- Top of page
- Disclosure of Conflict of Interests
Summary. Venous thromboembolism (VTE) is one of the most relevant causes of maternal death in industrialized countries. Low molecular weight heparin (LMWH), continued throughout the entire pregnancy and puerperium, is currently the preferred treatment for patients with acute VTE occurring during pregnancy. However, information on the efficacy and safety of anticoagulant drugs in this setting is extremely limited. We carried out a systematic review and a meta-analysis of the literature to provide an estimate of the risk of bleeding complications and VTE recurrence in patients with acute VTE during pregnancy treated with antithrombotic therapy. The weight mean incidence (WMI) of bleeding and thromboembolic events and the corresponding 95% confidence interval (CI) were calculated. Eighteen studies, giving a total of 981 pregnant patients with acute VTE, were included. LMWH was prescribed to 822 patients; the remainder were treated with unfractionated heparin. Anticoagulant therapy was associated with WMIs of major bleeding of 1.41% (95% CI 0.60–2.41%; I) antenatally and 1.90% (95% CI 0.80–3.60%) during the first 24 h after delivery. The estimated WMI of recurrent VTE during pregnancy was 1.97% (95% CI 0.88–3.49%; I2 39.5%). Anticoagulant therapy appears to be safe and effective for the treatment of pregnancy-related VTE, but the optimal dosing regimens remain uncertain.
- Top of page
- Disclosure of Conflict of Interests
Venous thromboembolism (VTE), comprising deep vein thrombosis (DVT) and pulmonary embolism (PE), is one of the most relevant causes of maternal death in developed countries, with a reported mortality rate of 1.56 per 100 000 maternities in the UK [1,2]. Symptomatic VTE is estimated to occur in 5–12 women per 10 000 pregnancies antepartum, and in 3–7 women per 10 000 deliveries postpartum [3,4]. During pregnancy and puerperium, the risk of developing VTE is five times higher than in the general female population of childbearing age .
Treatment of VTE in pregnant women poses some particular challenges, because it requires taking into account the safety of the selected drugs not only for the mother, but also for the fetus. Vitamin K antagonists cross the placenta and have the potential to cause fetal bleeding and teratogenicity . Unfractionated heparin (UFH) does not cross the placenta and does not cause fetal teratogenicity, but its use may be associated with additional maternal safety issues, including the risk of heparin-induced thrombocytopenia (HIT) and heparin-associated osteoporosis . Furthermore, the use of therapeutic doses of UFH requires regular laboratory monitoring of the activated partial thromboplastin time (APTT). Like UFH, low molecular weight heparin (LMWH) does not cross the placenta, and there is no evidence of teratogenicity or risk of fetal bleeding [6,7]. In addition, LMWH offers a number of advantages over UFH, thanks to its better bioavailability, longer plasma half-life, more predictable dose response, and improved safety profile with respect to osteoporosis and HIT . For these reasons, LMWH is currently recommended as the treatment of choice for patients with acute VTE occurring during pregnancy [2–4,6–10]. However, evidence to support this recommendation is largely based on case reports or case series of pregnant patients, or on data derived from studies carried out in non-pregnant patients. Consequently, limited data exist on the incidence rate of recurrent VTE or bleeding during treatment with LMWH or UFH in pregnancy and puerperium, and little is known about their optimal therapeutic dosage. Thus, different therapeutic strategies are proposed and used in clinical practice. [3,6,8,9]. Although a few systematic and narrative reviews have tried to assess the risk–benefit profile of anticoagulant treatment of VTE in pregnancy [3,9–11], these studies have often mixed together data from patients receiving anticoagulant therapy for different indications, such as VTE treatment or prophylaxis, prevention of obstetric complications, and prevention of arterial thrombosis in patients with mechanical cardiac valves, or, conversely, have focused on only a single therapeutic agent.
We therefore decided to carry out a systematic review of the literature and a meta-analysis of the studies that have reported on recurrent VTE or bleeding events in patients receiving anticoagulant drugs for the treatment of pregnancy-related VTE, with the aim of providing an estimate of the incidence rates of these complications.
- Top of page
- Disclosure of Conflict of Interests
A total of 5011 (2831 MEDLINE and 2180 EMBASE) citations, six abstracts from ISTH meetings and two abstracts from ASH meetings were identified by our systematic search. A total of 4988 studies, three ISTH abstracts and two ASH abstracts were excluded after review of the study titles or abstracts. We subsequently retrieved the full text of 31 potentially eligible articles and three ISTH abstracts. Among the studies published as full articles, we further excluded one article because no antithrombotic therapy was used, seven because they included ≤ 10 patients, two because no data on follow-up were provided, and six because data on patients treated for acute VTE were not distinguishable from data on women receiving antithrombotic prophylaxis. Thus, a total of 15 full articles [19–33] and three abstracts [34–36] were eligible for inclusion in our systematic review (Fig. 1).
The interobserver agreement for the study selection was optimal (k = 0.91). According to the Newcastle-Ottawa scale, no studies were of high quality and four studies were of medium quality [25,27,30,33]. Eight studies specified that the diagnosis of VTE was made with imaging techniques [20,21,23,25,26,28,32,33], two studies defined the diagnosis of recurrent VTE [20,32], and only one study used the classical definition of PPH to describe postpartum hemorrhages .
Among the selected studies, 17 were written in English [19,21–36] and one in German . The corresponding authors of 11 of the selected studies were contacted, with the aim of obtaining missing information, and six authors kindly provided the requested data.
A total of 981 pregnant women with acute VTE were included, with their age ranging from a minimum of 17 years to a maximum of 43 years. Only 13 of the selected studies, including a total of 632 patients, provided separate data on the number of DVTs and PEs (in total: 506 DVTs and 127 PEs) [19–21,23–26,29,30,32,33,35,36] (Table 1). The timing of VTE onset during pregnancy was not reported in all studies. Where available, this ranged from the 7th gestational week to the 38th gestational week.
Table 1. Baseline characteristics of the study populations
|Author||Type of study||Patients (no.)||Age (years)||Gestational age||BMI||DVT||PE|
|Aburahma and Boland ||Case–control||24||Mean 24 (17–39)||1 first trimester 5 second trimester 18 third trimester||NR||24||NR|
|Bahlmann et al. ||Cohort||12||NR||26.1 ± 6.2 weeks||NR||12||0|
|Barillari et al. ||Cohort||38||NR||NR||NR||NR||NR|
|Blanco-Molina et al. ||Cohort||173||Mean 31 ± 6||NR||NR||135||38|
|Clark et al. ||Cohort||17||NR||Median 19 weeks (7–34 weeks)||NR||NR||NR|
|Daskalakis et al. ||Case series||18||27–43||7–38 weeks||NR||18||0|
|Donnelly et al. ||Cohort||25||NR||NR||NR||21||4|
|Jacobsen et al. ||Cohort||20||Mean 31.8 ± 5.19||4 first trimester 6 second trimester 11 third trimester||NR||19||2|
|Mitic et al. ||Cohort||87||Mean 29.7 ± 4.86||Mean 22 ± 10.2 weeks||NR||82||5|
|Narin et al. ||Case–control||35 (18, group I 17, group II)||Mean 28.4 ± 3.5, group I Mean 30.0 ± 5.2, group II||Mean 29.3 ± 4.4 weeks, group I Mean 26.5 ± 6.5 weeks, group II||NR||35||0|
|Nelson-Piercy et al. ||Cohort||247||Mean 30.1*||NR||Mean 27.7*||NR||NR|
|O’Connor et al. ||Case series||34||NR||NR||NR||NR||NR|
|Parent et al. ||Cohort||39||Mean 32.5||NR||NR||20||19|
|Rodie et al. ||Case series||29||NR||NR||NR||23||6|
|Roshani et al. ||Cohort||13||Mean 32||NR||NR||8||5|
|Rowan et al. ||Cohort||13||NR||NR||NR||NR||NR|
|Ulander et al. ||Case–control||31 (10, group I 21, group II)||31.0 ± 5.7, group I 31.6 ± 4.3, group II||27 ± 8.5 weeks, group I 21 ± 9.8 weeks, group II||23.4 ± 4.7, group I 25 ± 4.6, group II||31||0|
|Voke et al. ||Cohort||126||Median 32 (16–42)||31 first trimester 37 second trimester 58 third trimester||Median 26 (19–43)||78||48|
|All 18 studies|| ||981|| || || ||506||127|
Overall, 822 patients received LMWH and 155 UFH for the acute-phase treatment of VTE. None of the studies reporting on patients treated with fondaparinux met our selection criteria. Dosing regimens of LMWH were heterogeneous among studies, and are summarized in Table 2. Only in two studies, during the acute phase, were the dosages adjusted according to the measurement of anti-factor Xa levels [25,32]. UFH was administered both intravenously and subcutaneously, in most cases, but not always, according to APTT values (Table 2).
Table 2. Treatment regimens for the acute phase of venous thromboembolism
|Author||LMWH||Dose of LMWH||UFH||Dose of UFH||Warfarin|
|Aburahma and Boland ||0||–||24||11 patients i.v. according to APTT (target 1.5–2.5 times normal); 13 patients i.v. bolus then 5000–10 000 U s.c. every 8–12 h||0|
|Bahlmann et al. ||1 dalteparin||5000 IU s.c. three times daily||11||According to APTT (target 1.5–2.0 times normal)||0|
|Barillari et al. ||38 nadroparin||100 IU kg−1 twice daily||0||–||0|
|Blanco-Molina et al. *||154||Mean 187 ± 51 IU kg−1 daily||16||NR||0|
|Clark et al. ||2 enoxaparin||100–110 mg kg−1 twice daily||15||6 patients s.c.||0|
|Daskalakis et al. ||0||–||18||According to APTT (target 2.0 times normal)||0|
|Donnelly et al. ||25||NR||0||–||0|
|Jacobsen et al. ||20 dalteparin||Initially 100 IU kg−1 twice daily, then according to anti-FXa levels (0.5–1.0 U mL−1)||0||–||0|
|Mitic et al. ||84 (83 nadroparin, 1 dalteparin)||100 IU kg−1 twice daily||3||NR||0|
|Narin et al. ||0||–||35||According to APTT (target 1.5–2.5 times normal)||0|
|Nelson-Piercy et al. ||247 tinzaparin||Median 13 000 IU (3500–28 000)‡||0||–||0|
|O’Connor et al. ||23||NR||11||NR||0|
|Parent et al. ||39 tinzaparin||18 119 UI kg−1 daily||0||–||0|
|Rodie et al. ||29 enoxaparin||1 mg kg−1 twice daily||0||–||0|
|Roshani et al. †||9 nadroparin, 2 dalteparin||100 IU kg−1 twice daily or 200 IU kg−1 once daily||1||NR||0|
|Rowan et al. ||13 enoxaparin||1 mg kg−1 twice daily||0||–||0|
|Ulander et al. ||21 dalteparin||Based on anti-FXa (target: 1–1.5 U mL−1)||10||According to APTT (70–100 s)||0|
|Voke et al. ||115||83 patients once daily, 39 patients twice daily||11||NR||0|
|Total||822|| ||155|| ||0|
Treatment regimens during the long-term treatment period are summarized in Table 3. The majority of patients initially treated with UFH were switched to LMWH, others were continued on UFH, in most cases administered subcutaneously, and a few (36 from a single study ) were switched to warfarin. In one study, four patients initially treated with LMWH were switched to UFH during the late phase of pregnancy . Six studies reported on dose changes to LMWH during the long-term treatment period according to anti-FXa levels in some patients [25,29,31–33,36].
Table 3. Treatment regimens for the long-term treatment period
|Author||LMWH||Dose||UFH||Dose||Warfarin||Duration of therapy|
|Aburahma and Boland ||0||–||24||5000–10 000 U s.c. every 8–12 h||0||6–8 weeks after delivery|
|Bahlmann et al. ||2||500–600 IU h−1 i.v.||108 pt s.c., 2 pt i.v.||5000–7000 U × 3 s.c.||0||NR|
|Barillari et al. ||38||80–100 IU kg−1 once||0|| ||0||3 months after delivery|
|Blanco-Molina et al. *||133||Mean 173 ± 59 IU kg−1 daily||0||–||36||NR|
|Clark et al. ||3||70 mg Kg−1 twice daily, 17 000 IU once daily||14 s.c.||9000–13 000 IU every 8–12 h||0||NR|
|Daskalakis et al. ||18||6150 anti-FXa IU once daily||0||–||0||1 month after delivery|
|Donnelly et al. ||25||NR||0||–||0||NR|
|Jacobsen et al. ||20||According to anti-FXa levels (0.5–1.0 U mL−1)||0||–||0||NR|
|Mitic et al. ||85||Intermediate dosage||2||NR||0||NR|
|Narin et al. ||35||1 mg kg−1 twice daily, group I 1.5 mg kg−1 once daily, group II||0||–||0||nr|
|Nelson-Piercy et al. ||247||Median 13 000 IU (3500–23 100)†||0||–||0||NR|
|O’Connor et al. ||23||NR||11||NR||NR||NR|
|Parent et al. ||39||According to anti-FXa levels||0|| ||0||2210 weeks|
|Rodie et al. ||29||According to anti-FXa levels (0.4–1.0 U mL−1)||0||–||0||Median 6 weeks (1–33 weeks)|
|Roshani et al. ||9 nadroparin, 2 dalteparin||100 IU kg−1 twice daily or 200 IU kg−1 once daily||1||NR||0||NR|
|Rowan et al. ||13||Some patients according to anti-FXa levels||0||–||0||NR|
|Ulander et al. ||31||100 UI kg−1 twice daily for 2 weeks, then once daily according to anti-FXa levels (0.7 U mL−1 and 0.5–0.6 U mL-1 at the end of pregnancy)||0||–||0||NR|
|Voke et al. ||118||Some patients according to anti-FXa levels||4||NR||0||NR|
|Total||874|| ||66|| ||36|| |
The duration of the follow-up period was heterogeneous among studies, and in some cases it was not reported. In one study, follow-up was limited to the first 3 months after the diagnosis of VTE . From the information provided, we could not estimate the total duration of the treatment period.
Bleeding complications in the antepartum period
During the antepartum period, a total of 28 bleeding events were reported in 16 studies in a total of 944 patients [19,21–23,25–36]. Regrettably, the available data were not sufficient to enable us to apply the ISTH classification of bleeding severity, so we could only use the definitions given by the authors in each study. Overall, five events were defined by the authors as major, one was defined as clinically relevant non-major, 16 were defined as minor, and six were not classified [27,28]. In one patient, major bleeding developed after severe eclampsia complicated by thrombocytopenia . Four of the five major bleeds occurred during the acute-phase treatment [19,21], and the only clinically relevant non-major bleed occurred after 2 weeks of treatment . Of the four major bleeds that occurred during the acute-phase treatment, two occurred in patients treated with LMWH and two in patients treated with UFH.
The use of anticoagulant therapy was therefore associated with an antepartum incidence of hemorrhagic complications of 3.28% (95% confidence interval [CI] 2.10–4.72; I2 14.6%). The antepartum incidence of major bleeding was 1.41% (95% CI 0.60–2.41%; I2 0%); the incidence of major bleeding during the acute treatment phase only was 1.00% (95% CI 0.4–2.0%; I2 0%) (Fig. 2 A-B). From the available data, it was not possible to separately estimate the incidence of bleeding complications for UFH and LMWH.
Figure 2. (A) All antepartum bleeds (WMD 3.28% [95% confidence interval (CI) 2.10–4.72; I2 14.6%]). (B) Major antepartum bleeds (WMD 1.41% [95% CI 0.60–2.41; I2 0%]). (C) Major bleeds in the first 24 h postpartum (WMD 1.00% [95% CI 0.4–2.0%; I2 0%]). (D) Major postpartum bleeds after 24 h (WMD 1.2% [95% CI 0.30–2.50%; I2 0%]).
Download figure to PowerPoint
Bleeding complications in the postpartum period
A total of 260 PPHs were reported in 13 studies [22–27,29,30,32–36], in a total of 725 patients. In most studies, information on blood losses was sufficient to enable us to use the classification of the RCOG. In total, 14 PPHs were defined as major, 41 were defined as clinically relevant non-major (four of these were atonic postpartum bleeds), and 205 were defined as minor (Table 4).
Table 4. Adverse events
|Author||Antepartum bleeding||Antepartum recurrent VTE||Other antepartum adverse events||Peripartum bleeding (first 24 h after delivery)||Other postpartum bleeding||Other postpartum adverse events||Follow-up|
|Aburahma and Boland ||1 retroperitoneal bleed in acute phase*||2 PEs, one fatal, in acute phase on UFH*||NR||NR||NR||NR||61 months (18–102)|
|Bahlmann et al. ||NR||4 (2 PEs): 3 in acute phase on UFH||NR||NR||NR||NR||NR|
|Barillari et al. ||0||0||0||0||1 epistaxis||NR||1 year after delivery|
|Blanco-Molina et al. ||3 major bleeds in acute phase: 2 on UFH; 1 on LMWH*||2 PEs during LMWH after acute phase*||NR||NR||NR||NR||3 months after VTE|
|Clark et al. ||0||1 PE in acute phase on LMWH*||NR||1 wound hematoma 1 minor bleed||1 clinically relevant non-major bleed 1 major bleed||NR||NR|
|Daskalakis et al. ||0||0||1 abortion (at 10 weeks‡)||0||NR||0||NR|
|Donnelly et al. ||NR||NR||NR||0 bleeds > 1500 mL||NR||NR||NR|
|Jacobsen et al. ||0||0||1 intrauterine death (at 37 weeks‡)||1 atonic postpartum bleed 1 wound hematoma||NR||0||NR|
|Mitic et al. ||3 minor bleeds||1 DVT in acute phase on LMWH*||4 skin reactions on nadroparin 1 fetal loss (at 8 weeks‡)||2 clinically relevant non-major bleeds 2 minor bleeds 1 wound hematoma||0||0||NR|
|Narin et al. ||1 vaginal bleed, 1 hematuria and 1 clinically relevant vaginal bleed on LMWH||0||1 abortion (at 9 weeks)||0||0||0||4–8 weeks after delivery in 57% of patients|
|Nelson-Piercy et al. ||4 unspecified†||5 (4 PEs)‡||NR||196 bleeds < 500 mL 31 bleeds > 500 mL and < 1000 mL 3 bleeds > 1000 mL§||NR||2 stillbirths 1 termination¶||NR|
|O’Connor et al. ||2 unspecified||NR||NR||NR||4 unspecified||NR||NR|
|Parent et al. ||1 hematuria 2 vaginal bleeds 1 major bleed on LMWH||0||1 fetal death 1 abortion 1 eclampsia with thrombocytopenia (gestational age not specified)||1 wound hematoma||NR||NR||NR|
|Rodie et al. ||0||0||2 skin reactions (on enoxaparin)||3 atonic postpartum bleeds 1 wound hematoma||NR||0||NR|
|Roshani et al. ‡||0||0||NR||2 major bleeds 2 bleeds < 1000 mL||1||NR||NR|
|Rowan et al. ||0||0||NR||NR||NR||NR||NR|
|Ulander et al. ||1 hematuria||1 during LMWH after the acute phase*||1 premature delivery (at 23 weeks‡)||3 abnormal bleeds > 1000 mL 1 wound hematoma||NR||NR||NR|
|Voke et al. ||7 minor bleeds||0||2 intrauterine deaths (at 38 and 23 weeks‡) 1 miscarriage (at 12 weeks) 1 abortion (at 12 weeks‡)||6 bleeds > 1000 mL 2 bleeds > 500 mL and < 1000 mL||2 major bleeds 4 clinically relevant non-major bleeds||1 neonatal death||NR|
|Total||28: 5 major bleeds; 1 clinically relevant non-major bleed; 16 minor bleeds; 6 unspecified||16: 11 PEs; 1 fatal‡||17: 6 skin reactions; 4 abortions; 1 miscarriage; 4 intrauterine deaths; 1 premature delivery; 1 eclampsia||260: 205 minor bleeds; 41 clinically relevant non-major bleeds; 14 major bleeds||14: 3 major bleeds; 5 clinically relevant bleeds; 5 unspecified; 1 minor bleed||1 neonatal death 2 stillbirths¶ 1 termination ¶|| |
The incidence of major PPH was 1.90% (95% CI 0.80–3.60%; I2 36.8%) (Table 5; Fig. 2 C-D). Only seven studies, with a total of 350 patients, provided information on bleeding events occurring after the first 24 h from delivery. The information provided on bleeding events was insufficient for the application of any classification, and the definition of severity provided by the authors was used. Information on the duration of follow-up was not available. A total of 14 hemorrhagic events were reported: three were defined as major (1.2%, 95% CI 0.30–2.50%; I2 0%).
Table 5. Bleeding complication rates
| ||Incidence (%)||95% CI|
|All antepartum bleeding complications||3.28||2.10–4.72|
|Antepartum major bleeding||1.41||0.60–2.41|
|Acute antepartum major bleeding*||1.00||0.40–2.00|
|Postpartum major bleeding (< 24 h)||1.90||0.80–3.60|
|Puerperium major bleeding†||1.2||0.30–2.5|
Information on recurrent VTEs occurring during the antepartum period was available in 16 studies including a total of 922 patients. There were 16 recurrent VTEs: 11 were PEs (one fatal) and five were DVTs (Table 4).
The estimated WMI of antepartum recurrent VTE was 1.97% (95% CI 0.88–3.49%; I2 39.5%) (Fig. 3). The estimated WMI of recurrent PE was 1.30% (95% CI 0.6–2.3%; I2 13.3%). Seven VTE recurrences occurred in the first 7 days after the start of treatment, and five of these occurred while patients were receiving UFH treatment [19,20,22,35].
The incidence rate of recurrent VTE during the acute-phase treatment period was 1.41% (95% CI 0.44–2.90%; I2 38%); during the long-term treatment period, this incidence was 0.95% (95% CI 0.37–1.82%; I2 0%). Once again, from the available data, we could not separately estimate the incidence of bleeding complications for UFH and LMWH.
Two studies [19,22], including 41 patients, reported on four cases of recurrent VTE that occurred during the postpartum period. Finally, we found no reported cases of HIT in the 13 studies that addressed this complication, in a total of 860 patients [21–23,25–27,29,31–36]. The occurrence of HIT was prospectively assessed in only two studies [21,33].
- Top of page
- Disclosure of Conflict of Interests
The results of this systematic review of the literature substantially support the efficacy and safety of currently used therapeutic strategies for the treatment of pregnancy-related VTE, but suggest that some work is needed to improve their efficacy and safety profile, in particular during the highest-risk periods. In fact, the estimated incidence of recurrent events during pregnancy remains substantial during the first week of treatment, as do the rates of major bleeding complications, again during the first week of treatment and in the first 24 h after delivery.
We observed substantial heterogeneity in the reported treatment regimens: although most patients received LMWH, various dosing regimens for the acute-phase treatment and, in particular, for long-term secondary prevention were prescribed in the studies, and no direct comparisons are currently available. There was also heterogeneity among definitions of major bleeding events before and after delivery in individual studies. Because the application of a single, more widely accepted definition was not always feasible, given the limited information provided in some studies, the observed rates of bleeding should be interpreted cautiously.
The 2012 version of the guidelines of the American College of Chest Physicians (ACCP) recommends the use of adjusted-dose LMWH for the treatment of VTE during pregnancy . LMWH is considered to be the best option, because of its greater bioavailability and its favorable safety profile. It is also recommended that anticoagulant therapy should be continued with the same dose throughout pregnancy, and until at least 6 weeks after delivery . Clear-cut indications about the need for dose adjustments over the course of pregnancy or the usefulness of routine measurement of anti-FXa activity were not provided . In the RCOG guidelines published in February 2007 , the authors suggested treating VTE with LMWH given twice daily throughout pregnancy, and discouraged physicians from the routine measurement of peak anti-FXa activity, with the exception of women at extremes of body weight or with other complicating factors, such as renal insufficiency. In our search, we found six studies in which, in some patients, the LMWH dosages were modified according to anti-FXa levels. Unfortunately, a separate assessment of clinical outcomes in this subgroup of patients was not feasible. During the acute-phase treatment period, most patients received weight-adjusted, full-dose LMWH, with an acceptable incidence of major bleeding events; during the long-term treatment period, doses were empirically reduced for most patients, and a minority of patients received dose adjustments based on the measurement of anti-FXa levels.
It is of note that the rates of major bleeding events in our study are consistent with the rates of major bleeding events reported in studies carried out in non-pregnant patients treated for acute VTE with standard anticoagulant therapy [4,37]. Conversely, the estimated incidences of bleeding and recurrent VTE events in our study differ somewhat from the results of a previous systematic review published in 2005 by Greer et al. . In this study, the authors assessed the safety and efficacy of LMWH during pregnancy, and included both patients treated for pregnancy-related DVT or PE (n = 146) and patients receiving antithombotic prophylaxis. The overall rates of what the authors defined as significant bleeding were 1.98% (95% CI 1.50–2.57%) in the antepartum and the postpartum periods combined, and 0.43% (95% CI 0.22–0.75%) in the antenatal period only. The rate of bleeding in the postpartum period in patients receiving LMWH for the treatment of acute VTE was 1.72%. These rates are lower than the rates reported in our study, probably because we included only patients receiving therapeutic doses of anticoagulants and patients receiving UFH, and possibly because of a different definition of bleeding events. Finally, the rates of recurrent VTE in our study were also higher than those reported in the review by Greer et al. (1.15%).
Our meta-analysis has a number of limitations. First, only case–control and cohort studies have been published, and the application of formal meta-analytic methods to observational studies is controversial, as the bias implicit in the study design may misrepresent the strength of associations within the data . In addition, we considered studies with different end-points and with different durations of follow-up. Moreover, the therapeutic strategies were highly heterogeneous among studies, and different compounds with different dosing regimens were used. Unfortunately, separate analysis of these therapeutic regimens was not feasible, owing to insufficient information being provided in the selected studies. Another limitation is the lack of standardized definitions of bleeding severity among selected studies, which mandates some caution when the reported figures are interpreted. This applies in particular to the definition of PPH. PPH is traditionally defined as any blood loss >500 mL from the genital tract during delivery [13,14,38], but other definitions also required the presence of clinical signs of blood loss . Furthermore, we could not obtain data related to the entire duration of the puerperium: first, because of the wide variability in follow-up duration among the selected studies; and second, because, in most studies, data on anticoagulant treatment and on related complications after the delivery were not provided. Similarly, the criteria used to diagnose recurrent VTE were not reported in most studies, and it is possible that some heterogeneity exists in the adjudication criteria that were applied. Moreover, we had insufficient information to identify early recurrences that may have been caused by thrombosis extension or inadequate anticoagulation. Finally, we acknowledge the existence of studies conducted with other anticoagulant agents, including danaparoid  and fondaparinux. However, treatment with a heparinoid is not standard practice, and is not listed in the recommended strategies in the most important guidelines (ACCP and RCOG); in addition, in our search, we identified one study on the use of danaparoid in pregnant patients who were intolerant to heparin (also in the case of previous HIT), but in this article we could not distinguish data from patients treated for an acute episode of VTE or data from patients treated for the prophylaxis of VTE, so we decided to not include the study in the meta-analysis; none of the studies with fondaparinux was eligible for inclusion in our study, because none had >10 included patients.
In conclusion, this is, to our knowledge, the first systematic review and meta-analysis specifically aimed at estimating the incidence of recurrent VTE and bleeding during anticoagulant treatment for pregnancy-related VTE. The results of this study suggest that anticoagulant drugs during pregnancy are effective in the prevention of recurrent VTE, with an acceptable safety profile. Most events, both bleeding and recurrences, occurred during the first week of treatment. We could not identify the optimal therapeutic regimen among the different approaches used in the selected studies.
- Top of page
- Disclosure of Conflict of Interests
E. Romualdi: conception and design of the study, acquisition of data, analysis and interpretation of data, drafting of the manuscript, and statistical analysis; F. Dentali: conception and design of the study, analysis and interpretation of data, critical revision of the manuscript, and supervision and statistical analysis; A. Squizzato: conception and design of the study, analysis and interpretation of data, critical revision of the manuscript, and supervision; E. Rancan: acquisition of data; L. Steidl: acquisition of data and critical revision of the manuscript; S. Middeldorp: critical revision of the manuscript; W. Ageno: conception and design of the study, analysis and interpretation of data, critical revision of the manuscript, and supervision.