We searched PubMed, Scopus, and Web of Science (WoS) from inception to May 2025, with supplementary checks in Google Scholar. Search strings combined DFU terms with TCM terms and a meta-analysis filter (full strategies in Supplementary file). Field limits were: PubMed Title/Abstract, Scopus TITLE-ABS-KEY, and WoS Topic, which searches Title, Abstract, Author Keywords, and Keywords Plus. No date limits were applied. Language was restricted to English at the umbrella-review level. We screened reference lists of eligible papers and related reviews for additional studies.

The PICO criteria for the present umbrella meta-analysis were structured as follows: Population/Patients (P), eligible individuals were adults aged 18 years or older DFU; Intervention (I) focused on the administration of TCM including orally and topically; Comparison (C) control group or placebo; Outcome (O) DFU related outcomes including ulcer area, cure rate, effectiveness rete, healing time, adverse events, hospitalization duration and Amputation rate. Only meta-analysis studies published in English that explored the impact of TCM on DFU outcomes and had reported effect sizes (ES) along with their corresponding confidence intervals (CI) were considered for inclusion. No restriction was placed on the year of publication. Studies published in languages other than English were excluded, as our team was unable to ensure accurate translation and quality assessment across multiple languages. Original studies, editorials, letters to the editor, and observational studies, were excluded from consideration.

Studies were independently screened and Two reviewers (XL, SW) extracted the data based on pre-established criteria. This included author’s name, year of publication, sample size, study location, study number, TCM supplementation duration, intervention type, consumption type, mean age, Heterogeneity, ES and CI for DFU related parameters. Any discrepancies were resolved with discussion with third reviewer (XX).

Two independent reviewers (XX, XL) assessed the methodological quality of the included articles using the Assessing the Methodological Quality of Systematic Reviews 2 (AMSTAR2) questionnaire (31). The AMSTAR2 questionnaire comprises 16 items that are answered with “Yes,” “Partial Yes,” “No,” or “Not a Meta-analysis.” The AMSTAR2 checklist is divided into four categories: “Critically low quality,” “Low quality,” “Moderate quality,” and “High quality.” A score of 7 or higher indicated that a meta-analysis was of high quality.

First, we retrieved 559 articles by searching databases. Second, 392 studies relevant to the effect of TCM on DFU outcome remained after deduplication. Then, after evaluating the titles and abstracts, 322 articles were excluded. Next, 64 studies were also excluded after the full-text screening. In the next step, 5 studies from the rest of the sources that met the inclusion criteria were entered. Finally, a total of 11 studies with 44 data-sets were regarded as eligible for the umbrella review (Figure 1).

The characteristics of 11 meta-analyses with 44 data sets are presented in Table 1. All studies have been conducted in China. All included meta-analyses synthesized RCTs; no eligible meta-analysis of purely observational primary studies was identified published from 2010 to 2024. Eight studies, including 15 data sets, were conducted to assess the effects of TCM on effectiveness rates (16, 18, 19, 22, 23, 25, 32, 33). Five studies, including eight data sets, have investigated the impact of TCM on the cure rate (19, 20, 23–25). A total of eight studies, including nine data sets, have examined changes in healing time as a dependent variable in response to the TCM (19, 20, 22–25, 32, 33). Five studies, including eight data sets, have been conducted to assess the effects of TCM on ulcer areas (17, 18, 20, 32, 33). A total of two studies, including two data sets, have examined changes in hospitalization duration as a dependent variable in response to the TCM (20, 22), one study has investigated the impact of TCM on Amputation rate (25), and finally one study including one data set has investigated the effect of TCM on adverse events (18). The sample size of the included data sets ranged from 86 to 1958 people. Also, the age range of the data sets was from 55 to 65 years, and the average duration of interventions was from 1.7 weeks to 27 weeks. Among the clinical trial studies included in the mentioned meta-analyses, in three articles, the interventions were oral (16–18), one study was both oral and topical (25), and the rest were topical (19, 20, 22–24).

Table 2 presents the findings of the quality assessment of meta-analyses according to the AMSTAR2 questionnaire. Among them, four studies received a high-quality score (16, 20, 24, 32), three received a moderate score (19, 21, 25), and five received a low score (17, 18, 22, 23, 33). Two studies in the search section used one researcher (19, 25), which may cause bias. Also, two studies have not examined the quality of the included studies and their risk of bias in their discussion (18, 22). Moreover, two studies did not assess publication bias or discuss its results (17, 23). None of the studies mentioned the funding of the included studies. On the other hand, five studies have mentioned the data related to clinical trials entered (dosage, duration of intervention, sample size, etc.) carefully and in detail (16, 17, 20, 22, 25). Also, four studies have searched entirely and accurately with reference checks (16, 20, 24, 25).

Eight studies comprising 15 datasets examined the impact of various TCM interventions on effectiveness rates. Fourteen datasets reported a significant increase in effectiveness, while one—using Astragalus membranaceus (Fisch.) Bunge (verified via The World Flora Online)—showed no significant effect (Risk Ratio (RR): 1.70; 95% CI: 0.86–3.36; I²: 94%) (19). Significant heterogeneity was observed in three datasets from Yan-Wu Yang’s study (19) but not in the others. Da-Yuan Zhong et al. (18) found that Buyang Huanwu decoction, primarily based on Astragalus membranaceus, combined with Western medicine, yielded the highest effect size (OR: 6.12; 95% CI: 4.23–8.86; I²: 0%). Similarly, Xiaoping Wan et al. (22), reported a marked improvement with Periploca sepium Bunge liquid (RR: 5.38; 95% CI: 3.52–8.24; I²: 0%). In contrast, Lizi Tan et al. (16), observed the smallest increase using Ginkgo biloba L. extract injection (RR: 1.17; 95% CI: 1.08–1.27; I²: 0%).

Five articles encompassing eight datasets assessed the impact of TCM on cure rate, all demonstrating a significant improvement. The greatest effect was observed in Yan-Wu Yang et al. (19), who used Astragalus membranaceus (Fisch.) Bunge (verified via The World Flora Online) as the intervention (RR: 2.37; 95% CI: 1.85–3.04; I²: 0%). In contrast, the smallest effect was reported by Li Ma et al. (24), using Jinhuang powder containing Curcuma longa L. (RR: 1.19; 95% CI: 1.06–1.33; I²: 0%). A large-scale study by Yan-Wu Yang et al. [21], involving 1,958 participants and using Phellodendron amurense Rupr, also showed a substantial increase in cure rate (RR: 2.01; 95% CI: 1.61–2.50; I²: 34%). Similarly, Chen Shuo et al.’s data set (25), with an average intervention duration of 14.8 weeks, reported a notable improvement (RR: 2.12; 95% CI: 1.63–2.77; I²: 0%). Significant heterogeneity was present in two datasets (19, 23), while the others showed no heterogeneity.

Eight studies comprising nine datasets evaluated the impact of various TCM interventions on healing time, all reporting significant reductions. The greatest reduction was observed in Ke Shen Qu (23), using Periploca sepium Bunge (verified via The World Flora Online; MD: –5.73; 95% CI: −6.95 to −4.52; I²: 0%), while the smallest was reported in Chen Shuo (standardized mean difference (SMD): –0.64; 95% CI: −0.89 to −0.40; I²: 13%), with an average intervention duration of 22.8 weeks—the longest among the studies. Both Xiaoping Wan and Ke Shen Qu employed Periploca sepium Bunge liquid (22, 23). The average intervention duration of the Chen Shuo dataset is 22.8 weeks, which is more than that of others (25). Significant heterogeneity was detected in the Ke Shen Qu and Chen Shuo datasets (23, 25), while no significant heterogeneity was observed in the remaining studies.

Eight datasets assessed the effect of TCM on ulcer area. Da-Yuan Zhong et al. (18) reported a non-significant reduction using Buyang Huanwu decoction—primarily Astragalus membranaceus (Fisch.) Bunge—combined with Western medicine (MD: –1.72; 95% CI: −4.67 to 1.23; I²: 96%; n = 184). Three datasets evaluated the rate of ulcer area reduction during TCM intervention. In the first two, Resina Draconis (Daemonorops draco [Willd.] Blume) applied for 4 and 4.5 weeks significantly increased the probability of achieving ≥70% (RR: 1.29; 95% CI: 1.10–1.59; I²: 0%) and ≥50% (RR: 1.38; 95% CI: 1.15–1.66; I²: 0%) reductions, respectively (20). In contrast, datasets from Min Chen’s study (17) using Chinese herbal medicine, showed reductions in the probability of minimal to complete ulcer area reduction: <30% (RR: 0.34; 95% CI: 0.21–0.53; I²: 76%), 30% (RR: 0.81; 95% CI: 0.71–0.92; I²: 0%), and 100% (RR: 0.62; 95% CI: 0.39–0.97; I²: 0%). Significant heterogeneity was reported in two datasets (20, 27), while the remaining four showed no significant heterogeneity (17, 18).

Two studies, each contributing one dataset, assessed the impact of TCM on hospitalization duration. Haoyue Feng et al. (20) used Resina Draconis (Daemonorops draco [Willd.] Blume) and reported a significant reduction (MD: –9.00; 95% CI: −9.81 to −8.19; I²: 0%) with an average intervention duration of 3.75 weeks and a sample size of 213. No significant heterogeneity was observed. Similarly, Xiaoping Wan et al. (22) found a significant effect using Periploca sepium Bunge (verified via The World Flora Online) liquid (SMD: –3.68; 95% CI: −5.38 to −1.97; I²: 97%) in a larger sample of 408 participants, though significant heterogeneity was present.

Chen Shuo et al. (25) reported that TCM significantly reduced the odds of amputation by 64% (OR: 0.36; 95% CI: 0.20–0.65), with an average intervention duration of 27 weeks and a sample size of 299. No significant heterogeneity was observed.

Additionally, Da-Yuan Zhong et al. (18) examined the effect of Buyang Huanwu decoction—primarily Astragalus membranaceus (Fisch.) Bunge—combined with Western medicine on adverse events. The intervention showed no significant effect (RR: 1.50; 95% CI: 0.57–3.91), and significant heterogeneity was reported.

Across included meta-analyses, fixed-effects models were typically used under low heterogeneity and random-effects otherwise. Examples include: Chen (25) (fixed if p > 0.10 and I² < 50%; Egger test + funnel; sensitivity analyses); Tan 2018 (16) (fixed if I² < 50%, random otherwise; funnel plot for primary outcomes); Qu (23) (fixed if I² < 25%, random for 25–85%); Wan (22) (random where p < 0.05 and I² > 50%; Begg/Egger/funnel); Zhong (18) (fixed if I² ≤ 50%/p > 0.05, random otherwise; funnel plot); and Feng (34) (random when I² > 50% or p < 0.10; sensitivity analyses; publication bias not assessed due to <10 studies). These reporting patterns are reflected in Table 1 and our narrative synthesis.

Our umbrella synthesis—spanning 11 meta-analyses—aligns with earlier English-language meta-analyses on DFU adjunctive TCM. For example, meta-analyses evaluating topical formulations (e.g., Kangfuxin liquid, Resina Draconis) (22, 23, 34) report improvements in healing time, healing/“cure” rates, and ulcer area compared with standard care alone, consistent with our findings for topical interventions. Meta-analyses of oral formulations (e.g., Jinhuang powder, Buyang Huanwu decoction, Danggui Sini decoction) (18, 24, 32) similarly report gains in effectiveness and cure rates, again matching the direction of effects we summarize. Because this is an umbrella review, we did not re-pool primary data; rather, we compare and contextualize the reported pooled effects (with CIs and heterogeneity) across meta-analyses, highlighting areas of agreement and inconsistency.

Across the included meta-analyses, TCM was typically evaluated as an adjunct to standard DFU care (e.g., offloading, wound-bed preparation/debridement, infection control, metabolic and vascular optimization). Accordingly, our conclusions pertain to adjunctive use, not replacement of standard care. The clearest signals of benefit appear for topical TCM (healing time, ulcer area), while some oral formulations are more often associated with cure/effectiveness outcomes. Given heterogeneity and variable methodological quality across the contributing meta-analyses, clinicians should apply shared decision-making, consider product quality and local availability, and interpret benefits with appropriate caution. To strengthen international relevance, pragmatic RCTs and meta-analyses from diverse settings outside China—and head-to-head comparisons with other guideline-supported adjuncts—are priorities.

To the best of our knowledge, this is the first umbrella meta-analysis to comprehensively examine the effects of various types of TCM on outcomes related to DFU. A wide range of TCM interventions and their effects were systematically analyzed and discussed. However, the study has some limitations. It is also noteworthy that several included meta-analyses were rated as low quality according to AMSTAR2. This raises concerns regarding the robustness of their findings. While we reported their results for completeness, greater weight should be placed on evidence from high- and moderate-quality meta-analyses. Consequently, our conclusions should be interpreted with caution, as the inclusion of low-quality reviews may overestimate treatment effects. Also, limited number of studies addressed key outcomes such as adverse events, amputation rate, and hospitalization duration, restricting the depth of analysis for these endpoints. Additionally, many included studies lacked detailed information on intervention dosage, which hindered more precise evaluation. Future high-quality clinical trials with standardized dosing protocols and extended intervention durations—particularly targeting underrepresented outcomes—are needed to enhance the validity and generalizability of findings. As this was an umbrella review, we synthesized evidence from published meta-analyses without pooled re-analysis; therefore, forest plots were not generated. While consistent with the purpose of umbrella reviews, future studies could strengthen interpretability by re-analyzing primary data and providing graphical synthesis. Another limitation is the absence of protocol registration in PROSPERO or a similar database, which—although this review focused on synthesizing existing meta-analyses—may reduce methodological transparency. Additionally, all included meta-analyses were conducted in China, which introduces a potential geographical bias and may limit generalizability to other healthcare contexts. Furthermore, the majority of the original trials lacked clear reporting of funding sources, creating risk of publication bias or selective reporting. Finally, substantial heterogeneity in TCM formulations, dosages, and intervention durations was observed, which complicates interpretation and comparison across studies.