Acupuncture for acute ischemic stroke: A systematic review and meta-analysis of randomized controlled trials

Abstract

Objective

Assess the safety and effectiveness of acupuncture for acute ischemic stroke (AIS).

Methods

We conducted a comprehensive search across the PubMed, Cochrane Library, Embase, Web of Science, China National Knowledge Infrastructure (CNKI), China Science and Technology Journal Database (VIP), Wanfang, and SinoMed databases from their inception until October 3, 2023. Two reviewers screened eligible randomized controlled trials (RCTs) according to criteria and extracted data using a pre-established form. Cochrane tool was used for risk of bias assessment, and Revman 5.3 was used for subsequent meta-analysis. The GRADE tool will be used to assess the quality of evidence.

Results

Thirty-one RCTs were included, involving 3604 patients. Meta-analysis showed that compared with conventional treatments (CTs), acupuncture combined with CTs could improve in National Institutes of Health Stroke Scale (NIHSS) and Efficiency (MD:1.70, 95 %CI [-2.27, -1.14], P < 0.00001;RR: 1.21, 95 %CI [1.12, 1.31], P < 0.00001,);On the Chinese Stroke Scale (CSS) and effectiveness based on CSS, acupuncture showed positive effects (MD:4.33, 95 %CI [-5.67, -2.98], P < 0.00001; RR: 1.26, 95 %CI [1.13, 1.41], P < 0.0001). Furthermore, ADL, Fugl-Meyer Assessment Scale (FMA), prognosis analyses also showed the effectiveness of acupuncture. (SMD: 0.98, 95 % CI [0.64, 1.31], P < 0.00001; MD: 16.46, 95 %CI [12.56, 20.35], P < 0.00001; RR: 0.38, 95 %CI [0.16, 0.89], P = 0.03). However, the certainty of evidence was low.

Conclusion

According to current evidence, Acupuncture may be effective and safe for AIS. The future still needs high-quality evidence to support this conclusion.

1. Introduction

Stroke is the main cause of death and disability worldwide.¹ Acute ischemic stroke (AIS) is a pathological subtype of stroke that is caused by a sharp decrease in blood flow to the brain, resulting in damage to brain cells.² In 2020, a cross-sectional study involving 676,394 Chinese adults reported that the prevalence, incidence, and mortality rates of stroke were 2.6 %, 505.2 per 100,000 person-years, and 343.4 per 100,000 person-years, respectively. Specifically, AIS accounted for 15.5 million cases of stroke, representing 86.8 % of all strokes in 2020. In developing regions worldwide, the incidence of AIS is progressively increasing, thus increasing the burden of stroke in these areas.^3,4 According to guidelines, the acute phase of AIS can last up to 2 weeks after onset.⁵ The first two weeks was an important period for prevention.^6,7 Currently, the recommended treatment method for AIS is improving cerebral circulation.⁸ However, thrombolysis has limitations, such as a short time window (within 4.5 h), a high risk of bleeding, and obvious reperfusion injury. Anticoagulants are also associated with bleeding risks and many contraindications.⁹ According to the Global Burden of Disease (GBD) survey results, even with routine medical treatment (including rehabilitation), the proportion of patients with ischemic stroke who rely on others for living after discharge remains high.¹⁰ Therefore, it is imperative to develop a complementary therapy aimed at accelerating the recovery of patients with AIS, thereby enhancing their quality of daily living. According to a 2019 World Health Organization (WHO) report, acupuncture was the most widely used traditional and complementary medicine.¹¹ The 2018 Chinese Guidelines for the Diagnosis and Treatment of Acute Ischemic Stroke recommend acupuncture as a promising complementary therapy for AIS.⁵ However, additional reliable evidence is needed to support this recommendation.

There are divergent views regarding the effects of acupuncture in managing AIS. Several trials^12,13 have indicated that acupuncture treatment has no beneficial effect on neurological function or activity of daily living (ADL) in acute stroke patients. Conversely, numerous clinical trials^14,15 have shown that acupuncture leads to meaningful improvements in neurological function, motor function, and ADL in patients with AIS, thereby contributing to a favorable prognosis.

Although numerous systematic reviews of acupuncture for ischemic stroke treatment have been conducted in recent years, most have focused on its effectiveness and safety during the stroke recovery phase. There were two systematic reviews and meta-analyses evaluating the effectiveness and safety of acupuncture in the treatment of AIS recently. One previous study was a Cochrane review¹⁶ that did not ascertain the role of acupuncture used as an alternative or complementary therapy in treatment. In fact, acupuncture generally serves as a complementary modality in the clinical management of AIS.^17,18 Therefore, a more pragmatic approach involves scrutinizing the safety and effectiveness of acupuncture within the context of its role as a complementary therapy. Another study¹⁹ investigated the differential effectiveness of Governor meridian acupuncture versus standard acupuncture in the treatment of AIS. Essentially, it examined the therapeutic differences attributable to acupoint specificity.

Therefore, we conducted the current meta-analysis to assess the effectiveness and safety of acupuncture as a complementary treatment for AIS compared with conventional treatments (CTs; i.e., routine medicine therapy and/or rehabilitation).

2. Methods

This systematic review was registered on the PROSPERO website on September 19, 2023 (ID: CRD42023464998) and was reported by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) for acupuncture.²⁰ The PRISMA-A checklist²¹ was shown in Supplement 1, Table S1.

2.1. Search strategy

We systematically searched the PubMed, Cochrane Library, Embase, Web of Science, China National Knowledge Infrastructure (CNKI), China Science and Technology Journal Database (VIP), Wanfang, and SinoMed databases from inception to October 3, 2023, to identify relevant studies. We conducted a search using the following keywords: ‘Acute stroke’, ‘Acupuncture’, and ‘Randomized Controlled Trial’. We also searched the reference lists of the included literature as well as relevant reviews and meta-analyses published in recent years. To ensure the comprehensiveness of our search, we consulted with an expert and explored the WHO International Clinical Trials Registry platform for additional eligible studies. The details of the search strategy are presented in Supplement 1, Table S2.

2.2. Eligibility criteria

2.2.1. Participants

Included were participants of both sexes who had AIS onset within the preceding 14 days, were diagnosed with ischemic stroke in accordance with the WHO diagnostic criteria, and had brain computed tomographic or magnetic resonance imaging scans available to exclude intracerebral hemorrhage,²² had no severe consciousness or speech communication disorders and provided informed consent.

2.2.2. Interventions

The included intervention in the experimental group was acupuncture treatment (AT), including electroacupuncture, abdominal acupuncture, body acupuncture, scalp acupuncture, and eye acupuncture. We excluded interventions lacking needle insertion (e.g., auricular pressure, acupressure, and laser stimulation), acupuncture influenced by extraneous factors, or interventions without needle retention (e.g., acupoint injection, warming needle, or fire needle). The included intervention of control group was sham acupuncture or no AT. Sham acupuncture refers to the use of (1) a needling prick on the surface of the skin at non-acupuncture points and (2) needles attached to the surface of the skin using a well-designed device. Both two groups had to receive the same conventional treatments (CTs). CTs refers to routine drug therapy and/or rehabilitation for AIS.^8,23

2.2.3. Outcomes

Included were studies reporting any of the following outcomes.

2.2.4. Primary outcomes

• 1.Neurological function: Neurological function was measured using the National Institutes of Health Stroke Scale²⁴ (NIHSS), with a total score ranging from 0 to 42. Higher scores indicate more severe neurological impairment.
• 2.Clinical effectiveness based on the NIHSS: Clinical effectiveness was evaluated based on the NIHSS score. According to the Clinical Neurological Function Deficit Scale for Stroke Patients (1995) ,²⁵ the efficacy index was calculated as follows: [(pretreatment NIHSS score – posttreatment NIHSS score) ÷ pretreatment NIHSS score] × 100 %. An efficacy index exceeding 18 % indicated clinical effectiveness.

2.2.5. Secondary outcomes

• 1.Neurological function: Neurological function was measured using the Clinical Neural Function Deficit Score Criteria for Stroke²⁵ (a Chinese Stroke Scale, CSS), with a total score ranging from 0 to 45. Higher scores indicate more severe neurological impairment.
• 2.Clinical effectiveness based on CSS: Clinical effectiveness was evaluated based on CSS scores. The efficacy index was calculated as described above. An efficacy index exceeding 18 % indicated effectiveness.
• 3.Activity of daily living (ADL) and disability: ADL were measured using Modified Barthel Index²⁶ (MBI) and the Barthel Index²⁷ (BI), with scores ranging from 0 to 100. Higher scores indicate better ADL. Disability was measured using modified Rankin Scale²⁸ (mRS), with scores ranging from 0 to 6. A higher score indicated more severe disability, and a score of 6 indicated death.
• 4.Motor function: Motor function was measured using the Fugl-Meyer Assessment²⁹ (FMA), with a total score ranging from 0 to 100. Higher scores indicate better motor function.
• 5.Prognosis: Adverse prognosis was defined as dependency on others for daily living, deterioration, or death based on BI scores at the six-month follow-up.
• 6.Adverse events (AEs): AEs related to acupuncture were recorded for each study.

2.3. Study design

We included RCTs examining the use of acupuncture to treat AIS. All eligible studies were published in English or Chinese.

2.4. Selection of studies and data extraction

The included literature was organized using EndNote X9. Following both automated and manual deduplication, two researchers (YF, PC) independently screened the titles and abstracts to identify potential studies. Subsequently, the researchers (YF, PC) thoroughly reviewed the full texts of these potential studies to determine eligibility. Any disagreements were resolved by consulting a third party (DY) to reach a consensus. Two researchers (SS, ZQ) independently extracted the data from the included articles, and a senior researcher (JC) checked the data. Any disagreements were resolved by discussion and consensus. The following data were extracted from the included studies: title, publication year, subjects, intervention, outcome, and follow-up visits. These data were recorded in Excel. See Table 1 for details.

Table 1.

Characteristics of included studies.

First authors (year) [ref]	Sample size (M/F), age (year) Onset	Interventions (Regimen), follow up		Outcomes	Results	Acupoints	AEs Funding
		Treatment group	Control group
Chu (2017)30	(A) 60(35/25), 71.0 (B) 60(34/26), 69.0 within 2 weeks	(A) AT (6 times/week for 2 weeks), plus B NR	(B) Edaravone injection (30mg/12 h), Piracetam injection (10 g/d, 2 times/d)	1) NIHSS 2) FMA 3) BI	1) MD, −1.84 [−2.27, −1.41], P < 0.05 2) MD, 17.00 [15.35, 18.65], P < 0.05 3) MD, 1.09 [0.69, 1.48], P < 0.05	LI15, LI11, LI10, TE5, LI4, GB30, ST31, ST32, ST36, SP6, ST41, LR3.	NR NR
Cui (2022)31	(A) 60(29/31), 53.1 (B) 60(25/35), 52.5 within 24 h	(A) AT (6 times/week for 2 weeks), plus B NR	(B) Butylphthalide sodium chloride injection (100ml/time, 2 times/day)	1) NIHSS 2) Clinical effectiveness (NIHSS)	1) MD, −1.59 [−1.82, −1.36], P= 0.001 2) MD, 1.14 [1.00, 1.29], P = 0.04	PC6, SP6, GV26, LU5, BL40, HT1.	Dry mouth (A:2, B:0), Abdominal distension (A:2, B:1), Headache (A:1, B:1), Dry vomiting and nausea (A:2, B:1) NR
Li (2016)32	(A) 30(20/10), 65.0 (B) 30(17/13), 59 1 to 14 days	(A) Scalp and body AT (5 times/6 days for 18 days), plus B NR	(B) Aspirin (100 mg/d), Nimodipine (30mg/d, 3 times/d), Edaravone injection (30mg/12 h)	1) Clinical effectiveness (NIHSS) 2) FMA-upper limb 3) FMA-lower limb	1) MD, 1.37 [1.01, 1.86], P < 0.05 2) MD, 10.48 [7.42, 13.54], P < 0.01 3) MD, 10.16 [6.86, 13.46], P < 0.05	Anterior oblique line (GV21 to GB6) and posterior oblique line (GV20 to GB7), BP-HN2, BP-LE6, BP-UE2.	NR NR
Li (2016a)33	(A) 65(37/28), 56.8 (B) 65(35/30), 56.9 5 to 22 h	(A) Scalp and body AT (5 times/week for 4 weeks), plus B NR	(B) Rehabilitation training + Drug therapy: NR	1) FMA 2) BI	1) MD, 17.70 [15.16, 20.24], P < 0.05 2) MD, 2.08 [1.65, 2.50], P < 0.05	Upper 2/5 of the contralateral motor area of the affected limb, the motor sensory area of the foot, SJ10, TE5, GB34, BL62, HT1, SP6, SP10.	NR NR
Li (2022)34	(A) 30(21/7), 62 (B) 30(18/8), 62 within 1 week	(A) Electric AT (1 time/day for 10 days), plus B NR	(B) Drug therapy: NR	1) NIHSS 2) MBI	1) MD, −0.69 [−1.33, −0.05], P < 0.05 2) MD, 0.65 [0.10, 1.20], P < 0.05	GV26, GV29, LI15, LI11, LI10, TE5, LI4, GB31, ST36, SP6, LR3.	NR Yes
Liang (2014)35	(A) 40(21/19), 52.3 (B) 40(22/18), 53.2 6 to 21 h	(A) Scalp and body AT (5 times/week for 4 weeks), plus B NR	(B) Clopidogrel (NR), Oxiracetam (NR)	1) FMA 2) MBI	1) MD, 17.13 [12.04, 22.22], P < 0.05 2) MD, 0.79 [0.34, 1.25], P < 0.01	The upper 1 / 5, middle 2 / 5 and foot movement areas contralateral to the affected limb, SJ10, TE5, HT1, LU5, GB34, BL62, SP6, SP10.	NR NR
Shen (2012)14	(A) 145(46/99), 60.2 (B) 145(52/92), 61.4 1 to 14 days	(A) AT (1 time/day for 4 weeks), plus B NR	(B) Sham acupuncture + Ebatroxobin (5 BU/2 days, 3 times), Aspirin (75mg/d)	1) NIHSS 2) BI 3) CSS	1) MD, −2.20 [−2.81, −1.59], P < 0.01 2) MD, 0.64 [0.40, 0.88], P = 0.01 3) MD, −3.69 [−4.89, −2.49], P < 0.001	PC6, GV26, SP6, HT1, BL40, LU5.	Transient fainting (A:1, B:0), Pain at needling area (A:10, B:17), Bleeding (A:0, B:4), High blood pressure (A:1, B:1) Yes
Song (2022)36	(A) 71(36/35), 63 (B) 30(17/13), 64 within 4.5 h	(A) AT (1 time/day for 6 days), plus B NR	(B) rt-PA (0.9 mg kg-1, maximum dose: 90 mg)	1) Clinical effectiveness (NIHSS)	1) MD, 1.12 [1.01, 1.23], P < 0.05	GV26, PC6, SP6, HT1, LU5, BL40.	Gingival bleeding (A:1, B:2), Fundus retinal hemorrhage (A:1, B:0), Skin bleeding around the lips (A:0, B:1), Cerebral hemorrhage (A:0, B:6) Yes
Tan (2013)37	(A) 32(15/17), 57.4 (B) 31(21/9), 60.3 within 72 h	(A) Electric AT (6 times/week for 2 weeks), plus B NR	(B) Aspirin (100 mg/d), Clopidogrel (75 mg/d), Atorvastatin (20 mg/d)	1) NIHSS 2) FMA 3) Clinical effectiveness (CSS)	1) MD, −1.65 [−3.04, −0.26], P < 0.05 2) MD, 11.10 [0.67, 21.53], P < 0.05 3) MD, 1.34 [1.04, 1.72], P < 0.05	LI15, TE14, LI11, LI4, LU5, PC6, ST36, GB34, ST40, GB39, SP6, LR3.	No serious AEs Yes
Wang (2008)38	(A) 60(32/28), 66 (B) 60(29/31), 64 within 7 days	(A) Eye AT (1 time/day for 2weeks), plus B NR	(B) Aspirin (100 mg/d), Nicholin (1.0 mg/d)	1) CSS 2) Clinical effectiveness (CSS)	1) MD, −5.56 [−7.15, −3.97], P < 0.01 2) MD, 1.46 [1.20, 1.78], P < 0.05	Upper Foci Area, Lower Foci Area	NR NR
Wang (2014)39	(A) 170(106/64), 65 (B) 170(109/61), 64 within 2 weeks	(A) Electric AT (NR for 4 weeks), plus B 6 months	(B) Rehabilitation training + Aspirin (100 mg/d)	1) NIHSS 2) BI<75	1) MD, −2.20 [−2.73, −1.67], P = 0.002 2) MD, 0.31 [0.21, 0.45], P = 0.000	PC6, GV26, SP6	Distending pain (A:5, B:0) Yes
Wang (2022)40	(A) 41(18/20), 67.2 (B) 41(19/20), 68.2 within 1 week	(A) Electric AT (1 time/day for 2 weeks), plus B NR	(B) Drug therapy: NR	1) NIHSS 2) Clinical effectiveness (NIHSS)	1) MD, −2.63 [−4.56, −0.70], P = 0.029 2) MD, 1.52 [1.09, 2.10], NR	SP6, SP9, LR3, ST34, SP10, ST36, LI4, LI10, TE5, GV26, GV20.	NR NR
Yang (2023)41	(A) 48(25/23), 61.7 (B) 48(29/19), 62.1 within 6 h	(A) AT (1 time/day for 3 weeks), plus B NR	(B) Ozone (25 ug/ml reach 40 ug/ml within 4 days), Clopidogrel (NR), Aspirin (NR)	1) NIHSS 2) BI	1) MD, −0.45 [−0.50, −0.40], P = 0.000 2) MD, 2.94 [2.36, 3.53], P = 0.000	PC6, GV26, SP6, 1 cun below HT1, BL40, LU5.	Headache (A:1, B:2), Nausea and vomiting (A:1, B:3), Cerebral hemorrhage (A:0, B:1), Others (A:1, B:3) Yes
Zhan (2023)15	(A) 51(24/27), 60.0 (B) 51(25/26), 59.0 Within 3 h	(A) AT (1 time/day for 2 weeks), plus B NR	(B) Rehabilitation training + rt-PA (0.9 mg kg-1, maximum dose: 90 mg)	1) NIHSS 2) MBI 3) mRS 4) Clinical effectiveness (NIHSS)	1) MD, −2.08 [−4.13, −0.03], P < 0.05 2) MD, 0.63 [0.23, 1.03], P < 0.01 3) MD, −0.37 [−0.71, −0.03], P < 0.05 4) MD, 1.25 [1.02, 1.53], P < 0.05	GV26, CV12, CV6, PC6, ST25, LI4, ST3, GB34, SP6, LR3.	NR Yes
Zhang (2010)42	(A) 40(26/14), 65.9 (B) 40(24/16), 69.2 within 1 week	(A) Warming-reinfrocing AT (1 time/day for 3 weeks), plus B NR	(B) Rehabilitation training + Drug therapy: NR	1) FMA-upper limbs 2) FMA-lower limbs 3) BI 4) CSS 5) Clinical effectiveness (CSS)	1) MD, 5.32 [0.68, 9.96], P < 0.01 2) MD, 3.33 [0.70, 5.96], P < 0.01 3) MD, 0.59 [0.14, 1.04], P < 0.01 4) MD, −3.50 [−6.22, −0.78], P < 0.01 5) MD, 1.30 [1.01, 1.66], P < 0.05	LI15, LI11, LI4, GB34, SP9, ST36, SP6.	NR Yes
Zhang (2015)13	(A) 427(242/185), 64.1 (B) 435(239/196), 64.6 3 to 10 days	(A) AT (5 times/week for 3 weeks), plus B 6 months	(B) Rehabilitation training + Drug therapy: NR	1) BI < 60	1) MD, 0.81 [0.62, 1.04], NR	GV26, PC6, SP6, GV20, ST36, ST40, LR3, LU5, GB20, CV6.	Faint (A:2, B:0), Dizziness (A:1, B:0), Pain (moderate or severe) (A:38, B:0), Recurrent stroke (A:7, B:7), Pneumonia and urinary tract infection (A:21, B:28), Others (A:5, B:4) Yes
Bi (2016)43	(A) 24(16/8), 62.3 (B) 24(17/7), 62.4 1 to 7 days	(A) Eye AT (1 time/day for 2 weeks), plus B NR	(B) Sodium ozagrel (160 mg/d), Citicoline for injection (0.5 mg/d)	1) Clinical effectiveness (CSS)	1) MD, 1.10 [0.89, 1.36], P < 0.05	Upper focal area, Lower focal area, Spleen area, Stomach area, Liver area	NR NR
Gao (2012)44	(A) 45(NR), NR (B) 16(NR), NR 24 h	(A) Contra-lateral AT (1 time/day for 1 month), plus B NR	(B) Drug therapy: NR	1) FMA	1) MD, 24.50 [22.02, 26.98], P < 0.01	LU5, LI4, LI11, SJ5, TE14, ST32, ST36, ST40, BL40, GB30	NR NR
Jiang (2019)45	(A) 20(NR), NR (B) 20(NR), NR 3 to 14 days	(A) AT (1 time/day for 4 weeks), plus B NR	(B) Rehabilitation training + Drug therapy: NR	1) BI	1) MD, 1.79 [1.04, 2.53], P < 0.05	GV26, PC6, SP6, HT1, BL40, LI11, LI4, ST32, ST34, ST36, ST40, LR3	NR NR
Li (2005)46	(A) 30(18/12), 65.5 (B) 30(19/11), 66.0 within 3 days	(A) scalp AT (2 times/day for 10 days), plus B NR	(B) Drug therapy: NR	1) FMA-upper limbs 2) FMA-lower limbs	1) MD, 23.06 [21.31, 24.81], P < 0.01 2) MD, 10.37 [6.94, 13.80], P < 0.01	Nie's three needles, GV20 through to GB7, Head's double needles	NR NR
Li (2011)47	(A) 44(24/20), 65.3 (B) 44(23/21), 66.4 6 to 48 h	(A) AT (NR for 90 days), plus B NR	(B) Drug therapy: NR	1) NIHSS 2) BI	1) MD, −0.82 [−1.75, 0.11], P < 0.01 2) MD, 1.36 [0.89, 1.82], P < 0.01	GB20, GV20, EX-HN1, EX-HN5, GV26, LI15, LI11, LI10, TE5, HT5, LI4	NR NR
Li (2020)48	(A) 45(25/20), 60.2 (B) 45(27/18), 60.2 2 to 7 days	(A) AT (1 time/day for 2 weeks), plus B NR	(B) Rehabilitation training + Drug therapy: NR	1) NIHSS	1) MD, −5.89 [−7.06, −4.27], P = 0.001	EX-HN1, GV26, GV20, LI15, EX-UE, LI13, LI4, LI10, TE5, LI11, EX-LE10, ST32, GB39, GB40, ST33, GB34, SP6, ST36, BL60, ST40	NR NR
Liu (2014b)49	(A) 40(22/18), 61.1 (B) 40(22/18), 63.7 within 72 h	(A) AT (1 time/day for 20 days), plus B NR	(B) Aspirin (100 mg/d), Xueshuantong Injection (500 mg/d)	1) Clinical effectiveness (CSS)	1) MD, 1.16 [0.95, 1.41], P < 0.01	PC6, SP6, GV26	NR Yes
Liu (2014)50	(A) 18(14/4), 65.6 (B) 20(10/10), 68.1 within 72 h	(A) AT (5 times/week for 2 weeks), plus B 3 months	(B) Aspirin (100 mg/d)	1) NIHSS 2) FMA 3) BI, 4) mRS	1) MD, −0.06 [−0.41, 0.29], P = 0.93 2) MD, 9.03 [4.75, 13.31], P = 0.33 3) MD, 0.51 [−0.21, 1.22], P = 0.72 4) MD, −0.14 [−0.35, 0.07], P = 0.75	LI4, TE5, LI11, ST36, SP6, LR3, Nie's three needles	NR NR
Liu (2016)51	(A) 18(14/4), 65.6 (B) 20(10/10), 68.1 within 72 h	(A) AT (1 time/day for 2 weeks), plus B 3 months	(B) Aspirin (100 mg/d)	1) NIHSS 2) FMA, 3) BI 4) mRS	1) MD, −0.06 [−1.49, 1.37], P > 0.05 2) MD, 9.03 [−8.61, 26.67], P < 0.05 3) MD, 0.13 [−0.58, 0.83], P > 0.05 4) MD, −0.14 [−0.99, 0.71], P > 0.05	LI11, TE5, LI4, ST36, SP6, LR3, EX-HN1, Nie's three needles	No serious AEs Yes
Liu (2021)52	(A) 20(12/8), 61.7 (B) 20(10/10), 60.8 2 to 14 days	(A) AT (NR for 6 weeks), plus B NR	(B) Sham acupuncture + Rehabilitation training + Drug therapy: NR	1) FMA-lower limbs	1) MD, 6.60 [4.97, 8.23], P < 0.01	PC6, GV26, SP6, GB34, BL40, GB40	Dizziness (A:1, B:0) NR
Rao (2006)53	(A) 20(12/8), 67.0 (B) 20(13/7), 67.3 3 to 10 days	(A) AT (5 times/week for 4 weeks), plus B 6 months	(B) Drug therapy: NR	1) BI 2) mRS	1) MD, 0.15 [−0.48, 0.77], P > 0.05 2) MD, 0.08 [−0.68, 0.84], P > 0.05	PC6, SP6, GV26	No serious AEs NR
Wang (2021)54	(A) 30(15/15), 64.5 (B) 30(16/14), 65.3 1 h to 14 days	(A) AT (NR for 2weeks), plus B NR	(B) Drug therapy: NR	1) NIHSS 2) BI 3) mRS 4) Clinical effectiveness (NIHSS)	1) MD, −0.82 [−1.82, 0.18], P < 0.05 2) MD, 1.40 [0.83, 1.96], P < 0.05 3) MD, −0.83 [−1.17, −0.49], P < 0.05 4) MD, 1.32 [1.05, 1.65], P < 0.05	PC6, SP6, GV26	No acupuncture-related AEs NR
Wang (2023)55	(A) 32(16/16), 59.1 (B) 32(18/14), 58.9 within 2 weeks	(A) AT (1 time/day for 2 weeks), plus B NR	(B) Drug therapy: NR	1) NIHSS 2) mRS 3) Clinical effectiveness (NIHSS)	1) MD, −4.87 [−5.69, −4.05], P = 0.000 2) MD, −1.15 [−1.40, −0.90], P = 0.000 3) MD, 1.29 [1.05, 1.59], P = 0.031	PC6, SP6, GV26	NR NR
Xing (2015)56	(A) 25(13/12), 63.5 (B) 25(11/14), 61.8 within 1 weeks	(A) electric AT (6 times/week for 2 weeks), plus B NR	(B) Drug therapy: NR	1) BI	1) MD, 0.72 [0.15, 1.29], P = 0.013	GV26, GV20, GB20, PC6, LI11, LI4, LI15, LI10, SP6, ST36, LR3, ST32, ST31	No acupuncture-related AEs NR
Wu (2009)57	(A) 29(16/13), 56.7 (B) 29(17/12), 58.5 within 2 weeks	(A) electric AT (6 times/week for 2 weeks), plus B NR	(B) Rehabilitation training + Aspirin (0.1 g/d), Cinepazide Maleate Injection (160mg/d)	1) NIHSS 2) FMA-lower limbs 3) BI	1) MD, −0.37 [−1.15, 0.41], P > 0.05 2) MD, 4.58 [2.82, 6.34], P < 0.01 3) MD, 0.15 [−0.36, 0.67], P > 0.05	LI15, LI11, LI10, LI4, ST32, ST34, ST36, ST40, ST41, ST31	NR NR

Notes: AT: acupuncture treatment, BI: Barthel index, CSS: Chinese Stroke Scale, FMA: The fugl-meyer assessment, MBI: The modified Barthel index, mRS: The modified Rankin scale, NIHSS: The National Institutes of Health Stroke Scale, NR: Not report. The location of acupoints adheres to the WHO standard of acupuncture point locations.

2.5. Quality assessment

We used the Cochrane Collaboration Network risk of bias assessment tool to evaluate the quality of the included studies.⁵⁸ Two researchers (SS, MJ) independently evaluated the quality of the studies across 7 dimensions: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other biases. A judgment of low, high or unclear risk of bias was made for each dimension. Any disagreements were resolved by consulting our superiors (YT). The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) framework was employed for a comprehensive assessment of the certainty of the evidence in the meta-analysis.⁵⁹

2.6. Statistical analysis

Review Manager software version 5.3 was used for data analyses. For dichotomous variables, we calculated the risk ratio (RR) and 95 % confidence intervals (CIs) to aggregate the outcomes. Continuous variables were pooled using the mean difference (MD) or standard mean difference (SMD) and 95 % CIs. Statistical heterogeneity was assessed using I² statistics and P values. We assigned adjectives of low, moderate, and high to I² values of 25 %, 50 %, and 75 %, respectively.⁶⁰ Due to the varying methods and populations across the studies, we believe that the heterogeneity among them was substantial. Therefore, we used random effects models to pool the outcomes. A sensitivity analysis was conducted by systematically omitted each RCT in succession until a state of low or moderate heterogeneity was attained. The meta-analysis results were presented in a forest plot. A funnel plot combined with Egger’s test was employed to assess publication bias. An asymmetric funnel plot visually suggests publication bias. Additionally, Egger's test was conducted using Stata 17.0, with a P value <0.05 statistically indicating the presence of publication bias.⁶¹ AEs related to acupuncture in each study were diligently recorded.

2.7. Subgroup analysis

The treatment course duration referred to a specific period during which the condition was effectively controlled after continuous acupuncture treatment. The time to acupuncture commencement was the temporal juncture at which the commencement of acupuncture occurred. Various treatment course durations and acupuncture commencement times were reported in the included studies, all of which showed improvements with acupuncture. However, ambiguity persists regarding the precise commencement of acupuncture and the duration required for the treatment to achieve substantial efficacy.62, 63, 64 Thus, the time to acupuncture commencement were categorized based on enrollment time (within 24 h, within 72 h, within 1 week, or within 2 weeks). Additionally, treatment courses duration was segmented into four intervals for analysis (10 days, 2 weeks, 3 weeks, and 4 weeks). Among the included studies, some used medicine therapy and rehabilitation together as baseline modalities, whereas others only used medicine therapy as a baseline modality. To elucidate the effects arising from the juxtaposition of various therapeutic measures, we performed additional subgroup analyses.

3. Results

3.1. Study characteristics

We initially retrieved 3874 potentially eligible studies. A total of 1430 duplicates were removed. After screening the titles and abstracts, 2294 studies were excluded, and 150 studies remained for full-text screening. Ultimately, 30 studies were included, and an additional study³⁸ was obtained from alternative sources. Therefore, 31 RCTs were included in this meta-analysis, with a total of 3604 patients with AIS. Acupuncture combined with CTs was typically regarded as an intervention in the experimental group, whereas sham acupuncture combined with CTs or CTs alone was regarded as an intervention in the control group. Notably, Shen 2012¹⁴ and Liu 2021⁵² implemented a sham acupuncture procedure near acupoints. Among the included studies, 9 RCTs^13,15,33, 39, 42, 45, 48, 52, 57 used routine medicine therapy and rehabilitation as the baseline treatment for both groups, and routine medicine therapy alone was used for the remaining RCTs. All studies reported that the baseline data could be comparable. Details were provided in Fig. 1 and Table 1. Excluded studies and reasons were shown in Supplement 1, Table S3.

Fig. 1.

Flow chart.

3.2. Quality assessment

We assessed the bias risk and quality of the included studies using the Cochrane risk of bias tool. Although all studies reported randomization, 2 studies^44,55 generated random numbers based on lottery or visit orders. And 19 of the included studies did not report the allocation concealment. Due to the intricacies of needle insertion, blinding for acupuncturists was challenging. One study⁵⁶ mentioned that the assessments were conducted by doctors, which could lead to detection bias. Fifteen studies reported complete outcome data and performed ITT analyses. We believe that reporting bias was absent in all publications. Other bias was low risk in 11 studies13, 14, 15, 34, 36, 37, 39, 41, 42, 49, 51 because they have reported the funding and stated no conflict of interest. In summary, the overall quality of the included studies was generally low primarily due to inadequate blinding methods and incomplete outcome data. Fig. 2 shows the details.

Fig. 2.

Risk of bias.

3.3. Primary outcome

3.3.1. Neurological function: national institutes of health stroke scale (NIHSS)

Sixteen studies reported posttreatment NIHSS scores, encompassing a total of 1658 subjects, with 831 patients in the experimental group. A significant difference was noted between the two groups in terms of the NIHSS score after treatment (MD: −1.70, 95 % CI [−2.27, −1.14], P < 0.00001, very low-certainty evidence) and high heterogeneity was detected (I² = 96 %) (Supplement 2A). To assess the sources of heterogeneity, we carried out a sensitivity analysis by removing RCTs in turn. We observed moderate heterogeneity (I² = 47 %) after eliminating 6 RCTs (Chu 2017, Cui 2022, Shen 2012, Li 2020, Wang 2014, Wang 2023) (Supplement 3a). This finding suggested that the studies of Chu 2017, Cui 2022, Shen 2012, Li 2020, Wang 2014, Wang 2023 might be potential sources of heterogeneity because of the patients or methodology of trial.

3.3.2. Neurological function: clinical effectiveness (NIHSS)

Seven studies recorded the clinical efficacy using the NIHSS score as a metric. Each group comprised 312 patients. After treatment, significant differences in clinical effectiveness were observed between the two groups (RR: 1.21, 95 % CI [1.12, 1.31], P < 0.00001, low-certainty evidence) (Supplement 2B). Moderate heterogeneity was observed (I² = 26 %).

3.4. Secondary outcomes

3.4.1. Neurological function: chinese stroke scale (CSS)

Three studies reported CSS scores after treatment. The experimental group consisted of 244 participants, whereas the control group comprised 243 participants. Compared with CTs, acupuncture combined with CTs improved neurological damage (MD: −4.33, 95 % CI [−5.67, −2.98], P < 0.00001; low-certainty evidence) (Supplement 2C). Moderate heterogeneity was detected (I² = 46 %).

3.4.2. Neurological function: clinical effectiveness (CSS)

Five studies delineated the clinical efficacy predicated on CSS scores. The respective group sizes were 195 patients and 194 patients. Notably, there were significant differences between the experimental group and the control group (RR: 1.26, 95 % CI [1.13, 1.41], P < 0.0001; low-certainty evidence) (Supplement 2D). Furthermore, low heterogeneity was observed (I² = 18 %).

3.4.3. Activity of daily living (ADL): (Modified) barthel index

Twelve studies reported BI scores, whereas 4 studies focused on MBI scores. The collective number of patients included in the analysis was 1340, with 672 patients allocated to the experimental group. When the BI and MBI were combined, the results revealed that acupuncture, as a complementary treatment, significantly enhanced the ability of patients with AIS to perform ADL compared to that of patients in the control group (SMD: 0.98, 95 % CI [0.64, 1.31], P < 0.00001; very low-certainty evidence) (Supplement 2E). However, substantial heterogeneity was observed (I² = 87 %). A sensitivity analysis was performed to discern the sources of this heterogeneity (Supplement 3b), indicating that Li 2016a, Li 2011, Rao 2006, Wang 2021 and Yang 2023 may potentially contribute to this heterogeneity. The nuances of trial design and the severity of the condition may contribute to sources of heterogeneity.

3.4.4. Disability: modified rankin scale

Eight studies reported mRS, among them, Yang 2023 and Xing 2015 reported the number of mRS ≤2. Six other studies would be pooled the outcomes, encompassing a total of 328 subjects, with 165 patients in the experimental group. The pooled result suggested that there were significant differences between two groups (MD: −0.48, 95 % CI [−0.91, −0.05], P = 0.03; very low-certainty evidence) (Supplement 2F) and a high heterogeneity (I² = 88 %). A sensitivity analysis was performed to identify the sources of this heterogeneity (Supplement 3c), indicating that Wang 2021 and Wang 2023 may potentially contributed to the heterogeneity due to the difference of trial design and the severity of the condition.

3.4.5. Motor function: FUGL-Meyer assessment

Seven studies assessed FMA scores, encompassing 270 patients in the experimental group and 242 patients in the control group. Compared with CTs alone, acupuncture with CTs more strongly alleviated motor symptoms in patients with AIS (MD: 16.46, 95 % CI [12.56, 20.35], P < 0.00001, very low-certainty evidence) (Supplement 2 G). Unfortunately, high heterogeneity was observed (I² = 87 %). After a sensitivity analysis, Gao 2012 and Liu 2014 might be a potential source of heterogeneity due to the nature of the patients and the distinction of study design (Supplement 3d).

3.4.6. Prognosis

Three studies reported the prognosis after six months based on the BI scores during the follow-up period. In this context, an adverse prognosis is defined as a dependency on others for daily living, deterioration, or death. There were 661 patients in the experimental group and 694 in the control group. Compared to the control group, acupuncture had favorable effects on the prognosis of patients with AIS (RR: 0.38, 95 % CI [0.16, 0.89], P = 0.03, very low-certainty evidence) (Supplement 2H). The high heterogeneity (I² = 92 %) may have been due to the discrimination of the prognosis standard. After excluding Zhang 2015, the heterogeneity decreased (I² = 12 %) (Supplement 3e).

3.5. Subgroup analysis

3.5.1. Effect of different acupuncture commencement time on the NIHSS scores

We divided the studies into 4 categories (within 24 h, within 72 h, within 1 week and within 2 weeks) for the meta-analysis. The analysis revealed statistically significant differences within 24 h (MD: −1.18, 95 % CI [−2.21, −0.16], P = 0.02) and within 2 weeks (MD: −2.06, 95 % CI [−3.04, −1.07], P < 0.0001). However, no statistically significant difference was observed within 72 h (MD: −0.51, 95 % CI [−1.19, 0.17], P = 0.14) and within 1 week (MD: −3.06, 95 % CI [−6.66, 0.54], P = 0.10) (Supplement 3f).

3.5.2. Effect of different acupuncture treatment course duration on NIHSS scores

We divided the duration of treatment into four parts (10 days, 2 weeks, 3 weeks, and 4 weeks) for the meta-analysis. A significant difference was found between the two groups at 2 weeks (MD: −1.95, 95 % CI [−2.82, −1.08], p < 0.0001) and 4 weeks (MD: −2.20, 95 % CI [−2.60, −1.80], P < 0.00001) (Supplement 3 g).

3.5.3. Effect of different conventional treatments on NIHSS scores

The 16 studies were categorized into two groups based on the inclusion or exclusion of rehabilitation training. Three studies juxtaposed pharmacotherapy and rehabilitation training as the baseline treatments, whereas the remaining studies exclusively employed drug therapy as their baseline treatment. Based on the outcomes of the meta-analysis, difference was observed between the studies subjected to drug therapy coupled with rehabilitation as a baseline treatment (MD: −2.62, 95 % CI [−4.69, −0.55], P = 0.01). A statistically significant difference was also observed between the studies in which only drug therapy was used as a baseline treatment (MD: −1.44, 95 % CI [−2.02, −0.85], P < 0.00001) (Supplement 3 h).

3.5.4. FMA scores of the upper or lower limbs

Three studies documented FMA scores concerning upper limbs functionalities, with 100 patients and 100 patients allocated to the two groups, respectively. Difference was noted in the FMA scores of the upper limbs (MD: 13.10, 95 % CI [2.04, 24.15], P = 0.02) (Supplement 3i). five studies reported FMA scores concerning lower limbs functionalities, with 149 patients and 149 patients allocated to the two groups, respectively. Significant difference for the FMA scores of the lower limbs (MD: 6.74, 95 % CI [4.41, 9.06], P < 0.0001) were noted (Supplement 3j).

3.6. Adverse events (AEs)

Only 12 studies described AEs related to acupuncture. There were no severe AEs related to acupuncture. AEs were shown in Table 1.

3.7. Publication bias

The funnel plot suggested the presence of publication bias concerning NIHSS. However, Egger's test showed no evidence of publication bias (P = 0.397). Consequently, we do not believe that publication bias is present (Fig. 3).

Fig. 3.

Publication bias for National Institutes of Health Strock scale.

3.8. Certainty of evidence

The results of the GRADE evaluation were presented in Supplement 1 Table S4. The certainty of the evidence was very low or low. The main reasons that the certainty of the evidence downgraded were the lack of blinding, unclear allocation concealment, small total sample size, high heterogeneity, and potential publication bias.

4. Discussion

4.1. Main findings

In this review and meta-analysis, we focused mainly on the effectiveness and safety of acupuncture for the treatment of AIS. We found that using acupuncture as a complementary therapy was beneficial for improving neurological and motor dysfunctions and enhancing the ability to live independently. Thus, acupuncture promotes the clinical effectiveness of AIS treatments, ultimately leading to a more favorable prognosis than CTs alone. However, the evidence examined herein was of very low or low certainty. Several intriguing subgroup effects were identified in our subgroup analysis. Augmenting the course of acupuncture treatment may result in a corresponding increase in therapeutic effect. The temporal windows for acupuncture might not be associated with the effect of acupuncture for AIS. Adding rehabilitation therapy to acupuncture and conventional treatment might yield better therapeutic outcomes. Acupuncture can improve overall FMA scores in AIS patients, meanwhile it could improve both upper-limb and lower-limb sub-scores of the FMA. These findings were derived from the synthesis of several studies, and their generalizability was limited. Further validation of these conclusions requires additional high-quality randomized controlled trials. There were no serious AEs related to acupuncture. The most prevalent AEs included pain, bleeding, headaches, dry vomiting, nausea, and fainting. These mild symptoms often exhibit a self-resolving nature without significantly affecting treatment adherence. Some severe AEs, such as gingival bleeding and fundus retinal hemorrhage, may be connected with thrombolytic therapy. Therefore, acupuncture was a relatively safe and effective therapy for treating AIS.

4.2. Explanation of the results

Acupuncture treatment has been widely used to promote the recovery of both sensory and motor functions after stroke in China.⁶⁵ As a traditional medical therapy, the efficacy of acupuncture necessitates a sufficiently extended course of treatment.⁶⁶ The subgroup analysis indicated that a favorable duration for acupuncture treatment may be 4 weeks. In a study⁶⁷ published in 2023, the occurrence of early neurologic improvement compared with baseline was defined as a decrease in the NIHSS score greater than or equal to 2. Hence, upon attaining a substantial treatment course (4 weeks, MD = −2.2; see Supplement 3 g), the incorporation of acupuncture therapy has the potential to enhance the efficacy of drug. However, due to the lack of contrast in treatment duration (over 4 weeks), this study could not determine whether a longer treatment duration would provide additional benefits. It remained uncertain whether the optimal treatment duration was indeed 4 weeks. Our findings indicated that acupuncture treatment could yield favorable effects for AIS, as long as it was administered during the acute phase (within 2 weeks). This finding suggested that acupuncture therapy for AIS patients may be not limited by temporal windows. Our research revealed that acupuncture combined with medical therapy and rehabilitation therapy could better improve symptoms in patients with AIS. This finding was consistent with the results of many other studies.^68,69 However, weak patient recovery might be inhibited by physical capacity.⁷⁰ Choosing appropriate and proper treatment may be more conducive to the recovery of stroke patients. Acupuncture enhanced the motor function of patients with AIS, significant improvement was observed in the function of upper limbs and lower limbs. However, several studies^69,71 showed that it may be difficulty for treating fine motor disorders in the extremities with acupuncture. More work was necessary to demonstrate the effects of acupuncture on fine movements of limbs.

4.3. Strengths and limitations of this study

The strengths of this meta-analysis and review were as follows: a comprehensive search for RCTs that found 31 RCTs investigating the effectiveness and safety of acupuncture in treating AIS; subgroup analyses to explore the optimal temporal arrangement of acupuncture; scrutinizing the synergistic impact of these interventions; and the use of the GRADE methodology for the evaluation of the quality of evidence.

However, there were several limitations in this review. First, all included trials were conducted in China, resulting in a lack of sample diversity. Second, the absence of extended follow-up data made it impossible to evaluate the long-term effects of acupuncture. A follow-up period of 6 months was not sufficient to reflect the enduring impact of acupuncture. Third, the methodological quality of the included studies was poor, resulting in limited strength of the evidence. Although all included RCTs noted the importance of random consequence generation, allocation concealment was not performed well. With the development of technology, the central random system can ensure the fairness of randomization.⁷² The implementation of the blinding method faced challenges, particularly in achieving double-blind conditions for RCTs of acupuncture, due to the unique nature of acupuncture. Nevertheless, blinding the subjects remains feasible. A skillfully designed cover or barrier placed on the skin can effectively prevent subjects from discerning that a needle has penetrated the skin during an acupuncture procedure.⁷³ Meanwhile, blinding evaluators and analysts was indispensable to objectivity of the assessment. Fourth, the studies included in this meta-analysis lacked proper division or stratification of the dataset. To confirm the counts of the ITT population and per protocol (PP) population, it is necessary to describe the number of inclusions, exclusions, dropouts, and eliminations using a CONSORT flowchart.⁷⁴ Researchers assess the average effect of a group of patients only through ITT analysis.⁷⁵ Fifth, There was publication bias because studies with negative results were less likely to be published. The clinical trial registration platform is a crucial channel for obtaining unpublished data from RCTs. Although numerous registration platforms exist worldwide, they lack interconnectedness. The search strategies used across these platforms remain unshared and incomplete.⁷⁶ Each platform updates its information independently and termly, resulting in delayed data acquisition.⁷⁷ Therefore, it is imperative to encourage clinical trial registration platforms to enhance their retrieval functions and impel trial registrants to promptly submit statistical findings, regardless of positivity or negativity, at the end of a trial. Sixth, the study was not strictly adhering to the pre-registered protocol. The initial age range was established at 18 to 80 years; however, numerous RCTs included participants who exceeded this range. Consequently, the age limit was removed. Meanwhile, to enhance the paper, secondary outcomes and subgroup analyses were incorporated while maintaining the primary outcomes unchanged.

4.4. Implications for clinicians and researchers

Acupuncture is a viable modality for the clinical management of AIS. A sustained treatment over a duration of 4 weeks during the acute phase (within 2 weeks of onset) may manifest considerable therapeutic effect. Due to the low or very low certainty of evidence, researchers have been compelled to formulate rigorous trial protocols for more robust investigations.

5. Conclusions

Compared with conventional treatments, acupuncture combined with conventional treatments is effective and safe for treating AIS. However, additional high-quality, multicenter, large-sample, well-designed RCTs need to be conducted to prove the effects of acupuncture.

CRediT authorship contribution statement

Shanshan Huang: Writing – original draft, Writing – review & editing. Yuting Wang: Writing – original draft, Writing – review & editing. Yifan Wu: Methodology. Puchen Huang: Methodology. Yu Dong: Methodology. Qiao Zhang: Methodology. Mengjiao Zhao: Methodology. Jingchun Zeng: Methodology. Liming Lu: Conceptualization, Writing – review & editing. Xinxia Liu: Conceptualization, Writing – review & editing.