In his commentary on the conduct of the PROactive Study(Clinical Diabetes 24:63-65, 2006), Dr. Jay Skyler reiterated his opinion of misconduct by some or all of the study statisticians, the members of the Data Safety and Monitoring Committee (DSMC), or the overall study chairman and his steering committee, and specifically that they colluded in suggesting a new principal secondary end point, having prior knowledge of the study's primary end point likelihood.1,2 By so doing, he impugns yet again the scientific integrity of all of the above. Despite our earlier unequivocal refutation of his allegations,3 we feel that his continued onslaught once again calls for an absolute and transparent rebuttal of his allegations.
Throughout the study, blinded patient data from the PROactive participating centers was transmitted directly to an independent data center, Nottingham Clinical Research (U.K.), under the directorship of Dr. Allan Skene. Here,still-blinded data was processed and forwarded to the DSMC statistician, Prof. Gordon Murray in Edinburgh, U.K. Prof. Murray was independent of the sponsor and all other aspects of the study. His unit alone had access to the treatment allocation codes, and, as agreed at study onset, prepared unblinded event tables for discussion by the DSMC at its scheduled meetings under the chairmanship of Prof. Pierre Lefèbvre. At no time during the study did the sponsor have access to the treatment codes.
The DSMC reviewed study progress with its principal role of ensuring patient safety in mind. As such, DSMC members assessed the study's declared primary and secondary end points and any other clinical or biochemical trends that they felt might impinge on patient safety as the trial progressed. They had no concern for the publication or marketing potential of the trial result,and at no time did they suggest any new analyses that might ultimately reflect favorably on the study drug.
Where and when did the previously unpublished new secondary end point cluster of mortality, nonfatal myocardial infarction (MI), or stroke so suspiciously reviled by Skyler appear on the scene? This evolved during discussions with members of the Executive Committee and study sponsors (but not the DSMC) under the auspices of the overall study chairman, Prof. John Dormandy. It was realized that the previously declared end points did not include a comparison of what had now become common in cardiovascular outcome trials, namely a composite of death, nonfatal MI, or stroke. This was debated in complete ignorance of any study events but was incorporated into the final analytical plan that was submitted to the U.S. Food and Drug Administration before database lock and study unblinding, despite uncertainty as to its importance, but in appreciation of its potential clinical relevance. The DSMC had met for the last time before this new addition to the analytical menu and had never seen, let alone discussed, data presented in such a composite, nor made surreptitious hints to that effect.
We believe that adding to a previously declared analytical plan in order to incorporate contemporary ideas in advance of knowledge of study progression is perfectly legitimate provided overall study integrity is preserved. That is what was done in PROactive. We accept that, for undeclared reasons, Skyler has difficulties accepting the findings of the PROactive study, but we unequivocally assure him and your readers that neither conspiracy nor the sacrifice of scientific integrity was involved.
The Executive Committee comprises J. Dormandy, B. Charbonnel, E. Erdmann,M. Massi-Benedetti, I. Moules, A. Skene, and M. Tan.
The Data and Safety Monitoring Committee comprises P. Lefèbvre, G. Murray, E. Standl, L. Wilhelmsen, and R. Wilcox.
For the primary sponsor John Yates, MD, President, Takeda Global Research& Development.
