Despite the Fixings and Garnish Atop FDA’s Proposed Risk Classification Scheme, the Obesity Device Industry Appears to Have Been Served a Potentially More Restrictive Diet

30% observed mean EWL over sham; AND

FDA recenlty held a general issues advisory panel meeting to present a proposed risk classification scheme designed to allow the agency to stratify medical devices intended to treat obesity into varying risk levels. Each risk level would be tied to a specific efficacy endpoint, with higher endpoints required for higher risk devices and lower efficacy endpoints for lower risk devices. The proposed risk classification scheme, presented to the panel as a straw man, was based on discussions at the Device Development in Obesity and Metabolic Disease Workshop that was co-sponsored by FDA last fall. The purpose of the general issues panel meeting was for FDA to gain input on the general concept of the proposed framework, as well as the specific risk levels, proposed efficacy endpoints and clinical trial design issues. As a general matter, the proposed risk classification scheme looks appetizing, particularly to an industry that has been requesting efficacy endpoints to be tailored to the risks of the device, as technologies have moved to less invasive, temporary and reversible devices. As a practical matter, however, the proposal as presented, provides little give when compared to historical endpoint requirements which are based on the high risk devices such as the Lap-Band® and Realize® adjustable gastric banding systems. Specifically, FDA’s proposed co-primary efficacy endpoint for a sham-controlled study for the highest risk level - Level 3 – would be as follows:

50% responders with 25% EWL in the treatment group.

 This is a higher bar than the following endpoint FDA has been holding companies to in the past:

25% observed mean EWL over sham; AND

50% responders with 25% EWL in the treatment group.

At the same time, the co-primary efficacy endpoint requirement for Level 2 devices was presented as:

20% observed mean EWL over sham; AND

50% responders with 15% EWL in the treatment group.

The above endpoints would likely not present as significant an issue for industry if there was comfort that only LapBand®-like devices would be subject to the Level 3 primary efficacy endpoint.  However, the statistical thresholds FDA proposed for the predefined risk categories, which are based on the percentage of expected and unexpected adverse events occurring within one year (as presented in Table 1 on page 32 of the Panel Pack), would likely push most devices into Level 3.  This is certainly a major concern for industry, and was noted by the panel. 

FDA appears willing to consider many of the panel’s proposed changes, and to consider revisions to the proposed scheme, including refinements to the statistical thresholds and risk categories, as well as perhaps the addition of another intermediate risk level - a Level 2b - to help avoid a sitation where everything lands in Level 3.  With respect to the latter, it is unclear how that would play out with the associated efficacy endpoint requirements.  Of course, even as modestly lowered, the co-primary efficacy requirement for Level 2 devices may still offer little relief for companies developing the temporary, reversible devices designed to be placed endoscopically, as it may remain difficult to observe this level of efficacy over the control within a reasonably sized pivotal study.  What is clear is that FDA is committed to working towards a new approach that all stakeholders can live with – at least until more data is gathered on the lower risk devices, so risk levels can be based on actual risk profiles. However, concerns remain with the complex framework, as proposed, which appears to lead most devices to the highest risk level that has an associated efficacy requirement that has only been made tougher.  Once tweaked, FDA’s revised framework could very well provide for the tailored approach on study design issues that industry has been starving for.  In the mean time, it remains to be seen how FDA will approach these issues with respect to incmoing IDE submissions, and whether they will begin to test some aspects of the proposed framework.

Share Back to main blog

Related blog posts

Loading data