Liuwei Dihuang Pills Enhance the Effect of Western Medicine in Treating Diabetic Nephropathy: A Meta-Analysis of Randomized Controlled Trials

Objectives. To assess the effectiveness and adverse effects of adding Liuwei Dihuang Pills (LDP) to Western medicine for treating diabetic nephropathy. Methods. Studies were retrieved from seven electronic databases, including PubMed, Embase, The Cochrane Library, CBM, CNKI, Chinese Scientific Journal Database (VIP), and Wanfang Data until November 2015. Study selection, data extraction, quality assessment, and data analyses were conducted according to Cochrane standards. Meta-analysis was performed on the overall therapeutic efficacy of hyperglycemia and renal functions, and the study also analyzed adverse events. Results. A total of 1,275 patients from 18 studies were included. The methodological quality of these included trials was generally low. We found that adding LDP can lower patients' FBG (MD: −0.36 [−0.46, −0.25], P < 0.00001), PBG (MD: −1.10 [−1.35, −0.85], P < 0.00001), and HbA1c (MD: −0.14 [−0.49, 0.21], P = 0.43). There were also improvements in lowering patients' BUN (MD: −0.67 [−0.89, −0.45], P < 0.00001), SCr (MD: −0.96 [−1.53, −0.39], P < 0.00001), 24 h UTP (SMD: −1.26 [−2.38, −0.15], P < 0.00001), UAER (MD: −26.18 [−27.51, −24.85], P < 0.00001), and UmAlb (SMD: −1.72 [−2.67, −0.77], P < 0.00001). Conclusion. There is encouraging evidence that adding LDP to Western medicine might improve treatment outcomes of diabetic nephropathy, including hyperglycemia and renal functions. However, the evidence remains weak. More rigorous high-quality trials are warranted to substantiate or refute the results.


Introduction
Diabetic nephropathy (DN) is a widely recognized microvascular complication of diabetes and almost the leading cause of end-stage kidney failure worldwide responsible for morbidity and mortality [1]. If the damage to the kidney and proteinuria is irreversible, it will evolve into End-Stage Renal Disease. However, exact pathogenesis of DN is still unclear and it is difficult for us to cure DN. At present diet management, control of blood pressure and blood sugar, and blood fat treatment are the foundation treatment for DN. Furthermore, an adequate control of high blood pressure and treatment of microalbuminuria are the major therapeutic targets [2]. To achieve adequate blood pressure control, a combination therapy with different classes of antihypertensive agents is often necessary, especially including angiotensin-converting enzyme inhibitors (ACEIs) and angiotensin receptor blockers (ARBs) [3]. ACEIs and ARBs have been demonstrated to protect renal function of DN but are not enough to delay or retard the progression of DN; therefore, exploring feasible drugs is the hotspot of medical research at present.
Currently, with increasing application of complementary and alternative medicine (CAM) worldwide, traditional Chinese medicine (TCM) has become more popular and has drawn more attention [4][5][6][7]. TCM has lots of advantages over the conventional medical approaches in the prevention of diabetic complications because of less toxicity and/or side effects [8][9][10]. TCM is becoming increasingly popular and widely used among patients with DN [11].
Liuwei Dihuang Pills (LDP), a traditional Chinese herbal formula containing six commonly used herbs (Rehmannia 2 Evidence-Based Complementary and Alternative Medicine glutinosa, Cornus officinalis Sieb., Common Yam Rhizome, Alisma orientalis, Tree Peony Bark, and Poria cocos), are widely used to DN-related symptoms in clinical practice for centuries in China. Three of the six ingredients in the formula are nutrients, while the other three facilitate drainage, enrich yin, nourish the kidney (Shen), and thereby address the root cause of diabetes according to the Chinese medicine theory [12]. Currently, LDP combined with Western medicine has been widely used as an alternative and effective method to treat or prevent diabetic nephropathy in China. Up to now, lots of studies have been published about the effects of LDP combined with conventional drugs for diabetic nephropathy. It is necessary for us to compare the effect of such combinations to the use of Western medicine alone. This report aims to evaluate the beneficial and adverse effects of LDP combined with conventional drugs for the treatment of DN in randomized trials. Additionally, we checked bibliographies of retrieved articles and prior reviews on the subject for additional references. We contacted the authors of included trials for missing information when necessary.

Inclusion and Exclusion
Criteria. All the randomized controlled trials (RCTs) based on LDP combined with Western drugs or Western conventional therapeutics compared with corresponding Western drugs or Western conventional therapeutics in patients with DN were included. No restrictions on language, population characteristics, or publication type were enforced. The primary outcome measure that the current study examined was the overall clinical efficacy of LDP for hyperglycemia and impairments of renal functions. A secondary outcome that was assessed was frequency of adverse events. Articles that duplicated the same groups of participants in another publication were excluded.   [13]. The items included the following 6 aspects: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), and selective reporting (reporting bias).

Data Synthesis.
All data analyses were conducted using the Review Manager 5.2 software provided by the Cochrane Collaboration. The outcomes were analyzed as continuous variables using fixed-or random-effect models, and the results for overall clinical efficacy were reported as weighted mean difference (WMD), 95% confidence intervals (95% CIs), and pooled odds ratio (OR). Meta-analysis was performed if the intervention, control, and outcome were the same or similar. It was considered to be indicators of a substantial level of heterogeneity when 2 > 50% or < 0.1. In the absence of significant heterogeneity, data were pooled using a fixed-effect model ( 2 > 50%), and, otherwise, data were pooled using a random-effect model (50% < 2 > 50% [14]. The relative strength of treatment efficacy was illustrated by forest plots.
The interventions included LDP combined with antihypertensive or routine treatment drugs. The controls included three types of groups, including antihypertensive drugs, routine treatment, and antihypertensive drugs combined   [17,23] described inappropriate methods of randomization by order of hospital registration. The others did not mention the sequence generation processes adequately. Thus we failed to judge whether it was conducted properly or not because of insufficient information. No trial stated the double-blind principle and allocation concealment. Only one trial [23] mentioned single blinding but gave no details of either participants' or investigators' or assessors' blinding. Only one trial [16] reported dropouts or withdrawals. Selective reporting was generally unclear in the RCTs due to lacking data for the current research team. All the trials did not mention follow-up. The pretrial estimation of sample size was not mentioned in all trials. We tried to contact the authors for further information but regrettably we have got no information.

Improvement of Hyperglycemia.
We analyzed the RCTs that studied the benefits of LDP for FPG, PBG, and HbA1c.
(1) Improvement of FPG. Six studies [17,18,22,23,28,29] including 371 participants used the levels of FPG as an outcome measure. These trials showed homogeneity in the consistency of the trial results (chi-square = 7.00; = 0.22; 2 = 29%). Therefore, a fixed-effect model should have been used for statistical analysis. A meta-analysis showed a significant difference for the LDP combined with conventional therapies group (MD: −0.36 [−0.46, −0.25]; < 0.00001), which demonstrated that LDP combined with conventional therapies group was superior to the conventional therapies taken by the control groups (Table 3).   (3) Improvement of HbA1c. Two studies [17,22] including 112 participants used the levels of HbA1c to measure the outcome. The trials showed homogeneity in the consistency of the trial results (chi-square = 0.07; = 0.79; 2 = 0%). Therefore, a fixed-effect model should have been used for statistical analysis. A meta-analysis showed that HbA1c measurements also showed greater efficacy for LDP combined with conventional therapies group (−0.14 [−0.49, 0.21]; = 0.43) ( Table 5).

Improvement of Renal Functions.
We analyzed the RCTs that measured blood urea nitrogen, serum creatinine, UAER, 24 h urine protein quantitation, and UmAlb.     [18,20,25,27,32] including 327 participants used the levels of 24 h UTP to measure the outcome. Four studies [18,20,25,27] including 207 participants compared the combination of LDP treatment plus conventional therapies with conventional therapies. The trials showed homogeneity in the consistency of the trial results (chi-square = 1.20; = 0.75; 2 = 0%). Therefore, a fixed-effect model should have been used for statistical analysis. A meta-analysis showed a significant beneficial effect of LDP plus conventional therapies compared with conventional therapies in decreasing the level of 24 h UTP (SMD: −0.67 [−0.95, −0.38]; < 0.00001) ( Table 8). One study [32] including 120 participants compared the combination of LDP plus telmisartan plus routine treatment with telmisartan plus routine treatment. The homogeneity in the consistency of the trial results is not applicable ( = 11.90; < 0.00001). A meta-analysis showed a significant beneficial effect of LDP treatment plus telmisartan plus routine treatment compared with telmisartan plus routine treatment in decreasing the levels of 24 h UTP (SMD: −3.48 [−4.05, −2.90]; < 0.00001) ( Table 8).
(4) Improvement of UmAlb. Five studies [16,19,22,24,29] including 321 participants used the levels of UmAlb to measure the outcome. Three studies including 202 participants [16,24,29] compared the combination of LDP treatment plus conventional therapies with conventional therapies. The trials showed homogeneity in the consistency of the trial results (chi-square = 0.66; = 0.72; 2 = 0%). Therefore, a fixed-effect model should have been used for statistical analysis. A meta-analysis showed a significant beneficial effect of LDP plus conventional therapies compared with conventional therapies in decreasing the level of UmAlb (SMD: −1.37 [−1.68, −1.06]; < 0.00001) ( Table 9). One study [19] including 44 participants compared the combination of LDP treatment plus Perindopril plus routine treatment with Perindopril plus routine treatment. The homogeneity in the consistency of the trial results is not applicable ( = 7.76; < 0.00001). A meta-analysis showed a significant beneficial effect of LDP treatment plus Perindopril plus routine treatment compared with Perindopril plus routine treatment in decreasing the levels of UmAlb (SMD: −4.70 [−5.89, −3.52]; < 0.00001). Another study [22] including 75

Publication
Bias. The number of trials was too small to conduct any sufficient additional analysis of publication bias.

Adverse Effects.
In total, five of the eighteen trials mentioned the presence or absence of adverse effects [18,22,24,31,32]. Two trials reported the incidence of adverse events, whereas 3 trials reported that no adverse events occurred in the treatment groups compared with the control groups. One trial [22] reported slight cough ( = 1) in the treatment 8 Evidence-Based Complementary and Alternative Medicine group. Another trial [18] indicated that the treatment group presented with hypoglycemia ( = 1) and gastrointestinal tract reaction ( = 1) and that the control group presented with hypoglycemia ( = 2) and gastrointestinal tract reaction ( = 2). In general, adverse events were rare and minor.

Discussion
There is no specific treatment of diabetic nephropathy. Currently, comprehensive treatments including diet management, control of blood pressure, control of blood sugar, and lipid adjusting treatment have been commonly used. Some corresponding measures taken to intervene actively against some risk factors may reduce proteinuria and delay the occurrence of albuminuria. It is considered that active treatment can prevent or delay the progress of diabetic nephropathy at any stage, especially in the early period. ACEI and ARB drugs are widely used in the treatment of DN, but they are not enough to delay or retard the progression of DN. TCM has lots of advantages over the conventional medical approaches in the prevention and treatment of DN [9,10]. Based on the meta-analyses of the outcome, LDP may have positive effects for lowering FBG, PBG, and HbA1c and improving renal function. LDP as an adjunctive treatment to conventional drugs significantly improved blood sugar and renal function in patients with DN. However, due to the low-quality methodology and potential publication bias, we cannot draw a definite conclusion of the beneficial effectiveness of LDP combined with Western drugs in preventing and treating diabetic nephropathy. Our positive findings should be interpreted conservatively.
Firstly, in terms of the current evaluative standards, all of the included studies were of low methodological quality. Bias may exist in many areas, such as an unclear method of random sequence, no mention of allocation concealment, and being double-blind. It is also possible that bias was produced by the link of distribution, implementation and measurement, statistics, and reporting.
Secondly, five of the eighteen trials mentioned the presence or absence of adverse effects. The three trials only mentioned that there were no adverse effects in their study. The other two trials reported the presence of adverse effects. One trial [22] reported slight cough ( = 1) in the treatment group. Another trial [18] indicated that the treatment group presented with hypoglycemia ( = 1) and gastrointestinal tract reaction ( = 1) and that the control group presented with hypoglycemia ( = 2) and gastrointestinal tract reaction ( = 2). The safety of LDP needs to be rigorously monitored in future clinical trials. Therefore, conclusions about the safety of LDP cannot be made from this study due to the limited and inadequate evidence in included trials. Large-scale clinical trials with long-term follow-up appear warranted.
Thirdly, none of the eighteen included trials mentioned health related quality of life, economic index, or compliance with treatments. A recent investigation of 20 public hospitals and integrative medical hospitals in Beijing demonstrated that $4 million can be saved in medical expenses if prescriptions of Chinese herbal medicine were increased by 1% [33].
In view of this finding, we should pay attention to the health economics indices of LDP treatment.

Conclusions
The results suggest that Liuwei Dihuang Pills added to other routine treatment have a therapeutic potential for people with diabetic nephropathy. However, due to the poor quality of included studies, the reported effectiveness and safety of Liuwei Dihuang Pills for diabetic nephropathy cannot be taken as confirmative conclusion. More rigorous RCTs will be needed to present a high level of evidence for the effectiveness of Liuwei Dihuang Pills in treating diabetic nephropathy.