BACKGROUND/OBJECTIVE: Several risk scores (RSs) have been used to stratify risk of cardiac complications (CCs) in pregnant patients with heart disease. We aimed to compare and contrast the accuracy of several RSs for predicting CC in this population.
METHODS: Prospective inclusion of all consecutive pregnant patients with heart disease, and follow-up until 6 months postpartum. CCs were defined as primary if admission was required due to heart failure, arrhythmia or thromboembolic events, and secondary if the decline in NYHA class compared with baseline was >2 or urgent invasive cardiac procedures were needed. The discriminatory power of each RS was assessed by the area-under-the receiver-operating characteristic (ROC) curve (AUC).
RESULTS: 179 patients, mean age: 32 years, accounted for 13.4% of CC (primary 11.7%, secondary 1.7%); the main diagnosis was congenital heart disease (CHD) in 68% followed by valvulopathies in 16%, arrhythmia in 7% and myocardiopathies in 5%. 22% (n=40) were classified as mWHO=1, 59% (n=105) mWHO=2 including subgroup 2-3, 14% (n=26) mWHO=3 and 4%(n=7) mWHO=4; 1 patient was unclassifiable. mWHO showed a better AUC (0.763) than CARPREG (0.67). For the CHD population, ZAHARA RS showed an AUC of 0.74, and Khairy an AUC of 0.632.
CONCLUSIONS: mWHO was better at predicting CC than CARPREG; mWHO was also better at predicting CC than the specific CHD RS in the CHD subgroup.
PRACTICE: There are an increasing number of pregnant women with HD.
IMPLICATIONS: Improved prediction of CC risk during pregnancy can provide better preconception assessment in women with HD.
