Journal of Thrombosis and Haemostasis, 6: 18681875

DOI: 10.1111/j.1538-7836.2008.03134.x

ORIGINAL ARTICLE

Factor V Leiden is associated with pre-eclampsia but not with

fetal growth restriction: a genetic association study and
meta-analysis
T. DUDDING,* J. HERON, A. THAKKINSTIAN, E. NURK,** J. GOLDING, M. PEMBREY,
S . M . R I N G , J . A T T I A and R . J . S C O T T
*Hunter Genetics, Hunter New England Health Service, NSW; Discipline of Medical Genetics, School of Biomedical Sciences, Faculty of Health,
University of Newcastle and the Hunter Medical Research Insitute, NSW, Australia; ALSPAC, Department of Social Medicine, University of
Bristol, Bristol, UK; Clinical Epidemiology Unit, Mahidol University, Bangkok, Thailand; Centre for Clinical Epidemiology and Biostatistics,
University of Newcastle, Newcastle, NSW, Australia; **Department of Nutrition, Institute of Basic Medical Science, University of Oslo, Oslo,
Norway; ALSPAC, Department of Community Based Medicine, University of Bristol, Bristol; Institute of Child Health, University College
London, London, UK; Division of Medicine, John Hunter Hospital, Newcastle, NSW; and Division of Genetics, Hunter Area Pathology Service,
John Hunter Hospital, Newcastle, NSW, Australia

To cite this article: Dudding T, Heron J, Thakkinstian A, Nurk E, Golding J, Pembrey M, Ring SM, Attia J, Scott RJ. Factor V Leiden is associated with
pre-eclampsia but not with fetal growth restriction: a genetic association study and meta-analysis. J Thromb Haemost 2008; 6: 186875.

Summary. Background: Adverse pregnancy outcomes have

been related to environmental and/or genetic factors. Of interest
are genes associated with the clotting system as any perturbation in the balance of thrombotic and thrombolytic cascades
could aect the placental circulation and hence the viability of
the developing fetus. Several previous reports using relatively
small numbers of cases and controls have suggested that there is
a relationship between poor pregnancy outcomes and two
polymorphisms, one in the factor V gene, the 1691G to A
change (rs6025) located on chromosome 1q23 (factor V Leiden,
FVL), and the other in the prothrombin gene, 20210G to A
change (rs1799963) on chromosome 11p11-q12 (PT). These
results, however, are conicting. Methods: We genotyped 6755
mother/infant pairs from the Avon Longitudinal Study of
Parents and Children (ALSPAC) to determine whether maternal or fetal FVL or PT, either alone or in combination, are
associated with fetal growth restriction (FGR) or pre-eclampsia
(PE). We also added the present results to previous cohort
studies using meta-analysis. Results: Smoking, primiparity and
lower body mass index (BMI) were all associated with FGR,
but neither maternal nor fetal FVL or PT, singly or in
combination, were associated with FGR in the ALSPAC
cohort. Meta-analysis conrmed the lack of association
between maternal FVL and FGR with a pooled odds ratio
(OR) of 1.15 [95% condence interval (CI) 0.951.39]. High
BMI, primiparity, diabetes and chronic hypertension were all
Correspondence: Rodney Scott, Discipline of Medical Genetics,
School of Biomedical Sciences, Faculty of Health, University of
Newcastle and the Hunter Medical Research Institute, NSW, Australia.
E-mail: rodney.scott@newcastle.edu.au
Received 1 April 2008, accepted 6 August 2008

associated with pre-eclampsia. Combining ALSPAC results

with previous studies in a meta-analysis indicated that maternal
FVL is signicantly associated with pre-eclampsia, with a
pooled OR of 1.49 (95% CI 1.131.96). Conclusion: Neither
maternal nor fetal FVL or PT, singly or in combination, are
associated with FGR; this contradicts previous casecontrol
studies and meta-analyses based on these studies. In a metaanalysis of all published cohort studies to date, maternal FVL
appears to increase the risk of pre-eclampsia by almost 50%.
This result is robust, homogeneous and does not appear to be
aected by publication bias.
Keywords: adverse pregnancy outcome, ALSPAC, factor V
Leiden, fetal growth restriction, genetic polymorphisms, preeclampsia, prothrombin, single nucleotide polymorphism, SNP.
Introduction
Adverse pregnancy outcomes including fetal growth restriction
(FGR) and pre-eclampsia (PE) remain a major cause of
maternal and infant morbidity and mortality. A successful
pregnancy is highly dependent on a complex interplay between a
variety of maternal, fetal and placental factors. Given the
importance of establishing and maintaining an adequate
placental circulation, hereditary thrombophilias have been
postulated as a possible cause of placental insufciency. Despite
early reports supporting an association between hereditary
thrombophilias and adverse pregnancy outcomes [26] a number of other studies have yielded conicting results [710]. To
address this discrepancy, three large meta-analyses [1113] were
published to help clarify the contribution of inherited thrombophilias to the risk of FGR. Although two studies [11,12] reported
an association between FGR and factor V Leiden (FVL), their
results were dominated by small, casecontrol studies.
 2008 International Society on Thrombosis and Haemostasis

Thrombophilias and adverse pregnancy outcomes 1869

Several large meta-analyses summarizing the association

between thrombophilia and pre-eclampsia report that FVL is
associated with an increased risk of hypertensive disease of
pregnancy [1315], and especially of severe PE [11,14]. Robertson et al. [13] also reported an association between the
prothrombin 20210G-A polymorphism and PE [odds ratio
(OR) 2.54 95% condence interval (CI) 1.524.23], but these
meta-analyses reported frequent heterogeneity between studies
and may be subject to publication and time-lag bias. Restricting
the literature to population-based studies yields conicting
results, which may be as a result of differences in study design,
study populations or power [1,9,16,18].
As placental circulation consists of both a fetal and maternal
component, it is possible that hypercoagulability not only
within the maternal circulation but also within the fetal placental
circulation may be associated with an increased risk of adverse
pregnancy outcome. A number of small casecontrol studies
and parent triad studies evaluating an association between fetal
FVL and adverse pregnancy outcomes have also produced
conicting results [7,17,1922]. To overcome the shortfalls
observed in the large number of small and possibility underpowered studies, we conducted a large cohort study evaluating
the association between both the maternal and fetal FVL and
prothrombin (PT) genotype and risk of FGR and PE. To
increase the power of detecting an association, data from other
published cohort studies were combined in the meta-analyses.
Methods
Study population and design for the nested case-control study

The Avon Longitudinal Study of Parents and Children

(ALSPAC) [23] recruited 14663 pregnant mothers with an
expected date of delivery between 1st April 1991 and 31st
December 1992 in the dened county of Avon in the UK. One
of the main aims of the ALSPAC cohort was to promote the
study of how genotype differences inuence the reactions of
individuals to the environment. Details of the ALSPAC study
area, eligibility criteria, enrolment, data collection and questionnaire validation are reported elsewhere [23], or can be
found at http://www.alspac.bris.ac.uk. Briey, data were
collected by (i) self-completion questionnaires during pregnancy and post-delivery completed by the mother and her
partner, (ii) review of medical records and (iii) biological
samples from the mother, her partner and child.
Outcomes

Outcomes were as follows:

1 FGR, as defined by 10th centile on gender-specific UK
population charts.
2 PE, defined as a systolic blood pressure of 140 mmHg or a
diastolic of 90 mmHg in addition to 3 g/24 h of proteinuria.
This was assessed independent of a previous history of
hypertension, although previous history was included as a
confounder in the analyses for this outcome.
 2008 International Society on Thrombosis and Haemostasis

The outcome of stillbirth was not analyzed as a result of

insufcient power.
DNA collection

Blood collection, DNA extraction and processing have been

described previously [24]. Maternal DNA samples were
extracted from blood collected in EDTA or heparin from
mothers during their antenatal care and childrens samples
from umbilical cord blood collected in heparin or from venous
samples taken later in life collected in EDTA. DNA is available
for approximately 70% of the cohort children. Ethical
approval for the study was obtained from the ALSPAC Law
and Ethics committee and the local ethics committees, as well
as the Newcastle University Human Research Ethics Committee (HREC). All participants had given consent to participate
in genetic association studies.
Of the initial ALSPAC sample of 14 663 pregnancies, 13 432
pregnancies were single pregnancies and had non-Down
Syndrome children. DNA was available for 10 232 children,
and 9117 mothers; because these were not completely overlapping, this represents 6755 motherchild pairs with DNA.
Genotyping

Genotyping on stored DNA was performed retrospectively by

KBiosciences (Hoddesdon, UK) using KASPar, an allelespecic PCR system developed by the company (http://
www.kbioscience.co.uk).
The variants that were genotyped were the factor V Leiden
1691G-A polymorphism (rs 6025), in F5 (coagulation factor V)
gene on chromosome 1q23 and a common prothrombin
variant 20210G-A (rs1799963) in the F2 [coagulation factor II
(thrombin)] gene on chromosome 11p11-q12.
The genotyping failure rates for each of the groups tested
were:
1 mothers: FVL 3.01% and prothrombin 3.26% and
2 children: FVL 9.58% and prothrombin 9.11%.
This difference in failure rates may be as a result of
differences in sample processing. The blood samples for
children were taken at follow-up visits, when they were older,
and the samples taken in EDTA rather than heparin as for the
mothers. In addition, many of the childrens samples were
frozen unseparated at )20 C for up to 3 weeks before DNA
extraction.
A series of duplicate assays were also performed to
determine the efciency of the genotyping assays. There were
805 duplicate samples of mothers and 226 childrens samples
tested and no differences in the genotyping were observed.
None of the genotypes in any of the groups tested deviated
from the HardyWeinberg equilibrium (P < 0.05).
Statistical analysis

Stata version 8.0 was used for logistic regression analysis to

examine the relationship between a number of predictors and

1870 T. Dudding et al

outcomes. Predictors included: (i) maternal factor V Leiden; (ii)

maternal prothrombin G20210A polymorphism; (iii) fetal
factor V Leiden; (iv) fetal prothrombin G20210A polymorphism; and (v) combined maternal and foetal genotype. Rare
homozygous results for FVL or PT were analyzed in combination with heterozygous results. Exclusion of homozygous
women or infants from the analysis did not change the results.
There were no individuals homozygous for both FVL and PT.
The results are presented as ORs with 95% CIs. Univariate
regression analysis was used to determine possible predictors of
the outcome prior to multivariate analysis. Age (<35, 3539,
40 years), body mass index (BMI) (30, 2529, <25),
smoking (20, 1019, 19, no cigarettes) and parity (primiparous, multiparous) were included in the FGR univariate
analysis; and age, BMI, smoking, parity, diabetes (yes, no) and
chronic hypertension (yes, no) were included in the PE
univariate analysis.
We also analyzed a possible interaction between the
maternal and fetal genotype. To reduce multiple subgroup
analyses, a decision was made to combine FVL and PT
genotypes together as a thrombophilic tendency. The
description either heterozygous includes mothers or infants
who were heterozygous for one or both FVL or PT
polymorphisms. Wild type refers to mothers and infants
who were homozygous for the wild-type FVL and PT
alleles.
Meta-analysis of cohort studies

We also performed a MEDLINE and EMBASE search (up to

February 2007) using the headings: (i) factor V leiden
(textword) and pregnancy, OR pre-eclampsia OR fetal growth
restriction (MeSH headings) and (ii) prothrombin (text word)
and pregnancy, OR pre-eclampsia OR fetal growth restriction

(MeSH headings). The search was limited to human studies

published in English. Inclusion criteria were:
1 cohort design;
2 outcomes clearly defined as fetal growth restriction or preeclampsia;
3 entire cohort, or at least both nested cases and controls,
tested for FVL or prothrombin G20210A mutation; and
4 sufficient data to enable the calculation of an odds ratio.
Data were extracted independently by two of the authors
(Attia and Dudding) on (i) the general characteristics of the
study, e.g. age ranges of mothers, ensuring that denitions of
predictors and outcomes were comparable, etc; (ii) research
questions and methodology and (iii) outcome data. Testing
for heterogeneity was performed using the BreslowDay
method with the P-value threshold for heterogeneity set at
P < 0.1, and using I2. If there was no heterogeneity, a xed
effects model was used to pool data; otherwise a random
effects model was used. Study size (publication) bias was
assessed using Eggers test. Pooling was performed using
StatsDirect (version 2.5.7, StatsDirect Ltd, Cheshire, UK,
http://www.statsdirect.com).
Results
Cohort study

Fetal growth restriction Among 13 601 pregnancies, 10 976

pregnancies where data for fetal growth and at least one of
FVL and PT gene for the mother/child were available. These
subjects were included in the analysis. Characteristics between
included and non-included subjects were similar except that
included mothers were slightly older and smoked less than nonincluded mothers (Table 1); all analyses were adjusted for these
factors. Less than 5% of the cohort was of non-Caucasian

Table 1 Characteristics of mothers included and not included in the associations between genotype and fetal growth restriction (FGR) or pre-eclampsia
(PE)
FGR
Maternal
characteristics
Parity
Primiparous
Multiparous
Smoking/day
20
1019
19
None
BMI
30
2529
<25
Age group
<35
3539
40

PE

Included

Not included

Included

Not included

P-value

4586 (44.4)
5753 (55.7)

1024 (45.4)
1233 (54.6)

0.380

2199 (51.9)
2037 (48.1)

3481 (40.9)
5029 (59.1)

<0.001

298 (2.8)
905 (8.5)
1149 (10.8)
8317 (77.9)

93 (3.9)
275 (11.6)
353 (14.8)
1660 (69.7)

<0.001

136 (77.0)
369 (11.5)
500 (8.4)
3366 (3.1)

262 (2.9)
831 (9.4)
1020 (11.6)
6721 (76.1)

0.313

589 (6.3)
1731 (18.4)
7065 (75.4)

121 (6.5)
327 (17.6)
1410 (75.9

0.664

271 (7.0)
759 (19.7)
2828 (73.3)

444 (5.9)
1319 (17.6)
5735 (76.5)

0.001

9833 (89.6)
1002 (9.1)
141 (1.3)

2419 (92.2)
181 (6.9)
25 (0.9)

<0.001

4028 (90.0)
389 (8.7)
60 (1.3)

8318 (90.1)
807 (8.7)
108 (1.2)

0.695

P-value

Inclusion was based solely on availability of DNA or availability of data from manual record review. BMI, body mass index.
 2008 International Society on Thrombosis and Haemostasis

Thrombophilias and adverse pregnancy outcomes 1871

Table 2 Genotype frequencies of (a) maternal factor V Leiden (FVL) (1691G-A) and fetal FVL (1691G-A) and (b) maternal prothrombin (PT) (20210GA) and fetal PT (20210G-A) in relation to the presence of birth weight <10th centile
(a) Maternal FVL

(b) Maternal PT

Birth weight < 10th centile

Birth weight > 10th centile
Fetal FVL
Birth weight < 10th centile
Birth weight > 10th centile

A/A or A/G
33 (5.6%)
368 (5%)

G/G
554 (94.4%)
6914 (95%)

A/A or A/G
16 (2.7%)
162 (2.2%)

G/G
575 (97.3%)
7089 (97.8%)

A/A or A/G
37 (6.8%)
396 (5.1%)

G/G
508 (93.2%)
7408 (94.9%)

Foetal PT
A/A or A/G
15 (2.7%)
211 (2.7%)

G/G
542 (97.3%)
7655 (97.3%)

Percentages are based on row totals.

ethnicity. Table 2 describes the maternal and fetal gene

frequencies of FVL and PT for the pregnancies where the
infant weighed < 10th centile for gestational age (based on
population growth charts). As can be seen from Table 2, the
gene frequencies in all the subgroups are consistent with
expected population frequencies of 5% for FVL and 2% for
PT. A number of predictors of FGR were explored by
univariate analysis including maternal age, BMI, smoking and
parity. From this analysis, smoking, parity and BMI emerged
as statistically signicant predictors of FGR.
These factors were included in multivariate logistic models in
which maternal and fetal genotypes were added separately. As
seen in Table 3, neither maternal nor fetal prothrombotic
genotypes were associated with FGR. The power of these tests
to detect the effect sizes seen for FVL and PT was only 45.6%
and 44.7%. Smoking, parity and BMI continued to be
associated, indicating independent effects. Given the maternal

Table 3 Odds ratios for maternal factor V Leiden (FVL) and prothrombin G20210A and (FGR) and odds ratio for fetal FVL and prothrombin
G20210A and fetal growth restriction (FGR): multivariate analyses
allowing for all variables in the table
Maternal
Characteristics
FVL
A/A and A/G
G/G
PT
A/A and A/G
G/G
Smoking/day
20
1019
19
None
Parity
Primiparous
Multiparous
BMI
30
2529
<25

OR (95% CI)

Table 4 Maternal and fetal gene effects on fetal growth restriction (FGR):
multivariate analysis
Characteristics
Maternal FVL-PT
Either heterozygous

Fetal
P-value OR (95% CI)

and fetal contribution to placental circulation, we went on to

explore a possible interaction between maternal and fetal
genotypes.
Considering the multiple possible combinations of maternal
and fetal FVL and PT (16 possible combinations) and the
ensuing loss of power with this many degrees of freedom, we
combined FVL and PT into one variable reecting thrombophilia, which was positive if the individual was heterozygous for
FVL or PT or both. This was supported not only on the basis
of power consideration, but also on the lack of any clear
difference in effect size, or indeed effect, between the two genes
and FGR. The term wild type was used to denote the
homozygous normal genotype for both FVL and PT. Table 4
describes the interaction between maternal and fetal thrombophilia dened in this way in a multivariate analysis on FGR.
Compared with the dyads where both had the wild type, when
the mother and the fetus had a thrombophilia allele, the OR

P-value
Either heterozygous
wild type

1.09 (0.71, 1.69)

1

0.693 1.25 (0.83, 1.89)

1

0.286

1.37 (0.79, 2.39)

1

0.266 0.97 (0.55, 1.74)

1

0.930

4.45 (2.86, 6.92) <0.001 3.45 (2.06, 5.80) <0.001

3.56 (2.70, 4.70) <0.001 2.68 (1.98, 3.64) <0.001
2.20 (1.67, 2.88) <0.001 2.16 (1.65, 2.83) <0.001
1
1
2.18 (1.78, 2.67) <0.001 2.57 (2.09, 3.16) <0.001
1
1
0.66 (0.42, 1.04)
0.69 (0.53, 0.91)
1

0.073 0.56 (0.33, 0.94)

0.009 0.79 (0.60, 1.03)
1

0.027
0.084

BMI, body mass index; CI, condence interval; OR, odds ratio, PT,
prothrombin.
 2008 International Society on Thrombosis and Haemostasis

wild type
Smoking/day
20
1019
19
None
Parity
Primiparous
Multiparous
BMI
30
2529
<25

Fetal FVL-PT
Either
heterozygous
wild type
Either
heterozygous
wild type

OR

95% CI
of OR

0.94

0.50, 1.78

0.858

1.92
1.09

1.13, 3.27
0.61, 1.96

0.016
0.774

3.71
3.23
2.05
1

1.20, 6.89
2.25, 4.65
1.46, 2.88

<0.001
<0.001
<0.001

2.53
1

1.96, 3.26

<0.001

0.65
0.70
1

0.37, 1.14
0.50, 0.98

0.131
0.039

P value

The description either heterozygous includes mothers or infants who

were either factor V Leiden (FVL) heterozygous or prothrombin (PT)
heterozygous or both FVL and PT heterozygous. Wild type refers to
mothers and infants who were homozygous wild type at both FVL and
PT alleles.

1872 T. Dudding et al

was 0.94 (0.51.78). When the mother had a thrombophilia

allele and the infant was homozygous wild type, the odds ratio
was 1.92 (95% CI 1.133.27, P = 0.016). When only the infant
had a thrombophilia allele, there appeared to be no effect on
FGR (OR = 1.09, 95% CI 0.611.96). Results were similar
when the denition of birth weight <10th centile was based on
customized growth charts, which incorporates maternal weight,
height, parity and ethnic group (data not shown).
Pre-eclampsia As previous meta-analyses have suggested an
association between thrombophilia and PE, we explored this
possible association. Only 4477/14 273 pregnancies had at least
one of FVL and PT genotypes and sufcient data for assessing
pre-eclampsia; the latter had to be extracted manually by
medical record review and this process is not yet complete for
the entire ALSPAC cohort. Characteristics between those
pregnancies that did/did not have data were similar except that
included mothers were more likely to be primiparous and have
slightly higher BMI (Table 1); all analyses were adjusted for
these factors. Table 5 describes the maternal gene frequencies
for FVL and PT in those who were diagnosed with PE
compared with controls. As a result of small numbers and some
missing data, this analysis was not run on fetal FVL and PT.
On univariate analysis, parity, BMI, diabetes and chronic
hypertension were independent predictors of pre-eclampsia.
Including these variables in a multivariate logistic model along
with maternal genotype showed that heterozygosity in the
mother for FVL or PT was not associated with a risk of PE
with OR of 1.19 (95% CI 0.642.23) and 1.39 (95% CI 0.54
3.58) respectively (Table 5). The power of these tests to detect
the effect sizes seen for FVL and PT was 33.1% and 3.4%

respectively. Parity, BMI, diabetes and chronic hypertension continued to be associated, indicating an independent
effect.
Meta-analysis of cohort studies

Fetal growth restriction In order to investigate whether or not

the lack of association noted above was as a result of
insufcient power (type 2 error), we pooled our results by
meta-analysis with data from other published cohort studies.
The study characteristics and OR for the cohort studies
evaluating the risk of FGR in unselected groups of FVL
positive mothers were available from six studies including ours.
Nurk et al. [16] provided data from the rst pregnancy of the
5874 women in the previously published Hordaland study. The
combined OR for studies where FGR was dened as a birth
weight <10th centile for gestational age were homogeneous
(BreslowDay P = 0.75, I2 = 0%) with a combined OR of
1.15 (95% CI 0.951.39) (Fig. 1). With the sample size in this
meta-analysis, and assuming a power of 80%, we would have
been able to detect an OR of 1.09. There was no evidence of
study size (publication) bias (Egger test P = 0.83).
Only two cohort studies evaluated the effect of fetal FVL on
the risk of FGR. These studies were homogeneous with a
pooled OR of 1.2 (95% CI 0.881.63). There were insufcient
studies to perform a meta-analysis of PT positive mothers or
infants and FGR.
Pre-eclampsia The same six cohort studies evaluated the
effect of maternal FVL on the risk of PE. These studies
were homogeneous (BreslowDay P = 0.93, I2 = 0%)

Table 5 Genotype frequencies of maternal factor V Leiden (FVL) (1691G-A) and prothrombin (PT) (20210G-A) and pre-eclampsia (PE)
Maternal FVL 1691G-A
Pre-eclampsia
No pre-eclampsia

Maternal prothrombin variant 20210G-A

A/A or A/G
17 (7%)
204 (4.8%)

G/G
226 (93%)
4002 (95.2%)

A/A or A/G
5 (2.1%)
85 (2%)

G/G
234 (97.9%)
4091 (98%)

Percentages are based on row totals.

Lindqvist (9)
9/88 cases 261/2392 controls

0.93 (0.41, 1.89)

Murphy (1)
0/9 cases 13/576 controls

2.20 (0.00, 24.55)

Dizon-Townson (17)
10/413 cases 114/4139 controls

0.88 (0.41, 1.69)

Kocher (18)
8/101 cases 13/303 controls

1.92 (0.67, 5.17)

ALSPAC cohort
33/587 cases 368/7282 controls

1.12 (0.75, 1.62)

Nurk (16)
79/887 cases 334/4828 controls

1.21 (0.92, 1.57)

Total pooled odds ratio

(heterogeneity P = 0.7)

1.15 (0.95, 1.39)

0.2

0.5
1
2
5
10
Odds ratio (95% confidence interval)

100

Fig. 1. Pooled odds ratio for the association between fetal growth restriction (FGR) (<10th centile) and maternal factor V Leiden (FVL).
 2008 International Society on Thrombosis and Haemostasis

Thrombophilias and adverse pregnancy outcomes 1873

Lindqvist (9)
5/39 cases 265/2441 controls

1.21 (0.37, 3.14)

Murphy (1)
0/12 cases 16/576 controls

1.36 (0.00, 13.90)

Nurk (16)
22/189 cases 401/5685 controls

1.74 (1.05, 2.75)

Dizon-Townson (17)
5/146 cases 129/4739

1.27 (0.40, 3.11)

Kocher (18)
14/231 31/693

1.38 (0.66, 2.72)

ALSPAC cohort
17/243 204/4206

1.48 (0.83, 2.48)

Total pooled odds ratio

Heterogeneity P = 0.9

1.49 (1.13, 1.96)

0.2

0.5

1
2
5
10
Odds ratio (95% confidence interval)

100

Fig. 2. Pooled odds ratio for the association between pre-eclampsia (PE) and maternal factor V Leiden (FVL).

Table 6 Odds ratio for maternal factor V Leiden (FVL) and prothrombin
(PT) and pre-eclampsia (PE): multivariate analysis total pooled odds ratio
(OR) (heterogeneity P = 0.7)
Characteristics
FVL
A/A and A/G
G/G
PT
A/A and A/G
G/G
BMI
30
2529
<25
Parity
Primiparous
Multiparous
Diabetes
Yes
No
Chronic HT
Yes
No

OR

95% CI of OR

1.19
1

0.64, 2.23

0.586

1.39
1

0.54, 3.58

0.500

1.92
1.12
1

1.23, 3.00
0.77, 1.62

0.004
0.550

2.32
1

1.66, 3.24

<0.001

4.18
1

2.00, 8.73

<0.001

5.38
1

3.78, 7.66

<0.001

P-value

BMI, body mass index; HT, hypertension.

and indicated a statistically signicant increased risk of

pre-eclampsia with a pooled OR of 1.49 (95%CI 1.131.96,
P = 0.003) (Fig. 2). There was no evidence of study size
(publication) bias (Egger test P = 0.20).
Two studies evaluated the effect of maternal PT on the
risk of PE. The studies were homogeneous with a pooled OR
of 0.96 (95% CI 0.481.88). There were insufcient studies to
perform a meta-analysis of fetal FVL or PT and PE
(Table 6).
Discussion
Fetal growth restriction

Overall the results of our ALSPAC cohort study show no

statistically signicant association between maternal FVL,
maternal PT, fetal FVL or fetal PT and birth weight <10th
 2008 International Society on Thrombosis and Haemostasis

centile. The one statistically signicant OR of 1.92 (1.13

3.27) (Table 4) when the mother has a thrombophilia
(heterozygous for fVL and/or PT) and the infant is wild
type at both loci may represent a true association; however,
the result is inconsistent with the other results and is more
likely to represent a chance statistically signicant result
owing to the multiple number of analyses. Furthermore, our
fetal growth restriction (FGR) cohort meta-analysis does not
support this lone result and nds no evidence of an effect of
maternal FVL on FGR. Given the size of the pooled sample,
we had 80% power to detect an OR of 1.09, indicating that
if there is an effect of FVL on FGR that has been missed by
this meta-analysis then it must be quite small clinically. It is
also possible that there is an association between thrombophilia and more severe FGR, e.g. <3rd centile, but the size
of cohort required to test this hypothesis would need to be
extremely large as a result of the small number of infants
with this outcome.
The result of this cohort meta-analysis contradicts the
conclusion of previous large meta-analyses [11,12], which
noted an association between FVL and FGR. However, these
previous meta-analyses reported statistically signicant heterogeneity between studies, possibly related in part to study
designs, i.e. casecontrol versus population cohorts. There are
a number of possible explanations for this discrepancy
between the cohort studies and casecontrol studies, which
include biased recruitment of cases, poor selection of controls
and interaction effects. It is possible that although cases with
FGR or PE are enriched for the FVL or PT mutations
compared with controls, this effect is manifest because of
other unmeasured genotypes or clinical factors [6,22]; this
comes about because cases are selected based on having had
an adverse event. In other words, FVL or PT are only part of
the high-risk prole but because the rest of the factors are
not measured, the entire risk is assigned to the FVL or PT
mutation. Those with FVL or PT but no other unmeasured
risk factors do not have events and are unrepresented in the
casecontrol studies. Conclusions about PT gene effects are
limited because of a lack of studies and there is room for
further investigation in this area. There is also a dearth of

1874 T. Dudding et al

studies addressing the effect of fetal thrombophilia on

pregnancy outcomes.
The results of our ALSPAC cohort study conrm previous
reports that smoking, parity and BMI are independent
predictors of FGR [25,26].
Pre-eclampsia

Overall, the results of previous meta-analyses, based largely

on casecontrol studies, show an association between PE and
FVL, particularly in relation to severe PE [11,1315];
however, a previous meta-analysis limited to cohort studies,
reecting an unselected group of women, showed no
association between FVL and PE with an OR of 1.1 (95%
CI 0.42) [11]. Although the results of our ALSPAC cohort
study show no statistically signicant association between
maternal FVL, maternal PT and PE, increasing the power by
combining our study with other cohort studies by metaanalysis conrmed a positive association between maternal
FVL and pre-eclampsia with an OR of 1.49 (95% CI
1.131.96, P = 0.003). There is insufcient data to make any
rm conclusions about maternal PT or fetal thrombophilias
and PE.
The results of our ALSPAC cohort study also conrm
previous reports that BMI, parity, diabetes and chronic
hypertension are independent predictors of PE [27].

Conclusions
We nd that previous estimates of FVL increasing the risk of
FGR were driven largely by small casecontrol studies and are
not supported by our cohort study or our meta-analysis of
cohort studies. Conversely, the results from our study combined with other cohort studies by meta-analysis does support
an association between maternal FVL and PE, and this is
consistent between casecontrol and cohort studies. The effects
of fetal thrombophilia on these outcomes are yet to be claried.
Addendum
While this manuscript was being submitted and reviewed,
another large cohort study addressing the association between
FVL and poor pregnancy outcomes was published online
[Clark P et al. The GOAL study: a prospective examination of
the impact of factor V Leiden and ABO(H) blood groups on
haemorrhagic and thrombotic pregnancy outcomes. Br J
Haematol (published online 19 November 2007)]. Consistent
with our ndings, the authors found no association between
FVL and fetal growth restriction (dened as <5th centile) with
an OR = 0.91 (95% CI 0.322.58) but did nd a higher point
estimate for association with PE, although this did not reach
signicance (OR = 1.83, 95% CI 0.516.13).
Acknowledgements

Clinical implications

The results of this study and the meta-analysis of cohort studies

are important in a number of contexts:
1) Women with FVL or PT mutations identied through
cascade testing. These results support the evidence that
maternal FVL is an independent risk factor for PE. The
magnitude of this risk may vary depending upon the presence
of other environmental or genetic risk factors, which means
that a pregnancy management plan should be formulated on
an individual basis combining history of other known risk
factors and evaluation for multiple thrombophilia. Whether or
not FVL carrier women should just be monitored more closely
or treated during pregnancy remains to be determined. There
appears to be no increased risk of FGR for FVL carriers. There
is still insufcient data specically relating to the risk of PT
carriers.
2) Screening women who have experienced an adverse
pregnancy outcome. There are no studies evaluating the
subsequent risk of PE in a second pregnancy for women
carrying an inherited thrombophilia compared with the
subsequent risk of PE in a second pregnancy of women
without a thrombophilia. However, these results support the
screening of women with a previous history of PE for FVL.
Whether those who test positive should simply be monitored or
treated during their subsequent pregnancy remains to be
determined. There is insufcient data to recommend any course
of action regarding PT genotyping.

We are extremely grateful to all the families who took part in

this study, the midwives for their help in recruiting them, and
the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research
scientists, volunteers, managers, receptionists and nurses. The
UK Medical Research Council, the Wellcome Trust and the
University of Bristol provide core support for ALSPAC. This
publication is the work of the authors and J. Heron and
A. Thakkinstian will serve as guarantors for the contents of this
paper. This research was specically funded by the National
Health and Medical Research Council of Australia, grant
number 858908.
Disclosure of Conflict of Interests
The authors state that they have no conict of interest.
References
1 Murphy RP, Donoghue C, Nallen RJ, DMello M, Regan C, Whitehead AS, Fitzgerald DJ. Prospective evaluation of the risk conferred by
factor V Leiden and thermolabile methylenetetrahydrofolate reductase
polymorphisms in pregnancy. Arterioscler Thromb Vasc Biol 2000; 20:
26670.
2 Dizon-Townson DS, Nelson LM, Easton K, Ward K. The factor V
Leiden mutation may predispose women to severe preeclampsia. Am J
Obstet Gynecol 1996; 1: 9025.
3 Grandone E, Margaglione M, Colaizzo D, Cappucci G, Paladini D,
Martinelli P, Montanaro S, Pavone G, Di Minno G. Factor VLeiden,

 2008 International Society on Thrombosis and Haemostasis

Thrombophilias and adverse pregnancy outcomes 1875

10

11

12

13

14
15

C > T MTHFR polymorphism and genetic susceptibility to preeclampsia. Thromb Haemost 1997; 77: 10524.
Martinelli P, Grandone E, Colaizzo D, Paladini D, Scianname N,
Margaglione M, Di Minno G. Familial thrombophilia and the
occurrence of fetal growth restriction. Haematologica 2001; 86: 42831.
Agorastos T, Karavida A, Lambropoulos A, Constantinidis T,
Tzitzimikas S, Chrisa S, Saravelos H, Vavilis D, Kotsis A, Bontis J.
Factor V Leiden and prothrombin G20210A mutations in pregnancies
with adverse outcome. J Matern Fetal Neonatal Med 2002; 12: 26773.
Kupferminc MJ, Eldor A, Steinman N, Many A, Bar-Am A, Jaa A,
Fait G, Lessing JB. Increased frequency of genetic thrombophilia in
women with complications of pregnancy. N Engl J Med 1999; 340: 9
13.
Livingston J, Barton J, Park V, Haddad B, Phillips O, Sibai B.
Maternal and fetal inherited thrombophilias are not related to the
development of severe preeclampsia. Am J Obstet Gynecol 2001; 185:
1537.
Currie L, Peek M, McNiven M, Prosser I, Mansour J, Ridgway J. Is
there an increased maternal-infant prevalence of Factor V Leiden in
association with severe pre-eclampsia? BJOG 2002; 109: 1916.
Lindqvist PG, Svensson PJ, Marsaal K, Grennert L, Luterkort M,
Dahlback B. Activated protein C resistance (FV:Q506) and pregnancy.
Thromb Haemost 1999; 81: 5327.
Infante-Rivard C, Rivard GE, Yotov WV, Genin E, Guiguet M,
Weinberg C, Gauthier R, Feoli-Fonseca JC. Absence of association of
thrombophilia polymorphisms with intrauterine growth restriction. N
Engl J Med 2002; 347: 1925.
Dudding TE, Attia J. The association between adverse pregnancy
outcomes and maternal factor V Leiden genotype: a meta-analysis.
Thromb Haemost 2004; 91: 70011.
Howley H, Walker M, MA R. A systematic review of the association
between factor V leiden or prothrombin gene variant and intrauterine
growth restriction. Am J Obstet Gynecol 2005; 192: 694708.
Robertson L, Wu O, Langhorne P, Twaddle S, Clark P, Lowe GD,
Walker ID, Greaves M, Brenkel I, Regan L, Greer IA. Thrombophilia
in pregnancy: a systematic review. Br J Haematol 2006; 132: 17196.
Lin J, August P. Genetic thrombophilias and pre-eclampsia: a metaanalysis. Obstet Gynecol 2005; 105: 18292.
Kosmas IP, Tatsioni A, Ioannidis JP. Association of Leiden mutation
in factor V gene with hypertension in pregnancy and pre-eclampsia: a
meta-analysis. J Hypertens 2003; 21: 12218.

 2008 International Society on Thrombosis and Haemostasis

16 Nurk E, Tell GS, Refsum H, Ueland PM, Vollset SE. Factor V Leiden,
pregnancy complications and adverse outcomes: the Hordaland
Homocysteine Study. QJM 2006; 99: 28998.
17 Dizon-Townson D, Miller C, Sibai B, Spong CY, Thom E, Wendel G
Jr, Wenstrom K, Samuels P, Cotroneo MA, Moawad A, Sorokin Y,
Meis P, Miodovnik M, OSullivan MJ, Conway D, Wapner RJ, Gabbe
SG. The relationship of the factor V Leiden mutation and pregnancy
outcomes for mother and fetus. Obstet Gynecol 2005; 106: 51724.
18 Kocher O, Cirovic C, Malynn E, Rowland CM, Bare LA, Young BA,
Henslee JG, Laer TG, Hu JB, Kruskall MS, Wong G, Bauer KA.
Obstetric complications in patients with hereditary thrombophilia
identied using the LCx microparticle enzyme immunoassay: a controlled study of 5,000 patients. Am J Clin Pathol 2007; 127: 6875.
19 Gibson CS, MacLennan AH, Janssen NG, Kist WJ, Hague WM,
Haan EA, Goldwater PN, Priest K, Dekker GA. Associations between
fetal inherited thrombophilia and adverse pregnancy outcomes. Am J
Obstet Gynecol 2006; 194: 947 e110.
20 Dekker JW, Lind J, Bloemenkamp KW, Quint WG, Kuijpers JC, van
Doorn LJ, de Groot CJ. Inherited risk of thrombosis of the fetus and
intrauterine fetal death. Eur J Obstet Gynecol Reprod Biol 2004; 117:
458.
21 Vefring H, Lie RT, Degard R, Mansoor MA, Nilsen ST. Maternal
and fetal variants of genetic thrombophilias and the risk of preeclampsia. Epidemiology 2004; 15: 31722.
22 von Kries R, Junker R, Oberle D, Kosch A, Nowak-Gottl U. Foetal
growth restriction in children with prothrombotic risk factors. Thromb
Haemost 2001; 86: 10126.
23 Golding J, Pembrey M, Jones R. ALSPAC The Avon Longitudinal
Study of Parents and Children 1. Study methodology. Paediatr Perinat
Epidemiol 2001; 15: 7487.
24 Jones RW, Ring S, Tyeld L, Hamvas R, Simmons H, Pembrey M,
Golding J. A new human genetic resource: a DNA bank established as
part of the Avon Longitudinal Study of Pregnancy and Childhood
(ALSPAC). Eur J Hum Genet 2000; 8: 65360.
25 Mayer M, Tonascia J. Maternal smoking, pregnancy complications,
and perinatal mortality. Am J Obstet Gynecol 1977; 128: 494502.
26 Hammoud A, Bujold E, Sorokin Y, Schild C, Krapp M, Baumann P.
Smoking in pregnancy revisited: ndings from a large populationbased study. Am J Obstet Gynecol 2005; 192: 185662.
27 Duckitt K, Harrington D. Risk factors for pre-eclampsia at antental
booking: systemic review of controlled studies. BMJ 2005; 330: 565.