RECORD interim results tabulated in response to rosiglitazone meta-analysis

Results ‘contradict’ the earlier finding of excess cardiovascular death that led to Congressional hearing.

Issue: July 2007

In direct response to results of the recent meta-analysis examining cardiovascular risk associated with rosiglitazone treatment, researchers of the Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycemia in Diabetes trial, or RECORD, have carried out an interim analysis of their trial.

Thus far in the trial, no statistically significant evidence of an increase in death from cardiovascular causes associated with rosiglitazone has been identified. The interim results were published in The New England Journal of Medicine, as was the meta-analysis which prompted it.

The publication of both studies and the subsequent reactions to them among physicians and patients alike prompted the American Diabetes Association to hold a special panel discussion of rosiglitazone at the association’s 67th Annual Scientific Sessions held in June in Chicago. Researchers from both trials gave presentations during the panel.

The RECORD interim results were published online on June 5th, a day before a Congressional hearing was held to determine why the FDA delayed releasing warnings about rosiglitazone. At that hearing, the House Committee on Oversight and Government Reform questioned FDA representatives, including FDA commissioner Andrew C. von Eschenbach, MD; a researcher of the meta-analysis, Steven E. Nissen, MD, chairman of the department of cardiovascular medicine at the The Cleveland Clinic; John Buse, MD, PhD, president of the American Diabetes Association and chief of the division of general medicine and clinical epidemiology at the University of North Carolina; and Moncef Slaoui, PhD, chairman of research and development for GlaxoSmithKline, among others.

Since the hearing, the FDA has called for black box warnings for congestive heart failure to be included in labeling on both rosiglitazone and competitor drug pioglitazone (Actos, Takeda).

RECORD data and MI

The data from RECORD are insufficient to determine whether rosiglitazone (Avandia, Glaxo SmithKline) is associated with an increased risk for MI, according to the researchers. The interim results demonstrate a lower number of overall cardiovascular events and a similar number of events between patients assigned rosiglitazone compared with patients assigned metformin plus sulfonylurea.

“There is no evidence of any increased mortality, either from any cause or from cardiovascular causes,” wrote the RECORD researchers. “The data do not allow a conclusion as to whether treatment with rosiglitazone results in a higher rate of MI than does therapy with metformin or a sulfonylurea.”

The interim results follow patients for 3.75 years and were released under “extraordinary circumstances,” Stuart J. Pocock, PhD, professor of medical statistics at London School of Hygiene and Tropical Medicine and a RECORD researcher, told Endocrine Today. “A borderline significant meta-analysis doesn’t usually hit the headlines,” he said.

Pocock was referring to the meta-analysis by Nissen and Kathy Wolski, MPH, a statistician at The Cleveland Clinic, which reflected a 43% higher risk for MI and 64% increased risk in cardiovascular death in patients assigned rosiglitazone vs. other diabetes medications.

RECORD is a randomized, multicenter, open-label, noninferiority trial of 4,447 patients with type 2 diabetes. The primary end point of the trial is hospitalization or death from cardiovascular causes associated with rosiglitazone. The investigators assigned 2,220 patients to receive add-on rosiglitazone and 2,227 a combination of metformin plus sulfonylurea.

The overall rate of hospitalization or death from cardiovascular causes seen in the interim results was 3.1% per year for adjudicated plus pending events.

Particularly striking, according to Pocock, were a fewer number of overall deaths and cardiovascular deaths in the rosiglitazone group compared with the control group (217 vs. 202; HR=1.08; 95% CI, 0.89-1.31; after inclusion of end points pending adjudication, HR=1.11; 95% CI, 0.93-1.32).

When the researchers compared the interim results with the excess risk for MI found in Nissen and Wolski’s meta-analysis, the RECORD data did not demonstrate any statistically significant difference between the two groups. In RECORD, there was a 23% (HR=1.23; 95% CI, 0.52-1.24) increased risk for MI in the rosiglitazone group; in the meta-analysis, there was a 43% (OR=1.43; 95% CI, 1.03-1.98) increased risk.

Pocock said these results are reassuring, albeit nonsignificant. “We found no hint of an excess risk for mortality, whereas Nissen and Wolski had suggested there might be an excess of cardiovascular death. Our interim results contradict that,” he said.

Excess risk for coronary heart failure in RECORD among patients assigned rosiglitazone occurred more frequently (38 vs. 17 adjudicated events; HR=2.24; 95% CI, 1.27-3.97). When the investigators included events pending adjudication, the number of events increased (47 vs. 22; HR=2.15; 95% CI, 1.30-3.57). This translates to an excess risk for HF in the rosiglitazone group of 3.0 per 1,000 patient-years of follow-up.

“We were confirming that there is an identified and fairly well-known increased risk with heart failure. That is the only excess risk we found,” Pocock said. “Every other cardiovascular outcome we looked at, there was no significant difference between the rosiglitazone group and the control group.”

According to the researchers, RECORD had limited statistical power due to withdrawal of patients from the trial and follow-up and the unexpected low rates of primary end point in both groups. The trial will not be completed until 2009.

Safety remains questionable

The safety of rosiglitazone for the treatment of patients with diabetes remains controversial, according to three editorials that accompanied the RECORD report in the The New England Journal of Medicine.

According to David M. Nathan, MD, professor of medicine, Harvard Medical School, the interim results do not provide any reassurance of the drug’s safety. He referenced the original meta-analysis (OR=1.43; 95% CI, 1.03-1.98), the independent analysis by GlaxoSmithKline (OR=1.31; 95% CI, 1.01-1.70) and the RECORD data (HR=1.08; 95% CI, 0.89-1.31) and said none should be considered conclusive. Rather, these analyses support a concern about the safety of rosiglitazone, he wrote.

“The jury may still be out with regard to the cardiotoxicity of rosiglitazone, but when it comes to patient safety, first do not harm should outweigh any presumption of innocence,” Nathan wrote.

Other editors, including Jeffrey M. Drazen, MD, editor-in-chief of the NEJM, and his colleagues also expressed concern about safety. “Even a small increase in cardiovascular risk in a fragile population of patients with type 2 diabetes is of considerable concern,” they wrote.

Others expressed concern about the weaknesses of RECORD. The primary weakness is the exceptionally low event rate in a high-risk population, according to Bruce M. Psaty, MD, PhD, professor of medicine, Cardiovascular Health Research Unit, University of Washington, and Curt Furberg, MD, PhD, professor of public health sciences, Wake Forest University, Winston-Salem. For example, the rate of MI events in the control group was 4.5 per 1,000 person-years.

“Even with the findings from the RECORD trial included [with those of the meta-analysis], the possibility of some benefit in terms of the risk for MI remains remote, and there is still significant evidence of harm,” Psaty and Furberg wrote.

No undue harm

Pocock said he and the RECORD investigators believe it was a prime time to release the interim findings.

“We felt it was important that everyone, the public, was faced with the total current evidence, rather than just what was published from a meta-analysis,” he said. “Second, we thought we would do more good than harm to the continued successful completion of RECORD; patients were more likely to continue in the study knowing that we were not seeing any undue great harm occurring.”

Nissen said that ultimately, making the right decision on behalf of the patients is what matters most. “The real question here is not the merits of the [meta-analysis] but what can be done to correct this regulatory failure in the future,” Nissen told Endocrine Today.

“We need to improve our system for drug surveillance. The issue is how we are going to do that,” Nissen said. “Some of the solutions revolve around transparency for clinical trials, some will require improvements in our drug surveillance. A large component will be mandating post-marketing studies with mandates that drug companies will follow. Most post-marketing requirements currently are not fulfilled.”

Compiled by Katie Kalvaitis, Jeremy Moore, Evan Young and Judith Rusk.

For more information:
Home PD, Pocock SJ, Beck-Nielsen H, et al. Rosiglitazone evaluated for cardiovascular outcomes: an interim analysis. N Engl J Med. 2007;doi:10.1056/NEJMoa073394. Accessed June 5, 2007.
Nathan DM. Rosiglitazone and cardiotoxicity: weighing the evidence. N Engl J Med. 2007;doi:10.1056/NEJMe078117. Accessed June 5, 2007.
Drazen JM, Morrissey S, Curfman GD. Rosiglitazone: continued uncertainty about safety. N Engl J Med. 2007;doi:10.1056/NEJMe078118. Accessed June 5, 2007.
Psaty BM, Furberg CD. The record on rosiglitazone and the risk of myocardial infarction. N Engl J Med. 2007;doi:10.1056/NEJMe078116. Accessed June 5, 2007.
For more about the NHLBI’s trials using rosiglitazone visit: www.nhlbi.nih.gov