Edward Tobinick, Güven Gürbüz
The PESTO trial's claim to provide Class I evidence that perispinal etanercept does not improve quality of life is precluded by the absence of publicly verifiable trial-wide inclusion/exclusion criteria and by its use of a novel primary responder definition that utilizes a non-accepted scoring pathway, of no established validity or reliability, that was first specified after data collection was completed. PESTO's Statistical Analysis Plan differs from the analysis its operative protocols and registration specify, and no amendment, registry update, or explanation discloses the change. PESTO's published primary quality-of-life measurement is unreliable, and its definitive negative conclusion cannot be relied upon as a statement of what the trial established. Whether the reported estimate would survive the prespecified analysis, the record cannot yet answer. Correction of PESTO's record and conclusions and reclassification of its Class I designation are due now. Further, independent multidisciplinary review, with separate statistical and research-ethics components, is required.