This protocol conforms to the Preferred Reporting Items for Systematic Reviews and Meta-Anadlyses statement and has been registered in the International Prospective Register of Systematic Reviews (no. CRD42024502006).

Several electronic databases will be used to conducte a literature search, including PubMed, Embase, Web of Science Core Collection and Cochrane Central Registry of Controlled Trials. A systematic search will be conducted from the database inception to January 2024. Besides, a manual review of reference lists for included studies will also be conducted to track potentially relevant studies. The search phrase included MeSH terms and keywords related to stroke, acupuncture, motor impairment and SRs/MAs. The draft searching keywords are listed in table 1. The language will be restricted to English.

Search strategy for PubMed

The literature comparing outcomes between acupuncture and non-acupuncture for post-stroke motor impairment will be included. Study patients who were diagnosed with post-stroke motor impairment will undergo acupuncture or conventional treatment (don't include acupuncture therapy). Post-stroke motor impairment is defined as a certain degree of dysfunction in limb movement compared with healthy individuals after ischaemic stroke, encompassing motor deficits caused by flaccid hemiplegias or dystonic disorders. The outcomes will include functional assessment and adverse events. The exclusion criteria will be the following: (1) non-human studies, (2) non-comparative studies, (3) studies that do not include post-stroke motor impairment, (4) studies without available data can be extracted and (5) non-original studies (letters reviews and editorials).

EndNote V.X9 software will be used to store and manage the retrieved studies, and duplicate publications will be excluded. Two independent reviewers will conduct a preliminary scan of the title and abstract of the retrieved articles for potentially relevant studies. Disputes arising during the review process will be resolved by a neutral third-party reviewer. The flow diagram of the study selection process is presented in figure 1.

Schematic diagram of the screening process. Notes: MAs, meta-analyses; SRs, systematic reviews; PSMI, post-stroke motor impairment.

The data extraction will be performed by two independent reviewers according to a predefined data collection form. The data extraction process will encompass the following considerations: (1) details of the SR/MA publication (authors’ name, publication dates, number of included studies and total participants), (2) participants’ characteristics (age, sex, type of stroke, and days from stroke onset), (3) details of interventions in the treatment and control groups (treatment period, type of intervention), (4) description of outcome measures (primary outcome measurement, secondary outcome measurement), (5) acupuncture-related adverse events, (6) results of risk-of-bias assessment, and (7) summary effect sizes (study-specific standardised mean difference (SMD), 95% CIs).

Two reviewers will cross check the extraction results of the extracted studies. Any discrepancies will be resolved by a third reviewer.

The primary outcomes will be defined as the average overall change in motor-related functional assessment. The included SRs/MAs will be required to report at least one of the following outcomes: (1) motor function-related assessment scales, including the Fugl–Meyer assessment (FMA),19 Modified Ashworth scale (MAS)20 or Berg balance scale;21 (2) measurement of the range of motion, including goniometer; and (3) muscle strength, including manual muscle testing. The secondary outcomes will consider activities of daily life (ADLs), pain intensity assessment, emotional state function and adverse events. The impairment of ADLs is one of the impacts of stroke, and the assessment of ADL can provide a comprehensive evaluation of the consequences of post-stroke motor impairment. Additionally, post-stroke limb pain and emotional disorders are common comorbidities in stroke patients, and attention to these outcomes can facilitate a multifaceted evaluation of the comprehensive effects of acupuncture.

Two independent reviewers will conduct back-to-back quality evaluation of the included studies. Any disagreement during the assessment procedure will be resolved through discussion or adjudication with a third reviewer.

The methodological quality of each included SR/MA will be assessed using the AMSTAR-2 tool.22 The AMSTAR-2 tool has been specifically developed for SRs encompassing randomised or non-randomised studies. AMSTAR-2 comprises 16 items, with each item offering four potential responses: ‘yes’ (definite completion), ‘no’ (definite non-completion), ‘can't report’ (uncertainty regarding completion) and ‘not applicable’. The quality of each item will subsequently be grouped into four distinct levels: critically low, low, medium and high.

Additionally, the Grading of Recommendations Assessment, Development and Evaluation (GRADE) system will be used to assess the evidentiary quality of each SRs/MAs,23 downgrading from five aspects: research limitations, inconsistency, indirectness, imprecision and suspected publication bias. According to the GRADE working group’s recommendations, we will grade the quality of evidence into four levels: high, medium, low and very low.

Statistical analysis will be performed with Review Manager V.5.4 software. The qualitative analysis of the summary effect size for each SR/MA will be conducted using the SMD and its corresponding 95% CI. If there were no SMDs expressed in the original studies, the results will be converted to SMDs using Review Manager V.5.4 software. If data are insufficient and cannot be converted, we will extract the data from the original randomised controlled trial and process it according to the instructions by the Cochrane Collaboration. Heterogeneity will be reported using the I2 statistic, where values ranging from 0% to 25% suggest no heterogeneity, 25% to 49% suggest low heterogeneity, 50% to 74% suggest moderate heterogeneity, and values above 75% suggest high heterogeneity.

The I2 statistic will be used for reporting heterogeneity: with 0% to 25% indicating no heterogeneity, 25% to 49% indicating low heterogeneity, 50% to 74% indicating moderate heterogeneity, and above 75% indicating high heterogeneity. Publication bias will be assessed through the funnel plots.

We will deliver the following subgroups analysis: (1) for the type of acupuncture therapies received by participants (eg, dry needling, electroacupuncture or other acupuncture therapy), (2) for the recovery of the upper and lower limb, and (3) for the duration of the intervention period.