The Efficacy and Safety of the Zhuyun Formula and Auricular Acupressure for the Infertile Women with Recurrent Implantation Failure: A Randomized Controlled Trial

Background Recurrent implantation failure (RIF), a clinical disorder characterized by failure to achieve pregnancy after repeated (≥3) embryo transfer, is a challenge for reproductive demands worldwide. In our preliminary work, the Zhuyun formula (ZYF) with auricular acupressure, a complementary and alternative medicine (CAM) with a small sample size for RIF, can improve the clinical pregnancy rate (41.2% vs. 26.7%, treatment group vs. control group, p < 0.05). Based on the toxicological and pregnancy-related pharmacological analysis of ZYF for RIF, the T-cell receptor signaling pathway might be involved in the pharmacological activity. This study aimed at evaluating the efficacy and safety of the CAM therapy according to pregnancy outcomes and maternal and child health and investigating the changes of T-helper (Th) cells in the peripheral blood of unexplained RIF women. Materials and Methods We conducted a prospective, two-arms, randomized, nonblinded study. All eligible women were randomly assigned to the treatment group (TG) and the control group (CG) according to a computer-generated randomization list in sealed opaque envelopes. Blood samples were collected from the two groups, and serum Th1, Th2, and Treg were detected by flow cytometry. The cytokines were detected by an enzyme-linked immunosorbent assay (ELISA). The TG was administrated with ZYF and auricular acupressure for three months before ovarian stimulation, while the control group was on a waiting list for the same period. The primary outcome was CPR. The second outcomes were the serum levels of immune parameters. For the safety evaluation, the perinatal outcomes of maternal and child were obtained by follow-up. Post-hoc sensitivity analyses were performed to assess the effect of missing data. Results One hundred and twenty-three women were randomized into the TG (n = 62) and CG (n = 61). The CPR was increased significantly in the TG (45.2%) than CG (26.2%) (p = 0.029). Twenty blood samples were collected, and the Th2/Th1 and Treg expression level was significantly higher in the TG than in the CG. IL-2, IL-10, and Foxp3 were higher significantly in the TG than in the CG. The maternal and child perinatal outcomes were not significantly different between the two groups. Conclusions The ZYF with auricular acupressure was effective and safe in improving the pregnancy outcomes of RIF. It might be related to balancing the level of cytokines related to the immune tolerance of the maternal-fetal interface to protect the embryo from the maternal immune system. Clinical Trial Registration: Clinical Trial Registry; date: 14/Dec/2013; no. NCT03078205.


Introduction
Numerous progress in reproductive technology has significantly increased the opportunity of being parents for infertile couples [1]. However, a new challenge has emerged: recurrent implantation failure (RIF) [2]. Although there is no universally agreed-upon definition [3], RIF is often defined as the failure to achieve a clinical pregnancy after the repeated transfer (≥3) of more than four good-quality embryos under the age of 40 years [3,4]. RIF is still one of the most challenging clinical dilemmas worldwide since the low clinical pregnancy rate (CPR) of in vitro fertilization and embryo transfer (IVF-ET). ere is 10% incidence of RIF patients among the IVF-ET [2,5,6]. Currently, the etiology search and treatment of RIF mainly focus on the following three aspects: 1) maternal-fetal factors mainly including reproductive organ lesions, endometrial receptivity, and thrombotic diseases; 2) embryo factors mainly include chromosomal abnormalities, zona pellucida sclerosis, embryo culture, and transfer, and the malefactors; and 3) the immune factors, which gained increased attention recently, such as immune cells including T-helper cells ( 1), 2, Treg, 17, mononuclear macrophages, and natural killer cells, which are all important cells involved in the process of implantation. Autoimmune disease antiphospholipid antibodies and antinuclear antibodies have been confirmed to have higher plasma concentrations in patients with recurrent implantation failure than in healthy women of childbearing age, especially for the unexplained RIF [7][8][9].
Traditional Chinese medicine (TCM) therapy with a rich theoretical basis and clinical application for treating infertility includes Chinese herbs [10], acupuncture [11,12], moxibustion [13], and taichi (exercise), which have been widely and increasingly applied in the clinic as a complementary and alternative medicine (CAM) [14]. Kidney deficiency, Qi stagnation, and blood stasis play crucial parts in infertility pathogenesis.
e Zhuyun formula (ZYF) is composed of Wuzi Yanzong Wan (WZYZW) [15,16], Sini San (SNS) [17], and Siwu decoction (SWD), which are the most classic recipes for kidney nourishment, liver regulation, and blood circulation. e components and molecules of ZYF are shown in Table S1 and Figure S1. Auricular acupressure is also widely used for treating infertility, especially for decreasing distress [18]. We found the significant efficacy of a CAM therapy (ZYF combined with auricular acupressure) in the preliminary trial with a small sample size (n � 30, each group) for RIF underwent IVF-ET (CPR: 41.2% vs. 26.7%, treatment group vs. control group p < 0.05). Furthermore, based on the therapeutic efficacy, we performed network pharmacology of ZYF for the RIF target protein and biological regulation process to identify the therapeutic mechanism of ZYF on specific target proteins involved in RIF (Tables S2 and S3). We also found out that the biological regulation process of immune tolerance during embryo implantation (ET) and the T-cell receptor signaling pathway (Tables S4 and S5), especially, might be involved in the pharmacological activity of ZYF on RIF. However, reliable evidence supporting the efficacy and safety of the CAM therapy use during the IVF-ET is still lacking. Whether the CAM therapy affects postpartum recovery and normal development of newborns in the long term is also widely of concern for medical workers and families. Consequently, focusing on evaluating the effects on perinatal outcomes and long-term health among these infants should be complemented in the follow-up phase.
Accordingly, considering the particularity of the population (women of childbearing and who are pregnant) and safety concerns, we conducted the randomized, nonblinded study to evaluate the efficacy and safety of ZYF and auricular acupressure to reduce the selection bias. Our findings provided a scientific, effective, and systematic theory and contributed to improving the evidence regarding the safety of the ZYF during IVF-ET.

Sample
Size. Combining our preliminary study, evidence from the systematic reviews, and clinical advice, we estimated that a 25% or greater increase in the proportion of clinical pregnancy will be clinically important. To obtain 80% power at a 5% significance level for a 2-sided test, we assume a proportion of 24.5% [19] clinical pregnancies in the control group and 49.5% clinical pregnancies in the treatment group. e minimal sample size calculated is 55 for each group. Considering about a 10% dropout rate, 121 participants in total are needed. e sample size was calculated by PASS 15.0 software.

Study
Design. Women of childbearing age or pregnancy are special, and the placebo setting may increase the likelihood of declining enrollment. To reduce the placebo effect, objective outcomes were observed. us, the study was a nonblinded, single-center, prospective, randomized controlled trial conducted in the West China Second University Hospital, Sichuan University. All eligible patients collected in the study agreed to participate and signed an informed consent form. e study was approved by the ethics committees (2016KL-013). e authors registered the trial with Clinical-Trials.gov (NCT03078205). Details of the study design ( Figure S2), rationale for the primary and secondary outcome measures, power analyses, and the statistical analysis plan are available in the protocol (Supplementary Materials 1). It complies with the guidelines prescribed by the Consolidated Standards of Reporting Trials (CONSORT) checklist (Supplementary Materials 2).

Patient Recruitment.
Participants diagnosed with RIF [5] and aged between 20 and 39 years who were undergoing ART will be included. Women who were planning cycles of preimplantation genetic testing (PGT), preimplantation genetic screening (PGS), reproductive malformation, and reproductive inflammation, as well as those with a diagnosis of congenital abnormality (such as a submucous myoma, intrauterine adhesion, or uterine malformation), autoimmunity, and endocrine disorders, and their husbands diagnosed with serious asthenozoospermia or oligospermia, were excluded from this trial.

Randomized and Masking.
Women were recruited at the time of the diagnosis of RIF without remaining embryos, and the decision to undergo a new IVF cycle or intracytoplasmic sperm injection (ICSI) cycle and randomization occurred before the new menstrual cycle. Subjects were randomly allocated to the treatment group by a selection of the sealed envelope in the sequence based on a computer-generated list. Nurses did the randomization and the procedure was performed blinded to both the patients and to the clinician who performed the embryo transfer.

Intervention.
In the treatment group, CAM treatment was utilized for three months before follicle stimulation, which included two kinds of interventions: ZYF and auricular acupuncture. ① Z YF was given three times daily, 200 ml per time.
In the control group, all participants were naturally waiting for 3 months before IVF/ICSI. For all subjects undergoing ART after three months, a standard long agonist protocol for ovarian stimulation was performed [20]. Egg retrieval, fertilization, and embryo transfer were determined by their treating clinician.
2.6. Outcomes. Information on demographics, fertility history, and health status were collected from subjects. To reduce the placebo efficacy of the CAM therapy as possible, we collected the objective outcomes in our study. e details of the definition and calculation of the outcomes were shown in the protocol (Supplementary Materials 1)

e Pregnancy Outcomes after IVF-ET.
For the primary analysis, we first compared the proportion of women with clinical pregnancy in the two groups for all participants using relative risks (RRs) with 95% confidential difference (CI) with a hypothesis test for no effect. Clinical pregnancy is defined as the presence of at least one intrauterine gestational sac or fetal heartbeat confirmed by ultrasound, 4∼6 weeks after embryo transfer. e second outcomes were natural conception, ongoing pregnancy rate, and pregnancy loss (including biochemical miscarriage, clinical pregnancy loss, and ectopic pregnancy).
2.8. e Expression Levels of 1 (IL-2, IFN-c), 2 (IL-4, IL-10), and Treg (Foxp3) in Peripheral Blood. To describe and analyze the benefit of CAM for unexplained RIF, the level of immune cells ( 1 (IL-2, IFN-c), 2 (IL-4, IL-10), and Treg (Foxp3)) was tested. Enrolled subjects diagnosed with unexplained RIF had serum levels checked. Serum levels of 1 (IL-2, IFN-c), 2 (IL-4, IL-10), and Treg (Foxp3) were measured before and after the intervention at the middle luteal phase (6∼7 days after LH peak, progesterone ≥5 ng/ml). It is known that successful pregnancy is associated with the maternal immune, and we also extracted the ratio of 2 and 1 ( 2/ 1). 4 ml of complete blood was collected by vacuum blood collection vessel anticoagulant with heparin sodium. e peripheral blood mononuclear cells (PBMC) were isolated and cultured. e samples were stored in 4°C refrigerators and analyzed within 24 hours. Flow cytometry was used to detect the level of Treg cells (BD Cytofix/Cytoperm) in peripheral blood mononuclear cells using ™ fixation/permeabilization solution kit (554714, BD company, CA, USA), CD4-FITC (eBioscience, 11-0049-41), and CD25-PE (eBioscience, 12-0259-42). e level of IL-2, IFN-c, IL-4, IL-10, and Foxp3 was detected in an enzyme-linked immunosorbent assay (ELISA). e specific steps were strictly followed as per the instructions mentioned in the kit. Additional details on the experimental instruments and equipment, sample collection steps, and cytokines detection steps are provided in Supplementary Materials 1.

Maternal and Child Health during the Perinatal Period.
e pregnancy and perinatal outcomes were obtained by followup with the patient as per the Society of Assisted Reproductive Technologies (SARTs) reporting guidelines through systematic medical records inquiry, face-to-face consultation, telephone, or WeChat inquiry. Maternal health indicators included Down's syndrome screening, the incidence of pregnancy complications and pregnancy-specific diseases, and delivery and postpartum conditions to evaluate maternal health. e neonatal outcomes will be recorded, and the neonatal development will be observed by follow-up in 3 months and 1 year.

Statistical Analysis.
To assess the effect of missing data, an intention-to-treat (ITT) analysis was planned and we performed post-hoc sensitivity analyses, fitting best-and worst-case scenarios. For the best-case scenario, we assumed all unknown events in the treatment group were positive (clinical pregnancy) and those in the control group were negative. For the worst-case scenario, we assumed none of the women with missing data in the treatment group became pregnant, and all of the women with missing data in the control group did become pregnant.
Evidence-Based Complementary and Alternative Medicine SPSS 25.0 statistical software was used for statistics, and the measurement data were expressed as X ± s. e data were normal distribution, the independent sample t-test was used for intergroup comparison, and paired sample t-test was used for intragroup comparison; if the data were skew distribution, the Mann-Whitney U test was used between groups, Wilcoxon signed rank-sum test, case/control association analysis, and chi-square test were used before and after the group, and the difference was statistically significant (p < 0.05).

Patient Enrollment.
A total of 150 women were assessed for eligibility in the trial. However, fifteen patients did not meet the inclusion criteria and twelve patients met the exclusion criteria ( Figure 1). Finally, 123 patients were randomly assigned to either the intervention group (62 cases) or the control group (61 cases). During the threemonth intervention phase, 5 participants in the TG and 1 in the CG dropped out and 8 participants in the TG and 1 in the CG got natural conception. In the IVF-ET phase, 2 participants (1 was frozen all embryos and 1 was no surviving embryos) in the TG and 10 (4 were frozen all embryos, 2 were no surviving embryos, 4 were canceled for personal reasons) in the CG were canceled ET ( Figure 1). erefore, 108 women completed the three-month intervention phase (49 in the TG, 59 in the CG). Ninety-six women completed the follow-up (47 in the TG, 49 in the CG). Recruitment took place between March 2017 and February 2020. e TG and the CG were comparable in baseline demographics ( Table 1). As for the outcomes of ovarian stimulation and embryo culture, the A-type endometrial pattern (EMP) on the hCG day was significantly higher in the TG (p � 0.033), and there were no significant differences between TG and CG (p > 0.05) regarding several mature ova, several D3 embryos, and stage of embryo transferred (Table 2). e mean (SD) age of the participants was 30.9 (4.02) years. 24.2% of women in the TG, and 23.0% in the CG were diagnosed with unexplained RIF.

e Pregnancy Outcomes after IVF-ET.
As the primary outcome, CPR was significantly higher in the TG than in the CG (28 of 62 [45.2%] in the TG vs. 16 of 61 [26.2%] in the CG; relative risk: 1.72; 95% CI: 1.04-2.85; p � 0.029) (Table 3). Among the second outcomes of pregnancy, natural pregnancy rate and ongoing pregnancy rate were also significantly higher in the TG than in the CG (Table 3). ere was a trend suggesting that the CAM therapy may perform more successfully in avoiding biochemical miscarriage (RR: 0.318; 95% [CI]: 0.09-1.10, p � 0.052). However, the rates of conception per woman, clinical pregnancy per IVF cycles, clinical pregnancy per ET cycles, and clinical pregnancy loss were similar between the two groups (Table 3).
For the post-hoc outcomes in clinical pregnancy, the best-case sensitivity analysis found significant improved differences in the TG than the CG (RR: 2.15; 95% CI: 1.34-3.46; p < 0.001), whereas the worst-case sensitivity analysis found no difference (RR: 1.02; 95% confidence interval (CI): 0.69-1.51; p � 0.92) (Table 3). Moreover, we did not determine any side effects associated with CAM treatment. 10), and Treg (Foxp3) in Peripheral Blood. For unexplained RIF, subgroup analysis was analyzed between the two groups, CPR was significantly higher in the TG than in the CG (9 of 15 [60.0%] in the TG vs. 3 of 14 [21.4%] in the CG; p � 0.035) (Figure 2(a)). Nineteen blood samples (9 from the TG and 10 from the CG) of unexplained RIF were collected from the two groups, respectively (5 patients in the TG and 4 patients in the CG refused blood sampling) for flow cytometry and the ELISA test. One sample in the TG for flow cytometry becomes hemolysis. When serum data were analyzed before and after the intervention, there were variations in different T lymphocyte expressions. Based on our observations on the flow cytometry outcome of PBMC, there were significant differences in 2/ 1, and Treg expressions in the TG than in the CG after the intervention (i.e., 2/ 1 13.18 ± 7.15 vs. 7.22 ± 3.52, TG vs. CG after the intervention, p < 0.05) ( Table 4, Figures 2(b) and 2(c)). Moreover, based on the observation of ELISA, IL-2, IL-10, and Foxp3 were significantly higher in the TG than in the CG after the intervention (i.e., Foxp3 220.66 ± 82.2 vs. 208.66 ± 91.36, the TG vs. the CG, after the intervention, p < 0.05) ( Table 5).

Maternal Child Health Outcomes during the Perinatal
Period.
e perinatal outcomes of 41 pregnant women (28 in the TG and 13 in the CG) and 56 neonatal outcomes were obtained by the follow-up (natural conceptions were included). ere was no significant difference in the Down's syndrome-related screening, the pregnancy-related complications and idiopathic diseases, and delivery, postpartum condition, and related complications. And the basic characteristics of newborns, neonatal complications, and neonatal development in 3 months and 1 year were similar between the two study groups (p > 0.05) (Tables 6 and 7).

Discussion
In this single-center, randomized trial involving 123 patients diagnosed with RIF who underwent IVF-ET, our study demonstrates that CAM therapy (ZYF with auricular acupressure) is efficient and safe in improving the pregnancy outcomes of RIF women who is undergoing IVF-ET, including the clinical pregnancy rate and ongoing pregnancy rate.
e balance of the immune system during embryo implantation is considered to be based on the change of T lymphocyte expression in unexplained RIF women. We believe this is the first time to evaluate the efficacy and maternal and child health of the CAM therapy (ZYF with auricular acupressure) for the RIF and a correlation between serum T lymphocyte expression and the use of the CAM therapy in IVF cycles has been reported. 4 Evidence-Based Complementary and Alternative Medicine About 1∼3% of women experience early recurrent implantation failure; about 50% of patients are associated with unexplained RIF [21][22][23]. Many studies have found that the cytokine balance of 1/ 2/ 17/regulatory T cells (Treg) plays an important role in the maintenance of embryo implantation and normal pregnancy, especially for unexplained RIF [23]. e disorder of the cytokine balance will lead to repeated implantation failure and recurrent abortion.
Evidence-Based Complementary and Alternative Medicine anti-inflammatory response. Treg cells mainly secrete IL-10, IL-35, and transforming growth factor (TGF)-β, which participate in maternal and fetal immune tolerance and maintain normal pregnancy. 1/ 2/ 17/Treg cytokines can affect the maternal immune system, regulate the angiogenesis of the maternal-fetal interface, and play an important role in all stages of pregnancy [7].
According to the TCM theory, RIF is associated with kidney essence deficiency, liver stagnation, and blood stasis [24,25]. erefore, improving the physical condition of the kidney, liver, and blood may improve the pregnancy outcomes of RIF. Significant efficacy has been demonstrated for subfertile women by Chinese herbs [25,26] and acupuncture [12,27] , which are classic herbs in nourishing the kidney, smoothing the liver, and circulating the blood to improve the CPR. Moreover, searching the HERB database of the herbs, ellagic acid, astragalin, acteoside, ferulic acid, and gallic acid were considered the most relative bioactive molecules for RIF (based on the gene target degree, Figure 3(a)), which are essential for the development of the embryo, such as astragalin and ferulic acid can regulate the balance the immune-related signaling way to suppress anti-inflammatory in uterine [28]. However, excessive amounts of ellagic acid (dose >1208 mg/kg b.w.) will induce abortion (Table 8).
Our findings of ZYF with auricular acupressure in regulating the process of immune tolerance during embryo transfer verified the network pharmacology analysis. Based 0.92 * Natural conception was calculated. ¶Assuming all unknown events in the treatment group were positive, and those in the control group were negative. § Assuming none of the women with missing data in the treatment group became pregnant, and all of the women with missing data in the control group did become pregnant. Values are mean ± SD or n(%). e definition and calculation of indicators are shown in Table S7. on the network pharmacology of ZYF for RIF, KEGG pathway enrichment showed that the T-cell receptor pathway, 1, and 2 cell differentiation are prominent pathways (Figure 3(b), Table S6). Furthermore, IL-4, IL-10, HLA-DQA1, NF-κB, and ANXA1 signal pathway genes are enriched in large numbers, which are the key pathways for  5.89 ± 1.54 ‡ e data showed a skewed distribution, the Mann-WhitneyU test was used for comparison between groups, and the Wilcoxon signed-rank test was used for comparison before and after treatment within groups; * p < 0.05, TG vs. CG (after intervention); # p < 0.05, ## p < 0.01, TG (after intervention) vs. TG (baseline). Values are mean ± SD. e data showed a skewed distribution, the Mann-Whitney U test was used for comparison between groups, and the Wilcoxon signed-rank test was used for comparison before and after treatment within groups; * p < 0.05, TG vs. CG (after intervention); # p < 0.05, ## p < 0.01, TG vs. TG (baseline); values are mean ± SD.

Evidence-Based Complementary and Alternative Medicine
ZYF to treat RIF, as shown in Figure 3(c). ere is a close synergy between key pathways, which can regulate the key biological processes of maternal-fetus balance during the embryo implantation, especially for immune regulation and other related targets.
Pregnancy is a physiological process greatly dependent on immune tolerance. In our study, the results of a higher level of 2/ 1 and Treg expression compared with the control group are consistent with the improvement of pregnancy outcomes. Furthermore, the secretion of IL-2 (produced by 1) was decreased. IL-10 (produced by 2) and Foxp3 (produced by Treg) were increased in the TG than in the CG. T lymphocytes play a critical role in regulating the immune response and maintaining immune Values are mean ± SD or n (%).    Evidence-Based Complementary and Alternative Medicine tolerance both physiological and pathological during embryo implantation [53]. Recent studies have shown that pregnancy is a process of mutual conversion between proinflammatory and anti-inflammatory conditions [54,55]. e status of pregnancy was divided into three distinct immunological stages: at the first beginning of ET, a proinflammatory stage is performed [56]; in the second status, for fetal growth, an anti-inflammatory-oriented stage is performed [57,58]; and third, an anti-inflammatory status switched to a pro-inflammatory stage for the initiation of labor [59]. Some studies have demonstrated in RIF patients, a predominant 2/Treg immunity (which produces antiinflammatory cytokines) underlies normal pregnancy [54], whereas a predominant 1 immunity (which produces proinflammatory cytokines) has been observed in women with recurrent miscarriage [60]. Accordingly, the balance of 2/ 1 and the shift from pro-inflammatory to antiinflammatory status has been linked to positive pregnancy outcomes in those undergoing IVF [61], which is also important for RIF in the early stage of ET.
In agreement with previous studies [10], we demonstrated no side effects of the CAM therapy in infertile women who underwent IVF-ET. In addition, our findings suggest that CAM therapy has no increased risk of maternal and child health in treating RIF who underwent IVF-ET. According to the Society of Assisted Reproductive Technologies (SARTs) reporting guidelines, the perinatal outcomes (Down syndrome and pregnancy or delivery-related complications and idiopathic diseases) and the neonatal outcomes (birth weight, Apgar score, and neonatal complications) were obtained by follow-up during the pregnancy, and three months and 1 year after born. Furthermore, the safety of the CAM therapy is demonstrated with durable effects 1 year after giving birth. To our best knowledge, several large epidemiological studies [62,63] have concerned about the perinatal outcomes such as preterm delivery and low birth weight of the ART-conceived infants without inconsistent results. However, enough follow-up needs more observation, especially for CAM therapy.  injection of over 40 mg/kg 5-12 days after conception may lead to abnormal development of the fetal musculoskeletal system Chlorogenic acid (50 μmol/L) supplementation can improve the maturation, fertilization, and development of oocytes, and improve the quality of embryonic development and electroporation-treated embryos [29] Ellagic acid Injection of excessive amounts of ellagic acid (dose >1208 mg/kg b.w.) induced abortion after 16 days of gestation in females, with specific developmental abnormalities in the blood and lymphatic systems Ellagic acid (10 μM) protects zebrafish embryonic development from oxidative DNA damage, improves embryo survival, and improves morphological parameters in larvae in vitro [30]; ellagic acid (60 mg/kg) shows anti-senescence protection in rat embryonic fibroblasts [31] Hyperoside No relevant toxicological reports have been published Hyperoside (40 mg/kg b.w.) can reduce pregnancy loss in anticardiolipin antibody-positive rats by regulating the mechanistic target of rapamycin (mTOR)/S6K and Toll-like receptor (TLR)-4/MyD88/nuclear factor-kappa B(NF-κB) signaling pathways [32].
Isoquercitrin Details of toxic effects have not been reported other than a lethal dose value (rat, ip >5 g/kg b.w.) Isoquercitrin (100 mg/kg b.w.) alleviates nitrosation, oxidation, and endoplasmic reticulum stress by inhibiting nitric oxide synthase (NOS)-2 and increasing the expression of superoxide dismutase (SD)-1, to treat neural tube defects in embryos of diabetic mothers [33] Acteoside Details of toxic effects did not report other than a lethal dose value (rat, oral (po) >5 g/kg b.w.) * e addition of acteoside (10～50 μM) during in vitro maturation (IVM) increases the blastocyst formation rate and significantly reduces reactive oxygen species (ROS). Furthermore, in acteoside-treated oocytes, cytoplasmic maturation is improved and mitochondria and lipid droplets are evenly distributed throughout the cytoplasm [34] Astragalin No relevant toxicological reports have been published Astragalin (100 μg/mL) over regulates the NF-κB/mitogen-activated protein kinase (MAPK) signaling pathway to suppress inflammatory responses in murine Leptospira-infected uterine and endometrial epithelial cells [35] Kaempferol Kaempferol at a dose of 25 mg/l can induce mutations in mammalian (hamster ovary) somatic cells Kaempferol (0.1 μM) improves embryonic development by inhibiting oxidative stress in porcine oocytes and reducing apoptosis [36]; kaempferol (0.05～1 μM) reduces mitochondrial dysfunction and abnormal oxidative stress induced by H 2 O 2 during embryonic development [37] Saikosaponin D Details of toxic effects were not reported other than a lethal dose value (ip > 1530 mg/kg b.w.) Saikosaponin D (5～50 mg/kg b.w.) has a weak estrogen-like effect in vivo and may be a potential plant estrogen [38]; its regulatory effect on reproductive diseases is not clear Paeoniflorin Excessive intake of paeoniflorin can lead to sleep behaviors in mice (ip > 3530 mg/kg b.w.) Paeoniflorin (8 mg/kg b.w.) can enhance endometrial receptivity of luteal insufficiency (RU486-induced model) mice through the leukemia inhibitory factor [39]; Paeoniflorin (60 mg/kg b.w.) can also improve depression-like behavior in offspring of prenatally stressed mothers by restoring hypothalamic-pituitary-adrenal (H-P-A) axis and glucocorticoid receptor related dysfunction [40] Naringin Excessive naringin use in guinea pigs is lethal when naringin administration reaches 1250 mg/kg b.w. per day; rats experience weight loss, etc. Naringin overdose in guinea pigs is lethal (ip >2 g/kg b.w.) e bi-directional regulatory functions of naringin's estrogenic and anti-estrogenic activities may present deleterious disturbances to endocrine regulation in women. However, it was reported that naringin does reduce fertility or cause early embryonic developmental toxicity even when the intervention dose reached 1250 mg/kg b.w. in rats [41]  A potential preventive effect of hesperidin (50 mg/kg b.w.) on formaldehyde (2 mg/kg b.w.) toxicity was found in pregnant rats [42]; the use of hesperidin (1 mg/kg b.w.) during pregnancy has a positive effect on the reflex motor behavior of mouse offspring, which may stem from its antioxidant activity [43] Liquiritin No relevant toxicological reports have been published Liquiritin and glycyrrhizic acid (0.2～0.4 mg/mL) can resist uterine contractile spasms caused by oxytocin by inhibiting the phosphorylation of heat shock protein (HSP)-27 [44] Glycyrrhizic acid Glycyrrhizic acid doses >662 mg/kg b.w. may cause seizures in humans. e lethal dose in mice (po) is >3 g/kg b.w.
Ferulic acid IP injection of 350 mg/kg ferulic acid can cause ataxic behaviors such as rigidity in mice; an intravenous dose >857 mg/kg b.w. resulted in mouse death Ferulic (40～120 μM) acid inhibits NK-κB and MAPK pathways and inhibits lipopolysaccharide-induced inflammation in endometrial epithelial cells [28]; ferulic acid (20 mg/kg b.w.) protects pancreatic islets in pregnant diabetes rats insulin β cell [45]. It can also improve placental inflammation and pathological apoptosis in preeclampsia rats [46] Amygdalin An oral dose of amygdalin >50 mg/kg b.w. causes nausea or vomiting in human infants and even gastrointestinal dyspnea; ip injection dose >167 mg/kg b.w. will lead to drowsiness or significant weight change in primates; ip injection >405 mg/kg b.w. causes mouse death Treatment of endometriosis rats with a combination of atorvastatin and amygdalin (5 mg/kg b.w.) modulates the expression of TNF-α, IL-6, matrix metalloproteinases (MMP)-2 and MMP-9 [47]; amygdalin is particularly important for female reproduction by modulating extracellular and intracellular signaling pathways involved in secretory activity, cell viability, steroidogenesis, proliferation, and apoptosis [48] Catalpol Oral catalpol administration >10 g/kg b.w. or intravenous injection (iv) >2500 mg/kg b.w. lead to mouse death Catalpol (10～100 mg/kg b.w.) can inhibit the transmission of TLR4 signals, prevent the expression of its downstream NF-κB/MAPK signal pathway, and significantly reduce the inflammatory cytokines IL-1β, IL-6, TNF-α, and chemokines C-X-C motif chemokine ligand (CXCL)-8 and CXCL-5, as well as myeloperoxidase (MPO) activity of uterine tissue, to implement anti-inflammatory and protective effects on endometrial tissue [49] Gallic acid Gallic acid, at oral doses >5 g/kg b.w., causes chronic pulmonary edematous liver in rabbits; subcutaneous injection at >5 g/kg b.w., iv >320 mg/kg b.w., or >4300 mg/kg b.w. ip is lethal in mice Subcutaneous injection of gallic acid (dose >5 mg/kg b.w.) 1 day before the mating of female animals can affect ovary and fallopian tube function; gallic acid can inhibit the production of pro-inflammatory and labor-promoting mediators related to myometrial contraction and rupture of fetal membranes. Future preclinical studies may clarify the efficacy of gallic acid in the prevention of inflammatory preterm birth [50] Ligustrazine Ligustrazine oral administration >1910 mg/kg b.w., ip >800 μg/kg b.w., or iv > 239 mg/kg b.w. will lead to mouse death Ligustrazine (100 mg/kg b.w.) can delay the development and fibrosis of endometriosis mice by reducing the pathological aggregation of platelets and the pathological expression of epithelial-mesenchymal transition, fibroblast-to-myofibroblast transdifferentiation, and fibrogenesis markers [51]; ligustrazine (60 mg/kg b.w.) induction reduces preeclampsia by inhibiting trophoblast autophagy and promoting its survival and migration by regulating the miR-16-5p/IGF-2 axis [52] Note: toxicological information with * was obtained from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/) and the CHEMSRC database (https://www.chemsrc.com/).

Limitation
is trial has limitations. First, placebo ZYF and sham auricular acupressure should be administrated for potential placebo efficacy. Second, successful pregnancy is a complex process, and we explore the mechanism based on the network pharmacology only.

Conclusion
In conclusion, the CAM therapy (ZYF with auricular acupressure) has achieved the balance of the immune system during ET for RIF. e mechanism may be mediated by upregulating the expression of Treg and 2/ 1 pattern, which can help embryo implantation. In addition, concerning the perinatal outcomes and the neonatal outcomes, the CAM therapy indicated the safety for the RIF women. For further study, we will screen the main active ingredients of ZYF in the treatment of RIF and combine metabolomics with proteomics to fuse various metabolic pathways and regulatory pathways to jointly elucidate the overall status of biological systems and construct the expression regulatory network of the final metabolites in the further study. Finally, the protocol of three months before IVF-ET might be a burden to patients in other countries.

RIF:
Recurrent implantation failure CPR: Clinical pregnancy rate ART: Assisted reproductive technology IVF-ET: In vitro fertilization and embryo transfer COS: Controlled ovarian stimulation TCM: Chinese traditional medicine ZYF: Zhuyun formula ET: Embryo transfer.

Data Availability
All data used to support the finding of the study can be obtained from the corresponding author upon reasonable request.
Ethical Approval e experimental project was approved by the research ethics committee of Sichuan Hospital of traditional Chinese medicine (Approval no. 2016kl-013). All participants gave informed written consent.

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

Authors' Contributions
Hang Zhou and Xiaoyan Zheng contributed equally to this work and should be regarded as the co-first authors. Wanting Xia and Jinmei Li collected the data. Hang Zhou and Xiaoyan Zheng conducted a statistical analysis of the data and contributed to the writing of the manuscript. Hang Zhou performed flow cytometry analysis. Ma Qianhong has made important contributions to the completion of the IVF-ET process. Zeng Qian and Jinzhu Huang designed the research project and have the final interpretation of the project. All authors read and approved the final manuscript.