NEW YORK – A team led by researchers at the University of Oxford has used genetics to show that treating patients with an anti-osteoporotic drug called romosozumab might increase their risk of cardiovascular complications.
The findings, reported on Wednesday in Science Translational Medicine, underscore the value of carrying out large-scale human genetic analyses as part of drug target identification and validation efforts, according to Jonas Bovijn, first author on the study and a researcher at the University of Oxford's Big Data Institute.
Romosozumab, marketed as Evenity by Amgen and UCB, is an anti-osteoporosis drug that belongs to a class of pharmaceuticals called sclerostin inhibitors. The US Food and Drug Administration cleared the drug last year, but with a warning that romosozumab might increase risk of myocardial infarction, stroke, and cardiovascular death, a finding seen in some but not all clinical trials of the drug. The European Medicines Agency has issued a similar warning.
"This finding was not consistently seen across all trials of romosozumab, which limited conclusions about the safety of this medicine," noted Bovijn in an email. His team therefore decided to look through existing clinical trial data as well as data from several biobanks to understand whether the reported effect was real or not.
Partnering institutions included the University of Tartu in Estonia, the Broad Institute and Massachusetts General Hospital in the US, and the MRC Integrative Epidemiology Unit at the University of Bristol in the UK. "We analyzed data from several large-scale biobanks, including UK Biobank, Estonian Biobank, Partners HealthCare Biobank, and China Kadoorie Biobank," Bovijn said.
The researchers collected cardiovascular safety data from clinical trials of romosozumab and looked at the average risk of cardiovascular complications across the trials. Then, turning to the biobanks, they looked at people in the general population who carry genetic variants in the sclerostin gene, SOST, that mimic the effect of romosozumab and examined their risk of developing cardiovascular disease. The analyses aggregated data from just over 11,000 clinical trial participants and more than a million subjects across several biobanks and genome-wide association study datasets, Bovijn noted.
He added that the work involved computational and logistical challenges, including systematically reviewing the literature, data extraction and aggregation, and implementing a range of analytical approaches. "The interpretation of findings from human genetics may also pose challenges; for instance, genetic variation may not always be a valid tool for predicting drug effects, although we have performed several sensitivity analyses to strengthen our conclusions," he said.
The authors found that people who carry genetic variants that mimic the effect of romosozumab had, on average, a 41 percent lower risk of sustaining a fracture but an 18 percent increased risk of heart attack, which supported the increased risk of heart attacks seen in the trials of romosozumab as well as the FDA and EMA's warning labels.
"We found that the totality of evidence from clinical trials and from human genetics supports the existence of an increased risk of cardiovascular complications arising from sclerostin inhibition," said Bovijn. "These findings support the warning labels issued by regulatory authorities such as the FDA and EMA, and suggest that the cardiovascular adverse effects are real. This also indicates that other therapies that inhibit sclerostin are likely to pose a similar risk."
It might also encourage pharmaceutical companies to run similar studies of drugs they have in development, he said, a trend that is already underway. "Most drug developers are likely already incorporating large-scale human genetic analyses in their drug target identification and validation efforts," said Bovijn. He cited GlaxoSmithKline's work with 23andMe as well as Regeneron Pharmaceutical's work to sequence the exomes of 450,000 individuals in partnership with the Colorado Center for Personalized Medicine as examples of such endeavors.
As for Bovijn's own group, he said they are "working on a range of other drug target-related questions, with the hope that our data brings important information from genetics in a timely way that can help guide the development of therapies."