Herbal Medicines for Treating Metabolic Syndrome: A Systematic Review of Randomized Controlled Trials

Objective. The aim of this systematic review is to evaluate the efficacy and safety of herbal medicines in the management of metabolic syndrome. Materials and Methods. On December 9, 2015, we searched PubMed, EMBASE, Cochrane Library, SCOPUS, AMED, CNKI, KoreaMed, KMBASE, OASIS, and J-STAGE with no restriction on language or published year. We selected randomized controlled trials that involved patients with metabolic syndrome being treated with herbal medicines as intervention. The main keywords were “Chinese herbal medicines”, “metabolic syndrome”, and “randomized controlled trials”. Herbal substances which were not based on East Asian medical theory, combination therapy with western medicines, and concurrent diseases other than metabolic syndrome were excluded. The risk of bias was assessed by Cochrane's “Risk of Bias” tool. The protocol or review was registered in PROSPERO (an international prospective register of systematic reviews) (CRD42014006842). Results. From 1,098 articles, 12 RCTs were included in this review: five trials studied herbal medicines versus a placebo or no treatment, and seven trials studied herbal medicines versus western medicines. Herbal medicines were effective on decreasing waist circumference, blood glucose, blood lipids, and blood pressure. Conclusion. This study suggests the possibility that herbal medicines can be complementary and alternative medicines for metabolic syndrome.


2
Evidence-Based Complementary and Alternative Medicine metabolic syndrome are abdominal obesity and insulin resistance [5,6]. Therefore, preventing atherosclerotic cardiovascular disease by controlling waist circumference and insulin resistance is the key to managing metabolic syndrome. Clinically, each treatment of hyperglycemia, hypertension, and hyperlipidemia is prescribed according to each patient's state.
However, ACE inhibitor that is a drug used for hypertension, including enalapril and captopril, may cause adverse events such as cough, increased serum creatinine, headache, and skin rash [7,8]. It has also been shown that metformin, a drug used to treat type 2 diabetes mellitus, can induce gastrointestinal symptoms and lactic acidosis [6]. Therefore, herbal medicines showing evidence of safety and efficacy can be alternative treatments for metabolic diseases.
Although there are several reviews of herbal medicines for obesity [9,10], hypertension [11], and type 2 diabetes mellitus [12], systematic review for metabolic syndrome has not been conducted yet. This study, however, reviews not only a single disease but also metabolic syndrome as a whole. The aim of this study is to evaluate the efficacy and safety of herbal medicines to help manage metabolic syndrome.

Data Source and Search Strategy
2.1.1. Data Source. This study included the following databases: PubMed, EMBASE, Cochrane Library, SCOPUS, AMED, China National Knowledge Infrastructure (CNKI), KoreaMed, KMBASE, OASIS, Electronic (J-STAGE), and Japan Science and Technology Information Aggregator.

Search Strategy.
The study used herbal medicine, metabolic syndrome, and randomized controlled trials for the basic search terms. A search strategy in PubMed is shown in Table 1. Language and publication date were not restricted. The date for the search was December 9, 2015. This review's protocol was registered in PROSPERO (an international prospective register of systematic reviews) (registration number: CRD42014006842).

Study Design. Randomized controlled trials (RCTs)
were included regardless of blinding. Other designs such as in vivo, in vitro, case report, and retrospective study and thesis were excluded.

Participants.
Participants were patients with metabolic syndrome and there was no restriction to sex or age. Diagnosis criteria of metabolic syndrome were restricted to international or national standard criteria such as AHA/NHLBI, National Cholesterol Education Program-Adults Treatment Panel (NCEP-ATP), and International Diabetes Federation (IDF) and should be clearly described in Section 2. Chinese pattern identification was optional. Concurrent diseases with metabolic syndrome were excluded.

Interventions. Single or mixed herbal medicines with
East Asian medical theory were included. Traditional Chinese Medicine, Traditional Korean Medicine, and Japanese Kampo Medicine are regarded as East Asian medicine. Ayurvedic medicine, crude plant, food, and dietary supplement were excluded. Combination therapy with western medicines, acupuncture, and moxibustion was also excluded. Exercise, diet-control, and health education were not restricted if they were applied to both intervention and control groups.

Comparisons.
There was no special restriction on comparisons. Placebo, no treatment, active-control, exercise, diet-control, health education, and usual care were allowed as control groups. Active-control means western medicines for metabolic syndrome, or herbal medicines other than intervention.

Outcome Measures.
The primary outcome measures were WC, FPG, TG, HDL-C, SBP, and DBP which are clinical parameters of metabolic syndrome. Secondary outcome measures included body mass index (BMI), body weight, hemoglobin A1c, 2-hour postprandial glucose (2 hPG), total cholesterol (TC), low density lipoprotein cholesterol (LDL-C), and clinical effective rate. Trials that assessed one or more outcome measures were included. However, trials which measured level of hormone or enzyme, such as leptin and adiponectin, were excluded in this review.

Study Selection and Data Extraction
2.3.1. Selection of Literature Articles. After excluding any duplication of literature reviews from 10 databases, two authors (S. Jang and J.-S. Park) reviewed titles and abstracts for the first exclusion. Then, full texts of the selected literature articles were subject to another review before the final selection of literature articles was made to make sure each article qualified using the inclusion criteria for this study. For excluded literature articles, the reason for exclusion was recorded. When two authors showed a difference of opinion, a third author (B.-H. Jang) intervened to help come to an agreement. The entire process was displayed by generating a flow diagram in PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) ( Figure 1).

Data Extraction.
One author (S. Jang) conducted data extraction, and a different author (Y. Ko) reviewed the data. Items extracted from each trial include the following: (1) general characteristics of the study: author, published year, language, and country; (2) participants: sample size, sex, and age, Chinese pattern identification; (3) interventions: intervention, compositions of intervention, formulation of intervention, control, dosage, and medication period; and (4) outcomes: outcomes, main conclusion, and adverse events.

Assessment with Risk of Bias.
Two authors (S. Jang and Y. Ko) assessed methodological quality using the Risk of Bias (RoB) tool, which was developed by Cochrane [7]. RoB was divided into 6 selection biases, including 2 selection biases Evidence-Based Complementary and Alternative Medicine 3 (random sequence generation and allocation concealment), performance bias, detection bias, attrition bias, and reporting bias. Each item of all the included RCTs was determined as "high risk," "unclear," or "low risk." A RoB graph was drawn using RevMan 5.3 program.

Data
Analysis. We used mean difference (MD) with 95% confidential interval (CI) to measure primary outcomes between trials. Analyses were divided into 4 subgroups depending on type of controls: no treatment, placebo, metformin, other western medicines. Heterogeneity was analyzed by the Cochrane and 2 test. 2 values of 25%, 50%, and 75% mean low, medium, and high levels of statistical heterogeneity. RevMan 5.3 program was used for analysis. We also made summary of findings (SoF) table to present results of review by Gradepro software.

Description of Included
Trials. From ten databases, 1,098 literature articles were identified. Among them, 826 records remained after eliminating duplications, and 733 records were excluded after screening titles and abstracts. By reviewing full texts of 93 records, 12 RCTs were included in this systematic review. The process of the study selection is shown in Figure 1.

3.1.2.
Participants. The number of participants for the trials varied from 43 [14] to 183 [13]. No trial was restricted to participants based on sex, but one trial [17] did not report sex distribution. Six trials included Chinese pattern identification as inclusion criteria: 2 trials [15,19] of Exuberance of Phlegm-Dampness Type, 1 trial [24] of Spleen Deficiency and Stagnation of Dampness Type, 1 trial [18] of Heart-Liver Stagnated Heat Type, 1 trial [20] of Flaming-Up of Fire of the Liver Type, and 1 trial [17] of Blood-Stasis Type. Two trials [14,20] were conducted on patients with hypertension and metabolic syndrome.

Comparisons.
Comparisons were divided into two types. One type was a placebo or no treatment, and 2 trials [13,14] were compared with a placebo, while 3 trials [15][16][17] were conducted under no treatment. Diet-control and exercise or health education was used for the no treatment group. The other type was western medicine, and 7 trials [18][19][20][21][22][23][24] followed this comparison type. Metformin was used as a comparison in 3 trials [18,19,21], and nifedipine was used in 1 trial [20]. Two trials [22][23][24] provided different conventional medicines according to the symptoms of the patients.

Effects of Interventions.
Overall efficacy of FPG, TG, SBP, DBP, WC, and HDL-C was presented in summary of findings (Table 3). Meta-analyses of FPG, TG, SBP, and DBP were shown in Figures 2-5. All 12 trials had different herbal medicine interventions; therefore, we also compared effects of each intervention. Table 4 shows the mean differences (MD) for each outcome measure. The unit mmol/L was converted into mg/dL.    The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MD: mean difference. 1 Heterogeneity and possible publication bias downgraded quality of the evidence. 2 Sparse data downgraded quality of the evidence. 3 Heterogeneity downgraded quality of the evidence.
Evidence-Based Complementary and Alternative Medicine 9
The mean SBP was 6.76 lower in the intervention groups compared to control groups within 11 trials (−7.72 to −5.81) (Figure 4). The mean DBP was 5.23 lower in the intervention groups than control groups within 11 trials (−5.77 to −4.86) ( Figure 5). [13,14,[20][21][22]24] reported 26 adverse events. Nine cases occurred in the herbal medicine group, and the remaining 17 cases occurred in the western medicine group. There was no adverse event in the placebo or no treatment control group. The Dia-No group [13] had 6 upper digestive disorders, the Huanglian Wendan decoction group [21] had 1 gastrointestinal disorder, and the Yiqi Huaju Recipe group [14] had 2 skin hypersensitivities. There was no observed adverse event in the Pinggan Jiangya pill group [20], the Xueguan Ruanhua decoction group [22], and the Shenling Jianpihuashi decoction group [24]. The most commonly reported symptoms were digestive disorders such as nausea, vomiting, and burning of the epigastrium (Table 5).

Assessment with Risk of Bias.
RoB of the 12 RCTs was assessed into 6 areas. Six RCTs [13][14][15][16][17][18] used a random number table to generate the random sequence. There were high risks of performance bias (blinding of participants and personnel) in 10 trials [15][16][17][18][19][20][21][22][23][24] due to the difference of drug formulation. Except in 2 RCTs [17,20] where the primary outcomes were blood pressure and blood-stasis symptom, the remaining 10 trials were assessed as "low risk" for detection bias (blinding of outcome assessment). There were low risks of attrition bias (incomplete outcome data) and reporting bias (selective reporting) in all 12 trials. Only 1 RCT [13] was assessed as "low risk" for all six items. Details of RoB are presented in Figure 6.

Discussion
As a result of searching 10 databases, 12 randomized controlled trials were included in the systematic review. Because the review's purpose was to determine the efficacy and safety of herbal medicines for metabolic syndrome, clinical trials that included herbal medicines combined with conventional western medicines were excluded. Five trials [13][14][15][16][17] studied herbal medicines versus a placebo or no treatment, and seven trials [18][19][20][21][22][23][24] studied herbal medicines versus western medicines. All 12 trials included controls for diet-control, exercise, or health education with medications.
According to Table 4, all the trials showed positive effects with the administration of herbal medicines, and most of them proved significant. Gegen Shanzha decoction [16] improved 5 metabolic indexes, including WC, FPG, TG, HDL-C, and BP. Yiqi Huaju Recipe [14] had an effect on lowering body weight, blood sugar, and blood pressure (except blood lipids). Because the outcome measured the blood-stasis symptom only in the trial for Xuefu Zhuyu decoction [17], the efficacy of Xuefu Zhuyu decoction on metabolic diseases could not be determined. The Pinggan Jiangya pill [20] was effective for the metabolic syndrome of Flaming-Up of Fire of the Liver Type. Flaming-Up of Fire of the Liver Type is the largest type of hypertension [25]; therefore, the Pinggan Jiangya pill would be suitable to treat obesity and hypertension. The Qinggan Jiangtang tablet [18], Modified Banxia Baizhu Tianma decoction [19], Huanglian Wendan decoction [21], Xueguan Ruanhua decoction [22], the Shengjiangtongmai powder [23], and Shenling Jianpihuashi decoction [24] were superior (or not inferior) to western medicines used to treat metabolic syndrome.
As for summary of findings (Table 3), mean differences of metabolic parameters were compared. Metabolic syndrome is not determined by single indicator, and comparing value of each parameter is not appropriate, strictly speaking. Relative risk (RR) of metabolic syndrome should be calculated. However, there was no study presenting difference of prevalence before and after treatment. Meanwhile, WC and HDL-C could not be calculated because they were not separated by men and women within included trials.
Forest plots of FPG, TG, SBP, and DBP indicate high heterogeneity although subgroup analysis was done ( Figures  2-5). It is assumed that heterogeneity did not result from controls. Instead, difference of each intervention would have been affected. It is also limitation of meta-analysis in this review.
Regarding safety, adverse events were reported less in herbal medicines than in western medicines (Table 5). In Zhang et al. 's trial [20], the nifedipine group had 4 facial flushes, while the Pinggan Jiangya group had no adverse events. Additionally, there was 1 gastrointestinal disorder with Huanglian Wendan decoction, but there were 5 adverse events when metformin was used in Guan et al. 's trial [21]. Therefore, herbal medicines would be an effective and safe Evidence-Based Complementary and Alternative Medicine 13   treatment for metabolic syndrome compared with western medicines. In 6 trials [15,[17][18][19][20]24], oriental pattern identification along with metabolic factors was set for inclusion criteria. A selection of treatments with pattern identifications would help reduce metabolic risk factors, improve general conditions, and decrease chances of adverse events. With the collection of such trial data, this would provide a ground for herbal medicines to be used as treatment for obesity or metabolic syndrome by pattern identifications.
There are more studies on herbal medicines for metabolic syndrome although they were not included in this review. Keishibukuryogan [26], Yiqi Huaju Qingli Formula [27], Ba-Wei-Wan [28], Heqi San [29], Baoling decoction [30], and Combination of Four Gentlemen Decoction and Sini Powder [31] showed effects on metabolic syndrome; however, trials on these were excluded because treatment group was also treated with conventional western medicines. Herbal supplements, for example, Ginseng [32], berberine, bitter melon [33], nigella sativa [34], and Gymnema sylvestre [35], are also used for management of metabolic diseases. In particular it is well known that ginsenosides which are compounds of ginseng have clear effect of regulating blood glucose and blood pressure [33]. This review has some limitations. First, metabolic syndrome is not a single disorder but rather a complex disease. The herbal medicines used as interventions and their efficacy need to be matched for the following indicators: waist circumference, body weight, blood glucose, blood lipids, and blood pressure. Second, the trials included in the study showed a relatively low level of quality because most of them failed to conduct a double-blinded technique, and only 1 trial [13] met the qualification for advanced protocol. Because of these limitations, there may be the possibility that therapeutic effects have been overestimated. Publication bias also needs to be taken into consideration. Third, there could be trials missing even though we tried to cover all of the RCTs from English, Korean, Chinese, and Japanese databases. However, the study is significant because it has reviewed RCTs on the administration of herbal medicines for treating metabolic 16 Evidence-Based Complementary and Alternative Medicine diseases. Further studies are needed to develop new herbal medicines for metabolic syndrome and to build evidence on their effectiveness and safety.

Conclusion
Herbal medicines showed therapeutic effects on regulating waist circumference, blood glucose, blood lipids, and blood pressure in this systematic review. This means herbal medicines have the potential to be complementary and alternative medicines for metabolic syndrome. However, more high quality trials are needed to prove the efficacy and safety of herbal medicines.