Pregnancy Benefit of Acupuncture on in vitro Fertilization: A Systematic Review and Meta-Analysis

Currently, more and more infertility couples are opting for combined acupuncture to improve success rate of in vitro fertilization (IVF). However, evidence from acupuncture for improving IVF pregnancy outcomes remains a matter of debate. To quantitatively summarized the evidence of the efficacy of acupuncture among women undergoing IVF by means of systematic review and meta-analysis. Four English (PubMed, Web of Science, EMBASE, and Cochrane Register of Controlled Clinical Trials) and Four Chinese databases (Wanfang Databases, Chinese National Knowledge Infrastructure, Chinese Science and Technology Periodical Database, and SinoMed) were searched from database inception until July 2, 2023. Randomized controlled trials (RCTs) that evaluated the acupuncture’s effects for women undergoing IVF were included. The subgroup analysis was conducted with respect to the age of participants, different acupuncture types, type of control, acupuncture timing, geographical origin of the study, whether or not repeated IVF failure, and acupuncture sessions. Sensitivity analyses were predefifined to explore the robustness of results. The primary outcomes were clinical pregnancy rate (CPR) and live birth rate (LBR), and the secondary outcomes were ongoing pregnancy rate and miscarriage rate. Random effects model with I2 statistics were used to quantify heterogeneity. Publication bias was estimated by funnel plots and Egger’s tests. A total of 58 eligible RCTs representing 10,968 women undergoing IVF for pregnant success were identifified. Pooled CPR and LBR showed a signifificant difference between acupuncture and control groups [69 comparisons, relative risk (RR) 1.19, 95% confifidence intervals (CI) 1.12 to 1.25, I2=0], extremely low evidence; 23 comparisons, RR 1.11, 95%CI 1.02 to 1.21, I2=14.6, low evidence, respectively). Only transcutaneous electrical acupoint stimulation showed a positive effect on both CPR (16 comparisons, RR 1.17, 95%CI 1.06 to 1.29; I2=0, moderate evidence) and LBR (9 comparisons, RR 1.20, 95%CI 1.04 to 1.37; I2=8.5, extremely low evidence). Heterogeneity across studies was found and no studies were graded as high-quality evidence. Results showed that the convincing evidence levels on the associations between acupuncture and IVF pregnant outcomes were relatively low, and the varied methodological design and heterogeneity might inflfluence the fifindings. (Registration No. PROSPERO CRD42021232430)

(9) However, as the most expected outcome for patients, evidence from acupuncture for improving IVF pregnancy outcomes remains a matter of debate.
Since the first randomized clinical trial (RCT) published in 1999, (10) numerous RCTs and concomitant meta-analysis have emerged to explore the impact of acupuncture on IVF outcomes.However, the conclusions were inconclusive.Both positive (3,(11)(12)(13)(14)(15)(16)(17) and negative (18)(19)(20)(21)(22)(23)(24)(25)(26) effects of acupuncture assisting IVF success were reported.The individualized treatment protocol, different control types, variate acupuncture timing, different searching databases, literature selection criteria, and outcome measures, all potentially contributed to the confl icting results. (9,27)Since most current trials used the original protocol of Paulus, et al, (28) the main infl uencing factor across acupuncture studies thereafter might be the types of acupuncture.For example, studies that evaluated only a few acupuncture methods would cause distinct results. (15,17,24))(19) Otherwise, most meta-analysis only considered clinical pregnancy rate (CPR) as a primary outcome but neglected the live birth rate (LBR).But for infertile couples, the outcome of "take-home" baby was the most important long-term and outcome of interest. (29,30)erefore, the results of LBR need to be further updated.As such, taken together with the continuously added new evidence, this study aimed to evaluate the impact of acupuncture IVF RCTs for both CPR and LBR by conducting comprehensive literature covering English and Chinese involving all common acupuncture methods.

METHODS Registration
The study has been registered (No.PROSPERO CRD42021232430) online and could be accessed by www.crd.york.ac.uk/prospero/.

Literature Search
PubMed, Web of Science, EMBASE, and Cochrane Register of Controlled Clinical Trials (CENTRAL) were searched to obtain articles addressing the effect of acupuncture on the outcomes of IVF published from database inception to July 2, 2023.Medical Subject Headings (MeSH) terms and free text terms were combined to screen for potentially relevant studies in PubMed, the detailed search algorithm was slightly modified and used in several other databases (Appendix 1).In addition, acupuncture originated from China and a large number of IVF RCTs were conducted every year.Therefore, main Chinese databases such as the Wanfang Databases, Chinese National Knowledge Infrastructure (CNKI), the Chinese Science and Technology Periodical Database, and SinoMed were used to collect related studies published in Chinese.Other potentially relevant papers were also identifi ed by cross-referencing.The present study complied with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Guidelines (Appendix 2).

Inclusion Criteria
The inclusion criteria were as follows: (1) RCTs that evaluated the acupuncture's effects for women undergoing IVF; (2) articles were published in English or Chinese; (3) pregnant data could be directly extracted or calculated from the original article; (4) any type of acupuncture was accepted, including traditional acupuncture (TA), electrical acupuncture (EA), warm acupuncture (WA), auricular acupuncture (AA), laser acupuncture (LA), transcutaneous electrical acupoint stimulation (TEAS) and others.No treatment, sham or placebo acupuncture, active treatments including Western medicines, dietary supplements, or lifestyle interventions were all allowed.However, if there were one or above co-intervention together with acupuncture in the treatment group, the co-intervention between the treatment and control group should be the same.Studies with treatments performed around the time of embryo transfer were also included.
If the study was reported in duplication, the version firstly published or provided more detailed information was included.Studies with review or metaanalysis articles, animal studies, abstracts, comments, letters, case reports, retrospective studies, full article unavailable, and not related studies were excluded.To avoid the small sample size effect, the studies with less than 30 sample size were excluded.

Data Extraction
The relevant literature was reviewed independently in pairs by 4 investigators (Liu XY, Pu ZQ, Sun CY and Lin YJ) and the consensus was reached on all of the items.Any disagreements were resolved by discussion or with the involvement of a third investigator (Ma PH).CPR (presence of at least one gestational sac or fetal heartbeat, confi rmed by transvaginal ultrasound) and LBR (a baby born alive after 24 weeks gestation) were the primary outcomes, and the secondary outcomes were ongoing pregnancy rate (OPR, pregnancy beyond 12 weeks of gestation, as confi rmed by fetal heart activity on ultrasound), miscarriage rate (MR) and adverse events related to acupuncture.The following information was also extracted: first author and published year, sample size, reasons for IVF, age of the participants, country, treatment, control, main objective, acupuncture time and sessions, and IVF outcomes.

Assessment for Risk of Bias
The risk of bias was assessed using Version 2 of the Cochrane tool for assessing the risk of bias in randomized trials (RoB 2). (31)Each item of every trial was categorized as "high", "some concern", and "low".If one of the domains was rated as high, the study was considered to be at high risk of bias.If all the domains were rated as low, the study was considered to be at low risk of bias.Two reviewers (Sun CY a and Sun CY b ) were independently involved in the assessment for risk of bias, and the consensus was reached on all the items.Any disagreements were resolved by a third investigator (Zhang HR).

Methodologies Assessment
The potential methodological quality was assessed using Grading of Recommendations Assessment, Development and Evaluation (GRADE). (32)The detailed assessment principle of GRADE including risk of bias, inconsistency, indirectness, imprecision, and publication bias.Synthesis of evidence for outcome were categorized as "high", "moderate", "low" and "extremely low".If the total scores of any of the 5 domains were reduced by 2 grades, it would be rated as low quality.Two reviewers (Sun CY a and Sun CY b ) were independently involved in the assessment.Any disagreements were resolved by a third reviewer (Zhang HR).

Statistical Analysis
Meta-analyses were carried out using STATA Meta-Analysis (V2.0, Biostat, Englewood, NJ, USA).The RoB 2 graph was drawn by Review Manager 5.3 (Cochrane Collaboration Software, RevMan).The measure of treatment effect was calculated by the pooled relative risk (RR) with 95% confidence intervals (CIs) with random effects models.Intention-to-treat (ITT) analysis would be used when available.
To examine the effect of different sub-populations, the subgroup was predefi ned with respect to the age of participants (<35 or 35 years (33) ), different acupuncture types, type of control (no treatment or sham/placebo control), acupuncture timing [(2 types were categorized: type A, the acupuncture was performed around the time of embryo transfer (ET)], which referred to the acupuncture treatment performed before and after ET; (34) type B, before the ET); geographical origin of the study (Asian or not Asian), whether or not repeated IVF failure (RIF), acupuncture sessions (1, 2 or more than 3 sessions).Otherwise, previous studies have shown that the IVF results would be better for those with lower baseline pregnancy rates, (20,30) therefore, the stratified analysis was conducted by the pregnancy rate of no treatment group.Sensitivity analyses were limited to the following studies: (1) the studies simultaneously with CPR and LBR results; (2) the studies mainly aiming to evaluate the effects of acupuncture on pregnancy rate from IVF. Heterogeneity was assessed using the I 2 test.Publication bias were shown by a funnel plot and assessed by the Egger weighted regression methods.P<0.05 was considered statistically signifi cant.

Search Results
The selection process was presented in Appendix 3. The initial search identifi ed 1,958 articles.After removing the duplicate articles, 1,223 were screened.Based on the title and abstract, 118 articles were further evaluated in detail.Finally, 60 articles were excluded and 58 RCTs (10,968 participants) with 32 published in English and 26 in Chinese were included in the final review (Appendices 4-5).If the included studies had multiple treatment or control groups, the pooled analysis was undertaken based on each comparison.Therefore, 69 comparisons for CPR and 23 comparisons for LBR from 58 trials were reported.

Risk of Bias of Included Studies
As shown in Appendices 6-7, there was no specific description of random sequence generation or deviations from intended interventions in 31 studies and it was determined to be high risk of bias.There were 14 studies with the uncertain defi nition of some items from all 5 main domains which were defi ned as unclear risk of bias.However, only 4 studies (44,47,66,85) were scored as low risk of bias.In all included studies, since results could be directly obtained, reporting bias was assessed as "low risk" for all included studies.

Certainty of Evidence
Evaluation of GRADE for each published metaanalysis by our study outcomes and acupuncture types were shown in Appendix 8.The detailed assessment principle was depicted in Appendix 9.Because the methods were not clearly described, results of CPR, LBR, OPR, and MR were all scored as low or extremely low in quality.With respect to varied treatment types of CPR outcomes, studies with TEAS (n=10) and LA (n=2) were recognized as moderate quality because of 1-degradation of the risk of bias.However, studies reported with TA, WA and EA were considered low or extremely low quality as severe methodological fl aws.In addition, for LBR outcomes, besides LA (n=2) being graded as moderate quality, other treatment types were scored as low to very low evidence.

Subgroup Analysis
Tables 1 and 2 showed the subgroup analysis results for CPR and LBR of TA and TEAS treatment.In subgroups classifi ed by control type (no acupuncture, sham or placebo), when compared with no treatment, TA did improve CPR (28 comparisons, RR 1.28, 95% CI 1.17 to 1.41, I 2 =18.9).However, no signifi cant difference was found between TA and sham control (10 comparisons, RR 1.09, 95% CI 0.94 to 1.28, I 2 =42.4).While according to acupuncture sessions, TA performed for 2 or more than 3 times showed the significant effect on CPR (22 comparisons, RR 1.14, 95% CI 1.07 to 1.22,I 2 =20.5; 14 comparisons, RR 1.30, 95% CI 1.17 to 1.44, I 2 =0; respectively), while performed for 1 session found no significance (2 comparisons, RR 1.28, 95% CI 0.96 to 1.71, I 2 =78.6).TEAS also had similar results, more improvements on CRP and LBR in the patients who received 3 or more times acupuncture (10 comparisons

Sensitivity Analysis
Appendix 12 showed the results of the sensitivity analysis.When restricting to studies that reported both CPR and LBR, no significant difference was found in the pooled CPR results (10 trials, RR 1.21, 95% CI 0.85 to 1.36, I 2 =69.2).While the analysis was confined to the main goal of evaluating the IVF effect, the pooled CPR results were observed to be signifi cant (33 trials, RR 1.35, 95% CI 1.15 to 1.54, I 2 =61.6).

Publication Bias
Evident publication bias was observed on both CPR and LBR results as evaluated by Egger's weighted regression analysis (both P<0.01) and funnel

DISCUSSION
Our studies covered a wide range of acupuncture methods and provided a broad analysis of the existing evidences, the summarized estimation showed that acupuncture might be a benefi cial treatment for infertile women undergoing IVF.TEAS showed a significant benefi t on both CPR and LBR improvement, while TA, WA, EA and LA were found signifi cant effectiveness in CPR, but not significant for LBR.Subgroup analyses showed that less than the age of 35 years, more than 3 sessions of acupuncture, no RIF history, and the lower baseline estimate of the pregnancy rate had benefi ts on CPR and LBR.The methodological quality of evidence for all acupuncture treatments was low or extremely low.
Our review suggested that the pooled CPR and LBR from all the acupuncture groups were signifi cantly higher than that from all the control group.It was consistent with the studies reported by Manheimer, et al, (11) Zheng, et al, (12) and Smith, et al. (17) TA was associated with signifi cantly increased CPR, however, there were significant differences compared with no treatment, but no signifi cant differences compared with sham/placebo acupuncture, which may be partially explained by the characteristics of sham/placebo acupuncture.Many studies had suggested that sham/ placebo acupuncture did not insert the body, (90) and the efficacy paradox presented by Walach (91) should give a better explanation for the present results.The non-specific effect is a large part of acupuncture effectiveness, (92,93) In this sense, no treatment, waiting list, or other standard care should be given priority to evaluate the general effects of acupuncture in the IVF trials. (24)so, we found that a dosage of 3 times for TA and TEAS was both associated with significant improvement on CPR and LBR.Insuffi cient acupuncture dosage in IVF has been a common opinion. (94)Some investigators suggested that a clinically valid dosage of acupuncture should be more than 6 acupuncture sessions. (27)However, Westergaard, et al (88) suggested that when the dosage of acupuncture was increased, the early pregnancy loss rate increased in the acupuncture group compared with the control group.Therefore, further studies should explore the specific effective dosage of acupuncture treatment.
Of note, TEAS had a signifi cant benefi t on CPR and LBR with low heterogeneity.TEAS intervention did not involve in the existed reviews.Our results of TEAS are of great clinical signifi cance for IVF, based on its advantages of noninvasiveness and painlessness, TEAS may have better potential application and be more acceptable.Furthermore, the safety of TEAS may be better than TA, because the miscarriage rate was not significant between TEAS and no treatment in our study.Meanwhile, unlike the TA method, sham TEAS is easy to be implemented by "pseudopercutaneous patch" method.
In our review, with respect to certainty of effect estimates measured by GRADE, certainty of results was from extremely low to moderate across different outcomes.It suggested that clinicians should counsel patients adequately and comprehensively considerate various factors, when they would consider using acupuncture treatment.Specifically, risk of bias and inconsistency were the main downgrade domains in the GRADE assessment.Not enough blinding and dispersion in effect estimate were the main downgrade reasons, which should be paid more attention to future studies.There were some limitations in our study.First, the significant heterogeneity among trials should be kept in mind.The different acupuncture prescriptions and control types might be important contributions.Second, the different effects between no treatment and sham/placebo acupuncture could not be addressed.Third, most included studies did not report the related health economic evaluation of acupuncture on IVF, which made it impossible to be analyzed in our study.The importance is that it is involved in the generalization of acupuncture among infertile women.
In conclusion, the association between acupuncture and IVF pregnant effi cacy has been extensively studied, but the convincing evidence level of all the associations was low.Still remarkably, our analysis indicated that acupuncture especially for TEAS improved CPR and LBR among IVF women.To give better guidance for the clinical operations, well-conducted research with more detailed assessments of pregnant outcomes is needed.

Figure 1. Forest Plot on Pooled Results of Clinical Pregnancy Rate
, RR 1.23, 95% CI 1.03 to 1.48; 2 comparisons, RR 1.91, 95% CI 1.07 to 3.39; respectively).The additional results of the effects of different acupuncture methods by baseline , OPR, or LBR of no treatment group were shown in Appendix 11.Studies with lower control group rates of CPR showed larger effects of TA (RR 1.47, 95% CI 1.22 to 1.77, I 2 =0) than that with higher control group rates of CPR (RR 1.22, 95%CI 1.09 to 1.37, I 2 =0).