walking ability which can increase stroke patients’ confidence in rehabilitation. In addition, good balance function was the prerequisite for recovering the ability of independent walking [55]. Our findings indicated that TCEs had positive effects on improving limb motor function, balance function and activities of daily living ability among stroke patients. More importantly, due to the positive effects, the 24

ACCEPTED MANUSCRIPT low cost, and being easy to learn without any equipment, TCEs are attracting an increasing interest in the rehabilitation field of chronic diseases. However, due to different types of TCEs and different lengths of intervention time, obvious statistical heterogeneity was showed in the results. After subgroup

RI PT

analysis, the positive effects of TCEs on the rehabilitation of limb function among stroke patients did not be affected. We continued to perform a sensitivity analysis to investigate the potential sources of clinical heterogeneity [29]. Results of the sensitivity analysis indicated that the I2 values obviously decreased after removing the

SC

dubious studies, and the study results were relatively stable. The studies of Zhou et al [38] and Gao et al [51] were the most frequently identified sources for contributing

M AN U

the heterogeneity. The reason could be that the intervention methods of these two trials were obviously different with most of the included studies which were performed TCEs either combined with or without routine rehabilitation exercise in intervention group. Zhou et al [38] used Tai Chi gait intensive exercise combining with acupuncture therapy in intervention group, and Gao et al [51] performed sitting

TE D

Tai Chi. Regarding to limb motor function, the present meta-analysis indicated that TCEs not only had significantly improvements on upper limb motor function but also the lower limb motor function. Tai Chi could produce more effects on limb motor

EP

function than Banduanjin, Yijinjing and Daoyin. However, due to the lack of enough number of included studies, a definite conclusion still cannot be drawn whether there

AC C

were positive effects on physical function between the intervention group and the control group.

The present study had some limitations including: 1) there was a big risk of bias

in the present study because of the poor methodology of some included studies. Only 14 studies reported randomization sequence generation methods, five studies reported allocation concealment methods, and 15 studies reported blinding processes; 2) the different types of TCEs, intervention duration and frequency in our included studies could contribute significantly clinical heterogeneity; 3) the different types of intervention, intervention duration and frequency in the control group might be 25

ACCEPTED MANUSCRIPT another cause of the significantly clinical heterogeneity; 4) There were also some causes of the significantly clinical heterogeneity, for example, the majority of the included studies were small sample sizes; all included studies were not performed for estimation of sample size as recommend by the CONSORT statement; only one study

RI PT

reported a reasonable drop-out rate; and only 11 studies used intention-to-treat analysis as well; 5) there might be a language bias because the included studies were only published in English and Chinese; 6) the funnel plot analyses about FMA and BBS indicated that there might be a publication bias. Publication bias could also arise

SC

from the design and execution of the single included studies and the selection for publication by authors, journal editors, and reviewers [65].

M AN U

Based on the above limitations, we put forward some implications for future related clinical research: 1) researchers should further improve the methodological quality, including rigorous design, correct random allocation and allocation concealment; although it is difficult for exercise therapy to take blinding of participants and clinicians, data collectors and outcome adjudicators should be

TE D

blinded at least; 2) the training scheme of the TCEs group and the control group should be standardized, including training type, training duration, training intensity and frequency; 3) in order to monitor the long-term effect of TCEs and guide patients

EP

how to exercise TCEs at home, it is important to consider follow-up in studies; 4) the sample size should be correctly estimated, especially the drop-out sample should be

AC C

given specific explain and intention-to-treat analysis; 5) researchers should ensure the quality of each part of the study and control the risk, such as estimating each implement of the study whether strictly follow study scheme, making ensure the authenticity and reliability of clinical data, and making data be traced as well.

5. Conclusion This present study indicated that TCEs produced positive significantly improvements on limb motor function, balance function, activities of daily living ability and neurological impairment among stroke patients. After subgroup analysis 26

ACCEPTED MANUSCRIPT and sensitivity analysis, the positive effects did not be affected by different types of TCEs and different lengths of intervention time. Due to the lack of enough number of included studies, a definite conclusion still could not be drawn whether there were positive effects on physical function between the intervention group and the control

RI PT

group. In view of the limitations of the methodological quality and sample size of the included studies, more large-scale, high-quality, multiple-center RCTs are required to further verify the efficacy of TCEs on the rehabilitation of limb function among

SC

stroke patients in the future.

Acknowledgements

M AN U

We are grateful to Professor David Brunt for revising the language in the article.

Appendix

Appendix A Selected Searching Strategies. Item Searching strategies Pubmed #1

TE D

Qigong Qigong[MeSH Terms]

#2

((((qigong[Title/Abstract]) OR Qigong[Title/Abstract]) OR Qi Gong[Title/Abstract]) OR Ch'i

Kung[Title/Abstract]) OR breathing exercise*[Title/Abstract] #1 OR #2

#4

stroke[MeSH Terms]

EP

#3 #5

((((((((((((((((Stroke*[Title/Abstract]) OR Cerebrovascular Disorder*[Title/Abstract]) OR Brain

Infarction*[Title/Abstract]) OR apoplexy*[Title/Abstract]) OR cerebral arterial

AC C

thrombosis[Title/Abstract]) OR CVA[Title/Abstract]) OR Cerebrovascular Diseases[Title/Abstract]) OR cerebral vascular accident*[Title/Abstract]) OR Brain Vascular Accident*[Title/Abstract]) OR cerebral vascular Diseases[Title/Abstract]) OR cerebrovascular accident*[Title/Abstract]) OR Cerebrovascular Occlusion*[Title/Abstract]) OR Brain Vascular Disorder*[Title/Abstract]) OR celebral vascular inciden*[Title/Abstract]) OR Cerebral Hemorrhage*[Title/Abstract]) OR Intracranial Hemorrhage*[Title/Abstract]) OR Brain Hemorrhage*[Title/Abstract] #6 #7

#4 OR #5 (((((((("randomized controlled trial"[Publication Type]) OR "controlled clinical trial"[Publication

Type]) OR "randomized"[Title/Abstract]) OR "randomised"[Title/Abstract]) OR "randomly"[Title/Abstract]) OR "trial"[Title/Abstract]) OR "groups"[Title/Abstract] #8

(animals[MeSH Terms] NOT (humans[MeSH Terms] AND animals[MeSH Terms]))

#9

#7 NOT #8

#10

#3 AND #6 AND #9 27

ACCEPTED MANUSCRIPT Cochrane Library Qigong #1

MeSH descriptor: [Qigong] explode all trees

#2

qigong:ti,ab,kw or Qigong:ti,ab,kw or Qi Gong:ti,ab,kw or Ch'i Kung:ti,ab,kw or breathing

exercise*:ti,ab,kw (Word variations have been searched) #1 or #2

#4

MeSH descriptor: [Stroke] explode all trees

#5

Stroke* or Cerebrovascular Disorder* or Brain Infarction* or apoplexy* or cerebral arterial

RI PT

#3

thrombosis or CVA or Cerebrovascular Diseases or cerebrovascular accident* or Brain Vascular

Accident* or cerebral vascular Diseases or cerebral vascular accident* or Cerebrovascular Occlusion* or Brain Vascular Disorder* or celebral vascular inciden* or (Cerebral near/3 Hemorrhage*) or

SC

Intracranial Hemorrhage* or (Brain near/3 Hemorrhage*):ti,ab,kw (Word variations have been #6

#4 or #5

#7

#3 and #6

#8

#7 in Trials

M AN U

searched)

Embase Qiqong 'qigong'/exp

#2

qigong:ab,ti OR 'qi gong':ab,ti OR 'breathing exercise*':ab,ti OR 'chi kung':ab,ti

#3

#1 OR #2

#4

'stroke'/exp

#5

'cerebrovascular accident'/exp

#6

stroke*:ab,ti OR (cerebrovascular:ab,ti AND disorder*:ab,ti) OR (brain:ab,ti AND

TE D

#1

infarction*:ab,ti) OR apoplexy*:ab,ti OR (cerebral:ab,ti AND arterial:ab,ti AND thrombosis:ab,ti) OR

EP

cva:ab,ti OR (cerebrovascular:ab,ti AND diseases:ab,ti) OR (cerebrovascular:ab,ti AND accident*:ab,ti) OR (brain:ab,ti AND vascular:ab,ti AND accident*:ab,ti) OR (cerebral:ab,ti AND

AC C

vascular:ab,ti AND diseases:ab,ti) OR (cerebral:ab,ti AND vascular:ab,ti AND accident*:ab,ti) OR (cerebrovascular:ab,ti AND occlusion*:ab,ti) OR (brain:ab,ti AND vascular:ab,ti AND disorder*:ab,ti) OR (celebral:ab,ti AND vascular:ab,ti AND inciden*:ab,ti) OR (cerebral NEAR/3 hemorrhage*):ab,ti OR (intracranial:ab,ti AND hemorrhage*:ab,ti) OR (brain NEAR/3 hemorrhage*):ab,ti #7 #8 #9

#4 OR #5 OR #6

'randomized controlled trial'/exp randomized:ab,ti AND controlled:ab,ti AND trial:ab,ti OR randomized:ab,ti OR randomly:ab,ti

OR trial:ab,ti OR (groupsrandomized:ab,ti AND controlled:ab,ti AND trial:ab,ti) OR (controlled:ab,ti AND clinical:ab,ti AND trial:ab,ti) OR groups:ab,ti #10

#8 OR #9

#11

'animal'/exp OR 'nonhuman'/exp OR 'animal experiment'/exp

#12

'human'/exp

#13

#11 AND #12 28

ACCEPTED MANUSCRIPT #14

#11 NOT #13

#15

#10 NOT #14

#16

#3 AND #7 AND #15

References 2017.

RI PT

1. World Health Organization. The Top 10 Causes of Death; http://www.who.int/mediacentre/factsheets/fs310/en/. (Accessed March 26, 2017)

2. Chumney D, Nollinger K, Shesko K, et al. Ability of functional Independence Measure to accurately predict functional outcome of stroke-specific population: systematic review. J Rehabil Res Dev 2010;47(1):17-29.

SC

3. Zhu ZH, Cui LL, Yin MM, et al. Hydrotherapy vs. Conventional land-based exercise for improving walking and balance after stroke: a randomized controlled trial. Clin Rehabil 2016;30(6):587–593.

M AN U

4. Liu L, Wang D, Wong K S L, et al. Stroke and Stroke Care in China: Huge Burden, Significant Workload, and a National Priority. Stroke 2011;42(12): 3651-3654.

5. Zhang Y, Huang L, Su Y, et al. The Effects of Traditional Chinese Exercise in Treating Knee Osteoarthritis: A Systematic Review and Meta-Analysis. PLoS One 2017;12(1): e0170237. 6. Luo X, Zhang J, Castelberg R, et al. The Effects of Traditional Chinese Exercise in Patients with Chronic Obstructive Pulmonary Disease: A Meta-Analysis. Plos One 2016;11(9):e0161564. 7. Jiang Y, Zou J. Analysis of the TCM theory of traditional Chinese health exercise. J Sport Health Sci 2013;2(4):204-208.

TE D

8. Lin LH, Jiang CY. The value and Significance of Traditional Chinese Health Exercises to Promote the Health of the Disabled Persons. Strike-Wushu Science 2011;08(2):102-104. 9. Yin H, Liu JB. Application of Chinese traditional Qigong in sanatorium. Chinese Journal of Convalescent Medicine 2010;(11):995-996.

10. Tsang H W, Mok C K, Au Yeung Y T, et al. The effect of Qigong on general and psychosocial

EP

health of elderly with chronic physical illnesses: a randomized clinical trial. Int J Geriatr Psychiatry 2003;18(5):441-449.

AC C

11. Zhang WC. Theoretical research on Qigong exercise on the basis of life view of Physique, Qi and Spirit [dissertation]. Nanjing: Nanjing University of Traditional Chinese Medicine; 2010. 12. Kim H, Kim Y L, Lee S M. Effects of therapeutic Tai Chi on balance, gait, and quality of life in chronic stroke patients. Int J Rehabil Res 2015;38(2):156-161. 13. Li GC, Yuan H, Zhang W. Effects of Tai Chi on health related quality of life in patients with chronic conditions: A systematic review of randomized controlled trials. Complement Ther Med 2014;22(4):743-755. 14. Zheng GH, Fang QY, Chen B, et al. Qualitative Evaluation of Baduanjin (Traditional Chinese Qigong) on Health Promotion among an Elderly Community Population at Risk for Ischemic Stroke. Evid Based Complement Alternat Med 2015;(4):1-10. 15. Chen MC, Liu HE, Huang HY, et al. The effect of a simple traditional exercise programme (Baduanjin exercise) on sleep quality of older adults: A randomized controlled trial. Int J Nurs Stud 2012;49(3):265-273. 29

ACCEPTED MANUSCRIPT 16. Ye Q. Clinical studies of Yijinjing intervention to the elderly hyperlipidemia [dissertation]. Nanjing: Nanjing University of Traditional Chinese Medicine; 2014. 17. Dou SD, Xu RX, Wu NX, et al. Research of modern application of health Qigong Yijinjing exercise. Chinese archives of traditional Chinese Medicine 2017;35(5):1083-1085. 18. Tian BW. A study on the effect of Wuqinxi exercise on the proprioception and balance function of elderly female patients with KOA [dissertation]. Beijing: Beijing Sport University; 2012.

RI PT

19. Jiang YX. The clinical study of health Qigong Wuqinxi exercises effect on nonspecific lower back pain [dissertation]. Nanjing: Nanjing University of Traditional Chinese Medicine; 2016.

20. Zheng L, Zhang C, Wu XK, et al. Influence of healthcare qigong-6-character formulas (Liuzijue) on heart function in patients with chronic heart failure. Chinese Journal of Evidence-Bases Cardiovascular Medicine 2017;9(6):659-662.

SC

21. Li DX. The effect of six-character formula respiratory gymnastics in relation to respiratory mechanics in patients with chronic obstructive pulmonary diseases [dissertation]. Fujian: Fujian University of Traditional Chinese Medicine; 2011.

M AN U

22. Zhang J, Wang Q. Analysis of the time value of “Dao yin” method in Zhu Bing Yuan Hou Lun. Chinese Journal of Basic Medicine in Traditional Chinese Medicine 2014;20(9):1203-1204,1275. 23. Liu Fz. The study of Guiding technique of TCM played in the recovery of the Bi syndrome of limbs [dissertation]. Guangzhou: Guangzhou University of Traditional Chinese Medicine; 2014. 24. Wang X, Pi Y, Chen P, et al. Traditional Chinese Exercise for Cardiovascular Diseases: Systematic Review and Meta-Analysis of Randomized Controlled Trials. J Am Heart Assoc 2016;5(3):e002562.

25. Wang X, Pi Y, Chen B, et al. Effect of traditional Chinese exercise on the quality of life and

TE D

depression for chronic diseases: a meta-analysis of randomised trials. Sci Rep 2015;5:15913. 26. Zheng GH, Fang QY, Chen B, et al. Qualitative Evaluation of Baduanjin (Traditional Chinese Qigong) on Health Promotion among an Elderly Community Population at Risk for Ischemic Stroke. Evid Based Complement Alternat Med 2015;(4):1-10. 27. Zheng G, Chen B, Fang Q, et al. Primary prevention for risk factors of ischemic stroke with

EP

Baduanjin exercise intervention in the community elder population: study protocol for a randomized controlled trial. Trials 2014;15(1):1-10. 28. Chen BL, Guo JB, Liu MS, et al. Effect of Traditional Chinese Exercise on Gait and Balance

AC C

for Stroke: A Systematic Review and Meta-Analysis. Plos One 2015;10(8):e0135932. 29.

Cochrane

Handbook

for

systematic

reviews

of

intervention

[Internet].

http://training.cochrane.org/handbook (Accessed March 27, 2017). 30. Taylor-Piliae R E, Hoke T M, Hepworth J T, et al. Effect of Tai Chi on Physical function, Fall Rates and Quality of Life Among Older Stroke Survivors - Archives of Physical Medicine and Rehabilitation. Arch Phys Med Rehabil 2014;95(5):816-824. 31. Taylorpiliae R E, Coull B M. Community-based Yang-style Tai Chi is safe and feasible in chronic stroke: a pilot study. Clin Rehabil 2012;26(2):121. 32. Au-yeung S S, Huichan C W, Tang J C. Short-form Tai Chi improves standing balance of people with chronic stroke. Neurorehabil Neural Repair 2009;23(5):515.

30

ACCEPTED MANUSCRIPT 33. Zheng DC, Li Y, Chen LY. The effects of Chao-yuan-fang Daoyin combined routine rehabilitation therapy on improvements of motor function among stroke patients. Guangdong Medical Journal 2013;34(20):3196-3198. 34. Jiao JK. The upper limb’s Daoyin exercises for the treatment of upper limb dysfunction in stroke patients with hemiplegia: clinical study. Journal of Emergency in Traditional Chinese Medicine 2011;20(1):16-17. 35. Wang F, Wang JH, ZHANG LJ, et al. Self-compiled Daoyin Combined with Manipulation & Rehabilitation Medicine 2014;5(2):34-36.

RI PT

Constraint-induced Movement Therapy on Dysfunction of Cerebral Infarction. Chinese 36. Zheng WC, Zhang YH, Jiang XQ, et al. Effects of Persistent Tai Chi exercise on Rehabilitation in Ischemic Stroke: A Prospective Randomized Controlled Tria. Chinese Journal of Integrative Medicine on Cardio /Cerebrovascular Disease 2015;13(3):304-307.

SC

37. Zhou L. Clinical observation on the effect of modified Taijiquan in the rehabilitation among ischemic stroke patients [dissertation]. Beijing: Beijing University of Chinese Medicine; 2015. 38. Zhou ZG. The effect of intensive training Taiji gait on walking ability and gait parameter of

M AN U

stroke patients with hemiplegia [dissertation]. Chengdu: Chengdu University of Traditional Chinese Medicine; 2013.

39. Wang XB, Hou MJ, Tao J, et al. The effect of Tai Chi Yunshou on gait of community-based hemiplegic patients after stroke. Chinese Journal of Rehabilitation Medicine 2016;31(12): 1328-1333.

40. Xie F. The effects of Taiji rehabilitation exercises on rehabilitation of balance disorder among stroke patients [dissertation]. Jilin: Jilin University; 2008.

41. Huang YD. The study of Taijiquan exercise on movement function and nerve excitability in

TE D

patients with stroke [dissertation]. Shanghai: Shanghai Institute of Physical Education; 2016. 42. Wang XY. Effect of Tai chi "yunshou" training

on balance function in

patients with hemiplegia after stroke in community-based rehabilitation [dissertation]. Henan: Henan University of Traditional Chinese Medicine; 2016. 43. Yang ZB, Liu D, Chang YS, et al. Clinical study on treatment of post stroke hemiplegia

EP

balance disorder with balance of Taiji. Contemporary Medicine 2013;24(19):5-7. 44. Li HY. The effect of Tai Chi training and the ProKin balance therapy system training on the equilibrium function in hemiplegic patient after stroke [dissertation]. Suzhou: Suzhou University;

AC C

2012.

45. Zhou QA, Xu JY, Hu AP, et al. Effect of Taijiquan on rehabilitation of patients with cerebral infarction. Chinese Journal of Practical Nervous Disease 2010;13(22):20-22. 46. Fu CX, Zhang QY. Effects of taijiquan on balance function and walking ability of stroke hemiplegic patients in convalescent phase. Chinese Journal of Rehabilitation Medicine 2016;31(5): 536-539. 47. Liu TJ, Qin P, Chen XZ. Effect of Taijiquan on balance function of stroke patients. Chinese Journal of Physical Medicine and Rehabilitation 2009;31(11):781-782. 48. Yang HX, Tan Q. Clinical observation of Taijiquan in rehabilitation of motor dysfunction in stroke patients. Chinese Journal of Rehabilitation Medicine 2016;31(10):1146-1148. 49. Li X, Yang SL, Li TJ, et al. Application of Taijiquan exercise imagination in rehabilitation of stroke patients. Fujian Journal of Traditional Chinese Medicine 2011;42(4):5-6. 31

ACCEPTED MANUSCRIPT 50. Xu XD, Bai J, Zhang HL, et al. Analysis of curative effect for Tai chi on the balance function of patients with hemiplegia after stroke. Hebei Journal of Traditional Chinese Medicine 2014;36(8):1149-1150. 51. Gao JB, Feng YZ, Li QN. Observation of the effect of Tai Chi ten type exercise on stroke patients with hemiplegia. Chinese Manipulation& Rehabilitation Medicine 2012;3(7):71-72. 52. Jia WZ. The effects of Yijinjing exercise on the motor function,balance ability and emotion in stroke patients [dissertation]. Hebei: Hebei Normal University; 2008.

RI PT

53. Zhang GD, Liang GQ, Hui RH. Acupuncture Combined with Muscle-bone Strengthening Exercise Training on Hemiplegia after Stroke Patients with Motor function and Psychological Impact. Journal of Changchun University of Chinese Medicine 2016;32(5):967-969.

54. Zhang Y, Li L. Treatment of 31 Cases of Balance function Disorder after Cerebral Stroke by Eight-section Brocade and Balance Training. Shandong Journal of Traditional Chinese Medicine

SC

2016;35(8)716-718.

55. Bai YJ, Mao HY, Guo J, et al. Research on Baduanjin Combined with function Training in Improving Balance functions of Stroke Patients. China Journal of Chinese Medicine

M AN U

2011;26(10):1231-1232.

56. Zhang B, Liu JR. Effects of Eight-section Brocade on motor function of upper and lower limbs and balance ability in stroke patients with hemiplegia. Modern Health 2015;8(16):246-247. 57. Zhang M, Guo J, Bai YJ, et al. Effects of Eight-section Brocade combined rehabilitation exercise on the balance function in stroke patients. Lishizhen Medicine and Material Medical Research 2013;24(10):2438-2439.

58. Guo J, Zhang M, Bai YJ, et al. Effects of Eight-section Brocade combined rehabilitation exercise on the the clinical effect in stroke patients. Lishizhen Medicine and Material Medical

TE D

Research 2013;24(12):2941-2942.

59. Cai W, Liang CY. Effects of Sitting Baduanjin on the ability of Activities Daily Living of patients

with

stroke

2011;11(11):810-811.

sequelae

in

community.

Journal

of

Nursing

Administration

60. Wang YR, Hu JP, Wang J, et al. The Formation and Development of Etiology and Theory

for

Stroke.

EP

Pathogenesis

Clinical

journal

of

traditional

chinese

medicine

2017;(03):303-306.

61. Hu LT, Cai FN, Wang YL. Analysis of etiology and pathogenesis of apoplexy. Chinese

AC C

Journal of Integrative Medicine on Cardio /Cerebrovascular Disease 2017;(07):883-885. 62. Li MX, Guan L, Ling ZC, et al. Progress of application of traditional Chinese medicine exercise therapy in rehabilitation of limbs after stroke. Chinese Journal of Convalescent Medicine 2017;26(3):241-244.

63. Zheng H, Liu WL, ShangGuan H, et al. Effect of acupuncture and moxibustion treatment on limb motor dysfunction among stroke patients: meta-analysis. Chinese Journal of Rehabilitation Medicine 2016;31(2):217-221. 64. Li H, Fu JM, Gu XD, et al.The effects of robot-assisted lower-limb rehabilitation plus psychological intervention on post-stroke depression. Chinese Journal of Physical Medicine and Rehabilitation 2013;35(8):630-633. 65. Thornton A, Lee P. Publication bias in meta-analysis: its causes and consequences. J Clin Epidemiol 2000;53(2):207-216. 32