No Cover Image

Journal article 94 views 25 downloads

Development and validation of a postpartum cardiovascular disease risk prediction model in women incorporating reproductive and pregnancy-related predictors

Steven Wambua, Francesca L. Crowe, Shakila Thangaratinam, Dermot O’Reilly, Colin McCowan, Sinead Brophy Orcid Logo, Christopher Yau, Krishnarajah Nirantharakumar, Richard D. Riley, Kym I. E. Snell, (the MuM-PreDiCT Group)

BMC Medicine, Volume: 23, Issue: 1

Swansea University Author: Sinead Brophy Orcid Logo

  • 70563.VoR.pdf

    PDF | Version of Record

    © The Author(s) 2025. This article is licensed under a Creative Commons Attribution 4.0 International License.

    Download (1.23MB)

Check full text

DOI (Published version): 10.1186/s12916-025-04229-1

Abstract

BackgroundEach year, over 700,000 pregnancies occur in the UK, with up to 10% affected by complications such as hypertensive disorders of pregnancy and gestational diabetes mellitus. Pregnancy-related complications and reproductive factors are associated with an increased risk of cardiovascular dise...

Full description

Published in: BMC Medicine
ISSN: 1741-7015
Published: Springer Science and Business Media LLC 2025
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa70563
Abstract: BackgroundEach year, over 700,000 pregnancies occur in the UK, with up to 10% affected by complications such as hypertensive disorders of pregnancy and gestational diabetes mellitus. Pregnancy-related complications and reproductive factors are associated with an increased risk of cardiovascular disease (CVD) later in life. Our aim was to determine whether adding pregnancy factors to a prediction model with established CVD risk factors improves 10-year risk prediction of CVD in postpartum women, using QRISK®-3 as a benchmark model.MethodsWe used a population-based retrospective cohort of women aged 15 to 49 who had been pregnant from the Clinical Practice Research Datalink (CPRD) primary care database. Women who were CVD-free were followed from 6 months postpartum. We evaluated the performance of QRISK®-3 and updated the risk prediction model using established risk factors for CVD from QRISK®-3 and additional risk factors specific to pregnancy. Models were developed using Cox-proportional hazards regression for CVD within 10 years. Models were evaluated and compared using measures of overall model fit, calibration, discrimination and clinical utility.ResultsAmong 567,667 eligible women, 2175 (0.38%) experienced a CVD event within 10 years. The median follow-up was 4 years. Of the additional pregnancy factors, gestational hypertension, preeclampsia, miscarriage, stillbirth, postnatal depression, gravidity, endometriosis and polycystic ovary syndrome remained associated with CVD after adjusting for other established risk factors of CVD. Adding pregnancy factors to those from QRISK®-3 led to marginal improvements in model performance (QRISK®-3 C-statistic: 0.703 (95% CI 0.687 to 0.718), updated model C-statistic: 0.726 (95% CI 0.711 to 0.740) Although calibration did not improve overall, subgroup analysis showed better calibration in women with a history of pre-eclampsia, postnatal depression and preterm birth using the updated model. The clinical utility was improved for updated models.ConclusionsThe updated risk prediction models resulted in marginal improvement in discrimination and calibration compared to QRISK®-3 in postpartum women. This could be due to the known association of pregnancy-related complications with established risk factors of CVD. Although the overall predictive performance and calibration of the updated models was similar, the updated model resulted in better clinical utility.
Keywords: Risk prediction; Pregnancy complications;; QRISK®-3; Cardiovascular disease
College: Faculty of Medicine, Health and Life Sciences
Funders: This work is funded by the Strategic Priority Fund “Tackling multimorbidity at scale” programme (grant number MR/W014432/1) delivered by the Medical Research Council and the National Institute for Health Research in partnership with the Economic and Social Research Council and in collaboration with the Engineering and Physical Sciences Research Council. SW PhD studentship is funded by the British Heart Foundation (BHF) Data Science Centre (BHF grant number SP/19/3/34678, awarded to Health Data Research (HDR)). His PhD is also supported through the HDR-UK-Turing Wellcome PhD Programme. RR and KS are supported by funding from the NIHR Birmingham Biomedical Research Centre at the University Hospitals Birmingham NHS Foundation Trust and the University of Birmingham.
Issue: 1

Similar Items