Barcelona — Research presented at the European Society of Cardiology Congress highlights a breakthrough in pregnancy-related heart failure detection, facilitated by an artificial intelligence (AI)-enabled digital stethoscope. The study, conducted by the Mayo Clinic and published in Nature Medicine, reveals that the AI-enhanced tool identified twice as many cases of heart failure compared to traditional obstetric care.
The trial, conducted in Nigeria—where pregnancy-related heart failure rates are among the highest globally—demonstrated that the AI-enabled stethoscope was 12 times more likely than conventional screening methods to detect heart pump weakness when evaluated at an ejection fraction threshold of less than 45%. This threshold is critical for diagnosing peripartum cardiomyopathy, a severe form of heart failure that can develop during or after pregnancy.
Dr. Demilade Adedinsewo, MD, a cardiologist at Mayo Clinic and the lead investigator of the study, emphasized the importance of early detection. “Peripartum cardiomyopathy can worsen as pregnancy progresses or following childbirth, potentially endangering the mother’s life if her heart becomes too weak. Early identification allows for timely treatment, which can range from medication to more intensive interventions like a mechanical heart pump or even a heart transplant in severe cases.”
The study involved nearly 1,200 participants and compared traditional obstetric care with AI-enhanced screening methods. The AI-enabled digital stethoscope, developed by Mayo Clinic in collaboration with Eko Health, builds on a foundational 12-lead AI-electrocardiogram (ECG) algorithm previously created to predict weak heart function. This enhanced version, FDA-cleared for detecting heart failure with low ejection fraction, significantly outperformed standard screening methods.
Findings indicate that the digital stethoscope, combined with the AI-based 12-lead ECG, detected weak heart function with high accuracy. It identified twice as many cases of low ejection fraction (below 50%) and was 12 times more effective at detecting an ejection fraction below 45% compared to usual care.
The study evaluated the AI tools across three levels of ejection fraction used in clinical diagnoses: less than 45% for peripartum cardiomyopathy, less than 40% for heart failure with reduced ejection fraction, and less than 35% for severely low heart pump function. Each participant in the intervention group underwent an echocardiogram to confirm the AI predictions.
Dr. Adedinsewo noted, “This study demonstrates that AI technology can significantly improve the detection of peripartum cardiomyopathy in Nigeria. However, further research is needed to assess the tool’s usability and adoption by healthcare providers in Nigeria and its overall impact on patient care. Given that peripartum cardiomyopathy affects approximately 1 in 2,000 women in the U.S., and up to 1 in 700 African American women, testing this AI tool in diverse populations and healthcare settings will be crucial for evaluating its broader applicability.”
