The protocol of this meta-analysis has been registered in PROSPERO (International Prospective Register of Systematic Reviews) with the registration number CRD42024589362. This review will be conducted and reported in line with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 statement [18].

The following electronic databases will be searched: Cochrane Central Register of Controlled Trials (CENTRAL, Ovid), MEDLINE (Ovid, from 1946 to present), Embase (Ovid, from 1974 to present), Allied and Complementary Medicine Database (Ovid, from 1985 to present), China National Knowledge Infrastructure (from 1949 to present), Chinese BioMedical Literature Database (from 1949 to present), and Wanfang Database (from 1985 to present). Search terms include (traditional medicine OR oriental medicine OR medicinal plants OR phytotherapy OR Chinese herbal drugs OR plant extracts OR herbal medicine) and (insulin resistance OR insulin sensitivity OR glucose intolerance OR glucose tolerance OR impaired glucose OR impaired glycaemia OR metabolic syndrome OR prediabetic state). The search strategy will be developed and modified for different databases (Table S1 in Multimedia Appendix 1). We will apply the Cochrane sensitivity-maximizing randomized controlled trial filter with adaptations [24]. There will be no restriction on the language. Gray literature will be searched with Google. We will also use the ClinicalTrials.gov and World Health Organization International Clinical Trials Registry Platform (ICTRP) Search Portal to identify ongoing or unpublished trials. In order to gather missing information and details of ongoing trials, we will reach out to the authors for clarification. We will further review the reference lists of included studies and any relevant systematic reviews to identify additional references. The screening process of the searched studies will be conducted according to the PRISMA flow diagram (for checklist, see Multimedia Appendix 2) [25].

Literature retrieved citations will be managed by EndNote (version 20; Clarivate) software. Eligibility evaluation will be done by title and abstract reviews and when abstracts do not provide enough information, the full text of the paper will be retrieved for evaluation. This will be performed independently by 2 reviewers. Discrepancies will be resolved by consensus and arbitration by a third reviewer when necessary.

A data extraction form will be constructed and include general information (author, year, title, journal, language, and country), participant characteristics, and study design (sample size, age, gender, IR-related diseases, diagnosis criteria, group assignment, and blinding), treatment regimen (dosage, form, administration route, duration, and frequency), clinical outcomes, adverse events, and assessments of risk of bias. Before implementing it for all included studies, we will conduct a pilot test using at least 1 study from the review. A total of 2 review authors will independently extract study characteristics from the included studies.

The risk of bias (RoB) of the included studies will be adopted by 2 reviewers independently using the Cochrane revised tool for randomized trials (RoB 2) [24]. Each trial will be scored as either high, low risk, or some concerns for 5 items: randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, and selection of the reported result. When there are disagreements, a third reviewer will be consulted.

The overall certainty of evidence will be assessed by 2 assessors using Grading of Recommendations Assessment, Development, and Evaluation (GRADE), as outlined by the GRADE Working Group [26]. The 5 items will be investigated, including limitations in study design, inconsistency, imprecision, indirectness, and publication bias. The certainty of evidence will be summarized in 4 categories: high, moderate, low, and very low certainty.

We will perform statistical analysis using the RevMan (version 5.4; The Cochrane Collaboration) software. We will calculate the mean difference or risk ratio with 95% CI, separately, for continuous and dichotomous outcomes. Data heterogeneity will be assessed using the chi-square test and the I² statistic. A fixed-effects model will be used to analyze pooled effects when heterogeneity is not significant (P≥.10 or I²≤50%); a random-effects model will be used when heterogeneity is statistically significant (I²>50% or P<.10). We plan to conduct the following subgroup analyses: single herbal medicine versus placebo, single herbal medicine versus no treatment, and single herbal medicine combined with cointervention versus cointervention alone. Sensitivity analysis will be conducted by excluding trials with a high RoB in more than 2 domains or with a dropout rate exceeding 20%. Egger regression test and funnel plot (if more than 10 trials are included) will be performed to assess the publication bias of the included studies.

Missing or ambiguous data will be addressed by contacting the first or corresponding authors via email twice (at 2-wk intervals) to request additional information. In cases where data remain unavailable despite repeated correspondence attempts, the affected studies will be excluded from quantitative synthesis. For included studies with partial missing data, intention-to-treat analysis data will be used.

IR is a central factor in the pathogenesis and progression of metabolic disorders, including T2DM, obesity, and hypertension [27]. Furthermore, a strong association has been established between IR and an increased risk of developing cardiovascular diseases [28]. The intricate mechanisms underlying IR may involve complex interplay among inflammation, endoplasmic reticulum stress, oxidative stress, and mitochondrial dysfunction [29-32]. CHM has emerged as a potential therapeutic modality for enhancing insulin sensitivity [13,33]. Numerous CHMs have demonstrated therapeutic effects in improving insulin sensitivity in peripheral tissues, enhancing glucose-stimulated insulin secretion, promoting browning of white adipose tissue, and lowering glucose and lipid levels [34-37]. The growing use of CHM for improving insulin sensitivity underscores the urgent need for a thorough assessment of its effectiveness and safety.

This systematic review and meta-analysis will synthesize existing evidence to comprehensively evaluate the therapeutic efficacy and safety profiles of single CHMs for insulin IR. Subgroup analyses will compare single herbs against placebo or no treatment and assess efficacy variations across IR-associated conditions (eg, T2DM and obesity), providing important insights into context-specific therapeutic potential. In addition, critical appraisal of the study and reporting quality in included trials will identify methodological limitations in current research, informing recommendations for future trial design and transparency.

While the HEC remains the gold standard for quantifying IR due to its direct measurement of insulin-mediated glucose uptake [38-40], its clinical utility is limited by technical complexity and invasiveness [41]. Consequently, multiple validated markers, such as fasting-based (eg, HOMA-IR and QUICKI), dynamic tests (eg, OGTT and IVGTT), and composite indices (eg, TyG index and METS-IR), are widely adopted in clinical research [20-23], Among these, HOMA-IR is prioritized in observational and interventional studies due to its simplicity and reproducibility [38]. To ensure methodological inclusivity, this review will incorporate all IR-associated outcomes.

This represents the first systematic evaluation of all single herbs and herbal extracts for IR, addressing a critical gap left by prior reviews focused on multiherb formulas [42,43] or isolated compounds [44,45]. While single-herb analyses deviate from traditional multiherb clinical practice, they provide clarity in therapeutic efficacy essential for evidence-based integration into IR management protocols. However, heterogeneity in treatment details (eg, administration routes and treatment duration) and the potential inclusion of low-quality studies may limit the generalizability of findings.

This study may inform the potential of single CHMs as alternative therapies for IR, bridging traditional knowledge with clinical needs. Identified high-priority herbs should undergo rigorous preclinical validation (eg, pharmacokinetic studies, in vivo mechanistic models) followed by confirmatory clinical trials with rigorous design. Concurrently, the quality assessment framework established here may guide future trial designs to minimize bias and enhance reproducibility.

This systematic review and meta-analysis will rigorously synthesize existing evidence to clarify the efficacy and safety of single CHMs in ameliorating IR, offering critical insights for their integration into evidence-based therapies for metabolic disorders. By focusing on single-herb therapy, the findings may promote the application of CHM for IR, bridge the gap between traditional applications and evidence-based practice, and ultimately optimize the role of CHM in integrative metabolic health management.

Results will be disseminated through peer-reviewed journals and presentations at international conferences focused on metabolic disorders, ensuring accessibility to clinicians, researchers, and policy makers.