AI-powered stethoscope doubles success in diagnosing heart failure during pregnancy
Heart failure during pregnancy is a life-threatening and often overlooked condition, primarily because its symptoms—such as shortness of breath, severe fatigue, and difficulty breathing while lying down—closely resemble typical discomforts associated with pregnancy. This confusion often leads to delayed diagnosis. However, a groundbreaking study presented at the European Society of Cardiology Congress, based on research by the Mayo Clinic, demonstrates that an artificial intelligence (AI)-enabled digital stethoscope allowed healthcare providers to diagnose twice as many cases of heart failure compared to conventional obstetric care methods. The full study has been published in Nature Medicine.
This clinical trial was conducted in Nigeria, where pregnancy-related heart failure occurs more frequently than in any other region worldwide. The findings revealed that the AI-enabled stethoscope was 12 times more likely to detect weakened heart function—specifically in cases with an ejection fraction of less than 45%—compared to traditional methods. An ejection fraction under 45% is a critical indicator of peripartum cardiomyopathy, a form of heart failure that can develop during the final months of pregnancy or soon after childbirth.
“Early detection of this form of heart failure is crucial for safeguarding maternal health and wellbeing,” explained Dr Demilade Adedinsewo, a cardiologist at the Mayo Clinic and the lead investigator of the study. “Symptoms of peripartum cardiomyopathy can progressively worsen as pregnancy advances, or more commonly after childbirth. If left undiagnosed and untreated, this condition can become life-threatening as the heart weakens further. Although medications can help when detected early, severe cases may necessitate advanced interventions, including intensive care, mechanical heart pumps, or even heart transplants in extreme situations.”
The randomised, controlled, open-label clinical trial involved nearly 1,200 participants. Each was screened for heart conditions using either standard obstetric care or AI-enhanced tools. Researchers at the Mayo Clinic had previously developed a 12-lead AI-electrocardiogram (ECG) algorithm, capable of predicting a weak heart pump, known clinically as low ejection fraction. This algorithm was further refined by Eko Health, which incorporated it into its point-of-care digital stethoscope. The stethoscope, cleared by the U.S. Food and Drug Administration (FDA), is designed to detect heart failure in patients with low ejection fractions.
The results of the study were compelling. The combination of the AI-based screening tools—comprising the digital stethoscope and the 12-lead ECG—enabled doctors to identify cases of weak heart function with a high degree of accuracy. Specifically, the AI-enhanced stethoscope doubled the number of heart failure cases detected with ejection fractions below 50%, and significantly increased detection rates for ejection fractions under 45%.
The researchers evaluated the AI-enabled screening tools across three different levels of ejection fraction, all of which are used in the clinical diagnosis of heart failure. An ejection fraction below 45% is the threshold for diagnosing peripartum cardiomyopathy, while a measurement below 40% indicates heart failure with reduced ejection fraction, for which specific medications are known to alleviate symptoms and lower the risk of mortality. Ejection fractions below 35% suggest critically low heart pump function, often requiring aggressive management, including advanced heart failure treatments or the implantation of a defibrillator if heart function fails to improve. Each participant in the intervention group underwent an echocardiogram at the start of the trial, which provided confirmation of the AI-predicted heart function.
“This research provides compelling evidence that AI-assisted tools can significantly improve the detection of peripartum cardiomyopathy, especially in Nigerian women, where the condition is more prevalent,” stated Dr Adedinsewo. “However, there are still important questions that need to be addressed. Our next step involves assessing the usability and adoption of these tools by Nigerian healthcare providers, including both doctors and nurses, as well as evaluating the impact of the AI-enabled stethoscope on patient outcomes. In the United States, peripartum cardiomyopathy affects approximately 1 in 2,000 women, but among African American women, the incidence is as high as 1 in 700. Evaluating the effectiveness of this AI tool in the U.S. will further test its capabilities across diverse populations and healthcare environments.”
The clinical trial received financial backing from several sources, including the Mayo Clinic’s Centres for Digital Health and Community Health and Engagement Research, the Mayo Clinic’s Building Interdisciplinary Research Careers in Women’s Health (BIRCWH) programme, which is funded by the National Institutes of Health (NIH), and the Mayo Clinic’s Centre for Clinical and Translational Sciences (CCATS), also funded by the NIH.
The study not only underscores the potential of AI in improving maternal healthcare, but also highlights the critical importance of early diagnosis in preventing life-threatening complications related to heart failure during and after pregnancy. With further refinement and wider implementation, this AI-enhanced tool could transform the way pregnancy-related heart failure is detected and managed, potentially saving countless lives in the process.
