Abstract 20910: Are Claims of Effectiveness of MitraClip Device Compared with Surgery for Severe Mitral Regurgitation Substantiated by the Evidence? A Critical Look at the EVEREST II Trial.
Background: The EVEREST II trial reported that the percutaneous MitraClip device demonstrated noninferiority (NI) with respect to effectiveness and improved safety compared with surgical treatment for severe mitral regurgitation (MR). However, 23% (41/178) of patients in the device arm were excluded from the analysis and 95% confdence interval (CI) of observed differences were not reported.
Objective: To reanalyze the trial data using appropriate NI methodology.
Methods: We performed Per Protocol (PP) analyses with (effectiveness 1) and without (effectiveness 2) excluding 21 patients that did not achieve procedural success despite device implantation. Similarly, intention-to-treat (ITT) analyses were performed with (effectiveness 1) or without (effectiveness 2) excluding 20 randomized patients who despite attempts did not receive the device. Two-sided 95% CI (one-sided 97.5% CI) were estimated for both absolute and relative risk differences.
Results: (Table): The trial met its primary safety endpoint for both PP and ITT analyses. In the PP analysis (effectiveness 1), clinical failure rate was 28% device vs. 12% surgery. Although NI was established using a liberal margin, the data also established device to be inferior to surgery (95% CI excluding zero difference). Using the PP analysis (effectiveness 2), the device was established to be inferior, but NI was no longer met. Similar results were observed with ITT analyses.
Conclusions: The MitraClip device is up to 2- to 2.5-fold worse (ITT cohort) and up to 4- to 6-fold worse (PP cohort) than surgery and yet deemed not “unacceptably worse” — an arguable conjecture. The data also reveal an apparent paradox of NI success with an inferior intervention. These issues plus the wide NI margin and 20% cross-over to surgery materially undermine the investigators conclusion of noninferiority, thereby calling into question the enthusiastic claims that the device is an important therapeutic option for patients with severe MR.
- © 2010 by American Heart Association, Inc.