Thrombophilia and stillbirth: possible connection by intrauterine growth restriction


Dr E. Shalev, Department of Obstetric and Gynecology, Ha'Emek Medical Center, Afula, 18101, Israel.


Objective  To define the association between thrombophilia and unexplained stillbirth.

Design  A case–control study.

Setting  Obstetric department in a university affiliated hospital (Ha'Emek Medical Center, Afula).

Population  A total of 53 women who delivered stillborns between March 1998 and June 2002 and 59 women with unremarkable obstetric history who delivered at the same period.

Methods  Presence of genetic and acquired markers of thrombophilia was investigated.

Main outcome measure  Presence or absence of thrombophilia.

Results  Thrombophilia was found in 34% of the women who delivered stillborns and in 20% of the 59 women with normal pregnancies (non-significant). However, significantly higher prevalence of thrombophilia (73%) was found in women who delivered small for gestational age stillborns compared with women who delivered normal birthweight stillborns (73%vs 18.4%, P < 0.0001).

Conclusions  There is no association between thrombophilia and stillbirth, overall. However, there is a clear association between thrombophilia and stillbirth of extremely growth restricted infants.


Intrauterine fetal death is a serious complication of pregnancy, which in many cases remains idiopathic in spite of extensive and thorough investigation. The stillbirth rate varies from 4.9 to 10.4 per 1000 births and its incidence has not changed in recent decades.

Previous studies reported a very high prevalence of thrombophilia in women with adverse fetal outcome.1–3 Several studies were specifically dedicated to fetal loss and thrombophilia.4–6 Although an association was found between fetal loss and thrombophilia, no attempt was made to distinguish between unexplained stillbirth and stillbirth linked with clinical risk factors associated with thrombophilia.

The present study was designed to define the association between thrombophilia and stillbirth. We therefore studied all women admitted to our department with stillbirth during the study period.


The rate of thrombophilia in the normal population in our region has been demonstrated to be 22%.6 In order to detect a prevalence of 50% thrombophilia in the study group, for a power of 0.80 and an alpha of 0.05, it is necessary to recruit 52 members each, in the study group and in the control group.

We prospectively studied all women who had singleton pregnancies complicated by stillbirth admitted to our department after 24 weeks of gestation from March 1998 to June 2002. The exclusion criteria were the presence of fetal congenital malformations or chromosomal abnormalities, hydrops fetalis, fetal–maternal haemorrhage, cord accident, uncontrolled maternal diabetes and pregnancy duration over 42 weeks. We also studied 59 normal pregnant women who delivered during the same period and had at least one normal pregnancy with no history of thromboembolic complications. The women in both groups were Israeli, either Jewish or Arab. The control group was chosen at random and matched the study group for maternal age and ethnicity. The study was approved by the Hospital Research Committee and the Ministry of Health. An informed consent was obtained from each woman.

The women were recruited after delivery and blood was drawn for DNA analysis. Assays for factor V Leiden, the prothrombin G20210A mutation, methylenetetrahydrofolate reductase (MTHFR) mutation, protein S, protein C, antithrombin III levels and for the presence of anticardiolipin antibodies and lupus anticoagulant were performed.

The centiles of birthweight were calculated by using the Brenner curves.7 Gestational age was confirmed in all cases by first trimester ultrasound. Small for gestational age was defined as birthweight below the 5th centile. Bad obstetric history was defined as an obstetric history with a minimum of three fetal losses during the first trimester or two fetal losses during the second trimester or one fetal loss during the third trimester.

The plasma protein C activity was measured by functional synthetic chromogenic substrate (Stachrom, France), and quantitatively by enzyme-linked immunosorbent assay (Asseachrom Stag, France). Protein S activity was measured functionally with synthetic chromogenic substrate (Il-Test, USA) and quantity of free and total protein S by enzyme-linked immunosorbent assay (Asseachrom Stag). Antithrombin III was measured by chromogenic substrate with factor Xa reagent (Il-test). Levels of IgG and IgM anticardiolipin antibodies were determined with enzyme-linked immunosorbent assay (Autozyme, Cambridge Life Science, England). Lupus anticoagulant assays were performed using the DRVVT kit (Screen and Confirm, Gradipore, North Ryde, Australia), KCT kit (Kaoclot, Gradipore), FSL/FS (Date Behring, Deerfield, Illinois, USA). The criteria for the diagnosis of deficiencies of anti-coagulation factors were activity levels below lower limit of normal, as determined by our laboratory normal values.

A genetic analysis was performed to determine the presence of mutations of G1691A factor V Leiden, the G20210A prothrombin gene mutation and the C677T MTHFR mutation. Genomic DNA was extracted from peripheral blood according to Miller and Polskey.8 The presence or absence of factor V Leiden (G1691A), MTHFR (C677T) and prothrombin gene (G20210A) mutations were determined by using GamidaGen-pronto Thrombo risk kit. For confirmation of the genotypes, an allele-specific restriction enzyme analysis was performed for factor V Leiden,9 MTHFR (C677T)10 and prothrombin (G20210A).11 The results of the allele-specific restriction enzyme analysis were in 100% concordance with the results of genotype analysis as determined by Gamidagen-pronto kit.

The SPSS statistical package was used for all statistical analysis. The associations between the categorical variables were assessed by χ2 or by Fisher's exact test for small groups. For non-categorical variables, we used the Student's t test. P≤ 0.05 was considered as statistically significant.


Seventy-six women who delivered stillborns after 24 weeks of gestation were admitted to our department during the study period. Excluded from the study were 13 cases with congenital malformations, seven cases with cord accident, two cases with uncontrolled maternal diabetes mellitus and one case of post term (43 weeks of gestation). The characteristics of the women with intrauterine fetal death and those with normal pregnancies are shown in Table 1.

Table 1.  Characteristics of the study and control groups. Values in parentheses represent standard deviation.
 Control group (n= 59)Study group (n= 53)
  • GA = gestational age.

  • *

    P < 0.01.

Maternal age (years)28.9 (5)30.3 (7)
Arab (n)3032
Jewish (n)2921
GA at delivery (weeks)39.8 (1.7)31.3 (5.2)*
Birthweight (g)3468 (452)1256 (612)*

The prevalence of thrombophilia in the study and the control groups is presented in Table 2. There were no significant differences between the different types of acquired or genetic thrombophilia comparing between the two groups. Thrombophilia was found in 34% of women who delivered stillborns and in 20% of women with normal pregnancies (non-significant).

Table 2.  A comparison between thrombophilia in women with stillbirth and women with normal pregnancy. Values are presented as n (%).
 Stillbirth (n= 53)Normal pregnancy (n= 59)
Antiphospholipid antibodies3 (5.7)2 (3.4)
Antithrombin deficiency01 (1.7)
Protein C deficiency2 (3.8)0
Protein S deficiency1 (1.9)1 (1.7)
Factor V Leiden9 (17)5 (8.5)
MTFHR (homozygote)6 (11.4)4 (6.8)
Prothrombin mutation02 (3.4)
Combined thrombophilia3 (5.7)3 (5.1)
Total no. of women with thrombophilia18 (34.2)12 (20)

Dividing the study group into two categories, we found that women with stillbirth and infants with birthweight below the 5th centile had a significantly higher prevalence of thrombophilia compared with women with stillbirth and normal birthweight infants (Table 3). In 7 out of 15 cases of small for gestational age infants, intrauterine growth restriction was diagnosed during pregnancy. In the remaining eight women, there was no documentation of fetal weight estimation prior to the admission to our department. However, they all felt fetal movements during the last week prior to their admission. Therefore, it is unlikely that the small birthweight was due to an extended period from fetal demise to delivery.

Table 3.  A comparison between thrombophilia in women with stillbirth and small for gestational age infants, women with stillbirth and normal birth weight infants and women with normal pregnancy. Values are presented as n (%).
 Stillbirth with SGA (n= 15)Stillbirth with AGA (n= 38)Normal pregnancy (n= 59)
  • SGA = small for gestational age; AGA = appropriate for gestational age.

  • *

    P < 0.05.

  • **

    P < 0.0001.

Antiphospholipid antibodies2 (13)1 (2.6)2 (3.4)
Antithrombin deficiency001 (1.7)
Protein C deficiency2 (13)*00
Protein S deficiency1 (1.9)01 (1.7)
Factor V Leiden5 (33)*4 (10.5)5 (8.5)
MTFHR (homozygote)4 (27)*2 (5.3)4 (6.8)
Prothrombin mutation002 (3.4)
Combined thrombophilia3 (20)*03 (5.1)
Total no. of women with thrombophilia11 (73)**7 (18.4)12 (20)

We further examined the prevalence of thrombophilia in the study group for complications of pregnancy. Two out of seven women with pre-eclampsia (29%), two out of eight women with bad obstetric history (25%) and three out of seven women with placental abruption (43%) had thrombophilia. These numbers did not reach statistical significance, although there were fewer cases of these complications than intrauterine growth restriction.


There is a vast amount of data in the literature associating thrombophilia and adverse pregnancy outcome.1–3 The complications reported are first and second trimester fetal loss,6 intrauterine fetal death,4 intrauterine growth restriction,1 pre-eclampsia12–16 and placental abruption.17 These studies have assumed a thrombotic-induced pathogenesis in the vascular placental bed adversely affecting normal uterine blood flow in pregnancy. Thrombotic lesions have been observed on pathologic examination of the placenta in women with severe pregnancy complications.18–21 Arias et al.18 documented placental thrombotic lesions in 11 women with severe pregnancy complications. Factor V Leiden was detected in 7 of 11 (54%) and protein S deficiency in 3 of 11 (23%) of patients. Dizon-Townson et al.19 evaluated placental infarction in fetuses carrying factor V Leiden mutation after miscarriage. They found a tenfold increase in the fetal carrier frequency in 10 of 24 (42%) placentas with more than 10% infarction compared with 7 of 372 (1.9%) placentas with less than 10% infarction. Moreover, Vern et al.,20 investigating thromboembolic events of the fetal side of the placenta, found a significantly higher rate of factor V Leiden and prothrombin G20210A mutations in placentas with thrombotic events compared with normal placentas.

Assuming that adverse pregnancy outcome in women with thrombophilia is associated with placental vascular damage, we hypothesised that any stillbirth that is not associated with placental vascular damage might not be associated with thrombophilia. However, in order to ascertain this hypothesis, it is necessary not only to define very carefully the group of unexplained stillbirth, excluding congenital malformations, cord accident and other complications associated with placental vascular damage, but also to provide histological proof for this assumption.

Typical pregnancy complications associated with thrombophilia and stillbirths are placental abruption and intrauterine growth retardation. Previous studies did not define clearly whether these other clinical risk factors for thrombophilia were present. In view of our results, this may be the reason why stillbirth was found to be associated with thrombophilia in these studies. Preston et al.5 found an increased fetal loss rate in women with heritable thrombophilia. According to their study, the effect of thrombophilia was especially pronounced for stillbirth. However, no information about the causes of stillbirth was provided and there was no evidence that they were unexplained. Martinelli et al.4 found 16% mutation in coagulation factors in late fetal losses compared with 6% in the control group. However, data were not provided concerning the prevalence of intrauterine growth restriction or placental abruption and therefore fetal loss may not be truly unexplained. Many et al.22 have also recently reported association between stillbirth and thrombophilia. However, as in the previous studies, they did not exclude from the study group other complications such as intrauterine growth restriction.

In our study, we were able to demonstrate that stillbirths were associated with thrombophilia only when the birthweight was below the 5th centile. The association between intrauterine growth restriction and thrombophilia has been discussed in previous studies. Kupferminc et al.1 found a 61% rate of thrombophilia in growth-restricted fetuses. In the study of Verspyck et al.,23 the prevalence of anticardiolipin and antinuclear antibodies, but not inherited thrombophilias, was higher in growth-restricted fetuses. Infant-Rivard et al.24 did not show any association between thrombophilia and intrauterine growth restriction. The different results obtained in different studies are probably due to different study groups of growth-restricted fetuses. Our subgroup of extremely small for gestational age stillborns probably represents the most severe cases of intrauterine growth restriction. Therefore, thrombophilia might be associated only with severe cases of intrauterine growth restriction. There is a clear association between thrombophilia and stillbirths of extremely small for gestational infants, suspected of suffering from intrauterine growth restriction.

An important finding was the absence of association of thrombophilia with unexplained stillbirth of normal birthweight infants. Therefore, a thrombophilia work-up is not indicated for investigating stillbirth unless there are other additional complications known to be associated with thrombophilia. In addition, while there are abundant data in the literature associating thrombophilia and pregnancy complications, a better definition of this association should be made for each complication.

Accepted 1 March 2004