AI could be used to guide drug therapy for heart failure

UK researchers have used artificial intelligence (AI) to develop a way to identify patients with heart failure who would benefit from treatment with beta blockers, by interrogating data from nine landmark clinical trials.

The study applied machine learning to trial data involving 15,659 patients who were being treated for heart failure with reduced ejection fraction (HFrEF), a form of heart failure where the left side of the heart isn't able to pump blood effectively around the body.

Among this group, 12,823 had normal heart rhythm, while 2,837 had atrial fibrillation, a disorder that is often associated with heart failure and typically means patients have a poorer prognosis, despite improvements in drug therapy for HFrEF.

Beta blockers are a standard therapy for people with heart failure, but haven't been able to improve survival in people with AF in trials. The researchers wanted to see if their AI could select the patients who would benefit the most from treatment with them.

In the study – which is published in The Lancet – the AI examined different underlying health conditions for each patient, as well as the interactions of these conditions, to identify the profile of a patient who would be expected to respond to beta blocker therapy.

In patients with AF, the research found a cluster of patients who had a substantial reduction in death with beta blockers – from 15% to 9%. The group included younger patients with lower rates of prior heart attack but similar heart function to the average AF patient.

Meanwhile, the AI also identified the group with normal heart rhythm most likely to get a benefit from beta blocker treatment as patients with older age, less severe symptoms and a lower heart rate than average.

The authors, from the University of Birmingham's cardAIc group, concluded that applying the AI prospectively to patient therapy could provide an individualised approach to treatment that "accounts for comorbidities, thereby contributing to improvements in patient wellbeing".

They also suggested the same approach could be applied across the spectrum of therapies for heart failure, as well as in other cardiovascular and non-cardiovascular conditions.

"We hope these important research findings will be used to shape healthcare policy and improve treatment and outcomes for patients with heart failure," said lead author Dr Andreas Karwath, Rutherford Research Fellow at the University of Birmingham.

AI has also been used to try to identify atrial fibrillation in primary care using data from electrocardiogram monitoring, although a recent study suggested that the approach still isn’t ready to replace trained specialists.