Novel approaches to the management of recurrent pregnancy loss: The OPTIMUM (OPtimization of Thyroid function, Thrombophilia, Immunity, and Uterine Milieu) treatment strategy

Abstract Purpose Does the OPtimization of Thyroid function, Thrombophilia, Immunity, and Uterine Milieu (OPTIMUM) treatment strategy, developed for treating repeated implantation failure (RIF), contribute to improving pregnancy outcomes in patients with a history of recurrent pregnancy loss (RPL)? Methods Between 2018 and 2019, women with RPL after two or more clinical pregnancy losses underwent RPL testing. We treated chronic endometritis with antibiotics, high Th1/Th2 cell ratios with vitamin D and/or tacrolimus, overt/subclinical hypothyroidism with levothyroxine, and thrombophilia with low‐dose aspirin. Of 168 consecutive women aged ≤43 years, 115 underwent RPL testing. We compared 100 pregnancies (90 women) and 46 pregnancies (41 women) with and without the OPTIMUM treatment strategy, respectively. Results RPL testing identified intrauterine abnormalities in 66 (57.4%), elevated Th1/Th2 cell ratios in 50 (43.5%), thyroid dysfunction in 33 (28.7%), and thrombophilia in 33 (28.7%). The live birth rate in the OPTIMUM group was significantly higher than that in the control group among women aged <40 years (78.1% and 42.3%, respectively; p = 0.002), but no significant difference was observed in women aged ≥40 years (55.6% and 30.0%, respectively; p = 0.09). Conclusions The OPTIMUM treatment strategy improved pregnancy outcomes in patients with not only RIF but also RPL.


| INTRODUC TI ON
Pregnancy loss is the most frequent complication during pregnancy. 1 Yet, ≥50% of women with a history of recurrent pregnancy loss (RPL) have no risk factors. [2][3][4] Women with unexplained RPL may suffer repeated sporadic miscarriage, but cannot give birth because of risk factors undetectable using common RPL screening. 5 Although many clinical trials for unexplained RPL have been performed, there remains no established treatment. 6,7 Pregnancy loss is considered a multifactorial disease, as complex multiple influences including lifestyle habits also affect pregnancy outcomes. [8][9][10] We previously reported a combination treatment for abnormalities of intrauterine circumstances, immune status, and thyroid function, referred to as the "OPTIMUM" (OPtimization of Thyroid function, Immunity, and Uterine Milieu) treatment strategy, for treatment of repeated implantation failure (RIF) after multiple embryo transfer cycles in assisted reproductive technology (ART) treatment. 11 In this strategy, we detected and treated the risk factors for reproductive failure including intrauterine disorders, elevated helper-T (Th)1/Th2 cell ratios, and thyroid dysfunction.
Thyroid dysfunction has been established as a risk factor for RPL. 12,13 but intrauterine organic abnormalities including endometrial polyps and submucosal myomas remain insufficient evidence. 12 However, hysteroscopic surgery dramatically improves pregnancy rate, because intrauterine disorders can impair the chance of embryo implantation. 14,15 Persistent endometrial inflammation (ie, chronic endometritis [CE]) is also not included among the risk factors of RPL because of a lack of reliable evidence. However, CE was recognized in the local endometrium of 24% to 56% of women with a history of RPL. [16][17][18] Previous research showed that unexplained RPL is involved in an endometrial proinflammatory response with an abnormally increased density of uterine natural killer (uNK) cells and aberrant angiogenesis in endometrial stromal cells [19][20][21] ; therefore, CE is a potential risk factor for RPL. The relationship between immunological abnormality and RPL is also unknown. Successful pregnancy requires balancing Th1 and Th2 cells with secretion of pro-and anti-inflammatory cytokines, respectively; thus, perturbations in the Th1/Th2 cell ratio with Th1 bias can cause not only implantation failure but also pregnancy loss. 22,23 Local and systemic abnormal inflammatory reactions are linked to unexplained RPL. Furthermore, in our previous study of the OPTIMUM treatment strategy, we examined and treated thrombophilia with low-dose aspirin, because the prevalence of thrombophilia is relatively high in women with an RIF history. 24,25 Therefore, the OPTIMUM treatment strategy can cover various risk factors for pregnancy loss.
In this study, we analyzed the prevalence of impaired intrauterine circumstances, elevated Th1/Th2 cell ratios, thyroid dysfunction, and thrombophilia, and the prognosis of clinical pregnancy in women with a history of RPL after the OPTIMUM treatment strategy.

| Patient selection
This study is a retrospective cross-sectional study. A total of 180 consecutive women with a history of RPL after two or more clinical pregnancy losses visited the Sugiyama Clinic Shinjuku from April 2018 to December 2019. After excluding 12 women aged ≥44 years, 168 women were recruited (Figure 1). To detect the prevalence of risk factors for RPL, 115 women including 67 aged <40 years and 48 aged ≥40 years underwent our RPL testing for the OPTIMUM treatment strategy in our clinic. The remaining 53 women without RPL testing including examinations for thyroid function and thrombophilia were also recruited as the control group. In the control group, some of the patients did not desire RPL testing or some doctors did not suggest RPL testing.
In the OPTIMUM group, 102 clinical pregnancies were confirmed in 92 women. We excluded two women who did not want to undergo our immunotherapy, tacrolimus; therefore, 100 pregnancies in 90 women were recruited for analysis of the therapeutic efficacy of the OPTIMUM treatment strategy. In the control group, 46 clinical pregnancies were confirmed in 41 women. Therefore, we compared the pregnancy prognosis in the OPTIMUM group with that of the control group. An intrauterine gestational sac by transvaginal ultrasound was used to diagnose clinical pregnancy. Pregnancy loss was defined as a loss of clinical pregnancy, not a biochemical pregnancy.
Infertility was defined as the failure to achieve clinical pregnancy after unprotected intercourse for 12 months or longer.
We compared women with a successful live birth (live birth group) and those with a pregnancy loss (miscarriage group) in their first pregnancies among 90 women who conceived after the OPTIMUM treatment strategy. To analyze the predictive factors in 113 women, including those unable to conceive for 6 to 14 months after the OPTIMUM treatment strategy, we also compared women who had a successful live birth within two pregnancies (success group) and those with no pregnancy or pregnancy loss (failure group). This study was approved by the local ethics committee of Juntendo University, Faculty of Medicine (No. 14-103) and Sugiyama Clinic (No. 18-002).

| OPTIMUM treatment strategy protocol
The design of OPTIMUM treatment strategy was described previously 11 (Figure 2). Our RPL testing consisted of the following: hysteroscopy; endometrial biopsy for CD138 immunostaining and intrauterine bacterial culture; measurement of serum levels of 25-hydroxyvitamin D 3 (25OHVD), interferon (IFN)γ-producing Th cell (Th1 cell) and interleukin (IL)-4-producing Th cell (Th2 cell); testing for thyroid function including thyroid-stimulating hormone (TSH) levels and thyroid peroxidase antibody (TPOAb); and thrombophilia screening for antiphospholipid syndrome (APS), protein C and S deficiency and factor XII deficiency. We did not perform preimplantation genetic testing for aneuploidy (PGT-A) and structural rearrangements (PGT-SR) or endometrial microbiome analysis.

| Medical interview for lifestyle habits
Pregnancy loss is caused by multiple environmental, genetic, and lifestyle factors. 8,9 Risk factors that increase the rate of pregnancy loss 1.5-to 2-fold include smoking 10 to 20 or more cigarettes daily, 26,27 daily caffeine intake of two to three cups of coffee, 28 drinking alcohol twice a week or more, 29 and obesity with a body mass index >30 kg/ m 2 . 30 Thus, we confirmed the women's lifestyle habits and recommended adjustments such as cessation of smoking, caffeine intake, and drinking as well as diet and moderate exercise for obesity. The experience of pregnancy loss is often physical burden, leading to depression, anxiety disorders, and insomnia. [31][32][33] Maternal stress is involved in an increased risk of pregnancy loss. 34,35 Therefore, counseling was provided as necessary to relieve their stress and anxiety.

| Tests and treatment for intrauterine circumstances
To confirm intrauterine milieu, we performed a hysteroscopy and endometrial biopsy for CE. When intrauterine diseases including submucosal myomas, endometrial polyps, and intrauterine adhesion were detected by hysteroscopy, we performed hysteroscopic surgery first. On the day of surgery, all women underwent endometrial sampling for intrauterine bacterial culture with antibiotic sensitivity testing using endometrial suction curette (Pipet Curet; Fuji Medical Corporation, Tokyo, Japan) after the vagina was washed sufficiently with physiological saline to prevent sample contamination. All intrauterine disorders and typical CE findings such as an erythrogenic surface, stroma, and micropolyps 36 were removed using a monopolar resecting loop (Olympus, Tokyo, Japan) without applying electrodes, as described previously. 37 The specimens were fixed in 10% formaldehyde for histological examination with CD138 immunohistochemistry staining. We sent the samples to BML, Inc. (Tokyo, Japan) for both CD138 immunostaining and bacterial cultures. Pathologists stained the specimens using anti-CD138 antibodies (M7228; Dako, Agilent Technologies Japan, Ltd., Tokyo, Japan) and counted the CD138positive plasmacytes in 10 nonoverlapping random stromal areas visualized at 400-fold magnification (BML, Inc.); CE was diagnosed with the presence of five or more CD138-positive cells. In our previous study, most CE in women with endometrial polyps were cured by polypectomy without antibiotic therapy. 37 Unnecessary antibiotic treatment decreased the CE recovery rates by polypectomy and pregnancy outcomes after surgery. 37 Therefore, CD138 immunostaining and bacterial cultures were reexamined without antibiotics during the luteal phase in the next menstruation cycle after surgery.

F I G U R E 1
Flowchart of patient selection. Of 180 women with recurrent pregnancy loss (RPL), we recruited 168 women, including 115 women who underwent our RPL testing (OPTIMUM group) and 53 women without RPL testing (control group); 12 women aged ≥44 years were excluded. To analyze the efficacy of the OPTIMUM treatment strategy, 100 pregnancies in 90 women in the OPTIMUM group and 46 pregnancies in 41 women in the control group were recruited Our treatment protocol for CE without intrauterine disorders was described previously. 38 Women underwent endometrial sampling during the luteal phase. When CE was diagnosed, patients received oral doxycycline (Vibramycin ® tablets; Pfizer Japan Inc., Tokyo, Japan), 100 mg twice a day for 2 weeks as the first choice.
If CE was not cured and specific bacteria were detected (except for Lactobacillus spp. or Bifidobacterium spp.), second-line therapy consisted of bacterium-sensitive antibiotics for 2 weeks, based on the results of the antibiotic sensitivity testing. If CE was not cured without specific bacteria, we administered a combination of amoxicillin (Sawacillin ® tablets, 250 mg; LTL Pharma Co., Ltd.), azithromycin (Azithromycin tablets, 250 mg; Fuji Pharma Co., Ltd.), metronidazole (Flagyl ® , 250 mg; Shionogi & Co., Ltd.), and antibiotic-resistant lactic acid bacteria (Biofermin-R ® tablets, 6.0 mg; Biofermin pharmaceutical Co., Ltd.) twice daily for 2 weeks. When CE was not cured by antibiotic treatment, dilatation and curettage (D&C) was performed to remove the endometrium with plasmacytes.

| Tests and treatment for Th1/Th2 cell balance
Immune testing consisted of measuring serum 25OHVD, IFNγ, and IL-4 using SRL Inc. (Tokyo, Japan) as described previously. 39,40 Vitamin D deficiency, insufficiency, and sufficiency were diagnosed as 25OHVD levels of <12, between ≥12 and <30, and Vitamin D was supplemented at 1000 or 2000 IU daily for 25OHVD levels of ≥20 and <30 ng/ml or <20 ng/ml, respectively, based on our previous report. 39 Based on the previous study, the definition of high Th1/Th2 cell ratios was determined to be >10.3. 40 Elevated Th1/Th2 cell ratios can be attenuated by vitamin D supplementation 39 ; therefore, when the women had high Th1/Th2 cell ratios and low 25OHVD levels, the serum levels of 25OHVD, Th1, and Th2 cells were retested after 3 months or more of vitamin D intake. When their 25OHVD levels were <30 ng/ml, the daily dose of vitamin D supplementation was increased by a further 1000 IU. When high Th1/Th2 cell ratios could not be controlled, we treated them using an immunosuppressive drug, tacrolimus (Prograf ® capsules, 1 mg; Astellas Pharma), as described previously. 40,42 Women with RPL who had Th1/Th2 cell ratios of 10.3-12.9, 13.0-15.7, and ≥15.8 were treated with 1, 2, and 3 mg of tacrolimus daily, respectively, from the day of positive pregnancy test (4-5 weeks of gestation). In some women with a history of RIF, the intake of tacrolimus was initiated from 1 day before the day of embryo transfer, as in our previous trial. 11 In women with aberrant elevated Th1 cell levels (≥28.8), the dosage of tacrolimus was increased by a further 1 mg. 42 F I G U R E 2 Recurrent pregnancy loss testing and treatment for risk factors of pregnancy loss. Recurrent pregnancy loss (RPL) testing included a hysteroscopy, endometrial biopsy for CD138 immunostaining and bacterial culture, and blood testing for 25-hydroxyvitamin D 3 , interferonγ-producing helper-T (Th1) cell, interleuking-4-producing helper-T (Th2) cell, thyroid function, and thrombophilia. We treated intrauterine organic disorders with hysteroscopic surgery, chronic endometritis with antibiotics and/or dilatation and curettage, high Th1/ Th2 cell ratios with vitamin D supplementation and/or tacrolimus, overt or subclinical hypothyroidism with levothyroxine and thrombophilia with low-dose aspirin and/or low-molecular-weight heparin [Colour figure can be viewed at wileyonlinelibrary.com]

| Tests and treatment for thyroid function
To examine thyroid function, we measured serum TSH and TPOAb levels using a commercial electrochemiluminescence immunoassay (normal range: 0.56-4.30 μIU/ml, Roche Diagnostics) and an enzyme-linked immunosorbent assay (normal range: <16.0 IU/ml; Roche Diagnostics), respectively. The treatment threshold for thyroid abnormalities using levothyroxine in the patients with RPL was TSH ≥2.5 μIU/ml. All women with levothyroxine treatment maintained TSH levels <2.5 μIU/ml and continued until live birth.

| Tests and treatment for thrombophilia
Thrombophilia tests consisted of measuring serum levels of protein C and S activities, factor XII, antiphospholipid antibodies including lupus anticoagulant, anticardiolipin antibodies including IgG and IgM and anti-β2-GP1 antibodies including IgG and IgM. When positive antiphospholipid antibody was confirmed, the detected antibody was remeasured 3 months later and the decision to administer lowdose aspirin (Bafferin Combination Tablet A81, 81 mg; Eisai Co., Ltd.) was made. Treatment with low-dose aspirin was started from the day of the positive pregnancy test. In some women with a history of RIF, daily aspirin intake was initiated from 10 days after ovulation or the start of progesterone administration in embryo transfer cycles, as in our previous trial. 11 We administered low-molecular-weight heparin to only one woman with APS who had a clinical pregnancy loss with normal karyotype even with treatment with low-dose aspirin.

| Treatment for women without risk factors
When any risk factors, except for vitamin D insufficiency, were not detected after RPL testing, we recommended supplementation of 30 mg dydrogesterone tablets (Duphastone 5 mg, Abbott Japan LLC, Tokyo, Japan) three times daily until 12 weeks of gestation. Deficient or inadequate progesterone secretion during the luteal phase is associated with pregnancy loss, 43 and unexplained RPL is involved in perturbed decidual change of the endometrium. 44,45 Delayed implantation timing of an embryo is also linked to an increased risk of subsequent pregnancy loss rates. 46 Therefore, luteal support with progesterone treatment has a potential therapeutic effect on inducing decidualization of the endometrium and optimizing the timing of implantation, leading to pregnancy loss prevention.

| Statistical analysis
All statistical analyses were performed using Statistical Analysis System version 9.4 (SAS Institute, Cary, NC, USA). Statistical significance was calculated using the t test for differences in continuous variables, and chi-square and Fisher's exact tests were used as appropriate to test the statistical significance of the categorical variables. To identify predictive factors that affect pregnancy outcomes after the OPTIMUM treatment strategy in women with RPL, we constructed a multivariable logistic regression model to determine the independent factors while controlling for confounders.
Odds ratios (ORs) and their 95% confidence intervals (CIs) were computed. The level of significance was defined as p < 0.05.  The RPL testing showed that 60 (52.2%) of the RPL women had two or more risk factors, including 5 (4.3%) who had all four risk factors.

| Prevalence of risk factors for pregnancy loss
The relationships among four risk factors for pregnancy loss were irregular and complicated. However, there were no risk factors in 11 women (9.6%).

| Treatment for CE and aberrant high Th1/Th2 ratios
Of the 60 women with CE, CE was detected in five and two women after hysteroscopic polypectomy and adhesiolysis, respectively.
After surgery, recovery from CE was confirmed with reexamination of endometrial biopsy without antibiotics in five women. In two women, CE was cured by treatment using bacterium-sensitive antibiotics, as in our previous report. 37

| Pregnancy outcomes after OPTIMUM treatment strategy
To investigate the efficacy of the OPTIMUM treatment strategy for preventing pregnancy loss in women with RPL, we compared pregnancy outcomes between those who received the OPTIMUM treatment strategy and those who did not. Table 2

| Predictive factors for pregnancy outcomes after the OPTIMUM treatment strategy
After the OPTIMUM treatment strategy, 87.7% of the women aged The OPTIMUM treatment strategy is the treatment procedure not only for RPL but also for infertility. 11 However, unfortunately, 23 women could not reach clinical pregnancy for 6-14 months after the OPTIMUM treatment strategy. Therefore, we also analyzed the predictive factors in 113 women including 70 women who had a successful live birth within two pregnancies (success group) and 43 women with no pregnancy or with pregnancy loss (failure group;  Infertility is a predictive factor for pregnancy prognosis in the OPTIMUM treatment strategy. In our study, 95.5% (21/22 women) of the women without infertility had their babies in the first pregnancy; therefore, the OPTIMUM treatment strategy has adequate preventive effects on repeated pregnancy losses among women without infertility. However, the miscarriage rate in the women with both TA B L E 3 Predictive factors for therapeutic effects of OPTIMUM treatment strategy in the women with clinical pregnancy Live birth group a n = 62 Miscarriage group a n = 28 p-value To identify the predictive factors for the therapeutic efficacy of the OPTIMUM treatment strategy, we compared the 62 women who had successfully childbirth (live birth group) and the 28 women who ended in pregnancy loss (miscarriage group) in the first clinical pregnancy after the OPTIMUM treatment strategy.

TA B L E 4 Predictive factors for therapeutic effects of OPTIMUM treatment strategy in the women with and without clinical pregnancy
Success group a n = 70 Failure group a n = 43 p-value  62 Although there have been only few reports on the pregnancy outcomes of women with untreated CE, a relationship between unexplained RPL and local endometrial proinflammatory response with aberrantly increased uNK cells and angiogenesis has been reported. [19][20][21] Among proinflammatory mediators in women with unexplained RPL, IL33/ST2 activity is disordered in the human endometrium. 63 Furthermore, nucleotide polymorphisms in NLRP, which is a key gene in the regulation of IL-1β, are also involved in RPL. 64  Progesterone is an essential hormone for decidualization of the human endometrium and pregnancy maintenance. Progesterone can activate the secretion of endometrial cortisone during decidualization of the endometrium, leading to direct and indirect regulation of abnormally increased uNK cells and angiogenesis and optimization of immune tolerance for an embryo. [65][66][67] In addition, progesterone can reduce contraction of uterine smooth muscle and prostaglandin production. 68 Therefore, progesterone is a potential inhibitor of pregnancy loss. In a randomized controlled trial of vaginal progesterone suppositories for women with unexplained RPL, no significant difference was found between pregnancy outcomes in women receiving progesterone and those receiving placebo. 69 However, a systematic review found that synthetic progestogen, dydrogesterone, has a therapeutic effect on unexplained RPL. 70 and has no adverse effects on pregnancy or the fetus. The use of progesterone and vitamin D can be considered as treatment for unexplained RPL.
The OPTIMUM treatment strategy is the first treatment procedure for not only RIF but also RPL. However, the OPTIMUM treatment strategy for RPL is different from that for RIF as follows: the initiation of tacrolimus intake from positive pregnancy test for RPL and 1 day before the day of embryo transfer for RIF; the start of lowdose aspirin intake from positive pregnancy test for RPL and 10 days after ovulation or the initiation of progesterone administration in embryo transfer cycles; medical interview and recommendation of lifestyle habit improvement for RPL; and progesterone treatment for RPL without any risk factors for pregnancy loss.
This study has some limitations. First, this is a retrospective cohort study. Therefore, the patients in the control group did not desire or were not suggested RPL testing. It might cause a selection bias of this study. Second, there are still no global standard diagnostic criteria for CE. The presence of five or more CD138-positive plasmacytes in 10 nonoverlapping random stromal areas was defined as CE.
The different criteria might result in different outcomes. Third, our recent report showed that, according to data from fertile women, a high Th1/Th2 cell ratio was >11.8. 55 However, subjects receiving tacrolimus treatment were defined as having a Th1/Th2 cell ratio >10.3 in this study, based on previous data in infertile women with a history of childbirth. 40 Fourth, data from the chromosomal analysis of the couples were not included in this study.
In conclusion, this is the first report of the OPTIMUM treatment strategy for women with a history of RPL. Combination treatment for abnormalities of intrauterine circumstances, Th1/Th2 cell balance, thyroid function, and thrombophilia can improve pregnancy outcomes in the women with RPL. However, among women with advanced age, diminished ovarian reserve or infertility, fertility treatment including ART should be performed to shorten the time to live birth. In addition, this treatment cannot cover pregnancy loss with embryonic chromosomal abnormalities; therefore, combining the OPTIMUM treatment strategy with PGT-A may produce further efficacy expected to prevent pregnancy loss in women with advanced age.

ACK N OWLED G EM ENTS
We thank all participating patients in this study of the OPTIMUM treatment strategy. We thank Ms. Sayaka Kitagawa of Sugiyama Clinic Shinjuku for assisting with data collection. Keiji Kuroda was supported by JSPS KAKENHI grant 18K09273.