In well over 50% of couples seeking consultation for REPL, a dominant, credible cause for pregnancy losses will not be found. Given this reality, our initial approach is to advise the couple that REPL is regarded by most experts as multifactorial in nature. We mention that advancing maternal age and the number of prior losses are among the factors associated with REPL and that professional guidelines recommend evaluations that we will complete whenever possible.
Lifestyle and Environmental Factors
Whether maternal smoking or alcohol consumption is associated with REPL is uncertain, although each has been associated with SEPL in retrospective studies.22
One retrospective, case-controlled study found that an estimated caffeine consumption of more than 150 mg/d was associated with REPL.23
General physical activity levels have not been associated with REPL and are not linked to sporadic miscarriage.
Obesity (body mass index [BMI] > 30 kg/m2
) is associated with REPL, with a recent systematic review noting an odds ratio (OR) of 1.75 (95% confidence interval [CI] of 1.24-2.47).24
In a prospective study, women with obesity were found to have a higher frequency of euploid miscarriages,25
and another study showed that women with obesity having REPL were more likely to have a subsequent miscarriage.26
Although being underweight has been linked to sporadic miscarriage, its relationship to REPL is unknown.
While chromosomal abnormalities are an obvious cause of a majority of SEPLs, the rate is modestly lower in REPL.27,28
In support of this, a case-controlled study of 420 abortus specimens showed that significantly more women who were older than 36 years with REPL had euploid products of conception compared to unselected controls.27
If, as most clinicians assume, an abortus with aneuploidy is a sporadic occurrence, its occurrence in a patient with REPL might predict a more favorable outcome in the next pregnancy.29
Emerging evidence suggests that EPLs are, at least in some cases, associated with gain or loss of subchromosomal amounts of genetic material, discernible using such techniques as microarray.30
Given that most of these so-called microdeletions or microduplications are of uncertain pathological significance, they are referred to as copy number changes (CNCs). To be sure, some CNCs are more likely to represent a causative factor in pregnancy loss, including those of large CNC size, deletions (rather than duplications), CNCs including genes
described in the Online Mendelian Inheritance in Man database, and changes not described in data-bases of healthy individuals.
In 3% to 5% of couples with REPL, one of the partners is found to have a balanced chromosomal rearrangement, a figure that is at least several-fold higher than in the general population.31
Most mutations are reciprocal balanced translocations; Robertsonian translocations, sex chromosome mosaicisms, and chromosomal inversions are less commonly found.32
In REPL couples, balanced translocations are more commonly found in the female partner, and it appears that the risk of sub-sequent EPL is increased if the translocation originates from maternal side.33
Moreover, a woman with a first-degree relative with REPL has a two- to sixfold increased risk of REPL herself.33,34
Although there is some evidence that otherwise unexplained REPL is associated with an increased rate of aneuploid sperm and apoptosis,35
the role of such paternal gamete meiotic errors in REPL is uncertain; this is an area that deserves further, more sophisticated study. A case-controlled study investigated the additional risk factors in REPL couples that increased the risk of parental karyotype abnormalities.36
Younger maternal age at second miscarriage, history of ≥3 miscarriages, or having a sibling or parent with >2 miscarriages were found to be associated with higher prevalence of parental karyotype abnormalities.
In addition to numerical and structural chromosomal abnormalities, some single gene disorders are known to be associated with REPL. The most common ones are metabolic disorders and hemoglobinopathies, although they contribute to a very small proportion of REPL cases.32
Skewed X inactivation, a phenomenon of preferential inactivation of one X chromosome instead of random inactivation of either, has been considered as a cause of REPL. One study found that the rate of skewed X inactivation was threefold higher in women with REPL compared to those in the general population,37
while another did not find any association between skewed X inactivation and REPL.38
Interpretations of the studies that examine an association between REPL and uterine malformations are limited by variation in the definition of recurrent loss and the use of different malformation classifications systems.39
That said, most experts would accept that uterine malformations are found in 10% to 25% of women with REPL40
with a mean of 12.6% based on results from various studies.22
By comparison, a malformed uterus (including arcuate uterus) is found in approximately 4% to 8% of women seeking gynecologic care for a variety of reasons other than pregnancy loss.39,41
It also has been shown that in women undergoing three dimensional (3D) ultrasound screening for reasons other than pregnancy loss, a history of pregnancy loss was more common in women with uterine malformations than in women with a structurally normal uterus42
; a history of first-trimester loss was found in 42% of women with a subseptate uterus and 16% of women with arcuate uterus. The most common malformations found in women with REPL are septate, subseptate, or arcuate anomalies, followed by bicornuate uterus (Figure 1.1
). Unicornuate uterus is uncommon.
The cause of REPL in women with uterine malformations may be implantation on a poorly vascularized septum, although this mechanism is supported only by indirect evidence. Whether the degree of uterine cavity distortion, for example, the length of the uterine septum or the reduction in the length of the remaining cavity, is associated with a higher likelihood of pregnancy loss is debatable. Work in murine genetic models suggests that dysregulation of stromal or endometrial tissues might play a role in miscarriages associated with uterine malformations. Murine Hox
gene mutants have reproducible uterine anomalies that also are associated with reduced uterine stromal tissue and reduced epidermal growth factor receptor expression.43 Hoxa
10 and Hoxa
11 genes mutants have
abnormal uterine and/or stromal development, and normal endometrial Hoxa
10 expression appears to be critical for implantation in adult mice.44
The possibility that some women with REPL and uterine malformations may be found to have altered gene expression in uterine tissues is intriguing, and variants of uncertain significance have been found in HOXA
10 and HOXA
11 in a small number of individuals with sporadic uterine malformations.45
Other gene variants have been associated with Müllerian abnormalities,45
but to date, each has been found only in a small number of individuals.
Figure 1.1 Antiphospholipid syndrome.
Whether acquired uterine abnormalities, such as leiomyomata, intrauterine adhesions, or endometrial polyps, are associated with REPL is not well established. One expert review concluded that multiple and intramural fibroids were associated with an increased proportion of pregnancies ending in miscarriage,46
but the relative risk was less than 2. A prospective cohort of over 5000 women with ultrasound assessment fibroids found no association of fibroids with pregnancy loss before 20 weeks.47
A systematic review of prospective studies concluded that nearly 20% of women suffering miscarriage have intrauterine adhesions, although these are mild in most cases.48
Moreover, it was found that women with recurrent miscarriages were more likely to have adhesions.48
Antiphospholipid syndrome (APS) is a widely accepted cause of REPL, with reports of from 5% to 15% of women affected by REPL testing positive for antiphospholipid (aPL) antibodies compared to 2% to 5% of women without miscarriages.16
International consensus criteria49
recognize three clinically relevant aPLs: lupus anticoagulant (LAC), anticardiolipin (aCL), and anti-β2
-glycoprotein 1 (aβ2
GP1) antibodies. Some experts also test for immunoglobulin (Ig) G and IgM antibodies to phosphatidylserine. APS is discussed in detail in Chapter 40
, but a brief overview is warranted here.
The predominant pathogenic mechanism by which aPL antibodies may cause pregnancy loss has been reviewed.50
Pathogenic aPL antibodies recognize β2
GP1, a glycoprotein constitutively expressed on the cell surface of trophoblast cells, as well as on many other cells, including maternal decidual endothelial cells. In extravillous trophoblasts, aPL antibodies promote pro-inflammatory, antimigratory, and antiangiogenic effects. aPL antibodies also displace annexin V, an inhibitor of coagulation, from the surface of trophoblastic cells.51
aPL antibody-induced immune activation, including complement activation, at the maternal-fetal interface results in pregnancy loss in animal models.52,53
Hypocomplementemia and increased complement activation products have been reported in women with APS and in association with adverse pregnancy outcomes, including miscarriage.54,55
The association of aPL antibodies with REPL is not without controversy.56
Overall, the published studies are highly heterogeneous, with many flawed by the inclusion of subjects with either poorly characterized pregnancy losses, especially regarding gestational age of loss, or the admixture of subjects with pre-embryonic, embryonic, and fetal deaths. Many studies also used nonstandard aPL antibody tests and/or “positive” results that do not meet international criteria definitions of positive. Some studies did not perform repeat testing to confirm positive aPL antibodies in accord with expert guidelines. Over the course of nearly a decade, two experienced groups of investigators have found fewer than 25 women with REPL as an isolated chief complaint (ie, without thrombosis, systemic lupus erythematosus, or fetal death), and with repeatedly positive aPL antibody titers, meeting international criteria.57,58