Risk factors and predictors of stillbirth

1
Risk factors and predictors of stillbirth

• Gordon C S Smith, MD, PhD, MRCOG
• Professor of Obstetrics Gynaecology,
• Cambridge University

2
Antepartum stillbirth

• Major cause of perinatal death
• Majority have no direct cause
• Associated with poor fetal growth

3
Risk assessment and prevention of antepartum
stillbirth

• Primary means of prevention is elective delivery
  of a fetus deemed at risk
• Risk of antepartum stillbirth balanced against
  risk of elective delivery
• Risks of elective delivery
• Preterm Morbidity and mortality related to
  prematurity
• Term Operative delivery Interference in
  natural process of birth

4
Planning intervention

• Identify high risk groups
• Increased level of surveillance (fetal and
  placental Doppler ultrasound)
• Elective delivery if surveillance indicates fetal
  compromise or pregnancy reaches given threshold
  of gestation
• Previous SB at 38 weeks
• IDDM at 39 weeks
• Post-dates pregnancy at term 10 days

5
Predicting risk

• High risk groups already identified
• IDDM
• Previous SB
• Connective tissue disease
• Problem most stillbirths occur to low risk
  women
• Solution identify factors associated with an
  increased risk of stillbirth

6
Where to start?

• Majority of stillbirths have a placental cause
• Abruption
• Pre-eclampsia
• IUGR
• In the absence of overt risk factors, may be able
  to identify high risk women within a low risk
  population by screening tests of placental
  function

7
Tests of placental function

• Assess resistance in uterine circulation by
  Doppler ultrasound of uterine arteries
• Measure circulating maternal concentrations of
  placentally-derived proteins

8
Uterine artery Doppler

• Measured at 20-23 weeks
• Indicates invasion of maternal circulation by
  placenta
• Increased resistance associated with increased
  risk of placentally-related events (fetal death,
  abruption, PET or IUGR delivered before 34 weeks)

9
Biochemical tests of placental function

• Alpha-feto protein (AFP)
• Pregnancy associated plasma protein A (PAPP-A)
• Human chorionic gonadotrophin

10
PAPP-A

• Derived from placenta
• Involved in the control of insulin-like growth
  factors
• Measured in screening for Downs syndrome

11
CUBS Study

• Non-interventional, prospective cohort study in
  central Scotland designed to evaluate predictors
  of Downs syndrome
• Measured PAPP-A and FbhCG 8-14 weeks
• Study group of 8839 women with outcome data
  manually retrieved
• Stillbirth excluded congenital abnormality but
  included all other causes

12
PAPP-A and stillbirth
Adjusted odds ratio for lowest 5 of PAPP-A 3.6
(95 CI 1.2 11.0)
Smith GCS et al, JCEM 2002 8717621767
13
Methods for follow-up study

• Record linked CUBS study database with
• Scottish Morbidity Record 2 (SMR2, national
  database maternity hospital discharge data)
• Scottish Stillbirth Infant Death Enquiry
  (national register of perinatal deaths)
• Limited to women assayed in first 10 weeks
  post-conception
• Linked database contained first trimester
  biochemistry and eventual outcome for 7934
  singleton births between 24-43 weeks gestation
• Independent ascertainment of exposures and events

Smith GCS et al, JAMA. 20042922249-2254
14
Results 1

• No association between low PAPP-A (5th
  percentile) and a range of maternal
  characteristics
• Age, marital status, socio-economic deprivation,
  ethnicity, smoking, parity, previous abortions,
  height and BMI (all Pgt0.05)

Smith GCS et al, JAMA. 20042922249-2254
15
Smith GCS et al, JAMA. 20042922249-2254
16
Smith GCS et al, JAMA. 20042922249-2254
17
Smith GCS et al, JAMA. 20042922249-2254
18
Summary of results

• Very strong relationship between low PAPP-A and
  subsequent risk of stillbirth
• Association due to stillbirth related to
  placental dysfunction
• No placental stillbirths in upper 60 of PAPP-A
• Association specific
• Not due to maternal confounding
• No relationship with FbhCG
• No relationship with other causes of stillbirth

Smith GCS et al, JAMA. 20042922249-2254
19
Aims of research studies

• To identify biological determinants of stillbirth
  
• To obtain clinically useful assessment of risk
• Requires analysis of multiple factors with a
  summary estimate of risk
• Gestational age dependent

20
Potential for combinations of tests
Smith GCS et al 2005, Submitted for publication
21
Gestational age
Grambsch Therneau test P0.04
Smith GCS et al, Lancet 2003 362 177984
22
Current work

• Data source (funded by FSID) record linkage of
• CUBS study (PAPP-A)
• West of Scotland regional biochemical screening
  databases (AFP and hCG),
• SMR2 (basic pregnancy outcome data)
• SSBIDE (perinatal deaths)

23
Aims of current study

• To assess predictive ability of combined maternal
  and biochemical assessment of stillbirth risk to
  identify high risk women
• To determine whether associations with maternal
  and biochemical factors vary according to
  gestational age
• Study cohort of gt200,000 women

24
Current data (unpublished)

• Factors associated with stillbirth and
  associations similar between 24-43 weeks
• Maternal age, deprivation category, height,
  smoking, body mass index and parity
• Factors associated with stillbirth and
  associations differ between 24-43 weeks
• Maternal serum levels of alpha-fetoprotein and
  human chorionic gonadotrophin
• Factors not independently associated with
  stillbirth risk
• Marital status

25
(No Transcript)
26
(No Transcript)
27
Screening performance of models
28
Requirements of model for population-based
screening

• Develop sensitive and specific methods for
  identifying women at increased risk of antepartum
  stillbirth
• Assess mass screening and intervention on the
  basis of such a test

29
Intervention

• Needs to be assessed in RCT
• Recruit population and screen all
• Reveal results and manage 50 per protocol and
  conceal results in other 50
• Primary outcome perinatal death at term
• Intervention would be elective delivery at 37
  weeks
• Less potential to cause harm than preterm
  elective delivery

30
Sample power calculations

• Incidence of antepartum stillbirth at term 1.8
  per 1000
• Trial of a screening test which took the top 5
  of predicted risk as screen positive
• 50 sensitivity, trial would require 60,000 women
• 75 sensitivity would require 22,000 women
• Current method top 5 has 16.4 sensitivity for
  term stillbirth

31
Mass screening

• Key to further progress is to develop a good
  screening tool
• Large scale prospective observational study is
  required to evaluate integrated assessment of
  risk
• Maternal history and characteristics
• Biochemical interrogation of the placenta using
  further analysis of samples already obtained at
  time of booking and triple test
• Uterine artery Doppler at 20 weeks
• ?growth and umbilical artery Doppler scan in
  third trimester
• Identify women with a high absolute risk of term
  stillbirth

32
Conclusions

• Many maternal factors associated with the risk of
  antepartum stillbirth
• Tests of placental function also predictive of
  stillbirth risk
• Currently available methods are better predictors
  of preterm than term stillbirth
• Current methods to assess stillbirth risk are not
  sufficiently discriminative to assess
  population-based screening
• Need to evaluate whether a clinically useful
  model can be created by adding further tests to
  AFP/maternal prediction

33
Acknowledgements

• Foundation for the Study of Infant Deaths
• Richard Dobbie at ISD of NHS Scotland for record
  linkage
• CUBS study group and West of Scotland Regional
  Genetics service (Drs Jenny Crossley and David
  Aitken)
• Prof Jill Pell (Public Health, Glasgow)
• Ian White and Dr Angela Wood at MRC Biostatistics
  Unit, Cambridge