Evidence and Potential Mechanisms of Jin-Gui Shen-Qi Wan as a Treatment for Type 2 Diabetes Mellitus: A Systematic Review and Meta-Analysis

Background: Type 2 diabetes mellitus (T2DM) is a subtype of diabetes mellitus characterized by progressive dysfunction of β-cell insulin secretion and insulin resistance. Jīn-Guì Shèn-Qì Wán (JGSQW) has for many years been widely used in clinical practice as a treatment for T2DM. However, its effect remains unknown. Objectives: This study aims to summarize the clinical evidence of the effect of JGSQW on glucose and lipid metabolism in T2DM and the potential mechanisms underlying this effect. Methods: Six databases were searched without language or publication status restrictions. Data were extracted to a predefined template for synthesis. Results: Fourteen studies with 1586 participants were included in this meta-analysis. All 14 studies were judged to be at high risk of bias. JGSQW is safe for T2DM patients. Pooled results indicated that combination treatment results in a reduction in glycated hemoglobin (HbA1c) (mean difference (MD) −0.49%; 95% CI −0.67 to −0.31), fasting blood glucose (FBG) (MD −0.84; 95% CI −1.19 to −0.49), and 2-hour postprandial glucose 2hBG (MD −1.38; 95% CI −1.60 to −1.16). No significant difference in glucose metabolism was observed between JGSQW and hypoglycemic agents. The available evidence was insufficient to determine the effects on lipid metabolism. Sensitivity analyses indicated that these results were robust. Conclusion: By combining the available evidence, we found that JGSQW is safe for T2DM patients. Compared with hypoglycemic agents alone, combination treatment with JGSQW enhances the effect on glucose metabolism in patients with T2DM. We found no difference in the efficacy of JGSQW alone compared to hypoglycemic agents alone. In terms of lipid metabolism, the current evidence is insufficient and too inconsistent for us to draw firm conclusions, so further studies are needed.


INTRODUCTION
Type 2 diabetes mellitus (T2DM) is a subtype of diabetes mellitus (DM) characterized by progressive dysfunction of β-cell insulin secretion and insulin resistance (2. Classification and Diagnosis of Diabetes: Standards of Medical Care in Diabetes-2021Diabetes- 2021. Over 75% of patients with T2DM have abnormal lipid metabolism (Athyros et al., 2018;Eid et al., 2019) which contributes to the development of T2DM by increasing the insulin resistance. With more than 425 million diabetic patients globally, T2DM has become one of the most prevalent chronic diseases and the prevalence may increase with the rising numbers of obese patients (Cho et al., 2018;Li et al., 2020). In 2019, despite the fact that nearly 10% of global healthcare costs were related to diabetes management, 4.2 million adult deaths were attributed to diabetes and its complications (Peters et al., 2017;Association 2018;Dall et al., 2019). Although a series of hypoglycemic agents have been developed, an effective treatment of T2DM remains a challenge.
In China, botanical drugs are widely used in the treatment of T2DM Teng et al., 2018). Jīn-Guì Shèn-Qì Wán  Table 1. To date, there are no high-quality systematic reviews published on JGSQW as a treatment for T2DM. The experimental studies have indicated that JGSQW may exert its effect on T2DM through multiple mechanisms (Alam et al., 2014;Luo et al., 2015;Chen et al., 2016;Suchal et al., 2017;Zhuang et al., 2018;Yao et al., 2019). However, no studies to date have systematically summarized its beneficial effects and potential mechanisms.
Given this situation, the aim of this study is to 1) conduct a systematic review and meta-analysis of the clinical evidence of the effect of JGSQW on glucose and lipid metabolism in T2DM and 2) summarize the beneficial effects and potential mechanisms of JGSQW on T2DM as demonstrated in the experimental studies. In this study, the herb-drug interaction is analyzed through careful examination of the clinical and preclinical evidence. This study presents an accurate update on this issue as a basis for future research and clinical practice.

MATERIALS AND METHODS
To optimize our approach, this systematic review was conducted and reported following the guidelines provided by the Cochrane Handbook for Systematic Reviews of Interventions version 6.0 (updated July 2019), methodological expectations for conduct, reporting and updating of systematic reviews of intervention (MECIR), the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement, and A Measurement Tool to Assess Systematic Reviews 2 (AMSTAR 2) (Shea et al., 2017;Page et al., 2021). The PRISMA 2020 checklist is provided in Supplementary Material S1. As demonstrated by the AMSTAR 2 assessment in Supplementary Material S2, the overall methodological quality of this meta-analysis is high.

Search Strategies
Three English language databases (PubMed, Embase, and the Cochrane Central Register of Controlled Trials (CENTRAL)) and three Chinese language databases (China National Knowledge Infrastructure (CNKI), Wanfang Data Knowledge Service Platform, and VIP information resource integration service platform (cqvip)) were searched from inception to February 1, 2021, without language or publication status restrictions. The ClinicalTrials.gov database and Chinese Clinical Trial Registry (CHiCTR) were also searched to find any ongoing or unpublished clinical trials. In addition, the reference lists of the reviews and metaanalyses on this topic were searched. All the names of JGSQW are interchangeable and were used in database searching. The detailed search strategies are provided in Supplementary Material S3.

Study Types
Randomized controlled trials (RCTs) were included. Observational studies were not included due to the high risk of bias and confounding factors.

Participants
Studies including adults (18 years or older) with an established diagnosis of T2DM or T2DM complications were included. No restrictions were placed on other demographic factors of participants and settings. Studies including only a subset of relevant participants were included and the sensitivity analysis was carried out to assess the impact of including these studies.

Interventions
In TCM, subtle changes in formula composition can lead to significant differences in indications. According to the basic principles of evidence-based medicine, different interventions cannot be combined in the same meta-analysis. The composition of JGSQW in the included studies must be the eight medicines mentioned above without modification. The medicine was taken orally. Ideally, the authors should report the composition of the medicine they used including the source of any Chinese patent medicine. Studies missing this information were included in the primary analysis and excluded from the sensitivity analysis. Studies using modified JGSQW were not included in this meta-analysis. Comparators could include a placebo, anti-diabetic agents, and lifestyle management. Co-interventions, if administered to the intervention group, should also apply to the control group.

Comparisons
The following comparisons were investigated separately in this research: The first is JGSQW alone vs hypoglycemic agents, and the second is the combination treatment of JGSQW and hypoglycemic agents vs hypoglycemic agents.
In addition, we performed further comparisons according to the specific medicines used in the experimental and control groups.
If multiple time points were reported, the results with the longest time point were included in the analysis.

Safety Outcome
Any adverse events during the trials can be included.

Study Selection
Two authors (Zhipeng Hu and Maoyi Yang) screened all the citations independently by reading the titles and abstracts. Studies meeting the inclusion criteria were obtained for further screening. Disagreements were resolved by discussion with the third author (Xiaoke Liu). Studies that could not be included for reasons such as incomplete information in the abstract were classified as "Studies Awaiting Classification." Studies excluded after reading the full text and reasons for exclusion are listed in Supplementary Material S4.

Data Extraction
A form was developed prior to data collection to record the following data about each study: first author and year, country, diagnostic criteria, participant age (treatment/control; years), number of participants (treatment/control), number of females (treatment/control), duration of T2DM (treatment/control; years), co-intervention, treatment, comparator, duration of treatment, funding, and outcomes. Data extraction was carried out by two authors independently (Zhipeng Hu and Maoyi Yang), and the discrepancies were resolved by discussion with the third author (Xiaoke Liu). We contacted study authors by email for more unpublished information if necessary. Transformation of reported data was not carried out in this study.

Risk of Bias Assessment
A risk of bias assessment was carried out using the 2019 Cochrane risk-of-bias tool (version 2) for randomized trials (RoB2) (Sterne et al., 2019). This risk of bias assessment includes the following five domains: bias arising from the randomization process, bias due to deviations from intended interventions, bias due to missing outcome data, bias in measurement of the outcome, and bias in selection of the reported result-as well as an overall risk-of-bias judgement. The nature of the effect of interest was an "intention-to-treat" effect. Discrepancies were resolved by discussion with the third author (Xiaoke Liu). To ensure transparency of assessment, judgments and agreement on these are provided in Supplementary Material 5.

Data Synthesis
Data analysis was carried out in accordance with the Cochrane Handbook for Systematic Reviews of Interventions version 6.0 (updated August 2019). Meta package (version 4.11-0) for RStudio (Version 1.2.1335, https://www.rstudio.com/) was used for data synthesis. Forest plots were used to present the results. mean difference (MD) with 95% confidence intervals (CI) were used as effect measures for continuous outcomes. No dichotomous outcomes were used in this study. Prediction intervals were calculated to indicate the extent of betweenstudy variation if the number of included studies was more than ten. Considering the likelihood of heterogeneity, a random-effects model was used to pool the studies and a fixed-effects model was used for the sensitivity analysis.

Heterogeneity and Subgroup Analyses
A chi-squared test was used to test the heterogeneity between the studies with a significance level of P < 0.1. I 2 statistics were used to quantify the heterogeneity with values of 0-40%, 30-60%, and 75-100% indicating low, moderate, and high heterogeneity, respectively. The following potential effect modifiers were considered when conducting subgroup analyses: Baseline HbA1c (> 7% or ≤7%), baseline body mass index (BMI) (> 24 kg/m 2 or ≤ 24 kg/m 2 ), comorbidity (yes or no), age (≥ 65 years old or <65 years old), and form of JGSQW (patent or decoction).
The following additional subgroup analysis was performed for FBG: level of control group (≥ 9 mmol/L or < 9 mmol/L).

Assessing Non-Reporting Biases
We evaluated the non-reporting biases by comparing the published literature and protocols or registration information if possible. The funnel plot and Egger's test were carried out if at least 10 studies could be included in the analysis.

Dealing with Missing Data
We contacted the corresponding authors for missing data. In the absence of a reply, we imputed replacement values.

Identification of Phytochemical Profile, Herb-Drug Interactions With Anti-Diabetic Drugs, and Potential Mechanisms
The phytochemical profile of eight herbal medicines and JGSQW was identified in the platform of the TCM System Pharmacology (TCMSP) database (http://lsp.nwu.edu.cn/tcmsp.php) with two screening criteria: oral bioavailability (OB) ≥ 30% and druglikeness (DL) ≥ 0.18.
The preclinical studies are an important source for understanding the beneficial effects and herb-drug interactions. In order to provide a more comprehensive understanding of JGSQW as a treatment for T2DM, we identified all relevant preclinical studies to date and extracted data on herb-drug interactions between JGSQW and antidiabetic drugs and on the potential mechanisms for tabulation.

Database Search
A total of 1,245 citations were retrieved from the database search. Four hundred and sixty-two citations were excluded due to duplication. Seven hundred and fifteen citations were excluded after reading the titles and abstracts. After reading the full-texts, 52 citations were excluded. No additional publication was identified by searching the references of the included studies, related reviews, and meta-analyses. Two studies were found to be awaiting classification due to insufficient information. Fourteen studies were included in the final quantitative analysis. The full list of excluded studies with reasons is provided in Supplementary material S4. The flow chart of the study selection is shown in Figure 1.

Characteristics of the Included Studies
Fourteen studies including 1,586 participants were included in this meta-analysis (Zhaoyi et al., 2010;Xiaoming 2011;Xinyi 2011;Lina 2012;Yu and Yuan 2012;Yuting 2012;Yinzhong 2013;Yang 2016;Cuirong 2017;Fan 2017;Hongyun 2018;Jingzu et al., 2019;Bin 2020;Shufang 2020). Three studies did not report the detailed composition of medicine, nor did they clearly state whether the Chinese patent medicine was used. Therefore, we excluded these three studies from the sensitivity analysis and reanalyzed the data to evaluate their impact on the pooled effect (Yinzhong 2013;Yang 2016;Fan 2017). Two studies compared the efficacy of JGSQW and hypoglycemic agents directly, while other studies used a combination of hypoglycemic agents and JGSQW in the experimental group (Zhaoyi, Guang and Hongwei 2010;Xiaoming 2011).

HDL-C
One study including 78 participants reported a significant increase in HDL-C with combination treatment compared Two studies including 310 participants reported the effects of JGSQW compared with hypoglycemic agents on HDL-C (Zhaoyi, Guang and Hongwei 2010;Xiaoming 2011). Zhaoyi et al. (2010) reported no difference compared with Xiaoke Pills, while Xiaoming (2011) reported significant difference compared with metformin hydrochloride tablets (Supplementary Material S6.22).

TC
Two studies including 248 participants reported the effect of combination treatment compared with hypoglycemic agents alone on TC (Xinyi 2011;Yang 2016). Due to significant heterogeneity among studies, the data were not pooled. Yang (2016) reported that combination treatment of JGSQW decoction and lifestyle intervention decreased the level of TC compared with lifestyle intervention alone (MD −2.10; 95% CI −2.37 to −1.83), while Xinyi (2011) showed no difference for JGSQW patent medicine compared with metformin + Xiaoke Pill  Two studies including 310 participants reported the effect of JGSQW compared with hypoglycemic agents on TC (Zhaoyi, Guang and Hongwei 2010;Xiaoming 2011). Zhaoyi et al. (2010) reported no difference compared with Xiaoke Pills, while Xiaoming (2011) reported significant difference compared with metformin hydrochloride tablets (MD −1.00; 95% CI −1.47 to −0.53) (Supplementary Material S6.25).

Publication Bias
A funnel plot of data relating to 2hBG showed symmetry (Supplementary Material S6.28), and Egger's test showed no statistical significance (p 0.9994). A funnel plot of the FBG data showed asymmetry (Supplementary Material S6.29) and the result of Egger's test showed no statistical significance (p 0.8944).

Adverse Events
Two studies including 138 participants reported adverse events as outcomes (Yuting 2012;Yinzhong 2013). The work of Yuting (2012), which included 60 participants, reported that five hypoglycemic events occurred in the experimental group, while six such events occurred in the control group. The work of Yinzhong (2013), which included 78 participants, reported that no adverse events occurred during the study.

Phytochemical Profile and Potential Mechanisms of JGSQW on T2DM in the Experimental Studies
A total of 102 components were retrieved by searching the ingredients of eight herbs. After removing 11 duplicates, we identified a total of 91 active components for JGSQW. The detailed information of the components of herbal medicines and JGSQW are provided in Supplementary Material S7.1-S7.9. A total of 40 preclinical studies investigated the interaction between JGSQW and anti-diabetic agents and summarized the potential mechanisms of JGSQW on T2DM. The beneficial effects and potential mechanisms are summarized in Table 3.

Main Results and Application of This Research
Some important findings have emerged from this analysis of data from 14 studies. Combining JGSQW with hypoglycemic agents does not increase the incidence of adverse events such as hypoglycemia, indicating that JGSQW is safe in the clinical practice. HbA1c, FBG, and 2hBG are the three parameters for evaluating the glycemic control. Among them, HbA1c reflects the average glycemic level over the past few months and is therefore considered the gold standard for evaluating the glycemic control. In this study, compared with a control group, the combined use of JGSQW was found to reduce the level of HbA1c, enhancing the glycemic control level of patients with T2DM. Changes in both FBG and 2hBG will cause changes in HbA1c levels. This study found less reduction of 2hBG than of FBG, which means that JGSQW mainly decreases these levels by reducing 2hBG. The glucose lowering effects of JGSQW were consistent regardless of the form of JGSQW. We did not find significant heterogeneity in the results of studies on HbA1c and 2hBG, suggesting that the effect of JGSQW is consistent across different settings. We found substantial heterogeneity in the results of studies on FBG. The subgroup analysis on different levels of control group showed that the heterogeneity decreased significantly and there was a significant difference in the effect size among different subgroups, indicating that the effect of JGSQW may depend on the patients' FBG levels. The significant heterogeneity in the results of direct comparisons between JGSQW and hypoglycemic agents may be due to the different hypoglycemic agents used in the studies. Due to the lack of a placebo control, the net effect for JGSQW remains unknown.
Inconsistency was also found among the results on lipid metabolism, limiting the clinical application of those findings. On examining the data, we found that the differences in lipidlowering efficacy may be due to different comorbidities (with and without hyperlipidemia) of participants. Overall, the influence of JGSQW on lipid metabolism is unclear and its effect on type 2 diabetic patients with hyperlipidemia deserves further investigation.
In this study, we identified 91 active ingredients of JGSQW, of which the most common are Sitosterol, Stigmasterol, (−)-taxifolin, Alisol B, and (+)-catechin. These components lay the material foundation for the effect of JGSQW. However, the components of JGSQW are not only the addition of compounds in various Chinese herbal medicine but also the possibility of new compounds in the preparation process, which may be neglected 3 | Herb-drug interactions with the anti-diabetic drugs and the potential mechanisms of JGSQW as a treatment for T2DM reported by preclinical studies.

Risk of Bias
Using the latest ROB2 tool, we found that all included studies had a high risk of bias, mainly caused by deviations from intended interventions. In addition, most studies failed to report in detail their randomization method, including the generation of randomized sequences, and none described allocation concealment. These factors and an absence of intention to treat analysis may have led to exaggerated treatment effects.
Overall, due to the clear methodological shortcomings of the included studies, there is uncertainty in our judgment of the overall results.

Non-Reporting Biases
In general, the presence of publication bias leads to a pooled result that favors the intervention group. In the present study, we tested the publication bias by the funnel plot and Egger's test for FBG and 2hBG. The funnel plot for 2hBG was asymmetric, while the statistical results suggested the absence of publication bias. This contradictory result may be due to the low power of the statistical test. We did not perform publication bias tests for other outcomes due to low numbers of studies reporting these. An effective method to assess the presence of publication bias is to compare the published studies with the study registration information. In view of the low statistical power of the funnel plot and Egger's test and the fact that none of the clinical trials included in this study were registered, we cannot fully exclude the possibility of publication bias.

Novelty and Limitations of This Research
This study is the first high-quality systematic review to summarize the effect of JGSQW in the treatment of T2DM. Compared with previous studies, 10 new studies were included in this systematic review. The methodological quality of the present study was evaluated using the AMSTAR 2 method and was found to be high. In addition, we have provided detailed Supplementary Materials, allowing repetition and further evaluation of this research. This is also the first study to comprehensively summarize the evidence on the mechanisms by which JGSQW may affect T2DM in the experimental studies. The results of our study will be of great value for future basic and clinical research. Despite the use of high-quality methodology, inevitable significant limitations remain. All of the clinical trials had clear methodological flaws and it is difficult to quantify the impact of these risks of bias on the pooled results. Secondly, in this study, there are differences in the numbers of studies and participants for different outcomes. This is partly because different studies observed and consequently reported different outcomes. In order to decrease the diversity in outcomes and facilitate standardization of the outcome measurement across the trials, efforts to develop core outcome sets for conditions have been made in recent years (Williamson et al., 2012;Boers et al., 2014;Bain et al., 2016;Kirkham et al., 2016;Kirkham et al., 2017;Williamson et al., 2017;Kirkham et al., 2019;Egan et al., 2020). Despite this significant progress, further work is needed in the standardization of outcome reporting. Consistency in the reporting of study results can be effectively improved by adopting a widely agreed core outcome set and by preregistering the studies before they are conducted. In addition to the limitations mentioned above, incomplete data reporting impedes further analysis. Since most of the included studies are published in Chinese journals, there may be a certain degree of language bias. Overall, the results of this study were limited by the low quality of the included studies.

CONCLUSION
In summary, by combining the available evidence, we found that JGSQW is safe for the T2DM patients. Compared with hypoglycemic agents alone, combination treatment with JGSQW enhances the effect on glucose metabolism in patients with T2DM. We found no difference in the efficacy of JGSQW alone compared to hypoglycemic agents alone. In terms of lipid metabolism, the current evidence is insufficient and too inconsistent for us to draw firm conclusions, so further studies are needed.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material; further inquiries can be directed to the corresponding author.

AUTHOR CONTRIBUTIONS
XL designed the research. ZH, MY, and XL conducted the metaanalysis. ZH drafted the original manuscript and XL revised this manuscript.