https://orcid.org/0000-0001-8981-4817
Figures
Abstract
Objective
To assess the risk of singleton intrauterine fetal death (IUFD) in women by the demographic setting of the online Fetal Medicine Foundation (FMF) Stillbirth Risk Calculator.
Methods
Retrospective single-centre case-control study involving 144 women having suffered IUFD and 247 women after delivery of a live-born singleton. Nonparametric receiver operating characteristics (ROC) analyses were performed to predict the prognostic power of the FMF Stillbirth risk score and to generate a cut-off value to discriminate best between the event of IUFD versus live birth.
Results
Women in the IUFD cohort born a significantly higher overall risk with a median FMF risk score of 0.45% (IQR 0.23–0.99) compared to controls [0.23% (IQR 0.21–0.29); p<0.001]. Demographic factors contributing to an increased risk of IUFD in our cohort were maternal obesity (p = 0.002), smoking (p<0.001), chronic hypertension (p = 0.015), antiphospholipid syndrome (p = 0.017), type 2 diabetes (p<0.001), and insulin requirement (p<0.001). ROC analyses showed an area under the curve (AUC) of 0.72 (95% CI 0.67–0.78; p<0.001) for predicting overall IUFD and an AUC of 0.72 (95% CI 0.64–0.80; p<0.001), respectively, for predicting IUFD excluding congenital malformations. The FMF risk score at a cut-off of 0.34% (OR 6.22; 95% CI 3.91–9.89; p<0.001) yielded an 82% specificity and 58% sensitivity in predicting IUFD with a positive and negative predictive value of 0.94% and 99.84%, respectively.
Citation: Muin DA, Windsperger K, Attia N, Kiss H (2022) Predicting singleton antepartum stillbirth by the demographic Fetal Medicine Foundation Risk Calculator—A retrospective case-control study. PLoS ONE 17(1): e0260964. https://doi.org/10.1371/journal.pone.0260964
Editor: Simone Garzon, Universita degli Studi dell’Insubria, ITALY
Received: June 30, 2021; Accepted: November 20, 2021; Published: January 20, 2022
Copyright: © 2022 Muin et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: We confirm that we have uploaded the dataset to following recommended repository: https://figshare.com/articles/dataset/FMF_Calculator_xlsx/17072672.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Intrauterine fetal death (IUFD) is a devastating event, which underlies a heterogeneous spectrum of causes, both acute and chronic [1]. Acute events, such as placental abruption, cord accidents, maternal or fetal trauma, are usually unpreventable, yet bear a low recurrence risk in future pregnancies, unless there is a background risk factor, such as a uterine malformation or maternal clotting and bleeding disorder precipitating placental abruption [2–4]. Conversely, chronic events are often accompanied by placental dysfunction leading to hypoxia, fetal growth restriction and ultimately IUFD, if the fetus is not delivered at an appropriate time [5, 6]. Placental dysfunction is often associated with maternal risk factors, such as antiphospholipid syndrome (APS), systemic lupus erythematosus (SLE), diabetes, obesity, vascular disorders as in preeclampsia, as well as maternal age and assisted reproduction [7–9]. Given these circumstances, the recurrence risk for fetal death is up to 22-fold [10, 11]. Whilst these risk factors are practically unmodifiable, fetal death may be prevented if women are subjected to therapeutic intervention, intensified antenatal surveillance, and eventually delivery [12, 13]. After all, from a socio-epidemiological standpoint, we ought to conceive risk of stillbirth as a continuum of “risk degrees”, rather than a dichotomy between “two extremes of exposure” [14].
Prediction models for stillbirths are commonly defined as “models, scores or clinical decision tools” which estimate the risk of adverse perinatal outcome upon certain variables [15]. To date, 69 prediction models for stillbirth have been described in the literature. A recent systematic review has identified maternal age, body-mass-index (BMI), and diabetes as the three most commonly used variables in such prediction tools [16]. The best evidence with consistently strong association with stillbirth is documented for maternal obesity, pre-existing hypertension, and nulliparity.
The Fetal Medicine Foundation (FMF) Stillbirth Risk Calculator is an online tool, which was designed to assess the risk of stillbirth in reference to either maternal history, or a combination of first and second trimester measurements, and biomarkers [17–26]. The demographic model is based upon maternal characteristics (i.e., weight, ethnicity, smoking), medical history (i.e., diabetes, chronic hypertension, SLE, APS), and obstetric history (i.e., parity, stillbirth and/or preeclampsia in previous pregnancies).
By the current study, we aimed to apply the demographic model of the FMF Stillbirth Risk Calculator in our cohort of women having suffered IUFD and matched live births as an independent dataset for external validation of this online prediction tool.
Methods
Study design and data collection
We retrospectively reviewed all cases of singleton IUFD and live birth that were delivered at our tertiary referral center between January 2003 and December 2019. Inclusion criteria were singleton IUFDs above 21+0 gestational weeks with documented fetal and maternal characteristics either at antenatal booking or at the time of delivery. Exclusion criteria for IUFDs were multiple pregnancies, medical or surgical terminations of pregnancy, perinatal fetal deaths, and cases with missing fetal and maternal demographics and medical history.
IUFD cases, which fulfilled the study’s inclusion and exclusion criteria, were matched with singleton live births for fetal sex, maternal age, gravidity, parity, and gestational age at delivery (S1 Appendix).
Maternal and fetal characteristics were retrieved from the electronic database ViewPoint Version 5.6.28.56 (General Electric Company, Solingen, Germany). Medical and obstetric history, including type 1 and 2 diabetes and respective treatment (i.e., diet, insulin, metformin), chronic hypertension, SLE, APS, and previous preeclampsia, were manually retrieved from the women’s medical records. Data were transferred into a study-file sheet, reviewed for accuracy, and made anonymous. Demographic variables from the included subjects were manually typed into the online FMF Stillbirth Risk Calculator (https://fetalmedicine.org/research/assess/stillbirth). For this study, we applied the "Maternal history” setting only. The individual risk score directly correlates with the degree of risk in each case and is presented in percent (%) or fractions (i.e., 1 in x).
Definitions
Maternal age was defined as age in years at the time of the delivery. Maternal weight (in kg) was obtained at the first visit. BMI was grouped as underweight (≤18.5 kg/m2), normal weight (18.6–24.9 kg/m2), pre-obesity (25.0–29.9 kg/m2), obesity class I (30.0–34.9 kg/m2), obesity class II (35.0–39.9 kg/m2), and obesity class III (≥40.0 kg/m2). Ethnicity was self-reported by the pregnant woman at the first visit and categorized for this study into white, black, South Asian, East Asian, and mixed. Smoking was defined as current smoker or non-smoker at antenatal booking or at the time of stillbirth. Gravidity was defined as the number of the current pregnancy. Parity was defined as the number of previous deliveries above 24+0 gestational weeks or with a fetal weight ≥500 g.
As per Austrian law, each case of fetal death underwent thorough pathology assessment. Upon receipt of all post-mortem results (i.e., fetal autopsy, genetic testing, placental histology, and maternal investigation), we concluded the cause of fetal death according to the taxonomy as described by Reddy et al [27]. Cause of fetal death was defined as the “initial, demonstrable pathophysiological entity initiating the chain of events that has irreversibly led to death” in recognition of potentially multiple competing risks and events, and was categorized according to the “Causes of death and associated conditions” (CODAC) classification [28].
Statistical analyses
The distribution of data was analyzed using the Kolmogorov-Smirnov test. Normally distributed data are expressed as mean ± standard deviation. Not normally distributed variables are expressed as median and minimum-maximum, or 25% and 75% interquartile range (IQR). Categorical data are given as absolute (n) and relative frequencies (%). Continuous data were compared with unpaired t-test and Wilcoxon signed-rank test, respectively. Categorical data were compared with Chi2 and Fisher’s Exact test, respectively, with a 99% Confidence Interval (CI). Univariable and multivariable logistic regression analyses were used to determine the odds ratio (OR) with a 95% CI from maternal characteristics and medical history provided for the prediction of stillbirth. Nonparametric receiver operating characteristics (ROC) analyses with an area under the curve (AUC) were performed to predict the prognostic power of the risk score, and to generate a cut-off value that best discriminated between the event of stillbirth versus live birth at the point where the sum of sensitivity and specificity was the highest [29].
Considering a prevalence of 0.3% for IUFDs in Central Europe [30], the positive predictive value (PPV) and negative predictive value (NPV) of the FMF prediction tool was calculated as follows: PPV = (sensitivity x prevalence)/[(sensitivity x prevalence) + ((1–specificity) x (1-prevalence))], and NPV = (specificity x (1–prevalence))/[(specificity x (1–prevalence)) + ((1–sensitivity) x prevalence)].
All reported p-values are two-sided, and the level of significance was set at <0.05. Statistical tests were performed by SPSS Statistics Version 26.0.0.0 (IBM Corporation, Armonk, NY, USA), and figures were designed by GraphPad Prism 9 for macOS Version 10.14.6 (GraphPad Software, LLC).
Details of ethical approval
The study was approved by the Ethics Committee of the Medical University of Vienna (registration number 1759/2018) and complied with the principles outlined in the Helsinki Declaration of 1975, as revised in 2013. Women’s written consent was not required as per the Austrian Federal Act (Protection of Personal Data Regulation, §46, Paragraph 1; 2000).
Results
Baseline characteristics
Our study cohort consisted of 144 cases of fetal death and 247 matched live births. Compared to matched controls, women after IUFD had a greater BMI (p = 0.002), were more often smokers (p<0.001), and suffered more frequently medical conditions, such as hypertension (p = 0.015), APS (p = 0.017), and diabetes (p<0.001), with a more frequent need for insulin (p = 0.006; Table 1).
The IUFD cohort involved 73 (50.7%) male and 71 (49.3%) female fetuses between 21+1 and 41+3 gestational weeks at a median age of 31+4 gestational weeks. The median fetal weight at stillbirth was 1397 g (180–4450 g). Abnormal fetal post-mortem findings and/or prenatally confirmed congenital malformations were reported in 72 (50.0%) cases. Causes of death were confined as placental pathologies in 26 (18.1%) cases; unknown despite thorough investigation or due to lack of important information in 25 (17.4%) cases; fetal conditions in 9 (6.3%) cases; umbilical cord complications in 6 (4.2%) cases; maternal conditions in 4 (2.8%) cases, and IUFD due to infection in 2 (1.4%) cases.
The matched live birth cohort consisted of 138 (55.9%) male and 109 (44.1%) female newborns delivered between 21+1 and 41+3 gestational weeks at a median age of 31+6 gestational weeks. Median weight at delivery was 1880 g (341–4540 g) with a median Apgar score of 8 points at 1 minute, and 9 points at 5 and 10 minutes, respectively, with a median arterial pH of 7.29 (6.89–7.46). From the five matched live births under 22+0 gestational weeks, Case 1 was noted to have a univentricular heart (male, birth weight 482g, Apgar 1/1/1). Case 2 had a preterm delivery following premature rupture of membranes (female, 354g, Apgar 2/1/1). Case 3 had bilateral kidney agenesis with oligohydramnios (male, 341g, Apgar 1/1/1. Case 4 suffered amniotic infection followed by preterm delivery (male, 440g, Apgar 1/0/0), and Case 5 had commissural agenesis and vermis hypoplasia (male, 541g, Apgar 2/1/1). None of those extremely premature newborns survived the first two weeks of life.
Prediction of stillbirth
The results of univariable and multivariable regression analyses to predict the risk of antepartum stillbirth in our cohort are shown in Table 2. In univariable regression analysis, the risk of stillbirth was higher in women with type 2 diabetes, nulliparous women, and those with a history of stillbirth. In multivariable regression analysis, the risk of stillbirth increased with chronic hypertension, nicotine consumption, and diabetes.
Fetal Medicine Foundation stillbirth risk score
Fig 1 shows the distribution of the FMF risk score in the IUFD cohort and matched controls. The median risk score in the IUFD cohort was 0.45% (i.e., risk 1 in 222; IQR 0.23–0.99), whilst the median score in the group of matched live births was 0.23% (i.e., risk 1 in 435; IQR 0.21–0.29%; p<0.001].
Red line at 0.34% represents the cut-off value to predict IUFD with a sensitivity of 58% and a specificity of 82%. Red dotted lines represent median values of the FMF risk score; black dotted lines represent 25% and 75% interquartile ranges; x-axis in Log 10 scale.
Fig 2 shows the FMF Stillbirth risk scores stratified per causes of fetal death.
To evaluate the discriminative power of the FMF risk score we performed a ROC analysis and calculated an AUC of 0.72 (95% CI 0.67–0.78; p<0.001) for predicting antepartum stillbirth in the total cohort. The cut-off level of 0.34% yielded a sensitivity of 57.64% (95% CI 49.47–65.41) and specificity of 81.78% (95% CI 76.49–86.10) in identifying risk of IUFD (OR 6.22; 95% CI 3.91–9.89; p<0.001; likelihood ratio 3.16) with a PPV of 0.94% and NPV of 99.84%.
After exclusion of fetal anomalies (n = 72) from the IUFD cohort, the ROC analysis showed an AUC of 0.72 (95% CI 0.64–0.80; p<0.001) with an optimal cut-off at 0.41% to predict IUFD with a sensitivity of 59.72% (95% CI 48.18–70.28) and specificity of 81.36% (95% CI 73.38–87.35; likelihood ratio 3.20) with a PPV of 0.95% and NPV of 99.85% (Fig 3a).
a) Receiver operating characteristics (ROC) curve for the Fetal Medicine Foundation (FMF) Stillbirth risk score to predict intrauterine fetal death (IUFD). ROC was performed in the total IUFD cohort (n = 144; blue line) and the IUFD cohort excluding cases of fetal anomaly (n = 75; green line), respectively. b) ROC curve for the FMF risk score to predict IUFD, as stratified by causes of fetal death (maternal n = 4; unknown n = 25; fetal anomaly n = 72; placenta n = 26; fetal condition n = 9; infection n = 2; cord accident n = 6).
ROC analyses stratified by causes of fetal death showed highest accuracy for prediction of IUFD caused by maternal disease [cut-off 0.72%: sensitivity 75.00% (95% CI 30.06–98.72); specificity 100.00% (95% CI 64.57–100.00)], placental dysfunction [cut-off 0.34%: sensitivity 57.69% (95% CI 38.95–74.46); specificity 68.18% (95% CI 53.44–80.00); likelihood ratio 1.82], fetal anomalies [cut-off 0.34%: sensitivity 52.78% (95% CI 41.40–63.87); specificity 84.50% (95% CI 77.26–89.73); likelihood ratio 3.40], and of unknown etiology [cut-off 0.38%: sensitivity 76.00% (95% CI 56.57–88.50); specificity 88.10% (95% CI 75.00–94.81); likelihood ratio 6.38; Fig 3b].
Discussion
Main findings
In this study, we independently assessed and validated the FMF Stillbirth Risk Calculator based upon demographic characteristics in a Middle-European case-matched cohort involving IUFDs ≥ 21+0 weeks. The online prediction tool achieved similar performance as in the reference group [17], and maternal, medical, and obstetric history yielded a high specificity and satisfactory sensitivity in discriminating between the outcomes live birth versus stillbirth, especially in cases of IUFD due to maternal disease, placental dysfunction, fetal anomaly, and of unknown cause. Most relevant conditions increasing the risk of IUFD in our cohort were type 2 diabetes, previous stillbirth, chronic hypertension, smoking, APS, and SLE.
As the NPV of the demographic FMF prediction tool was high (99.8% for the total IUFD cohort and excluding congenital anomalies), the clinical value of the demographic setting lies in identifying women who are at small risk of antepartum stillbirth and providing reassurance that these women do not require close antenatal monitoring, medical interventions or hospital admission. The inclusion of the demographic FMF Risk Calculator as early as at antenatal booking may facilitate efficient triage of women and help allocate resources of antenatal surveillance until timely delivery may be required.
Due to the retrospective design of this study, women in our cohort had not been risk-evaluated by the online FMF prediction tool in the antenatal or peripartum period. Therefore none of the women had undergone any specific intervention, other than the routine management of their known medical and obstetric conditions according to the best evidence at that time.
Clinical implications
In the 1980s, the British epidemiologist Geoffrey Rose coined the term Prevention Paradox by proving that “a large number of people at a small risk may give rise to more cases of disease than the small number who are at high risk" [31]. Leveling off somewhere between a “high-risk” and “population-based” strategy, there lie the strategies to prevent fetal deaths in high-income countries. With respect to the diversity of the underlying pathomechanisms, the risk of stillbirth is considered a condensation of multiple accumulating risk factors. Reducing the risk of stillbirth therefore requires primarily the collection of maternal, fetal and obstetric characteristics, which may then allow the identification of variables connected with adverse perinatal outcome with highest sensitivity and specificity. Given the relatively low prevalence in Central Europe, yet high cost of false-positive prediction of antepartum stillbirths, it is therefore recommended to choose a prediction cut-off at the lower part of the ROC curve to maximize specificity [32].
It is well recognized, that maternal demographic characteristics may influence fetal outcome. The “history only” setting of the FMF Stillbirth Calculator utilizes those demographic patterns and calculates the individual risk based upon a reference population. Whilst it is largely acknowledged that, e.g., maternal diabetes, SLE, APS, sickle cell disease, increased age and previous preeclampsia or stillbirth may hazard the current pregnancy, these women should be naturally considered as high-risk patients by their obstetricians and treated as such with timely intervention when required [13]. Yet, at the same time, the FMF Stillbirth Calculator allows objectifying and quantifying this risk in each individual case and may facilitate clinical decision-making regarding close antenatal surveillance and monitoring. With a cut-off risk score of 0.34%, or 1 in 294, women in our cohort were identified as high-risk for placental dysfunction, and stillbirth, likewise. This information may endorse the initiation of e.g. calcium or aspirin during the first trimester or at least enroll the woman into a close surveillance program with regular fetal growth scans, utero-placental Doppler ultrasound measurements, and computerized antenatal cardiotocography [12].
Recent systematic reviews reveal the superiority of maternal characteristics in risk-evaluation [13, 16]. They confirm a robust body of evidence for the link of nulliparity, pre-existing hypertension, and increased maternal BMI to antepartum stillbirth [16, 33]. The demographic findings from our cohort support these data, as all but nulliparity were significantly more prevalent in women affected by fetal death.
In acknowledgment of the challenge to predict adverse events in the presence of competing risks at later stages in pregnancy, the demographic maternal patterns, which may remain unaltered most often during pregnancy, may assist clinical risk assessment at the first booking appointment. Whilst these demographic high-risk characteristics are well recognized among clinicians, it may be useful to provide women with precise risk-estimates, which this readily available and free online tool provides on the basis of over 113 400 singleton pregnancies.
Two characteristics in our cohort are of note: First, we included both stillborn and live-born fetuses with congenital anomalies. Despite this variation to the reference cohort, the prediction model remained accurate with and without the consideration of congenital anomalies in our population. A recent secondary analysis of a case-control study confirmed that among stillbirths, 23.4% had one or more major anomalies compared to 4.3% of live births [34]. Yet, taking these data together, it seems justified to prospectively use this model for risk stratification in stillbirth at preconception, thus ahead of any eventual detection of fetal congenital anomalies later in pregnancy. Also, we considered all IUFD cases from gestational week 21+0 onwards. One reason for this is the international heterogeneity in definitions of stillbirth by gestational age and to address those who define fetal death as of week 20+0 [15].
Research implications
Townsend et al. have proposed that a future robust risk prediction tool for stillbirth should incorporate the following candidate variables: maternal age, BMI, parity, pre-existing hypertension, diabetes, previous stillbirth, nicotine consumption, uterine artery Doppler, pregnancy-associated plasma protein PAPP-A, and placental growth factor PlGF [16]. The merit of such a clinical model would be twofold: primarily, the accurate discrimination of high- from low-risk pregnant women, and secondarily, recognizing the variables that may require early enough alteration, if they are modifiable. Whilst maternal age, parity, previous stillbirths, and biomarkers cannot be adapted by intervention, maternal weight, hypertension, type 2 diabetes, and nicotine consumption can be improved through lifestyle modifications [12]. To extrapolate this concept to the demographic model of the FMF Stillbirth Risk Calculator, only four variables may be potential subjects to change and therefore possible risk reduction (weight, smoking, diabetes, hypertension). As with many other risk-adjustment models, however, social and behavioral variables, such as domestic abuse, stress, employment, and deprivation are hard to capture and should be further considered within a population-based conceptual framework [35]. Additional research into prediction models may objectify the true preventability of stillbirth by adaption of modifiable risk factors in the future.
Strengths and limitations
The strengths of this study are first, the strict inclusion and exclusion criteria in the cohort of singleton antepartum stillbirths, each of which was subject to extensive post-mortem investigations to define the cause of death. Our institutional database on antepartum stillbirths includes valid and accurate data, which are regularly checked and updated. This ascertains a continuous degree of data accuracy. The single-center setting of the study also reduces bias and heterogeneity in reporting pathology findings and collection of pregnancy data.
However, we acknowledge the relatively small sample size due to the study design and its retrospective setting. Also, our cohort is ethnically non-diverse, which may further limit the generalizability of the results. Lastly, certain maternal parameters (e.g., smoking, parity, previous preeclampsia) were self-reported by the woman at the time of antenatal booking or delivery and therefore subject to recall bias and potential language barrier in case of foreign nationality, which we cannot control for.
Conclusion
Raising awareness for maternal epidemiological risk factors is an important measure in antenatal care. The demographic setting of the online Fetal Medicine Foundation Stillbirth Risk Calculator may be a useful tool to identify women at low risk of antepartum stillbirth. Its implementation into clinical practice as early as in the preconception period might support obstetrical counseling and further antenatal management. Prospective validation in larger multinational cohort studies is recommended.
Supporting information
S1 Appendix. Flow diagram on the selection of fetal death cases and matched controls.
https://doi.org/10.1371/journal.pone.0260964.s001
(DOCX)
Acknowledgments
The authors are grateful to Mag. Dr. Christoph Krall and Mag Irene Steiner for their statistical review.
References
- 1. Silver RM Fetal death. Obstet Gynecol 2007 Jan;109(1):153–67. pmid:17197601
- 2. Page JM, Silver RM (2020) Stillbirth: Evaluation and Follow-up. Obstet Gynecol Clin North Am 47: 439–451. pmid:32762929
- 3. Reddy UM (2007) Prediction and prevention of recurrent stillbirth. Obstet Gynecol 110: 1151–1164. pmid:17978132
- 4.
Heinonen S, Kirkinen P Pregnancy outcome after previous stillbirth resulting from causes other than maternal conditions and fetal abnormalities.
- 5. Bukowski R (2010) Stillbirth and fetal growth restriction. Clin Obstet Gynecol 53: 673–680. pmid:20661051
- 6. Ptacek I, Sebire NJ, Man JA, Brownbill P, Heazell AE (2014) Systematic review of placental pathology reported in association with stillbirth. Placenta 35: 552–562. pmid:24953162
- 7. Bedell S, Hutson J, de Vrijer B, Eastabrook G (2021) Effects of Maternal Obesity and Gestational Diabetes Mellitus on the Placenta: Current Knowledge and Targets for Therapeutic Interventions. Curr Vasc Pharmacol 19: 176–192. pmid:32543363
- 8. Sarmon KG, Eliasen T, Knudsen UB, Bay B (2021) Assisted reproductive technologies and the risk of stillbirth in singleton pregnancies: a systematic review and meta-analysis. Fertility and Sterility 116: 784–792. pmid:34023069
- 9.
Ventolini G Conditions associated with placental dysfunction.
- 10. Bhattacharya S, Prescott GJ, Black M, Shetty A (2010) Recurrence risk of stillbirth in a second pregnancy. BJOG 117: 1243–1247. pmid:20573152
- 11. Lamont K, Scott NW, Jones GT, Bhattacharya S (2015) Risk of recurrent stillbirth: systematic review and meta-analysis. BMJ 350: h3080. pmid:26109551
- 12. Ota E, da Silva Lopes K, Middleton P, Flenady V, Wariki WM, et al. (2020) Antenatal interventions for preventing stillbirth, fetal loss and perinatal death: an overview of Cochrane systematic reviews. Cochrane Database Syst Rev 12: CD009599. pmid:33336827
- 13. Ishak M, Khalil A (2021) Prediction and prevention of stillbirth: dream or reality. Curr Opin Obstet Gynecol 33: 405–411. pmid:34459791
- 14.
Rose G, Khaw, K., & Marmot, M (2008) Rose’s Strategy of Preventive Medicine. Oxford University Press.
- 15. Townsend R, Manji A, Allotey J, Heazell A, Jorgensen L, et al. (2020) Can risk prediction models help us individualise stillbirth prevention? A systematic review and critical appraisal of published risk models. BJOG. pmid:32894620
- 16. Townsend R, Sileo FG, Allotey J, Dodds J, Heazell A, et al. (2020) Prediction of stillbirth: an umbrella review of evaluation of prognostic variables. BJOG. pmid:32931648
- 17. Yerlikaya G, Akolekar R, McPherson K, Syngelaki A, Nicolaides KH (2016) Prediction of stillbirth from maternal demographic and pregnancy characteristics. Ultrasound Obstet Gynecol 48: 607–612. pmid:27561693
- 18. Akolekar R, Bower S, Flack N, Bilardo CM, Nicolaides KH (2011) Prediction of miscarriage and stillbirth at 11–13 weeks and the contribution of chorionic villus sampling. Prenat Diagn 31: 38–45. pmid:21210479
- 19. Akolekar R, Machuca M, Mendes M, Paschos V, Nicolaides KH (2016) Prediction of stillbirth from placental growth factor at 11–13 weeks. Ultrasound Obstet Gynecol 48: 618–623. pmid:27854388
- 20.
Akolekar R, Machuca M, Mendes M, Paschos V, Nicolaides KH (2016) Placental growth factor in prediction of stillbirths at 11–13 weeks. Ultrasound Obstet Gynecol.
- 21. Akolekar R, Syngelaki A, Gallo DM, Poon LC, Nicolaides KH (2015) Umbilical and fetal middle cerebral artery Doppler at 35–37 weeks’ gestation in the prediction of adverse perinatal outcome. Ultrasound Obstet Gynecol 46: 82–92. pmid:25779696
- 22. Akolekar R, Tokunaka M, Ortega N, Syngelaki A, Nicolaides KH (2016) Prediction of stillbirth from maternal factors, fetal biometry and uterine artery Doppler at 19–24 weeks. Ultrasound Obstet Gynecol 48: 624–630. pmid:27854387
- 23. Aupont JE, Akolekar R, Illian A, Neonakis S, Nicolaides KH (2016) Prediction of stillbirth from placental growth factor at 19–24 weeks. Ultrasound Obstet Gynecol 48: 631–635. pmid:27854395
- 24. Bakalis S, Akolekar R, Gallo DM, Poon LC, Nicolaides KH (2015) Umbilical and fetal middle cerebral artery Doppler at 30–34 weeks’ gestation in the prediction of adverse perinatal outcome. Ultrasound Obstet Gynecol 45: 409–420. pmid:25684172
- 25. Chaveeva P, Carbone IF, Syngelaki A, Akolekar R, Nicolaides KH (2011) Contribution of method of conception on pregnancy outcome after the 11–13 weeks scan. Fetal Diagn Ther 30: 9–22. pmid:21346323
- 26.
Fetal Medicine Foundation (2018) Risk Assessment—Risk for stillbirth.
- 27. Reddy UM, Goldenberg R, Silver R, Smith GC, Pauli RM, et al. (2009) Stillbirth classification—developing an international consensus for research: executive summary of a National Institute of Child Health and Human Development workshop. Obstet Gynecol 114: 901–914. pmid:19888051
- 28. Froen JF, Pinar H, Flenady V, Bahrin S, Charles A, et al. (2009) Causes of death and associated conditions (Codac): a utilitarian approach to the classification of perinatal deaths. BMC Pregnancy Childbirth 9: 22. pmid:19515228
- 29. Krämer W (2011) Wojtek J. Krzanowski and David J. Hand: ROC curves for continuous data. Statistical Papers 52: 979–980.
- 30.
EURO-Peristat (26 November 2018) Stillbirth and infant death rates are decreasing, but still vary widely across Europe.
- 31. Rose G (1981) Strategy of prevention: lessons from cardiovascular disease. Br Med J (Clin Res Ed) 282: 1847–1851. pmid:6786649
- 32. Greiner M, Pfeiffer D, Smith RD (2000) Principles and practical application of the receiver-operating characteristic analysis for diagnostic tests. Prev Vet Med 45: 23–41. pmid:10802332
- 33. Flenady V, Koopmans L, Middleton P, Froen JF, Smith GC, et al. (2011) Major risk factors for stillbirth in high-income countries: a systematic review and meta-analysis. Lancet 377: 1331–1340. pmid:21496916
- 34. Son SL, Allshouse AA, Page JM, Debbink MP, Pinar H, et al. (2020) Stillbirth and Fetal Anomalies: secondary analysis of a case-control study. BJOG. pmid:32946651
- 35. Heazell AEP, Budd J, Smith LK, Li M, Cronin R, et al. (2020) Associations between Social and Behavioural Factors and the risk of Late Stillbirth—Findings from the Midland and North of England Stillbirth Case-Control Study. BJOG.