The literature was searched using a combination of subject terms and extra terms. The search formula was: ((Foot Bath OR Fumigation OR Fumigations OR Washing OR Steaming OR Lavipeditum) AND (Diabetic Foot OR Diabetic Feet OR Diabetes Foot OR Diabetes Feet)). Relevant studies were retrieved from four English-language databases (Embase, PubMed, Cochrane Library, and Web of Science) and four Chinese-language databases (China National Knowledge Infrastructure [CNKI], China Science and Technology Journal Database [VIP], Wanfang Data, and Sinomed). Additionally, gray literature was explored by searching the World Health Organization International Clinical Trials Registry Platform (WHO-ICTRP) and ClinicalTrials.gov. References from relevant reviews were also reviewed to ensure comprehensive coverage. The search was current up to 24 May 2025, with no language restrictions. The detailed search strategies for each database used in the initial search are provided in Supplementary Table S1.

Inclusion criteria: (i) Participants were required to meet the diagnostic standards for T2DM as specified by the WHO in 1999, as well as the diagnostic criteria for diabetic foot established by any national or international authoritative organization such as the Chinese Guidelines for the Diagnosis and Treatment of Diabetic Foot (The Diabetic Foot Branch of the China Healthcare International Exchange Promotion Association, 2017) and Traditional Chinese Medicine Standards for Diabetic Foot Diagnosis and Treatment (Fan et al., 2011). Additionally, participants needed to be classified as Wagner grade 0 or 1. (ii) The control group received a warm water foot bath, while the experimental group received Chinese herbal foot baths. (iii) The primary efficacy outcomes included the clinical effective rate, common peroneal nerve sensory nerve conduction velocity (SNCV), common peroneal nerve motor nerve conduction velocity (MNCV), and ankle-brachial index (ABI). The clinical effective rate was assessed using a symptom scale based on the Guidelines for Clinical Research of New Traditional Chinese Medicine (Zheng, 2002). A reduction of 30% or more in the symptom scale score was defined as effective. The clinical effective rate refers to the proportion of participants rated as effective relative to the total number of participants, used to evaluate the improvement of diabetic foot-related symptoms (Zheng, 2002). The ABI was used to assess blood circulation in the lower limbs, with a decrease indicating insufficient blood flow and an increased risk of foot ulcers (Brito-Zurita et al., 2013). Common peroneal nerve SNCV and MNCV were measured to evaluate the sensory and motor functions of the peroneal nerve, with decreases indicating the extent of nerve damage (Lawrence and Locke, 1961). Secondary efficacy outcomes include blood glucose levels (hemoglobin A1c [HbA1c], fasting blood glucose [FBG], and 2-h postprandial blood glucose [PBG]) and lipid levels (total cholesterol [TC] and triglycerides [TG]). HbA1c provides an average of blood glucose levels over the past 2–3 months, serving as a marker of long-term blood glucose control, while FBG and PBG reflect immediate control of fasting and postprandial blood glucose levels, respectively. TC and TG are indicators of lipid metabolism, which are particularly important in managing diabetic complications. The safety outcome was defined as the adverse events. (iv) The study design was a randomized controlled trial (RCT).

Exclusion criteria: (i) Duplicate publication of data. (ii) Incomplete data. (iii) Data not available.

The literature screening process began with the organization of all retrieved articles using Endnote X9. A systematic approach was employed to select studies based on predefined inclusion and exclusion criteria. These criteria were established to ensure that only relevant studies addressing the research question were included. Each article was carefully reviewed for eligibility, focusing on aspects such as study design, population characteristics, and intervention details. The screening process was conducted in multiple stages, including title and abstract screening followed by full-text review.

Once the relevant studies were identified, essential data were compiled into a comprehensive table using Excel 2010. This table included critical information such as the primary author’s name, publication year, sample sizes, male-to-female ratio, average age of participants, course of disease, Wagner grade, frequency of foot baths, duration of treatment, and various clinical parameters. The clinical parameters collected encompassed the clinical effective rate, ABI, MNCV, SNCV, HbA1c, FBG, 2h-PBG,TC, TG, and adverse events. This structured data collection process ensured that all relevant information was systematically organized for further analysis.

Risk of Bias 2 (RoB 2) tool was used to assess the risk of bias in each included study (Sterne et al., 2019), which evaluates five specific domains: the randomization process, deviations from intended interventions, missing outcome data, outcome measurement, and selection of reported results. The risk of bias for each domain was classified as low, high, or of some concern according to the Cochrane Handbook for Systematic Reviews of Interventions (Sterne et al., 2019). This thorough risk assessment aimed to provide a clear understanding of the quality and reliability of the included studies. The above assessment was conducted independently by two reviewers, and any disagreement was resolved through discussion and adjudication by a third reviewer.

The meta-analysis was conducted using RevMan 5.3 software. Effect measures for continuous variables were represented as the mean difference (MD), while the risk ratio (RR) and the Peto-odds ratio (PetoOR) were utilized for dichotomous variables. Heterogeneity among studies was assessed using the I² statistic, with I² values interpreted as follows: I² < 50% indicated low to moderate heterogeneity, while I² ≥ 50% suggested high heterogeneity. Given the inherent variability in interventions, particularly with individualized treatments such as polyherbal prescriptions, random-effect models were applied for all analyses to account for this variability.

To explore potential sources of heterogeneity, meta-regression analyses were conducted when the number of included studies was greater than 10 to ensure sufficient statistical power and reliability. These analyses aimed to assess the impact of covariates such as baseline characteristics, treatment duration, or intervention specifics on outcomes with significant heterogeneity. If the number of studies was insufficient (n ≤ 10), meta-regression was not performed, and heterogeneity was instead explored via sensitivity analyses. Sensitivity analyses were performed to evaluate the robustness of the meta-analysis results and to explore potential sources of heterogeneity. This involved a leave-one-out approach, where each individual study was sequentially excluded from the pooled analysis. The impact of each exclusion was assessed by noting changes in the overall effect size, confidence interval (CI), and heterogeneity metrics. A significant reduction in heterogeneity after removing a specific study suggested that this study might be a primary contributor to heterogeneity. Additionally, if the exclusion of any individual study did not materially alter the overall effect estimate, the results were considered robust and stable. This process helped confirm the reliability of the findings and contributed to understanding the consistency of the evidence. Subgroup analyses were performed based on clinical parameters such as Wagner grade, treatment frequency, treatment duration, and the composition and dosage of Chinese herbal medicine, with the aim of investigating the impact of clinical heterogeneity on the primary outcome of clinical effective rate.

The TSA was conducted to assess the robustness of cumulative evidence in the meta-analysis while controlling for type I and II errors from repeated significance testing. TSA was performed using version 0.9.5.10 beta, with a two-sided alpha of 5% and 80% power. The required information size was calculated based on the anticipated effect size, control event rate, and heterogeneity-adjusted diversity. A cumulative Z-curve was plotted against predefined monitoring boundaries to determine whether the evidence was sufficient to confirm or refute the intervention effect. If the Z-curve crossed the boundaries before reaching the required sample size, the evidence was considered conclusive; otherwise, further trials were likely needed.

Funnel plots were created to visually assess publication bias by examining the symmetry of effect sizes against their standard errors; any asymmetry may indicate potential bias. To enhance this visual assessment, Egger’s tests were performed to statistically quantify asymmetry by regressing the standard normal deviate (effect size divided by standard error) against precision (inverse of the standard error). A significant result (p < 0.05) suggests possible publication bias. By combining the visual evaluation from the funnel plot with the results from Egger’s test, we aimed to more accurately assess the likelihood of publication bias in the meta-analysis.

The certainty of the evidence for outcomes was assessed using the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) approach, considering factors such as risk of bias, inconsistency, indirectness, imprecision, and publication bias. This process was performed independently by two reviewers, with disagreements resolved through discussion. We used the GRADEpro Guideline Development Tool (gradepro.org) to facilitate transparent and consistent ratings. Randomized trials started as high certainty, and adjustments were made based on the evaluation of each domain. The final ratings were categorized as high, moderate, low, or very low certainty, providing a clear measure of confidence in the evidence supporting our findings.

Frequency analysis and association rule analysis were performed using SPSS Modeler 18.0 to identify key candidate herbs (Wang et al., 2024b; 2024c). In the frequency analysis, a threshold of ≥30% was established to pinpoint high frequency herbs utilized in foot-bath treatments. This threshold signifies that at least 30% of the analyzed cases included a specific herb, underscoring its prevalent use within the dataset (Yu et al., 2025). Subsequently, these high frequency herbs would be used for association rule analysis.

For the association rule analysis, the Apriori model was employed. The parameters were set with a support threshold of ≥30%, a confidence threshold of ≥80%, and a lift threshold of ≥1.0 to investigate key candidate herbs combinations. The support threshold represents the proportion of transactions in the dataset that include both the antecedent and consequent items, thereby indicating how frequently these items co-occur. The confidence threshold measures the likelihood that the consequent occurs given the presence of the antecedent, reflecting the strength of the association. Meanwhile, the lift threshold assesses the degree of independence between the antecedent and consequent, with lift >1.0 indicating a positive correlation. These thresholds were selected based on established practices in the literature (Yu et al., 2024b; 2024a), which indicate that a support level of 30% ensures that the identified associations are not merely coincidental, while a confidence level of 80% suggests a strong likelihood of the consequent occurring when the antecedent is present. The herbs identified within these combinations are defined as key candidate herbs.

A total of 2,767 articles were identified from various databases, including 548 from CNKI, 487 from VIP, 707 from Wanfang, 495 from Sinomed, 39 from PubMed, 63 from Embase, 211 from the Cochrane Library, 122 from Web of Science, 42 from ClinicalTrials.gov, 46 from WHO ICTRP, and 7 from reference lists. During the screening process, 1,399 articles were eliminated due to duplication, and additional 1,355 articles were excluded for failing to meet the inclusion criteria. Ultimately, 13 articles (Gao, 2009; Zhang et al., 2011; Zhou et al., 2011; Su, 2012; Chen and Zhao, 2014; Huang, 2014; Huang and Wang, 2015; Zhou et al., 2015; Ding and Huang, 2016; Liu, 2016; Gao, 2018; Jiang et al., 2018; Zhang et al., 2018) were included in this study, as shown in Figure 2.

A total of 13 clinical studies (Gao, 2009; Zhang et al., 2011; Zhou et al., 2011; Su, 2012; Chen and Zhao, 2014; Huang, 2014; Huang and Wang, 2015; Zhou et al., 2015; Ding and Huang, 2016; Liu, 2016; Gao, 2018; Jiang et al., 2018; Zhang et al., 2018) were included in this review, all of which were conducted in China between 2009 and 2018, encompassing a total sample size of 921 participants. Among these, 451 participants underwent warm water foot baths, while 470 received Chinese herbal foot baths (Table 1). The composition and dosage of the Chinese herbs used in the included studies are detailed in Supplementary Table S2.

In the randomization process, seven studies utilized random number tables, while six studies did not report details regarding the random sequence generation; however, all studies did not provide information on allocation concealment. Therefore, 13 studies were all assessed as having some concerns regarding the randomization process. In the deviations from intended interventions, three studies employed a placebo design, while 10 studies did not describe the blinding details. Consequently, 10 studies were rated as having some concerns in this area. Additionally, all studies reported dropout rates of less than 20%, and the outcome measurements were not influenced by blinding, with all pre-specified outcomes reported. Therefore, the domains of missing outcome data, measurement of the outcome, and selection of the reported result for all studies were assessed as low risk. In summary, the overall risk of bias for the 13 studies was rated as some concerns, as shown in Figure 3.

The I² statistics indicated a high level of heterogeneity for the results of ABI and MNCV, while moderate heterogeneity was observed for HbA1c. Consequently, a meta-regression analysis was employed to investigate the sources and impacts of heterogeneity for ABI, MNCV, and HbA1c. However, due to the inclusion of fewer than ten studies for these outcomes, the meta-regression analysis was limited and could not be conducted. Additionally, considering that sensitivity analyses had already identified potential sources of heterogeneity for ABI, MNCV, and HbA1c, a meta-regression was not performed in this study.

Subgroup analysis, using the clinical effective rate as the outcome measure, investigated the impact of factors such as Wagner grade, treatment frequency, treatment duration, and the composition and dosage of Chinese herbal medicine on the efficacy of Chinese herbal foot baths in the treatment of diabetic foot, as shown in Table 2.

For the Wagner grade, the statistical threshold corrected by Bonferroni method is p < 0.0167. The reusults showed that Chinese herbal foot baths significantly improved the clinical effective rates of diabetic foot patients with Grade 0 (RR 1.39, 95% CI 1.26 to 1.53, p < 0.00001), Grade 1 (RR 1.37, 95% CI 1.09 to 1.72, p = 0.007), and Grades 0–1 (RR 1.47, 95% CI 1.26 to 1.72, p < 0.0001).

Regarding treatment frequency, the statistical threshold corrected by Bonferroni method is p < 0.01. The following results were observed: 30 min bid (RR 1.61, 95% CI 1.20 to 2.15, p = 0.001), 30 min qd (RR 1.41, 95% CI 1.26 to 1.58, p < 0.00001), 20 min bid (RR 1.53, 95% CI 1.12 to 2.10, p = 0.008), 20 min qd (RR 1.32, 95% CI 1.14 to 1.53, p = 0.0001), and 20 min q2d (RR 1.62, 95% CI 1.19 to 2.22, p = 0.002) of the Chinese herbal foot bath all significantly improved the clinical effective rate.

In terms of treatment duration, the statistical threshold corrected by Bonferroni method is p < 0.025. The results suggested that both ≤30 days (RR 1.39, 95% CI 1.25 to 1.54, p < 0.00001) and >30 days (RR 1.48, 95% CI 1.29 to 1.70, p < 0.00001) of the Chinese herbal foot bath significantly improved the clinical effective rate.

However, due to the variability in the composition and dosage of the Chinese herbal formulas used for the foot bath, it was not possible to conduct a subgroup analysis based on the herbal composition or formula.

The TSA demonstrated that the cumulative Z-curve for the clinical effective rate, ABI, common peroneal nerve MNCV, and common peroneal nerve SNCV crossed the monitoring boundaries, indicating that the meta-analysis results for these outcomes have reached the required sample size and are conclusive. In contrast, the cumulative Z-curve for HbA1c, FBG, 2h PBG, TC, and TG did not cross the monitoring boundaries, suggesting that more studies are needed to further evaluate and validate these findings, as shown in Figure 8.

Funnel plots for ABI, common peroneal nerve SNCV, 2h PBG, TC, and TG exhibited symmetric scatter distributions on both sides, indicating no evidence of publication bias. In contrast, the funnel plots for the clinical effective rate, common peroneal nerve MNCV, HbA1c, and FBG displayed asymmetric scatter distributions, suggesting potential publication bias, as shown in Figure 9.

Furthermore, Egger’s tests indicated no significant potential publication bias for ABI (p = 0.806), common peroneal nerve MNCV (p = 0.328), HbA1c (p = 0.553), and FBG (p = 0.793), suggesting that the publication bias observed in the funnel plots is not severe. However, the Egger’s test for the clinical effective rate showed evidence of potential publication bias (p < 0.001). It is important to note that Egger’s tests could not be performed for common peroneal nerve SNCV, 2-h PBG, TC, and TG, as these analyses included only two studies.

The GRADE approach indicated that the certainty of evidence for the clinical effective rate, common peroneal nerve SNCV, HbA1c, FBG, 2h PBG, TC, TG, and adverse events was low. In contrast, the certainty of evidence for ABI and common peroneal nerve MNCV was very low, as shown in Table 3.

The frequency analysis included 13 regimens of Chinese herbal foot baths and 61 Chinese herbal medicines. Among these, the frequencies of [Cinnamomum cassia Presl] (69.23%), [Conioselinum anthriscoides ‘Chuanxiong’] (46.15%) [Carthamus tinctorius L.], (46.15%) [Prunus persica (L.), Batsch] (46.15%), [Angelica sinensis (Oliv.) Diels] (38.46%), [Asarum heterotropoides F. Schmidt] (38.46%), [Salvia miltiorrhiza Bunge] (30.77%), and [Paeonia lactiflora Pall.] (30.77%) all exceeded 30%, which were considered as high frequency herbs for diabetic foot with Wagner grade of 0 or 1, as shown in Table 4.

The network diagram of association rule analysis is shown in Figure 10. The association rule analysis reported 11 results and seven combinations of Chinese herbal medicine, as recorded in Table 5. The combinations of “[Cinnamomum cassia Presl] - [Conioselinum anthriscoides ‘Chuanxiong’]”, “[Paeonia lactiflora Pall.] - [Angelica sinensis (Oliv.) Diels]”, “[Prunus persica (L.) Batsch] - [Carthamus tinctorius L.]”, “[Cinnamomum cassia Presl] - [Carthamus tinctorius L.]”, “[Cinnamomum cassia Presl] - [Prunus persica (L.) Batsch]”, “[Cinnamomum cassia Presl] - [Asarum heterotropoides F. Schmidt]”, and “[Cinnamomum cassia Presl] - [Carthamus tinctorius L.] - [Prunus persica (L.) Batsch]” emerged as key candidate herbs combinations for foot baths. These combinations encompassed the key candidate herbs such as [Cinnamomum cassia Presl], [Conioselinum anthriscoides ‘Chuanxiong’], [Paeonia lactiflora Pall.], [Angelica sinensis (Oliv.) Diels] [Prunus persica (L.), Batsch], [Carthamus tinctorius L.], and [Asarum heterotropoides F. Schmidt].

The potential of Chinese herbal foot baths in the treatment of diabetic foot has received more and more attention. To the best of our knowledge, this study is the first meta-analysis and data mining of Chinese herbal foot baths in the treatment of diabetic foot with Wagner grade of 0 or 1. This meta-analysis indicates that, compared to warm water foot baths, Chinese herbal foot baths significantly improve the clinical effective rate, ABI, and MNCV and SNCV of the common peroneal nerve, with these results being robust. Although the meta-analysis showed a slight reduction in FBG and HbA1c levels with Chinese herbal foot baths, these outcomes were not robust. Additionally, the analysis supports that Chinese herbal foot baths have no significant impact on 2h PBG, TC, or TG levels, and do not increase the incidence of adverse events.

The meta-analysis revealed that Chinese herbal foot baths, compared to the warm water foot baths, improved the clinical effective rate by 42%, the ABI by 0.19, the common peroneal nerve MNCV by 4.09 m/s, and the common peroneal nerve SNCV by 3.83 m/s. These findings were confirmed to be robust and conclusive through sensitivity analysis and TSA. The clinical effective rate refers to the proportion of participants rated as effective relative to the total number of participants, used to evaluate the improvement of diabetic foot-related symptoms. These symptoms included numbness, tingling, cold skin, dullness or loss of sensation, and superficial ulcers on the foot, which are the result of a combination of peripheral vascular and neuropathy in the diabetic foot. The significant improvement in clinical effective rate suggests that Chinese medicine foot baths can effectively relieve clinical symptoms in patients with diabetic foot, thereby reducing their pain and overall disease burden. ABI serves as an assessment tool for evaluating the severity of lower limb arterial stenosis, with values below 0.9 indicating possible stenosis or obstruction of lower limb arteries. Given that diabetic patients have a higher risk of lower extremity arterial stenosis or obstruction, ABI represents an important indicator to evaluate the blood flow and vascular function of their feet (Brito-Zurita et al., 2013). The common peroneal nerve MNCV and SNCV are indicators for evaluating the nerve function of the lower limbs and are of great significance in evaluating the degree of neuropathy in diabetic foot (Lawrence and Locke, 1961). The benefits observed in ABI, MNCV, and SNCV indicate that Chinese herbal foot baths have the effect of improving peripheral vascular and neuropathy, and it may be the direct mechanism by which it alleviates symptoms and signs related to diabetic foot.

The results of meta-analysis also indicated that Chinese herbal foot baths reduced HbA1c and FBG compared with warm water foot baths, while 2h PBG, TC and TG were all equivalent. However, the Chinese herbal foot bath only reduced HbA1c by 0.15% and FBG by 0.28 mmol/L, which did not meet the MCID, indicating that these differences are not clinically significant. Additionally, sensitivity analysis revealed that the benefit of Chinese herbal foot baths in blood glucose levels was driven by the study of Chen and Zhao (2014). After removing this study, the pooled results showed that there was no significant difference in HbA1c (MD -0.07, 95% CI -0.32 to 0.19, p = 0.62) and FBG (MD -0.09, 95% CI -0.44 to 0.26, p = 0.62) between two groups. This finding underscores the fragility of the observed glycemic benefit. Moreover, TSA indicated that the cumulative sample sizes for HbA1c, FBG, 2h PBG, TC, and TG did not reach the required thresholds, highlighting the need for further studies to investigate the effects of Chinese herbal foot baths on blood glucose and lipid levels.

The subgroup analysis revealed that the clinical effective rate of Chinese herbal foot baths for treating diabetic foot with grade 0, grade 1, and grade 0–1 was significantly higher than that of warm water foot baths. This indicates that Chinese herbal foot baths offer added benefits for diabetic foot patients with Wagner grade of 0 or 1. From the perspective of treatment frequency, regimens such as 30 min bid, 30 min qd, 20 min bid, 20 min qd, and 20 min q2d all showed significantly higher clinical effective rates than warm water foot baths. This suggests that Chinese herbal foot baths administered within this range—from 20 min q2d to 30 min bid—consistently demonstrate superior efficacy. In terms of treatment duration, both short-term (≤30 days) and long-term (>30 days) use of Chinese herbal foot baths were associated with better clinical outcomes compared to warm water foot baths. These findings suggest that the therapeutic benefits of Chinese herbal foot baths are not limited by the severity of the condition, treatment frequency, or duration, allowing clinicians to tailor regimens based on patient tolerance and clinical need.

Our meta-analysis found that among the limited number of studies reporting safety outcomes, 129 patients with diabetic foot who received Chinese herbal foot baths and 110 who received warm water foot baths experienced no treatment-related adverse events. This suggests that Chinese herbal foot baths may not increase the risk of additional adverse events. As a form of topical therapy, Chinese herbal foot baths theoretically reduce risks associated with oral absorption and hepatic or renal metabolism. Prior systematic reviews in other conditions, such as dysmenorrhea and hypertension, have also reported no additional adverse events, indicating a potentially favorable safety profile (Tian et al., 2024; Wu T. et al., 2024). However, only a small number of included studies reported on adverse events, and most lacked detailed safety data. As such, current evidence remains insufficient to draw firm conclusions regarding the safety of Chinese herbal foot baths for diabetic foot. Further high-quality studies with rigorous safety monitoring and standardized adverse event reporting are needed to better evaluate their safety profile.

The data mining revealed that [Cinnamomum cassia Presl], [Conioselinum anthriscoides ‘Chuanxiong'], [Paeonia lactiflora Pall.], [Angelica sinensis (Oliv.) Diels] [Prunus persica (L.), Batsch], [Carthamus tinctorius L.], and [Asarum heterotropoides F. Schmidt] were the key candidate herbs for foot baths in treating diabetic foot with Wagner grades of 0 or 1.

Cinnamaldehyde, the main active compound of [Cinnamomum cassia Presl], exerts neuroprotective effects against diabetes-related neuronal damage by inhibiting advanced glycation end-products (AGE)-induced abnormal cell cycle re-entry and apoptosis, thereby preserving neuronal cells and maintaining their terminally differentiated state (Wu Y. et al., 2024). Another significant compound from [Cinnamomum cassia Presl], eugenol, promotes angiogenesis in diabetic model mice and accelerates wound healing in diabetic wounds (Han et al., 2024). Additionally, Z-ligustilide extracted from [Conioselinum anthriscoides ‘Chuanxiong'] can reduce inflammation and oxidative stress by activating Nuclear factor erythroid 2-related factor 2 (NRF2), indicating its potential for treating diabetic foot (Yang et al., 2018; Qi et al., 2024). The role of [Paeonia lactiflora Pall.] in diabetic neuropathy has garnered increasing attention due to their antinociceptive, anti-inflammatory, antioxidant, and anti-apoptotic activities (Wiegand et al., 2024). Its major compound, paeoniflorin, promotes the transition of macrophages from the pro-inflammatory M1 phenotype to the anti-inflammatory/pro-healing M2 phenotype, thereby accelerating wound healing (Yang et al., 2021). Angelica sinensis polysaccharide, a key component of [Angelica sinensis (Oliv.) Diels], exhibits hypoglycemic, hypolipidemic, anti-inflammatory, and hepatoprotective effects by improving insulin sensitivity, enhancing glycogen synthesis, modulating adipokine secretion, and inhibiting apoptosis in diabetic models (Wang et al., 2015; 2016). Amygdalin, the main component of [Prunus persica (L.) Batsch], protects against high-glucose-induced ferroptosis and oxidative stress in retinal endothelial cells by activating the NRF2/antioxidant response element (ARE) signaling pathway, highlighting its therapeutic potential in alleviating diabetes-related microvascular damage (Li S. et al., 2023). Hydroxysafflor yellow A, an active component of [Carthamus tinctorius L.], protects pancreatic β-cells by alleviating oxidative stress-induced damage in T2DM rat models (Xue et al., 2021; Alshareef et al., 2024). Asarone, an extract from [Asarum heterotropoides F. Schmidt], improves insulin levels and mitigates inflammation, fibrosis, and carcinogenesis, thereby reducing the risk of diabetic complications (Das et al., 2019). In summary, these herbs and their active compounds demonstrate potential in managing diabetic foot and related complications through various mechanisms, including anti-inflammatory, antioxidant, and neuroprotective effects.

Apart from the composition of Chinese herbal medicine, the regimen for Chinese herbal foot baths involves factors such as the temperature and depth of the medicinal solution, the soaking duration, the commencement time, and the treatment duration. (i) The temperature of the medicinal solution. The 2011 International Diabetic Foot Working Group Guidelines for the Disposition and Prevention of the Diabetic Foot suggest that the water temperature of foot baths should be lower than 40°C, with the recommendation to test the water temperature by hand and adjust accordingly until it feels suitable (Xu and Li, 2013). The clinical study by Hu Y and colleagues (Xu and Li, 2013) demonstrated that the effects of Chinese herbal foot baths at 36°C, 38°C, and 40°C were comparable, indicating that the range of 36°C–40°C is suitable for treating diabetic foot. In practice, the temperature of the medicinal solution should be flexibly adjusted within the range of 36°C–40°C according to the specific conditions of patients. (ii) The depth of the medicinal solution. A related study showed that the effect of Chinese herbal foot baths when the depth of the medicinal solution reached 20 cm above the ankle was significantly better than that of 10 cm above the ankle and just submerging the ankle (Hu et al., 2013), suggesting that 20 cm above the ankle may be the optimal depth of the medicinal solution. (iii) The soaking duration. A literature survey on the duration of foot baths showed that the optimal soaking time for foot baths was 20–30 min (Zheng, 2012). A subsequent clinical trial found that the group immersed for 30 min in Chinese herbal foot baths significantly improved the ABI and toe-brachial index better than the 25-min and 20-min groups (Hu et al., 2013), suggesting that 30 min may be the optimal soaking time. (iv) The commencement time. Hu et al. recommended arranging Chinese herbal foot baths between 19:00 and 21:00 because foot baths at this time of day help patients relax and improve sleep quality (Hu et al., 2013). (v) The treatment duration. The included studies showed that Chinese herbal foot baths from 14 days to 90 days were significantly more effective than warm water foot baths, so the duration of Chinese herbal foot baths is recommended to be set at 14 days and above.

In conclusion, we recommend the following regimen for treating diabetic foot with Wagner grade of 0 or 1 using Chinese herbal foot-baths: key medicinal herbs should be added to 4,000 mL of water and boiled for 30 min to prepare the therapeutic solution. The solution should be maintained at a temperature of 36°C–40°C and should reach a depth of approximately 20 cm above the ankle. Each soaking session should last for 30 min, preferably between 19:00 and 21:00. The treatment should be continued for at least 14 days to ensure optimal therapeutic outcomes. It is important to emphasize that the recommended herbal formula is exploratory in nature, derived from preliminary data mining and literature review. While the findings of the current study are encouraging, the pharmacological effects of the formulation have not been rigorously validated. Moreover, the considerable variability in herbal formulations across clinical settings may influence their therapeutic outcomes. As such, these recommendations should not be interpreted as definitive clinical guidance but rather as hypothesis-generating insights. Rigorous validation through high-quality RCTs is necessary to establish both the efficacy and safety of the proposed regimen. Future research should also aim to standardize herbal compositions to improve the reproducibility and reliability of clinical outcomes.

This study has several notable limitations. First, all included studies did not report details of allocation concealment, which increases the risk of selective bias. More seriously, the lack of blinding in the majority of the included studies (10 out of 13) raises significant concerns regarding performance bias, which may undermine the reliability of the meta-analytic outcomes. These factors contribute to a high risk of underreporting and undermine confidence in the reported results, particularly by increasing the risk of false positives in subjective outcomes such as the clinical effective rate. Second, while the focus on patients with Wagner Grade 0–1 diabetic foot enhances internal validity, it restricts the external generalizability of the findings. Third, all included studies were conducted in China with exclusively Chinese participants, which raises concerns about ethnogeographic bias and the applicability of the results to non-Asian populations. Fourth, the adverse events were infrequently reported, with only three studies documenting zero events. This limitation restricts our ability to draw robust conclusions about the safety of the intervention. Fifth, due to substantial variations in herbal compositions across the included studies, we were unable to conduct subgroup analyses or meta-regression to further evaluate the associations between specific herbs and clinical outcomes, as well as herb-drug interactions, standardization, and pharmacokinetic properties. Sixth, our recommended herbal formula was derived from data mining and literature review, and while its pharmacological mechanisms and optimal dosage have not been rigorously validated. Furthermore, the cultural acceptance and regulatory barriers associated with implementing herbal foot baths therapy globally, along with the heterogeneity in herbal combinations even within the included studies, complicate the interpretation of the findings.

Future research should aim to address these limitations comprehensively. To improve methodological quality, future studies should incorporate rigorous designs with clear randomization, allocation concealment, and blinding procedures to reduce the risks of bias. Establishing research centers across different countries and diverse populations will help evaluate the efficacy and safety of Chinese herbal foot baths therapy beyond the current ethnogeographic context, enhancing external validity. Multi-center clinical trials using standardized herbal formulations are essential to confirm the therapeutic benefits and to facilitate reproducibility. Moreover, improved reporting of adverse events and safety data is crucial for a comprehensive safety profile. Standardization of herbal components and dosing regimens should be prioritized to minimize heterogeneity and ensure consistent results. Additionally, mechanistic studies are needed to elucidate the pharmacological actions underlying observed clinical effects, which could improve acceptance and integration into broader healthcare systems. Addressing cultural acceptance and navigating regulatory barriers at an international level will be vital for the global implementation of herbal foot baths therapy. Collectively, these efforts will help overcome current methodological, safety, and applicability limitations and advance the evidence base for this traditional intervention.