Effectiveness of Electroacupuncture for Simple Obesity: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Objective To evaluate the effectiveness of electroacupuncture in the treatment of simple obesity. Methods Randomized clinical trials concerning electroacupuncture as a treatment of simple obesity published prior to October 31, 2019, were searched in the following Chinese and English databases: Chinese National Knowledge Infrastructure (CNKI), WanFang Database, China Science and Technology Journal Database (VIP), Chinese Biomedical Literature Database (CBM), PubMed, Cochrane Library, Web of Science, and Scopus. After data collection and quality evaluation, meta-analysis was performed using RevMan 5.3 software and Stata 15.0 software. Results A total of 13 studies involving 937 patients with simple obesity were included in the meta-analysis. Results revealed that the total effective rate (RR = 1.29, 95% CI [1.13, 1.48]; P=0.0002), BMI (MD = −1.82, 95% CI [−2.21, −1.43]; P < 0.000), waist circumference (MD = −2.39, 95% CI [−3.95, −0.84]; P=0.003), hip circumference (MD = 0.31, 95% CI [−2.37, 2.99]; P=0.82), waist-hip ratio (MD = −0.05, 95% CI [−0.07, −0.03]; P < 0.00), and body fat rate (MD = −1.56, 95% CI [−2.35, −0.78]; P=0.0001) in the electroacupuncture group were superior to those in the control group. Analysis of acupoint clustering and correlation using SPSS 24.0 and Clementine 12.0 revealed the highest statistical support for acupoint groups CV12-CV4 and CV12-ST25-CV4, while ST36-CV12-ST25, SP6, and ST40-ST24-SP15-ST37-CV4 were found to be validly clustered acupoints. Conclusion For treating simple obesity, electroacupuncture is superior to other interventions such as acupuncture, acupoint catgut embedding therapy, and simple lifestyle modification for improvement in body fat rate, waist circumference, and waist-hip ratio, although not hip circumference. Acupoint analysis revealed that ST25, CV12, CV4, SP6, and ST36 can form the basis for electroacupuncture therapy for the treatment of simple obesity.


Introduction
Recent decades have seen soaring obesity worldwide, in no small part due to lifestyle changes. According to the WHO, 13% of the world's adult population (11% of men and 15% of women) were obese in 2016, and worldwide obesity has nearly tripled since 1975 [1,2]. Simple obesity refers to the absence of obvious neurological and endocrine etiologies and is primarily due to either excessive caloric intake or suboptimal energy consumption.
is results in excess Furthermore, obesity has been found to affect the incidence and prognosis of various carcinomas [7][8][9]. Current treatments of obesity mainly include behavioral therapy, drugs, and surgery, with the latter two modalities especially possessing a large number of side effects. As one case study reported, use of certain weight-loss pharmacotherapies may affect control of HIV viral load [10]. Since obesity greatly impacts people's life and work habits and currently has limited treatment options available, researching safe and effective interventions with low rates of adverse reactions is of great significance. Acupuncture, a form of traditional Chinese medicine (TCM), is widely accepted as an effective treatment for simple obesity due to its lack of significant side effects and relatively good tolerability. Electroacupuncture, a form of acupuncture where a small electric current is passed between pairs of acupuncture needles, is a common variant of this therapeutic modality. Studies have shown that electroacupuncture combined with treadmill exercise significantly increases expression of PGC-1α, restores healthier levels of fatty acid oxidation, and improves skeletal muscle mitochondrial function, thus contributing to weight loss [11]. Animal studies have suggested that electroacupuncture can reduce weight by decreasing the expression of IL-6 [12], changing the expression of irisin [13], and inhibiting food intake when electroacupuncture treatment at acupoint ST36 upregulated anorexigenic proopiomelanocortin production in the solitary and hypoglossal nuclei [14]. Although the aforementioned studies provided noteworthy evidence regarding the treatment of obesity by electroacupuncture, we searched for randomized clinical trials (RCT) concerning such a treatment regimen in eight databases and performed a systematic review as well as meta-analysis in order to report a more advanced objective validation of available data.

Study Selection.
e included studies satisfied the following criteria: (1) clinical controlled trial (CCT) and RCT studies with at least 10 cases of simple obesity treated with electroacupuncture; (2) studies including patients diagnosed with simple obesity; (3) studies in which treatment group patients underwent electroacupuncture at any acupoints, were treated with any needling technique, and were treated for any duration, while control group patients were treated with any other regimen; (4) studies with full text available.
Exclusion criteria were as follows: (1) conference papers, reviews, and studies with control groups treated by electroacupuncture and those with more than two study groups; (2) studies with obvious errors or major shortcomings; (3) studies with unavailable full text.

Data Extraction.
Author(s), publication year, sample sizes, interventions, randomization methods, blinding methods, baselines, diagnostic criteria, evaluation standards of therapeutic effects, inclusion and exclusion criteria, follow-up periods, dropout or withdrawal rates, adverse events, and safety indices were extracted from included literature by two individuals.

Quality Assessment.
Bias risks of included studies were assessed according to the Cochrane Handbook version 5.1.0. Assessment included six aspects: (1) selection bias, (2) performance bias, (3) detection bias, (4) attrition bias, (5) selective reporting, and (6) other biases. Each aspect was evaluated and classified into three groups: (1) low risk (bias did not significantly affect results), (2) high risk (bias significantly impacted result credibility), and (3) unclear risk (the possibility of bias affecting results was suspected). Study quality was assessed by two evaluators; any disagreement was discussed with a third reviewer.

Statistical
Analyses. RevMan 5.3 software and Stata 15.0 software were used for data analysis and publication bias assessment by detailing forest and funnel plots. e weighed mean difference (MD) for measured data and relative risk (RR) for enumerated data were used to assess study heterogeneity. If there was low heterogeneity between trials (P > 0.10; I 2 <50%), a fixed effects (FE) model was used; otherwise, a randomized effects (RE) model was applied. e RR with a 95% confidence interval (CI) was calculated; P values less than 0.05 were considered to be statistically significant.

General Literature Description.
Our search strategy resulted in a total initial extraction of 721 papers from all of the aforementioned eight databases, of which 387 were in Chinese (158 from CNKI; 137 from the WanFang Database; 42 from VIP, and 50 from CBM) and 334 were in English (68 from PubMed; 72 from the Cochrane Library; 54 from Web of Science, and 140 from Scopus). After ruling out duplication and study failure to meet our inclusion criteria, 13 papers, of which 13 were in Chinese ultimately included in our analysis (Figure 1). Table 2 details characteristics of included studies. Altogether, the 13 papers we ultimately analyzed studied 937 participants, 470 of which were in an electroacupuncture group and 467 of which were in a control group. One paper compared the efficacy of electroacupuncture with that of acupoint catgut embedding [15], five with basic treatment [17,19,20,23,24], six with regular acupuncture [16,18,21,22,26], and one with sham electroacupuncture [25]. In all of these studies, 33 acupoints were used to treat simple obesity: Zhongwan (CV12), Qihai  Table 3). Support was defined as the ratio of literature mentioning the antecedent acupoints to that of all included (1) Study did not concerned with simple obesity or electroacupuncture (n = 268) (2) Study evaluated more than three groups (n = 30) (3) Study involved therapeutic regimens combining electroacupuncture with other modalities (n = 156) (4) Studies in which the control group was treated with electroacupuncture (n = 49) (5) Study was a conference paper or review or animal experiment of literature (n = 152) (6) Full-text study was unavailable (n = 34) (n = 689)

Study Characteristics.
Studies excluded after being deemed as duplicate reports (n = 20) Full-text articles assessed for ultimate inclusion (n = 13) Full-text articles excluded due to lack of a control group and/or lack of complete data (n = 7) Evidence-Based Complementary and Alternative Medicine   Table 3, for example, states that 40% of the 13 papers mentioned CV4 as a treatment acupoint. Confidence was defined as the accuracy and credibility of any given association; there thus was a 100% probability of papers mentioning CV12 as a treatment acupoint also mentioning CV4 as one. Greater confidence and support values revealed stronger association among acupoints. As seen in Table 3, acupoint groups with the highest support of 40% were CV12-CV4 and CV12-ST25-CV4. SPSS 21.0 software was used to analyze the clustering of acupoints selected to treat simple obesity by electroacupuncture at least five times. e icicle plot of Figure 2 details that there were three effective clusters. ey were ST36-CV12-ST25, SP6, and ST40-ST24-SP15-ST37-CV4. e dendrogram of

Quality of Included Studies.
All 13 studies ultimately included in our analysis were of RCT and CCT design. Among them, six studies [15-17, 22, 25, 26] were determined to be at low risk for bias with use of a random number table; one study [23] with an odd/even random number and another [21] with the sequence of seeing doctors were determined to be at high risk for bias. Five studies [18][19][20][21]24] that did not describe randomization methods were determined to be at unclear risk for bias. Only one study [25] describing allocation concealment was determined to be at low risk for bias while the remaining 12 not describing allocation concealment were considered to be at unclear risk. Two studies [22,25] that reported blinding of both participants and personnel were determined to be at low risk for bias; the remaining studies not mentioning blinding methods were considered to be at unclear risk. Only one study [25] reported blinding of personnel to outcomes and was determined to be at low risk for bias; the remaining studies not mentioning such measures were considered to be at unclear risk.
ree articles [21,23,25] that described dropout and loss to follow-up rates, as well as causes for these phenomena, and two [15,17] not reporting such conditions were determined to be at low risk for bias. e remaining eight studies [16, 18-22, 24, 26] that did not describe outcome data in detail were considered to be at unclear risk. Although all fourteen studies described results, as primary statistical materials were not available, all 13 trials were determined to be at unclear risk for bias. e baseline between electroacupuncture and control groups in all trials was considered to be at low risk ( Figure 4).
(2) Sensitivity Analysis. Sensitivity analysis was performed to identify the source of high heterogeneity. As shown in Figure 5, data from a prior study conducted by Chen et al.   [20] revealed negligible overlapping of confidence intervals with data from another eight studies, suggesting data heterogeneity. As shown in Figure 6, one study's data was distributed outside the 95% confidence interval, likewise suggesting data heterogeneity. After excluding the study by Chen et al., heterogeneity decreased (P � 0.026; I 2 � 21%), as shown in Figure 7. In the detailed review of Chen et al.'s study in its entirety, we found that the data heterogeneity was primarily influenced by variation in studied parameters and differing intervention measures. Differences in studied parameters included varying subject inclusion and exclusion criteria or the differences caused by study sizes; disparities in intervention measures included data discrepancies caused by variations in patient compliance.

Reduction of BMI.
Nine trials [15-17, 21-23, 25, 26] mentioned BMI in their outcomes and meta-analysis revealed high heterogeneity among trials (P < 0.000; I 2 � 80%). As illustrated in Figure 9, electroacupuncture was more effective than other interventions among control group patients in reducing BMI (pooled MD � −1.82, 95% CI [−2.21, −1.43]; P < 0.000). Due to high heterogeneity, subgroup analysis was conducted in terms of intervention ( Figure 10) and treatment course ( Figure 11). Heterogeneity was found to remain unaltered, although no source for it was identified.

Reduction of Waist and Hip Circumference.
Waist circumference (WC) was mentioned in the outcomes of six studies [15,16,[21][22][23] and low heterogeneity was revealed in subsequent meta-analysis (P � 0.13; I 2 � 41%). Results revealed that electroacupuncture was superior to other interventions among control group patients in the reduction of waist circumference, as shown in Figure 12 Figure 13.

Improvement of TCM Syndrome
Score. Statistical analysis of four articles [15,21,22] reporting TCM syndrome scores revealed high heterogeneity (P � 0.007; I 2 � 75%) and a significant difference among electroacupuncture and control group patients in the improvement of TCM syndrome scores (pooled MD � −2.38, 95% CI [−3.96, −0.81]; P � 0.003). As illustrated in Figure 15, electroacupuncture was more effective than other interventions in control group patients.

Discussion
Obesity, one of the most serious conditions harming public health over the world, is the main etiology of type 2 diabetes mellitus and cardiovascular disease [27]. Unfortunately, to date, no safe and effective pharmacotherapy has been found to treat obesity [28]. erefore, study of novel and alternative  therapeutic modalities is of great importance. Acupuncture, a form of TCM, has been previously reported effective in treating obesity [29][30][31][32]. Furthermore, recent years have witnessed an increasing number of clinical studies concerning the treatment of obesity using electroacupuncture [33][34][35]. Electroacupuncture was found to promote fat metabolism via sympathetic nervous system activation, thus decreasing appetite [36]. To confirm the effectiveness of this therapy as a treatment of simple obesity and help establish a foundation for clinical guideline formation, we conducted this systematic review and meta-analysis evaluating published RCT and CCT data available in major scientific article databases. irteen trials were included in analysis after rigorous screening. Meta-analysis was employed to analyze response rates reported in nine trials, BMI in nine, WC in six, HC in four trials, WHR in two, TCM syndrome scores in four, and body fat rate in seven. e results revealed electroacupuncture to be more effective than other therapies in improving the TCM syndrome score (P � 0.003) and overall response rate (P � 0.0002) as well as reducing BMI (P < 0.000), WC (P � 0.003), WHR (P < 0.00), and the body fat rate (P � 0.0001). However, no superiority of electroacupuncture in the reduction of HC was noted (P � 0.82). Overall response rate analysis revealed high heterogeneity (P � 0.001; I 2 � 69%). Detailed, individual analysis of nine trials revealed that high heterogeneity was likely due to a study by Chen et al. [20]; heterogeneity decreased (P � 0.26; I 2 � 21%) after exclusion of this study. In addition, there was high heterogeneity (P < 0.000; I 2 � 89%) among trials in regard to BMI reduction. To identify the cause of this phenomenon, subgroup analysis was performed, revealing no source for the heterogeneity and   Evidence-Based Complementary and Alternative Medicine that it remained unchanged in terms of intervention (P < 0.000; I 2 � 87%) and treatment course (P � 0.0001; I 2 � 78%). Overall analysis of 13 studies revealed electroacupuncture to be more effective than acupuncture, acupoint catgut embedding, and basic treatment in the therapeutic management of simple obesity. We also studied the association among main acupoints selected for treatment. Acupoint groups with the highest statistical support were found to be CV12-CV4 and CV12-ST25-CV4, followed by CV6-SP15, CV6-CV12-SP15, ST26-ST24, ST26-ST25-ST24, ST24-ST26, ST24-ST25-ST26, SP15-CV6, SP15-CV12-CV6, ST26-ST25-SP15, ST26-CV12-ST25-SP15, ST24-ST25-SP15, ST24-CV12-ST25-SP15, CV6-ST25-SP15, and CV6-CV12-ST25-SP15. Clustering analysis revealed three effective clusters; namely, ST36-CV12-ST25, SP6, and ST40-ST24-SP15-ST37-CV4. Our findings also suggested that ST25-CV12-CV4-SP6-ST36 serves as the primary acupoint group for effective electroacupuncture treatment of obesity, with ST24, ST26, ST37, SP15, and ST40 serving as supplementary acupoints capable      Evidence-Based Complementary and Alternative Medicine 11 of treating different TCM syndromes associated with simple obesity. is study did not lack shortcomings with the potential to limit our above conclusions. As all included trials were written in Chinese and no articles in other languages were initially searched for, the possibility of publication bias must be considered. All RCT data containing negative results were excluded; this could have potentially caused overestimation of response rates and given rise to result bias. Only one included study employed allocation concealment and two   Heterogeneity: tau 2 = 0.00; chi 2 = 3.07, df = 6 (P = 0.80); I 2 = 0% Test for overall effect: Z = 3.89 (P = 0.0001)
employed blinding, implying that the majority of studies we evaluated were low in quality. Inclusion of low-quality data compiled using nonstandardized diagnostic and evaluation criteria may have likewise led to publication bias. When high heterogeneity was detected, sensitivity and subgroup analyses were successful in only partially revealing sources of heterogeneity.
Rigorous study design including characteristics such as allocation concealment, adequate blinding, employment of randomization, and consideration of adverse effects, as well as detailed discussion of negative results, should be encouraged in future electroacupuncture research. More detailed information regarding electroacupuncture therapy, such as needling sites, insertion depths, needle types, and acupoints selected are likewise of great importance in the future establishment of clinical electroacupuncture therapy guidelines.

Conclusion
In the treatment of simple obesity, electroacupuncture is superior to acupuncture, acupoint catgut embedding, and basic treatment in improving the body fat rate, BMI, WC, WHR, and TCM syndrome scores, but not HC. Acupoint analysis revealed ST25, CV12, CV4, SP6, and ST36 to be most effective in the treatment of obesity with electroacupuncture.

Conflicts of Interest
e authors declare that there are no conflicts of interest.

Authors' Contributions
Yanling Gao and Yi Wang contributed equally to the design and conception of the study. e manuscript was drafted by Yanling Gao and Jing Zhou and reviewed by Yin Shi. e search strategy was improved by all authors and performed by Yi Wang and Zhihai Hu, who also independently screened the potential studies. Jing Zhou and Yi Wang extracted data from the included studies, and Yanling Gao assessed the risk of bias and finished data synthesis. Zhihai Hu arbitrated any discrepancies and ensured that no errors occur during the study. Besides, all authors have read and approved the final version of the manuscript.