ACC: When should devices or drugs stop being used in practice?

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon

CHICAGO—When medical drugs or devices begin to go awry, does the company know beforehand about the potential issues? This is the question Steve E. Nissen, MD, chairman of the department of cardiovascular medicine at the Cleveland Clinic, asked March 24 at the 61st annual American College of Cardiology (ACC) scientific session. The presentation aimed to answer one question: When should devices or medicines stop being used in practice?

Nissen referenced the rosiglitazone (Avandia, GlaxoSmithKlein) saga, in which the diabetes drug was shown to raise the relative risk of ischemic events. “The agency and company knew that this drug [rosiglitazone] raised LDL-cholesterol an average of 18.6 percent in patients,” Nissen said.  

“The whole approval went awry,” he offered. “This was a bad drug right out of the box.”

He asked why anyone would have approved a drug prescribed to diabetics when it was shown to raise LDL-cholesterol by 18.6 percent. In the fall of 2009, Nissen said that GSK “secretly” began a study whether its diabetes drug rosiglitazone was safer than its competitor Takeda Pharmaceutical’s pioglitazone (Actos). The company did not post the results on its website or to the public. It took 11 years for the study to be released and the industry to learn that the drug may be harmful to patients, Nissen said.

He said that it was realized that there was another drug in the thiazolidinedione (TZD) class, pioglitazone, that was being looked at. “It was found that pioglitazone had lipid effects that were relatively benign,” he said. In 2006, the World Health Organization urged GSK to do a meta-analysis including 42 randomized controlled trials that enrolled a total of 12,183 patients. It was found that there was a 30 percent risk of ischemic events in the rosiglitazone arm.

“No one told us,” Nissen said. That same year rosiglitazone became the best-selling diabetes drug on the market. The fact that it could raise the risk of MI was swept under the rug, he said.

In an analysis of the drug, Nissen and Kathy Wolski, MPH, found a 43 percent increase in MI with rosiglitazone (N Engl J Med 2007; 356:2457-2471). As the manuscript was published, GSK managed to get a copy and did an analysis to try to prove otherwise, Nissen said. The RECORD trial published in Lancet did just that; Home et al suggested that the data published on the effect rosiglitazone has on MI were “inconclusive” and said that the drug did not increase the overall risk of CV morbidity or mortality compared with standard glucose-lowering drugs (Lancet 2009;373:2125-2135).

Nissen said it was found that there was “case after case” where MIs occurred in the rosiglitazone arm of the trial, but had been deleted in trial results. Post-RECORD, the FDA performed an analysis using the same database without removing the MI cases and found a hazard ratio of 1.14, a number similar to Nissen and Wolski’s findings, he said.

To add to this saga, a 2010 analysis showed that the rates of CV death, MI or stroke with pioglitazone did not increase the risk of ischemic events. "If anything it was cardioprotective,” said Nissen. “We sat here for 11 years and allowed people to prescribe a drug where it was increasing the risk of MI from 30-40 percent when a comparably efficacious drug had the exact opposite effect.

“The FDA didn’t act, the company didn’t act and the public wasn’t protected."

Nissen added that “Pharmageddon” for rosiglitazone occurred Sept. 23, 2010, when the European Medicines Agency ordered rosiglitazone off the market in the European Union.

What are the lessons learned? Nissen asked. “When drug safety problems emerge, the company making the drug sometimes has known about the problem for many years,” he offered. “Drugs in the same class may have widely different biological effects."

Nissen urged that large clinical outcomes trials be conducted that are ethically performed to help shed light on these problems and ensure that these situations are avoided in the future.

“Right now our system does not protect it,” Nissen said. “This example didn’t result in 40 deaths … the estimate here is that 50,000 patients had an MI or died on rosiglitazone.”

Going back to 1999, the company knew that there may have been trouble surrounding this drug, he said.

“So I ask you, how can we make good decisions for our patients when we don’t have the totality of the information about the benefits and adverse effects of these drugs?” Nissen asked.

Nissen said that the only way