Seminar

Stillbirth
Gordon C S Smith, Ruth C Fretts

In the UK, about one in 200 infants is stillborn, and rates of stillbirth have recently slightly increased. This recent rise Lancet 2007; 370: 1715–25
might reflect increasing frequency of some important maternal risk factors for stillbirth, including nulliparity, Department of Obstetrics and
advanced age, and obesity. Most stillbirths are related to placental dysfunction, which in many women is evident from Gynaecology, Cambridge
University, Cambridge, UK
the first half of pregnancy and is associated with fetal growth restriction. There is no effective screening test that has
(Prof G C S Smith MD);
clearly shown a reduction in stillbirth rates in the general population. However, assessments of novel screening Harvard Vanguard Medical
methods have generally failed to distinguish between effective identification of high-risk women and successful Associates, Wellesley, MA, USA
intervention for such women. Future research into stillbirth will probably focus on understanding the pathophysiology (R Fretts MD); and Harvard
Medical School, Boston, MA,
of impaired placentation to establish screening tests for stillbirth, and assessment of interventions to prevent stillbirth
USA (R C Fretts)
in women who screen positive.
Correspondence to:
Prof Gordon C S Smith,
Introduction reporting systems, but a strategy that includes pregnancy Department of Obstetrics and
Stillbirth accounts for 60% of all perinatal deaths and losses at earlier gestations improves the reliability of Gynaecology, Cambridge
University, Box 223, The Rosie
75% of all potentially preventable losses (defined as reporting stillbirth rates at later gestations. Obtaining
Hospital, Robinson Way,
perinatal death of a normally formed infant weighing reliable estimates of the number of stillbirths in Cambridge CB2 2QQ, UK
1000 g or more).1 Stillbirth is ten times more common developing countries is especially difficult since most gcss2@cam.ac.uk
than sudden infant death syndrome.1 Moreover, although births take place in the home and, in some remote areas,
rates of sudden infant death syndrome have greatly fallen data are completely absent.7 Even in developed countries,
over the past 10–15 years,2 there has been a recent slight there is an inconsistent approach to inclusion of
rise in the rate of stillbirth in England and Wales, and the therapeutic terminations for fetal anomalies that are
causes are unknown.3 The aim of this Seminar is to prenatally diagnosed.8 Therefore, although perinatal
provide an overview of the causes of stillbirth and to death rates (early neonatal death and stillbirth rates
summarise present practice and future strategies to combined) are useful indicators of access to, and quality
reduce the number of stillbirths. We focus on stillbirths of, antenatal care, these rates should be compared with
in the developed world where fetal death occurs before caution.
the onset of labour.
Classification
Definition Stillbirths can be subclassified according to the gestational
WHO defines stillbirth as a fetal death late in pregnancy, age at birth, typically into early stillbirths (20–28 weeks’
and individual countries define the gestational age at gestation) and late stillbirths (after 28 weeks). Although
which a miscarriage becomes a stillbirth. The perinatal this division is somewhat arbitrary, this stratification
period is defined as 22 weeks or more of gestation allows for fairly reliable international comparison of late
(154 days) or, if the gestational age is unknown, it includes losses, and allows stillbirths to be divided into those that
infants with a birthweight of 500 g or more and ends are difficult to prevent (ie, early losses) and those that are
7 days after birth. For international comparison, stillbirths potentially preventable (ie, late losses). Stillbirths are also
are defined as infants born showing no signs of life in subclassified by whether death occurred before or after
the perinatal period.4,5 Although this definition is useful, the onset of labour—termed antepartum and intrapartum,
many developed countries register stillbirths at earlier respectively. However, the primary method for
weeks of gestation, some as early as 16 weeks.6 classification of stillbirth is according to the presumed
Underestimates of losses at early gestations occur in all cause or associated obstetric disorders. There are,
however, more than 30 reported systems for classification
of perinatal deaths.9 Early classifications included only a
Search strategy and selection criteria few subtypes—congenital malformations, immaturity,
We searched the PubMed database, using the word “stillbirth” asphyxia, and others.10,11 Recent systems have attempted
in all fields and the Medical Subject Headings term “fetal to obtain more information including on aberrations of
death”. We assessed all records using the title and abstract fetal growth, pathological changes of the placenta, and
over the past 10 years and obtained full copies of relevant maternal disorders.9,12 However, there is debate about
articles. We searched the titles in the reference lists of whether hierarchical systems should be used and whether
literature reviews published in the past 5 years. We searched conditions such as growth restriction and hypertension
the Cochrane Library using the terms “stillbirth” and are causes of or risk factors for stillbirth. Nevertheless, a
“perinatal mortality”. Finally, we used the Science Citation systematic approach to classification of stillbirths is a
Index to identify articles that subsequently cited important crucial step in design of prevention strategies. Panel 1
original research studies. shows a classification system based on obstetric criteria
with 27 categories, focused around eight major groups.1

www.thelancet.com Vol 370 November 17, 2007 1715

 Seminar

Panel 1: Modified version of the Wigglesworth system for Fetal abnormality

10%
classification of perinatal deaths by obstetric causes Unexplained SGA
23%
Congenital anomaly* Pre-eclampsia
7%
1 CNS
2 Cardiovascular system
3 Renal
4 Alimentary (excluding diaphragmatic hernia)
5 Chromosomal Haemorrhage
6 Biochemical 15%
7 Other (including musculoskeletal)
Isoimmunisation
Mechanical
8 Rhesus incompatibility 1%
9 Non-rhesus incompatibility Maternal
6%
Toxaemia† Unexplained AGA Miscellaneous
10 Severe—diastolic blood pressure of ≥110 mm Hg on 36% 2%
two or more occasions and >20 weeks with proteinuria
of ≥300 mg/24 h Figure 1: Causes of stillbirth with modified version of Wigglesworth
classification for all singleton births in Scotland, 1992–2001
11 Other toxaemia
Data from 2635 antepartum stillbirths, from a total of 563 719 births
Antepartum haemorrhage† (rate 4·7 per 1000). Over the same period, there were 320 intrapartum stillbirths
in 561 084 singletons alive at the onset of labour. 75% of these stillbirths were
12 Abruptio placentae anoxic, 17% were classified as caused by congenital abnormality, and the
13 Placenta praevia remaining 8% had diverse other causes. SGA=small for gestational age (smallest
14 Other (with evidence of recurrent bleeding after the first decile of birthweight for sex and week of gestation). AGA=appropriate for
gestational age (rest of population). The data sources are as described in
trimester)
reference 13. Smith GCS, unpublished data.
Mechanical‡
15 Breech In general, the study of specific causes of stillbirth has
16 Cord prolapse been slowed by the scarcity of uniform protocols for
17 Other mechanical assessment and classification of stillbirths and falling
autopsy rates. In most cases, death certificates are filled
Maternal disorder out before the results of postnatal investigations are
18 Maternal trauma available. The most useful information about specific
19 Essential hypertension causes of stillbirth comes from hospitals or regions that
20 Diabetes systematically review and classify these deaths. One such
21 Abdominal operations in pregnancy study, using data from McGill University (Canada), showed
22 Other (including maternal infection) differences in the causes of stillbirth in relation to
Miscellaneous gestational age. Between 24 weeks and 27 weeks of
23 (Specify) gestation, the most common causes were infection (19%),
abruption (14%), or fetal anomalies (14%).14 However, the
Unexplained contribution of infection to stillbirth rates can be technically
24 Birthweight <2500 g and <37 weeks difficult to define. There are some pathogens that are
26 Birthweight ≥2500 g and <37 weeks clearly causally associated with stillbirth, such as parvovirus
25 Birthweight <2500 g and ≥37 weeks B19, cytomegalovirus, and toxoplasma. However, there are
27 Birthweight ≥2500 g and ≥37 weeks others that might be associated with an increased risk of
*Any structural or genetic defect incompatible with life or potentially treatable but
stillbirth, but strong evidence of a causal relation is absent
causing death. †In deaths with antepartum haemorrhage (APH) secondary to toxaemia, (eg, colonisation with Ureaplasma urealyticum, Mycoplasma
toxaemia is classified first and antepartum haemorrhage second. ‡Any death from hominis, and group B streptococci).15,16
uterine rupture, cord compression, birth trauma, or intrapartum asphyxia that is
associated with disproportion, malpresentation, or breech delivery of babies weighing After 28 weeks of gestation, the most frequent types of
1000 g or more. Deaths from anoxia or cerebral trauma should be classified as stillbirth were those that were unexplained, including
unexplained (codes 24–27) if there is no evidence of difficulty in labour. Antepartum
deaths associated with cord entanglement in the absence of strong circumstantial
those associated with growth restriction, and placental
evidence that cord compression caused death (eg, fetal death soon after external abruption. A fetal death that is unexplained by fetal,
version) should be classified as unexplained (codes 24–27). placental, maternal, or obstetric factors is the most
common, representing between 25% and 60% of all fetal
Figure 1 shows the proportion of antepartum stillbirths deaths.17–20 Variation in the proportion defined as
in Scotland attributed to these groups, between 1992 unexplained generally reflects whether the classification
and 2001. system allows risk factors to be included as causes—in

1716 www.thelancet.com Vol 370 November 17, 2007

 Seminar

particular whether unexplained losses where the

birthweight was small for gestational age are defined as Panel 2: Commonly reported maternal risk factors for and
being due to growth restriction or as being unexplained. causes of stillbirth in developing and developed countries
A definitive classification system will probably continue by ranking of estimated attributable risk or importance
to be elusive until the pathophysiology underlying the Developing countries
large number of cases without a clear direct cause is • Obstructed or prolonged labour and associated asphyxia,
elucidated. infection, and birth injury (low availability of
caesarean-section)
Stillbirth in the developing and developed worlds • Congenitally acquired infections, especially syphilis and
A detailed discussion of stillbirth in the developing world gram-negative infections
is beyond the scope this Seminar, and has been reviewed • Hypertensive disease, especially poor management of
elsewhere.21 However, the scale and nature of the problem pre-eclampsia and eclampsia
will be compared with that in developed countries. The • Poor nutritional status
greatest risk factor for stillbirth is being born in the • Previous stillbirth
developing world.22,23 Within developed countries, the • Congenital anomalies
stillbirth rate is estimated to be between 4·2 and • Malaria
6·8 per 1000 births, whereas in the developing world this • Sickle-cell disease
rate is between 20 and 32 per 1000 births (table 1).23
Where good data exist, rates of stillbirth and neonatal Developed countries
death in the developing world are roughly equivalent. As • Congenital or karyotypic anomalies
care improves, there is generally a greater reduction in • Growth restriction or placental thrombosis
neonatal deaths, increasing the proportion of perinatal • Medical diseases such as diabetes, systemic lupus
deaths attributed to stillbirth.5 The causes of stillbirth erythematosus, renal disease, thyroid disorders,
vary in developed and developing countries (panel 2).21 thrombophilias, cholestasis of pregnancy
An estimated 27% of stillbirths worldwide occur during • Hypertensive disease/pre-eclampsia
labour at term or near term.22 In developed countries, • Congenitally acquired infections such as Group B
fetal death during labour is rare—rates are less than one streptococcus and parvovirus B19
per 1000 births.25 However, in many areas of the • Smoking
developing world intrapartum deaths occur as frequently • Multiple gestation
as one per 100 births.24 Hence, most of these deaths Reproduced from reference 21 with permission from author and publisher.
happen in the developing world and could be prevented
with adequate obstetric care. Furthermore, common
causes of antepartum losses in the developing world, that make delivery of basic obstetric care difficult. These
such as syphilis and malaria, are also largely preventable include, but are not limited to, geography, financial
by screening and treatment for pregnant women.16,26 resources available to governments, political power to
These observations suggest that stillbirth prevention in effect change, availability of facilities and trained
the developing world could be medically fairly simple personnel, and cultural factors. This conclusion is
and that obstacles to prevention largely relate to factors supported by the findings of a cluster randomised
controlled trial undertaken in a rural area of Pakistan
Stillbirth rate per Number of (Larkana in the province of Sindh).27 Subdistricts were
1000 deliveries (95% CI) stillbirths randomly assigned to intervention or control groups. The
World 23·9 (18·8–30·5) 3 219 428 intervention consisted of 3 days of training for traditional
Developed countries 5·3 (4·2–6·8) 57 865 birth attendants in conduct of delivery and provision of
Developing countries 25·5 (20·1–32·5) 3 161 563 three antenatal visits. Outreach clinics were provided by
North Africa 18·6 (14·1–24·7) 66 785 local obstetricians because of an absence of traditional
Sub-Saharan Africa 32·2 (25·4–40·9) 889 697
birth attendants. In the intervention subdistricts, the
Latin America/Caribbean 13·2 (10·4-16·7) 153 162
stillbirth rate was 50 per 1000 births, compared with
East Asia 23·2 (18·3–29·5) 483 436
71 per 1000 in the control districts (adjusted odds ratio
0·69, 95% CI 0·57–0·83).
South Asia 31·9 (25·1–40·7) 1 286 231
Southeast Asia 12·7 (10·0–16·0) 144 681
West Asia 18·9 (14·3–24·9) 94 810
Maternal characteristics and risk of stillbirth
In developed countries, the most prevalent risk factors
Eurasia 12·2 (9·5–15·5) 39 236
for stillbirth are nulliparity,28 advanced maternal age, and
Oceania 15·8 (12·4–20·1) 3524
obesity (table 2). From a public-health perspective, obesity
Reproduced from reference 23 with permission from Elsevier. and smoking are the two most prevalent modifiable risk
factors for adverse pregnancy outcome.29–32 Social factors
Table 1: Rates of stillbirth in different regions of the world
such as educational attainment, racial group, and access

www.thelancet.com Vol 370 November 17, 2007 1717

 Seminar

Prevalence Odds Estimated stillbirth rate

ratio* (per 1000)
All pregnancies 1·0 All pregnancies 6–7
Nulliparity 40% 1·2–1·4 Hypertensive disorders
Smoking >10 cigarettes per day 10–20% 1·7–3·0 Chronic hypertension 5–25
Obesity (before pregnancy) Superimposed pre-eclampsia 52
BMI 25·0–29·9 kg/m2 21% 1·9–2·7 PIH/mild pre-eclampsia 9
BMI >30 kg/m2 20% 2·1–2·8 Severe pre-eclampsia 21
Low educational attainment (<12 years vs ≥12 years) 30% 1·6– Eclampsia 18–48
2·0 HELLP syndrome 51
Previous growth-restricted infant (<10%) 6·7% 2·0– Diabetes mellitus
4·6
Gestational diabetes 5–10
Previous stillbirth 0·5–1·0% 1·4–3·2
Type 1 diabetes 6–10
Multiple gestation (reference singleton) 2–3·5%
Type 2 diabetes 35
Twins 2·7% 1·0–
2·8 Systemic lupus erythematosus 40–150

Triplets 0·14% 2·8–3·7 Chronic renal disease

Advanced maternal age (reference <35 years) Mild renal insufficiency 15

35–39 years 15–18% 1·8–2·2 Moderate and severe renal insufficiency 32–200

≥40 years 2% 1·8–3·3 Thyroid disorders

Black (reference white) 15% 2·0–2·2 Stable treated hyperthyroidism 0–36

Uncontrolled thyrotoxicosis 100–156
BMI=body-mass index. *Odds ratio of the factor present compared with the risk Subclinical hypothyroidism 0–15
factor absent. Modified from reference 14 with permission from Elsevier.
Overt hypothyroidism 15–125
Table 2: Systematic review of epidemiological associations with stillbirth Cholestasis of pregnancy 12–30

PIH=pregnancy induced hypertension. HELLP=haemolysis, elevated liver enzyme

to and quality of care are all greatly associated with risk of levels, and a low platelet count. Modified from reference 45 with permission from
stillbirth. In the USA, black women are four times more Elsevier.
likely to have no prenatal care than are white women.
Table 3: Medical disorders associated with stillbirth risk
The combination of no prenatal care and black race puts
these women at a seven-fold higher risk of stillbirth than
white women who received prenatal care.33,34 Even when gestations.42,43 A strategy of antepartum testing late in
studies are confined to women who have received care pregnancy has the potential to decrease late unexplained
beginning in the first trimester, black women still have a stillbirth in older women but also predicts increased
greater than three-fold risk of perinatal death compared induction and caesarean section rates.44
with white women.33,34 The excess of stillbirth was Hypertension and diabetes are two of the most common
attributed to higher rates of diabetes, hypertension, medical disorders to complicate pregnancy (affecting
placental abruption, and premature rupture of 7–10% and 3–5% of women, respectively).14,32,45 Population-
membranes.33–35 An increased stillbirth rate of based studies showed a two-fold to four-fold risk of
non-majority women compared to white women is a stillbirth in women with diabetes.46,47 However, reports
consistent finding even in countries that have universal from referral centres suggest that with optimum
access to medical care.36,37 management, including preconception care and close
Advanced maternal age is associated with an increased medical supervision, the risk of perinatal death is only
risk of stillbirth in both nulliparous and multiparous marginally raised above that of the general population.42
women.38 Historically, a substantial proportion of perinatal Table 3 lists other important medical conditions associated
deaths seen in older women was related to lethal with an increased risk of stillbirth.45
congenital and chromosomal anomalies.39,40 However, the The relation between inherited abnormalities of blood
introduction of population-based screening for chromo- clotting and stillbirth is poorly understood and different
somal abnormalities and the availability of elective studies have produced conflicting results. The results of
abortion has contributed to reduced rates of this type of a meta-analysis48 of small studies suggested that the
perinatal death.41 In developed countries, the increased presence of a thrombophilia increased the risk of stillbirth
perinatal mortality for older women is largely related to overall (odds ratio 3·6; 95% CI 1·4–9·4). Table 4
non-anomalous losses and losses related to multiple summarises the results of a comprehensive meta-analysis49
gestations.20,40 Large-scale studies clearly show an of the risk of stillbirth and other obstetric complications.
increased risk of unexplained stillbirth late in pregnancy, There is no distinctive placental lesion associated with
even after controlling for risk factors such as hypertension, thrombophilic defects. However, mechanistically, these
diabetes, placenta praevia, previous abortion, and multiple defects are thought to be associated with adverse outcome

1718 www.thelancet.com Vol 370 November 17, 2007

 Seminar

through effects on the maternal or fetal vasculature, or study unlikely to be a chance finding.54 Studies from
both.50 The associations between prothrombotic other countries, with data sources of variable quality,
mutations and stillbirth risk are the first clear example of have been inconsistent.56–58 The strengths of the Scottish
genetic predisposition towards stillbirth. The possibility dataset are that the population is fairly racially and
of genetic predisposition is supported by the tendency economically homogeneous, there is universal free
for stillbirth recurrence, although clearly other factors access to health care, and the country obtains complete,
could also explain this effect. The genetic epidemiology quality assured, and detailed information on maternal
of stillbirth will probably be a major area of future characteristics, pregnancy outcome, and the cause of
research. However, the conduct of adequate studies is perinatal death. Further studies from other high quality
potentially difficult, since the rarity of the outcome means databases will be needed to establish whether the
that large sample sizes are needed. Moreover, as discussed association is consistent.
above, stillbirth is the endpoint of diverse mechanisms,
and informative analyses will need to use very well Multiple gestations
defined phenotypes. In cardiovascular medicine, many Over the past two decades, rates of twin pregnancies have
studies that address obvious candidate genes for common more than doubled and higher order multiples have
conditions of well understood pathophysiology might be increased by six-fold to 12-fold.59–61 The stillbirth rate for
needed to establish the presence or absence of an multiples is four-fold higher than it is for singletons
association.51 (19·6 per 1000 vs 4·7 per 1000, respectively) with all types
The fact that women with a previous stillbirth are at of death more common for multiples than for singletons.62
increased risk of stillbirth in future pregnancies is well These higher rates are because of complications specific
known.52 Moreover, women with previous complicated to multiple pregnancy (such as twin to twin transfusion
pregnancies that resulted in a livebirth have a raised syndrome) and increased risks of complications common
risk of future stillbirth,53 both explained and to singletons and multiples, especially fetal abnormalities
unexplained.54 A large-scale study of more than and growth restriction. The determinants of increased
100 000 second births in Scotland from 1992 to 1998 perinatal mortality in twins have been extensively
showed an association between delivery by caesarean reviewed elsewhere.63 Multiple gestations are a substantial
section in a first pregnancy and the risk of stillbirth in contributor to overall perinatal mortality rates, and
the second.55 The association was with unexplained reduction in the proportion of births that are multiples
stillbirth, in particular those associated with growth represents an important area for prevention of
restriction. The association was also evident when stillbirths.59–61 The number of multiple births has risen
studies were confined to women whose previous because of an increased use of assisted reproductive
caesarean section was done at term and after more than technologies and a growing proportion of older mothers.61
10 h of labour, making previous caesarean delivery Higher order multiples are associated with even greater
unlikely to be merely a marker for women with rates of perinatal death,61 and many are attributable to
pre-existing medical complications. A follow-up study assisted reproductive technologies. An international
from Scotland has confirmed that the same association strategy of lowering the in-vitro fertilisation transfer rate
is present for births from 1999 to 2001, making the first to two embryos could substantially reduce the number of
perinatal deaths associated with higher order multiple
Stillbirth Pre-eclampsia Abruption IUGR pregnancies.61
Factor V Leiden + + + –
heterozygote Trends in stillbirth
PT heterozygote + + ++ + Rates of stillbirth fell greatly throughout the developed
MTHFR homozygote 0 0 – 0 world in the second half of the 20th century.64 A
Protein C deficiency – – – – longitudinal study65 of a single centre in Canada, where
Protein S deficiency ++ – – –
detailed information was available on the cause of
Anticardiolipin + + – –
perinatal death over 40 years, showed that the greatest
antibodies reductions in stillbirth took place when strategies were
Lupus anticoagulant – – – ++* developed to intervene in specific causes of fetal demise.
For example, there was a 95% reduction in stillbirths
–=Insufficient information. No studies or non-significant association, but
because of rhesus isoimmunisation after introduction of
upper limit of 95% CI >2. 0=Weak or no association. Upper limit of 95% CI <2.
+=Moderate association. Point estimate of odds ratio between 2 and 5 and rhesus immune prophylaxis and much the same
95% CI excludes 1. ++=Strong association. Point estimate of odds ratio >5 and reduction in deaths caused by intrapartum anoxia, which
95% CI excludes 1. IUGR=intrauterine growth restriction. PT=prothrombin coincided with developments in fetal monitoring and the
G20210A. MTHFR=methylenetetrahydrofolate reductase C677T. *p=0·05,
lower limit of 95% CI 0·96, point estimate 18·6. more liberal use of caesarean section. In northeast
England, there has been a 50% reduction in perinatal
Table 4: Acquired and inherited thrombophilia and the risk of stillbirth deaths due to congenital abnormality between 1982–90
and comparison with other adverse outcomes of pregnancy
and 1991–2000,62 which is assumed to result from the

www.thelancet.com Vol 370 November 17, 2007 1719

 Seminar

introduction of population-based screening for have focused on growth restriction as a proxy of stillbirth.
chromosomal and non-chromosomal congenital abnor- Before pregnancy, the uterus is a high resistance
mality. However, an analysis of data from Sweden found circulation. During the first half of pregnancy, trophoblast
that the progressive reduction in overall stillbirth rate invades the maternal spiral arteries, reducing resistance
stopped in the early 1980s.66 Moreover, data from England, to blood flow in the uterine circulation.73 Clinically this
Wales, and Northern Ireland have shown a significant process can be assessed by Doppler flow velocimetry of
increase in stillbirth rates from 2001.67 Data to directly the uterine arteries. A high-resistance pattern of flow at
establish the cause of this increase are scarce, but it could the end of the second trimester of pregnancy is associated
be due to improved reporting of data. However, a rising with an increased risk of growth restriction and stillbirth,
incidence would also be consistent with the known as well as other perinatal complications.74 It is much more
increase in prevalence of important risk factors, such as strongly associated with the risk of stillbirth at preterm
advanced maternal age, nulliparity, obesity, and multiple gestations (figure 2).75 High-resistance flow on the fetal
gestations.68–70 side of the placenta is also associated with an increased
risk of stillbirth.76 Detailed study of placental ultrastructure
Stillbirth and the placenta in fetuses with high resistance patterns of umbilical artery
The pathophysiology of stillbirths caused by congenital Doppler flow velocimetry has shown that this is associated
abnormality and infection will depend on the specific with maldevelopment of the villous tree.77
condition or organism, respectively. Many of the other The mechanisms that underlie impaired placental
broad categories of cause of death, including perfusion remain unclear. However, in some cases, the
pre-eclampsia, abruption, and unexplained stillbirth, are determining factors are probably related to placental
thought to be related to placental function. For abruption, dysfunction originating in very early pregnancy.78 In the
the role of placentation is self evident and for first 10 weeks after conception, both growth of the fetus
pre-eclampsia, there is much evidence to link the disease and maternal circulating concentrations of the placentally
to placentation.71 Similarly, there is evidence that attributes derived regulator of the insulin-like growth factor system,
many cases of unexplained stillbirth to the placenta. pregnancy associated plasma protein-A (PAPP-A), are
About half of unexplained stillbirths have a birthweight associated with the risk of delivering an infant with low
less than the tenth percentile corrected for gestational age birthweight.79,80 Women with PAPP-A concentrations in
and parental characteristics.72 Hence, stillbirth associated the lowest 5% in the first 10 weeks after conception had a
with intrauterine growth restriction, but without any 40–50-fold risk of stillbirth attributable to growth
other obvious direct cause, is one of the major types of restriction or abruption.81 Pathological analysis of the
stillbirth. Poor fetal growth, without other environmental placenta in otherwise unexplained stillbirths related to
causes, is assumed to indicate poor function of the growth restriction is consistent with chronic placental
placenta. Whether poor fetal growth is simply a marker of dysfunction, which is associated with focal lesions seen
placental dysfunction or whether it is causally associated on macroscopic examination.82
with the mechanism of death is unclear. The risk of unexplained stillbirth in the absence of
Perfusion of the placenta from both the maternal and growth restriction was not related to concentrations of
fetal side has been studied in detail, although most studies PAPP-A in early preganancy.81 However, studies of
placental pathological changes from such cases are also
suggestive of a placental origin to this type of loss.
10 Histopathological examination of placentae from a series
≥33 weeks
<33 weeks of such cases showed changes suggestive of acute
Adjusted likelihood ratio for stillbirth

8 placental dysfunction, specifically “severely reduced

vascularisation of the chorionic villi and lack of
6 syncytiocapillary membranes” and this feature was
associated with a 70-fold risk of stillbirth.82 The biological
4
mechanisms underlying both acute and chronic placental
dysfunction are unclear, but might include determinants
of placental function such as genomic imprinting83 or
2
immune interaction.84 Assessment of the detailed
published work on placental pathological changes and
0 stillbirth is impeded by a general absence of standard
0 0·5 1 1·5 2 2·5
definitions and nomenclature between studies.
Mean uterine artery pulsatility index

Clinical tests of stillbirth risk

Figure 2: Adjusted likelihood ratios for stillbirth, not caused by congenital
abnormality, before 33 weeks and at 33 weeks and after, in relation to
Understanding the pathophysiology of impaired
uterine artery doppler mean pulsatility index at 22–24 weeks’ gestation placentation and fetal growth restriction has led to the
Data from 30 519 unselected women in seven London hospitals.75 identification of tests that are associated with stillbirth

1720 www.thelancet.com Vol 370 November 17, 2007

 Seminar

risk. These tests include circulating concentrations of high-quality, randomised controlled trial. We are unaware
placentally derived proteins in the mother’s blood of any other adequately investigated medical therapies
(PAPP-A,85,86 and α-fetoprotein [AFP]87), Doppler flow shown to reduce the risk of stillbirth.
velocimetry of the uterine74 and umbilical76 arteries, and Grant and colleagues94 have assessed the use of kick
ultrasonic assessment of the appearance of the placenta88,89 charts to reduce antepartum stillbirth, but they noted no
(calcification and visible lesions). None of these tests is in difference in the stillbirth rates when the intervention
routine clinical use for assessment of stillbirth risk in and control groups were compared. Nonetheless, the
unselected populations. However, some are done for stillbirth rate decreased from 4·0 per 1000 to 2·8 per 1000
other purposes—eg, the biochemical tests are part of in both groups, which is probably because of the
population-based screening for Down’s syndrome. There Hawthorne effect. Other aspects of the published work
is clearly the potential to use the information that a suggest that fetal activity could be clinically important,
specific test result confers an increased risk of stillbirth and further research is needed to delineate the role of
to inform intervention. For example, it has been maternal assessment of fetal activity.95 Many biochemical
suggested that women with raised serum concentrations and biophysical tests of fetal wellbeing have been assessed
of AFP or human chorionic gonadotropin (hCG) in their as a means of modifying the risk of stillbirth and these
second trimester should have close surveillance of their are summarised elsewhere.96 However, simply doing a
pregnancies, such as growth scans every 2–4 weeks.90 test cannot directly affect the risk of stillbirth. Tests of
However, there is no direct evidence that this approach is fetal wellbeing can change the risk of stillbirth by
beneficial in an unselected population. Moreover, the informing decisions about the timing of delivery to
nature of the association with stillbirth risk is usually prevent fetal death. However, delivery of the fetus incurs
imprecisely known. The association between increased the risk of maternal or neonatal morbidity or mortality.
AFP concentration and stillbirth in nulliparous women Therefore, assessment of these methods includes the
has proved confined to preterm losses.13 Without this effect of interventions on total perinatal mortality—ie,
information, interventions such as routine late pregnancy the sum of stillbirths and neonatal deaths.
growth scans or induction on achieving term gestation Results from a meta analysis97 of randomised controlled
might be considered, but, in view of the gestational age trials shows that the use of umbilical artery Doppler flow
dependence of the association, these interventions would velocimetry may reduce overall perinatal mortality in
not be expected to be effective. high-risk pregnancies. However, only a trend towards a
reduction in perinatal mortality is reported, and the
Prevention of stillbirth analysis is also consistent with no effect on mortality
Prevention of intrapartum stillbirth is a cornerstone of (findings from a previous meta-analysis had shown a
the management of labour and delivery and has been significant reduction—significance was lost when a
extensively reviewed.91 Understanding the pathophysi- dubious trial was excluded). Meta-analyses of methods of
ology and aetiological factors for some causes of fetal monitoring do not suggest any methods of fetal
antepartum stillbirth has led to assessment of several assessment that reduce the risk of stillbirth when used
medical treatments, but none is in routine practice. In for screening in an unselected population. Some trials
view of the association between thrombophilia and the seem to show possible beneficial effects, such as
risk of stillbirth, strategies could include use of low assessment of placental maturity in the third trimester,88
molecular weight heparin or administration of high but this has not been confirmed (or refuted) by any
doses of folic acid (used to return homocysteine further trials.
concentrations in women with the methylenetetrahydro- Many methods of fetal assessment have been investi-
folate reductase mutation, C677T, to normal). However, gated in unselected populations. However, interpretation
no high quality data exist on the effects of these of the negative results is not straightforward, and the
interventions, and the present recommendations for meta-analysis of umbilical artery Doppler in low-risk
pregnant women in the second half of pregnancy with a pregnancies is a good example of the difficulties of
thrombophilia are that anticoagulant treatment should interpretation.98 First, the trials in this meta-analysis were
be for prevention of thromboembolic disease only.92 In designed without reliable information about how the test
view of the association between fetal hypoxia and performed as a predictor of stillbirth in a population of
stillbirth, some studies have assessed supplemental low-risk women. The adequate design of an interventional
maternal oxygen therapy as a means of reducing perinatal trial needs knowledge of how well the test can identify
death in women with a growth-restricted infant. The women at increased risk. In the case of stillbirth, this
results from a meta-analysis of three studies with includes both the discriminative power of the test and
94 women reported a 50% reduction in perinatal the gestational age dependence.99
mortality.93 However, only one of the studies was blinded. The second challenge in interpretation of these data is
This intervention is not in routine use, would be the failure to distinguish between the two major com-
impractical in many settings, and widespread application ponents of successful screening—namely, effective
would need confirmation of this finding in a large-scale, detection of women at increased risk and effective inter-

www.thelancet.com Vol 370 November 17, 2007 1721

 Seminar

randomised controlled trials that directly support routine

Odds ratio for unexplained stillbirth*
delivery. However, these interventions might be justified
1 2 3 4 5 on the basis of epidemiological evidence of an increased
Fetal deaths per 1000 per week 0·4 0·8 1·2 1·5 1·9 risk of fetal death47,52 and by the very low risk of neonatal
Fetal deaths averted per week† 1·2 2·4 3·5 4·7 5·9 death associated with delivery at term.102 Indeed, even in
Tests per pregnancy 3·4 3·4 3·3 3·3 3·3 unselected pregnancies, the lowest risk of perinatal death
Tests per fetal death averted 2862 1418 950 711 569 at term is associated with delivery at 38–39 weeks.102
Inductions per fetal death averted 233 116 78 58 47 However, routine induction of all women would be highly
Caesarean deliveries per fetal death averted 44 22 15 11 9 invasive and would need very large numbers of elective
deliveries to prevent each loss.
*Assumes base-case test characteristics (70% sensitivity, 90% specificity). †Fetal deaths averted per 1000 pregnancies
compared with no testing. ‡Outcomes from week 37 to week 41 of gestation. Reproduced from reference 44 with
The lower risk of neonatal death at term does, however,
permission from author and publisher. logically lead to a focus on interventions at term, since
these approaches have less potential to cause harm than
Table 5: Predicted effect of structured programme of antepartum fetal assessment at term on stillbirth
rate and rates of obstetric intervention‡
do those undertaken preterm. A detailed model that has
assessed the probable effect of routine induction of
women on the basis of a test of stillbirth risk has been
vention in high-risk women. A trial of an effective described (table 5). This model shows that losses might
prenatal screening method might yield a negative result be prevented by a structured programme of fetal
because the intervention is ineffective at preventing the assessment.44 The effectiveness of such a programme is
outcome in high-risk women. Only one of the trials raised with an increasing background risk of stillbirth.
included in the Cochrane meta-analysis of umbilical However, this analysis assumes a test with 70% specificity
artery Doppler had a protocol for the treatment of women and 90% sensitivity for stillbirth (hence a positive
who screened positive.98 Therefore, whether these trials likelihood ratio of 7). There are no data to suggest that
yielded a negative result because of failure of the any of the present tests of fetal wellbeing at term have
screening method or failure of the intervention is this amount of predictive value. Early elective delivery is
unknown. Umbilical artery Doppler is used as an done primarily on the basis of risk factors in a woman’s
example. The same shortcomings are widespread in previous obstetric or medical history. Perhaps the most
studies that have assessed other novel methods for promising approach to population-based screening is to
population-based screening for stillbirth, such as routine focus on development of predictors of stillbirth at term.
ultrasound in late pregnancy.100 Future work on If an effective screening method can be developed for
population-based screening for stillbirth should be this event, elective delivery at 37–38 weeks would be a
preceded by high-quality, non-interventional prospective simple intervention that should prevent placentally
cohort studies characterising the screening properties of related losses but which would carry a low risk of
new methods of risk assessment in an unselected neonatal death.
population. Such an approach should be feasible, as was
achieved for the assessment of novel methods of Management of stillbirth
screening for risk of Down’s syndrome101— a disorder Clinical management of a stillbirth is dealt with in detail
that is less common than stillbirth. After identification elsewhere,104 and only selected aspects will be discussed.
of effective screening methods, candidate interventions After intrauterine fetal death, induction of labour within
could be assessed in randomised controlled trials in 24 h is usual practice. Delay of induction has theoretical
women who screen positive. risks of disseminated intravascular coagulation and
An alternative approach to assessment of fetal wellbeing infection, but these risks are small in the absence of
is to schedule delivery of women who are considered at further obvious complications, and some parents want to
high risk of stillbirth, irrespective of the results of fetal delay induction. Administration of mifepristone before
assessment. The primary example of this approach is oxytocics is widespread in the UK, although the evidence
routine induction of labour post-term. The risk of is largely by extrapolation from its use in termination of
antepartum stillbirth increases from one per 2000 women pregnancy rather than direct evidence relating to
per week at 37 weeks, to one in 500 at 42 weeks, and one non-therapeutic losses.105,106 After delivery, the infant
in 200 by 43 weeks.102 Several trials have compared routine should be carefully inspected and findings documented
elective delivery beyond term with fetal assessment, and in detail. Parents should be encouraged to hold their
results from a meta-analysis of these trials showed that infant, although there is some evidence that this practice
routine delivery reduced perinatal mortality.103 Other could be associated with increased adverse psychological
clinical situations in which routine scheduled delivery at and social sequelae in some women with very preterm
term might be considered include women with a previous stillbirths.107 Consent should be sought for autopsy. The
stillbirth and those with pregestational diabetes mellitus. procedure and the potential usefulness of the information
In these women, because of the difficulty in doing should be explained in detail, and consent should be
adequately powered studies, there are no data from explicit in relation to any retention of tissue.

1722 www.thelancet.com Vol 370 November 17, 2007

 Seminar

Other postmortem investigations should be 17 Yudkin PL, Wood L, Redman CW. Risk of unexplained stillbirth at
different gestational ages. Lancet 1987; 329: 1192–94.
recommended, including karyotyping, gross and
18 Cnattingius S, Haglund B, Kramer MS. Differences in late fetal
histopathological examination of the placenta and death rates in association with determinants of small for
membranes, external examination of the infant by a gestational age fetuses: population based cohort study. BMJ 1998;
pathologist when autopsy is declined, radiograph of the 316: 1483–87.
19 Huang DY, Usher RH, Kramer MS, Yang H, Morin L, Fretts RC.
infant and, if available, MRI examination. The use of MRI Determinants of unexplained antepartum fetal deaths.
remains to be fully ascertained, and although it is unlikely Obstet Gynecol 2000; 95: 215–21.
to replace full autopsy, it can provide useful information.108 20 Froen JF, Arnestad M, Frey K, Vege A, Saugstad OD,
Stray-Pedersen B. Risk factors for sudden intrauterine
Generally, maternal blood is obtained to test for direct unexplained death: epidemiologic characteristics of singleton
causes (such as congenital infection, fetomaternal cases in Oslo, Norway, 1986–1995. Am J Obstet Gynecol 2001;
transfusion, and isoimmunisation) and risk factors 184: 694–702.
21 McClure EM, Nalubamba-Phiri M, Goldenberg RL. Stillbirth in
(thrombophilia screen and glycosylated haemoglobin). developing countries. Int J Gynaecol Obstet 2006; 94: 82–90.
Counselling should be offered. Finally, it is usual practice 22 Say L, Donner A, Gulmezoglu AM, Taljaard M, Piaggio G. The
in the UK for a consultant to see the woman and her prevalence of stillbirths: a systematic review. Reprod Health 2006; 3: 1.
partner (if appropriate) 6 weeks after the loss to discuss 23 Stanton C, Lawn JE, Rahman H, Wilczynska-Ketende K, Hill K.
Stillbirth rates: delivering estimates in 190 countries. Lancet 2006;
their questions, to review the results of investigations, to 367: 1487–94.
discuss management of a future pregnancy (if planned), 24 Lawn J, Shibuya K, Stein C. No cry at birth: global estimates of
and to arrange further referral if needed. intrapartum stillbirths and intrapartum-related neonatal deaths.
Bull World Health Organ 2005; 83: 409–17.
Conflict of interest statement 25 Smith GCS, Pell JP, Cameron AD, Dobbie R. Risk of perinatal
We declare that we have no conflict of interest. death associated with labor after previous cesarean delivery in
References uncomplicated term pregnancies. JAMA 2002; 287: 2684–90.
1 Information and Statistics Division NHS Scotland. Scottish 26 Adam T, Lim SS, Mehta S, et al. Cost effectiveness analysis of
perinatal and infant mortality report 2000. Edinburgh: ISD Scotland strategies for maternal and neonatal health in developing countries.
Publications, 2001. BMJ 2005; 331: 1107.
2 Smith GCS, Wood AM, Pell JP, White IR, Crossley JA, Dobbie R. 27 Jokhio AH, Winter HR, Cheng KK. An intervention involving
Second-trimester maternal serum levels of alpha-fetoprotein and traditional birth attendants and perinatal and maternal mortality in
the subsequent risk of sudden infant death syndrome. N Engl J Med Pakistan. N Engl J Med 2005; 352: 2091–99.
2004; 351: 978–86. 28 Raymond EG, Cnattingius S, Kiely JL. Effects of maternal age,
3 Fleming K. Confidential enquiry into maternal and child health. parity, and smoking on the risk of stillbirth. BJOG 1994; 101:
Stillbirth, neonatal and post neonatal mortality, 2000–2002, 301–06.
England, Wales and Northern Ireland. London: CEMACH, 2004. 29 Cnattingius S, Bergstrom R, Lipworth L, Kramer MS. Prepregnancy
4 World Health Organization. Definitions and indicators in family weight and the risk of adverse pregnancy outcomes [see comments].
planning maternal & child health and reproductive health. Geneva: N Engl J Med 1998; 338: 147–52.
WHO Press, 2001. 30 Cnattingius S, Lambe M. Trends in smoking and overweight during
5 World Health Organization. Neonatal and perinatal mortality: pregnancy: prevalence, risks of pregnancy complications, and
country, regional and global estimates. Geneva: WHO Press, 2006. adverse pregnancy outcomes. Semin Perinatol 2002; 26: 286–95.
6 Kowaleski J. State definitions and reporting requirements for 31 Stephansson O, Dickman PW, Johansson A, Cnattingius S.
livebirths, fetal deaths and induced terminations of pregnancy (1997 Maternal weight, pregnancy weight gain, and the risk of
revision). Hyattsville: National Center for Health Statistics, 1997. antepartum stillbirth 3. Am J Obstet Gynecol 2001; 184: 463–69.
7 Setel PW, Sankoh O, Rao C, et al. Sample registration of vital events 32 Mokdad AH, Ford ES, Bowman BA, et al. Prevalence of obesity,
with verbal autopsy: a renewed commitment to measuring and diabetes, and obesity-related health risk factors, 2001. JAMA 2003;
monitoring vital statistics. Bull World Health Organ 2005; 83: 611–17. 289: 76–79.
8 Sachs BP, Fretts RC, Gardner R, Hellerstein S, Wampler NS, 33 Healy AJ, Malone FD, Sullivan LM, et al. Early access to prenatal
Wise PH. The impact of extreme prematurity and congenital care: implications for racial disparity in perinatal mortality.
anomalies on the interpretation of international comparisons of Obstet Gynecol 2006; 107: 625–31.
infant mortality. Obstet Gynecol 1995; 85: 941–46. 34 Vintzileos AM, Ananth CV, Smulian JC, Scorza WE, Knuppel RA.
9 Korteweg FJ, Gordijn SJ, Timmer A, et al. The Tulip classification of Prenatal care and black-white fetal death disparity in the United
perinatal mortality: introduction and multidisciplinary inter-rater States: heterogeneity by high-risk conditions. Obstet Gynecol 2002;
agreement. BJOG 2006; 113: 393–401. 99: 483–89.
10 Wigglesworth JS. Monitoring perinatal mortality: a 35 Kallan JE. Rates of fetal death by maternal race, ethnicity, and
pathophysiological approach. Lancet 1980; 316: 684–86. nativity: New Jersey, 1991–1998. JAMA 2001; 285: 2978–79.
11 Chan A, King JF, Flenady V, Haslam RH, Tudehope DI. 36 Raleigh VS, Botting B, Balarajan R. Perinatal and postneonatal
Classification of perinatal deaths: development of the Australian and mortality in England and Wales among immigrants from the
New Zealand classifications. J Paediatr Child Health 2004; 40: 340–47. Indian subcontinent. Indian J Pediatr 1990; 57: 551–62.
12 Gardosi J, Kady SM, McGeown P, Francis A, Tonks A. Classification 37 Parsons L, Duley L, Alberman E. Socio-economic and ethnic
of stillbirth by relevant condition at death (ReCoDe): population factors in stillbirth and neonatal mortality in the NE Thames
based cohort study. BMJ 2005; 331: 1113–17. Regional Health Authority (NETRHA) 1983. BJOG 1990; 97: 237–44.
13 Smith GCS, Shah I, White IR, Pell JP, Crossley JA, Dobbie R. 38 Reddy UM, Ko CW, Willinger M. Maternal age and the risk of
Maternal and biochemical predictors of antepartum stillbirth stillbirth throughout pregnancy in the United States.
among nulliparous women in relation to gestational age of fetal Am J Obstet Gynecol 2006; 195: 764–70.
death. BJOG 2007; 114: 705–14. 39 Hook EB. Rates of chromosome abnormalities at different maternal
14 Fretts RC. Etiology and prevention of stillbirth. Am J Obstet Gynecol ages. Obstet Gynecol 1981; 58: 282–85.
2005; 193: 1923–35. 40 Fretts RC, Usher RH. Causes of fetal death in women of advanced
15 Gibbs RS. The origins of stillbirth: infectious diseases. maternal age. Obstet Gynecol 1997; 89: 40–45.
Semin Perinatol 2002; 26: 75–78. 41 Liu S, Joseph KS, Kramer MS, et al. Relationship of prenatal
16 Goldenberg RL,Thompson C. The infectious origins of stillbirth. diagnosis and pregnancy termination to overall infant mortality in
Am J Obstet Gynecol 2003; 189: 861–73. Canada. JAMA 2002; 287: 1561–67.

www.thelancet.com Vol 370 November 17, 2007 1723

 Seminar

42 Fretts RC, Schmittdiel J, Mclean FH, Usher RH, Goldman MB. 66 Odlind V, Haglund B, Pakkanen M, Otterblad OP. Deliveries,
Increased maternal age and the risk of fetal death. N Engl J Med mothers and newborn infants in Sweden, 1973–2000. Trends in
1995; 333: 953–57. obstetrics as reported to the Swedish Medical Birth Register.
43 Nybo Andersen AM, Wohlfahrt J, Christens P, Olsen J, Melbye M. Acta Obstet Gynecol Scand 2003; 82: 516–28.
Maternal age and fetal loss: population based register linkage study. 67 Confidential enquiry into maternal and child health. Perinatal
BMJ 2000; 320: 1708–12. mortality surveillance, 2004: England, Wales and Northern Ireland.
44 Fretts RC, Elkin EB, Myers ER, Heffner LJ. Should older women London: CEMACH, 2006.
have antepartum testing to prevent unexplained stillbirth? 68 Parrish KM, Holt VL, Easterling TR, Connell FA, LoGerfo JP. Effects
Obstet Gynecol 2004; 104: 56–64. of changes in maternal age, parity and birth weight distribution on
45 Simpson LL. Maternal medical disease: risk of antepartum fetal primary Cesarean section rates. JAMA 1994; 271: 443–47.
death. Semin Perinatol 2002; 26: 42–50. 69 Kanagalingam MG, Forouhi NG, Greer IA, Sattar N. Changes in
46 Smulian JC, Ananth CV, Vintzileos AM, Scorza WE, Knuppel RA. booking body mass index over a decade: retrospective analysis from
Fetal deaths in the United States. Influence of high-risk conditions a Glasgow Maternity Hospital. BJOG 2005; 112: 1431–33.
and implications for management. Obstet Gynecol 2002; 100: 1183–89. 70 Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F,
47 Macintosh MC, Fleming KM, Bailey JA, et al. Perinatal mortality Kirmeyer S. Births: final data for 2004. Natl Vital Stat Rep 2006;
and congenital anomalies in babies of women with type 1 or type 2 55: 1–101.
diabetes in England, Wales, and Northern Ireland: population based 71 Redman CW, Sargent IL. Latest advances in understanding
study. BMJ 2006; 333: 177. preeclampsia. Science 2005; 308: 1592–94.
48 Saade GR, McLintock C. Inherited thrombophilia and stillbirth. 72 Froen JF, Gardosi JO, Thurmann A, Francis A, Stray-Pedersen B.
Semin Perinatol 2002; 26: 51–69. Restricted fetal growth in sudden intrauterine unexplained death.
49 Wu O, Robertson L, Twaddle S, et al. Screening for thrombophilia Acta Obstet Gynecol Scand 2004; 83: 801–07.
in high-risk situations: systematic review and cost-effectiveness 73 Lyall F. The human placental bed revisited. Placenta 2002; 23: 555–62.
analysis. The Thrombosis: Risk and Economic Assessment of 74 Lees C, Parra M, Missfelder-Lobos H, Morgans A, Fletcher O,
Thrombophilia Screening (TREATS) study. Health Technol Assess Nicolaides KH. Individualized risk assessment for adverse
2006; 10: 1–110. pregnancy outcome by uterine artery Doppler at 23 weeks.
50 Redline RW. Thrombophilia and placental pathology. Obstet Gynecol 2001; 98: 369–73.
Clin Obstet Gynecol 2006; 49: 885–94. 75 Smith GCS, Yu CK, Papageorghiou AT, Cacho AM, Nicolaides KH.
51 Bonnici F, Keavney B, Collins R, Danesh J. Angiotensin converting Maternal uterine artery Doppler flow velocimetry and the risk of
enzyme insertion or deletion polymorphism and coronary stillbirth. Obstet Gynecol 2007; 109: 144–51.
restenosis: meta-analysis of 16 studies. BMJ 2002; 325: 517–20. 76 Baschat AA, Gembruch U, Weiner CP, Harman CR. Qualitative
52 Samueloff A, Xenakis EM, Berkus MD, Huff RW, Langer O. venous Doppler waveform analysis improves prediction of critical
Recurrent stillbirth. Significance and characteristics. J Reprod Med perinatal outcomes in premature growth-restricted fetuses.
1993; 38: 883–86. Ultrasound Obstet Gynecol 2003; 22: 240–45.
53 Surkan PJ, Stephansson O, Dickman PW, Cnattingius S. Previous 77 Krebs C, Macara LM, Leiser R, Bowman AW, Greer IA, Kingdom JC.
preterm and small-for-gestational-age births and the subsequent Intrauterine growth restriction with absent end-diastolic flow
risk of stillbirth. N Engl J Med 2004; 350: 777–85. velocity in the umbilical artery is associated with maldevelopment
54 Smith GCS, Shah I, White IR, Pell JP, Dobbie R. Previous of the placental terminal villous tree. Am J Obstet Gynecol 1996;
pre-eclampsia, preterm delivery and delivery of a small for 175: 1534–42.
gestational age infant and the risk of unexplained stillbirth in the 78 Smith GCS. First trimester origins of fetal growth impairment.
second pregnancy: a retrospective cohort study, Scotland 1992–2001. Semin Perinatol 2004; 28: 41–50.
Am J Epidemiol 2007; 165: 194–202. 79 Smith GCS, Smith MFS, McNay MB, Fleming JEE. First-trimester
55 Smith GCS, Pell JP, Dobbie R. Caesarean section and risk of growth and the risk of low birth weight. N Engl J Med 1998;
unexplained stillbirth in subsequent pregnancy. Lancet 2003; 339: 1817–22.
362: 1779–84. 80 Smith GCS, Stenhouse EJ, Crossley JA, Aitken DA, Cameron AD,
56 Taylor LK, Simpson JM, Roberts CL, Olive EC, Connor JM. Development: early-pregnancy origins of low birth
Henderson-Smart DJ. Risk of complications in a second pregnancy weight. Nature 2002; 417: 916.
following caesarean section in the first pregnancy: a 81 Smith GCS, Crossley JA, Aitken DA, et al. First-trimester placentation
population-based study. Med J Aust 2005; 183: 515–19. and the risk of antepartum stillbirth. JAMA 2004; 292: 2249–54.
57 Richter R, Bergmann RL, Dudenhausen JW. Previous caesarean 82 Stallmach T, Hebisch G, Meier K, Dudenhausen JW, Vogel M.
or vaginal delivery: which mode is a greater risk of perinatal death Rescue by birth: defective placental maturation and late fetal
at the second delivery? Eur J Obstet Gynecol Reprod Biol 2006; mortality. Obstet Gynecol 2001; 97: 505–09.
132: 51–57. 83 Coan PM, Burton GJ, Ferguson-Smith AC. Imprinted genes in the
58 Salihu HM, Sharma PP, Kristensen S, et al. Risk of stillbirth placenta–a review. Placenta 2005; 26 (suppl A): S10–20.
following a caesarean delivery: black-white disparity. Obstet Gynecol 84 Hiby SE, Walker JJ, O’Shaughnessy KM, et al. Combinations of
2006; 107: 383–90. maternal KIR and fetal HLA-C genes influence the risk of
59 Kiely JL, Kiely M. Epidemiological trends in multiple births in the preeclampsia and reproductive success. J Exp Med 2004; 200: 957–65.
United States, 1971–1998. Twin Res 2001; 4: 131–33. 85 Smith GCS, Stenhouse EJ, Crossley JA, Aitken DA, Cameron AD,
60 Smulian JC, Ananth CV, Kinzler WL, Kontopoulos E, Connor JM. Early pregnancy levels of pregnancy-associated plasma
Vintzileos AM. Twin deliveries in the United States over three protein A and the risk of intra-uterine growth restriction, premature
decades: an age-period-cohort analysis. Obstet Gynecol 2004; birth, pre-eclampsia and stillbirth. J Clin Endocrinol Metab 2002;
104: 278–85. 87: 1762–67.
61 Salihu HM, Aliyu MH, Rouse DJ, Kirby RS, Alexander GR. 86 Dugoff L, Hobbins JC, Malone FD, et al. First-trimester maternal
Potentially preventable excess mortality among higher-order serum PAPP-A and free-beta subunit human chorionic
multiples. Obstet Gynecol 2003; 102: 679–84. gonadotropin concentrations and nuchal translucency are associated
62 Bell R, Glinianaia SV, Rankin J, Wright C, Pearce MS, Parker L. with obstetric complications: a population-based screening study
Changing patterns of perinatal death, 1982–2000: a retrospective (the FASTER Trial). Am J Obstet Gynecol 2004; 191: 1446–51.
cohort study. Arch Dis Child Fetal Neonatal Ed 2004; 89: F531–36. 87 Waller DK, Lustig LS, Smith AH, Hook EB. Alpha-fetoprotein: a
63 Blickstein I, Keith LG. Multiple pregnancy: epidemiology, gestation biomarker for pregnancy outcome. Epidemiology 1993; 4: 471–76.
and perinatal outcome. Abingdon: Taylor & Francis, 2005. 88 Proud J, Grant AM. Third trimester placental grading by
64 Kalter H. Five-decade international trends in the relation of ultrasonography as a test of fetal wellbeing. BMJ (Clin Res Ed) 1987;
perinatal mortality and congenital malformations: stillbirth and 294: 1641–44.
neonatal death compared. Int J Epidemiol 1991; 20: 173–79. 89 Viero S, Chaddha V, Alkazaleh F, et al. Prognostic value of placental
65 Fretts RC, Boyd ME, Usher RH, Usher HA. The changing pattern ultrasound in pregnancies complicated by absent end-diastolic flow
of fetal death, 1961–1988. Obstet Gynecol 1992; 79: 35–39. velocity in the umbilical arteries. Placenta 2004; 25: 735–41.

1724 www.thelancet.com Vol 370 November 17, 2007

 Seminar

90 Coppens M, James DK. Planning prenatal care and identification of 100 Bricker L, Neilson JP. Routine ultrasound in late pregnancy (after
risk (screening). In: James DK, Steer PJ, Weiner CP, Gonik B, eds. 24 weeks gestation) (Review). Cochrane Database Syst Rev 2000;
High risk pregnancy: management options. Philadelphia: Elsevier, 2: CD001451.
2006: 67–83. 101 Malone FD, Canick JA, Ball RH, et al. First-trimester or
91 Creasy RK. Management of labor and delivery. Malden: Blackwell second-trimester screening, or both, for Down’s syndrome.
Science, 1997. N Engl J Med 2005; 353: 2001–11.
92 Bates SM, Greer IA, Hirsh J, Ginsberg JS. Use of antithrombotic 102 Smith GCS. Life-table analysis of the risk of perinatal death at term
agents during pregnancy: the Seventh ACCP Conference on and post term in singleton pregnancies. Am J Obstet Gynecol 2001;
Antithrombotic and Thrombolytic Therapy. Chest 2004; 184: 489–96.
126: 627S–44S. 103 Crowley P. Interventions for preventing or improving the outcome
93 Say L, Gulmezoglu AM, Hofmeyr GJ. Maternal oxygen of delivery at or beyond term. Cochrane Database Syst Rev 2000;
administration for suspected impaired fetal growth. 2: CD000170.
Cochrane Database Syst Rev 2003; 1: CD000137. 104 Schott J, Henley A, Kohner N. Pregnancy loss and the death of a
94 Grant A, Elbourne D, Valentin L, Alexander S. Routine formal fetal baby: guidelines for professionals. London: Stillbirth and Neonatal
movement counting and risk of antepartum late death in normally Death Society, 2007.
formed singletons. Lancet 1989; 334: 345–49. 105 Neilson JP. Mifepristone for induction of labour.
95 Froen JF. A kick from within--fetal movement counting and the Cochrane Database Syst Rev 2000; 4: CD002865.
cancelled progress in antenatal care. J Perinat Med 2004; 32: 13–24. 106 Neilson JP, Hickey M, Vazquez J. Medical treatment for early fetal
96 ACOG Clinical management guidelines for death (less than 24 weeks). Cochrane Database Syst Rev 2006;
obstetrician-gynecologists. ACOG practice bulletin. Antepartum 3: CD002253.
fetal surveillance. Number 9, October 1999. Int J Gynaecol Obstet 107 Hughes P, Turton P, Hopper E, Evans CD. Assessment of
2000; 68: 175–85. guidelines for good practice in psychosocial care of mothers after
97 Neilson JP, Alfirevic Z. Doppler ultrasound for fetal assessment in stillbirth: a cohort study. Lancet 2002; 360: 114–18.
high risk pregnancies. Cochrane Database Syst Rev 2000; 108 Griffiths PD, Paley MN, Whitby EH. Post-mortem MRI as an
2: CD000073. adjunct to fetal or neonatal autopsy. Lancet 2005; 365: 1271–73.
98 Bricker L, Neilson JP. Routine doppler ultrasound in pregnancy.
Cochrane Database Syst Rev 2000; 2: CD001450.
99 Smith GCS. Estimating risks of perinatal death. Am J Obstet Gynecol
2005; 192: 17–22.

www.thelancet.com Vol 370 November 17, 2007 1725