Application of the RCOG Risk Assessment Model for Evaluating Postpartum Venous Thromboembolism in Chinese Women: A Case-Control Study.

A total of 65 501 women gave birth in our hospital during the study period, and 39 thrombotic events occurred. The incidence of DVT was 0.060% and the incidence of PE was 0.018%. Among the VTE cases, 1 patient was excluded because delivery occurred at less than 28 gestational weeks. Therefore, 38 VTE cases (VTE group) and 152 women without VTE (control group) were finally included in this study ( Figure 1 ). Two women in the control group used low-molecular-weight heparin (LMWH) during pregnancy, one due to significant elevation in D-dimer and the other due to a history of 3 spontaneous abortions. In the VTE group, 1 woman with antiphospholipid syndrome (APS) used unfractionated heparin prenatally. Three women with antenatal DVT used LMWH during pregnancy, and the thrombus was dissolved before delivery. Among the 190 women, only the 3 with antenatal DVT used LMWH for thromboprophylaxis after delivery. Details of the VTE events are shown in Supplementary Table 2 . Table 1 shows the comparison of socio-demographic characteristics of the women in the 2 groups. There were no significant differences in BMI, educational level, and occupation between the 2 groups. Significant differences in age ( P =0.006) and birthplace ( P =0.024) were observed. Risk factors contained in the RCOG RAM were compared between the 2 groups ( Table 2 ). We found that the proportions of women in the VTE group with previous VTE ( P =0.001), a family history of VTE ( P =0.010), known low-risk thrombophilia ( P =0.014), age >35 ( P =0.007), smoking habit ( P =0.046), multiple pregnancy ( P =0.025), cesarean section in labor ( P <0.001), and surgical procedure in pregnancy or puerperium ( P =0.004) were higher than those in the control group. Other risk factors showed no significant differences between the 2 groups. Among the 190 women, the most common risk factors were elective cesarean section (88/190), age >35 years (39/190), cesarean section in labor (38/190), and family history of VTE (18/190). Table 3 shows that the RCOG scores of women in the VTE group were much higher than that of the control group (3.74±0.74 vs 1.49±0.22, P <0.001). A difference in risk stratification was also found between the 2 groups ( P <0.001). The univariable logistic regression analysis revealed that the risk of postpartum VTE was 9.75-fold greater in the intermediate-risk group, and 90.00-fold greater in the high-risk group than in the low-risk group. Meanwhile, as the RCOG score increased, the risk of developing postpartum VTE increased almost linearly ( Figure 2 ). The ROC curve analysis determined that the AUC of the RCOG RAM was 0.828 (95% CI: 0.762–0.894, Figure 3A ), and the Youden index was 0.50, with the best cut-off value being 2, which exactly matches the cut-off value recommended by the RCOG guidelines for pharmacological thromboprophylaxis after delivery. The sensitivity of the model was 89.47%, and its specificity was 60.53%. The positive predictive value was 36.17%, and the negative predictive value was 95.83%. In addition, the calibration plot showed good agreement between the prediction of the RCOG RAM and actual observation of VTE events (Hosmer-Lemeshow test p=0.214, see Figure 3B ). The DCA also showed a certain net benefit value when the threshold probability is between 0.05–0.74 and 0.87–0.99, indicating that clinical decisions based on the predicted result of this model can benefit patients ( Figure 3C ). Other factors that might be associated with postpartum VTE were also analyzed. We found that endometriosis ( P =0.024), a family history of diabetes ( P =0.010), recurrent pregnancy loss ( P =0.024), glucocorticoid therapy during pregnancy ( P =0.004), previous use of oral contraceptive pills ( P =0.005), and intrauterine devices ( P =0.001) for contraception were more common in women with postpartum VTE, while the previous use of condoms, calendar rhythm method, or withdrawal method for contraception ( P =0.009) was more common in the control group ( Table 4 ). Women in the VTE group had significantly higher D-dimer levels before delivery, and on the first and third day after delivery ( P <0.05, Table 4 ). RCOG risk stratification and factors (not included in the RCOG model) found to be significantly different in the univariable analysis were further entered into a multivariable logistic regression analysis. The results show that RCOG intermediate risk (adjusted OR=8.39, 95% CI: 2.64–26.60), RCOG high risk (adjusted OR=118.23, 95% CI: 18.05–774.46), previous use of IUD (adjusted OR=7.11, 95% CI: 1.45–34.93), and glucocorticoid therapy during pregnancy (adjusted OR=6.72, 95% CI: 1.56–28.91) were associated with elevated risk of postpartum VTE ( Table 5 ).

Venous thromboembolism (VTE) comprises deep venous thrombosis (DVT) and pulmonary embolism (PE). Because of increased concentrations of coagulation factors, decreased anticoagulant and fibrinolytic components, and increased compression of the uterus to the iliac vein and inferior vena cava [ 1 ], the risk of VTE for women during pregnancy and the puerperium period is approximately 4–5 times higher than that in non-pregnant women of the same age [ 2 , 3 ]. Meanwhile, the risk of postpartum VTE is 20 times higher than that of antepartum VTE [ 4 ]. The incidence of VTE is 1–2 per 1000 pregnancies in Europe and America [ 5 , 6 ]. Nearly 80% of pregnancy-related VTEs are isolated DVTs, and 20% are PEs or coexisting PE and DVT [ 7 ], whose fatality rate can be as high as 15% [ 8 ]. Currently, VTE has become one of the leading causes of maternal mortality worldwide [ 9 ]. In the past, insufficient attention was given to the prevention of pregnancy-related VTE in China, and until now, there has been no epidemiological analysis based on national data. The Wells score and the Caprini score are usually used in clinical practice for VTE risk assessment in the obstetric department, but neither is designed specifically for pregnant women [ 10 ]. In 2015, the Royal College of Obstetricians and Gynecologists (RCOG) released the latest edition of guidelines for the prevention of VTE during pregnancy and puerperium and updated the VTE risk scoring method (mentioned as RCOG risk assessment model below). The detailed scoring method for VTE is shown in Supplementary Table 1 . This model has been widely used in Western countries for the assessment of the risk of pregnancy-related VTE in both the antenatal and postnatal periods, and a thromboprophylaxis management process based on the results of the model has been recommended by the RCOG guidelines. If the total score is ≥2 postnatally, thromboprophylaxis for at least 10 days is recommended [ 11 ]. However, the data used for building the RCOG RAM were mostly obtained from Western countries. At present, whether this model is suitable for Chinese maternity care remains poorly understood. In view of the higher prevalence of VTE during the postnatal period, this study aimed to evaluate the predictive value of the RCOG RAM for postpartum VTE and its guiding significance in postpartum thromboprophylaxis in China. Furthermore, we aimed to explore related factors for postpartum VTE other than those already included in the model.

The RCOG risk assessment model is one of the most complex scales for pregnancy-related VTE. Since its first publication in 2004, it has been demonstrated to be effective in reducing the maternal mortality rate from PE in the UK [ 11 ], but its applicability in Chinese maternity care is uncertain. The traditional conception is that Asian people have a relatively low risk of VTE compared with people in Western countries [ 14 ]. As the data used to establish the RCOG model were obtained from Western countries, it is worrying that the use of RCOG RAM in China might lead to the overestimation of VTE risk as well as the overuse of LMWH. Several domestic studies have compared the RCOG RAM and other VTE scoring models not specifically designed for pregnant women. Liang et al found that the discrimination and accuracy of the Wells score were better than that of the RCOG RAM, but they used the 2009 version of the RCOG model [ 15 ]. At the same time, Zhang et al came up with an opposite result [ 16 ]. Moreover, a prospective observational study has been carried out in Beijing to evaluate the feasibility of the RCOG RAM [ 17 ]. In our case-control study, we found that the RCOG RAM could screen out the majority of patients who eventually developed postpartum VTE (34/38), and for those assessed as low-risk by the model, nearly 96% would not develop VTE (90/94). Therefore, this model has a clear predictive value for postpartum VTE in Chinese women. During its application in other Asian countries, it was found that the RCOG RAM might lead to the overuse of anticoagulant drugs in women who are less likely to develop postpartum VTE. A retrospective study in Malaysia suggested that 30.62% of postpartum women without VTE met the criteria for thromboprophylaxis [ 18 ]. Our study reached a similar conclusion. Currently, the main concerns for the large-scale use of LMWH are its high price and the increased risk of hemorrhage and poor wound healing. However, compared with other anticoagulants such as heparin, LMWH is much safer [ 19 ]. Some studies have also confirmed that the use of LMWH does not increase the risk of postpartum hemorrhage and wound complications [ 20 , 21 ]. In view of the fatal consequences of VTE, it is critical to ensure that as many women as possible with potential risks of VTE receive anticoagulation treatment. In our study, elective cesarean section, age >35 years, cesarean section in labor, and family history of VTE were the most common risk factors for postpartum VTE. Therefore, to reduce the incidence of VTE events, we strongly recommended that efforts should be made to prevent cesarean sections without indication and to avoid pregnancy in advanced age. We also found that in addition to the factors already contained in the RCOG RAM, previous use of IUD and glucocorticoid therapy during pregnancy might also increase the risk of postpartum VTE. IUDs containing levonorgestrel (Mirena) are thought to increase the risk of VTE, but a copper IUD seems to be a safe alternative for contraception, even for people with a history of DVT or PE [ 22 , 23 ]. However, we surprisingly found that previous use of a copper IUD might increase the risk of postpartum VTE. Among the 7 women who had used IUD in the VTE group, only 1 used Mirena, and the rest used copper IUDs. Previous studies have not detected changes in coagulation-related parameters in IUD users, but the placement of an IUD has been shown to cause abnormal uterine bleeding by elevating fibrinolysis [ 24 ], which might cause further potential abnormalities in coagulation function that lead to thrombosis during pregnancy and the peripartum period. Furthermore, extensive microthrombosis in stromal capillaries has been observed in the endometrium in contact with an IUD [ 25 ]. The main purpose of using glucocorticoids during pregnancy is to promote fetal lung maturity in women who are at risk of preterm birth, or to treat thrombocytopenia during pregnancy. The use of glucocorticoids can cause the increased circulation level of fibrinogen, Factor VII, Factor VIII, and other hemostatic components [ 26 ]. Previous epidemiological studies also found that the use of systemic glucocorticoids can increase the risk of VTE by up to 3 times [ 27 ]. However, in the currently widely used risk assessment models for VTE, whether designed for pregnant women or not, the use of glucocorticoids is not regarded as a risk factor [ 10 ]. Therefore, it is recommended that clinicians strengthen the prevention and treatment of VTE in women who use glucocorticoids. To our knowledge, this study is one of the few to assess the applicability of the RCOG RAM in China, and contains a relatively large number of VTE cases. At the same time, our study also has some limitations. Due to the low prevalence of postpartum VTE and the single-center design, although we enrolled all VTE cases within 4 years, the number of VTE events was relatively small. Additionally, the retrospective study design may limit the interpretation of the results and make it difficult to establish causal conclusions. To solve these problems, a prospective, large-sample, multicenter study will be required in the future. As far as we know, a large prospective observational study on the applicability of the RCOG RAM in Chinese women has already been carried out in Beijing, and the researchers hope to revise the model based on the study results to make it more suitable for Chinese population [ 17 ]. We hope that this study can provide a reference for the use of the RCOG RAM in Chinese maternity care, and provide valuable experience for subsequent large-scale studies.

Our results indicate that the RCOG RAM is effective in identifying women who were at risk of postpartum VTE, and that following the thromboprophylaxis methods recommended by the RCOG guidelines can effectively reduce the risk of VTE. But at the same time, we found that previous use of IUD and use of glucocorticoid during pregnancy might also increase the risk of postpartum VTE. Therefore, for the better prevention of pregnancy-related VTE, a risk assessment model based on national data needs to be established in the future.

This retrospective case-control study was conducted at the International Peace Maternity and Child Health Hospital (IPMCHH), a specialized hospital of obstetrics and genecology in Shanghai, from June 2016 to June 2020. Women diagnosed with VTE during the postpartum period constituted the VTE group. For each VTE patient, 4 women who gave birth at the same day and confirmed as without VTE were randomly selected as the control group. Inclusion criteria included women who gave birth beyond 28 weeks of gestation, regardless of delivery mode, and completed regular antenatal and postnatal examinations as well as diagnostic examinations for VTE. Exclusion criteria were missing or untraceable previous medical records. The diagnosis of DVT was confirmed by compression ultrasonography of the lower limb veins [ 12 , 13 ], and once thrombotic events were detected, the patient was routinely checked by echocardiography and computer tomography pulmonary angiography (CTPA) to investigate the presence of PE. All women were encouraged to walk as soon as possible after delivery. For those who underwent cesarean section, a pneumatic compression device was routinely used. Very few women received pharmacological prophylaxis, which was mainly based on empirical medication. Data collected from medical records included socio-demographic characteristics, reproductive history, gynecologic history, previous contraceptive use, obstetric characteristics, and the results of laboratory tests (coagulation-related parameters) and imaging examinations. All women returned to the hospital for routine check-ups 42 days after delivery, and relevant data were also collected. Based on the above information, the RCOG RAM was used to assess the risk score and risk level of postpartum VTE. The risk of postpartum VTE was divided into 3 levels: low risk (<2 points and without intermediate or high-risk factors, no need for thromboprophylaxis), intermediate risk (≥2 points and without high-risk factors, need for thromboprophylaxis), and high risk (with at least 1 high-risk factor, need for thromboprophylaxis) [ 11 ]. All scores were calculated independently by 2 researchers. If any difference in the RCOG score was present, a third researcher recalculated the score. This study was approved by the Institutional Review Board of IPMCHH (GKLW-2020-03). The Kolmogorov-Smirnov test was used to evaluate the normal distribution. Continuous variables were assessed using the t test or Mann-Whitney U test. Categorical data were analyzed using chi-square tests or Fisher’s exact test. RCOG risk stratification and factors (not included in the RCOG model) that were significantly different according to the univariable analysis were entered into a multivariable logistic regression analysis by stepwise selection. ROC curve analysis, calibration curve, and decision curve analysis (DCA) were used to evaluate the predictive value of the RCOG RAM for postpartum VTE, as well as the applicability of the thromboprophylaxis recommended by the RCOG guidelines. A 2-sided P value <0.05 was considered statistically significant. All statistical analyses were performed using SPSS (version 22.0; SPSS, Inc., Chicago, IL, USA) and R statistical software version 4.0.2 (packages ‘pROC’, ‘caTools’, ‘rms’, ‘rmda’).

The RCOG risk assessment model for VTE. VTE – venous thromboembolism; BMI – body mass index; ART – assisted reproductive technology; IVF – in vitro fertilization; PPH – postpartum hemorrhage. Risk factors with score ≥4 are considered as high-risk factor; risk factors with score of 2–3 are considered as intermediate-risk factors; risk factors with score of 1 are considered as low-risk factors. ART/IVF is regarded as a risk factor for antenatal VTE, and ovarian hyperstimulation syndrome (OHSS) is regarded as a risk factor for VTE only in the first trimester. VTE events. VTE – venous thromboembolism; DVT – deep venous thrombosis; PE – pulmonary embolism; APS – antiphospholipid syndrome; LMWH – low-molecular-weight heparin. “Confirmed Time” refers to the time of initial diagnosis of VTE before and/or after delivery. Day 2, Day 3, etc. indicate the second or third day, respectively, after childbirth. Artificial reproductive technology/in vitro fertilization and ovarian hyperstimulation syndrome are risk factors for VTE antenatally, so they are not included in this table for postpartum VTE assessment.