Predicting cardiac and pregnancy outcomes in women with adult congenital heart disease using the Anatomic and Physiological (AP) Classification System: How much does physiology matter?

Background: Pregnancy in women with congenital heart disease (CHD) is associated with an increased risk of adverse maternal and fetal events. Despite the physiological impact of CHD on pregnancy, current risk stratification scores primarily consider anatomical lesions. We assessed the performance of the novel American Heart Association Anatomic and Physiological (AP) classification system in predicting adverse maternal cardiac, obstetric and fetal events, and compared it with established risk models.Methods: This retrospective cohort study enrolled pregnant women with CHD managed by the Westmead Hospital high-risk pregnancy team. Preconception risk stratification scores (AP classification, mWHO classification, CARPREG II and ZAHARA scores) were retrospectively assigned to each pregnancy by an adult CHD cardiologist and compared with the primary outcome measures, which were maternal cardiac, obstetric and fetal complications.Results: We analysed 176 pregnancies in 120 women with CHD. Maternal cardiac risk significantly increased between AP class 2 and 3 (p = 0.001). Within class 3, higher physiological status correlated with maternal cardiac events (p < 0.001). Increasing AP severity correlated with lower fetal birthweight percentiles (p = 0.003). The AP classification was similar to mWHO at predicting maternal cardiac outcomes (AUC 0.787 vs 0.777, p < 0.001), but the CARPREG II (AUC 0.852, p < 0.001) and ZAHARA scores (AUC 0.864, p < 0.001) had higher discriminatory ability within our cohort.Conclusion: The AP classification system shows non-inferior preconception maternal cardiac risk prediction compared to current validated scores. Consideration of physiological status has additive predictive value in the most complex patients (Stage III). Prospective, multicenter studies are required for further validation for preconception risk estimation.