Frequency of thrombophilia-associated mutations and polymorphisms in pregnant women with a history of thrombosis or pregnancy complications

Objectives: Toassessthefrequencyofmultiplethrombophilia-associated mutations and polymorphisms in a selected population of high-risk pregnancies. Methods: Thrombophilia screening was performed for 1,500 pregnantwomenwith prior pregnancycomplicationsor thrombotic events. Nine thrombophilia-associated mutations or polymorphisms were screened: factor V Leiden, factor V H1299R, prothrombin G20210A, MTHFR C677T, MTHFR A1298C, factor XIII V34L, PAI-1 4G/5G polymorphisms, EPCR G4600A, EPCR C4678G. Results: Out of the 1,500 patients, 1,291 ful ﬁ lled the criteria for data interpretation. All patients had low-risk thrombo-philia-associated genetic variants. Only 1.24 % of cases presented high-risk abnormalities (homozygous factor V Leiden/ prothrombin G20210A, or both mutations in heterozygous form). Heterozygous factor V Leiden occurred in 10.38 % of Ana-Maria


Introduction
Thrombophilia refers to an inherited or acquired predisposition for thrombotic events [1].The predisposing factors for thrombosis include surgery, multiple traumas, prolonged immobilization, malignancy, cardiovascular diseases, various treatments/drugs, prior venous thromboembolism, old age, or pregnancy [1,2].Inherited thrombophilia may increase the risk for thrombotic events in selected cases.Pregnancy, by itself, is a prothrombotic state characterized by an imbalance between the coagulation and fibrinolysis systems, leading to a hypercoagulable state.A combination of a preexisting thrombophilia-associated genetic abnormality and pregnancy may trigger an increased hemostasis imbalance, leading to negative outcomes for both mother and fetus.
Maternal risk for developing thrombotic events (such as stroke, myocardial infarction, or venous thromboembolism) presents since conception, has a peripartum peak, and extends to at least 6-12 weeks postpartum [3].Pregnancy complications include preeclampsia, placental abruption, fetal growth restriction, miscarriage, intrauterine fetal death, and stillbirth [2,4,5].Clinically relevant inherited thrombophilia mutations include factor V Leiden (Arg506Gln or FV Leiden, G1691A polymorphism in the gene encoding factor V) and prothrombin G20210A, widely recognized as contributors to thrombotic events, being also associated with adverse pregnancy outcomes (APO) [6].Of these mutations, the most common mutation in the general population known to date is factor V Leiden (FVL).While the point mutation in the prothrombin gene is also quite often encountered, the deficiencies of the anticoagulant proteins (protein S, C, and antithrombin) are less commonly found [7].Women with homozygous FVL, homozygous prothrombin G20210A, and those presenting both heterozygous FVL and prothrombin G20210A have a particularly high risk for pregnancy-associated thrombosis [7].
Thrombophilia screening should be performed whenever significant factors or events are encountered.The presence of a single screening abnormality is not always a predictor for thrombotic events and is therefore not usually considered for risk evaluation [1].It is important to consider the cumulated importance of all risk factors when deciding to initiate thromboprophylaxis or not.Furthermore, we have encountered in our clinical practice pregnancies with negative outcomes despite normal results of routine thrombophilia testing.Starting from this point of view, we have chosen mutations or polymorphisms that are not commonly checked when evaluating thrombophilia.
The aim of this study was to screen for hereditary thrombophilia in pregnant women with at least one previous pregnancy complication or thrombotic event by using an extended thrombophilia associated mutation/variant panel.To the best of our knowledge, this is the largest thrombophilia screening study performed in our country up to this date.Our results highlight the necessity of implementing a wider thrombophilia screening panel in this selected group of women.

Materials and methods
Through the RO19.10 project ("Improvement of health services for high-risk pregnancy, premature birth and hematological diseases"), the multidisciplinary group from the Emergency University Hospital Bucharest investigated 1,500 pregnant women with ages between 13 and 50, with a history of at least one pregnancy loss, infertility, preeclampsia, or thrombotic events admitted in our hospital between 2015 and 2017.The patients and the legal guardians for patients under 18 signed a written informed consent to be included in this health program.Of the 1,500 cases, data was collected from 1,291 patients, while 209 patients were excluded due to improper sampling/biological samples, inconclusive results, and/or registration errors.This project was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of Emergency University Hospital Bucharest.
Peripheral whole blood was collected from each pregnant woman and genomic DNA was isolated with either the GENXTRACT Blood DNA Extraction kit (ViennaLab Diagnostics GmbH, Vienna, Austria), or the QIAsymphony DNA Mini Kit (Qiagen, Hilden, Germany) using a QIAsymphony automated system (Qiagen, Hilden, Germany).A multiplex PCR amplification of DNA samples, with simultaneous biotin-labeling (CVD Strip-Assay T, Vienna, Austria) was performed for the identification of thrombophilia-related mutations.Resulting amplicons were then hybridized onto a strip support.A streptavidin-conjugated alkaline phosphatase reaction with a chromogenic substrate was performed, in order to detect amplicon hybridization onto strips.The resulting strips were then scanned and interpreted according to kit instructions.

Statistical analysis
Statistical analysis was performed using IBM SPSS Statistics 25 (IBM Corporation, Armonk, NY, USA).The frequency results are presented as proportions.The Mantel-Haenszel common odds ratio estimate and the Chi-square tests have been used.A p-value lower than 0.05 was considered statistically significant.

Results
All 1,291 patients analyzed were positive for at least one prothrombotic genetic variant or mutation.The number of thrombophilia-associated variants and mutations per person ranged from one to eight, with a median of four (Figure 1).A total percentage of 77.92 % of patients were diagnosed with three, four or five mutations, while 98.11 % were identified as having at least two mutations (Figure 1).
The frequency of heterozygous FVL mutation was higher at 10.38 % (134 out of 1,291 patients) than that of the heterozygous prothrombin G20210A mutation at 5.81 % (75 out of 1,291 patients).The frequency of the homozygous form for both mutations was found to be the same, at 0.08 % (one patient with homozygous FVL and one patient with prothrombin G20210A) (Figure 2).The frequency of double heterozygosity for both FVL and prothrombin G20210A mutations was 1.08 % (14 out of 1,291 patients).
Two polymorphisms for the MTHFR gene have been analyzed: C677T and A1298C.The frequency for both heterozygous and homozygous variants was comparable for the two types: 46.86 % and 45.39 % for heterozygous MTHFR C677T and, respectively, A1298C variants; and 10.46 % and 10.15 % for homozygous MTHFR C677T and, respectively, A1298C variants (Figure 2).Most patients had at least one variant of the MTHFR gene, in 89.39 % of cases, (1,154 out of 1,291 patients).The presence of both heterozygous variants of the MTHFR gene (C677T and A1298C) was detected in 23.47 % of cases (303 out of 1,291 patients).
The A1 haplotype of the EPCR gene (C4678G) had a higher frequency (48.41 % for heterozygous cases and 21.22 % for homozygous cases), than A3 haplotype (G4600A) (20.06 % for heterozygous cases and 1.63 % for homozygous cases) in the studied population (Figure 2).
In another analysis, we segregated the studied population into two groups: a FVL positive group (135 patients), and a FVL negative group (1,156 patients).The prothrombin G20210A genetic mutation has been found to be 2-fold more frequent in the FVL positive group (14 prothrombin G20210A positive cases out of 135 patients, 10.37 %), compared to the FVL negative group (62 prothrombin G20210A positive cases out of 1,156 patients, 5.36 %)odds ratio 2.04; 95 % confidence interval, 1.11-3.76,p=0.019 (Figure 3).
We evaluated the associations between heterozygous and homozygous forms of FVL and prothrombin G20210A with other low-risk genetic variants.Although not statistically  significant, the comprehensive data of this analysis were presented in Table 1.

Discussion
Since thrombophilia is a well-known contributor to thrombotic events occurrence, many studies have shown an association of inherited thrombophilia with APO [6,[9][10][11].However, there are also several studies failing to demonstrate a cause-effect relationship [12][13][14].
The present study described the inherited thrombophilia genetic profile in a population of pregnant women with a history of APO or thrombotic events.In this high-risk population all participants had at least one thrombophiliaassociated variant.Further, 98.11 % were identified as having at least two variants/mutations.The most frequent genetic variants identified in over 50 % of patients were: We performed a literature survey for studies assessing the association between inherited thrombophilia and APO occurrence (such as recurrent pregnancy loss, preeclampsia, or repeated implantation failure).The results are summarized in Table 2.It is perhaps noteworthy that we found quite similar variants/mutation frequencies to those reported by these studies.
We detected a frequency of 10.38 % for heterozygous, and 0.08 % for homozygous FVL mutation.Several studies included in Table 2 performed on persons mainly from the Balkan peninsula and Middle East, reported a frequency of heterozygous FVL mutations ranging from 2.25 to 15 % (with a median of 7.8 %), which is in line with that found by our study [6,9,11,13,14].The frequency of homozygous FVL mutation in our study (0.08 %, representing one single case out of 1,291) was lower than that reported in two of the aforementioned studies -1 % [9], or 2.2 % [11].Given the larger number of cases screened and the low frequency we reported, we can assume that our patient population is more similar to the patient population in which no homozygous mutations were reported [6,13,14].In the three previously cited articles, it is possible that no cases of homozygous FVL mutations were reported due to a smaller patient population.
Regarding the other high-risk thrombophilia-associated mutation, the prothrombin G20210A mutation, we detected a total frequency of 5.89 % (heterozygous and homozygous mutations), of which 5.81 % of patients were heterozygous for the prothrombin G20210A mutation.The frequency with which we encountered this mutation is slightly higher than those previously reported, ranging from 1.12 to 5 %, with a median of 4 % [6,[12][13][14][15].Regarding the frequency of homozygous prothrombin G20210A mutations, we observed a relatively low frequency (0.08 %, one single case from 1,291), which is in accordance with that previously reported [6,12,13,15].We detected this rare mutation most probably due to our large patient number, as presented above for the homozygous FVL.
Compared to the prevalence previously reported in the general population for FVL (average of 3.5 % in Europe and 2.2-3.6 % in Balkan countries) [33], respectively for prothrombin mutations (1.7-3 % in Europe) [34], the frequencies detected in the analyzed population are much higher, which is more likely a result of investigating a highrisk population of pregnant women with previous APO.
Regarding the frequency of heterozygous FVL, it can be mentioned that Clark et al. [33] also included a small number of Romanian subject (n=42) and reported a relatively high frequency for FVL of 8.3 %.This was interpreted as more likely due to the sampling bias in the population, given the small number of subjects.However, when comparing the frequency in our population of high-risk patients with that described in the Romanian general population (10.38 % vs. 8.3 %, respectively) it can be seen that the discrepancy is not so striking.
Regarding the prevalence of prothrombin mutations, the data presented by Rosendaal et al. [34] showed that with respect to the general population in Europe, the further south the prothrombin G20210A mutation is tested, the higher prevalence is observed, with the highest frequency in Ferrara, Italy at about 4 % of subjects (this was the most southern European area included in the study).When comparing the latitude of Ferrara and Bucharest, it can be seen that these two cities are situated around the 44 th degree (44.83 latitude for Ferrara and, respectively, 44.43 latitude for Bucharest).This observation can help explain the higher prevalence of 5.89 % observed for prothrombin G20210A mutation we found in our patients.
From a pathophysiologic point of view, these two genetic factors induce a hypercoagulable state by increasing the resistance to activated protein C (APC), respectively by increasing prothrombin level [1].This suggests that FVL and prothrombin G20210A mutation may represent important contributors in pregnancy complications development such as APO and thrombotic events.
As expected, we found a higher frequency of thrombophilia-associated genetic variants than that of highrisk mutations, which is supported by other studies [10,14].The heterozygous form of PAI-1 4G/5G variant (56.86 %), and the heterozygous form of EPCR C4678G variant (48.41 %) registered the highest frequency.PAI-1 polymorphism is one Represents the studies reporting an additional outcome, the venous thrombosis.
of the factors leading to PAI-1 level increase, and therefore to a reduced fibrinolytic activity.The role of PAI-1 4G/5G in RPL, recurrent implantation failure (RIF) and female infertility is still unclear, some of the studies suggesting a positive involvement of PAI-1 polymorphism in APO occurrence [35,36], and others not [30].Soluble EPCR (sEPCR) plays an essential role in pregnancy maintenance.EPCR variants can cause a diminished level of sEPCR, leading to a tendency for implantation failure [21].A3 haplotype has been reported to act as a risk factor for RPL, whereas A1 haplotype has been noted to exert a protective effect, preventing RPL [22,23].In our study, A1 haplotype of EPCR had a higher frequency (69.63 % for both homozygous and heterozygous variants), than A3 haplotype (21.69 % of cases) in the studied population.Assessment of the two polymorphisms of the MTHFR gene mutation revealed a high frequency of the heterozygous forms.MTHFR gene polymorphisms can induce impaired enzymatic activity, leading to hyperhomocysteinemia, known as an independent risk factor for thrombosis and pregnancy-related disorders [16].Available data linking MTHFR polymorphism to RPL are however inconsistent, with pros [19,20,35,36] and cons [17,18] regarding this association.Factor V H1299R and factor XIII mutations frequencies were among the lowest.Factor V H1299R variant is associated with a mild increase of the thrombotic risk.However, data regarding the influence of factor V H1299R on pregnancy are discordant, some of the studies supporting an association with RPL [9,15,35], while others not [29,37].A higher frequency of factor XIII genetic variants in women with RPL has been noted [9,20,27,28,35], as well as an association of factor XIII V34L variant with RIF [27].
Regarding combined thrombophilia, 1.08 % of patients tested positive for both heterozygous forms of FVL and prothrombin G20210A, a percentage higher than that estimated in the general population (<0.1 %) [38].When a subgroup analysis was performed, we found that the frequency of prothrombin G20210A mutation is higher (10.37 %) in the group with FVL mutation, than in the group without presenting this mutation (5.36 %).This finding is consistent with published data [39], and may explain the increased thrombophilic risk in our patients, leading to the development of more APO and/or thrombotic events.
Finally, no associations were identified between FVL or prothrombin G20210A mutations and other low-risk genetic variants.The association of prothrombin G20210A mutation with factor V H1299R, was found close to the limit of statistical significance (Table 1).The overlap of prothrombin G20210A and factor V H1299R can increase the risk of thrombosis, based on the cumulative effect resulted from the concomitant presence of a high prothrombin level and APC resistance [40].
The main limitation of our study is the lack of a control group.This is because the RO 19.10 project, which was a health program, did not include funding for a healthy control population.Another limitation is that whilst all tested women had at least one risk factor, a history of pregnancy complications (at least one pregnancy loss, preeclampsia) or thrombotic events, we did not stratify them according to the type of complications.
In conclusion, our study described the frequency of prothrombotic mutations and polymorphisms in a large population of pregnant women with a relevant medical history, and to our knowledge, it is the largest study of this kind in our country.All patients were found positive for at least one thrombophilia-associated genetic mutation or variant.FVL and prothrombin G20210A mutations were encountered with a higher frequency in the studied population than in the general population.We also noted an almost 2-fold higher frequency of prothrombin G20210A mutation in patients with FVL, than in patients without FVL mutation.Low-risk variants had a higher frequency in the studied population.
Our data revealed that the patients with prior APO presented various thrombophilia associated variants/ mutations more frequently that in the general population, and sometimes more frequently than in other high-risk patient populations (see Table 2).Due to the limitations of our study, we could not calculate the odds ratio of the variants generating APO, thrombosis, or other complications.Further studies are needed, preferably prospective cohort studies, to better describe both the prevalence of these variants and to calculate the risk for complications they impart, particularly in Romanian pregnant women.The ultimate goal could be to implement a wider thrombophilia testing panel in pregnant women with previous pregnancy-related complications and/or thrombotic events.

Figure 1 :
Figure 1: Distribution of the patients by number of thrombophilia-associated genetic variants and mutations in the studied population.The included patients presented at least one, and up to eight thrombophilia-associated variants and mutations, with a median of four per patient.

Figure 2 :
Figure 2: The frequency of the thrombophilia-associated genetic mutations and variants (expressed as homozygous and heterozygous forms), and the frequency of the corresponding wild-type gene in the studied population.MTHFR, methylenetetrahydrofolate reductase; PAI-1, plasminogen activator inhibitor 1; EPCR, endothelial protein C receptor.

Figure 3 :
Figure3: Association between the presence of factor V Leiden and prothrombin G20210A in the studied population.Prothrombin G20210A was found to be almost 2-fold more common in the group with factor V Leiden mutation, than in the group without this mutation.

Table  :
Associations between factor V Leiden or prothrombin GA mutations and low-risk genetic variants in the studied population.

Table  :
Association between inherited thrombophilia and adverse pregnancy outcomes occurrence in the selected case-control studies.