Pregnancy complications in women with recurrent miscarriage associated with antiphospholipid antibodies treated with low dose aspirin and heparin

Authors


Correspondence: Professor L. Regan, Department of Obstetrics and Gynaecology, Imperial College School of Medicine at St Mary's, Praed Street, London W2 1PG, UK.

Abstract

Objective To study the obstetric course of women with a history of recurrent miscarriage associated with antiphospholipid antibodies, lupus anticoagulant and anticardiolipin antibodies, treated with low dose aspirin and low dose heparin.

Design Prospective observational study.

Setting University based tertiary referral clinic.

Population One hundred and fifty pregnant women with a history of recurrent miscarriage associated with persistently positive tests for antiphospholipid antibodies.

Methods Lupus anticoagulant was detected using the dilute Russell's viper venom time together with a platelet neutralisation procedure. IgG and IgM anticardiolipin antibodies were detected using a standardised enzyme linked immunosorbent assay. An IgG anticardiolipin level ≥5 per litre units and an IgM anticardiolipin level ≥3 per litre units was considered positive. Aspirin (75 mg daily) was commenced at the time of a positive pregnancy test and heparin (5000 units subcutaneously 12 hourly, or enoxaparin 20 mg daily) was started when fetal heart activity was demonstrated on ultrasound. Treatment was stopped at the time of miscarriage or at 34 weeks of gestation.

Results One hundred and seven pregnancies (71%) resulted in a live birth. Forty-one pregnancies (27%) miscarried, the majority in the first trimester. One woman had a stillbirth, and one a premature baby who died in the neonatal period. One pregnancy was terminated for a fetal anomaly. Gestational hypertension complicated 17% (18/108) of ongoing pregnancies and antepartum haemorrhage 7% (8/108). Twenty-six babies (24%) were delivered before 37 weeks of gestation. Fifty women (46%) were delivered by caesarean section. The median birthweight of all live born infants was 3069 g (range 5314300); however 15% (16/108) of the infants were small for gestational age.

Conclusion Combination treatment with aspirin and heparin leads to a high live birth rate among women with recurrent miscarriage and antiphospholipid antibodies. However, successful pregnancies are prone to a high risk of complications during all trimesters. Close antenatal surveillance and planned delivery of these pregnancies in a unit with specialist obstetric and neonatal intensive care facilities are indicated.

INTRODUCTION

Antiphospholipid antibodies, lupus anticoagulant and anticardiolipin antibodies are a family of auto-antibodies associated with recurrent miscarriage1, severe pre-eclampsia2, fetal growth restriction3,4 and placental abruption5.

The pathogenesis of poor pregnancy outcome associated with antiphospholipid antibodies is uncertain.

Initial reports, based on the histological examination of placentae obtained from women with antiphospholipid antibodies, suggested that thrombosis of the uteroplacental vasculature leading to placental insufficiency was the primary mechanism of action of these antibodies6,7. However, these findings are neither universal nor exclusive to pregnancies complicated by antiphospholipid antibodies7,8. More recently, in vitro studies have suggested that antiphospholipid antibodies directly impair trophoblast function leading to defective implantation and invasion of the uterine decidua9–12.

We have previously reported that 15% of women with recurrent miscarriage have persistently positive tests for antiphospholipid antibodies1. In future pregnancies in which no pharmacological treatment is given, these women experience a live birth rate as low as 10%13. This poor obstetric performance has led to the use of a variety of treatments, including steroids14,15, aspirin15,16, heparin14,16,17 and immunoglobulin18,19, in an attempt to improve the live birth rate among women with antiphospholipid antibodies. Steroids have fallen into disfavour due to the significant fetal and maternal morbidity associated with their use15,20. The mainstay of treatment now rests with anti-platelet and anti-thrombotic treatments. Our recent randomised prospective study reported that the live birth rate in women with recurrent miscarriage and antiphospholipid antibodies increases to 40% when they are treated with low dose aspirin (75 mg daily) alone, and that this is further, and significantly, increased to 70% when they are treated with low dose aspirin together with low dose heparin (calcium heparin 5000 IU 12 hourly) [OR 3.37, 95% CI 1.40–8.10]17. However, successful pregnancies were complicated by prematurity and growth retardation. The current study was designed to explore the late pregnancy and perinatal course in women with recurrent miscarriage associated with antiphospholipid antibodies treated with low dose aspirin and low dose heparin during pregnancy.

METHODS

The pregnancies of 150 women (median age 33 years; range 19–45), with a history of recurrent miscarriage (median 4; range 3–15) and persistently positive tests for antiphospholipid antibodies, were studied. All women tested positive on at least two occasions more than eight weeks apart for either lupus anticoagulant or anticardiolipin antibodies. No woman had systemic lupus erythematosus21, and none had a history of previous thromboembolic disease. The demographic details of these women are shown in Table 1. The pregnancy outcomes of 45 of the 150 women included in this study have previously been reported in an earlier study17.

Table 1.  Demographic and laboratory details of the study population (n= 150).
 n [range] or (%)
Clinical details 
 Median age (years)33 [19–45]
 Median no. of previous miscarriages4 [3–15]
 First trimester miscarriages only99 (66)
 Second trimester miscarriages only7 (5)
 First and second trimester miscarriages44 (29)
 Previous live birth66 (44)
Laboratory details 
 Lupus anticoagulant only82 (55)
 IgG anticardiolipin antibody only17 (11)
 IgM anticardiolipin antibody only16 (11)
 Lupus anticoagulant + anticardiolipin antibody35 (23)

All women were screened for antiphospholipid antibodies on at least two occasions prior to pregnancy. None was taking an oral contraceptive preparation at the time of testing. Screening for antiphospholipid antibodies was performed as previously described1. In brief, lupus anticoagulant was detected using the dilute Russell's viper venom time together with a platelet neutralisation procedure. Patient samples with a dilute Russell's viper venom time ratio (test:control) of ≥ 1:1 were re-tested with a platelet neutralisation procedure. A decrease of 10% or more in the ratio was considered to be positive for lupus anticoagulant. Anticardiolipin antibodies were identified using a standardised enzyme-linked immunosorbent assay (ELISA). An IgG anticardiolipin level ≥5 per litre units and an IgM anticardiolipin level ≥3 per litre units were considered to be positive22. Only women with persistently positive tests for antiphospholipid antibodies were included in this study (Table 1). The median IgG anticardiolipin titre was 14.5 per litre units (range 5.3–140) and the median IgM anticardiolipin titre 7 per litre units (3.1–57).

All women commenced low dose aspirin (75 mg daily) as soon as the urine pregnancy test was positive (urinary human chorionic gonadotrophin level > 50 IU/mL; Clearview, Unipath, Bedford, UK). Women were invited to attend for serial first trimester ultrasound scans from 5 weeks of amenorrhoea. When fetal heart activity was seen heparin was started. The first 97 women received self-administered subcutaneous calcium heparin (5000 IU 12 hourly; Calciparine, Sanofi Winthrop, Surrey, UK) and the remaining 53 women received self-administered subcutaneous low molecular weight heparin, enoxaparin sodium (20 mg daily; Clexane, Rhone-Poulenc Rorer, Sussex, UK). Treatment with aspirin and heparin was stopped at the time of miscarriage or at 34 weeks of gestation.

Both the mother and the fetus were closely monitored. Women attended for antenatal care at fortnightly intervals until 34 weeks of pregnancy and weekly thereafter. As long term heparin treatment can be associated with the development of thrombocytopenia and osteopenia, a platelet count was performed every four weeks and dual energy X-ray bone densitometry (Lunar Corporation, Madison, Wisconsin, USA) of the lumbar spine, femoral neck and the forearm performed at 12 and 34 weeks of gestation and postnatally. All pregnancies were monitored with serial ultrasound scans from 24 weeks of gestation to assess fetal growth

RESULTS

The live birth rate was 71% (107/150). Forty-one pregnancies (27%) miscarried before 24 weeks of gestation.

One baby, who had a hypoplastic left heart, was stillborn, and one, born at 24 weeks of gestation, died in the early neonatal period as a result of extreme prematurity. One pregnancy was terminated at 18 weeks of gestation following an abnormal fetal karyotype, diagnosed after amniocentesis. The live birth rate was significantly higher (79/99; 80%) among those with a history of only first trimester miscarriage, compared with those with a previous second trimester loss (28/51; 55%: P= 0.001).

One hundred and eight pregnancies progressed beyond 24 weeks of gestation. The antenatal complications in this group are shown in Table 2. The most common medical complication was gestational hypertension, the development of a diastolic blood pressure of 90 mmHg or more after the 20th week of pregnancy in a previously normotensive woman. In the majority of these cases, (12118; 66%) hypertension was accompanied by proteinuria (> 300 mg per 24 hour), i.e. pre-eclampsia. Two women with hypertensive disease developed the HELLP syndrome (haemolysis, elevated liver enzymes and low platelets) and were delivered at 24 and 27 weeks of gestation. Two women with proteinuric hypertension were delivered at term of babies weighing less than the 10th centile for gestational age. There were no other cases of maternal or fetal complications associated with hypertensive disease. Eight women experienced an antepartum haemorrhage and four developed non-insulin dependent gestational diabetes. Three of these four had a significantly elevated body mass index immediately prior to pregnancy.

Table 2.  Complications among pregnancies progressing beyond 24 weeks of gestation (n= 108).
Complicationn (%)
Antepartum haemorrhage 
 Placental abruption5 (5)
 Placenta praevia3 (3)
Gestational diabetes4 (4)
Gestational hypertension 
 Proteinuric12 (11)
 Non-proteinuric6 (6)
Preterm delivery26 (24)

Preterm delivery (< 37 weeks of gestation) complicated 26 of the 108 pregnancies (24%) which progressed beyond 24 weeks of gestation. Causes of preterm delivery were as follows: spontaneous preterm labour (24%); preterm prelabour rupture of the membranes (20%); pre-eclampsia (20%); placental abruption (16%); intrauterine growth restriction (12%); and abnormal fetal heart rate (8%). The majority of the preterm deliveries (18/26; 69%) occurred between 34 and 36 weeks of gestation (Table 3). The median birth-weight of all live born infants was 3069 g (range 5314300 g). However, 15% (16/108) of the infants were small for gestational age (birthweight below the 10th centile for gestational age). Eighteen babies (17%) were of low birthweight (< 2500 g). Twelve of these babies were delivered before 37 weeks of gestation and six after this time.

Table 3.  Indications for caesarean section (n= 50).
 n (%)
Elective caesarean section 
 Poor obstetric history11(22)
 Fetal malpresentation8(16)
 Fetal growth restriction4(8)
 Placenta praevia2(4)
 Pre-eclampsia2(4)
 TOTAL27(54)
Emergency caesarean section 
 Fetal distress11(22)
 Failure to progress in labour6(12)
 Anteparturn haemorrhage3(6)
 Maternal HELLP syndrome2(4)
 Breech presentation in labour1(2)
 TOTAL23(46)

The overall caesarean section rate was 46% (50/108), the majority being elective procedures (27/50; 54%). In a substantial proportion of this subgroup (41%), poor previous reproductive history was cited as the main indication for elective caesarean section. Fetal distress, based on an abnormal fetal heart rate pattern and fetal blood sampling in labour, was cited in almost half of the cases as the main indication for emergency caesarean delivery. Other indications for caesarean section are listed in Table 4. The rate of instrumental vaginal delivery was 5% (5/108).

Table 4.  Gestational age at birth (n= 108).
Gestation (weeks)n(%)
24–274(4)
28–310
32–334(4)
34–3618(17)
31+82(76)

There were two perinatal deaths. One was due to extreme prematurity. The infant, delivered at 24 weeks of gestation by emergency caesarean section, effected in response to maternal HELLP syndrome, weighed 531 g. He died at five days of age due to severe respiratory distress syndrome. One infant was stillborn. His mother went into spontaneous labour at 27 weeks of gestation. A postmortem examination revealed left sided cardiac hypoplasia.

No women suffered a thromboembolic episode and none developed thrombocytopenia. The median change in lumbar spine bone mineral density between 12 weeks of gestation and the immediate postnatal period was −3.4% (range −11.7% to +9%). The density at the femoral neck changed by a median of −1.1% (range –11.3% to +9.4%). There was no significant change observed in the forearm bone mineral density. There was no significant difference in bone density at the three sites between those women treated with low molecular weight heparin compared with those receiving unfractionated heparin.

Of the 41 women who miscarried, 28 (68%) experienced an early pregnancy loss (< 14 weeks of gestation) and 13 (32%) had a late miscarriage (14–24 weeks of gestation). Four of the late pregnancy losses were intrauterine fetal deaths. Karyotyping of the products of conception was attempted in 32 of the 41 unsuccessful pregnancies. The fetal karyotype was normal in 19 and abnormal in 6. Cell culture failed in the remaining 7 cases.

DISCUSSION

The results of this large cohort study confirm that pregnant women with a history of recurrent miscarriage associated with antiphospholipid antibodies have a high rate of live births in subsequent pregnancies in which they are treated with low dose aspirin together with heparin. However, a number of successful pregnancies were complicated by gestational hypertension, antepartum haemorrhage, prematurity and low birthweight babies.

The mechanism(s) of adverse pregnancy outcome in women with recurrent miscarriage associated with antiphospholipid antibodies is unclear. A substantial body of evidence from both the histological examination of placentae obtained from women with antiphospholipid antibodies and in vitro studies support a thrombotic basis for pregnancy loss. De Wolf's original description of widespread placental thrombosis and infarction of the placenta obtained from a woman with antiphospholipid antibodies has since been confirmed by others studying both first and second trimester placentae6,7,23. In vitro studies also suggest that fetal loss in women with antiphospholipid antibodies has a thrombotic basis. Peaceman et al.24 reported that the IgG fraction obtained from patients with lupus anticoagulant, when incubated with placental explants from normal human pregnancies, leads to a significant increase in thromboxane synthesis (prostacyclin synthesis being unchanged) when compared with the immunoglobulin fraction obtained from normal controls. Annexin V (also known as placental anticoagulant protein-I) is a 35 kD protein which binds to exteriorised anionic phospho-lipids to render their surfaces non-thrombogenic, by preventing the binding of activated factor X and pro-thrombin. Rand et al.25 have reported that women with antiphospholipid antibodies have a significantly lower distribution of annexin V covering the intervillous surfaces of their placentae compared with both normal controls and antiphospholipid antibodies negative women with a history of miscarriage. This would be expected to render the syncytiotrophoblast surface thrombogenic. In addition, it was reported that the addition of antiphospholipid antibodies decreases annexin V immunostaining of cultured placental villi26.

Despite the considerable amount of data supporting a thrombotic aetiology for pregnancy loss associated with antiphospholipid antibodies, there are several conceptual difficulties with this hypothesis. First, although placental thrombosis is a frequent finding in antiphospholipid antibody positive women who have miscarried, this is neither a universal nor a specific finding7,8. Secondly, for thrombosis of the umbilical circulation to occur, antiphospholipid antibodies must cross the placenta. This does not occur efficiently until the second trimester, after the time when antiphospholipid antibody mediated damage is most likely to have taken place. Finally, overt systemic thrombosis is rare in women with antiphospholipid antibody associated pregnancy complications 1,17. Alternative mechanisms by which antiphospholipid antibodies might mediate fetal loss have consequently been sought.

Evidence from murine models of the antiphospholipid syndrome suggests that antiphospholipid antibodies impair trophoblast function via mechanisms unrelated to thrombosis and have a direct effect on the placenta. It has been reported that antiphospholipid antibodies bind to surface phospholipids on the trophoblast, resulting in direct cellular injury, inhibition of syncytia formation, decreased human chorionic gonadotrophin production and defective invasion of the uterine decidua10,12.

Two recent studies, one randomised17 and one non-randomised16, have reported that treatment with aspirin together with heparin results in a significantly higher live birth rate than that obtained with the use of aspirin alone. Aspirin may improve pregnancy outcome in women with antiphospholipid antibodies by irreversibly blocking the action of cyclo-oxygenase in platelets, thereby inhibiting platelet thromboxane synthesis and preventing thrombosis of the placental vasculature27. Heparin, in addition to its anticoagulant action, may act to reduce fetal loss by binding to antiphospholipid antibodies, thereby protecting the trophoblast phospholipids from attack28, and so promoting successful implantation in early pregnancy.

The unifying cause of the pregnancy complications, hypertensive disease, antepartum haemorrhage and preterm labour, observed in this study is abnormal trophoblast invasion leading to defective implantation. It is, therefore, of paramount importance that a diagnosis of antiphospholipid antibodies is made prior to pregnancy, and treatment started as soon as the pregnancy test is positive.

An important observation is that the incidence of preterm delivery in the study group was four times greater than that in the normal obstetric population (24%vs 6%)29. Although we considered a preterm delivery to be one that occurs before 37 weeks of pregnancy, most mortality and morbidity occurs in babies born before 34 weeks. In this study, the majority of the preterm births (69%) occurred between 34 and 36 weeks. This may account for the exceptionally good infant survival noted in the study. There appears to be a temporal relationship between the cessation of treatment at 34 weeks of gestation and an increase in the number of babies delivered between 34 and 36 weeks of gestation. We are addressing this point by conducting a randomised trial to determine whether continuing aspirin and heparin treatment until delivery decreases the incidence of late pregnancy complications.

The high rate of elective caesarean section in this study reflects the informed choice made by couples who have had a poor obstetric history. Their decision is supported by the 21% incidence of emergency caesarean section.

No woman suffered a thromboembolic episode during pregnancy or in the postpartum period. Women tolerated heparin well and no woman developed either heparin-induced thrombocytopenia or an osteoporotic fracture. The mean decrease in the bone mineral density at the lumbar spine observed (3.4%) is less than the estimated bone loss as a result of breastfeeding for six months (5%)30.

The increased incidence of pregnancy complications and operative deliveries in women with recurrent miscarriage and antiphospholipid antibodies has a significant impact on the provision of maternity and neonatal services. The identification of these pregnancies as being at increased risk necessitates close antenatal surveillance and planned delivery in a unit with specialist obstetric and neonatal intensive care facilities.

Acknowledgement

The authors would like to thank Ms T. McGrath, clinic co-ordinator, for her dedication to the clinic and its patients.

Ancillary