Dr Steven Nissen

Dr Steven Nissen (Cleveland Clinic, OH), who presented the results during a late-breaking trial session here, called the results a "huge surprise."

"What we saw was that the people who got one of the most widely used therapies—glimepiride, a sulfonylurea—had unequivocal progression of coronary atherosclerosis by [intravascular ultrasound] IVUS, while pioglitazone had a little less plaque at the end of the study, and the difference between the two therapies was highly statistically significant. . . . To our knowledge, this is first time that a diabetes study has been shown to slow progression of coronary atherosclerosis."

Nissen emphasized that while no one study should change clinical practice, particularly one based on a surrogate end point—in this case, atherosclerosis progression as measured by IVUS. The trial did not address impact on clinical events. But Nissen also pointed out that the PERISCOPE results come in the wake of PROACTIVE, which showed a nonsignificant 10% reduction in its primary end point of all macrovascular events and a significant 16% reduction in its secondary end point of death, MI, and stroke with pioglitazone. "I think the totality of information suggests this is a beneficial therapy, but PERISCOPE alone doesn't answer all the questions," he said.

But commenting on the study for heartwire, Dr Roger Blumenthal (John Hopkins University, Baltimore, MD) warned against making too much of a small imaging study on top of a larger clinical trial that failed to meet its primary end point. "We need more supportive data. Right now the totality of evidence is not enough to change guidelines," he said. "The chance of this having a significant impact on clinical practice is the same as a snowball's chance in Hades."

The PERISCOPE results have also been published online March 31, 2008 in the Journal of the American Medical Association.

In PERISCOPE, 543 patients with type 2 diabetes underwent coronary IVUS and then were randomized to receive either glimepiride (1-4 mg) or pioglitazone (15-45 mg) for 18 months, at which time IVUS studies were repeated. According to study investigators, mean percent atheroma volume decreased by 0.16% in pioglitazone-treated subjects but increased by 0.73% in glimepiride-treated patients. When the analysis was repeated to include patients who had not completed the study, the results also showed an increase for glimepiride and a decrease for pioglitazone. Both agents lowered glycohemoglobin and fasting insulin levels, although pioglitazone's effects on these end points were statistically greater. Pioglitazone also produced statistically meaningful changes in HDL and triglyceride levels.

Adverse events in the trial were clearly different between the two drugs. More patients taking glimepiride developed hypoglycemia and angina, while patients taking pioglitazone were more likely to develop edema, gain weight, or suffer bone fractures.

Commenting on the study, Dr Salim Yusuf (McMaster University, Hamilton, ON) highlighted the fracture results, which occurred in 3% of the pioglitazone-treated patients.

"This was a significant excess in fractures with pioglitazone, and no matter how good a surrogate end point, even if it is truly related to the outcome you're interested in, which may be CV events, it doesn't tell you the totality of the benefit/risk," he said.

Likewise, Dr Darren McGuire (UT Southwestern, Dallas, TX), speaking with heartwire, also acknowledged that the fracture rate was "surprisingly high"—higher than the signal of fracture risk seen before with this drug.

"I think it is something to pay attention to," he said. "These drugs are not completely benign, but in total their benefit-to-risk ratio in select patients remains favorable."

Indeed, McGuire points out that PERISCOPE is not a standalone imaging study, because of the promising secondary results in PROACTIVE. "What the PERISCOPE study does is provide proof of principle that the drug is to some degree modifying atherosclerosis."

Several observers have pointed out that the absolute changes in atherosclerosis progression are small, a comment Nissen rejects. He cites research in progress at his own institution that is examining clinical outcomes in relation to IVUS results from statin trials. "I can tell you that, with a p value with a lot of zeros before the one, that changes of around 0.8% to 1.0% are associated with a very substantial reduction in hard end points across all the trials we've done. The differences we saw here are really very statistically robust and they will translate into clinical benefits," he said.

Indeed, an editorial [2] by Drs P Gabriel Steg (Centre Hospitalier Bichat-Claude Bernard, Paris, France) and Michel Marre (Universite Paris VII, France) points out that, while small, the apparent affect of pioglitazone is "well within the range of what is achieved with some therapies demonstrated to improve cardiovascular outcomes, such as high-dose statins."

Following Nissen's presentation, one of the session moderators, Dr Greg Brown (University of Washington, Seattle), asked Nissen if the PERISCOPE results have "changed [his] opinion about the glitazone class," referring to the 2007 hullabaloo over rosiglitazone, sparked in large part by a meta-analysis that Nissen coauthored.

Nissen called this a "fair question" but emphasized that, while technically in the same class, rosiglitazone and pioglitazone affect different genes. "They both affect a gene that is involved in lowering blood sugar, but they have otherwise extraordinarily different effects. We have to study each of these compounds individually."

In the press conference, Nissen acknowledged that he and his coauthors cannot yet explain the mechanism by which pioglitazone alters atherosclerosis progression but that its effects on blood pressure, lipids, triglycerides, and CRP are major candidates. "What this study now tells us is: we must do a comparator effectiveness trial looking at different diabetes treatment strategies. We can't just focus on pricking the finger, getting the blood sugar down, and saying, that's the goal of therapy. The goal in therapy is to prevent complications of diabetes, and the most feared, most serious complication is heart disease, which will kill 75% of diabetics."