Electroacupuncture for Gastrointestinal Function Recovery after Gynecological Surgery: A Systematic Review and Meta-Analysis

Background Evidence for the efficacy and safety of electroacupuncture (EA) on gastrointestinal function recovery after gynecological surgery is unclear. Objective This meta-analysis aimed to evaluate the effects of EA on recovery of postoperative gastrointestinal function for patients receiving gynecological surgery. Data sources: PubMed, Cochrane Central Register of Controlled Trials (CINAHL), Embase, China National Knowledge Infrastructure (CNKI), Weipu (CQVIP), and Wanfang databases were systematically searched from the inception dates to May 30, 2020, for relevant randomized controlled trials (RCTs). Study selection: RCTs that evaluated EA for postoperative gastrointestinal function directly related to gynecological surgery in adults aged 18 years or over. Data extraction and synthesis: paired reviewer independently extracted the data and assessed study quality. Standardized mean differences (SMD) were calculated as the effect measure from a random effects model. Main outcomes and measures: time to first flatus (TFF), time to bowel sounds recovery (TBS), and time to first defecation (TFD) were recorded as primary outcomes; postoperative nausea and vomiting (PONV), motilin (MTL), gastrin (GAS), pH value of gastric mucosa (pHi), gastric mucosal partial pressure of carbon dioxide (PgCO2), vasoactive intestinal peptide (VIP), and adverse event were reported as secondary outcomes. Results We included eighteen RCTs (1117 participants). Our findings suggested that compared to the control group (CG), electroacupuncture group (EG) showed significant effects on TFF (SMD = −0.98, 95% CI: [−1.28, −0.68], P < 0.00001, I2 = 69%), TBS (SMD = −0.98, 95% CI: [−1.84, −0.12], P=0.03, I2 = 92%), and TFD (SMD = −1.23, 95% CI: [−1.59, −0.88], P < 0.0001, I2 = 0%). Moreover, the incidence of PONV at postoperative 6 h (OR = 0.42, 95% CI: [0.27, 0.64], P < 0.0001, I2 = 0%) and 24 h (OR = 0.46, 95% CI: [0.32, 0.68], P < 0.0001, I2 = 0%) was lower in the EG than that in the CG, whereas no significant difference in ratio of PONV at postoperative 48 h (OR = 0.55, 95% CI: [0.20, 1.51], P=0.25, I2 = 0%) was detected between the two groups. Meanwhile, there was a significant effect in favor of EA on the level of MTL at postoperative 6 h (SMD = −0.93, 95% CI: [−1.36, −0.61], P < 0.0001, I2 = 21%), while no significant effect was observed at postoperative 24 h (SMD = −0.43, 95% CI: [−0.89, 0.02], P=0.06, I2 = 69%) in the EG when compared to the CG. Additionally, a large significant effect on decreasing PgCO2 was found in the EG in comparison to the CG, but no significant effect in favor of EA on GAS, VIP, or pHi was observed. It was reported that there was one participant with pain at the needling sites and bruising, and three participants withdrew because they were not intolerant to EA. Conclusions EA could be a promising strategy for the prevention and treatment of gastrointestinal dysfunction after gynecological surgery, including shortening TFF and TFD, TBS, regulating MTL, and decreasing the ratio of PONV within postoperative 24h. The effects on MTL and PONV varied with different intervention points, and EA used at 30 min prior to surgery might be recommended. However, the evidence quality ranged from low to very low, and large-scale and high-quality RCTs were warranted.


Introduction
Postoperative recovery of gastrointestinal function was considered as one of the most important parts for the rehabilitation after surgery, which was a condition that mainly related to surgical stress, anesthesia regimen, surgical treatment, and postoperative analgesia method [1]. e short-term gastrointestinal dysfunction after surgery was referred to as postoperative ileus (POI), which could cause undesirable consequences, including abdominal distension, lack of flatus and defecation, and nausea and vomiting. As was reported, POI would influence postoperative experience, increase the length of hospital stay, and even raise the risk of morbidity and mortality, which not only strongly prevented rehabilitation after surgery, but also posed a substantial economic burden on family and society [2].
POI was frequently observed after gastroenterological surgery, and it was followed by gynecological surgery [3]. It was reported that the prevalence of POI ranged from 5% to 25% among patients undergoing gynecological surgery [4], and even the incidence was up to 50% in patients receiving surgical therapy for gynecological cancers [5]. In recent years, minimally invasive surgery, such as laparoscopy surgery, has become more commonly used for gynecologic indications [6]. However, the incidence of gastrointestinal dysfunction following surgery is not decreased [7]. Even though increasing studies focusing on improving recovery of postoperative gastrointestinal function have been conducted, strategies with satisfactory efficacy are still rarely reported, especially for gynecological population.
us, apart from improving anesthesia regimen and surgical technique, so far it is still urgent and essential to develop an effective and safety method of promoting the return of gastrointestinal function for patients receiving gynecological surgery.
Acupuncture, as one of the conventional Chinese medical therapies, has been applied to promote gastrointestinal function for thousands of years [8]. Electroacupuncture (EA) is a modified technique involving traditional acupuncture and electrical stimulation to achieve a greater response, which has been proved to be a promising approach to reduce complications and accelerate rehabilitation after surgery in the field of orthopedic, abdominal, and gynecological diseases. Moreover, there are an increasing number of studies focusing on EA for treatment and prevention of POI. But the results concerning the efficacy of EA on POI are inconsistent [9,10]. It was reported in the previous meta-analyses [11][12][13][14] that supported the benefits of EA/acupuncture to POI, but the efficacy of EA on POI for gynecological patients still lacked evidence basis. Gynecological surgery was unlike gastroenterological surgery or other surgery in terms of pathology and surgical methods, which might lead to difference outcomes of gastrointestinal function recovery after surgery. Consequently, the purpose of this study was to evaluate the safety and efficacy of EA on postoperative recovery of gastrointestinal function for the patients receiving gynecological surgery, which could provide new evidence for promoting rehabilitation after gynecological surgery.

Methods
is study was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [15]. Moreover, the protocol of this study was registered in PROSPERO, and the registration number is CRD42021260096. In the process of retrieval, we found that most of the literatures were published in Chinese, so we retrieved with increased numbers of Chinese database.
However, due to the delay of preliminary preparation time, the deadline for literature retrieval was extended to May 30, 2021. We had submitted the amendments in the registration system.

Search Strategy.
In order to identify relevant studies, we systematically search the electronic databases including PubMed, Cochrane Central Register of Controlled Trials (CINAHL), Embase, China National Knowledge Infrastructure (CNKI), Weipu (CQVIP), and Wanfang from inception to May 30, 2021. e search terms, such as "postoperative ileus," "postoperative gastrointestinal motility disorder," "postoperative gastrointestinal function recovery," "postoperative gastrointestinal dysfunction," "postoperative gastrointestinal function," "postoperative nausea and vomiting," and "electroacupuncture," were used to search in each database without language or disease restrictions. e search strategy was described in Supplementary Figure 1.

Inclusion Criteria.
We formulated inclusion and exclusion criteria based on PICOS (patients, intervention, comparator, outcomes, and study design) approach [

Data Extraction.
e relevant data were independently extracted from the included studies by the two reviewers, including study characteristics (e.g., author names, publication year, study design, and sample size), participant characteristics (e.g., age, year), intervention type, intervention characteristics (e.g., acupoints, model, and intensity), and outcomes. During the process of screening and data extraction, any discrepancies would be resolved by discussion, or consultation with a third reviewer (Y. Guo) until a consensus was reached.
2.6. Quality Assessment. In accordance with the Cochrane Collaboration's Risk of Bias tool V 2.0, the two reviewers assessed the quality of the included literatures from seven dimensions: bias arising from the randomisation process, bias from deviations from the intended interventions, bias from missing outcome data, bias due to measurement of the outcome, bias from the selection of the reported results, and overall risk of bias [16]. e risk of each item is divided into three levels: high, unclear, and low. Meanwhile, the two researchers evaluated the methodological quality of included studies according to the Physiotherapy Evidence Database (PEDro) scale [17], which includes 10 items for assessment of trial quality from various aspects, including randomisation procedure, concealed allocation, similar baseline, patients blinding, therapists blinding, assessors blinding, adequate follow-up (dropout rate <15%), intention to treat analysis, between-group statistical analysis, and point and variability measures; the total score ranged from 0 to 10. According to the total score, quality of study was categorized into three degrees, including high (10 ≥ PEDro score ≥ 6), fair (6 > PEDro score ≥ 4), and low (0 ≤ PEDro score ≤ 3) [18].

Statistical Analysis.
We conducted the data analysis by using review manager (version 5.3, the Nordic Cochrane Centre, Copenhagen, Denmark) and Stata (version 13.0, the StataCorp LP, TX, USA). For continuous variables, standardized mean differences (SMDs) and 95% confidence intervals (95% CI) were calculated as the effect measure by using a random effects model, and standard mean effect sizes were divided to different categories, including no change (0), small effect (0.2), moderate effect (0.5), and large effect (0.8) [18]. Enumeration data was analyzed by using a random effects method and expressed as the odds ratios (OR) and 95% CI. All the corresponding meta-analysis results were illustrated by the forest map intuitively. Cochran's Q-test and I 2 index were employed to estimate heterogeneity [19], and I 2 statistic greater than 50% was considered as substantially heterogeneous. When a substantially heterogeneous was observed, subgroups analysis or sensitivity analyses were used to determine the risk factor resulted in the high heterogeneity. Grading of recommendations, assessment, development, and evaluation (GRADE) approach was utilized to evaluate the quality of evidence [20]. In addition, publication bias was assessed by Begg's and Egger's tests [19], and the visual inspection of funnel plots would be applied if the index was reported in more than 10 included studies. e significant difference level was set at P < 0.05.

Risk of Bias in Included Studies.
e detailed results are depicted in Figure 2. All the included studies were reported as random generation, and four of the RCTs [21,22,29,38] were Evidence-Based Complementary and Alternative Medicine conducted with concealed allocation. Participants were blinded to the group allocation in four studies [21,22,29,38], and assessment blinding was reported in seven RCTs [21,22,24,27,29,36,38]. Meanwhile, based on the evaluation of PEDro, all the scores of the included studies ranged from 5 to 9 (mean score + standard deviation � 6.61 ± 1.42; Supplementary Table 1), which indicated that the methodological quality of included studies was fair to high. All the included studies were described as random generation, while blinding was recorded only in four trials [21,22,29,38]. Additionally, none of the studies reported a follow-up rate of more than 15%, and intention-to-treat analysis was not found in any of the included studies.
(8) Adverse Event. One participant in one study [21] reported pain at the needling sites and bruising, while the side effects were alleviated spontaneously without any medical assistance. In addition, it was reported in two studies [30,38] that three participants withdrew because they were not intolerant to EA.

Sensitivity Analysis.
In this study, we conducted the sensitivity analysis to evaluate the robustness of primary outcome indicators with high heterogeneity by removing studies from the analysis individually. After sensitivity analysis for TFF, no substantial change about overall heterogeneities and results was detected (Supplementary Table 2). Meanwhile, we performed sensitivity analysis for TBS, and the results were listed in Supplementary Table 3. e overall effect indicated no statistically significant difference   Table 4). Moreover, visual inspection of funnel plots seemed to be relatively symmetrical for TFF and PONV (Figure 13), indicating no evidence for publication bias.

Evidence Quality Assessment According to GRADE.
Based on GRADE guidelines, we evaluated the quality of evidence from five aspects, including risk of bias, inconsistency, indirectness, imprecision, and publication bias. During the treatment period, EA therapists directly contacted to participants, which led to the high risk of performance bias, and consequently lowered the evidence quality. In this study, there was low evidence in TFF, TFD and PONV, and very low evidence in the remaining. Details were recorded in Supplementary Table 5.

Discussion
In spite of continuous improvement for surgical approach and technique, postoperative gastrointestinal dysfunction still commonly occurred after gynecological surgery. To resolve this clinical concern, EA has been proposed to promote gastrointestinal function recovery, while relevant evidence is sparse. To the best of our knowledge, this is first systematical review and meta-analysis regarding the efficacy and safety of EA on recovery of gastrointestinal function for patients receiving gynecological surgery. We found that EA was able to shorten the time to first flatus, first defecation after gynecological surgery. ese findings were in accordance with those reported by the previous meta-analyses [11][12][13][14] for other patient populations, but there were some new findings after subgroup analysis in the current study. Our findings suggested that EA was beneficial to reduce the time to first of bowel sounds recovery when EA was used at 30 min prior to surgery. However, no significant effect on TBS was observed when patients undergoing gynecological surgery received EA treatment after surgery. Additionally, the evidence in favor of EA for decreasing the ratio of PONV was observed only at postoperative 6 h and 24 h, which echo the results reported by Lee et al. [39], while the benefits were unobvious at postoperative 48 h. Similarly, the effect of EA on MTL was detected only within postoperative 6 h. us, our findings showed that the benefits of EA for PONV and MTL concentrate within 24 h after surgery, and even earlier. MTL stimulated gastrointestinal motility and accelerated gastric emptying, which could induce or aggravate nausea and vomiting [40,41].
us, it could be the reason that EA decreased MTL level and consequently resulted in lowering the ratio of PONV at with postoperative 24 h. After subgroup analysis for PONV by intervention points, we found that EA presented a largest effect in lowering the ratio of PONV when it was applied at 30 min prior to surgery, whereas no or very small effect was found when patients were treated with EA after gynecological surgery. It might be the reason that factors contributing to postoperative gastrointestinal dysfunction mainly generated during intraoperative period; it should be more proper to applied EA before the start of surgery until the end of surgery, while the effectiveness of EA would be weakened when used too early, and postoperative application seems to be late. PgCO 2 is a sensitive indicator reflecting the changes of blood oxygen in gastric mucosa [42]. It was proved that pHi was closely related with intestinal oxygen consumption, and decreased gastric pHi level prognosticated morbidity and mortality [43]. e results of this meta-analysis revealed that EA deceased the PgCO 2 both at 30 min after the start of pneumoperitoneum and 30 min after the end of pneumoperitoneum, whereas it was ineffective for mediation of pHi after surgery. Additionally, GAS was the hormone primarily responsible for gastric acid secretion [44], and VIP was a gut peptide hormone regulating gut motility [45]; hence, they were associated with postoperative gastrointestinal function [46]. However, EA showed no significant effect on GAS and VIP after gynecological surgery. erefore, EA accelerated the recovery of postoperative gastrointestinal function through other pathways, not depending on the mediation of GAS or VIP.
Several potential limitations should be considered in the present review had. Firstly, the EA parameters including acupoints, and times and frequency of electroacupuncture were selected without a consolidated standard in the included studies, which could be a potential source of clinical heterogeneity; secondly, blind methods of the included studies were rarely provided in detail; thirdly, as was reported in the previous basic experimental studies, EA mediated of the autonomic nervous system, improved dysmotility and local inflammation, and consequently ameliorated POI to restore gastrointestinal function [47], and the therapeutic effects was different when using lower limb and abdomen acupoints [48]. However, few clinical trials focus on examining the different effect of EA by using different acupoints.
Given the limitations of this work, large-scale RCTs with more rigorous and robust methods are still needed in future studies. Most importantly, the majority of the studies were not blind to participants and acupuncturists; sham-controlled studies should be performed to avoid performance bias. Hence, more studies should be conducted strictly following standard reporting guidelines such as CONSORT Evidence-Based Complementary and Alternative Medicine 13 [49]. Furthermore, more RCTs in the future should investigate the therapeutic effects of EA using different acupoints in order to identify the definitely effective acupoints.

Conclusion
In this analysis, we systematically reviewed and quantified the effect of EA on gastrointestinal function after gynecological surgery. Overall, EA was an effective and safe treatment for promoting recovery of postoperative gastrointestinal function, such as shortening TFF and TFD, TBS, regulating MTL, and decreasing the ratio of PONV within postoperative 24h, for patients receiving gynecological surgery through abdominal and laparoscopic approaches, while the effects on MTL and PONV varied with different intervention points, and EA used at 30 min prior to surgery might be recommended. Moreover, EA could regulate PgCO 2 during anesthesia process, which was associated with the recovery of postoperative gastrointestinal function. However, EA exerted no significant impact on mediating GAS, VIP, and pHi. us, EA could be a promising strategy for the prevention and treatment of gastrointestinal dysfunction after gynecological surgery. Notably, evidence quality ranged from low to very low; large-scale and high-quality RCTs were needed.
Data Availability e original data analyzed in this study are presented in the article; further inquiries should be directed to the corresponding authors.