Senseonics 2Q16 – Strong 8.8% MARD over 90 days in US pivotal, Eversense FDA PMA submission in next few months; launched in Sweden, with five EU countries coming in 2H16 with Roche and Rubin – August 9, 2016

Senseonics reported 2Q16 financial results this afternoon in a call led by CEO Dr. Tim Goodnow, the company’s second update following a ~$45 million IPO in March on the NYSE and reverse merger in December to go public. We enclose the top six highlights from the update, followed by a pipeline summary.

Top Six Highlights

1. Moments before the call, Senseonics released very positive topline results from the 90-day US pivotal trial (PRECISE II) of its implantable Eversense CGM, on-body transmitter, and mobile app: an outstanding MARD of 8.8% vs. YSI values taken at days 1, 30, 60, and 90 (range of 40-400 mg/dl). The study had an impressive 16,000+ paired CGM-YSI points, enrolling 90 patients at eight centers. An FDA PMA submission expected in the “next few months” – a very fast turnaround. An insulin-dosing claim is a possibility, though modeling data will be needed. Presumably this will be easier for companies with good accuracy data (especially few large sensor errors) to obtain following FDA action on Dexcom, which we expect this fall. While we do not expect a delay, the timing won’t be certain until we hear about it from FDA.   

2. The call reported “positive experiences” from the June launch of Eversense in Sweden with distributor Rubin Medical. No metrics were shared, but conservative guidance for 2016 was maintained: EU sales of “less than $1 million” for the full year. A controlled launch through new distribution partner Roche is expected in Germany in September, followed by Italy and Netherlands in early 4Q16. Launches in Norway and Denmark (with Rubin Medial) are also expected in 4Q16. Following Germany’s positive ruling on CGM reimbursement in June, Senseonics is working with Roche to apply for product-specific reimbursement; a code is expected in 4Q16. We’re looking to hear more on early impressions now from our patient and HCP friends in Europe.

3. A CE Mark amendment was filed in 2Q16 to extend the sensor life to 180 days (no approval timeline given). Dr. Goodnow anticipated that the “majority” of sensors will last to 180 days by the time of approval, and “all” Eversense sensors will last 180 days within “12-15 months.” Median sensor life was 149 days in the EU pivotal, with 40% lasting out to day 180.

4. Senseonics plans to file a CE mark application in 3Q16 for its next-gen smart transmitter, which would be 55% smaller than the existing version and “more ergonomically attractive” – we see this as a mission critical product improvement, given the current bulky size of the on-body transmitter (and the competition in Europe). Some patients have been surprised, given this is implantable, that there is an on-body element. We know the plan is to change this longer term – it’s not unlike the early versions of Dexcom’s CGM that improved dramatically but were quite different over time from the first generation.

5. Health Canada has approved a 180-day pilot pediatric study (n=30) that is slated to begin at a “leading diabetes center” in Toronto in October. No precise age range was given. Data from this study is expected to support an IDE application for a subsequent pediatric trial in the US.

6. As of June 30, Senseonics reported a cash balance of $36 million, expected to sustain operations through 3Q17. Cash was nominally down only $1 million from 1Q16, following an expanded debt facility agreement signed in the quarter ($15 million drawn upon completion). Actual cash burn was not disclosed. Net loss was $12 million, up from $7 million in 2Q15.

Top Six Highlights

1. Moments before the call, Senseonics released positive topline accuracy results from the 90-day US pivotal trial (PRECISE II) of its implantable Eversense CGM: an 8.8% MARD vs. YSI. The trial wrapped up in July (it only started in 1Q16!), and a PMA submission is expected in the “next few months.” Assuming a 12-month PMA process (the 1Q16 expectation), launch could occur in roughly late 2017. According to Dr. Goodnow, this was the largest primary CGM study completed to date – 90 adults were enrolled at eight centers across the country, generating over 16,000 comparative CGM-YSI data points (40-400 mg/dl) and over two million glucose readings! Patients were blinded to the actual real-time CGM readings, but the device used two fingerstick calibrations per day and calculated values prospectively. We believe the request to blind patients to the readings came from the FDA – perhaps because the 90-day wear time might change patients’ behavior. The accuracy surpassed management’s expectations of a 10.5% MARD, exceeded the ~11% MARD in the EU trial (algorithm improvements), and compares favorably to Dexcom G5’s industry-low MARD of 9.0% and Abbott’s ~11% MARD for the factory calibrated FreeStyle Libre. As mentioned in 1Q16, full PRECISE II results will be shared at DTM 2016 in November.

• Without going into detail, the highly regarded Mr. Goodnow expressed that the safety data was positive and that the implantable CGM is very durable. We assume the durability was similar to or slightly better than the EU 90-day data, where 88% of sensors lasted to 90 days. The median sensor life in the full EU 180-day study was just shy of five months, with 40% of sensors lasting out to 180 days. The company is working on improving this figure (see more below), though as we understand it, the initial FDA submission will follow the path in the EU – a 90-day indication first.
• Dr. Goodnow suggested an insulin-dosing claim is a possibility, given this MARD, though it was not clear if it will be part of the initial PMA. More modeling data will likely be needed. The Eversense system has now shown clinical accuracy in humans based on PRECISE (EU) and PRECISE II (US), but Dr. Goodnow was honest in Q&A that more in silico data will probably be necessary. The company has not spoken to the FDA about these pivotal results or expectations for modeling. Dexcom presented extensive modeling data at the FDA meeting on a replacement claim last month, and FDA-Dexcom discussions are now ongoing regarding the expected training materials and post-market study. It is not clear when these will end for Dexcom. It would be impressive if Senseonics could generate enough modeling data and finalize training materials in the next few months to support a dosing claim in the initial PMA filing – as a reminder, Dexcom has had ongoing discussion with the FDA since fall 2014. The first company to get this claim will always have the most work to do – this is probably going to be a situation where Senseonics will benefit from this work by Dexcom, assuming they have enough modeling data on large sensor errors, .
• Following PRECISE, the EU pivotal trial that yielded a MARD of 11.6% vs. YSI (180-day results presented at ADA 2016), Senseonics tweaked the sensor’s algorithm to generate the improved MARD in PRECISE II. A nearly-three percentage point improvement in MARD from an algorithm adjustment is very strong, but not unheard of – Dexcom went from 13% with G4 Platinum to 9% with Software 505. Still, 8.8% is very strong, particularly for a 90-day sensor with this many paired CGM-YSI points. For context, the US 90-day trial length vs. the EU 180-day trial length was likely not a factor: MARD results in the EU pivotal were consistent (~11%) at both 90 and 180 days.
• 15 of the 90 subjects had two sensors inserted. We wonder why this design was implemented – perhaps to investigate intra-subject sensor variability? We look forward to finding out at DTM 2016.

2. Senseonics launched Eversense in June in Sweden with distributor Rubin Medical. No launch metrics were shared, but guidance shared in 1Q16 was maintained: full year 2016 revenue of under $1 million. A controlled launch through new distribution partner Roche is expected in Germany in September, with Italy and the Netherlands to follow in early 4Q16. Norway and Denmark launches with Rubin Medial will occur in 4Q16, on par with expectations from 1Q16. Dr. Goodnow stated that reps had positive experiences training healthcare providers on how to insert the device, and the procedure took under five minutes. VP of Global Sales and Marketing Ms. Mirasol Panlilio added that the insertion training was very easy, and the Swedish patients are all “up and running” and using the systems at home. Neither speaker commented on how many patients had been fitted with the device, but Dr. Goodnow mentioned that enrollment in the required post-market study has not been a deterrent to acceptance. As for the Roche partnership in Germany, Italy, and Netherlands, local commercial teams are in the midst of completing or initiating the in-market controlled rollout and setting up commercial infrastructure to support initial clinics and users.

• The press release reported $19,000 in revenue in 2Q16 (not mentioned in prepared remarks), though we assume this was not from Eversense sales. Senseonics actually reported a similar $23,000 in revenue in 2Q15 at a time when Eversense was not available commercially; we assume both sales figures came from technology licensing or other small sources, but are confirming with management.
• Senseonics is working with Roche on an application for product-specific reimbursement in Germany and expects that a code will be available in 4Q16. Management was delighted with the German Federal Joint Committee’s (G-BA) ruling in mid-June, which allows the Eversense, among other CGMs, to be reimbursed. Pricing has not yet been established, as the company will have to negotiate on a payer-by-payer basis, but management has suggested comparable pricing to current CGM in the past. Dexcom’s remarks last week suggested there are a few steps remaining in the process, and revenue from Germany isn’t expected to ramp until 2017. We assume Senseonics is in a similar boat, though Roche is obviously a terrific ally to move reimbursement quickly. 

3. A CE Mark amendment was filed in 2Q16 to extend the sensor life to 180 days. No approval timeline was shared, given the delays in the original CE Mark. Dr. Goodnow anticipated that the “majority” of sensors will last to 180 days by the time of approval, and “all” Eversense sensors will last 180 days within “12-15 months.” Median sensor life was 149 days in the EU pivotal, with 40% lasting out to day 180 – there is presumably some R&D or manufacturing work to do to improve the reliability further. We do wonder how frustrating it will be for patients if the sensor unexpectedly dies, at which point they would need to go back into the clinic for a re-insertion.

4. Senseonics plans to file a CE mark application in 3Q16 for its next-gen smart on-body transmitter, which would be 55% smaller than the existing version and “more ergonomically attractive.” We see this as a mission critical product improvement, given the current bulky size of the on-body transmitter. In his closing remarks, Dr. Goodnow added that the company will continue to work on sensor accuracy, and has planned for improvements in the next 12-18 months. As for calibration, the current goal is to cut down to one per day, but during Q&A, Dr. Goodnow shared his dreams of seeing calibrations cut down to one per week, and eventually eliminated altogether. There were no comments on the compelling smartphone “swipe” technology that could eliminate the on-body transmitter entirely. See the pipeline summary below for more details.

• The press release added that Senseonics plans to release enhancements to the mobile application later this year. The app is currently available for Android and iOS. It receives and displays sensor glucose readings every five minutes and provides trend information. We’re not sure what to expect from the updates.

5. Health Canada has approved a 180-day pilot pediatric study (n=30) that is slated to begin at a “leading diabetes center” in Toronto (Banting & Best Diabetes Center? St. Michaels?) in October. The trial is currently not posted on clinicaltrials.gov (we presume still in the planning phase), but the design will reportedly be very similar to that of both PRECISE studies. No precise age range was given, but patients will use the device at home at all times, with regular clinic visits for accuracy measurements. The safety and accuracy data from this study will be used to support a subsequent IDE application for a pediatric trial in the US.

• Uptake in the pediatric population will be interesting – though the initial application is more invasive, it will eliminate painful and annoying weekly/bi-weekly sensor changes, with no apparent detriment in accuracy. Assuming the external on-body transmitter gets smaller, we wonder if the potential in pediatrics is actually higher than adults. Alternatively, parents may be more conservative about their child trying an implanted device. It’s hard to know, though it sounds like it will be some time before a pediatric indication is approved in the US.

6. As of June 30, Senseonics reported a cash balance of $36 million dollars (including net proceeds from the recently expanded debt facility), expected to sustain operations through 3Q17. The runway is identical to the 1Q16 call. Management did not disclose specific cash burn (we assume around $5-$10 million), but cash is only down $1 million nominally from last quarter following proceeds from the new debt agreement (+$15 million minus an unspecified sum used to pay off previous debt).

• Net loss increased to $12 million for 2Q16 vs. $7 million in 2Q15, due in large part to product development expenses for future versions of Eversense and costs associated with the US pivotal trial. Sequentially, net loss increased slightly from 1Q16’s $11 million.
• In June, Senseonics established a new debt facility arrangement with Oxford Finance and Silicon Valley Bank providing loans up to $30 million dollars. The company borrowed $15 million upon signing the agreement, some of which was used to pay off existing debt in 2Q16. Three installments of $5 million each can be subsequently drawn “based on achievement of specific milestones.” Presumably FDA approval is one of them.
• Senseonics is currently trading on the NYSE at $3.00 (market cap ~$290 million). The market cap is consistent with 1Q16. Senseonics has only been publicly traded for two full quarters following the IPO in March that raised $45 million – see our detailed coverage here.

Pipeline Summary

Close Concerns Questions

Q: What were Eversense sales in the quarter?

Q: How many patients have had Eversense inserted in Sweden?

Q: How is Eversense priced in Sweden? Will Eversense be accessible in markets where there is no CGM reimbursement?

Q: What is the status of the smartphone swipe project?

Q: How long will FDA review take? Will the company obtain an insulin-dosing claim?

Q: Can Eversense expand the CGM market? What type of patient will it most appeal to?

-- by Brian Levine, Adam Brown, and Kelly Close