FDA Panel Supports Novel Implantable Glucose Sensor

Miriam E. Tucker

April 02, 2018

A US Food and Drug Administration (FDA) advisory panel voted unanimously on March 29 to recommend approval of a novel implantable continuous glucose monitor (CGM) for use in adults with diabetes.

Votes of the FDA's Clinical Chemistry and Clinical Toxicology Devices Panel were 8 to 0 with no abstentions for safety, efficacy, and positive benefit/risk for the Eversense (Senseonics, Inc). If approved, it would be the first-ever implanted CGM available in the United States. A version of device was granted a CE Mark in Europe in 2016, and nearly 1700 patients are currently wearing it in 14 countries.

The fluorescence-based system comprises a sensor that is implanted subcutaneously into the upper arm during a 5- to 10-minute office-based procedure, an external transmitter taped onto the skin above it, and a mobile smartphone or tablet-based application that displays and stores the glucose readings.

The implanted sensor lasts 90 days (with a 7-day grace period), at which point it is removed with another short procedure, and a new one is inserted on the opposite arm. In contrast, most current external CGMs are approved only for 6 to 14 days. (The European device is approved for 180 days.)

As with other CGMs, the Eversense provides real-time glucose readings every 5 minutes, glucose trend information, and alerts for high and low blood glucose levels. In contrast to current CGMs, the device itself also vibrates to alert the user to high and low glucose levels, even when the reader isn't nearby.

Unlike the Dexcom and Abbott Libre CGMs, the proposed indication for the Eversense is for adjunctive use, meaning the user would still need to perform fingerstick monitoring for treatment decisions, in addition to twice-daily calibrations.

A long-lasting implantable CGM addresses an unmet need, according to endocrinologist Jeremy H. Pettus, MD, from the University of California, San Diego, speaking on behalf of Senseonics. At this time, 76% of patients with type 1 diabetes are not using CGMs, and 27% who start, discontinue them within a year because of problems with accuracy, insertion or adhesives, cost, or discomfort.

"There is a need for more CGM options to increase patient access," Pettus said. That notion was echoed by a majority of 19 speakers who addressed the panel during a public comment session, several of them study subjects for whom travel had been paid by the company.

Panel member George Grunberger, MD, medical director of the Grunberger Diabetes Institute, Bloomfield Hills, Michigan, told Medscape Medical News, "Clearly, this expands the pool of people. There are reasons why patients haven't used the currently available systems. The more options the better. There's no one that fits everybody. Some people had bad experiences, some are physically challenged in placement, or the sensor doesn't stick. There are many issues."

If approved, Grunberger said, the Eversense will "expose more patients and more clinicians to the world of CGM."

Accuracy Demonstrated in Pivotal Trial

The Eversense clinical program included three prospective clinical trials: PRECISE, which was conducted in 81 patients in Europe for the 180-day version approved there and used as supportive data for the FDA application; the pivotal PRECISE II, which was a 90-day US-based trial involving 90 patients; and PRECISION, which was a 90-day trial in 35 patients that tested a version of the system with revised features. In addition, data on the European patients are being collected in a registry.

In the nonrandomized, single-arm, multicenter PRECISE II, the system was accurate through 90 days, with an absolute mean relative difference of 8.5% compared with laboratory reference glucose measures, and 87% of the readings came within either 15 mg/dL or 15% of reference (the so-called "15/15 rule").

In addition, the system detected 96% of hypoglycemic excursions and 98% of hyperglycemic excursions, and 91% of the sensors functioned for the entire 90 days.

As with other CGMs, accuracy was somewhat less on day 1, with just 77% of readings meeting the 15/15 rule, Senseonics Chief Executive Officer Tim Goodnow, PhD, reported.

In PRECISION, patients wore two Eversense devices to compare a smaller second-generation transmitter that is now part of the proposed system with the first-generation transmitter used in PRECISE II, and 85% of readings met the 15/15 rule.

Some Safety Issues Mitigated With Design Changes

There were no device-related adverse events in either PRECISE II or PRECISION. In one patient in PRECISE II, the sensor couldn't be removed after 2 attempts, and the patient underwent surgery under general anesthesia to have it removed. In two other events, the removed devices were found to have missing end caps that were presumed to have been left in the patients' arms but were deemed not to present a danger.

Subsequent to that, revisions were made to both the sensor end caps and the blunt dissector tool to reduce those risks. Changes were also made to the glucose determination algorithm after PRECISE II, but that did not affect the sensor's output, according to Goodnow.

One of the questions FDA posed to the panel was whether they had concerns about the lack of data for the Eversense subsequent to the revisions. The consensus was that more data were needed, but that they could be collected postmarketing.

Panel member Kathleen Wyne, MD, PhD, an endocrinologist in Columbus, Ohio, told Medscape Medical News, "The algorithm seems to be an improvement, as do the smaller transmitter, the end cap and insertion device. I think the committee felt comfortable that those were positive changes. I do share the concern that we do need to test out the new device, but they really look like positive changes, so that's why it's reasonable to go ahead with the new device as it is, but obviously collect data as it comes into use."

There were no infections in either PRECISE II or PRECISION, and overall infection rates are less than 1%.

The Eversense would be contraindicated for use with intravenously administered (but not dietary) mannitol or sorbitol, because those can cause falsely elevated blood glucose readings. In addition, the proposed label would also include a warning about tetracyclines, which can falsely lower the readings.

Clinician Trainings Planned

Senseonics is developing a certification program to train clinicians in the insertion procedure, including a didactic session led by company-approved trainers, practice with simulated skin, and observed initial insertions and removals. Several endocrinologists who had participated in the studies testified that the procedure was easy to learn. Steven J. Russell, MD, from Harvard Medical School, Boston, Massachusetts, noted that as a diabetes-focused endocrinologist, "it's nice to do procedures again."

Senseonics is proposing a US postapproval study following serial insertions and removals for 2 years, involving 175 patients at up to 20 clinical sites. The primary safety endpoint would be the rate of device-related and insertion/removal procedure-related serious adverse events through 12 months of 7% or below, and the primary effectiveness endpoint, time in range (70 - 180 mg/dL) at month 12 vs month 1.

Several panelists also endorsed the establishment of a US-based registry similar to the one in Europe to track unexpected outcomes. Several panel members also advised collecting data on patient satisfaction. Grunberger added that, given that this device would a new element to usual endocrinology practice, clinician satisfaction should be assessed as well.

FDA advisory panel members are vetted for conflicts of interest and waivers are granted if necessary. None were required for this meeting.

For more diabetes and endocrinology news, follow us on Twitter and on Facebook


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: