Protocol and registration
This protocol follows the preferred reporting items for systematic reviews and meta-analysis protocols (PRISMA-P) guidelines (Additional file 1)  and is adapted from the structure provided in the Systematic Review Protocol for Animal Intervention Studies . This protocol was registered on OSF (Registration DOI: 10.17605/OSF.IO/FNM37).
Types of studies
This systematic review will include both randomized and non-randomized controlled studies evaluating the effect of acupuncture on animal models of PCOS. There will be no restriction on language, date, or publication status.
Types of animal models
This systematic review will examine all animal models of PCOS regardless of species or size of the animal.
Types of comparators
The comparison group will include PCOS animals induced by the same method as the intervention group but have not undergone intervention.
Types of intervention
We will include studies using both traditional acupuncture and contemporary acupuncture (in which needles were not inserted in classical meridian points). We will adopt hand stimulation, electrical stimulation, or warming needle with moxibustion. Acupuncture without needling will be excluded such as acupressure, acupoint injection, tap-pricking, and cupping.
We will exclude case reports, cohort studies, congress abstracts, letters to editors, in vitro experiments, and human studies. Studies comparing animal models to healthy controls will also be excluded.
Types of outcome measures
Preliminary searches have shown that parameters concerning hormone and insulin resistance have been reported in experimental studies of acupuncture in treating PCOS animal models. The following outcome measures will be used:
- Primary outcome: homeostatic model assessment- insulin resistance (HOMA-IR: (fasting insulin (μU/mL) ×fasting glucose (mmol/L))/22.5).
- Secondary outcome: testosterone (T), LH, LH/follicle stimulating hormone (FSH) ratio, AMH, body mass index (BMI: calculated using weight (kg)/height (m2)).
Search and study identification
The following sources will be searched: PubMed, Web of Science, China National Knowledge Infrastructure (CNKI), Chinese Science and Technology Periodical Database (VIP). The main terms “Polycystic ovary syndrome”, “Acupuncture”, and “Animal Experimentation”, indexed in the MeSH system, will be combined. We will search the terms and their synonyms separately first and then combine and cross the search terms. Details of the Pubmed search strategy present in Additional file 2.
Two independent reviewers (JJG and YYL) will screen titles and abstracts for eligibility. Disagreements between reviewers will be resolved by third review (YL). Full texts will be obtained and evaluated by the same reviewers using a pre-designed form with inclusion and exclusion criteria. For duplicated publication, we will use the paper with more complete data. The study selection progress will be presented using PRISMA flow diagram (Additional file 2).
Two reviewers (YL and JJG) will extract data independently using a pre-approved form. The following information will be recorded: study design, characteristics of included, characteristics of animal model, details of the intervention, and outcome measures. If the data is not available in the text but mentioned in the manuscript, the reviewer will contact the authors via email.
Risk of bias assessment
Risk of bias will be assessed by the SYRCLE’s tool for assessing risk of bias (SYRCLE ROB), which includes domains for selection bias (sequence generation, baseline characteristics, allocation concealment), performance bias (random housing and blinding), detection bias (random outcome assessment and blinding), attrition bias, reporting bias, and other biases.
Data will be presented in summary tables to describe the study population and intervention. We will perform a meta-analysis using random effects models. Mean differences (MD) among groups with their respective 95% confidence intervals and p values will be presented in forest plots. Data will be analyzed using Review Manager (RevMan) 5.4. The presence of heterogeneity will be evaluated by I2 and chi-square statistical analyses. Sources of heterogeneity will be explored by sensitivity analysis via systematic removal of individual trials. We will construct summary tables and provide a narrative synthesis if quantitative synthesis is not appropriate. Funnel graphics will be performed to evaluate publication bias if there are enough studies included.
We will use the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach  to assess the quality of evidence generated from each outcome. The GRADE contains five domains including risk of bias, inconsistency (heterogeneity), indirectness, imprecision, and publication bias. Two reviewers will assess the quality independently and the quality will be graded as “high”, “moderate”, “low”, or “very low.