Their study, published October 11 in the journal Cell Reports Medicine, is related to a $14.2 million National Institutes of Health grant to investigate new atrial fibrillation treatments.

Atrial fibrillation, the world's most common heart arrhythmia, can lead to stroke and heart failure. Advanced computation and genetic sequencing enabled researchers to determine that metformin's targets overlap significantly with genes dysregulated in atrial fibrillation.

Metformin targeted 30 genes associated with atrial fibrillation, directly effecting gene expression for eight. Testing and reviewing outcomes in large stores of patient data showed metformin to be the most promising candidate.

A network medicine approach to find and repurpose candidate drugs enabled researchers to winnow down 2,800 U.S. Food and Drug Administration-approved treatments. They analyzed three data sources: a map of interactions, or an "interactome," between proteins; a gene network associated with atrial fibrillation; and each medicine's molecular or genetic targets.

Previous atrial fibrillation treatments have been directed toward preventing arrhythmia using drugs targeting the electrical system or using catheter ablation to isolate the pulmonary veins where initiating atrial fibrillation beats occur. However, serious side effects, complications, and limited success limit these treatments.

"Finding drugs or procedures to treat atrial fibrillation is difficult because of potential serious side effects," co-senior author Dr. Mina Chung of Cleveland Clinic, said in a statement, while noting that there is a "significant need for new treatments for atrial fibrillation as there have been no new drugs approved in more than a decade."

"We can cut off 10-plus years in the drug development pipeline. We already have the information there. We just have to test it in a very computationally efficient way, such as artificial intelligence technology," added Cleveland Clinic's co-senior author Feixiong Cheng, PhD.