A prediction model for aberrant cardiac remodelling in women after preeclampsia; Queen of Hearts study

Abstract

Abstract Background Preeclampsia, a hypertensive pregnancy disorder, is strongly associated with aberrant left ventricular remodelling up to ten years after delivery and heart failure later in life. To detect occult cardiac abnormalities, echocardiography is necessary but not feasible for all these women. A prediction model for aberrant cardiac remodelling to be used in primary care may guide towards intensified follow-up for those at risk or tempered follow-up for low-risk individuals. Purpose To develop a prediction model for aberrant left ventricular remodelling in former pregnant women weighing their complicated pregnancy. Methods In this large cohort study, we included women (aged ≥18 years) with a history of preeclampsia or normotensive pregnancy at a postpartum interval of 6 months until 30 years. Comprehensive cardiovascular assessment was performed, including echocardiography, 30-minutes blood pressure measurements and circulating biomarkers. The procedures were in accordance with institutional guidelines and adhered to the principles of the Declaration of Helsinki. Aberrant cardiac remodelling based on echocardiography was defined as either left ventricular concentric remodelling or left ventricular hypertrophy. A prediction model based on clinical and circulating markers was developed using univariable and multivariable logistic regression with backward elimination. Internal validation was performed using Heuristic shrinkage factor. Performance of the final model was evaluated in terms of discrimination by the area under the receiver operating characteristic (AUC-ROC) curve. Results We included 1466 women, of which 93 (6.3%) with left ventricular concentric remodelling and/or left ventricular hypertrophy (mean ± SD age, 43±10 years) and 1373 (93.7%) without left ventricular concentric remodelling and/or left ventricular hypertrophy (40±8 years). The final prediction model was based on the predictors age, waist circumference, systolic blood pressure, Homeostatic Model Assessment of Insulin Resistance (HOMAIR), and preeclampsia (y/n) (Table 1). After internal validation, this prediction model showed a good discriminative ability of AUC-ROC-curve 0.71 (95% CI 0.661–0.758) (Figure 1). Conclusion We developed a good-performing prediction model for aberrant cardiac remodelling in former pregnant women based on five conventional variables that could contribute to risk-guided follow-up after preeclampsia aiming at intensifying follow-up for those at risk and tempered follow-up for low-risk women. Funding Acknowledgement Type of funding sources: None.