Dr. Preston Dunnmon, a medical reviewer in the FDA’s division of cardiovascular and renal drug products, said that the adjudication was "well conceived, well executed, and comprehensive" and that the small number of additional MACE events identified during adjudication did not change the original results reported for RECORD. The conclusions do not consider the impact of the open-label design of the study, but he said there was no convincing evidence of any manipulation, intentional or otherwise, of the safety outcomes in RECORD.
However, one FDA reviewer adamantly disagreed. Dr. Thomas Marciniak, medical team leader in the FDA’s division of cardiovascular and renal products, presented what he said were his own professional opinions, not the FDA’s official views. He said that RECORD was inadequately designed and not reliable, and "confirms and extends the recognized concerns regarding increased heart failure and heart failure deaths with rosiglitazone," and suggests that rosiglitazone increases the risk of MI.
He said that the readjudication process was not truly independent, did not collect much additional information, and that it did not overcome the flaws in the study’s design and what he said was mishandling of data.
Several panelists remarked that the unblinded design of the study was a limitation.
At the July 2010 meeting, 12 of the panelists voted to take the drug off the market, and 17 voted to continue marketing the drug, but with revisions. Compared to the year before the REMS program was instituted, the drug was available in about 55,000 pharmacies and was prescribed by almost 109,000 clinicians, which has dropped to 4 pharmacies and about 2,700 prescribers, according to GSK.