The relevance of prothrombotic genetic variants in women who experienced pregnancy loss or embryo implantation failure: A retrospective analysis of 1922 cases

The aim of our study was that to assess the allelic and genotype frequencies of nine prothrombotic gene variants in patients with a history of pregnancy loss and recurrent pregnancy loss (RPL). Women who underwent assisted reproductive technology (ART) with ongoing pregnancy and those with recurrent implantation failure (RIF) were also included.


| INTRODUC TI ON
More than half of all conceptions are lost within the first trimester of gestation.In about 75% of cases, this is due to implantation failure that is frequently not perceived, while remaining cases occur after implantation and are recognized as pregnancy loss (PL).About 15%-20% of pregnant women experience PL within the first trimester, and 1%-5% have two or more consecutive losses.Recurrent pregnancy loss (RPL) is still a matter of debate.According to the American Society for Reproductive Medicine 1 at least two, not necessarily consecutive, miscarriages are required, while the Royal College of Obstetricians and Gynaecologists propose three consecutive PLs (not necessarily intrauterine) to define RPL. 2,3Discrepancies in the definition of RPL regarding gestational weeks were also observed among international guidelines and it is still uncertain whether recurring pregnancy losses should be consecutive or not. 3 In addition to implantation failure in women who had a natural conception, failed implantation was observed in more than 60% of assisted reproductive technology (ART) cycles.Recurrent implantation failure (RIF) is defined as the failure to achieve clinical pregnancy after transfer of at least four good-quality embryos in a minimum of three fresh or frozen cycles, in a woman under the age of 40 years. 46][7] Despite chromosomal, uterine, immunological, infectious, and hormonal alterations being responsible for about half of RPLs, 8 the causes of the other half are unknown.
Conversely, for implantation failure, causes seem to be more related to the impaired endometrial receptivity 9 that may be constitutive 10 or due to immunological dysregulation. 11percentage of RPLs are due to placental thrombosis and infarction. 12Similarly, the altered development of the novel architecture of uterine vasculature, which is necessary to establish uteroplacental circulation, may impair the implantation process, causing RIF.
Inherited thrombophilia may predispose to pregnancy loss or to implantation failure by impairing placental function. 13,14Some variants in genes involved in hemostasis and its inhibition display a procoagulant effect. 15The most well-known prothrombotic variants include factor V Leiden (FVL, G1691A variant), which causes resistance to protein C; the G20210A variant of prothrombin, which enhances levels and activity of factor II (FII); and the C677T variant of methylene-tetrahydrofolate reductase (MTHFR), which reduces activity of the enzyme, causing an increase of serum homocysteine, a known trigger for coagulation at endothelial level. 16,17In addition, other gene variants have been studied in females who experienced various forms of obstetric complications, such as factor V, H1299R variant (FVR2), MTHFR A1298C, factor XIII (FXIII) V34L, human platelet antigen (HPA)-1 L33P, beta-fibrinogen -455G>A variants, and plasminogen activator inhibitor (PAI)-1 4G/5G alleles. 15ear evidence of a pathogenic role of thrombophilia could allow targeted therapies, but results on the role of such mutations as risk factors for pregnancy loss so far have been controversial, and frequently antithrombotic overtreatments are used. 18Similarly, screening for thrombophilia in women who undergo ART is controversial. 19ere are several reasons for these discordances: (1) heterogeneous criteria have been used to define RPL and RIF in most studies to date; (2) some studies compared the frequency of prothrombotic gene variants in patients and controls of different geographical areas; and (3) most studies analyzed a scarce number of cases.[22] Moving on from the idea that prothrombotic status may impair the outcome of the pregnancy, the tendency is to widely investigate thrombophilia in all women who experience obstetric complications and RIF.Nevertheless, information from adequately sized subgroups of patients is lacking.
The aim of our study was that to assess allelic and genotype frequencies of nine prothrombotic gene variants in patients with obstetric complications, including PL and RPL, and in women who underwent ART, including those with a history of RIF.A large group of subjects from the general population of the Campania region (southern Italy) served as control group.

| Patients
The study was approved by the Ethical Committee of the University Federico II, Naples, Italy (protocol no.370/18).Our laboratory acts as the reference laboratory for molecular diagnostics in the Campania region (which has about 5 million inhabitants).During the last 12 years (i.e., 2007-2018), we have received several thousand requests for molecular analysis of thrombophilia for different diseases.Here we report the results of a retrospective analysis of data deriving from women who underwent ART or who were referred due to various obstetric complications; these were compared with a group of subjects from the general population of the same geographical area (Campania region).Informed consent was obtained at the time of enrollment for molecular analysis of all samples to be used anonymously for research purposes.
For each subject, we recorded all personal and clinical data.In cases of patients belonging to the same family, we considered only the first case temporally analyzed.In our study population, we excluded women over the age of 40 years, considering the impact assisted reproductive technology, genetic variants, ovarian stimulation, polymorphism, recurrent implantation failure, recurrent pregnancy loss, thrombophilia of advanced maternal age on miscarriage and defective implantation. 23 (Table 1).

| Genotype analysis
A blood sample was collected by venipuncture using the Vacutainer system in EDTA tubes.DNA was extracted from leukocytes using a commercial automated procedure (Roche, Basel, Switzerland).The DNA was spectrophotometrically quantified (also to verify the purity) and analyzed by real-time polymerase chain reaction (Roche) for FVL (G1691R); FVR2 (H1299R variant); FII, G20210A; MTHFR, C677T and A1298C; beta-fibrinogen, -455G>A; FXIII, V34L; HPA-1, L33P variants; and PAI-1 4G/5G alleles.For some subjects, data on some prothrombotic variants are missing because the molecular analysis was not carried out.

| Statistical analysis
Both allele and genotype frequency were reported as absolute numbers and percentages.Differences among groups were accordingly assessed using the Chi-squared test.Moreover, in order to quantify the effect of each variant on the disease risk, univariate odds ratios (ORs) with the corresponding 95% confidence intervals (95% CIs) were computed.Multivariable logistic regression models were built to assess whether age could act as effect modifier of the variant-disease association.As no interaction reached statistical significance, ORs were reported for the whole study population without any adjustment.Statistical analyses were performed with the statistical platform (Core Team 2018, a language and environment for statistical computing Foundation for Statistical Computing, Vienna, Austria; https:// www.r-proje ct.org/ ).A P-value <0.05 was considered statistically significant.

| General population
The allele and the genotype frequency of each gene variant are displayed in Tables 2 and 3 (column "GP").No differences were observed between males and females in either the allelic or the genotype frequency (data not shown).

| Women who experienced one or two PLs
For each of the two groups, none of the allele frequencies of the nine variants were significantly different with respect to subjects from the general population (Table 2).This result was mirrored by the distribution of the genotype frequency of all nine variants, which was not significantly different between the groups of women who experienced one or two PLs and the general population (Table 3).

| Women who experienced RPL
The allele frequency of FVL was significantly higher in females who experienced RPL than in subjects from the general population.For all other variants, we did not find any significant difference (Table 2).This result agrees with the higher (even if not significant) frequency of subjects bearing the variant in the group of women with RPL as compared with the general population (Table 3).Furthermore, we observed a significant difference of the beta-fibrinogen -455G>A genotypes between females who experienced RPL and the general population due to the higher frequency of the heterozygous haplotype in the RPL group (Table 3), even if the allele frequency of the variant was higher (though not significantly) in this group than in the general population.Finally, we observed a higher occurrence (although not significant) of the Abbreviations: 1PL, an episode of pregnancy loss before 14 weeks of gestation; 2PL, two episodes of pregnancy loss before 14 weeks of gestation; GP, general population; MAPP, medically assisted procedure followed by pregnancy; RIF, recurrent implantation failure; RPL, recurrent pregnancy loss.
heterozygous genotype for PAI-1 among women who experienced RPL as compared with the general population.Each of the three variant genotypes (i.e., FVL, -455G>A of beta-fibrinogen, and PAI-1) determined a significant increase in the OR of disease in subjects bearing the variant (Table 4).

| Women who underwent MAPP
None of the allele frequencies of the nine variants were significantly different in women who underwent MAPP with respect to subjects from the general population (Table 2).This result was confirmed by the distribution of the genotype frequency of all the nine variants, which was not significantly different between the groups of women who experienced MAPP and the general population (Table 3).

| Women who experienced RIF
The allele frequency of FVL, FII G20210A, and MTHFR 677T variants were significantly higher in women who experienced RIF than in subjects from the general population.For all other variants, we did not find any significant difference (Table 2), even if the frequency of FV H1299R was higher (although not significant) in the group of women who experienced RIF.These results were mirrored by a different distribution of the genotype frequency of FVL, FVR2, FII G20210A, and MTHFR 677 T between the group of women with RIF and the general population.(Table 3).Each of the four variants determined a significant increase in the OR of disease (Table 4).

| DISCUSS ION
The investigated prothrombotic genetic variants did not increase the risk of developing one or two episodes of PL.Such findings implied that overtreatment and the request for expensive prothrombotic genetic evaluation should be avoided in women who have experienced one or two PLs.
Conversely, FVL, the -455G>A variant of beta-fibrinogen, and PAI-1 4G were significantly associated with the risk of RPL.Various studies concluded that FVL and PAI-1 4G confer a higher risk of RPL, while other variants, such as FII G20210A and MTHFR C677T, do not. 24,25In contrast to our data, Yenicesu et al., 26 who studied 272 women with RPL (defined as two PLs) and only 56 controls, found that FVR2, FII G20210A, and PAI-1 4G enhance the risk of RPL.The same findings were supported by another study. 27Concerning the beta-fibrinogen -455G>A variant, the present study confirms our previous results regarding the higher risk of PL in women with celiac disease who had the variant 28 and those of another group for whom it was revealed that such a variant is a risk factor for recurrent miscarriage. 29th regard to ART, our data confirm that women in which the technique was followed by pregnancy do not have any increase in the frequency of prothrombotic variants, 30,31 thus excluding prothrombotic tendency as a leading cause of infertility.While FVL, FVR2, FII G20210A, and MTHFR significantly enhance the risk of RIF, our data partially agree with a study that demonstrated a higher frequency of FII G20210A and MTHFR C677T, but not of FVL, in 45 women undergoing in vitro fertilization compared with 44 control women 32 ; yet our data disagree with another study which excluded the role of prothrombotic variants as risk factors for RIF, comparing only 50 women with RIF with 50 control women. 33Finally, the present study partially agrees with Q ublan et al. and other authors, 9,34,35 who found that inherited thrombophilia such as FVL and MTHFR C677T could be considered risk factors for RIF, suggesting that women who experience these conditions can be screened for thrombophilia.
Indeed, we suggest that the evaluation of specific genetic traits,  to the nodal role of placental vascularization in the process of implantation, which may be impaired at different levels.FVL could enhance the risk of thrombi within the placental microcirculation; FII G20210A and PAI-1 4G could cause an excess of fibrin deposit in the intervillous space; and MTHFR mutations could reduce the production of folates, necessary for embryo development. 9e main limitation of the present study is the low stringency in the exclusion criteria; therefore, further studies are needed to characterize the role of prothombotic variants in relation to the risk of PL even in the presence of specific clinical aspects.In addition, we were not able to add other variables (apart from age) to the logistic regression analysis, due to the inconsistency and lack of information in our records.
In conclusion, some prothrombotic variants could influence the initial stage of gestation, 36 promoting the development of polimicrothrombosis at the site of implantation, impairing the invasion of maternal vessels by syncytiotrophoblast, and causing underperfusion of the embryo.Thus, in women who access ART, it may be useful to test for a panel of selected gene variants, helping to reduce the risk of RIF, while the prothrombotic variants are less involved in reducing perfusion to the intervillous space occurring around the 10th week and impairing the vascular intercommunication between the mother and the fetus, predisposing to PL.These data could have a relevant impact, by allowing healthcare providers to avoid expensive genetic investigations and unjustified anticoagulant therapies, 37 which in some cases may be harmful. 38A B L E 4 Odds ratios (ORs) and 95% confidence intervals (95% CIs) of the factor V (FV) G1691R, factor V (FV) H1299R, factor II (FII) G20210A, methylene-tetrahydrofolate reductase (MTHFR) C677T, beta-fibrinogen -455G>A and plasminogen activator inhibitor-1 (PAI-1) 4G variants for recurrent pregnancy loss (RPL, n = 468) and recurrent implantation failure (RIF, n = 282).