Memorandum

FDA approves Dexcom G5 mobile system ahead of schedule; will ship in late September – August 25, 2015

Executive Highlights

  • This morning, Dexcom announced FDA approval of the G5 mobile system well ahead of expectations. It will begin shipping in late September.
  • Dexcom’s G5 mobile system uses the same G4 Platinum sensor and Software 505 algorithm (MARD: 9.0%, seven-day wear), but a new smart Bluetooth transmitter that can talk directly to a new Apple iOS mobile app (Android is said to be coming in early 2016) – this eliminates the requirement for a receiver, meaning patients can get real-time CGM data and alarms directly on their phone. The transmitter can pair with two devices so people who do not want to be looking at their phone can use a transmitter (and also presumably have it as a backup).
  • Current Dexcom users will be able to upgrade to the new transmitter for a low-cost cash price – if they bought Share since August 1 (seems like yesterday!) it’s free. More details will come in the 3Q15 call in November.
  • The G5 approval is an important convenience win for patients (viewing data right on the phone) and families, and the big question is to what extent it can expand the CGM market and improve diabetes management.

This morning, Dexcom announced FDA approval of the G5 mobile system, which will begin shipping in late September (this company wastes no time!). The approval comes months ahead of the previous “end of 2015” expectation, reflecting a remarkably speedy six-month (!) PMA review from the February submission. The approval is a landmark for the CGM field – a class III PMA device sending data directly to a consumer mobile platform – and a testament to a forward-thinking FDA and Dexcom’s smart stepwise approach to connectivity (Share cradle last fall, Share receiver in March, now G5). This project has been a long time coming – the first public mention of Dexcom’s Gen 5 mobile system came more than four years ago at JPM 2011!

As we covered previously, Dexcom’s G5 mobile system is all about connectivity: it uses the same G4 Platinum sensor and Software 505 algorithm (MARD: 9.0%, seven-day wear), but a new smart Bluetooth transmitter that can talk directly to a new Apple iOS mobile app (Android is coming in early 2016) – this technically eliminates the requirement for a receiver, meaning patients can get real-time CGM data and alarms directly on their phone. A receiver can still be used if desired, as the new transmitter can speak to two separate devices via Bluetooth. (And based on prior calls, we believe patients will be required to have a “backup receiver.”) The approval stretches down to patients two year old (retaining the Software 505 pediatric approval secured in June), and G5 still includes the Share remote monitoring features. We saw the first pictures of the new G5 user interface at JPM 2015, which are included below.

Upgrade details are scarce at this point, with more detail expected in the 3Q15 call in November. Dexcom has always been generous on the upgrade front. For now, here’s what we know:

  • Current Dexcom patients under warranty will be able to upgrade to G5 for a “low cost cash upgrade.” We assume ~$199 based on the Share receiver upgrade program, but that may depend on what receiver patients currently have (more below).
  • New purchasers of a G4 Share receiver (Aug 1 – late Sept) will get a no-cost upgrade to the G5 system. We’re not sure what is magic about the August 1 date.

Data from G5 will be integrated with “Dexcom Clarity,” the SweetSpot-developed cloud-based reporting software alluded to in the 2Q15 call. That’s a cool name, Clarity! Today’s announcement said this will include “personalized, easy-to-understand analysis of trends that may improve diabetes management.” We cannot wait to see what that entails, but assume it will give Dexcom users and HCPs a more competitive offering to Medtronic’s CareLink. We hope they have studied the data that Abbott’s Libre customers get “on-screen” – it is fantastic, including average glucose, time in zone, daily chart, etc. all over the last 7 days, 14 days, 30 days etc. There is no hiding trends with this system – we have found it very motivating.

The G5 approval is an important convenience win for patients (viewing data right on the phone), and the big question is to what extent it can expand the CGM market – how many patients are unwilling to go on CGM right now because they don’t want to carry an additional receiver? That is the pie that G5 can really expand CGM. For current Dexcom users, this is all about a better product experience and improved retention. We see the approval as a potential win for HCPs, since G5 should automatically upload the data to what sounds like a more robust cloud-based data platform (Dexcom Clarity).

From a competitive point of view, the G5 approval continues Dexcom’s dominance on the CGM connectivity front, which comes ahead of improvements from Medtronic, Abbott, and others. On the positive side, it does signal a very open FDA and could accelerate future products from these other companies. We share more details below based on what we know about G5 from previous Dexcom calls, the approval’s impact on patients, HCPs, Dexcom, pump partners, and CGM competition. We conclude with our key questions and screenshots of G5 from JPM.

Gen 5 Details

  • We learned in the 1Q15 call that the new G5 mobile transmitter will have a labeled useful life of three months, down from six months for the G4 Platinum and 12 months for Seven Plus. The reason is because the Bluetooth radio requires more power, and Dexcom wanted to keep the Gen 5 transmitter on-body footprint the same as the current G4 transmitter – a tough tradeoff, but we think the company made the right call, all things considered (we’re betting patients would rather have a smaller, shorter-lasting transmitter than something clunkier and longer lasting though obviously that’s easier for patients with higher disposable income). What is unclear is how long the G5 transmitter will last “in real life”, since the current G4 transmitter tends to last much longer than the labeled six months. Nevertheless, this is a negative overall.
  • The 1Q15 call also disclosed that the FDA would likely require each new patient to purchase a Dexcom receiver – we assume that is still the case, but are confirming with management. The G4 Platinum Share Receiver is upgradeable to work with the Gen 5 Bluetooth transmitter, so those who already have a Share receiver shouldn’t need a new one (as far as we understand). In practice most patients will probably use their iPhone/smartphone as the primary receiver anyways, so the required receivers may simply sit in patients’ drawers (our wording) ... Of course, those on Android phones will need a receiver for the time being, as a launch for that platform isn’t slated until early 2016. We wonder if the new partnership with Google could help Dexcom advance more quickly on an Android version of G5. 
  • There looks to be variability on the upgrade front, depending on what patients have. Those on the Share receiver (launched in March) will need a new transmitter, a free download of the iOS app, and based on prior Dexcom calls, a remote software update to make their Share receiver compatible with the new Bluetooth transmitter. Those on the prior G4 receiver (non-Share version) will need the new transmitter, the new app, and presumably, the offer to purchase an “optional” receiver.
  • Dexcom conducted “many, many hours of human factors studies” to develop and refine the G5 mobile app’s user interface. In the 4Q14 call, Mr. Sayer said it is “completely different” from what current patients experience, and it “looks like other apps in the mobile environment.” We agree based on screenshots shown at JPM 2015 – these are attached below.  
  • The G5 app leverages the already-launched Share remote monitoring system, where up to five followers can receive alerts/notifications of CGM data. Otherwise, G5 use the G4 Platinum sensor and the new G4AP algorithm.
  • Dexcom plans to launch Android applications for G5 “early next year” – this is the most specific timing we’ve heard on this front. As a reminder, there is no current Android version of the Share app (what patients would have on their phone), though Dexcom has launched the Follow app for Android (what caregivers would have on their phone). As we understand it, the wider variety of Android phones means more software testing is required than iOS to gain approval.
  • Since G5 uses the same G4 Platinum sensor and Software 505 algorithm, Dexcom can immediately launch an insulin-dosing label claim – based on the 2Q15 call, an insulin dosing claim is expected to launch sometime in 2016. Dexcom is now “in sync” with the FDA regarding the pre-market and post-market data required to obtain such a claim for CGM. For current users, we think this label claim may not be that impactful, since many Dexcom users dose off the CGM readings anyways. However, it could persuade more patients to wear CGM, since the need for confirmatory fingersticks does dilute the value proposition of CGM somewhat.
  • We’re not sure of the EU and global timing for a G5 launch. Per the 2Q15 call, international sales represented 13% of total revenue, so the US is the obvious focus for now.

Impact on Patients

  • G5 mobile brings a slew of meaningful product improvement benefits to patients: (i) greater discretion (for the most part!) and data viewing options with the new transmitter (smartphone only, smartphone + receiver, smartphone + pump, receiver only) – we do note some people do not want more reasons to pull out their phones; (ii) a much better user interface on the smartphone than on the receiver; (iii) the addition of alarms and notifications for the CGM user (those currently using the Share receiver and Share2 app cannot get notifications on their own phone); (iv) potentially lower out-of-pocket expenses for some Dexcom users (less need to buy new receivers when broken/lost, since data can be accessed via the phone – although it may be required that patients buy their first one), though some may spend more since the transmitter has a shorter useful life; (v) presumably improved data analytics through the G5 mobile app and Dexcom Clarity; and (vi) a growing ecosystem of apps that leverage data flowing directly to the cloud; this has already started with the Share receiver posting data to Apple Health, but will accelerate greatly once Dexcom patients are widely on G5. Presumably apps will only continue to get better though there is not a business model with most.
  • Among CGM users in the diabetes market research company dQ&A patient panel (1Q15), one in three patients (32%) listed smartphone integration in their top three most desired improvements (only exceeded by better accuracy and longer sensor wear). Of course, Dexcom is working on the latter with Gen 6. We also wonder how much perceptions of inaccuracy are based on comparing the CGM value to the BGM value, which will gradually decrease over time with an insulin dosing claim and the elimination of calibrations (with Gen 6 and beyond).
  • About a third of users in the dQ&A patient panel find that the Share feature has helped them avoid hypoglycemia (39%) and hyperglycemia (33%) and achieve more time in zone (36%). We assume this stems at least in part from caregivers getting notifications in addition to patients seeing directional arrows when they are in zone; presumably, the addition of alerts and alarms on the G5 mobile app for users will further enhance these numbers.
  • More than half (53%) of Dexcom Share users in the dQ&A panel agree that the system helps others worry less about their diabetes, and nearly half (46%) feel safer using Share. As noted above, G5 will preserve the Share remote monitoring features.
  • For desired improvements from patients not on CGM, as well as lots more patient feedback on CGM contact Richard Wood at Richard.wood@d-qa.com.

Impact on Healthcare Providers

  • G5 CGM data should flow directly to Dexcom’s new Clarity software, offering HCPs less hassle for downloading. Assuming Clarity brings more robust pattern recognition (and it sounds like it will, based on the name alone!), this advance would certainly be welcome for HCPs. Medtronic’s CareLink Pro gets high marks from providers on this front, and Dexcom has lagged behind in lieu of improving its sensor accuracy and connectivity. The challenge remains integrating CGM data and interpretation into clinical practice, particularly notoriously hated electronic medical records – we’ll be interested to hear HCPs’ thoughts once Clarity launches.
  • Presumably, G5 will make HCPs more likely to prescribe CGM. The value proposition for CGM becomes more compelling with the ability to get data directly on one’s phone, along with an insulin dosing claim expected sometime next year. As well, it should be much easier for HCPs to get data and for patients to understand it.

Impact on Dexcom

  • Dexcom has said in the past that G5 will be a “major catalyst in driving new patient growth.” Management also expects that the improved user interface and data sharing capabilities could have a positive impact upon patient retention and sensor utilization. We have no doubt on either front – sending the data directly to the phone is a huge leap forward, especially for those reticent to get on CGM.
    • As of the 4Q14 call, Dexcom’s patient base totaled roughly 90,000 patients (60% pumps/40% MDI). We wonder if the G5 approval could dramatically expand the MDI population on Dexcom CGM, shifting it more towards the split in the overall population (~30-40% on pumps, 60-70% on MDI in the US).
    • Dexcom has experienced nine consecutive quarters of 55%+ YOY growth, and we expect that to continue in 2H15 with this approval and launch. In the 2Q15 call, management increased the full-year 2015 revenue guidance to $350-$375 million (36%-46% YOY growth). CEO Kevin Sayer was candid in characterizing the updated guidance as “pretty ambitious. We love this guy – he calls a spade a spade. Dexcom had a back-loaded year last year, and with the upcoming launch of G5, a similar performance is to be expected this year. That said – the comparisons aren’t easy!
  • G5 might improve Dexcom’s profitability, as the company’s sales mix should shift more towards higher margin sensors and away from lower margin receivers. Receiver upgrades are expected to decrease as patients use their phone as the primary display device. Management expects total transmitter revenue to possibly increase, given the shorter useful life of three months, down from six months for the G4 Platinum. We assume the margins are better on the transmitters than receivers, but we cannot recall management giving this level of granularity. As well, we’re not sure if patients will be required to buy receivers – that’s still to be figured out. As noted earlier, we’re not sure how often new transmitters will be required “in real life” – but if that is more than previously, that’s also a benefit to the revenue side.
    • On the plus side, the shorter transmitter lifecycle will shorten patient upgrades to new platforms. For example, when Gen 6 is launched, the new smart transmitter and corresponding app will be available to all existing patients during their normal purchasing cycles within a few short months.
  • Once Gen 5 rolls out, Dexcom will be able to run real-time analytics across its installed base – that should bring tremendous insight into how CGM is being used in the real world. Per the 2Q15 call, Dexcom recently hired a Senior Data Executive from outside the medical technology field who has years of experience in Big Data in Silicon Valley – we believe this person is Annika Jimenez, formerly the Global Head of Data Science Services at a cloud-computing company called Pivotal. She is now listed on LinkedIn as Dexcom’s SVP of Data.
  • Today’s news continues an impressive cadence of new product updates and partnerships this year:

Impact on Pump Partners

  • Near-term, the Animas Vibe (already available) and the t:slim G4 (2H15 launch) are not compatible with the G5 transmitter, meaning those patients will have to choose between getting data on their phone (G5) or pump (G4). Management has expressed skepticism that patients will want to pull their pump out to check their CGM vs. getting the same information (in a much better user interface) on their phone. We believe different patients will value different display devices, and the option value of multiple viewing screens is a major plus of Gen 5. That said – we’ve heard from several people that pulling out their phone makes them a little uncomfortable since it seems like they are focused on email or social media, not the person to whom they are speaking. As well, some have said that they do not like the alarms at night on the iPhone. Overall, the reviews for the G5 for those people have used it in trials are not uniformly positive – that said, this is the start, and it’s only going to get better.
  • In the medium term, Animas, Bigfoot Biomedical, Bionic Pancreas, Insulet, and Tandem will presumably use Gen 5 and/or Gen 6 in future products. As a reminder, Dexcom is committing future technology (Gen 5, Gen 6) to pump partners only if they can commit to products that go beyond CGM display and offer some measure of insulin control. The 1Q15 and 2Q15 calls shared Dexcom’s strongest comments yet on the artificial pancreas: “...automated insulin delivery is going to come ...We want a product in that space. We want a partner to be in that space.”
    • Insulet will likely be the first out with a Dexcom G5 integration, assuming its next-gen PDM launches as expected in 2016 – there is not complete certainty to this, however, as we discuss in our Insulet 2Q15 report. Tandem has not given timing on a G5 integration (the G4 integration is still under review!), and J&J has not formally disclosed plans for a G5 integration, though a next-gen automated system is reportedly in the works.
    • The big question for pump partners is how quickly they will move on pivotal studies, considering that Medtronic has already started the pivotal study of the MiniMed 670G/Enlite 3. It also remains unclear how tightly the 670G will manage blood sugar, how good the Enlite 3 sensor will be, and how the integrated 670G will compare to Dexcom-integrated offerings from the aforementioned contenders. Lots of uncertainty, but unquestionably plenty of exciting movement and competitive offerings for patients!

Impact on CGM Competitors

  • This approval continues Dexcom’s dominance on the CGM connectivity front relative to competitors.
    • Medtronic’s MiniMed Connect keychain device will launch this fall for $199 (approval announced in June), but it will require patients to carry a small fob to relay CGM/pump data from the pump to a nearby phone. Medtronic’s Guardian Mobile system (more like G5 with a Bluetooth transmitter that would talk straight to a phone) is in a pivotal trial that was slated to wrap up in July; we’re not sure if it has ended, per the ClinicalTrials.gov posting. Medtronic has never given timing on this product.
    • Abbott has never disclosed official plans to make FreeStyle Libre or Navigator II cloud-connected, but we assume these must be in the consideration set. Given that many phones have NFC chips built in, we wonder if Abbott could create an app that reads and displays data from the small FreeStyle Libre sensor patch – that is speculation on our part! In general, we’re imagining they’re pretty on it.
    • Senseonics’ implantable CGM will communicate with a smartphone app, with a launch possible in Europe later this year. We’re not sure of the EU timing for G5, and it’s hard to know how much patients will balance the upside of Senseonics’ implantable CGM with the downside of a fairly large on-body transmitter.
    • Echo Therapeutics and Sano are working on “wellness” CGMs that pair with smartphone apps, though both are behind G5. Sano hopes to launch in 2016, while Echo is still finalizing its design details. Assuming these make it to market (a big ‘if”), we do not believe these will compete successfully with G5, since they presumably will aim for wellness indications to avoid the scrutiny of a PMA submission. It’s hard to know what that will look like.

G5 Mobile App Screenshots

 

Close Concerns Questions

Q: Is FDA actually requiring a receiver, or is it simply for redundancy? 

Q: To what extent will G5 expand the number of patients on CGM? Will G5 dramatically open up the MDI market for Dexcom?

Q: To what extent will G5’s Bluetooth connectivity pose an additional drain on smartphone battery?  

Q: How will G5 affect Dexcom’s profitability?

Q: What is the global launch timing on G5, particularly in the EU?

Q: If a patient is using G5 app and the phone dies, can they seamlessly transition to the receiver (since the calibration is in the transmitter)?

Q: Does the G5 app completely replace the receiver in terms of functionality (pairing with the transmitter, start, stop, calibration and error messages)?

Q: If the FDA has achieved “comfort” with the idea of full-function CGM on commercial smartphones (i.e., not just secondary display like Share), how far behind are they from running a pump off a commercial smartphone? What does this portend for the use of commercial smartphones as a component of artificial pancreas devices?

Q: Dexcom keeps saying that their task is not only to provide patients with access to their glucose data (the original premise of CGM), but also to help them better manage their diabetes. What are they actually doing in this area? Will the focus of progress in diabetes technology shift to pump companies?

Q: How quickly will Dexcom’s pump partners move on G5 integration? How will patients weigh pump vs. phone integration?

Q: Will the “smart transmitter” concept become the industry standard? Will Abbott and Medtronic go in this direction? What about CGM start-ups? If any pump company can grab any sensor’s Bluetooth signal, how will that effect the competitive market for CGM?

 

-- by Adam Brown and Kelly Close