Jeanine Roeters van Lennep, Erasmus Medical Center, Department of Internal Medicine-D425, PO Box 2040, 3000 CA Rotterdam, the Netherlands. Tel.: +31 10 70 95 193; fax: +31 10 70 33 639. E-mail: email@example.com
Summary. Background: The optimal approach for venous thrombosis (VTE) prophylaxis during pregnancy and postpartum in women with an increased risk of VTE is not established. Objectives: To evaluate the effectiveness, represented as the incidence of pregnancy-related VTE, and safety, represented as incidence of postpartum hemorrhage (PPH), of a protocol recommending prophylaxis with low-dose low-molecular-weight heparin (LMWH) in women at intermediate to high risk of VTE. Patients/methods: In this retrospective cohort study, we analyzed 34 women (44 pregnancies) with intermediate risk of VTE who received low-dose LMWH for 6 weeks postpartum and 57 women (82 pregnancies) with high risk of VTE who received low-dose LMWH during pregnancy and for 6 weeks postpartum. Pregnancy-related VTE was defined as VTE during pregnancy or ≤ 3 months postpartum. PPH was defined as blood loss > 500 mL and severe PPH as blood loss > 1000 mL. Results: The incidence of pregnancy-related VTE was 5.5% (95% CI, 2.4–12.3) despite prophylaxis with low-dose LMWH. All events occurred in women at high risk, with a postpartum incidence of 7.0% (95% CI, 2.9–16.7) and antepartum incidence of 1.8% (95% CI, 0.4–9.2). The risk of PPH was 21.6% (95% CI, 14.3–31.3) and severe PPH 9.1% (95% CI, 4.7–16.9), which was not different in women who started LMWH postpartum and those who used LMWH during pregnancy. Conclusions: Although prophylaxis with low-dose LMWH during pregnancy and postpartum proved to be safe, the risk of pregnancy-related VTE is considerable in women with a high risk of VTE. VTE prophylaxis with low-dose LMWH may not be sufficiently effective in these women.
Pregnancy-related venous thrombosis (VTE) is one of the leading causes of maternal mortality in developed countries. Pregnancy and the postpartum period carry a five-fold risk of VTE compared with the non-pregnant state [1,2]. Guidelines from the American College of Chest Physicians (ACCP)  provide recommendations for VTE prophylaxis with low-molecular-weight heparin (LMWH) for women who have an increased risk of VTE. These are based on a low level of evidence and are therefore mostly weak recommendations; that is, with unclear magnitude of the benefits and risks, burden and costs. Most recommendations are based on data from cohort studies [4–7] in which the underlying risk of VTE in the study population and the dose of VTE prophylaxis were heterogeneous. The various alternatives for management of women at intermediate risk of VTE, ranging from clinical surveillance to intermediate-dose LMWH during pregnancy, further reflect the lack of high-level comparative, prospective or interventional studies.
As the optimal approach for this population is unknown, it is not surprising that numerous local hospital protocols exist, which differ in risk stratification and LMWH dosage. One of the options in the ACCP guidelines is to use low-dose LMWH for VTE prophylaxis in women with intermediate to high risk of pregnancy-related VTE. Low-dose LMWH has several advantages over high-dose LMWH. There is evidence that the risk of bleeding complications is lower . Also low-dose LMWH is convenient because adjustment for bodyweight and measurement of anti-factor Xa is not necessary. In our hospital protocol (http://www.hematologieklapper.nl) pregnant women at intermediate risk of VTE are managed without pharmacological intervention during pregnancy and are prescribed low-dose LMWH for 6 weeks postpartum; women at high risk of VTE are given low-dose LMWH during the entire pregnancy and for 6 weeks postpartum. We performed a retrospective cohort study to evaluate the effectiveness, represented by the incidence of pregnancy-related VTE, and safety, represented by the incidence of postpartum hemorrhage (PPH), of a protocol recommending prophylaxis with low-dose LMWH in women at intermediate to high risk of VTE.
Study population and data collection
Pregnant women who were considered to be at intermediate or high risk of VTE were treated according to the local hospital protocol (http://www.hematologieklapper.nl) stated in Table 1. We searched the computerized medical files of all pregnant women in the obstetric department of our hospital for the words ‘thrombosis’ or ‘embolism’ or ‘anticoagulants’ between 1996 and 2009. We excluded women who used anticoagulant therapy for other reasons than thrombosis prophylaxis because of an increased risk of VTE, or who did not meet the criteria to be treated according to the hospital protocol (Fig. 1). We also excluded women who were treated with intermediate or therapeutic doses of LMWH at the discretion of their treating physician (Fig. 1). If women were not treated according to the hospital protocol, but received low-dose LMWH at any time during pregnancy or the postpartum period, we analyzed women according to the regimen that they had actually received.
Asymptomatic women with thrombophilia (other than high-risk thrombophilic defects mentioned in high-risk group.) AND positive family history (first degree) for VTE Asymptomatic women without thrombophilia but VTE in a first-degree family member without thrombophilia Women with a single episode of VTE related to a temporary risk factor (immobilization, postoperative, pill-related or postpartum period).
Clinical surveillance during pregnancy Low-dose LMWH (nadroparin 2850 IE anti-Xa once daily or an equivalent dose of another LMWH) for 6 weeks postpartum
Asymptomatic women with high-risk inherited thrombophilic defects (antithrombin deficiency, homozygosity for the factor V Leiden or prothrombin 20210A mutation, or combinations of thrombophilic defects) Women with a single episode of previous idiopathic VTE (not related to a temporary risk factor) Women with a single episode of previous VTE and thrombophilia Women with ≥ 2 episodes of previous VTE (not using long-term anticoagulation)
Low-dose LMWH (nadroparin 2850 IE anti-Xa once daily or an equivalent dose of another LMWH) during the entire pregnancy and 6 weeks postpartum
For each woman we recorded the obstetric and medical history (in particular previous VTE and known thrombophilia), weight and height at first visit during pregnancy, time of initiation of LMWH, dose, type of LMWH, concomitant use of aspirin or vitamin K antagonists, dose or regimen changes, (recurrent) VTE, pregnancy outcome, delivery route, estimated blood loss postpartum and blood transfusions during the first 24 h postpartum. Blood loss was estimated by the gravimetric method. This method consists of weighing materials such as soaked pads on a scale and subtracting the known weight of these materials to determine the blood loss. Follow-up up to 3 months postpartum was available for all women.
The primary endpoint of our study was pregnancy-related symptomatic VTE. VTE was considered pregnancy related if it occurred during pregnancy or within 3 months after delivery or after termination of pregnancy after a gestational age of 16 weeks due to other reasons. Episodes of VTE were adjudicated by two independent observers (JERVL and SM) as definite, probable or absent. A definite VTE was defined as a new or extended area of a non-compressible deep venous segment seen on compression ultrasound (CUS), or a diagnosis of pulmonary embolism made by computed tomography scan or ventilation-perfusion scan. In case the diagnosis of recurrent VTE could not be firmly established, but the treating physicians had decided to treat as a new VTE, the event was classified as probable. The secondary endpoint was postpartum hemorrhage (PPH) as representative of peripartum bleeding complications. PPH was defined by WHO criteria  as blood loss of > 500 mL within 24 h following vaginal delivery or Caesarean section. Because these criteria have been challenged , we classified severe PPH as > 1000 mL.
Results are presented as mean ± standard deviation if the data were normally distributed. Non-normally distributed data are presented as medians with ranges. We calculated the absolute risk of pregnancy-related VTE and its 95% confidence interval (95% CI) in two ways. First, we included only the first pregnancy of each unique woman in the database, because multiple pregnancies in one woman are not independent in terms of risk. Second, we analyzed all pregnancies and divided the number of VTE events by the total number of pregnancies. If a pregnancy-related VTE occurred, the subsequent pregnancies of that woman were censored. We performed a Kaplan–Meier survival analysis to visualize the timing of pregnancy-related VTE. The occurrence of PPH and blood transfusion within the first 24 h after delivery was compared between women who used low-dose LMWH during pregnancy and postpartum and women who only used low-dose LMWH postpartum using the chi-squared test. The difference in median blood loss between these two groups was compared by the Mann–Whitney test. A P-value of < 0.05 was considered to be statistically significant.
With our search strategy, we identified 316 unique women. After exclusion of 220 women for various reasons, 96 women were considered at intermediate or high risk of VTE (Fig. 1). Of these, 91 (94.8%) women who had had 126 pregnancies that could be analyzed were treated with low-dose LMWH. Thirty-four women at intermediate risk of pregnancy-related VTE had 44 pregnancies, for which they received low-dose LMWH for 6 weeks postpartum. Eight women had two pregnancies and two women had three pregnancies. Fifty-seven women at high risk of VTE had 82 pregnancies, for which they received low-dose LMWH during the pregnancy and 6 weeks postpartum. Twenty-four women had two pregnancies and one woman had three pregnancies.
The characteristics of the study population are shown in Table 2. The majority of women (85, 93.4%) had a history of at least one previous episode of VTE. Fifty-nine (69.4%) of these were hormone related, that is, related to the use of oral contraceptives, or had occurred during pregnancy or the postpartum period. Thirty-eight (41.8%) of the 91 women were carriers of a thrombophilic defect.
Table 2. Characteristics of the study population
All women (n = 91)
Intermediate risk (LMWH postpartum) (n = 34)
High risk (LMWH during pregnancy and postpartum) (n = 57)
*Homozygous factor V Leiden mutation, homozygous factor II G20210A mutation, a combination of thrombophilic defects or antithrombin deficiency. †First pregnancy in database. VTE, venous thrombosis; LMWH, low-molecular-weight heparin.
Number of pregnancies
Age at pregnancy (years) mean ± SD
31.9 ± 4.6
32.1 ± 4.3
31.7 ± 4.8
Previous VTE episodes (%)
≥ 2 (%)
Risk factor previous VTE
Oral contraceptives (%)
Inherited thrombophilia (%)
In combination with previous VTE (%)
Mild thrombophilia (%)
Type of prophylaxis
Nadroparin 2850 anti-Xa IU (%)
Dalteparin 5000 anti-Xa IU (%)
Enoxaparin 20 mg per day (%)
Vitamin K antagonist (postpartum) (%)
Live births (%)
Gestational age at delivery (days), median (range)
Birth weight (gram), mean ± SD
3272 ± 801
3366 ± 600
3216 ± 901
Vaginal delivery (%)
Caesarean section (%)
Of the 34 women who received only postpartum prophylaxis, nine should have received LMWH prophylaxis throughout pregnancy according to the protocol. Of the 57 women who were treated with low-dose LMWH during pregnancy and postpartum, seven should have received low-dose LMWH only in the postpartum period.
Incidence of VTE
When limiting the analysis to the first pregnancy in the database, five of 91 women had an episode of pregnancy-related VTE, resulting in an overall incidence of 5.5% (95% CI, 2.4–12.3) (Table 3). All events occurred in women who were considered at high risk of VTE and had used LMWH during pregnancy and the postpartum period (See Supplementary Table S1 for characteristics of women who experienced a pregnancy-related VT). One (1.1%; 95% CI, 0.3–5.9) event occurred antepartum, in a woman with a history of postpartum deep venous thrombosis (DVT) in the left pelvic vein after a previous pregnancy. Nadroparin 2850 IE administrated subcutaneously once daily had been started in the 22nd week of the current pregnancy. In the 28th week, she developed new-onset pain in the left leg. A CUS of the left leg showed a pelvic vein thrombosis and partial non-compressibility of the femoral vein. She was treated with therapeutic-dose LMWH during the remainder of her pregnancy and for 6 weeks postpartum. Because the possibility of residual DVT could not be entirely excluded, we defined this event as probable. Four (4.4%; 95% CI, 1.8–10.8) of the pregnancy-related episodes of VTE occurred in the postpartum period. In three cases, the episode of VTE was a recurrent DVT of the left leg. Two events were classified as definite, whereas one case was adjudicated as probable. One woman had a DVT of the upper extremity. This woman was hospitalized because of endometritis. She had used nadroparin during her hospital stay and had been given intravenous medication through a cannula in the cubital vein of the arm in which she subsequently developed DVT.
When we analyzed all 126 pregnancies in which low-dose LMWH was used, seven (5.6%; 95% CI, 2.8–11.0) pregnancy-related episodes of VTE were diagnosed. Again, all events occurred in those women who were considered at high risk and had used low-dose LMWH during pregnancy and postpartum. Two (1.6%; 95% CI, 0.4–5.6) events were diagnosed antepartum. In addition to the above-mentioned episode, another event occurred in a woman who was heterozygous for the FV Leiden mutation who had a history of recurrent episodes of DVT in the left leg. Despite the use of nadroparin 2850 IE anti-Xa subcutaneously once daily since the 9th week of pregnancy, she developed a definite DVT in the popliteal vein in the right leg in the 11th week of pregnancy. Five events developed postpartum, of which four events have been mentioned previously. The other event consisted of a recurrent DVT of the left leg, which was assessed as definite. Figure 2(A,B) shows the timing of events in the postpartum period for the first eligible pregnancies in our database (2A), and for all pregnancies (2B) in women at intermediate and high risk of VTE. Two of the five (40%) postpartum events occurred after discontinuation of the 6 weeks postpartum LMWH prophylaxis.
All seven events occurred in women with a previous VTE; six (85.7%) women had a previous episode of DVT and three (42.9%) had a history of pulmonary embolism. Five (71.4%) of these previous VTEs were hormone related. Four events occurred in women with an underlying thrombophilic disorder (one heterozygous and two homozygous for the FV Leiden mutation, one heterozygous for the prothrombin 20210A mutation). Six of the seven women had weight recorded at the first visit during pregnancy. Three women had a weight of ≥ 80 kg [(83 kg, BMI 29.4; 84 kg, BMI 26.8; 86 kg, BMI 28.4).
Incidence of peripartum bleeding
PPH complicated 21.6% (95% CI, 14.3–31.3) and severe PPH complicated 9.1% (95% CI, 4.7–16.9) of the deliveries (Table 4). This risk was not different between women who had used LMWH postpartum only and those who had used LMWH during pregnancy (30.3% vs. 16.4%, P = 0.12, and 9.1% vs. 9.1%, P = 1.00, respectively). PPH was observed in 16.9% of the vaginal deliveries and 41.2% of the Caesarian sections, and severe PPH in 8.5% of the vaginal deliveries and 11.8% of the Caesarian sections. The incidence of PPH or severe PPH was not significantly different between women who started PPH postpartum or used LMWH during pregnancy after vaginal delivery (P = 0.24 and P = 0.43, respectively) or Caesarean section (P = 0.49 and P = 0.16, respectively). The median estimated blood loss was 350 mL both for vaginal deliveries (range 50–2200) and Caesarean sections (range 100–3500) and was comparable between women who had used LMWH postpartum and those who had used LMWH during pregnancy (P = 0.43 and P = 0.42, respectively). A blood transfusion within 24 h after delivery was given to three (9.1%) women who had started LMWH in the postpartum period and three (5.5%) women who had been using LMWH during pregnancy, which was not different between the two groups (P = 0.51). These results did not change materially when we analyzed all 126 pregnancies.
Table 4. Peripartum bleeding complications
All women (n = 88*)
Intermediate risk (LMWH postpartum) (n = 33†)
High risk (LMWH during pregnancy and postpartum) (n = 55‡)
In this retrospective cohort study of 91 women at intermediate or high risk of pregnancy-related VTE, we observed an incidence of pregnancy-related VTE of 5.5% (95% C.I, 2.4–12.3) despite prophylaxis with low-dose LMWH. All events occurred in women considered to be at high risk, and were notably high postpartum (7.0%; 95% CI, 2.9–16.7) compared with antepartum (1.8%; 95% CI, 0.4–9.2). PPH occurred frequently but its incidence did not differ between women who had started LMWH postpartum and those who had used low-dose LMWH during pregnancy.
Several observational studies [6,11–13] found that the absolute risk of recurrent VTE without thromboprophylaxis in pregnant women with a previous VTE varied between 2.4–25% antepartum and 5.1–8.3% postpartum. The lowest incidence of recurrent VTE antepartum was seen in the study of Brill-Edward . Women with known thrombophilia were excluded and in the majority (67%) of the previous episodes of VTE a temporary risk factor was present. In our study the incidence of recurrent antepartum VTE was only slightly lower (1.8%) despite the use of LMWH prophylaxis. The risk of recurrent VTE was higher in our population, however, only 12.9% of the previous episodes of VTE were related to a temporary risk factor and 41.8% of the women had thrombophilia. Two studies [6,14] restricted their analysis to thrombosis prophylaxis in pregnant women with a history of thrombosis. In the study by Rozanski , which appeared as an abstract only, pregnant women with a previous VTE were divided into an intermediate risk group (previous VTE secondary to a transient risk factor) and high risk group (previous idiopathic VTE). The first group only received LMWH or warfarin for 6 weeks postpartum, the latter group received LMWH during pregnancy and LMWH or warfarin for 6 weeks postpartum. The risk of antepartum VTE was 2.7% (95% CI, 0.5–13.8) and 3.0% (95% CI, 1.0–8.5). No postpartum events occurred. Compared with our study, they found a higher risk of antepartum VTE and a lower risk of postpartum VTE. Although the study population seems comparable with ours we cannot yet compare our results because of the lack of information regarding LMWH dose. In another observational study  of pregnant women with a previous VTE, 29.7% of the women received thromboprophylaxis antepartum (heparin 5000 Units two to three times per day or LMWH 4000–5000 anti-Xa Units per day). Recurrent VTE antepartum did not occur in the group who received LMWH prophylaxis and was 6.2% (95% CI, 1.6–10.9) in the group without thrombosis prophylaxis. Interestingly, in this study the risk of postpartum recurrent VTE was 6.3% in 63 women who had received low-dose LMWH prophylaxis, and 6.7% in 75 women who had not received prophylaxis. It is not clear on what grounds women received prophylaxis or not, therefore confounding by indication is possible. Of the women who used antepartum prophylaxis, 94% had a previous episode of VTE associated with a temporary risk factor, which is markedly higher than in our population.
Most studies [4,7,14] conclude that the risk of recurrent VTE with the use of LMWH prophylaxis is low, although in these studies there were still women who experienced a VTE despite prophylaxis. However, the interpretation of what risk is acceptable is a matter of opinion, where the burden of daily injections and skin reactions at the injection site  needs to be balanced against the absolute magnitude of a beneficial effect, as well as bleeding complications. In our study, low-dose LMWH prophylaxis was associated with considerable treatment failure both antenatally and postpartum, in particular in women with a high risk of recurrent VTE. Forty per cent of the postpartum events occurred after the 6 weeks of LMWH prophylaxis. Although the risk of VTE is highest during the first 6 weeks postpartum, VTE risk remains elevated up to 3 months postpartum . Our study suggests that in the high-risk group, postpartum prophylaxis for 6 weeks might be too short.
We observed no recurrent VTE in women who were considered at intermediate risk of VTE who only used low-dose LMWH postpartum, indicating that it is possible to use risk stratification to identify a group of women in whom antepartum clinical surveillance and low-dose LMWH prophylaxis for 6 weeks postpartum is sufficient.
We observed PPH in 30.3% of women who started LMWH postpartum and 16.4% of women who used LMWH during pregnancy. Severe PPH complicated 9.1% of the deliveries in both groups. This is relatively high compared with other studies [16–19], which report a 2.1–12% risk of PPH in women using LMWH, though comparable to other Dutch studies [20,21]. In a population-based cohort study of nulliparous women who had a vaginal delivery, the incidence of PPH (> 500 mL blood loss) was 19% and severe PPH (> 1000 mL blood loss) was 4.2% . An explanation of the higher rates of PPH observed is that compared with other countries, active management during the third stage of delivery, such as prophylactic administration of oxytocics, is not routinely performed in the Netherlands. A problem regarding the classification of PPH is that the definition of PPH by the WHO criteria  of blood loss of > 500 mL within 24 h of delivery has been challenged because most women are not hemodynamically compromised with a blood loss of 500 mL. Therefore other criteria have been proposed that are more liberal, ranging from a 1000 to 2000 mL . We chose the commonly used definition of severe PPH as blood loss of > 1000 mL.
Although we did not have a control group, women who started LMWH postpartum could serve as a control group for women who used LMWH during pregnancy. The distribution of mode of delivery was comparable between these groups. Although PPH occurred frequently, its incidence did not differ between women who had started LMWH postpartum and those who had used low-dose LMWH during pregnancy. Therefore we considered the use of low-dose LMWH during pregnancy to be safe.
Several considerations warrant comment. The strength of our study is that we analyzed a relatively homogenous population treated with a consistent regimen of low-dose LMWH. We consider the risk of ascertainment bias for women with a recurrent VTE low because of the nature of our search strategy. A limitation of our study is its retrospective design. In the absence of a control group, we cannot formally assess the effectiveness of low-dose LMWH compared with no prophylaxis in high-risk women; the risk of recurrent VTE might be higher if no prophylaxis is given. Furthermore, we could not get objective diagnoses for all recurrent episodes of VTE. Finally, we cannot entirely exclude that events, particularly postpartum, were not recorded. However, this is unlikely because the obstetrician and/or internist instructed all women to report to our hospital with any complaints in the postpartum period, and if present, information bias would lead to an under-estimation of the true incidence.
In conclusion, VTE prophylaxis with low-dose LMWH in women at high risk of pregnancy-related VTE leads to considerable treatment failure. Although current guidelines support the use of low-dose LMWH during pregnancy and postpartum to prevent VTE in pregnant women, this dose might not be sufficient. A randomized trial comparing low-dose LMWH with high-dose LMWH in pregnant women at high risk of VTE would solve this issue.
Disclosure of Conflict of Interests
The authors state that they have no conflict of interest.