Acupoint Catgut Embedding for Insomnia: A Meta-Analysis of Randomized Controlled Trials

Objectives A Meta-analysis was carried out to evaluate the efficacy and safety of acupoint catgut embedding (ACE), a procedure of embedding sutures made of absorbable materials into the skin tissue of acupoints, on insomnia. Methods Relevant clinical randomized controlled trials (RCTs) were comprehensively searched from eleven electronic databases (up to 1 March 2020). Two authors independently screened literature, extracted data, and assessed the risk of bias of included studies. Stata 12 and RevMan 5.3.0 software were used for meta-analysis. PyCharm 2019 and Gephi software (version 0.9.2) were used for complex network analysis. Results Thirty-four RCTs involving 2,655 patients were included. The meta-analysis suggested that ACE induced a better clinical efficacy compared with that in the estazolam tablets (EZ) group (RR = 1.22, 95% CI: 1.13, 1.31) or in the acupuncture (ACU) group (RR = 1.21, 95% CI: 1.14, 1.28) and could significantly reduce the score of Pittsburgh Sleep Quality Index (P < 0.05). ACE resulted in better long-term efficacy compared to that in the EZ group (RR = 1.87, 95% CI: 1.58, 2.22) and ACU group (RR = 1.30, 95% CI: 1.14, 1.48). ACE could significantly reduce the incidence of adverse events (RR = 0.30, 95% CI: 0.15, 0.60) compared with that in the EZ group. Complex network analysis indicated that acupoints of BL23, SP6, PC6, BL15, BL20, BL18, and HT7 were the core acupoints selected in ACE for insomnia. Conclusion The clinical efficacy of ACE for insomnia is better than that of other interventions (EZ and ACU) in both short-term and long-term observations. Considering the efficacy and reduced visits to the clinic by ACE, the present study provides a practical and convenient complementary and alternative therapy for insomnia. This trial is registered with PROSPERO CRD 42020169866.


Introduction
Sleep disorders, especially insomnia, have become an emerging global pandemic. World Health Organization estimates that more than 1/3 of the population worldwide are sleep deprived [1]. Epidemiological studies show that approximately 10% to 23% of adolescents and 15% to 40% of adults worldwide experience different degrees of insomnia with an increased prevalence among the elderly, females, and people with medical or psychiatric disorders [2][3][4][5][6][7][8][9]. In China, 38.2% of adults suffer from insomnia, which is well above the international level that the World Health Organization reports for adults worldwide, and more than 300 million Chinese suffer from sleep disorders [10].
Studies have found that insomnia is an important risk factor for various diseases, including cardiovascular diseases, depression, diabetes, a weakened immune system, suicidal tendencies, and other mental disorders [11][12][13][14][15]. Furthermore, insomnia also increases the socioeconomic burden with more than $100 billion a year lost due to poor workplace performance and more traffic and fall accidents [16,17].
Fortunately, insomnia is a treatable sleep disorder and a modifiable risk factor for a variety of illnesses [18]. At present, treatment of insomnia mainly includes drug therapy, nondrug therapy, or a combination of both. However, drug therapies have adverse effects such as headache, dizziness, withdrawal reaction, bitter mouth, forgetfulness, lethargy, falls, and hangover [19]. ese side effects may increase the risk of developing cancers such as oral cancer, liver cancer, and breast cancer [20]. e firstline nondrug therapy is cognitive behavioral therapy for insomnia (CBT-I) [21], but it may not be a suitable treatment for everyone [22]. It is usually practiced by a licensed psychologist trained in behavioral sleep medicine or by a master-level clinician. One of the main obstacles to the expansion of CBT in the treatment of insomnia is the lack of trained clinicians and the availability of experts outside urban centers [18]. Acupuncture, defined as the use of an acupuncture needle by hand manipulation, has been proved to be an effective treatment for insomnia [23][24][25][26]. However, it has some disadvantages, such as it usually needs to be performed once a day or every other day [27].
Acupoint catgut embedding (ACE) refers to the procedure of embedding sutures made of absorbable materials into the skin tissue of acupoints which are closely related to different physiological processes or diseases [27]. It is a combination of traditional acupuncture and modern tissue therapy [28]. As a form of acupuncture, ACE can persistently stimulate acupoints for a week or longer until sutures are absorbed through softening, liquefaction, and absorption [28]. Its advantages are the long duration of acupoint stimulation, fewer times of treatment and less expense, short treatment cycle, and being acceptable for patients, etc. It reduces the frequency of patients seeking medical treatment and saves medical resources to a certain extent. It is more convenient and easier to be performed than traditional acupuncture. erefore, it is used to treat insomnia in China, and its efficacy for insomnia has been proved in previous clinical trials [29]. At the same time, a meta-analysis [30] found that acupoint catgut embedding is a potential alternative therapy for insomnia compared to estazolam tablets by analyzing results from 7 randomized controlled trials (RCTs). However, only Chinese databases were searched in their study. Another study [31] analyzed the effective rate of 22 RCTs, but not the Pittsburgh Sleep Quality Index (PSQI) score, and did not take different treatment methods in the control group into account. erefore, it is necessary to confirm the efficacy of ACE for insomnia with more rigorous methodologies. Besides, a few questions remain to be answered: Is ACE better than drugs such as benzodiazepines? Is ACE better than traditional acupuncture? Is ACE is safer than other types of treatment for insomnia? What are the core acupoints to perform ACE for insomnia? e present study aimed to answer these questions by conducting a comprehensive updated meta-analysis.  Table S1). Equivalent searching terms were employed in Chinese databases. A manual search was conducted for eligible studies listed in references of certain studies.

Inclusion and Exclusion
Criteria. Studies meeting the following criteria were included in the meta-analysis. (a) RCTs were included without restriction about whether using blind methods. (b) Individuals were diagnosed with insomnia, regardless of the subtype of insomnia, gender, age, duration of the disease, ethnicity, and education status; e diagnosis of insomnia needs to be consistent with the international classification of diseases or other guidelines for the diagnosis and treatment of insomnia in China [33][34][35][36]. (c) e intervention group received ACE while the control group received placebos, medications, sham ACE, no treatment, or acupuncture/electro-acupuncture. (d) e main outcome measures included clinical therapeutic effect and PSQI score.
Exclusion criteria included (a) studies that compare the effect of different catgut lengths, catgut placing procedure, or acupoints for ACE; (b) studies in which catgut embedding acupoints were not clearly reported in the intervention measures; (c) studies with incomplete or questionable data; and (d) repeated publications or duplicate data.

Study Selection and Data Extraction.
Two authors (WL and ZL) independently retrieved and reviewed all the eligible studies and extracted data. Disagreements between two authors were resolved by discussions or consulting a third author (HZ). e following information was extracted from the included studies and expressed in an excel table: research details (name of the first author, title, and year of publication), participant details (age, course of the disease, diagnostic and exclusion criteria), interventions (sample sizes, random sequence generation, allocation concealment, blinding methods, selective reporting and other biases, interventions in treatment and control groups, acupoints for ACE, duration of treatment, frequency of ACE and interventions of the control group, number of ACE, and followup visits), and primary and secondary outcomes (clinical therapeutic effect, PSQI scores, sleep quality (SQ) scores, fall asleep time (FAT) scores, sleep time (ST) scores, sleep efficiency (SE) scores, sleep disorder (SD) scores, daytime dysfunction (DD) scores, hypnotic drug (HD) scores, and adverse events).
We also considered the possible different types of missing data. First, we contacted the authors of the included studies to obtain the missing data. Second, we dealt with missing studies (and the associated risk of bias) by assessing for publication bias through the nonparametric trim-and-fill methods. e studies which did not report on either of our primary outcomes were excluded. e studies with missing data were also excluded for those we did not receive a corresponding response from study authors.

Quality Assessment.
Two authors (HZ and YW) independently used the Cochrane Collaboration's risk of bias tool [37] to evaluate each included study, and differences between the two were resolved through discussion with a third author (WL). e following domains were assessed: randomization sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other biases. According to the quality assessment tool, each study was rated as low risk, high risk, or unknown risk, respectively.

Statistical
Analysis. Stata 12 and RevMan 5.3.0 software were used for meta-analysis. PyCharm 2019 and Gephi (version 0.9.2) were used for complex network analysis. e dichotomous data used RR and 95% confidence interval (CI) as the statistical indicator. e standardized mean difference (SMD) and 95% CI were used for continuous data.
e Forest plot was built by considering P < 0.05 as statistically significant. Statistical heterogeneity was measured using the chi-squared test and I 2 statistics. For the chisquared test, P < 0.1 indicated statistical significance. And the I 2 values of 25%, 50%, and 75%, respectively, denoted cut-off points for low, moderate, and high levels of heterogeneity [38]. If a low or moderate statistical heterogeneity was found (P > 0.1, I 2 < 50%), the fixed-effect model was employed. When significant heterogeneity was observed in the studies (P < 0.1, I 2 ≥ 50%), a random-effect model was used to analyze the data. Sensitivity analysis, univariate metaregression analysis, and subgroup analysis were carried out to explore sources of potential heterogeneity and examine the robustness of primary results [39]. A publication bias analysis was performed when more than 10 studies were pooled. Publication bias was measured with a funnel plot and Egger's test [40]. e nonparametric trim-and-fill methods [41] were utilized to further evaluate the potential publication bias.

Searching Results and Study Selection.
A total of 705 studies were initially identified. After a series of procedures (removal of duplicate publications, screening of titles and abstracts, and assessment of full text), 34  articles were finally included for analysis. e searching and selecting process was shown in Figure 1.

ACE versus EZ
(1) Primary outcomes (a) Clinical erapeutic Effect. 21 studies involving 1440 patients were included for analysis .

Sensitivity
Analysis. Sensitivity analysis was used to identify the potential sources of heterogeneity and to examine the stability of the quantitative synthesis results. In the leave-one-out analysis by omitting one study each time, the overall combined results of clinical therapeutic effect (ACE versus EZ) and PSQI score (ACE versus EZ/ACU) did not change substantially. ese indicated that the synthetic results were robust and no single study had a significant impact on the overall results (Supplementary file 2: Figure  S9).

Metaregression Analysis and Subgroup Analysis.
Univariate metaregression analysis and subgroup analysis were used to analyze the internal factors of these studies (mean age, mean course of the disease, duration of treatment, frequency of intervention, and the number of intervention). It was found that the above factors were not the main reason for the heterogeneity of these studies in the clinical therapeutic effect (ACE versus EZ) and PSQI score (ACE versus EZ/ACU) (all P values >0.05; Supplementary file 2: Figures S10-S27).

Publication
Bias. Significant asymmetry was found upon visual inspection of the funnel plots of clinical therapeutic effect (ACE versus EZ/ACU) and the funnel plot of PSQI score (ACE versus EZ) (Figure 4). And significant publication bias was confirmed by Egger's test (all P values <0.05; Table 3). e sensitivity analysis using the trim-andfill method was carried out and produced symmetrical funnel plots for them. And the corrections for potential publication bias did not alter the significant associations ( Figure 4 and Table 3). It can be considered that the conclusions of the meta-analysis were robust and reliable.

Core Acupoints of ACE for Insomnia.
A total of 29 acupoints of ACE for insomnia were extracted from the 34 RCTs. Complex network analysis showed that acupoints of BL23, SP6, PC6, BL15, BL20, BL18, and HT7 were the core acupoints of ACE for insomnia according to the degree centrality and betweenness centrality ( Figure 5; Supplementary file 1: Table S4). ACE was often performed once every 10 days with a duration of 1 month (Supplementary file 1: Table S2).

Discussion
As an alternative form of traditional acupuncture, acupoint catgut embedding (ACE) has advantages in the treatment of insomnia [76]. e present study has some findings. Firstly, ACE has better efficacy in managing insomnia and reducing the scores of PSQI than estazolam tablets (EZ) or acupuncture (ACU) in both short-term and long-term observations (Tables 1 and 2). Secondly, ACE is safer than EZ for insomnia. irdly, the core acupoints of ACE for insomnia are BL23, SP6, PC6, BL15, BL20, BL18, and HT7 ( Figure 5). Lastly, ACE can reduce the visits to the clinic compared with traditional acupuncture.
Neurophysiological studies have proved that the sleepwake circadian rhythm is maintained by the normal activities of participating structures including the central nervous system, endocrine system, and immune system [77]. A variety of neurotransmitters are involved in the regulation of the sleep-wake cycle, such as monoamines, amino acids, cytokines, choline, hormones, and skin substances [78,79]. ACE, acupuncture, and estazolam tablets can regulate the balance of neurotransmitters, inhibit awakening and excitement, promote sleep, and regulate the rhythm of the sleep-wake cycle [80,81]. ACE and acupuncture also can regulate sleep by regulating the content of sleep-related cytokines, which suggests that both methods can take effect through multidirectional and multitarget pathways [82]. However, it is worth noting that the use of hypnotics is not to cure insomnia but to induce patients to sleep. at is, taking hypnotics before going to bed can only help patients sleep a few hours, and when they need to sleep again, they need to take the medicine again [82]. In addition, only short-term use of benzodiazepines is recommended [83]. According to the result of adverse events (ACE versus EZ), the number of adverse events (such as pain, dizziness, nausea, dry mouth, and allergy) due to ACE was lower than that in the EZ group. is may be due to the fact that EZ is a kind of benzodiazepine. e use of benzodiazepine is always associated with higher mortality and multiple adverse events, such as drug abuse and dependence, rebound insomnia, amnesia, lung diseases, muscle relaxation, and vulnerability to fracture [83,84]. Although ACE may also have adverse events such as low fever, subcutaneous induration, and rashes, these can be prevented or cured without special treatments [85]. us, ACE is superior to EZ in its efficacy for insomnia, long-term effect, and safety.
Previous studies have shown that acupuncture can regulate the sleep-wake cycle and play a sedative and hypnotic role in promoting sleep by affecting a variety of neurotransmitters, immunomodulators, hormones, skin substances, and nitric oxide [86]. Many chemicals such as 5-HT, NE, DA, GABA, Glu, IL-1, IL-6, melatonin, TNF, PGE2, and NO are closely related to the sedation and hypnosis effects of acupuncture [82]. Mechanisms of acupoint catgut embedding in the treatment of insomnia are similar to traditional acupuncture. ACE increases the contents of 5-   Evidence-Based Complementary and Alternative Medicine HT, IL-1, TNF-α, and decreases the contents of DA and NE in the hypothalamus [81,87]. It is found that the efficacy of ACE for insomnia is superior to traditional acupuncture and has a better long-term effect, which may be related to the therapeutic characteristics of ACE. ACE keeps stimulating acupoints with needles and catgut, which leads to a stronger and long-lasting therapeutic effect than traditional acupuncture. erefore, it is suitable for chronic and intractable diseases such as insomnia. Catgut, an allogenic protein, is gradually softened and absorbed by the body, and this longlasting effect can better regulate the relative balance of the internal environment and improve the regulatory ability of the body than traditional acupuncture. ACE overcomes the drawbacks of traditional acupuncture, such as short duration of therapeutic effect, difficulty in consolidating the therapeutic effect, ease to relapse, and poor patient compliance.
Acupoints used in these 34 RCTs were in different combinations, but there were seven acupoints that were the core acupoints to perform ACE for insomnia. Stimulation of BL23, SP6, PC6, BL15, BL20, BL18, and HT7 increases the expression of c-aminobutyric acid and c-aminobutyric acid A receptor in the hypothalamus, through which the quality of sleep is increased [88][89][90]. However, molecular biological mechanisms of acupoint catgut embedding in the treatment of insomnia need to be further studied due to the slight difference between ACE and ACU.
Subjective questionnaires are the main tools used to assess insomnia. Among them, PSQI is the most widely used questionnaire across different age groups in both nonclinical and clinical settings worldwide [91,92]. PSQI has 19 individual items for the assessment of subjective sleep quality, which can be aggregated into seven components that evaluate key aspects of sleep [93]. erefore, in the present  Figure 5: e core acupoints of ACE for insomnia analyzed by a complex network. is network contains 29 nodes and 137 edges. e higher the degree centrality value of the acupoint is, the larger the node is, the larger the font is, the redder the color of the node is, and the more important it is in the network. e thicker the edge is, the more frequently the acupoints are used together. study, the PSQI score and clinical therapeutic effect are primary outcomes. e seven components of PSQI are secondary outcomes.

Comparison with Previous Studies.
Our findings were generally consistent with results of previous reviews written by Liu et al. [30] and Zhang et al. [31], which also showed that ACE has certain advantages in the treatment of insomnia. Besides, the present study further supplements findings in a few important respects. Firstly, previous reviews did not perform any univariate metaregression analysis and subgroup analysis. We did them to explore whether the results were affected by some confounding factors. e consistent results of sensitivity, metaregression, and subgroup analysis show that our findings are robust and reliable. Secondly, the present meta-analysis included 12 new RCTs with larger sample sizes and added about twice the number of patients of previous studies, which significantly enhanced the capability of confirming the clinical efficacy of ACE in the treatment of insomnia. irdly, the present study not only compares ACE and EZ but also ACE and ACU. e studies mentioned in the present paper are more comprehensive, covering more comparisons and providing more useful information through thorough statistical analyses. e present study analyzed not only the clinical therapeutic effect and the PSQI score but also the SQ, FAT, ST, SE, SD, DD, and HD scores as well as adverse events. Fourthly, the present study analyzed not only the immediate results after treatment but also the results at a one-month followup. More importantly, compared with previous studies, the odds ratio (OR) was modified to the risk ratio (RR) in the present study. As a result, our risk estimation is more accurate and reliable. In addition, it is worth noting that this study also explored the core acupoints to perform ACE for insomnia through complex network analysis, which showed that BL23, SP6, PC6, BL15, BL20, BL18, and HT7 are the core catgut embedding acupoints for insomnia.

Limitations.
ere are also a few limitations to this meta-analysis. Firstly, the included RCTs were mainly conducted in China reporting short-or medium-term outcomes. erefore, there is a lack of international RCTs and long-term outcomes. Secondly, differences were found in acupoint selection and operation methods between RCTs on testing ACE, which may lead to clinical heterogeneity. For some assessments, answers to questionnaires were used to test the effective rate of ACE. is method has a strong subjectivity and lacks data on objective physiological and chemical measurements. irdly, there was a lack of comparison between ACE and placebo or between ACE and cognitive-behavioral therapies. ese limitations may influence the reliability of our meta-analysis results.
Based on our findings, we suggest that future multicenter, large sample, randomized, double-blind, and longterm RCTs are necessary to confirm the efficacy of ACE in managing insomnia.

Conclusions
In summary, our study has three findings. (1) Acupoint catgut embedding (ACE) is more effective than estazolam tablets (EZ) or acupuncture (ACU) in the treatment of insomnia and has a better long-term efficacy based on the assessment of the pooled results (clinical therapeutic effect, reduction of PSQI, SQ, ST, SE, SD, DD, and HD scores). (2) ACE is safer than EZ for insomnia. (3) BL23, SP6, PC6, BL15, BL20, BL18, and HT7 are the core acupoints of ACE for insomnia. Considering the increasing prevalence worldwide and the advantages of enhanced efficacy and reduced visits to the clinic by ACE, the present study provides a practical and convenient alternative therapy for insomnia.

Data Availability
All data used in this study are included in this article.

Conflicts of Interest
e authors declare that they have no conflicts of interest.

Authors' Contributions
LX and WL conceived the study. WL and ZL searched and checked the databases according to the inclusion and exclusion criteria. LX and HC helped to develop search strategies. WL and ZL extracted the data. HZ and YW assessed the quality of the included studies. WL analyzed the data and drafted the first version of the manuscript. All authors contributed critically to subsequent revisions and approved the final manuscript. LX is the guarantor of this study.
Acknowledgments is work was supported by grants from the 2019 Key Specialties of Medicine in Shanghai (ZK2019B13).

Supplementary Materials
Supplementary file 1. Table S1: searching strategy for PubMed search. Table S2: basic characteristics of included studies in the present study. Table S3: quality assessment of included studies. Table S4: topological structure value of ACE acupoints for insomnia. Figure S1: quality assessment of included studies (risk of bias summary). Supplementary file 2. Figure S1: the forest plot of clinical therapeutic effect (one-month follow-up). Figure S2: the forest plot of reduction of PSQI score (one-month follow-up). Figures S3 and S4: forest plots of reduction of SQ, FAT, ST, SE, SD, DD, and HD scores (ACE versus EZ). Figure S5: the forest plot of reduction of SQ, FAT, ST, SE, SD, and DD scores (onemonth follow-up) (ACE versus EZ). Figures S6 and S7: forest plots of reduction of SQ, FAT, ST, SE, SD, DD, and HD scores, respectively (ACE versus ACU). Figure S8: the forest plot of reduction of SQ, FAT, ST, SE, SD, and DD scores (one-month follow-up) (ACE versus ACU). Figure S9: the forest plot of sensitivity analysis. Figure S10: the graph of univariate metaregression analysis (ACE versus EZ: clinical therapeutic effect). Figures S11-S15: forest plots of subgroup analysis by mean age, mean course of disease, duration of treatment, frequency of intervention, and number of intervention, respectively (ACE versus EZ: clinical therapeutic effect). Figure S16: the graph of univariate metaregression analysis (ACE versus EZ: reduction of the PSQI score). Figures S17-S21: forest plots of subgroup analysis by mean age, mean course of disease, duration of treatment, frequency of intervention, and number of intervention, respectively (ACE versus EZ: reduction of the PSQI score). Figure S22: the graph of univariate metaregression analysis (ACE versus ACU: reduction of the PSQI score). Figures  S23-S27: forest plots of subgroup analysis by mean age, mean course of disease, duration of treatment, frequency of intervention, and number of intervention, respectively (ACE versus ACU: reduction of the PSQI score). (Supplementary Materials)