May 29th, 2013

Report on the RECORD Trial Released

An independent re-adjudication of the RECORD trial has confirmed the trial’s original finding that rosiglitazone does not increase cardiovascular risk. But critics of the trial and the drug are unlikely to be appeased by the new result.

The re-adjudication of RECORD will be the subject of an extraordinary 2-day FDA advisory committee meeting next week on June 5 and 6. GlaxoSmithKline posted a summary of the results on its website last October, but its presence was not generally known until the publication of an article in BioCentury on Tuesday.

The meeting next week may turn out to be the last major episode of the rosiglitazone controversy that began in 2007 when a New England Journal of Medicine meta-analysis raised concerns that people taking rosiglitazone might be at increased risk for myocardial infarction. The first author of that paper, Steve Nissen, has remained the most vocal critic of the drug and of the RECORD trial. RECORD was designed to demonstrate the safety of rosiglitazone, but critics of the trial — including Nissen and, ultimately, FDA reviewers — raised numerous questions about the conduct of the trial and the validity of the data. Following a 2-day FDA advisory panel in 2010, rosiglitazone was withdrawn from the market in Europe and its use was severely restricted in the U.S.

The FDA also required GSK to commission the independent re-adjudication of RECORD that now resides on the company’s website. The re-adjudication was performed by the Duke Clinical Research Institute (DCRI). According to BioCentury, “GSK said it does not typically publicize data put on its clinical trial registry and added that the Duke researchers are planning to submit the data for publication.”

Here is the conclusion of the DCRI re-adjudication:

As concluded by the independent DCRI RECORD CEC group, there was no evidence of treatment effect on all-cause mortality or cardiovascular plus unknown mortality. The results of the primary analyses and extensive sensitivity analyses using the re-adjudicated outcomes were consistent with the original RECORD results and did not show statistically significant differences between the RSG and the MET/SU groups for the primary composite of CV death (+unknown), MI or stroke and the individual components. Using both original definitions and new contemporary definitions for re-adjudication, RSG group had a numerically lower incidence of the composite MACE endpoint, higher incidence of MI events, and lower incidence of strokes compared with the MET/SU group.

As reported here last week, Nissen will not be participating in the panel and he is accusing the FDA of stacking the committee in favor of rosiglitazone. Nissen doesn’t believe the re-adjudication can truly confirm the safety of rosiglitazone. As he told me last week:

GSK supplied the files used by Duke to re-adjudicate the RECORD Trial. Given the unblinded nature of the study, how can this process be viewed as independent? Duke could only adjudicate what was provided to them by GSK. Given the pending court cases against GSK, the company was strongly incentivized to provide the information in a way designed to bias the outcome.


Partial Rosiglitazone/RECORD Chronology

Comments are closed.