ADA 2010: Near agreement on CV risk assessment for diabetes drugs

June 28, 2010

It was billed as a debate, but played as a mild disagreement. Should new diabetes drugs submitted to the FDA have a higher bar for cardiovascular safety than other drugs?

It was billed as a debate, but played as a mild disagreement. Should new diabetes drugs submitted to the FDA have a higher bar for cardiovascular safety than other drugs?

Yes, said Steven Nissen, MD, chair of cardiovascular medicine, Cleveland Clinic, Cleveland, Ohio, during a current issues debate Saturday afternoon.

Yes for some drugs, but not as a blanket rule, said David Orloff, MD, executive director, regulatory affairs, Medpace, Cincinnati, Ohio.

Too many diabetes drugs have been approved on the basis of surrogate markers rather than clinical outcomes, Nissen explained. If a drug lowered serum glucose, it would improve diabetes and warranted approval.

"I invented a word to describe what was going on at FDA," he continued. "Glucocentricity. It's the irrational belief that lowering blood sugar by any pharmacologic means will improve clinical outcomes. That's why we got drugs like rosiglitazone."

The agent was known to increase LDL by 20%, but it was approved because it prevented progression to diabetes. Meta-analysis showed that it also conferred a hazard ratio of 1.43 for myocardial infarction. A similar analysis of pioglitazone showed no effect on MI or other cardiovascular events.

"Drugs to lower blood sugar can help cardiovascular outcomes or they can seriously harm cardiovascular outcomes," Nissen pointed out. "How are we going to know the difference if we don't require a cardiac analysis? We have a real challenge if we just look at glycemia."

His proposal, eventually adopted by FDA, was to require cardiovascular data for all new diabetes drugs. A preapproval trial must show that any cardiac risk is no greater than 1.8. That equates to about 120 cardiac events, he noted, and adds 6 to 12 months to the typical drug submission.

A postapproval outcomes trial must show cardiovascular risk of less than 1.3.

"That is a very small price to pay for good outcomes data," Nissen said. "With 13 classes of drugs that lower blood sugar, we don't need more drugs, we need better drugs. It matters as much how you lower sugar as how much you lower it."

Very true, Orloff agreed, but requiring cardiovascular data on every new diabetes drug is untested.

"Today's debate is about the wisdom and practicality of imposing an experimental regime on drug approval," Orloff said. "To put a diabetes drugs through cardiovascular testing regardless of class or mechanism of action is not a rational expenditure of resources.

"We should be wary of any adverse events," Orloff continued. "An obligatory focus on any single aspect has the strong possibility of becoming an untenably expensive approach. It isn't about cutting corners, it is about finite resource allocation. We should make that decision case by case."