Combination of Astragalus membranaceus and Panax notoginseng as Main Components in the Treatment of Diabetic Nephropathy: A Systematic Review and Meta-Analysis

Objective This meta-analysis evaluated the curative effect of the compatibility of Astragalus membranaceus and Panax notoginseng (ARPN) as main components on diabetic nephropathy. Methods We used various Chinese and English databases, including the Cochrane Library, PubMed, Embase, Web of Science, the China National Knowledge Infrastructure (CNKI), China Biology Medicine Disc (SinoMed), VIP, and Wanfang, to search for randomized controlled trials on the compatibility of Astragalus membranaceus and Panax notoginseng as main components. After data extraction, meta-analysis was performed with Review Manager 5.4.0 and Stata 15, and the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) framework was used to evaluate the quality of the evidence. Result A total of 17 studies involving 1342 patients with diabetic nephropathy were included. Compared with the control group, ARPN can significantly improve the clinical effective rate of diabetic nephropathy (OR 5.12, 95% CI 3.42 to 7.66, P < 0.00001), and the curative effect of reducing UAER (MD −26.67, 95% CI −31.30 to −22.04, P < 0.00001) and 24 h urinary protein (SMD −0.58, 95% CI −0.75 to −0.41, P < 0.00001) is also significantly better than that of the control group, and it can also improve the renal function(Scr: MD −13.78, 95% CI −25.39 to −2.17, P=0.02; BUN: MD −0.74, 95% CI −1.27 to −0.20, P=0.007). In addition, it can also reduce glycosylated hemoglobin (SMD −1.30, 95% CI −2.33 to −0.27, P=0.01) and blood lipid(TC: SMD −0.62, 95% CI −0.95 to −0.29, P=0.0002; TG: SMD −0.47, 95% CI −0.75 to −0.19, P=0.0009; LDL: SMD −0.43, 95% CI −0.68 to −0.18, P=0.0008), and improve the TCM syndrome score (MD −4.87, 95% CI −6.17 to −3.57, P < 0.00001). Subgroup analysis suggested that the treatment plan of the control group could be the sources of heterogeneity. All the included studies had no obvious adverse effects. Conclusions The compatibility of Radix Astragali and Radix notoginseng as the main components can effectively improve the renal function of patients with diabetic nephropathy and delay the progress of diabetic nephropathy. However, the results of this study need further research to be confirmed because of the uncertainty of the evidence and the suboptimal risk bias.


Introduction
Diabetic kidney disease (DKD) is considered as a serious complication of diabetes mellitus (DM) and the main cause of end-stage renal disease (ESRD) [1]. Traditional treatment methods have not achieved satisfactory results in many clinical treatment practices of DKD. It should be noted that TCM has a long history of treating diabetic complications [2], especially diabetic nephropathy. Because of its multitarget efect, Chinese medicine, as the main or alternative treatment for diabetic nephropathy, has good clinical efcacy. In the preclinical research of treating diabetic nephropathy, TCM play a therapeutic role in many ways, such as targeting the infammatory response, apoptosis, autophagy, and cell stress [3].
Astragalus membranaceus (AM) is a traditional Chinese medicine from Mongolian milkvetch or Membranaceus milkvetch. It has a longstanding history and has gained widespread clinical applications in China, with the functions of decreasing blood lipid, lowering blood sugar, eliminating edema, and so on [4,5]. In addition, it has been proven to have wide pharmacological efects on diabetes and its complications, particularly in DN, where it has a unique curative efect [6]. Panax notoginseng (Burk) F.H. Chen (Sanqi in Chinese) is a commonly used Chinese medicine. Pharmacological studies have shown that notoginseng and its extracts have many functions, such as antioxidation, anti-infammatory, regulation of blood pressure and blood glucose, inhibition of platelet aggregation, inhibition of neuronal apoptosis, and neuronal protection [7][8][9][10][11].
At present, both Radix Astragali and Radix notoginseng are widely used in the clinical treatment of diabetic nephropathy, but there is still no rigorous systematic evaluation to evaluate the safety and efectiveness of their clinical use in the treatment of diabetic nephropathy with the combination of the two drugs. Now, systematic evaluation is being conducted to provide some reference for clinical popularization and use.

Search Strategy.
Te databases Cochrane Library, PubMed, Embase, Web of Science, the China National Knowledge Infrastructure (CNKI), Wanfang, VIP, and China Biology Medicine (CBM) were systematically searched from the date of their inception until June 2022. No language restriction was imposed. Te following search terms were used: diabetic kidney disease, diabetic nephropathy, and Astragali, Astragalus, Astragalus membranaceus, Astragalus plant, Astragalus root, milkvetch root, Radix Astragali, and Sanqi, Sanqizongzaogan, Panax notoginseng, Panax notoginseng saponins, Lulutong, Xuesaitong, and Xueshuantong. An example search using PubMed is provided in Table 1.

Inclusion and Exclusion
Criteria. Studies meeting the following criteria were included: (1) Types of studies: clinical RCTs using ARPN as a treatment for DKD; (2) Types of participants: the diagnostic criterion for DKD included adults with primary diabetes and persistent albuminuria/ proteinuria, details are as follows: urinary albumin-tocreatinine ratio(UACR) ≥ 30 mg/g or urinary albumin excretion rate (UAER)≥ 30 mg/24 h and repeatedly check UACR or UAER within 3 to 6 months, and the critical value was reached or exceeded twice in 3 times; (3) Types of interventions and comparisons: RCTs comparing ARPN therapy versus control therapy were included (control therapy were ACEI/ARB or monotherapy including AR or PN or basic treatment.). (4) Outcomes: clinical efcacy, albuminuria, 24 h urinary protein, serum creatinine (Scr), blood urea nitrogen (BUN), glycosylated hemoglobin (HbAlc), total cholesterol (TC), triglycerides (TG), lowdensity lipoprotein cholesterol (LDL), and TCM symptom score.
Te exclusion criteria were: (1) duplicate studies; (2) reviews, animal, or cell experiments; (3) non-RCTs, case reports, meta-analysis, meeting abstracts, and editorials; (4) incomplete, incorrect, or not available studies data, or the study did not report at least one of the primary outcomes; (5) studies, in which control groups received an intervention that treatment groups did not receive; and (6) improper outcome measures.
Te primary outcomes were albuminuria and 24 h urinary protein. Secondary outcomes included clinical efcacy (CE), serum creatinine (Scr), blood urea nitrogen (BUN), HbA1c, total cholesterol (TC), triglyceride (TG), lowdensity lipoprotein (LDL), and TCM symptom score (TSS). Studies that failed to report at least one of these primary outcomes were excluded.

Selection of Studies and Data Extraction.
Two reviewers independently conducted the comprehensive search, and a third reviewer checked for consistency. Discrepancies about screening results were resolved by a discussion and reaching consensus. Te reviewers deleted duplicate records, screened the relevance of titles and abstracts, and identifed those that they were excluded or requiring further evaluation. We then read the full texts to make sure that all the selected articles meet the inclusion criteria. Te following data were extracted from each study: frst author and study year, interventions, comparisons, number of patients, duration, and results report.

Statistical
Analysis. Data analysis was performed by using the Review Manager 5.4.0 and the Stata 15 statistical package. For continuous outcomes, mean and standard mean (SD) were obtained from each study and presented as standard mean diference (SMD) or mean diference (MD). Te random efects model was used to calculate the pooled SMD when substantial heterogeneity occurred (P < 0.05 or I 2 > 50%). Publication/reporting biases were performed by using funnel plots. Subgroup analyzes were conducted to explore potential sources, where results were stratifed by factors such as diferent control group treatments, including ACEI/ARB, TCM, and basic treatment.

Study Identifcation and Selection.
A total of 315 records were retrieved in the initial database search. After deleting 23 duplicate articles, 292 records are eligible for further analysis. According to the inclusion and exclusion criteria, 246 articles were excluded after reading the titles and abstracts. Te remaining 76 full-text articles were evaluated. Studies with an irrelevant study design, non-RCTs, meta-analysis, and studies that reported only combination treatment were excluded.
Finally, a total of 17 RCTs were included in the metaanalysis. Te selection process is carried out according to the Preferred Reporting Items for Systematic Reviews and Metaanalyzes (PRISMA) guidelines, as shown in Figure 1 [12].

Study
Characteristics. Details of the study characteristics were presented in Table 2. A total of 17 studies involving 1342 patients with diabetic nephropathy were included in this study, which covered the period from 2002 to 2018. Among the 17 RCTs, 9 studies evaluated ARPN plus ACEI/ ARB compared with ACEI/ARB; six studies assessed ARPN compared with single Astragalus membranaceus or Panax notoginseng, and two studies evaluated ARPN plus basic treatment, including keep on a diet, lower high blood sugar and high blood pressure, etc.

Quality Assessment.
Te overall quality of the methodology of the included studies was low because of unmasked participants and personnel, unclear randomized and allocated procedures. Te risk of bias outcomes is summarized in Figure 2. Eight articles were randomized by a random number table. Te remaining studies did not mention the details of randomization. All the included trials did not describe the concealment, blinding, and caseshedding method.

Primary
Outcomes. Efects of ARPN on albuminuria and 24 h urinary protein Twelve studies with 861 participants reported data on urinary albumin excretion rate at the end of treatment. We classifed the studies into three groups based on the treatment of the control group: TCM, ACEI/ ARB, and basic treatment. Because of high heterogeneity, the random-efect model was used, as shown in Figure 3(a).

Efect of ARPN on Clinical Efcacy.
Eleven studies analyzed the total efective rate of ARPN. According to the clinical symptoms and laboratory examination, the curative efect of ARPN can be divided into obvious efective, effective, and invalid, and the total efective rate includes the proportion of obvious efective and efective. Compared with the control group, the total efective rate of the treatment group increased, with statistical signifcance (SMD 5.12, 95% CI 3.42 to 7.66, P < 0.00001) with good homogeneity (I 2 � 0%), as shown in  Te similar result was found in the BUN analysis. BUN levels at the end of treatment were evaluated in 570 participants from six studies. Compared with control group, ARPN group apparently reduced the BUN level (SMD −0.74, 95% CI −1.27 to −0.20, P < 0.01) with high heterogeneity (I 2 � 82%). Subgroup analysis shown that there was no signifcant diference between ACEI/ARB group and control group (SMD −0.40, 95% CI −0.93 to −0.14, P � 0.15, I 2 � 45%), while the TCM group had statistical signifcance (SMD −1.28, 95% CI −1.52 to −1.04, P < 0.00001, I 2 � 0%). Subgroup analysis revealed that diferent control group treatments may have been the source of heterogeneity ( Figure 5(b)).

Efect of ARPN on TCM Symptom
Score. Five studies with 335 participants compared TCM symptom score between the ARPN group and control group. Compared with control group, ARPN group apparently reduced TCM symptom score (MD −4.87, 95% CI −6.17 to −3.57, P < 0.00001) without heterogeneity (I 2 � 29%). As shown in Figure 8.

Publication Bias.
Publication bias was not detected in the primary outcomes of UAER (P > 0.05, Figure 9(a)) and 24 h proteinuria (P > 0.05, Figure 9(b)). Te results show that there is no publication ofset.

Quality Evaluation of GRADE.
Based on the main relevant outcome indicators included in the study, and referring to the GRADE evaluation standard of evidence quality, this study comprehensively evaluates the evidence quality from fve aspects: bias risk, evidence consistency, evidence directness, evidence accuracy, and publication bias.  Tau      When the combined evidence results have one of the above situations, the corresponding evidence quality will be degraded, and the evidence quality of the main outcome indicators in this study will be comprehensively judged based on the aforementioned possible degradation factors. Te evaluation results suggest that the evidence quality of the main therapeutic outcomes related to DKD treated by ARPN is "moderate" or "low." As shown in Table 3. Gao    Evidence-Based Complementary and Alternative Medicine 3.8. Safety Evaluation. Safety outcomes were reported in 3 out of 17 included studies [15,21,27]. Te authors described that the participants had a slight adverse efect, such as dizziness, but the symptoms gradually decreased with the treatment, and there was no statistical diference between the two groups.

Discussion
Because the pathogenesis of DN is complex, it is difcult to delay the progress of DN efectively by single drug acting on a certain target. Traditional Chinese medicine (TCM) is playing an increasingly important role in the prevention and treatment of DN because of its advantages of multitarget and combined medication. It is believed that the pathogenesis of diabetic nephropathy is qi defciency and blood stasis, so the method of invigorating qi and activating blood circulation is the fundamental treatment of DN, which should run through the whole treatment of DN [30]. Astragalus membranaceus and Panax notoginseng, as traditional Chinese medicines, have their main functions of invigorating qi and activating blood, respectively [31,32]. Te combination of them accords with the basic treatment of invigorating qi and activating blood and has a solid theoretical basis.
Toxicological studies show that the combination of Astragalus membranaceus and Panax notoginseng is safe, and no genetic mutation, target organ damage, or other toxic reactions have been observed in cell and animal experiments [33]. Radix notoginseng Compound (A&P) is a traditional Chinese medicine compound composed of Radix Astragali, Radix notoginseng, Radix Angelicae sinensis, Achyranthis Radix, and Tallus laminariae, which is widely used in China to treat CKD. It has been found that A&P combined with Bifdobacterium can downregulate the infammatory reaction of macrophages in the kidney and intestine by inhibiting a small molecular signaling pathway, thus protecting the kidney from the infuence of CKD, which provides a new idea for traditional drug treatment of CKD [34]. In diabetic nephropathy, Astragalus mongholicus Bunge and Panax notoginseng (Burkill) F.H. Chen formula (APF) signifcantly improved the renal damage of DN mice, especially recovered blood urea nitrogen, serum creatinine, and 24hour proteinuria. In STZ-induced diabetic nephropathy mice, APF also decreased the mRNA and protein expression of infammatory factors such as TNF-α, IL-1-β, and IL-6, and also improved the damage of high glucose to renal mesangial cells by regulating autophagy [35]. In addition, the combination of Astragalus membranaceus and notoginseng Radix can also improve kidney podocyte injury in diabetic rats, and the curative efect is signifcantly better than that of Astragalus membranaceus or notoginseng Radix alone [36]. Many clinical studies have shown that the combination of Astragalus membranaceus and Panax notoginseng can improve the renal function of diabetic nephropathy patients.
In this study, we aimed to analyze the efect of ARPN on renal function of diabetic nephropathy patients and included 17 clinical studies involving 1342 diabetic nephropathy patients, which involved compatibility of Astragalus membranaceus and notoginseng Radix or traditional Chinese medicine compound mainly composed of Astragalus membranaceus and notoginseng Radix. Te analysis shows that ARPN can signifcantly improve the clinical efective rate of diabetic nephropathy, and the curative efect of reducing UAER and 24 h UP is also signifcantly better than that of the control group, and it can also improve the renal function. In addition, it can also reduce glycosylated hemoglobin and blood lipid, and improve the TCM syndrome score. However, subgroup analysis showed that compared with the control group's basic treatment or only using Astragalus membranaceus and notoginseng Radix, when the control group's treatment plan was ACEI/ARB drugs, the curative efects of Scr, BUN, and HbA1c in the treatment group and the control group decreased. In the analysis of Scr, the P value of ACEI/ARB group is higher than that of TCM group. Te analysis of BUN shows that there is no statistical signifcance between the ACEI/ARB treatment group and control group, but there is signifcant statistical signifcance between the TCM treatment group and the control group. Te same result also appears in the analysis of HbA1c, which may be due to the renal protective efect of ACEI/ARB drugs. As we all know, ACEI/ARB drugs have been widely used in DKD treatment. Teir efectiveness in reducing or preventing microalbuminuria has been documented [8,9]. ACEI/ARB drugs can reduce proteinuria and improve renal function in patients with DKD. Terefore, with the intervention of ACEI/ARB drugs, the protective efect of ARPN on the kidney cannot be fully exerted.
Overall, the quality of the evidence in this study is moderate or low and the results of the risk bias analysis show that many studies have high or unclear risks. Although all the included studies indicated that a randomized controlled design was carried out, only 8 studies described the randomization method in detail, that is, using computergenerated random numbers or random number tables. Two studies described the blind method in detail, and adopted the standard double-blind method. Other studies lacked the description of blind method. None of the included studies mentioned distribution concealment. In addition, some test indexes are highly heterogeneous. After using randomized model or subgroup analysis, it is found that the treatment plan of the control group is the source of heterogeneity of some indexes. In addition, due to the limitation of language barrier, this study only searched Chinese and English databases, and all the included studies were conducted in China, which afected the fnal results to some extent. Terefore, the applicability of the results in other countries needs further study.

Data Availability
All data that support the fndings of this study are available from the corresponding author upon reasonable request.