Logistic regression analysis of key markers of prethrombotic state screening in RSA patients

Purpose To screen the important markers of prethrombotic status (PTS) in patients with recurrent abortion (RSA) and establish a multi-factor Logistic regression model, and to predict the risk of PTS in patients with RSA, so as to provide evidence for the early clinical prevention and treatment of RSA. Methods A total of 193 patients with who 2021 were selected as the case group and 131 normal pregnant women who visited our hospital during the same period as the control group. The age, body mass index (BMI) and PTS-related markers were compared between the two groups. Including coagulation factor V1691, prothrombin gene (F(cid:0))20210, plasminogen activator inhibitor-1(PAI-1)675,methylene tetrahydrofolate reductase (MTHFR)677 and 1298, methionine synthase reductase (MTRR)66 gene locus polymorphism, homocysteine (HCY) level, protein C(PC) and protein S(PS) activity. The above markers were screened to establish a multi-factor Logistic regression model, and the predictive value of the model was assessed by ROC curve.


Introduction
Recurrent spontaneous abortion (RSA) refers to three or more consecutive spontaneous abortions with the same partner, but most experts believe that two consecutive abortions should be paid attention to and evaluated, because its risk of re-abortion is close to that of three abortions.1 Recurrent spontaneous abortion has great di culty in clinical diagnosis and treatment because of complex etiology and lack of clear clinical manifestations. Prethrombotic state (PTS) is a pathological process of dysfunction or disorder of coagulation and anticoagulant system, brinolytic system and anti-brinolytic system caused by a variety of factors. PTS is a risk factor for thrombosis. Studies have shown that when the coagulation function is abnormal, the blood ow state of uterine and placenta will change, and local tissues are prone to forming microthrombus, which will lead to placental brosis deposition, infarction and eventually lead to various adverse pregnancy outcomes. 2There is increasing evidence that PTS plays an important role in the pathogenesis of RSA. Brenner believed that PC defect, PS defect, FV and F mutations, and hyperhomocysteinemia had signi cant correlation with abortion.3 Studies have con rmed that genetic polymorphisms of MTHFR, MTRR, F V, F II and PAI-1 can promote thrombosis and induce abortion.4 Screening the MTHFR and MTRR gene polymorphisms and monitoring the HCY level have important guiding signi cance for perinatal health care.5 Therefore, many scholars in home and abroad have proposed to increase the screening of laboratory diagnosis of PTS in RSA patients to prevent the occurrence of abortion.6 However, there is no clear and uniform standard for the diagnosis of PTS and the screening of markers in RSA patients. In this study, important markers of PTS in RSA patients were screened and a multivariate Logistic regression model was established to predict the risk of PTS in RSA patients.

Research design and subjects
A total of 193 cases with two or more spontaneous abortions who visited the Pre-natal Diagnosis Center of the Second A liated Hospital of Harbin Medical University from December 2019 to May 2021 were selected as the case group, and 131 cases of normal pregnant women in the same period were selected as the control group. Inclusion criteria: patients with two or more spontaneous abortions were selected as the case group; The control group was selected from the population with no previous pregnancy history or normal pregnancy history. Patients in the two groups had not taken any medicine that affected the observation effect recently. Sign informed consent. Exclusion criteria: (1) Various acute and chronic diseases such as heart, liver, kidney, brain, and pelvis; (2) Chromosome abnormalities existing in both husband and wife; (3) Immune diseases, hematological diseases and endocrine diseases; (4) Malignant diseases; (5) Hysterosalpingography and Doppler ultrasound diagnosis of reproductive anatomy deformity.
All participants signed the study protocol and informed consent for voluntary enrollment. The study was approved by the Institutional Review Board of Second A liated Hospital of Harbin Medical University.

Methodology
The non-anticoagulant blood was collected from the elbow vein in the fasting state. After standing at room temperature for 15 min, the supernatant serum was collected by centrifugation to detect the HCY level, PC and PS activity. The PTS-related genes were ampli ed by polymerase chain reaction (PCR) and sequenced by Sanger method. The genetic polymorphisms of age, BMI, Fv (1691), F (20210), PAI-1(675), MTHFR(677), MTHFR(1298), and MTRR(66) sites in the two groups were analyzed and compared. The detection of each item was completed in strict accordance with the experimental operating procedures.
Statistical Analysis SPSS 25.0 soft ware was used for statistical analysis. Enumeration data were expressed as the percentage of cases, and inter-group comparison was performed using χ2 test. Age, BMI, PAI-1(675), MTHFR(677), MTHFR(1298), MTRR(66) locus polymorphism, HCY level, and PC and PS defects in the case group and the control group were compared, and the differences were considered to be statistically signi cant when P < 0.05. Independent variables with P < 0.1 in the univariate analysis were included in the multivariate logistic regression analysis, and the OR value and 95% CI were calculated. If P < 0.05, statistical signi cance existed. The ROC curve was used to evaluate the multivariate logistic regression analysis model.

Establishment of logistic regression model
A univariate analysis was performed by χ2 test on age, BMI, PAI-1(675), MTHFR(677), MTHFR(1298), MTRR(66) locus polymorphism, HCY level, and PC and PS defects in the two groups (see Table 1). Risk factors with P<0.1 in univariate analysis were assigned (see Table 2) and included in Logistic multivariate analysis. The construction of multi-factor logistic regression risk prediction model: 0+11+22+...+NN was converted to risk probability prediction model of recurrent spontaneous abortion: P =. The results showed that BMI, PAI-1(675) and MTHFR(677) gene locus polymorphism were the independent risk factors for RSA, and the differences were statistically signi cant (P < 0.05) (see Table 3)

Evaluation of logistic regression model
The binary Logistic regression analysis showed that the model had statistical signi cance (p <0.001). The predictive variables were obtained according to the Logistic regression model, and the ROC curve of the model was drawn with RSA as the state variable. The results showed that the ROC curve was far from the reference line, indicating that the model was credible. The area under the ROC curve (AUC) of the model for predicting the risk factors related to PTS in RSA patients was 66.4%, the 95% con dence interval (95% CI) was 60. 5-72.3%, and the standard error was 0.030. Compared with the area under the curve of 0.5, the difference was statistically signi cant (P < 0.001), as shown in Fig. 1.

PTS-related genetic markers in patients with RSA
The most common pathogenesis of hereditary PTS in western populations is FV mutation, followed by F II mutation. The mutated FV can not only continue to express procoagulant activity, but also decrease the cleavage ability of activated protein C, thereby generating activated protein C resistance, manifested as PTS8. The research by Poort et al has shown that the G-A conversion of nucleotide 20210 at the 3' untranslated region of coagulation F II can increase the plasma prothrombin level of the body, 9 thus causing venous thromboembolic diseases. In this study, no F V 1691 or F II 20210 polymorphic mutation was found in the tested women, 10 suggesting that Chinese women with RSA have rare F V 1691 or F II 20210 polymorphic mutations in prethrombotic state. This may be due to the uneven distribution of PTS across ethnic groups and regions.
PAI-1 is the main inhibitor of plasma brinolytic system and the main substance that inhibits brinolytic activity in the blood circulation. The increased level of PAI-1 leads to the decreased brinolytic activity and enhanced coagulation process. It is one of the risk factors for thrombotic diseases. Sheppard believed that PAI-1 expression products increased the deposition of vascular brin in endometrium during early pregnancy, and decreased uterine and placental blood ow, resulting in abortion.14 Glueck et al found that the PAI-1 (-675) 4g/5g polymorphism was the susceptible genotype for unexplained recurrent spontaneous abortion and had a strong independent correlation with adverse pregnancy outcomes. 15 In this study, we found that PAI-1(675) was closely related to RSA in a multi-factor regression analysis (P < 0.05; OR=1.508), indicating that the PAI-1(675) locus polymorphism was an independent risk factor for RSA.
MTHFR and MTRR are the two most important genes in folic acid metabolism. The mutation at site 677 of MTHFR gene and the mutation at site 66 of MTRR gene at site 1298 make the serum HCY level too high and have a direct toxic effect on the embryo, which can damage the vascular endothelium, stimulate the proliferation of vascular smooth muscle cells, and lead to the imbalance of coagulation and brinolysis in the body.16 The results of this study suggested a correlation between the MTHFR(677) locus polymorphism and RSA (P<0.05), but no correlation was found between the MTHFR(1298) locus polymorphism and RSA (P > 0.05), which was consistent with the results of a meta-analysis.17

PTS related coagulation markers in RSA patients
The most common genetic factor in the Han population is congenital anticoagulant protein de ciency, including protein C, protein S and antithrombin, mainly PC or PS, and antithrombin de ciency is rare.11 Because the levels of PC and PS uctuate signi cantly during pregnancy, it is more reliable to monitor the nonpregnant levels in women. 12 In this study, the χ2 test showed that both PC defects and PS defects had no correlation with RSA (P > 0.05). However, the positive rate of PS defects in the case group (21.8%) was signi cantly higher than the incidence rate of PS defects in the control group (14.5%), which might be related to the insu cient sample size in this study. Therefore, the correlation between PC and PS defects and RSA needs further large sample data analysis.
Some studies have suggested that the increase of HCY is closely related to RSA, and its mechanism may be related to coagulation abnormalities. The main reason for the increase of HCY is folic acid de ciency. After standardized folic acid intervention, the metabolism of folic acid can be regulated to reduce the level of HCY. 18,19 In this study, no correlation was found between HCY and RSA, which might be related to the increased publicity of the national policy of free folic acid distribution, the increased awareness of pre-pregnancy health care for women of childbearing age in the region, and the adequate supplementation of folic acid during the pre-pregnancy period.
Some studies have suggested that the increase of BMI is related to the increased risk of spontaneous abortion,13and its molecular mechanism may be that excessive insulin acts on the ovarian theca cells through insulin receptors to affect the embryo implantation. Logistic regression analysis in this study showed that the OR value of BMI was 3.264, which meant that for every 1 kg/m2 increase in body mass index after other factors were controlled, the probability of illness was 3.264 times the original one. Therefore, controlling reasonable BMI might reduce the risk of illness of RSA.

Conclusion
BMI ≥24, PAI-1(675) and MTHFR(677) gene locus polymorphism in patients with RSA are independent risk factors for multiple spontaneous abortions. In the process of clinical work, we should improve the alertness to these patients and carry out early prevention and treatment. The shortcoming of this study was that the included sample size was small, and no validation group was set to further verify the accuracy of the model. In the future, large-sample, multi-center research was required. In addition, the screening of PTS-related markers in this study is not comprehensive, and joint monitoring of coagulation function and platelet aggregation function is required in the future. In conclusion, the multivariate Logistic regression model that can effectively predict the prethrombotic state of RSA patients assessed by ROC curve has been established in this study, which will provide new reference for clinicians to prevent and treat RSA among women of childbearing age.