Monday, March 5 is the Final Day to Submit Your Opinion to the FDA on the Recent Artificial Pancreas Guidance!
As we covered in the diaTribe #39 Learning Curve and the diaTribe #38 Letter from the Editor, the FDA recently released its much awaited draft guidance on the artificial pancreas (AP). This document contains the FDA’s recommendations for effective and safe testing of an artificial pancreas, which will culminate in an FDA review and hopefully approval of the system for use by patients. The current version of the guidance document is a “draft” because it is open for public comment from patients, researchers, and industry representatives. You can submit a comment here until March 5, presumably 5 pm EST. Overall, we here at diaTribe are pleased that the FDA recognizes the importance of developing AP systems and has provided a feasible path to develop such systems. However, as a diabetes patient advocates, we have also taken it upon ourselves during the draft guidance comment period to think of ways that the guidance could be further improved. The JDRF has a great blog post summarizing the comments it will be submitting, and we here at diaTribe echo their sentiments and hope to see the FDA address the following areas as it finalizes the draft guidance:
Time in Range (TIR): This is the one we care about the most. The FDA’s guidance document does not accept time in range (e.g., 70% of the time between 70-180 mg/dl) as a primary measure to evaluate an AP’s performance (called a ‘primary endpoint’). Unfortunately, this is the major measure by which most artificial pancreas studies are currently evaluated and compared. One reason TIR is so widely used is that it’s a more appropriate measure than A1c for short-term artificial pancreas studies of a few days or weeks. Of course, it still doesn’t capture every aspect of glucose control and we believe that its validity should be studied in long-term trials. However, compared to a random A1c number, we think TIR remains the best way to concisely and logically summarize blood sugar control. As a result, based on TIR’s intuitive importance and its status as a standard within AP research, we think that FDA should identify how this measure can be used as a primary endpoint in studies.
Study Recruitment: We are concerned about study recruitment and cost given the safety bar outlined in the guidance: sponsors must demonstrate that the AP does not increase the rates of severe hypoglycemia, severe hyperglycemia, or DKA. Given the low rates of these events, we believe that long and large studies will be needed and/or that sponsors must enroll individuals with recurrent severe events (a fairly small population). On top of this, the Agency prefers enrollment of experienced pump/CGM users (>3-6 months), or at minimum, pumpers who undergo a four- to six-week CGM training period. With all of this mind, we anticipate that recruitment for these studies will be challenging, time-consuming, and expensive. We urge the FDA to consider safety requirements and study designs that can speed recruitment and reduce study cost and length while gathering the necessary data.
Superiority Margins: To gain FDA approval, studies must show that those wearing the artificial pancreas do not increase their A1c compared to those wearing a traditional pump and CGM (this is known as ‘non-inferiority’). Superiority studies, going beyond non-inferiority, can demonstrate that the artificial pancreas is better than wearing just a standard insulin pump and CGM. For this sort of claim, the guidance document states that the artificial pancreas must improve A1c by at least 0.4% and/or improve measures of hypo- or hyperglycemia (as measured by CGM) by at least 30%. Unfortunately, these margins represent a fairly high bar and will require either very large differences between study groups or a high number of trial participants. Additionally, many of the FDA’s recent device guidance documents do not include specific values for non-inferiority or superiority. As a result, instead of defining specific margins, we believe the FDA should work with each sponsor on an individual basis to define the statistical margins for their particular artificial pancreas study. –AB/JS/MY