fullPIERS (Pre-eclampsia Integrated Estimate of RiSk): external validation and recalibration
The fullPIERS model was developed and internally validation in a cohort of 2023 women with pre-eclampsia admitted to tertiary perinatal units in Canada, the UK, New Zealand and Australia. With good stratification capacity and predictive performance (AUC ROC: 0.88), the fullPIERS model needs to be validated externally, and then recalibrated, using external datasets.
Using the 2012 CIHR Knowledge Translation Award to Peter von Dadelszen, Vivian Ukah, herself a recipient of a UBC Four Year Fellowship (FYF), has been recruited to read for her PhD by undertaking the external validation and recalibration of the fullPIERS model, including and excluding translational biomarkers.
Following baseline recalibration of the fullPIERS model using the initial dataset with novel methodology developed by Beth Payne primarily for her miniPIERS-focussed PhD thesis, the plan is to update the BC Women’s Hospital cohort of women admitted with any hypertensive disorder of pregnancy, and to import data from extant datasets that will inform the model.
The integration of placental growth factor (PlGF) into fullPIERS will become a priority should the Brazilian Grand Challenges preterm birth operating grant be awarded (see Objective Four), as that study will use both PlGF and fullPIERS to risk stratify women with pre-eclampsia in that study.
Data collection of external, relevant datasets has been completed. These include the PREP UK study, and the FINNPEC Finland data. We have also received partial data from Alere-funded PELICAN and the CHIPS multicentre study. Ethical approval and data sharing agreements are almost complete to allow access to the Alere-funded PETRA, and the John Radcliffe Hospital, Oxford, UK datasets.
For the extension of the fullPIERS model with biomarker, we will be incorporating PlGF to the model using the PELICAN and PETRA studies which have measured maternal blood levels of PlGF. We hypothesise that PlGF will improve the identification of women at risk of complications from pre-eclampsia. This has been supported by our recent finding using data on PlGF collected in Mozambique from women with suspected pre-eclampsia. In that study, low PlGF was associated with adverse outcomes such as stillbirth, Caesarean delivery and delivery within seven and fourteen days of bPlGF testing. The manuscript for this study has been submitted for publication.
Once we finalize data transfer from the Alere PETRA and PELICAN, we will assess any improved prediction capacity upon the addition of PlGF to the original fullPIERS model. In addition, a preliminary temporal validation of the fullPIERS model has been carried out using data collected from the BC Women’s Health Centre. The fullPIERS model had good performance, AUC ROC of 0.84 (95% CI 0.78–0.90) and a moderate calibration.
We plan to finalise all data analyses and writing of manuscripts. Expected publications include:
- The diagnostic performance of placental growth factor in women with suspected pre-eclampsia attending antenatal facilities in Maputo, Mozambique.
- Temporal validation of the fullPIERS model using data collected from the BC Women’s Hospital, Canada.
- External validation of the fullPIERS model using the combined PREP, FINNPEC, Alere-funded PETRA and PELICAN and Oxford datasets.
- Addition and assessment of PlGF in the fullPIERS model.
- Recalibration of the fullPIERS model for better performance, in the context of a methodology paper, using external datasets.