Recurrent miscarriage
Definition
The definition of RM has been subject of much debate as evidenced by differences in definitions from the European Society of Human Reproduction (ESHRE), the American Society of Reproductive Medicine (ASRM), and the Royal College of Obstetricians and Gynaecologists (RCOG).
More recently, the term recurrent pregnancy loss (RPL) has been adopted. The ASRM defines RPL ‘as a disease distinct from infertility, defined by two or more failed pregnancies' (19). The International Committee for Monitoring Assisted Reproductive Technology and the WHO revised glossary of ART terminology define recurrent spontaneous loss/miscarriage as the spontaneous loss of two or more clinical pregnancies (20). Previously ESHRE's Special Interest Group in Early Pregnancy defined recurrent abortion as three consecutive first- trimester clinical losses or two consecutive miscarriages, however a more recent revision states that “A diagnosis of RPL could be considered after the loss of two or more pregnancies” (21, 22, 23). RM is defined by the RCOG as the loss of three or more consecutive pregnancies (24). These differences in definition are important as a much greater number of women have two consecutive early pregnancy losses, followed by a successful pregnancy, than have three consecutive losses. While it is much easier to recruit women with two consecutive losses into research studies on RM, the background (non-treatment related) successful pregnancy rate will be much higher in these studies than in those that recruit only according to the more strict ‘three consecutive losses' definition. This must be kept in mind when interpreting research data.Epidemiology
Unlike sporadic miscarriage with a prevalence of 12-15% in the reproductive population, RM is a much rarer phenomenon occurring in 1-3% of couples attempting conception. Distinct from sporadic miscarriage, the repetitive nature of RM introduces a heavy psychosocial burden to the couple and a unique challenge to clinicians involved.
However, this prevalence is much higher than the statistical permutation of consecutive sporadic miscarriage, suggesting a unique pathophysiological entity. Patients with RM have poor reproductive performance across gestation, with higher prevalence of adverse pregnancy outcomes such as preterm delivery, intrauterine growth restriction, and pre-eclampsia in the latter part of pregnancy.Various medical conditions have been associated with RM although the strengths of these associations are variable. The two strongest associations with RM are embryonic karyotype abnormalities and APS. Other described associations, albeit of weaker strength, include endocrine disturbances, autoimmune disorders, hereditary thrombophilia, and structural uterine abnormalities.
Age (25) and the number of miscarriages are the two most important patient-a ssociated risk factors for RM. However other patient-related risk factors such as obesity (26) and exposure to toxins such as smoking, alcohol, and caffeine are increasingly being recognized as contributory to RM.
The immediate and long-term sequelae associated with RPL are often underestimated and include anxiety, depression, post- traumatic stress disorder, and relationship breakdown (27, 28).
Associated factors
In about 50% of patients investigated for RM, an underlying medical condition will be found. Commonly investigated conditions include genetic, endocrine, autoimmune, thrombotic, and uterine structural abnormalities. However, apart from embryonic karyotype abnormalities, APS, and uterine structural abnormalities, the strength of association between endocrine, autoimmune, and hereditary thrombophilia and RM is comparatively weak.
Genetic
Similar to sporadic miscarriage, the commonest cause of RM, particularly recurrent early miscarriage, is embryonic karyotype abnormalities (29, 30). Commonly identified karyotype aberrations in RM include trisomies 21, 16, and 18 as well as triploidy and monosomy X (31).
Detection of an embryonic karyotype abnormality in RM not only unravels underlying aetiology but has prognostic value, as patients with aneuploidy have better livebirth outcomes in subsequent pregnancies (29, 31, 32). More recent and cost-effective approaches to investigating RM recommend initial karyotyping of products of conception and only proceeding to exclude conventional associations (e.g. endocrine, thrombophilia, and APS), in euploid karyotypic losses (33, 34). A practice of routine embryonic karyotype analysis reduces the prevalence of truly unexplained RM from 50% to 24.5% (29).Investigating for parental karyotype abnormalities to exclude balanced and/or Robertsonian translocations has not been shown to be cost-effective (35). ASRM recommends screening for parental karyotype abnormalities (36), RCOG recommends reactive screening based on a finding of an unbalanced structural embryonic karyotype abnormality (24), while ESHRE recommends selective screening based on maternal age at second miscarriage, number of miscarriages and family history of RM (in parents and siblings) (22). However ESHRE's recent guideline revision states that “Parental karyotyping is not routinely recommended in couples with RPL. It could be carried out after individual assessment of risk” (23).
Endocrine factors
Endocrine factors with reported associations with RM include luteal phase insufficiency, poorly controlled or occult diabetes, polycystic ovary syndrome (PCOS), thyroid dysfunction, and hyperprolactinaemia.
Luteal phase insufficiency
Progressive oestrogen production in the follicular phase is a key mediator of increased luteinizing hormone production and the ultimate surge required for ovulation. The hallmark of the luteal phase is progesterone production and establishment of the corpus luteum. The corpus luteum continues oestrogen and progesterone production and the latter is responsible for decidualization of the endometrium in preparation for implantation (37).
Progesterone- driven decidualization involves adaptations in the endometrial glandular epithelium and stroma, resulting in, among others, mucin and glycogen production as well as secretion of prolactin, growth factors, and extracellular matrix proteins (collagen, laminin, and fibronectin), which are all involved in enhancing implantation (38).Although luteal phase defect or insufficiency remains a controversial entity with a lack of consensus definition and diagnostic criteria, poor follicular growth, oligo-ovulation, inadequate corpus luteal function, and altered endometrial response to oestrogen are thought to play a role in luteal phase defect (39-41). Luteal phase defect has long been associated with RPL and supplementation with progesterone (progestogens), the key endocrine mediator of the luteal phase, has long been the focus of many studies in RM (42, 43). Due to inconsistencies in the definition of RM (two or three losses), population demographics (age and number of previous miscarriage are confounders for future reproductive performance), interventions (different progestogen formulations, dosages and routes of administration), and primary outcome measures (mostly used miscarriage rate rather than live-birth rate), it is unsurprising that these studies have yielded conflicting results (44).
Until recently, the majority of trials examining the effect of progesterone supplementation dated back to the 1950-1960s with considerable methodological limitations (42, 43). The first metaanalysis to examine this subject was by Daya in 1989, which included three studies (45). Different progesterone formulations and routes of administration were used and none of the included studies were sufficiently powered to detect a clinically significant difference in outcome, but pooling resulted in an odds ratio (OR) for pregnancies reaching at least 20 weeks' gestation of 3.09 (95% CI 1.28-7.42). In 2013, a subgroup analysis of patients with RM (three or more consecutive miscarriages) in a Cochrane review that investigated progestogen for preventing miscarriage, showed a statistically significant reduction in miscarriage in favour of those randomized to the progestogens group (Peto OR 0.39; 95% CI 0.21-0.72) (46). This subgroup analysis was based on four trials, three of which were from the 1950s and 1960s.
The authors highlighted interpretation of results with caution as numbers were small and the trials were of poor methodological quality. In a systematic review examining the effect of dydrogesterone (Duphaston) on RM by Carp (47), only 3 out of13 studies were eligible for analysis due to study quality and methodological inadequacies in the others. Of the eligible studies, one was randomized, one quasi-randomized and the other was an open-label study. The conclusion of this systematic review with 409 women for analysis was that administration of dydrogesterone was associated with a 29% odds of miscarrying (OR 0.29; CI 0.13-0.64) (47).Recently, the results of the Progesterone in Recurrent Miscarriages (PROMISE) trial, the first multicentre, double-blind, placebo- controlled, randomized trial to investigate whether 400 mg of mi- cronized vaginal progesterone in the first trimester of pregnancy would improve pregnancy outcomes in women with unexplained RM, have been published (48). With live-birth rate as the primary outcome and 836 women enrolled, the PROMISE trial reported no significant differences in live-births or maternal and neonatal outcomes in women who were treated with placebo or progesterone in the first half of pregnancy. A recently published randomized controlled trial involving 700 women with unexplained RM (350 in the placebo group and 350 administered 400 mg vaginal progesterone twice daily) started periconceptionally (in the luteal phase prior to pregnancy confirmation), reported significant reduction in miscarriage rate (12.4 vs 23.3% in the placebo group; P = 0.001) and improvement in pregnancy continuation and live-birth rate in the progesterone group (87.6 vs 76.7% and 91.6 vs 77.4%, respectively; P antibodies (67). Various mechanisms have been postulated to explain the association between thyroid autoimmunity and RM including binding of autoantibodies to placental or trophoblast antigens, thyroid-stimulating hormone receptor antibodies blocking luteinizing hormone receptors on the corpus luteum, and an association with infertility and delayed age at conception.
ESHRE and ASRM recommend thyroid function testing as part of RM workup (22, 36). Additionally, ESHRE strongly recommends testing for thyroid peroxidase (TPO) antibodies in women with RPL (24).Obesity
Increased body mass index (BMI) is a recognized risk factor for sporadic miscarriage in both spontaneous and assisted conceptions (68, 69). The role of BMI in RM still needs further elucidation but is likely mediated, among others, by hyperinsulinaemia, hyperandrogenism, and leptin. In a nested retrospective case-control study, Lashen et al. found that obese women (BMI >30 kg/m2) had a higher spontaneous early and recurrent early miscarriage rate in comparison to normal weight, age-matched controls (26). Furthermore, Metwally and colleagues reported a significantly higher risk of miscarriage in subsequent pregnancies of obese and underweight patients with RM (70). Recently, we reported a very high prevalence of obesity in our RM population (64). Obesity does not seem to increase the frequency of aneuploidy embryos in RM (71), but poor oocyte quality and/or endometrial molecular and endocrinological disturbances may be implicated (72). With increasing global burden of obesity, RM may become a less rare phenomenon, and simple strategies such as prepregnancy weight reduction will likely have an impact in obviating pregnancy loss.
Antiphospholipid syndrome
According to the international consensus criteria for APS, APS is present when one clinical and one laboratory criteria are met (73). The clinical criteria consist of vascular thrombosis or pregnancy morbidity (one or more unexplained deaths of a morphologically normal fetus after the tenth week of gestation; one or more premature births of a morphologically normal neonate before 34 weeks' gestation as a result of eclampsia, pre- eclampsia, or recognized features of placental insufficiency; and three or more consecutive spontaneous miscarriages before the tenth week of gestation with maternal anatomical, hormonal, and chromosomal abnormalities excluded). The laboratory criteria consist of positive plasma lupus anticoagulant (LAC), anticardiolipin antibodies (ACA) (>40 GPL or MPL units or >99th centile), and/or anti-beta-2 glycoprotein-1 antibodies (>99th centile) on two or more occasions at least 12 weeks apart.
Antiphospholipid antibodies are directed against anionic phospholipid-binding plasma proteins. Although previously thought to be implicated in pregnancy loss through a thrombotic mechanism, there is recent evidence that antiphospholipid antibodies inhibit trophoblast function and differentiation, suppress expression of key adhesion molecules (alpha-1 integrins, E and VE cadherins) (74), and/or activate complement pathways at the maternal-fetal interface resulting in a local inflammatory response (75). The risk of RPL at less than 13 weeks' gestation in the presence of immunoglobulin G ACA has been reported as 3.16 (95% CI 1.48-8.59) and for anti-beta-2 glycoprotein-1 antibodies 2.12 (95% CI 0.69-6.53) (76). The risk of RPL at less than 24 weeks' gestation with positive ACA antibodies is 5.39 (95% CI 3.72-7.82) and with positive LAC 7.79 (95% CI 2.30-26.45). Among medical conditions associated with RM, APS has the strongest association with RM and is the only one for which treatment has been shown to improve chances of livebirth.
Heparin decreases antiphospholipid antibody binding to trophoblasts, increases cleavage of beta-2 glycoprotein-1, decreases complement activation and trophoblast apoptosis, and enhances trophoblast invasiveness and expression of essential growth (77). There is now abundant evidence that heparin, in combination with aspirin, improves chances of livebirth in APS-related RPL (78-81). However, the combination of heparin and aspirin, or aspirin alone, does not improve pregnancy outcomes in non-APS-related losses, such as in women with unexplained RM (82-84). Most societal guidelines, including those from ESHRE, ASRM, and RCOG, recommend testing for APS in patients with RM (22, 24, 36).
Hereditary thrombophilia
Unlike acquired thrombophilia (APS), the strength of association between RM and hereditary thrombophilias such as factor V Leiden (FVL) mutation, prothrombin gene mutation (G20210A), methylene tetrahydrofolate reductase (MTHFR) gene mutation (C677T), and deficiencies in natural anticoagulants (protein C, protein S and antithrombin III), is rather weak and controversial. While some studies have reported association between FVL, prothrombin gene mutation (PGM), and RPL (85, 86), others have found no association between hereditary thrombophilic defects and RPL (87-89).
The two most common hereditary thrombophilias are PGM and FVL. The first meta-analysis to examine the association between hereditary thrombophilias and fetal loss (not exclusively RPL) was by Rey et al. (90). In this meta-analysis, FVL (OR 2.01; 95% CI 1.133.58), activated protein C resistance (3.48; 95% CI 1.58-7.69), and PGM (2.56; 95% CI 1.04-6.29) were associated with early recurrent fetal loss (a higher prevalence has been reported in association with RM (94, 96). Of the congenital abnormalities, septate and subseptate uteri were most commonly identified in patients with RM (95, 97). Saravelos and colleagues showed that women with septate or bicornuate uterus suffered from significantly increased second-trimester miscarriages when compared with controls (13.2% and 13.8% vs 1.0%; P underlying molecular pathophysiological mechanism in specific phenotypic categories of RM is the fundamental requisite for the advancement of this field.