IPPIC 

Full title: Accuracy of clinical characteristics, biochemical and ultrasound markers in the prediction of preeclampsia: an Individual Patient Data (IPD) Meta-analysis

Project Title: : International Prediction of Pre-eclampsia IPD Collaborative Network (IPPIC)

Acronym: IPPIC

Research Funder: Health Technology Assessment Programme (HTA)

Research status: Started

OVERVIEW

Pre-eclampsia, a condition in pregnancy with raised blood pressure and protein in the urine is a major cause of complications in the mother and baby. There is emerging evidence that pre-eclampsia is not a single disease but a syndrome which can arise from different causal pathways. When pre-eclampsia occurs early in pregnancy, before 34 weeks of gestation, it is more severe than late onset disease and contributes disproportionately to adverse maternal and fetal outcomes. The treatment for the condition is delivery of the baby. In women with early onset pre-eclampsia, mothers are often delivered early to improve their condition, contributing to prematurity associated complications including death, and long term neurological disability in the children. This leads to significant costs to the NHS in caring for the preterm baby and societal costs for their long-term care.

Early identification of mothers at risk of pre-eclampsia, especially early onset, will allow appropriate targeted management of mothers at risk. This includes frequent monitoring and commencement of aspirin early in pregnancy. There is no single test that can accurately predict the risk of pre-eclampsia in pregnant women. Current national and international guidelines provide a list of risk factors based on clinical characteristics of the mother to assess their risk with limited accuracy. Many studies have found an association between abnormal biochemical tests in first and second trimester and onset of pre-eclampsia. Abnormal ultrasound findings on blood flow to the womb of the mother have also been shown to have some accuracy in detecting mothers at risk. However, the performance of these individual tests is not sufficient to warrant their implementation in routine clinical practice.

The HTA call is for synthesis of available evidence to identify the most accurate tests, separately and in combination in the prediction of pre-eclampsia, including the early onset type. We have identified over 50 published evidence synthesis projects on this topic, and they are unable to provide clear conclusions on the performance of the tests due to the limitations in the published data. We therefore propose to obtain the individual data of all participants in relevant studies, through our International Prediction of Pre-eclampsia IPD Collaborative (IPPIC) Network. The Network comprises of researchers involved in studies on this topic and we have the support of 61 researchers, with access to data from over 400,000 women to-date.

Access to the individual data will allow us to take into account multiple factors that predict risk of preeclampsia and develop a scoring system (prediction model) to provide women with their individualised risk. Furthermore, our approach will allow us to test the performance of this scoring system to ensure that it is reliable.

The team comprises of researchers experienced in evidence synthesis, particularly use of IPD to collate the evidence, expertise in the area of pre-eclampsia, methodological experts in prediction studies and patient representatives. APEC (Action Against Pre-eclampsia) is a charity representing and helping women affected by pre-eclampsia, and are partners in our proposal and will provide input from project design to dissemination of findings.

STAFF

Investigators

Chief investigator:

Professor Shakila Thangaratinam, Professor of Maternal and Perinatal Health, Women’s Health Research Unit, Multidisciplinary Evidence Synthesis Hub (mesh), Email: This email address is being protected from spambots. You need JavaScript enabled to view it.

Co-investigator:

Dr Julie Patricia Dodds,Senior Clinical Trials Manager, Women’s Health Research Unit, Barts & The London Queen Mary’s School of
Medicine & Dentistry, Email: This email address is being protected from spambots. You need JavaScript enabled to view it.

Ms Ewelina Rogozinska, IPD Project Coordinator, Centre for Primary Care and Public Health, Queen Mary University of London Barts and The London School of Medicine and Dentistry Email: This email address is being protected from spambots. You need JavaScript enabled to view it.

Dr Jenny Myers, Clinical Senior Lecturer, Maternal & Fetal Health Research Centre, The University of Manchester Email: This email address is being protected from spambots. You need JavaScript enabled to view it. 

Professor Lucilla Poston, Head of Division of Women’s Health, Maternal and Fetal Research Unit, King’s College London, Email: This email address is being protected from spambots. You need JavaScript enabled to view it.

Dr Lucy Chappell, Reader in Obstetrics, Women’s Health Academic Centre, King’s College London, Email: This email address is being protected from spambots. You need JavaScript enabled to view it.

Professor Khalid Khan, Prof of Women’s Health and Clinical Epidemiology,Queen Mary University of London Barts and The London School of Medicine and Dentistry, Email:This email address is being protected from spambots. You need JavaScript enabled to view it.

Professor Asif Ahmed, Pro-Vice-Chancellor for Health, Professor of Vascu, Aston University, Email:This email address is being protected from spambots. You need JavaScript enabled to view it.

Professor Baskaran Thilaganathan, Professor and Director Fetal Medicine Unit, St Georges, University of London, This email address is being protected from spambots. You need JavaScript enabled to view it.

Dr Richard Riley, Professor of Biostatistics, Institute of Primary Care and Health Sciences Keele University, Email: This email address is being protected from spambots. You need JavaScript enabled to view it.

Dr Asma Khalil, Senior Lecturer in Fetal Medicine and Obstetrics, St George’s University of London, Email: This email address is being protected from spambots. You need JavaScript enabled to view it.

Professor Ben Mol, The University of Adelaide, Email: This email address is being protected from spambots. You need JavaScript enabled to view it.

Professor Louise Kenny, University College Cork, Email: This email address is being protected from spambots. You need JavaScript enabled to view it.

Mr John Allotey, Trial Co-ordinator, Centre for Primary Care and Public Health, Queen Mary University of London Barts and The London School of Medicine and Dentistry Email: This email address is being protected from spambots. You need JavaScript enabled to view it.

Mrs Ann Marie Barnard, Action on Pre-eclampsia (APEC), Email:This email address is being protected from spambots. You need JavaScript enabled to view it.

Professor Peter von Dadelszen, Associate Professor of Obstetrics and Gynaecology, Obstetrics and Gynaecology University of British Columbia, Email: This email address is being protected from spambots. You need JavaScript enabled to view it.

Professor Karel Moons, Professor of Clinical Epidemiology, Julius Center for Health Sciences and Primary Care, UMC Utrecht, This email address is being protected from spambots. You need JavaScript enabled to view it.

OTHER INFORMATION

Scientific Summary

Aims and Objectives:

We will develop, externally validate and update separate prediction models for (i) early (<34 weeks’ gestation), (ii) late (>= 34 weeks) and (iii) any onset pre-eclampsia.

Primary

  1. To estimate the prognostic value of individual clinical, biochemical and ultrasound markers for predicting pre-eclampsia by IPD meta-analysis
  2. To validate, and improve or tailor the performance of existing models in relevant population groups, for predicting early, late and any onset pre-eclampsia in our IPD dataset based on

- Clinical characteristics only

- Clinical and biochemical markers

- Clinical and ultrasound markers

- Clinical, ultrasound and biochemical markers

  1. Using IPD meta-analysis, to develop and externally validate (using internal-external cross validation) multivariable prediction models for early, late and any onset pre-eclampsia in the following circumstances: existing predictive strategies cannot be adjusted for the target population, no such models exist, or the relevant pre-eclampsia outcomes are not studied.

Secondary

  1. To assess the differential performance of the models in various predefined subgroups based on population characteristics (unselected; selected) and timing of model use (first trimester; second trimester)
  2. To study the added role of novel biomarkers on the accuracy of the developed models

Scientific Abstract:

Pre-eclampsia remains a major cause of maternal, fetal, and neonatal mortality and morbidity. Preeclampsia is a syndrome: Early onset disease, occurring before 34 weeks’ gestation, is more severe, and is considered to have a different pathophysiology than the late onset disease. It is unlikely that a single model will accurately predict both early and late onset disease. The HTA brief calls for a systematic review on the predictive accuracy of markers, separately and in combination, and of models for predicting pre-eclampsia, including early onset disease. Our review of reviews has identified over 50 published systematic reviews to-date, on the accuracy of clinical characteristics, biochemical and ultrasound markers for predicting pre-eclampsia, including our own HTA report (No. 01/64/04); and we have identified a further 69 risk prediction models for pre-eclampsia, with inadequate external validation.

Clinical applicability based on the findings of aggregate meta-analyses is limited due to the observed heterogeneity in population, in combinations of predictors, in outcome definitions (e.g. most published models focussed on any pre-eclampsia and not on the much more clinically relevant early onset preeclampsia) and by the lack of robust methods for aggregating data of published models. The Brief particularly calls for prognostic model development if appropriate, and to explore its use for screening pre-eclampsia. Aggregate meta-analysis is not an appropriate method to achieve these objectives.

Furthermore, prior to the use of prediction models in clinical practice, there is a need to successfully validate the model in multiple datasets external to the model development phase. This often takes many years to accomplish in a primary study.

Individual Participant Data (IPD) meta-analysis can overcome many of the above limitations by accessing the raw data of the individual participants. A large scale IPD meta-analysis will enable us to predefine the desired clinically relevant endpoints (e.g. timing of pre-eclampsia onset) and cut off values of clinical parameters; standardise the definitions of predictors and outcomes; take into account the performance of many candidate prognostic variables; directly handle missing data on both predictors and outcomes; account for heterogeneity in baseline risks; and most importantly, to develop, validate and tailor the use of the most accurate prediction models to the appropriate population. By allowing for the clustering within studies, the developed model will avoid performance deterioration encountered in aggregate meta-analysis, when the individual’s baseline risk is different from the average estimated during model development.

We have established an International Prediction of Pre-eclampsia IPD Collaborative Network (IPPIC) of global researchers with access to the IPD from existing studies and large databases (>400,000 women). Our collaborative approach achieves consensus towards well developed, and externally validated prognostic models.

Using our large IPD dataset, we will assess the performance of clinical characteristics alone or in combination with biochemical and ultrasound markers for the prediction of early, late and any onset pre-eclampsia. We will develop, externally validate and if required update separate prediction models for the above outcomes. Additionally, we will study the added value of novel biomarkers to the performance of these models.