Aim To assess current methods of prediction of adverse pregnancy outcomes, develop a prediction model and assess diet and life style in preventing preeclampsia. Methods Meta-analyses performed to assess the role of abnormal 1st trimester biomarker levels in predicting PE and the predictive accuracy of 2nd trimester UAD indices for stillbirth. A prospective observational study was performed to assess the efficacy of maternal characteristics, biomarkers, arteriography and UADs for predicting adverse pregnancy outcomes. Previously published 1st trimester PE prediction models were validated using data collected from the observational study. A systematic review on the effect of diet and life style based metabolic risk modifying interventions on PE was performed. Results The review of biomarkers found that abnormal levels were particularIy associated with early onset PE. The stillbirth review demonstrated a three-four fold increased risk of still birth with abnormal UAD. 1045 women were included for analysis in the prospective observational study. Our models' detection rate (false positive rate of 15%) was 72% for PE; 48% PIH; 30 % SGA < 10th centile; 57% SGA < 5th centile and 67% stillbirth. In the validation study the observed discrimination ability in the derivation studies ranged from 0.70 to 0.954. When validated against the study cohort, the AUC varied importantly, ranging from 0.504 to 0.833. Dietary interventions were shown to reduce the risk of PE by 33%, with no reduction in risk with mixed interventions or fatty acid supplementation. Conclusion The high heterogeneity of studies in the systematic reviews makes it difficult to draw firm conclusions regarding the use of biomarkers or UADs in screening for pregnancy complications. Our prospective study showed a role for haemodynamics as part of routine 1st trimester screening for assessing the risk of hypertensive disease in pregnancy.

Gestational diabetes (GDM) is a potentially serious disorder requiring timely diagnosis and management to prevent adverse maternal and fetal outcomes. Of increasing concern today, when treating the woman with GDM, is the need to provide every woman with an intensive management plan to optimise the likelihood of favourable pregnancy outcomes. Early identification of those women with GDM who require insulin therapy in addition to diet therapy would be beneficial in the planning and standardisation of clinical management protocols, to enhance pregnancy outcomes and increase cost benefits with improved allocation of resources. The aim of this study was to evaluate the ability of the fasting plasma glucose level (FPG) at diagnosis to predict an increased risk to the fetus and the need for insulin therapy in a pregnancy complicated by GDM. A prospective longitudinal study design and recruitment by convenience sample was used. Data were obtained from 327 women and their babies. Diagnosis of GDM was made by a 75 gram oral glucose tolerance test (OGTT) using Australasian Diabetes in Pregnancy Society (ADIPS) criteria with the exception of seven women diagnosed on a blood glucose level >11.1mmol/l. Following consent of the women data were collected by a self report questionnaire and the medical record system at three points; at first intervention, following delivery and at the postpartum OGTT. Demographic, social, medical, maternal and neonatal outcome data were collected. The management protocol was similar for all of the women. Following nutritional intervention any woman who could not meet the glycemic targets of <= 5mmol/l fasting and/or <= 6.5mmol/l two hours postprandial was commenced on insulin therapy. The women had a mean age of 32 years, body mass index (BMI) of 25.7 and parity of 2 (range 1-12). Diagnosis was made at an average of 30 weeks and 70 women required insulin therapy with a mean dose of 34 IU per day, commencing at a mean of 31 weeks gestation. Mean birthweight was 3400G. Of the babies 12% were >4000G. Congenital abnormalities occurred in 3%, neonatal morbidities in 2% and there was 1 death in utero. Logistic regression analysis found the following significant associations: Increasing maternal BMI was related to increasing FPG levels at diagnosis and the requirement of higher insulin doses. There was a negative linear relationship to weight gain. Ethnicity was associated with maternal BMI and ethnicity with BMI was associated with birthweight in the specific ethnic group. BMI with insulin therapy as a covariate and the FPG value at OGTT were predictive of persistent glucose intolerance in 14% of women postpartum. Each value of the OGTT was a significant predictor of the need for insulin therapy as a function of the week of gestation. The FPG level was the statistical model of best fit. A 50% probability for requiring insulin was reached with a FPG at diagnosis of 4.0 mmol/l if tested at 10 weeks gestation, 5.1mmol/l at 20 weeks and 6.1 mmol/l at 30 weeks (p<.001). These results support the substantive research aim of the study. The model has the power to predict the probability (risk) of requiring insulin therapy based on the maternal FPG level at the OGTT according to the week of gestation. The study results demonstrate that glucose intolerance is linked to a number of adverse maternal and fetal outcomes in a continuous and graded fashion. The degree of reversibility of maternal and fetal risk through therapeutic interventions such as nutrition therapy, blood glucose monitoring, exercise and active patient participation aimed at improving glucose tolerance is unknown. Therefore, the rationale for, and feasibility of, new treatment strategies such as the application of this statistical model as a management approach require large scale randomised intervention studies, oriented toward measuring maternal and fetal outcomes amongst different populations.

Philosophiae Doctor - PhD / The aim of the study was to detect the prevalence of AV and its associated bacteria with preterm delivery in the Western Cape, South Africa. Furthermore, it sought particularly to examine and investigate the predictive value of GBS and E. faecalis for preterm delivery (PTD). It also aimed to establish other factors which may predict adverse pregnancy outcomes. Three hundred and one pregnant women were recruited from four different antenatal in the Western Cape, South Africa. The study conformed with the Declaration of Helsinki (2013). Maternal data was collected from a questionnaire and maternal medical records. Vaginal and rectal swabs were collected and microscopically examined for AV, followed by culture characterization of GBS and E. faecalis. Antimicrobial susceptibility testing was also performed. In this study, AV was detected in 79 (26.2%) of the 301 pregnant women, and GBS and E. faecalis isolated from 50 (16.6%) and 21 (7.0%) respectively. GBS serotype V was the predominant serotype, followed by serotype III. Pulse field gel electrophoresis (PFGE) profile analysis for both GBS and E. faecalis yielded a total of 24 restrictions profiles for GBS and 16 for E. faecalis. Multivariable analysis revealed that parity, gravidity, vaginal discharge, urinary tract infection, and smoking were significantly associated with PTD. The results from the study provides improved guidelines maternal screening of pregnant women. The early detection of AV-related bacteria may significantly reduce maternal and neonatal morbidity.

This thesis aims to undertake health technology assessments in high risk pregnancies through the following objectives: 1. In women with pre-eclampsia, a) To evaluate the association of maternal genotype and severe pre-eclampsia b) To assess the accuracy of tests in predicting adverse pregnancy outcomes c) To develop composite outcomes for reporting in trials on late onset pre-eclampsia 2. In women with multiple pregnancy, a) To study the association between chorionicity and stillbirth b) To identify the optimal timing of delivery in monochorionic and dichorionic twin pregnancies 3. In the field of prediction research in obstetrics a) To provide an overview of the existing prognostic models and their qualities b) To evaluate the methodological challenges and potential solutions in developing a prognostic model for complications in pre-eclampsia Methods The following research methodologies were used: Delphi survey, systematic reviews and meta-analyses. Results 1. a) Maternal genotype and severe pre-eclampsia: 57 studies evaluated 50 genotypes; increased risk of severe pre-eclampsia with thromobophilic genes. b) Accuracy of tests in predicting pre-eclampsia complications: 37 studies evaluated 13 tests. No single test showed high sensitivity and specificity. c) Delphi survey of 18/20 obstetricians and 18/24 neonatologists identified clinically important maternal and neonatal outcomes and maternal and neonatal composite outcomes were developed. 2. Prospective risk of stillbirth and neonatal deaths in uncomplicated monochorionic and dichorionic twin pregnancies: 32 studies were included. In dichorionic twin pregnancies, the risk of stillbirths was balanced against neonatal death at 37 weeks' gestation. In monochorionic pregnancies, there was a trend towards increase in stillbirths after 36 weeks but this was not significant. 3. a) From 177 studies included, 263 obstetric prediction models were developed for 40 different outcomes, most commonly pre-eclampsia, preterm delivery, mode of delivery and small for gestational age neonates. b) The obstetric prognostic model challenge of dealing with treatment paradox was explored and seven potential solutions proposed by expert consensus. Conclusion I have identified the strength of association for genes associated with complications in pre-eclampsia, components for composite outcomes for reporting in studies on pre-eclampsia, and the optimal timing of delivery for twin pregnancies. My work has highlighted the gaps in prediction research in obstetrics and the limitations of individual tests in pre-eclampsia.

Objective: Short interpregnancy interval (IPI), <18 months between pregnancies, is a potential cause of adverse delivery and birth outcomes, and may be a particular issue among those with other risks. Our goal was to examine IPI and delivery/infant complications in Tennessee. Methods: Birth certificate/vital records data included 101,912 women with a previous delivery. IPI groups (<6, 6-12, 12-18, 18-60 months) were compared on outcomes. Results: Thirty-nine percent of the deliveries had IPI <18 months, 9% were <6 months, rates 11% and 27% higher than nationally. Women with IPI <18 months were younger, lower educated with lower income, had higher BMIs, and were more likely to be unmarried, smokers, and have begun prenatal care later (p <.001). In adjusted analyses, IPI <18 months predicted elevated risk for precipitous labor, low-birth weight, preterm delivery, NICU admission, and infant mortality, with effects strongest for IPI <6 months. Finally, risks related to IPI <6 months were substantially higher for the lowest income women. Conclusions: Rates of less than optimal IPI were high in this population already experiencing significant maternal-child health disparities, with short IPI a particular risk factor for poor out-comes for the most disadvantaged women, suggesting yet another precursor for adverse birth outcomes in those already most at risk.

Maternal prenatal exposure to hurricanes and tornadoes could contribute to an increased risk for adverse birth outcomes. Little is known about the effects of Hurricane Katrina of August 2005, on pregnancy outcomes in Mississippi. Additionally, little is known about the influence of the April 2011 Alabama tornado disaster on births in that state. The purpose of this study was to bridge this knowledge gap by examining the relationship between maternal prenatal exposure to these storms and adverse infant health outcomes. The theoretical framework guiding this retrospective, cross-sectional study was the life course approach. Data for this investigation included 2,000 records drawn from the Linked Infant Births and Deaths registers. Chi-square and logistic regression analyses were performed. Results indicated hurricane exposure was not a predictor of preterm birth (OR = .723, 95% CI = [.452, 1.16]; p = 1.76) or low birth weight (OR = .608, 95% CI = [.329-1.13]; p = .113). However, an association was observed between tornado exposure and preterm birth (OR = 1.68, 95% CI = [1.19-2.39]; p = < 0.05) and low birthweight (OR = 1.91, 95% CI = [1.27-2.87]; p = < 0.05). Findings suggest pregnant women are vulnerable to natural disaster storms, and are at risk for adverse pregnancy outcomes. The implications for social change include informing preparedness efforts to reduce vulnerability to increased pregnancy risk factors and adverse birth outcomes, consequential to hurricane and tornado disasters.

Doctor Pharmaceuticae - Dpharm / The use of efavirenz (EFV) in the first trimester of pregnancy remains controversial. In South Africa, the use of EFV-containing antiretroviral therapy (ART) as part of a Fixed Dose Combination (FDC) during the first trimester of pregnancy started in April, 2013. Literature to date has reported conflicting outcomes following the use of EFV-containing ART during the first trimester of pregnancy. The objectives of the study were to determine the prevalence of adverse pregnancy outcomes among HIV-positive pregnant women treated with EFV-containing ART and compare these results with those of pregnant women treated with NVP-containing ART and HIV-negative pregnant women in resource-limited settings. In addition, the study also aimed to determine the effect of the time of initiation of ART on the prevalence of adverse pregnancy outcomes.

Nigeria is one of the countries with the highest Under-5 Mortality rates (U5M) estimated at 117 deaths/1000 live births. Despite public health control initiatives, no significant improvement in U5M has been demonstrated. The purpose of the study was to determine whether history of Adverse Pregnancy Outcomes (APO) and maternal socioeconomic factors could predict the death of children before their fifth birthday, using the life course health development and fetal programming theories. The study population was women in their reproductive age (15- 49 years). The study was a secondary data analysis of the datasets obtained from three Nigeria Demographic and Health Surveys (2003, 2008, and 2013). Complex samples multivariate logistic regression was used to determine the associations among variables. The results showed that lower education level (p < 0.001), lower income level (p <0.05), rural residential setting (p< 0.01), and lower socioeconomic status index (p < 0.001) of women were statistically significant predictors of U5M. APO was not statistically associated with U5M (p > 0.05). This concludes that children of women with low socioeconomic factors and status index could be at higher risk of death within the first 5 years of their lives, and women with history of APO stand no greater risk of losing their under-5 children. The study would contribute to positive social change among women in Nigeria through early identification of women whose children may be at risk of U5M and provision of evidence-based advocacy to urge increased government and public attention to women and child welfare.

Disinfection by-products (DBPs) are a major group of water contaminants and their role in causing adverse health outcomes, including adverse pregnancy outcomes, endocrine disruption, respiratory related adverse health outcomes and cancer has been subject to extensive epidemiological and toxicological research and review. Determination of safe exposure to DBPs, particularly within drinking water supplies, has been a topic of extensive debate, with a wide range of acceptable levels set across the industrialized world. The focus of the research in this thesis was on two of the main health outcomes associated with DBP exposure, namely adverse pregnancy outcomes and asthma related symptoms. To assess adverse pregnancy outcomes in Perth, an extensive classification quantification of the major DBPs in Perth drinking water was conducted. A registrybased prevalence study was carried out to assess birth defects in relation to high, medium and low DBP areas (defined by the water sampling and analysis). It was found that women living in high THM areas are 22% (odds ratio (OR) 1.22, 95% confidence interval (95% CI) 1.01-1.48) more likely of having a baby with any birth defect. High exposure was also strongly associated with an increased risk of having a baby with a cardiovascular defect (62% increased risk). Low birth weight and prematurity were also assessed; however these outcomes were not associated with an increased risk through an increase in exposure. Following on from this analysis, a population risk assessment model was developed for DBPs in high exposure environments. This involved a three step process: (i) Firstly a questionnaire-based validation and reliability study was used to assess water consumption patterns of a population of pregnant women in Perth. (ii) Secondly a prediction model for teratogenic burden of DBPs in Perth was developed, related to the exposure patterns of the population of pregnant women involved in the validation and reliability study. (iii) Finally, combining the information collected in (i) and (ii), along with the regression slope estimates for birth weight from the prevalence study (defined in Section 2.2.1), a dose-response model for THMs and birth weight was developed. Predictive simulations for birth weights at given THM levels were then conducted. It was estimated that pregnant women in Perth are exposed to between 0.3 – 4.10 µg/day ingested TTHM, and of this, the more toxic brominated forms accounted for between 0.27 – 3.69 µg/day. Based on a dose-response model used, birthweights calculated for the ‘hypothetical’ exposures ranged from 3403.2g for the highest exposure to 3503.5g in the lowest exposure, which is a difference of over 100g. Although the resulting reduction in birth weight is not extreme, there is still a significant reduction in birth weight present as exposure to TTHMs increases. This is the first doseresponse model to be developed to assess an adverse pregnancy outcome based on pregnant women exposure data, and will be a useful tool for assessing varying exposures throughout not only Australia but also throughout the industrialised world, where DBP exposure is highly prevalent.

Les pathologies vasculaires placentaires sont fréquentes et graves. La forme maternelle prédominante est la pré-éclampsie et la forme fœtale le retard de croissance intra-utérin. Les questions posées autour de ce sujet concernent tout d'abord la prédiction de la survenue de ces pathologies suffisamment tôt afin de permettre une surveillance rapprochée, une administration de corticoïdes et une prise en charge dans une maternité de niveau adaptée. La prévention de la survenue et de la récidive ainsi que le traitement de ces pathologies la phase constituée sont aussi des problématiques encore non résolues. Notre objectif était de travailler sur ces différentes questions pa l'intermédiaire de trois études : l'étude ANGIOPRED), l'étude VOLUPLA et l'étude GROWTH. Les résultats de ces travaux et une revue d la littérature mettent en évidence une perturbation des facteurs de l'hémostase et des facteurs angiogéniques dans la pré-éclampsie et dans le retard de croissance. L'association des facteurs maternels, échographiques, angiogéniques et sériques constitue un modèle prédictif efficace principalement du fait d'une excellente valeur prédictive négative. Le volume placentaire est corrélé au taux de D- Dimères et est intéressant pour la prédiction des pathologies vasculaires placentaires. De nouveaux travaux devront poursuivre l'étude d la prédiction, de la prévention et du traitement des pathologies liées au placenta. Le traitement est notamment l'objet de l'étude Growth qui vise à évaluer l'efficacité de l'énoxaparine dans le traitement du retard de croissance vasculaire constitué. / Placenta-mediated adverse pregnancy outcomes are frequent and severe pathologies whose predominant maternal form is preeclampsia and fetal form, intrauterine growth retardation. The questions asked about this subject concern first of all the prediction of the occurrence of its pathologies in a sufficiently early way to allow for close monitoring, administration of corticosteroids, and management in an appropriate level of maternity. The prevention of the occurrence and recurrence and the treatment of its pathologies in the constituted phase are also unresolved problems. Our objective was therefore to work on its various questions through three studies: the ANGIOPREI study, the VOLUPLA study and the GROWTH study. The results of his work and of the literature show that the factors of haemostasis anc angiogenic factors are disturbed in preeclampsia and in growth retardation. The association of maternal, ultrasound, angiogenic and serum factors constitutes a predictive model that is effective mainly by an excellent negative predictive value. The placental volume is correlated with the D-dimer level and is interesting for placenta-mediated adverse pregnancy outcomes prediction. New studies will have to continue the exploration of the prediction, prevention and treatment of this pathologies related to the placenta. The treatment is notably the object of the study Growth which aims to evaluate the effectiveness of the Enoxaparin for the treatment of constituted vascular growt retardation.