Intravenous immunoglobulin use in patients with unexplained recurrent pregnancy loss

Recurrent pregnancy loss (RPL) affects up to 4% of couples attempting to conceive. RPL is unexplained in over 50% of cases and no effective treatments exist. Due to the immune system's pivotal role during implantation and pregnancy, immune‐mediated RPL may be suspected and immunomodulatory treatments like intravenous immunoglobulin (IVIg) have been administered but remain controversial. The goal of our study was to evaluate our center's 6 year‐outcomes and to develop a framework for IVIg use in RPL.


INTRODUCTION
Recurrent pregnancy loss (RPL) is commonly defined as ≥2 consecutive intrauterine pregnancy losses prior to a gestational age (GA) 20−24 weeks. 2 Primary RPL occurs in the absence of previous live birth, while secondary RPL occurs after a previous pregnancy is carried ≥24 GA. Depending on maternal age, the method employed to detect pregnancy and the definition of RPL used, RPL is estimated to affect up to 4% of couples. 3 Unfortunately, half of RPL cases remain unexplained despite extensive workup for genetic, anatomic, endocrine, and coagulation abnormalities, further compounding the emotional and financial burdens associated with this disease. 4 While women with 2−3 previous miscarriages have a good prognosis in a subsequent pregnancy; the live birth rate (LBR) is inversely proportional to the number of previous miscarriages and women with ≥5 consecutive RPL can expect a LBR below 45% in future attempts. 5,6 Apart from continued attempts, possibly early progesterone supplementation and low dose aspirin therapy (LDA), there are no treatments available to improve the prognosis in patients with unexplained recurrent pregnancy loss (uRPL). 3 A concept gaining in popularity is the immune contribution to RPL pathogenesis. Indeed, maternal immune cells present in the cycling endometrium play crucial roles in decidualization, embryo recognition, implantation, and spiral artery remodeling. Defective decidual immune remodeling may thus underlie RPL by interfering with the early establishment of maternal tolerance to the semi-allogenic fetus. 7 Unfortunately, there are no validated biomarkers to confirm the diagnosis of immune-mediated RPL when it is clinically suspected, and often, immunomodulatory drugs are started empirically with little scientific evidence of benefit. 8 Amongst these, intravenous immunoglobulin (IVIg) is prominently featured.
IVIg is polyclonal immunoglobulin G (IgG) preparation pooled from donor plasma that been used for over 40 years as an off-label treatment for patients with uRPL. IVIg is well tolerated by an obstetric population 9,10 and acts by multiple mechanisms to decrease inflammation and promote a tolerogenic immune environment ( Figure 1). 11 Despite initial success in small cohorts, IVIg has failed to show a consistent beneficial effect for RPL in a randomized controlled trial (RCT) setting (Table 1). However, many methodological challenges complicate the interpretation of the literature: available RCTs often underpowered and heterogenous in terms of patients recruited, RPL definition as well as IVIg protocols; the lack of biomarkers to diagnose immune mediated RPL renders patient selection difficult. 8 Meta-analysis data have shown that only studies initiating IVIg prior to conception demonstrate improved LBR (Table 1) 12,13 consistent with the theory that an appropriate immune response during implantation dictates the outcome of pregnancy. 14 While some studies have found that IVIg administration is most effective for women with RPL and peripheral blood immune abnormalities, 15 the validity of such testing has been called into question. 16 More recently, a high quality RCT by Yamada et al. has shown that high dose IVIg (0.4 g/kg/day for 5 days) improves LBR in patients with ≥4 unexplained RPL when started at pregnancy diagnosis but not after GA 6 ( Table 1). 17 This introduces the concept that an appropriately timed intervention may rescue an abortive phenotype irrespective of previous immune cell testing. In this study, we aimed to evaluate the safety and efficacy of IVIg in a "real world" RPL setting and retrospectively review our clinic's 6-year IVIg outcomes, compared to expected outcomes without treatment from a separate natural history cohort.

MATERIALS AND METHODS
In the Province of Quebec (Canada), IVIg treatment is publicly funded but tightly controlled, accessible only through hospital-run blood-  (Figure 1, 1), suppressing Th1 cytokine production and skewing towards a Th2 response 44 (Figure 2, 2), as well as by suppressing NK cell cytotoxicity ( Figure 2, 3), 45,46 IVIg's ability to enhance adaptive regulatory responses has gained much interest in the last decade. Indeed, several studies have shown that RPL women have measurable defects in peripheral T regulatory (T reg ) 47 ; women successfully treated with IVIg show increased peripheral T reg numbers as well as enhanced T reg activation status and suppressive function (Figure 2, 4). 48,49,50 While IVIg may improve regulatory adaptive responses through indirect effects such as pro-inflammatory cytokine 51 and complement neutralization ( Figure 2, 5), 52 inhibition of antibody dependent cytotoxicity 45 and restoration of the anti-idiotype network, 53 other mechanisms implicating tolerogenic antigen-presenting cells (APC) have also recently been recently described. One proposed mechanism is through T regulatory cell epitopes (Tregitopes) which are highly conserved moieties contained within all IgG allotypes (Figure 2, 6). Tregitopes exhibit high-affinity binding to human HLA class II complexes and are processed onto APC-MHCII after internalization. This causes the downregulation of APC co-stimulatory molecules CD80 and CD86 and induction of Treg upon antigen presentation. 42 Recent murine studies have shown that Tregitopes effectively maintain immature DC phenotypes, induce T reg and modulate crosstalk between B and Tcells, leading to an expansion of B and T regulatory cell pools, anti-inflammatory cytokine secretion (IL-10, TGF-) as well as enhanced live birth in an abortion-prone mating model. 54 Another similar mechanism (which has not been substantiated in humans or in animal models of RPL) involves a fraction of IgG that contains sialic acid residues on either the fragment crystallizable (Fc) or fragment antigen binding (F(ab) 2 ) portion. The sialylated IgG moiety binds to SIGN-R1 (specific intercellular adhesion molecule 3 [ICAM-3]-grabbing nonintegrin-related 1), a C-type lectin expressed on macrophages ( Figure 2, 7). This causes the macrophage to release soluble mediators which act on effector macrophages to increase their expression of FcγRIIB inhibitory receptors. 24 Similarly, sialylated IgG has been shown to bind to DCIR (dendritic cell immunoreceptor), a C-type lectin expressed on dendritic cells (Figure 2, 8). 55 IgG binding to DCIR RIIB leads to IgG internalization, decreased DC pro-inflammatory cytokine production 56 and negatively regulates DC differentiation and proliferation through immunoreceptor tyrosine-based inhibitory motif (ITIM) signalling. 57 The overall effect is thus enhanced T reg induction through modulation of APC phenotype and reciprocal maintenance of APC tolerogenic phenotype by T reg . Further studies are required but are likely to conclude that IVIg acts by multiple synergistic and non-mutually exclusive mechanisms to restore the ability to tolerate fetal alloantigen in patients with immune-mediated RPL.
-Normal uterine cavity prior to index pregnancy -Absence of endocrinopathy or coagulopathy or failure of ≥1  Patients were excluded from the IVIg cohort if they did not follow treatment protocol (non-compliance or co-treatment with other new immune modulators), if they did not conceive during the treatment period or if they had an explained pregnancy loss on IVIg (aneuploidy, cervical incompetency). Patients were allowed to use co-treatment with other immune modulators (granulocyte colony stim-ulating factor, low molecular weight heparin, corticosteroids) provided they had previously failed ≥1 pregnancy with such treatments. For patients requiring ART for subfertility, infertility or undergoing preimplantation genetic screening (PGT-A), patients were excluded if resorted to donor oocytes, produced poor quality embryos, or with substantial loss of embryo quality upon thaw. Patients undergoing IVF for PGT-A were only included if they had failed a previous euploid transfer or if they had evidence of a previous euploid miscarriage.
Patients with a history of thrombosis, unprovoked cytopenia (anemia, neutropenia or thrombocytopenia) or renal failure were excluded.

Statistical analysis
Characteristics and outcomes between patients exposed to IVIg and controls were compared within the primary RPL group and the sec- patients were allocated to the secondary RPL control group (Figure 1).
The natural history cohort will henceforth be referred to as the "control group".
Patients' baseline characteristics are presented in Table 2

Pregnancy outcome
Overall, patients with primary RPL treated with IVIg had a nonsignificant trend towards more live births than their controls ( Of note, nine patients switched to scIg after receiving 1−2 doses of IVIg. Six patients in the primary RPL group received scIg (two because of IVIg side effects, two because of geographical distance from the clinic and two because of unaccommodating work schedules), five patients had a live birth and one miscarried at 9 weeks gestation. Three patients in the secondary RPL group received scIg because of geographical distance from the clinic, two patients miscarried (biochemical, 6 weeks) and the other had a live birth. Because of low numbers and co-founding factors (IVIg being administered prior to scIg), we cannot comment on the efficacy of scIg for RPL.

IVIg safety
No reported serious adverse effects related to infusion occurred in the IVIg group. Eight patients (8/37, 21.6%) had predictable post-infusion symptoms including headaches (4/37, 10.8%) and mild infusion reaction consisting of myalgia, malaise, fatigue, skin erythema, and pruritus (4/37,10.8%) which resolved with supportive therapy. One patient had a vaso-vagal reaction during her first IVIg infusion and was able to continue subsequent infusions without issue; another had chest pain with normal electrocardiogram and troponin levels during her first IVIg infusion, she was switched to sub-cutaneous immunoglobulin (scIg) without reaction. No patients discontinued IVIg therapy prematurely.

DISCUSSION
Immune dysregulation causing defective establishment or early loss of fetal alloantigen tolerance has been proposed as a mechanism of RPL, 20,21 leading to the use of immunomodulatory treatments in such patients. 22 Amongst these, IVIg is a leading candidate 23 because of its favorable safety profile during pregnancy 9,10 and broad immunomodulatory potential. 24 The exact mechanisms by which IVIg can reverse an abortive phenotype are still being elucidated (Figure 1).
In women with 2−3 previous unexplained miscarriages, live birth on a subsequent pregnancy is favorable, approaching 70% without treat-ment; the expected LBR then plummets below 50% for patients with ≥5 previous consecutive RPL. 34  Currently, it may be reasonable to consider IVIg use in carefully selected patients with ≥5 unexplained RPL. However, we must continue to refine eligibility criteria for IVIg prescription. Thankfully, patients with high order unexplained RPL are a rare patient population, but it is unlikely that a single center will accumulate enough patients for a well powered study over an acceptable amount of time.
Therefore, as a scientific community, we must develop multi-center collaborative efforts implicating both clinicians and basic scientists to understand how IVIg works in RPL, and which patients are most likely to benefit from IVIg. As physicians, to ensure optimal resource allocation we must develop guidelines to regulate IVIg access for RPL patients.
Patients receiving IVIg should be included in national/international registries, characterized according to predetermined clinical and biochemical profiles, and prospectively followed during their treatment.
This enables periodic practice audits, more rapid identification of clinical factors influencing IVIg response and better characterization of birth and neonatal outcomes. As basic scientists, we must prioritize elucidating immune mechanisms of RPL. This requires a narrow collaboration with physicians to obtain relevant patient samples and access to a healthy control population for comparative analysis. Several new technologies are becoming more accessible, namely, single cell RNA sequencing, exome sequencing, special transcriptomics and multiparametric flow cytometry; these will likely provide important insight into normal and pathologic immune response and how this pertains to RPL.
Similarly, while much is known about mechanisms of action of IVIg, little is known about how it improves reproductive outcomes in RPL.
We may discover that the IVIg effect in RPL is dependent upon the highly sialylated IgG moieties or perhaps determined by Tregitopes contained within the IgG molecule. This may lead to the development of engineered IVIg preparation enriched sialic acid residues or in the development of humanized Tregitopes 42 ; reducing our overall use of human plasma. Nevertheless, refining our understanding of IVIg as an immune modulator in RPL and developing markers of IVIg success in this patient population is paramount for appropriate resource allocation.

ACKNOWLEDGMENTS
This research did not receive any specific grant from public funding agencies, commercial, or not-for-profit sectors.

CONFLICT OF INTEREST STATEMENT
None of the authors have any conflicts of interest to declare. This study was approved by the MUHC ethics board, study number MUHC REB # 2022-8157.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.