12th Annual Diabetes Technology Meeting

November 8-10, 2012: Bethesda, MD Full Commentary – Draft

Executive Highlights

The enclosed report brings you our full coverage of the 12th Annual Diabetes Technology Meeting (DTM), held November 8-10, 2012 in Bethesda, Maryland. The annual update on all things diabetes tech gets better and better every year, and this year’s meeting attracted more than 400 attendees from 24 different countries. In his closing remarks, DTM chair Dr. David Klonoff highlighted his key takeaway from this year’s meeting: “the importance of information” – including how to obtain, transmit, and analyze information, with the ultimate (and often challenging) goal of improving treatment.

Our favorite sessions of the conference provided the latest updates and smartest thinking on the artificial pancreas (AP). We heard about the newest upcoming studies and enjoyed two rich panel discussions where some of the biggest names in AP research weighed in on system design considerations - Do we need glucagon? Should nighttime systems relax control? Should closed-loop systems be fully reactive, or include manual prandial insulin dosing? While certainly many challenges and debates are yet unsolved (both design- and regulatory-related), there’s no question that outpatient studies are now increasingly the norm (amazing!), and will become more and more ambitious over the coming years.

FDA was highly present at DTM 2012, both through four different Agency presentations and a number of questions during Q&A – this active involvement was great to see and encouraging progress from more passive (or even silent or even nonexistent) participation in the past. Big kudos to the FDA for being willing to join in the discussion. Notably, FDA’s CDRH Director Dr. Jeffrey Shuren’s keynote address highlighted the steps being undertaken to improve its very underfunded regulatory review process, including a public-private partnership to address this issue and the ongoing development of better tools and software. Notably, FDA published the final guidance on the artificial pancreas during DTM (see our report at https://closeconcerns.box.com/s/d2vqcjbpmgvxf77crqjn).

The conference also looked at some of the challenges and successes of currently available technologies. Continuous glucose monitors and bolus calculators were at the forefront of the discussion with an award winning CGM comparative accuracy study (Dexcom’s G4 Version A (the Animas Vibe version, not the G4 Platinum) vs. the Abbott FreeStyle Navigator I vs. the Medtronic Enlite) and two sessions exploring the use of algorithms and clinical impact of bolus calculators. There were other sessions devoted exclusively to insulin delivery, novel insulins (lots of commentary on degludec) and glucagon, glycemic variability, sensor design, and social media.

This report contains our full coverage of DTM 2012. Below, we discuss the major themes and our big- picture takeaways from the conference, followed by our coverage of individual presentations. Talk titles highlighted in yellow were not previously published in our daily reports, while talks titles highlighted in blue were some of the most memorable presentations we heard.

  • We were encouraged to hear updates on multiple ongoing and upcoming outpatient artificial pancreas studies. Notably, Dr. Edward Damiano (Boston University, Boston, MA) hopes to begin his team’s five-day, outpatient transitional study (the Beacon Hill study) this month, which will test an iPhone 4S controller/user interface that communicates with a Dexcom G4 CGM and two Tandem t:slim pumps (insulin and glucagon). The Investigational Device Exception (IDE) for the system was submitted to the FDA in the week preceding DTM. In the same session, Dr. Claudio Cobelli (University of Padova, Padova, Italy) gave updates on theAP@home project: the consortium has begun a set of overnight, partially outpatient experiments (n=12) that use the University of Virginia’s latest DiAs mobile-phone-based controller, a specialized remote monitoring system, an algorithm that incorporates recent insulin delivery information in its control decisions, and a simplified protocol for open-loop insulin dosage at meals. Additionally, in mid-2014, Dr. Frank Doyle (University of California, Santa Barbara, Santa Barbara, CA) expects to begin an eight-week outpatient trial as part of the NIH DP3-grant funded Ambulatory Control project, which will assess a new periodic zone model predictive control (MPC) closed-loop algorithm. Further, Dr. Boris Kovatchev (University of Virginia Health System, Charlottesville, VA) presented encouraging first-patient results from an outpatient, closed-loop efficacy trial using a new modular algorithm, the Dexcom G4 Platinum CGM, and Tandem’s t:slim insulin pump.

  • To what extent should artificial pancreas algorithms loosen control overnight? Dr. Frank Doyle explained that while wide agreement exists that closed-loop control should be relaxed overnight in order to minimize the risk of hypoglycemia, there is not yet consensus on the best approach to do so. UCSB researchers have addressed this issue with a periodic zone model predictive control (ZMPC) closed-loop algorithm that smoothly adjusts the boundaries of the zone from their daytime values (80-140 mg/dl) to wider nighttime values (110-220 mg/dl). Looser control that minimizes the risk of hypoglycemia is of course much safer for patients, though it could also raise A1c – we look forward to better understanding how this will be viewed from a regulatory, reimbursement, physician, and especially a patient perspective. That said, looser overnight range control could also leave A1c unchanged, or even reduce A1c if the reduction in hypoglycemia brings a consequent reduction in rebound hyperglycemia. We are glad to see that the FDA’s final AP guidance (see our report at http://www.closeconcerns.com/knowledgebase/r/d8ed7b95) provides flexibility for study sponsors to demonstrate different endpoints.

  • Will we see insulin-glucagon control in the near future? Turning to bihormonal control with insulin and glucagon, Dr. Damiano’s group is “relying on the importance of glucagon” for their fully closed-loop system; however, the development of a stable, pumpable glucagon certainly has drug and device R&D challenges, not to mention a potentially higher regulatory bar. We were encouraged to hear progress on the glucagon front from Dr. Steve Prestrelski (Chief Science Office, Xeris Pharmaceuticals, Austin, TX) – Xeris is developing a non-aqueous stabilized glucagon formulation for use in an auto-injector pen for severe hypoglycemia (G-Pen), a mini dose pen for mild/moderate hypoglycemia (G-Pen Mini), and a formulation for the bi-hormonal artificial pancreas. The compound is currently preclinical, though an IND enabling program has been agreed upon with the FDA and the quicker 505(b)(2) regulatory pathway will be used. Xeris will begin a phase 2a clinical trial in 1Q13 under PI Dr. Ralph DeFronzo (Texas Diabetes Institute) and data is hopefully expected early next year. Dr. Prestrelski explained that if the product can be smoothly licensed, it could come on the market as soon as 2014. (This would be ahead of Biodel’s stable glucagon formulation, which is expected to be filed with the FDA in mid-2014 per comments at the October 12 Analyst Day.)

  • Is fully reactive closed-loop control equivalent/superior to hybrid control that still requires patient input? Dr. Damiano’s aforementioned Beacon Hill study will also investigate whether fully reactive control (i.e., no meal boluses from the user) is better than hybrid control using pre-meal priming boluses. Dr. Roman Hovorka (University of Cambridge, Cambridge, United Kingdom) was outspoken during the subsequent Q&A, expressing his belief that “without prandial dosing [i.e., via patient input], I’m not sure it’s possible to get as good of control as with open-loop therapy. Dr. Damiano “entirely disagree[d] that you cannot beat open-loop controlusing a fully reactive system.” Dr. Boris Kovatchev added a brilliant regulatory perspective, noting that by using a range controller on top of traditional basal-bolus therapy (which includes prandial dosing and a basal rate controlled by the individual), an AP controller then has an “adjunct” claim (vs. a “replacement” claim). That strategy could accelerate the time it takes to get closed-loop systems through regulators and into the hands of patients.

  • FDA’s presence was certainly felt during DTM, with Dr. Jeffrey Shuren (Director, Center for Devices and Radiological Health) delivering the meeting’s keynote address. Dr. Shuren spoke about the challenges the FDA faces in medical device regulation (e.g., a measly $15 million budget for medical device regulatory science and a very inefficient and decentralized evaluation system) as well as CDRH’s efforts to improve the device regulation process and bring products to market in a more timely fashion – we certainly viewed the FDA’s publication of final guidance on the artificial pancreas, which occurred during DTM 2012, as a sign of major progress on this front (see our November 14 Closer Look at http://www.closeconcerns.com/knowledgebase/r/d8ed7b95). Further, Dr. David Klonoff presented the CDRH with this year’s Diabetes Technology Society Leadership award. Considering the significant impact FDA can have on product development and time to market, we were encouraged to see the Agency’s accessibility to industry and research representatives in attendance. FDA’s Lieutenant Quynh-Nhu Nguyen and Dr. Patricia Beaston also gave the Agency’s perspective on human factors and bolus calculators, respectively. Dr. Beaston emphasized that incorporating multiple diabetes management systems onto a single smartphone platform (a holy grail from a patient perspective) introduces error that would not be present in a standalone system – in her view, such a design introduces the possibility for various softwares to corrupt or interfere with the diabetes software. FDA’s caution on this front seems to be influencing device development; indeed, Dr. Damiano commented during a panel discussion that moving his artificial pancreas system to a pivotal trial will “require a custom medical grade hardware/software platform that is dedicated to the task of controlling glucose.” (Currently, the controller for the system involves an iPhone 4S.) Broadly, we hope that researchers, companies, and FDA can collaborate on this front – given the shortage of providers and tightening reimbursement, we continue to believe mobile phones will be a major part of future diabetes care. Everyone is awaiting the FDA’s publication of final guidance on mobile medical applications, which we understand is on the CDRH’s “A-List” for release in FY13 (http://www.fda.gov/MedicalDevices/DeviceRegulationandGuidance/Overview/MDUFAIII/ucm 321367.htm?source=govdelivery).

  • Presenters called for better metrics to assess glycemic variability (GV) and a system to validate these metrics. Dr. Thomas Peyser (VP, Science and Technology, Dexcom) proposed two unconventional methods to measure glycemic variability. First, glycemic variability index, which measures the relative length of the line that would be shown in a CGM trace, and second, patient glycemic status, which would weigh both glycemic variability and overall glycemic control. We continue to hear new glycemic variability metrics being proposed, though we truly hope some consensus can be reached in the coming years – of course, it’s tough to build agreement since the clinical data is still ambiguous. [We continue to look forward to seeing the results of Dr. Irl Hirsch’s (University of Washington, Seattle, WA) 120-patient, eight-month pilot study called FLAT-SUGAR (ClinicalTrials.gov Identifier: NCT01524705). While this study’s main aim is feasibility as we understand it (showing that two groups can have the same A1c but different glycemic variability), it will pave the way for a larger outcomes study that could fundamentally change the way diabetes is treated and reimbursed.] More broadly, Dr. Natalie Wisniewski’s (Medical Device Consultancy, San Francisco, CA) expressed her belief that a widerarray of accepted clinical endpoints will allow more diabetes technologies to come to market. Given the FDA’s recently published finalized AP guidance, we believe regulatory may be trending in the right direction here in the US; of course, continued payor and association focus on A1c is still a limitation to overcome. Further, Dr. Wisniewski proposed that a working group, made up of industry, academic, government, and foundation leaders, should be established to standardize and validate endpoints. Dr. Klonoff was quick to respond and announced that the Diabetes Technology Society would organize such a group. We certainly look forward to following progress on this front, and hope that endpoints better suited to evaluate glycemic variability and other parameters like hypoglycemia will be addressed.

  • Companies have the potential to expand their social media presence in ways that benefit patients. Ms. Amy Tenderich (DiabetesMine; Alliance Health Networks, San Francisco, CA) explained that even if companies do not actively participate on social media platforms, people with diabetes do – they are constantly sharing opinions on products and studies regardless of whether the company is part of the conversation. She suggested that industry can benefit from this dialogue by gaining insight into patient needs and product likes and dislikes in order to build better next-generation devices. Ms. Kelly Close (Close Concerns, San Francisco, CA) further encouraged companies to take a patient-centric approach to social media, suggesting that the primary goal of their involvement should be to strengthen and sustain patients’ engagement in diabetes care. Further, Kelly highlighted that companies can potentially benefit by achieving this goal as better diabetes management often involves greater product adoption. To view the slides that Kelly presented, visit http://www.closeconcerns.com/knowledgebase/r/bb10d3a9.

  • Speakers expressed concerns about the discrepancies between commercially available bolus calculators and the impact of fat and protein content on insulin requirements. In a comprehensive comparison of available insulin pump bolus calculators, Dr. Bruce Bode (Atlanta Diabetes Associates, Atlanta, GA) underscored that no two bolus calculators are the same. He suggested that at the very least, the terminology used between bolus calculators needs to be standardized in order to avoid confusing patients switching from one pump to another. Dr. Howard Wolpert (Joslin Diabetes Center, Boston, MA) drew attention to the very real difficulties of carbohydrate counting and the effect of fat and protein intake on carbohydrate absorption patterns. He suggested that proper dietary education that directs patients with diabetes to meal and food choices that have less glycemic impact should supersede perfecting an insulin dose calculation algorithm. In a session on the clinical applications of bolus calculators, mainly those for MDI patients, the sentiment was mixed on whether calculators were uniformly efficacious; however, there was general agreement that bolus calculators should be made available because they will definitely be beneficial for some. We certainly agree that more options benefit patients, and we hope that Abbott can eventually get the European version of the FreeStyle InsuLinx here in the US (as a reminder, the European version includes a full-fledged bolus calculator, while the US version only allows logging of insulin doses).

Table of Contents 


1. Artificial Pancreas

Artificial Pancreas: Engineering Aspects


Edward Damiano, PhD (Boston University, Boston, MA)

Dr. Edward Damiano provided the latest update on his team’s quickly moving bi-hormonal work, a very valuable update from what we last heard at Children with Diabetes in July. Most notably, Dr. Damiano discussed the status of the five-day outpatient study that we’ve been looking forward to for over a year. The IDE for the controller device (an iPhone 4S communicating with a Dexcom G4 CGM and two Tandem t:slim pumps) was submitted to the FDA last week and the hope is the study can start in December. It will take place in Boston’s Beacon Hill neighborhood, a three square mile area downtown where subjects will able to roam freely while wearing the device with a nurse chaperoneduring the day. Dr. Damiano also presented interim results from the third inpatient study that is finishing up 51-hour experiments in 12 adolescents and 12 adults. Half of the study participants are getting adaptive pre-meal priming boluses while the other half are using a fully reactive closed-loop system (i.e., no meal boluses from the user). Notably, the closed-loop algorithm adapts over time, starting quite conservatively and then fine-tuning insulin dosing. The interim results look very solid in nine adolescents and 11 adults by day two of the study – mean blood sugars [projected A1c] of 143 mg/dl [6.6%] in adults and 171 mg/dl [7.6%] in adolescents receiving no meal boluses, improving to 138 mg/dl [6.4%] and 157 mg/dl [7.1%] when adaptive pre-meal boluses were used. We hope Dr. Damiano’s team can continue the momentum pending positive feedback from the FDA and IRB for the outpatient study.

  • Dr. Damiano’s much awaited five-day, transitional, outpatient closed-loop study will hopefully begin in December. The IDE for the controller device (an off-the-shelf iPhone 4S that communicates with a Dexcom G4 CGM and two Tandem pumps) was submitted to the FDA last week. He is awaiting feedback from the Agency and pending IRB approval (fingers crossed!), the study will start next month. It will take place in Beacon Hill, a three square mile neighborhood in downtown Boston. Patients (20 adults) will have the ability to roam freely (with a chaperone) during the day with unrestricted eating and exercise and point of care blood glucose testing. At night, they will sleep with a GlucoScout for reference blood glucose checks.

    • The iPhone 4S will run the control algorithm and communicate with two low-energy Bluetooth Tandem t:slim pumps (insulin and glucagon) and a Dexcom G4 CGM. The G4 will wirelessly stream data into the iPhone through a new custom hardware attachment connected through the 30-pin connector. This was an update over the system we saw at ADA in June and Children with Diabetes in July, which was hardwired to Abbott’s FreeStyle Navigator CGM receiver. Dr. Damiano was wearing the system during the presentation and showed the audience his real-time streamed blood glucose value from the G4 along with the Tandem pumps dosing saline.

    • The Beacon Hill study will test both fully reactive (no meal boluses) closed- loop control and closed-loop control with adaptive pre-meal priming boluses. For the latter, patients will select whether a meal is small, medium, or large, and pre-meal doses will be adapted over time by the algorithm. More broadly, the control algorithm itself will also adapt over time and fine tune dosing based on its performance and changing insulin requirements.

  • Dr. Damiano reviewed the design and interim results from his group’s ongoing third clinical feasibility study in 12 adults and 12 adolescents. The trial involves 51-hour experiments using the Abbott Navigator CGM as the input to laptop-driven insulin and glucagon control. The laptop directs dosing on two Insulet OmniPods. Participants ate six high carbs meals (the level of control achieved given the carb content is quite impressive) and had 30-40 minutes of structured exercise (4,000 heart beats). The algorithm initializes with only the subject’s weight and adapts over time – notable robustness considering both adults and adolescents are taking part in the study. Half of the adolescents and half of the adults receive adaptive priming boluses at meal presentation (i.e., the algorithm automatically changes the size of the pre-meal priming bolus over the course of the study), while the other half are on fully reactive control with no priming bolus.

    • Similar to previous trials, Dr. Damiano’s group tested CGMs head to head: Dexcom’s G4 Platinum and Abbott’s FreeStyle Navigator (first gen) – accuracy was very comparable. Dexcom’s G4 had a MARD of 12.3%, very comparable to the Abbott FreeStyle Navigator’s MARD of 12.6%. The CGMs werecompared to blood sampling every 15 minutes. Data was used from eight to 48 hours of closed loop experiments. The CGMs were inserted 24 hours before the first calibration. The system to be used in the new outpatient study (see above) will use the Dexcom G4.

    • Dr. Damiano displayed interim study results, demonstrating good average control and a low prevalence of hypoglycemia. He urged the audience to pay more attention to the slightly better day two numbers since the algorithm takes six to 12 hours to adapt to the patient and establish optimal control. In his view, these numbers are more predictive of how the system would perform for several months. Dr. Damiano also emphasized that the A1c’s achieved in adults and adolescents in both experimental conditions were much better than standard of care. Additionally, hypoglycemia was infrequent, though it remains to be seen if there will be an increase once the outpatient study gets going and patients are not so sedentary.


CGM Average


BG Average (mg/dl)

[Projected A1c]

% BG Values

< 70 mg/dl



  Day 1 Day 2 Day 1 Day 2 Day 1 Day 2  
Adults No Meal Bolus (n=5) 141 136





2.0 5.7 3.6
Adolescents No Meal Bolus (n=6) 165 156





1.6 0.5 4.5
Adults Auto Meal Bolus (n=5) 125 126





5.5 0.0 4.2
Adolescents Auto Meal Bolus (n=3) 162 147





0.0 0.7 5.2
  • Dr. Damiano briefly touched on his team’s second clinical feasibility study (just published in Diabetes Care), highlighting that children are much different from adults. Initially, the closed-loop algorithm performed well in six adults: an overall mean blood glucose of 158 mg/dl (68% in the range of 70-180 mg/dl, 0.7% <70 mg/dl) and a mean of 123 mg/dl overnight (93% in the range of 70-180 mg/dl, 0.5% <70 mg/dl). However, when the same system was brought into children, it could not get them in range – average BGs were 180-190 mg/dl with the same controller. The team iterated the algorithm and eventually generated a more adaptive system, which has since been used in the third feasibility study (see above) and will be part of the Beacon Hill study (also described above). It is initiated with only the subject’s weight and comes online with conservative dosing that adapts over time.


Frank Doyle III, PhD (University of California, Santa Barbara, Santa Barbara, CA)

Laying out the near-term goals of the NIH DP3-grant funded Ambulatory Control project, Dr. Frank Doyle III described the control algorithm developed for the project’s first outpatient studies. The algorithm, called periodic zone model predictive control (PZMPC), is similar to UCSB’s initial zone MPC algorithm in that insulin delivery reverts back to a fixed basal rate unless CGM values start to trend outside of a specified glycemic zone. However, with PZMPC the boundaries on this zone gradually shift at night, from the daytime zone of 80-140 mg/dl to an overnight zone of 110-220 mg/dl. To further mitigate the risk of nocturnal hypoglycemia, overnight insulin delivery is constrained to be no more than 150% of the basal rate. In silico modeling suggests that such an algorithm should enable excellent overnight safety, and on November 2 PZMPC was successfully tested for the first time in a human patient. Dr. Doyle said that next steps include additional feasibility evaluations at UCSB/Sansum in the coming weeks and months, larger in-clinic studies in the spring and summer of 2013, supplemental in silico testing in late 2013, and eight-week outpatient experiments starting in mid-2014.

  • Starting with the basic on-off controller described by Kadish in 1963, Dr. Doyle reviewed the history of algorithm development in closed-loop glucose control research. Later the precursors to today’s proportional-integrative-derivative controllers were developed by Albisser et al. (1974), Clemens (1979), and Fischer et al. (1980). In 1996 Dr. Doyle worked on the first application of model predictive control (MPC) for glucose control, and in 2001 Dr. Roman Hovorka’s group introduced non-linear MPC to artificial pancreas research. Promising work has been done with other algorithmic approaches such as pole-placement, H- infinity, adaptive, and fuzzy logic, Dr. Doyle noted. To supplement this overview he presented a slide with all but the most recent published clinical trials of artificial pancreas studies: these included four studies of PID algorithms, one with a PD/PI algorithm, one with a PD-based controller that also dosed glucagon, one hybrid MPC-/PD-driven insulin/glucagon system, 11 using MPC-based algorithms, and one with zone MPC.

  • One of the latest developments in MPC-based glycemic management has been UCSB’s zone MPC algorithm (Grosman et al., J Diabetes Sci Tech 2012), whereby insulin delivery changes from a pre-set basal rate only if the patient’s CGM values leave the target zone or are predicted to leave the target zone (80-140 mg/dl, in initial applications). Dr. Doyle noted that zone MPC has been demonstrated feasible in both a 12-patient UCSB/Sansum study of fully closed-loop control (Zisser et al., ADA 2012) and in larger industry trials of an artificial pancreas precursor product (Mackowiak et al., ADA 2012).

  • Dr. Doyle described the rationale for and design of a periodic zone model predictive control (ZMPC) closed-loop algorithm. Wide agreement exists that closed-loop control should be relaxed overnight in order to minimize risk of hypoglycemia, he explained; the question is just what approach to use. One approach would be to turn off the controller altogether or “de- tune” it so dramatically the controller never takes action (e.g., by widening the target zone to go all the way up to 1,000 mg/dl). As an alternative, UCSB researchers have designed the PZMPC algorithm that smoothly adjusts the boundaries of the zone from their daytime values (80-140 mg/dl) to establish a wider zone overnight (110-220 mg/dl). The PZMPC algorithm also puts a hard constraint on how much insulin can be delivered, even in hyperglycemia: no more than 50% of basal rate. The UVa/Padova simulator was used to compare PZMPC, traditional zone MPC, and a regime that switches to a fixed basal rate at night; PZMPC led to more overnight hyperglycemia but less overnight hypoglycemia (Gondhalekar, Dassau, Doyle III Eur Control Conf 2013).

  • The initial clinical evaluation of PZMPC closed-loop control will enroll 5-12 of the 12 patients who participated in the first study of UCSB’s original zone MPC algorithm (Zisser et al., ADA 2012). Except for the difference in control algorithms, the experimental designis identical to that of the zone MPC study (day-and-night study with unannounced meals, unannounced exercise, skipped lunch). The first patient tested PZMPC on November 2, 2012 and experienced favorable glycemic control (including less overnight insulin delivery – possibly safer if a patient were not frequently testing, Dr. Doyle noted).

  • Dr. Doyle closed with a look to the future of the NIH DP3-grant-funded Ambulatory Control project, a collaboration of artificial pancreas researchers at UCSB, the Sansum Diabetes Research Institute, the University of Virginia, and the Mayo Clinic. He reminded the audience that the five-year, $4.5-million initiative is designed to develop closed-loop systems that respond to glucose on a scale of minutes, adapt to day-to-day and week-to-week glycemic changes, and “monitor and supervise” glucose control over months and months.

    • Communication between the hardware and algorithms will occur through the UCSB/Sansum Artificial Pancreas System (APS) platform, which Dr. Doyle said is being ported to a new iDevice framework so that it can run on mobile phones (as opposed to laptops or tablets as previously used). In addition to the PZMPC controller, the system will incorporate the UCSB/Sansum Health Monitoring System (HMS) safety algorithm, which can send text messages and graphical alerts to physicians for remote monitoring. Dr. Doyle also noted that the FDA has approved the inclusion of fingersticks with Bayer’s Contour Next BG meter in an upcoming UCSB/Sansum closed-loop study – hopefully the first step toward outpatient studies that fingerstick tests as the sole reference values.

    • Dr. Doyle looked forward to upcoming clinical studies in the DP3 project, which will begin inpatient closed-loop studies at the University of Virginia, Sansum Diabetes Research Institute, and Mayo Clinic during the spring and summer of 2013. Each study will include two closed-loop sessions: one for behavioral initialization of the individual patients, and another with “behavioral adaptation.” Additional in silico are slated for late 2013 as a prelude to the main event: eight-week, case-controlled outpatient comparisons of closed-loop and open-loop control.


Boris Kovatchev, PhD (University of Virginia Health System, Charlottesville, VA)

In an engaging, data-driven presentation, Dr. Boris Kovatchev presented the first-patient results from a closed-loop efficacy trial using a new modular algorithm and the latest device technologies. The algorithm used an enhanced control-to-range system comprised of: 1) an insulin on board (IOB) tracking module; 2) a range control module; and 3) a safety supervision module. The algorithm ran on a DiAs smart phone (a portable AP system developed at the University of Virginia by Dr. Patrick Keith- Hynes that has a modified medical-grade Android OS platform designed for AP application), which connected to a Dexcom G4 Platinum receiver via USB and wirelessly to the Tandem t:slim pump via low energy Bluetooth. (The DiAs also could be monitored remotely over 3G or Wi-Fi connection.) The study, which just commenced a week ago, is being conducted at four sites, with five patients at each site. In a crossover design, patients were randomized to either open- or closed-loop control (run on the DiAs) for a 40-hour session. Dr. Kovatchev presented the very first patient experience. Over 40 hours with closed- loop control, the participant was in range (70-180 mg/dl) 83.4% of the time, never below 60 mg/dl, in range overnight (80-140 mg/dl from 11:00 pm to 7:00 am) a notable 100% of the time, and >180 mg/dl 14.4% of the time. While Dr. Kovatchev said he knew he shouldn’t present data from just one person, we were sure glad he did! After seeing the promising first tracing, we can’t help but look forward to when full results emerge.

  • Dr. Kovatchev began by detailing the specifics of the Diabetes Assistant (DiAs) portable AP platform. Notably, the system has a medical grade Android operating system designed for AP applications. (The operating system and graphical user interface are deposited in the FDA master file MAF 2109, “AP Mobile Medical Platform.”) DiAs was developed at the University of Virginia by Dr. Patrick Keith-Hynes. The system can wirelessly communicate with an insulin pump and CGM and can operate multiple control algorithms. It’s color touch screen features a home screen with hypoglycemia and hyperglycemia “traffic lights” to inform the patient whether intervention is needed, and various system statuses (e.g., battery time, whether there is connection to the pump or sensor). DiAs can be used for closed-loop or open-loop control, and can enable remote monitoring (even simultaneous real-time remote monitoring of several patients, as Dr. Howard Zisser [Sansum Diabetes Research Institute, Santa Barbara, CA] demonstrated – he controlled several patients from a single iPad). For a deeper delve into the user interface, please see page three of our DTM 2011 Day #2-3 report at http://www.closeconcerns.com/knowledgebase/r/5f40a09f.

  • Early closed-loop feasibility studies with DiAs demonstrated the ability to maintain inter-device communication. The system consisted of DiAs, a communication box (Google Galaxy Nexus phone), and iDex (an Insulet OmniPod PDM integrated with the Dexcom Seven Plus CGM). Across four centers (UVA, Padova, Montpellier, and Sansum Diabetes Research Institute), patients received both open- and closed-loop control using DiAs. Inter-device communication was maintained 98.9% of the time in open-loop control (out of 277 patient hours) and 97.1% of the time in closed-loop control (out of 550 patient hours).

    • After 13 hours of open-loop control, participants had closed-loop control for 29 hours. The closed-loop control was two-fold: during the day the system implemented control-to-range and overnight the system was in safety mode (i.e., more relaxed control to reduce hypoglycemia risk).

  • Remote monitoring using DiAs connected to Dexcom’s G4 sensor reduced nocturnal hypoglycemia in a trial in young children at three summer camps sessions (n=20/camp [n=10 G4 + DiAs; n=10 G4 only]). Total study time was 1360 hours, of which remote monitoring was operational for 1314 hours (97%). For a deeper delve into the study, please see our coverage of Dr. Bruce Buckingham’s (Stanford University, Stanford, CA) dedicated presentation on the trial on page 12 of our EASD Day #2 Highlights report at http://www.closeconcerns.com/knowledgebase/r/ee283b0b.

  • Just last week, a multi-center efficacy trial of closed-loop control using a control-to- range algorithm and the newest generation devices commenced. Participating centers include UVA Center for Diabetes Technology, Sansum Diabetes Research Institute (UC Santa Barbara), Padova (Italy), and Montpellier (France). This randomized crossover study consists of one 40-hour session each of open- and closed-loop control (DiAs runs both). Five patients are enrolled per site; patients are responsible for system communications.

    • The modular control-to-range algorithm is comprised of three modules: 1) an IOB tracking module (UCSB); 2) a range control module (Pavia); and 3) a safety supervision module (UVA). Importantly, the algorithm allows for enhanced control-to- range during the day for intensive treatment, but relaxes control overnight.

  • The closed-loop system consists of DiAs smart phone, which connects by USB to the Dexcom G4 receiver and by low power Bluetooth to the Tandem t:slim. The G4 Receiver, of course, wirelessly communicates to the G4 sensor. Dr. Kovatchev said that to the best of his knowledge, this was the first time a closed-loop used the G4 and t:slim.

  • Results from the first patient tracing were encouraging, with 83.4% of time in target range (70-180 mg/dl) and 100% of time in target range (80-140 mg/dl) overnight. Dr. Kovatchev drew attention to the accuracy of the G4 – the 12 fingersticks shown seemed to fall closely in line with the G4 tracer. Further, Dr. Kovatchev highlighted the “traffic light” system of DiAs, with a color charting beneath the tracer showing hypoglycemia lights. Dr. Kovatchev noted two examples: 1) when the blood sugar was rapidly declining the safety system picked up the event at 140 mg/dl, the yellow hypoglycemic light came on, and insulin delivery was cut; 2) when the blood sugar reached ~90 mg/dl, the red late came on indicating that carbohydrates were needed and hypoglycemia was avoided.

Closed Loop Control

Time in range of 70-180 mg/dl


Time above 180 mg/dl


Time in range of 80-140 mg/dl overnight (11:00 pm to 7:00 am)


Number of hypoglycemic episodes below 60 mg/dl



Claudio Cobelli, PhD (University of Padova, Padova, Italy)

Dr. Claudio Cobelli described three ongoing projects to improve closed-loop control: the latest AP@home trials, an improved Dexcom sensor, and an updated simulator. The AP@home consortium is in midway through a set of overnight, partially outpatient experiments (n=12) that use UVa’s latest DiAs controller, a specialized remote monitoring system, an algorithm that incorporates recent insulin delivery information in its control decisions, and a simplified protocol for open-loop insulin dosage at meals. Meanwhile Padova engineers are working with Dexcom to develop a “smart” CGM transmitter with new onboard processing algorithms to detect noise and enhance calibration. Dr. Cobelli also described several recently submitted modifications to the UVa/Padova metabolic simulator, which include a more-physiological nonlinear response to hypoglycemia, a model of glucagon counterregulation, and revised definitions of insulin-to-carbohydrate ratio and correction factor. Ongoing research on the simulator will introduce a new model of sensor error, improve the model of subcutaneous rapid-acting insulin, and attempt to “clone” the results of the initial AP@home studies.

  • Dr. Cobelli explained that several improvements have been introduced in the latest AP@home clinical trial, which is targeted to complete by the end of the year. The crossover-design study uses outpatient closed-loop during the day and inpatient closed-loop control at night; the design also includes exercise and video games. The first four patients have completed the study at Padova, and the trial is planned to conclude with four patients in Montpellier and four in Amsterdam. The DiAs system used in the new trial has been improved over that in the first outpatient European studies, and the MPC “observer” module has beenmodified to monitor the pump for information on insulin delivery (enabling more accurate glycemic predictions). Also, integration of open-loop meal control in the closed-loop scheme has been simplified. Pre-clinical simulations were run on a recently modified simulator (see below), and improvements have been made to the “worst-case analysis” CVGA grid used to tune the controller’s aggressiveness. (Basically, the previous grid scored a particular simulated patient’s performance based only on whichever was worse in a given experiment, the highest hyperglycemic excursion or the lowest hypoglycemic excursion. By contrast the new curvilinear grid would rate an algorithm’s performance differently if a patient’s respective maximum and minimum values were 300 mg/dl and 110 mg/dl, instead of 300 mg/dl and 70 mg/dl, for example.)

    • Dr. Cobelli presented data from one of the patients in the study (“as you can imagine,” he smiled, “I chose the best.”) This patient’s time in target range (70-180 mg/dl) was improved dramatically with closed-loop control (99.9%) compared to open- loop control (72.7%); the mean time in target for closed-loop control in all four patients was in the mid-80% range.

  • In collaboration with Dexcom, Padova’s bioengineering team is exploring improved CGM algorithms for better noise detection (Facchinetti et al., IEEE Trans Biomed Eng 2011) and enhanced calibration (Guerra et al., IEEE Trans Biomed Eng 2012). The published work on these algorithms has involved post-processing sensor data that had already been converted to a glucose signal. However, by building algorithms directly into a future Dexcom “smart” transmitter, Dr. Cobelli hopes to further improve sensor performance and simplify wireless communication in closed-loop systems.

  • Modifications to the Virginia/Padova metabolic simulator were submitted to the FDA on October 17, 2012, and subsequent improvements are already underway. The changes currently under FDA review include a non-linear response to hypoglycemia, a counterregulation model that includes glucagon secretion, kinetics, and action), a new way to define insulin-carbohydrate ratio and correction factor (to mimic the way that real patients would determine these values), and an altered model of absorption parameters. In a CE-EGA of how well actual patient data (n=96) agreed with the old and new simulations, the new simulation performed significantly better in hypoglycemia (and agreed more closely with real data on interquartile range and high and low blood glucose indices, as well).

  • Ongoing work on the simulator includes a new model of CGM error, which is based on data that the Oregon researchers shared from a recent clinical trial (Castle et al., Diabetes Care 2012). This work includes individualized models of blood to interstitial glucose kinetics as well as models of the calibration function, sensor variability, and measurement noise. These components can be analyzed individually to see how errors in each might affect the sensor result. The researchers can also assign various probabilities to the likelihood transient artifacts, error codes due to noise, and disconnection of the sensor from the body, to see what these problems would mean for glycemic control.

  • Padova engineers are also changing their module of subcutaneous insulin kinetics to incorporate data from a clamp study of insulin lispro in 41 patients with type 1 diabetes. Dr. Cobelli thanked Biodel’s Alan Krasner for donating these data, and he expressed hopes that the updated module would be completed by the end of 2012.

  • To conclude, Dr. Cobelli explained that the simulation is being adjusted so that it can “clone” the clinical data from the AP@home CAT Trial. (As a reminder, this dataset includes a total of 141 traces from eight patients.) For example, these modifications would allow the simulation to incorporate intraday variability in glucose absorption and insulin sensitivity, as occurs physiologically.


Jeffrey Joseph, DO (Thomas Jefferson University, Philadelphia, PA), Eyal Dassau, PhD (University of California, Santa Barbara, Santa Barbara, CA), Frank Doyle III, PhD (University of California, Santa Barbara, Santa Barbara, CA), Edward Damiano, PhD (Boston University, Boston, MA), Boris Kovatchev, PhD (University of Virginia Health System, Charlottesville, VA), Claudio Cobelli, PhD (University of Padova, Padova, Italy)

Dr. Joseph: It seems feasibility has been demonstrated out of the clinic and in the home. What has to happen to make this a tool that patients can use in the real world setting? Is it two years off? Five years off? Is it just refinement that is needed or a major breakthrough?

Dr. Kovatchev: There will be advancements needed to bring this to seamless outpatient use, and there will be stages. We have a clear line of sight. The first step is to have reliable, robust wireless connection between devices in the closed-loop system and I believe that is coming sometime next year. Then, there will be some miniaturization of the devices, but the hardware will be to a large extent set. And we are beginning with relaxed safety modules taking care of hypoglycemia over night that will progress to intensive therapy with enhanced control. As soon as we get wireless connections, we can do larger studies.

Dr. Damiano: We’ve staged out our approach to include experiments at three levels. First, are feasibility studies in carefully monitored inpatients, which we have been conducting over the past four months and which will conclude for us next month. Then, we have transitional studies where you loosen the reigns a little bit. You might have nursing attention of one-on-four or one-on-eight. Experiments are longer in duration – five days, then two weeks. Beginning next month, we expect to do these over the course of the next 18 months in our group. These will include the Beacon Hill study, camp studies in children, and studies with adults wearing the device for a couple weeks and bringing it home and taking it to work. We’ll have close reigns on the participants, but not to the extent of feasibility trials. The last major step would be to move to a pivotal study. That will require a custom medical grade hardware/software platform that is dedicated to the task of controlling glucose. We want to build this over the next 18 months. The pivotal trial would have hundreds of subjects, be a parallel study and take about six months. I envision something fully closed-loop perhaps available in the next four to five years. In our case we are relying on the importance of glucagon, but we still need stable pumpable glucagon. We need reformulation of glucagon, not a glucagon analog, so that there can be a faster regulatory path, and then we need a dual-chamber pump that can deliver both insulin and glucagon.

Dr. J. Hans DeVries (Academic Medical Center, Amsterdam, The Netherlands): When you presented your third experiment results, you divided them between day one and day two. I like that. But perhaps it needs some justification from a statistical point of view. Why is better in the second 24 hours?

Dr. Damiano: When we talk about day one and day two, we’re talking about the system adapting in real time over the course of the experiment. To begin, the system does not know anything about the subject other than the subject’s weight; it is not informed about carb-to-insulin ratios, correction factors, basal rates, total daily dose, etc. The system begins dosing insulin conservatively. It then adapts to the higher insulin needs in adolescents, if that is required, or it might stay close to where it started, which is more typical of the insulin needs of adults. While it’s always adapting basal infusion rates every few minutes, the degree it may dial up may be greater in one subject than another. The point is we have a time scale of adaptation that looks like six to 12 hours to converge upon the subject’s insulin requirement. On the first day, control is not as good as what we expect on day two. We don’t know how much better it will get on days three, four, and five. We do see that control on day two is better and anticipate that it might continue to improve somewhat more over time.

Q: Are you planning to compare dual-hormone to single-hormone control in clinical trials? Maybe starting from a simulation mode?

Dr. Damiano: The studies we have done have been limited to dual hormone control. In order to assess the true potential of glucagon, we have to design a study that uses our own bi-hormonal and insulin-only control systems or work collaboratively with another group that follows the same clinical protocol and does insulin-only control. It’s hard to make comparisons between closed-loop systems when protocols are not handled in the same way.

Dr. Gary Steil (Children’s Hospital Boston, MA): Dr. Damiano, I think you told me in your last diabetes paper that the variability between subjects on any two days was as large as between any two subjects. In other words, each patient is a new patient on a different day. When you find these new parameters over the first six hours, what is the stability of that new identification? And for Dr. Cobelli, in the past, it’s been very difficult to identify new virtual patients. If the patient is changing very rapidly, how identifiable is the UVa/Padova model for all of the parameters? Are the patients really that variable? Is the same patient tomorrow different from the one today?

Dr. Damiano: When we described the variability from day one to day two in our previous study, we were referring to the variability in insulin dosing by the controller, not in glycemia. We analyzed this variability in insulin dosing between day one and day two in the same subject and compared that to the variability in insulin dosing between day one in that subject and day two in each of the other subjects. That paper was restricted to our adult cohort. What we observed from this analysis in our adult cohort was that intra- subject variability was not less than inter-subject variability. In the adult study, both are comparable and not very large – there is relatively low variability in both. There was not nearly as much inter-subject variability in that adult cohort as there is in the cohort now that includes both adolescents and adults. The variability between adolescents and adults is profound. For adolescents, we’re seeing insulin requirements that are two-to-three fold higher than for adults of the same weight.

Nevertheless, if you take the same adult subjects and run them in a study for three months, they might develop an inter-current illness along the way. They may require significantly more insulin during the illness and can start to look like an adolescent, in terms of their insulin requirement. The system must be able to automatically adapt up to meet those insulin-dosing demands. Alternatively, they may also have a vomiting illness, where the system must be able to automatically dial down the insulin. That is what I think our system must be able to demonstrate, and I think our data suggest that we can achieve this. Variability might not be that significant from day to day, but it very well can be from one week to another.

Dr. Cobelli: Thanks to the Bayesian identification strategy it is possible to identify the UVa/Padova model on the 24-hour glucose traces of the AP@home project. Important changes on glucose absorption parameters and insulin sensitivity were observed in each individual. The ability to well describe the traces starting from the a priori information of the UVa/Padova simulator provides an indirect validation of the simulator. We are midway in the analysis of the 141 traces.

Q: As to comparing insulin alone to insulin plus glucagon, we’ve published that study. It was a crossover design so people started on insulin plus placebo then switched to insulin plus glucagon. There was less hypoglycemia with the glucagon. So that study has been done, but as Ed pointed out, there is no pumpable glucagon.You each mentioned exercise. Have any of you considered measuring something like heart rate and how it would input into the system? Maybe it could turn on glucagon earlier based on heart rate?

Dr. Cobelli: Heart rate is an important signal but may not be the signal. Sometimes it’s good in detecting the immediate change. But as exercise progresses, I don’t think heart rate catches up with physical activity. There is a variety of physical activity – low, medium, high. It is not just the exercise you are doing on a bike. We are doing experiments with the Mayo group. They have a sophisticated system to capture physical activity, and we have presented the first results but are still working to better understand the relationship between physical activity and the glucose signal (CGM and plasma). I’m not sure if heart rate is the signal. It is important, but not the totality.

Dr. Damiano: In particular, while you did the dual-hormone versus insulin-only comparison study, it has to be taken in the context of how that study was done. That study was conducted in sedentary subjects and will likely understate the importance of glucagon. What I think we’re going to see is greater importance of glucagon with physical activity. On the question of using input signals to the controller other than glucose, I think heart rate is very ambiguous and I don’t think that is something that should drive control. Various levels of exertion impact the way glucose is cleared. Extreme exercise can actually require enhanced insulin secretion while moderate exercise results in tremendous glucose clearance. I think the verdict is born out in glucose.

Dr. Steil: Can we identify all the model parameters without a glucose tracer?

Dr. Cobelli: Yes, we have been able to reproduce the traces of the AP@home patients by using the simulator. Obviously, we are using the a priori information of the simulator. The distribution of the parameters of the simulator probably resembles reality since we were able to describe a variety of traces in a variety of situations.

I certainly agree with all the comments that Boris made in terms of the timeline. Keep in mind, that there are ongoing studies in parallel in terms of understanding the physiology – the intra-day and inter-day variability of insulin sensitivity. And what is the model for stress? What is the model for exercise? These studies are being done in parallel with the technological advancements.

Q: Do any algorithms have the ability to discern hyperglycemia from eating versus hyperglycemia by pump failure? It warrants very different responses from algorithms and patients.

Dr. Doyle III: To distinguish between this, you need a fault diagnosis module – something to detect for abnormal events. It is impossible to do blindly from glucose alone. By marrying this with fault diagnosis tool you could manage it, but not with a straight MPC algorithm.

Dr. Cobelli: It is a different module; it has nothing to do with MPC. A paper is in press of our group in IEEE Transaction of BME describing this module.

Dr. Yogish Kudva (Mayo Clinic, Rochester, MN): Earlier this year in August we had a paper in Diabetes Care in low grade physical activity using accelerometers and it has a different effect as far as glucose lowering is concerned after a meal. We also compared heart rate and accelerometers in the same patients. We have a poster, number 39. And we have data being analyzed now on V02 max.

Dr. Howard Wolpert (Joslin Diabetes Center, Boston, MA): Dr. Cobelli, you said that you were integrating the glucagon counterregulatory response in the hypoglycemia module. But in type 1 diabetes, patients lose the glucagon response to hypoglycemia.

Dr. Cobelli: The glucagon model is based on the tracer data that has been done. It is what is the state of the art in terms of glucagon secretion, kinetics, and action. Incorporation of glucagon into the model allows the simulator to have glucagon in it.

Dr. Wolpert: So it’s only relevant to dual hormone closed loop systems?

Dr. Cobelli: Yes, exactly.

Dr. Hovorka: On the issue of prandial insulin dosing – there seems to be a question about whether to give priming boluses or have fully reactive systems. It’s quite an important decision about which way to go. Without prandial dosing, I’m not sure it’s possible to get as good of control as with open-loop therapy.

Dr. Damiano: We have not included prandial dosing to the exclusion of testing our fully reactive system. Our first study tested a fully reactive system using venous blood glucose as the input to the control system. In our second study, we added a weight-based prandial dose at the presentation of the meal because of the 10- to 15-minute delay in the CGM signal. In the study we’re doing now, which will finish in a couple weeks, there are two arms. One arm includes adults and adolescents receiving an adaptive prandial bolus at the beginning of each meal. The other parallel arm includes adults and adolescents receiving entirely reactive control with no prandial insulin boluses. I entirely disagree that you cannot beat open loop control using a fully reactive system. In adults, we were able to achieve average blood glucose levels of 143 mg/dl with entirely reactive control, versus 138 mg/dl with a prandial insulin bolus. So we didn’t see much difference with or without a prandial insulin bolus in adults; in either case, this would correspond to an A1c of about 6.5%. In adolescents, we achieved average glucose levels of 171 mg/dl with hypoglycemia less than 1% of time. This is in a sedentary condition. You see that even moderate exercise causes greater glucose clearance, so we should see these numbers drop further in the outpatient setting. That corresponds to an A1c of 7.6%. That’s way better than the standard of care. It turns out it’s not a hard bar to beat open-loop therapy in adolescents.

Dr. Doyle: We need to inform the decision of the best strategy. We’re pushing hard. Data from ADA showed that we can get to within 70% time in range with a purely reactive algorithm. I think that competes favorably with prandial dosing.

Dr. Kovatchev: I have a slightly different take on prandial dosing. It can also be beneficial from a design and regulatory point of view. If you have prandial dosing and the basal rate controlled by the person, then a range controller on top of this basal-bolus therapy, that controller has an adjunct claim – that’s opposed to a replacement claim in regulatory terms. That’s important for closing the loop and taking this out there.

Comment: Overnight glucose control is the time to get better control. It’s prime time for the closed loop to improve control over the open loop. I am concerned that by relaxing control, you are losing the greatest benefit.

Dr. Doyle III: I don’t dispute that at all. We relax measurements over night as we begin to worry more about the robustness of the system. My personal comment is that the greatest challenge is getting away from clean in-clinic environments and having truly free living. That will be a real tough test. In anticipating this, we need to minimize the risk of hypoglycemia and we offer this tool to relax control. I completely agree with you, but this is an alternative to turning it off.

Q: On zone MPC, you mentioned that between the range of 80-140 mg/dl, the algorithm would keep a constant basal. Does that mean when the subject eats a meal, the controller won’t give anything but basal until the glucose goes over 140 mg/dl? What’s the rationale for using a zone rather than single set point?

Dr. Doyle: It starts delivering when the prediction of glucose leaves the zone. If it’s showing a trajectory that will leave the zone, dosing will commence. In our years of working with doctors at Sansum, we’ve learned that there is no single magic value. Glucose is always in a range.

Artificial Pancreas: Clinical Aspects


Bruce Buckingham, MD (Stanford University, Stanford, CA)

The esteemed Dr. Bruce Buckingham presented pilot-study results assessing a predictive low glucose suspend (PLGS) system, one that Dr. Buckingham characterized as “home-grown.” The system was designed with the intention of eliminating prolonged episodes of nocturnal hypoglycemia, while working “in the background” with a reduced number of alarms. The system only alarms if blood glucose is less than 60 mg/dl or greater than 250 mg/dl and does not alarm for system failure (instead, it reverts back to usual basal). “Glucose is important for you brain,” he said, “but so is sleep.” These home trials included 19 patients (375 nights total), and tested three iterations of the algorithm. Nights were randomized such that 2/3 of nights were controlled by active PLGS. After finding that PLGS reduced nocturnal hypoglycemia by ~50% and nights with prolonged lows by ~70%, Dr. Buckingham said that they are moving forward with the third version of the algorithm (which conferred the best overall nighttime glucose management) and are enrolling 44 subjects in a multicenter study with a goal of obtaining 1,600 nights of system use.

  • Nineteen subjects completed 375 nights, with 2/3 of nights randomized to predictive low glucose suspend (PLGS); the system is run from a bedside computer. Studies were conducted in homes, with no nurse present or remote monitoring. Participants ranged in age from 18-56 years old and in A1c from 6.0% to 7.7%.
  • The algorithm went through three stages of fine-tuning, such that three versions of the algorithm were tested. The first had a prediction horizon of 70 minutes with basal insulin being resumed when glucose was predicted to rise above 100 mg/dl. The second used a prediction horizon of 50 minutes with basal insulin resumed with any positive rate of change; there was no insulin suspension if glucose was above 230 mg/dl. The third iteration was the same as the second, but used a prediction horizon of 30 minutes.
  • After the third fine-tuning of the algorithm, the system had the lowest mean glucose at first shutoff and the lowest median peak glucose following fist shutoff. As Dr. Buckingham explained, one of the concerns with PLGS is that insulin suspension leads to rebound hyperglycemia, which did occur in the study he presented. However, the third iteration of the algorithm seemed to mitigate this risk somewhat.



Algorithm 1 (n=5)

Algorithm 2 (n=12)

Algorithm 3 (n=77)

# Intervention nights




# Nights with Pump Suspension

52 (78%)

84 (78%)

41 (53%)

# Pump Suspensions Per Night

















Mean Glucose at First Shutoff (mg/dl)







Median Peak Glucose Following First Shutoff (mg/dl)







  • Dr. Buckingham presented safety data showing that the number of nights with nocturnal hypoglycemia, and the duration of that hypoglycemia, was reduced with PLGS compared to control nights. However, the reduction in hypoglycemia seemed to some extent to come at the cost of higher overnight glucose. Dr. Buckingham polled the audience (see below) to see how much of a rise in overnight glucose people would be willing to tolerate in order to reduce hypoglycemia. He would be willing to accept about ~12 mg/dl. Of course, as this was a pilot study, we await the larger, better-powered study Dr. Buckingham is currently enrolling to determine whether these findings will be replicated, statistically significant, and clinically significant.

Algorithm 1

Algorithm 2

Algorithm 3


Control (n=5)

Active (n=5)

Control (n=10)

Active (n=12)

Control (n=8)

Active (n=9)

Number of nights







Mean Bedtime Glucose (mg/dl)







Mean Overnight Sensor Glucose (mg/dl)













Nights with CGM values 71-180 mg/dl













Nights with CGM values > 180 mg/dl













Nights with CGM values >250 (mg/dl)













Nights with CGM values < 60 mg/dl













> 60 min duration







> 120 min duration







  • Dr. Buckingham asked the audience: to avoid severe nocturnal hypoglycemia, what is the highest increase in mean overnight glucose values that would be acceptable to you? Surprised by the majority response, Dr. Buckingham noted that a 45 mg/dl increase for 24 hours would raise your A1c 1%. We were surprised expectations were this high.

    • 5 mg/dl: 8%

    • 12 mg/dl: 18%

    • 20 mg/dl: 33%

    • 25 mg/dl: 41%

  • He closed his presentation with a second audience response question: for nocturnal closed-loop control, what is the maximum number of times you would be willing to wake up to do a calibration on a nightly basis?

    • 0: 82% (“Me too,” said Dr. Buckingham.)

    • 1: 17%

    • 2: 1%

    • 3: 1%


J. Hans DeVries, MD (Academic Medical Center, Amsterdam, The Netherlands)

Dr. J. Hans DeVries presented an overview of the AP@home Project’s CAT trial, including new secondary analyses as well as both intent-to-treat and per-protocol results (previously presented at ADA 2012 and EASD 2012, respectively). As a reminder, CAT was a 23-hour, crossover comparison of open-loop control to two different MPC algorithms (one from Cambridge and another from the iAP consortium, both “de-tuned” to prioritize hypoglycemia reduction). Compared to open-loop control, both algorithms significantly reduced time spent in hypoglycemia but significantly increased time spent in hyperglycemia. The Cambridge algorithm delivered significantly less insulin than that from iAP, and both algorithms used less insulin than open-loop control. Results were equivalent regardless of whether CGM was calibrated with YSI or glucose meter measurements. Control was generally similar whether centers used manual input of the algorithm’s commands (control decisions made every 15 minutes) or automated control with the Artificial Pancreas System (decisions every five minutes), though time in hyperglycemia was significantly lower with manual control.

  • In addition to reviewing the main findings of the CAT Trial, Dr. DeVries unveiled several secondary analyses. For the primary intent-to-treat and per-protocol analyses, see page 64 of our ADA 2012 coverage at http://www.closeconcerns.com/knowledgebase/r/c6afb200 and page 6 of EASD 2012 coverage at http://www.closeconcerns.com/knowledgebase/r/9f88794c.

    • Compared to YSI reference values, the Dexcom Seven Plus sensors used in the CAT Trial had a mean absolute relative difference (MARD) of 15.1% in the intent-to-treat analysis. The MARD fell only slightly to 14.1% in the per-protocol analysis, though the most common reason for data exclusion from the per-protocol analysis was CGM failure (defined as a MARD above 50% for 45 minutes or more). Such instances of CGM failure accounted for 0.4% of sensor values in the open-loop setting but 13% and 17% of time under control by the iAP and Cambridge algorithms, respectively. He alluded to an ADA 2012 oral presentation from the Cambridge group in which the Dexcom Seven Plus was found to have more long lag times than Abbott’s FreeStyleNavigator, but he added that these problems seem to have been largely addressed in the Dexcom G4 Platinum.

    • Overall control did not differ when CGM was calibrated with self-monitoring of blood glucose (SMBG) as opposed to YSI. Dr. DeVries indicated that this bodes well for the transition to outpatient studies. The next challenge will be to replace YSI reference values with a glucose-sensing technology that can be used by patients at home.

    • Dr. DeVries also compared centers using “automated” vs. “manual” closed- loop control. In three of the six CAT centers, sensors and pumps were linked to the control algorithms via a laptop running the Artificial Pancreas System (APS), which enabled control every five minutes. The other three centers were located in countries where regulatory clearance of APS was thought unlikely. In these centers, communication from the sensor to the control algorithm to the pump was carried out by clinical trial staff every 15 minutes. The results in these sub-studies were similar to the main results, except that a lower percentage of time was spent in hyperglycemia in the manual vs. automated centers (31.6% vs. 44.4%). We speculate that automated closed-loop control may have placed a strain on the devices involved, limiting the systems’ functionality, whereas the15-minute vs. five-minute difference was relatively inconsequential given that thealgorithms had been designed not to be aggressive. Dr. DeVries noted that in a sub-sub- analysis of manual-control centers, the Cambridge algorithm outperformed open loop for time in range (64.7% vs. 51.7%); he did not hypothesize as to why this result was seen.


Eric Renard, MD, PhD (Hospital Lapeyronie, Montpellier, France),

Dr. Eric Renard gave a unique presentation focused on two major questions: 1) how much can we trust the artificial pancreas today? and 2) who are the ideal candidates for an artificial pancreas? On the first, Dr. Renard outlined what has been proven thus far with the AP: feasibility (Steil et al., Diabetes 2006); reduction of nocturnal hypoglycemia (Hovorka et al., Lancet 2010); improved mean blood glucose (Breton et al., Diabetes 2012); and the possibility of outpatient use (Cobelli et al., Diabetes Care 2012). He also outlined the biggest current weak points of the AP: CGM sensors (the one he most emphasized), delays in insulin absorption, and mealtime control. Dr. Renard outlined two groups of patients that will be ideal candidates for the AP: those with recurrent hypoglycemia and those that are afraid of hypoglycemia (“hypophobic”) and purposely keep their blood glucose high. Although patients with high glycemic variability might be good candidates for the AP, Dr. Renard believes the subcutaneous delivery route may be challenging –intraperitoneal may be a better route in these cases. He concluded with a review of how to prepare candidates for use of the AP.

  • Dr. Renard believes the major benefit of the artificial pancreas is reduction in hypoglycemia. He explained that this has been demonstrated in nearly all closed-loop studies Dr. Renard specifically referenced the results of the CAT trial, which found equivalent overall mean glucose control between open loop and closed-loop control, but a significant reduction in hypoglycemia. Encouragingly, the CAT trial also took place at several centers without any closed- loop experience, demonstrating the potential feasibility of rolling the AP out to a broader population.

  • In Dr. Renard’s opinion, CGM remains one of the weakest points in the artificial pancreas. He explained that CGM was responsible for the highest number of problems in theCAT trial. Pump errors occurred less than 1% of the time, while software problems occurred 1.6% of the time. By contrast, 7% of time was spent without a sensor signal and almost 5% of time was spent without a sensor reading due to sensor failure. We note that the Dexcom Seven Plus was used in this trial. We expect that once Dexcom’s G4 Platinum is used in AP studies, these rates will drop precipitously (as a reminder, one of the benefits of the G4 Platinum is improved signal transmission).

  • The AP is most ideal for those with recurrent hypoglycemia (i.e., overzealous insulin delivery) and those who are afraid of hypoglycemia (“hypophobic”). Dr. Renard showed CGM trace examples of both patients, explaining that automated insulin delivery stands to benefit both problem areas. However, he feels that different experiments are needed for these two populations and they should not be mixed. Dr. Renard did agree that patients with high glycemic variability might be good candidates for the AP, though he thinks “the subcutaneous route is challenging.” The big question is whether the subcutaneous route – fraught with its own challenges – will be able to manage variability. In his view, intraperitoneal delivery is probably the best route for this population.

  • How to prepare candidates for AP use? 1) Move the patient to CSII; 2) experiment with sensor use; 3) train the patient to carbohydrate count (patients will still need to input carbs at mealtime); 4) train patients on the AP platform; 5) start closed loop at the hospital (or a similar controlled environment) and check remote monitoring before leaving for home use.


Roman Hovorka, PhD, MSc, BSc (University of Cambridge, Cambridge, United Kingdom), Arleen Pinkos, BS (U.S. Food and Drug Administration, Silver Spring, MD), Bruce Buckingham, MD (Stanford University, Stanford, CA), J. Hans DeVries, MD (Academic Medical Center, Amsterdam, The Netherlands), Eric Renard, MD, PhD (Hospital Lapeyronie, Montpellier, France)

Dr. Hovorka: If you are moving patients with the AP to the home, in your experience how much more knowledge and training would you say the closed loop requires compared to what they already know?

Dr. Renard: For patients who have used a pump, a sensor, and know how to carbohydrate count, within two days they can go home. We need to select patients who want to improve their diabetes management and if they are dedicated, I don’t expect it to be a long time in the hospital before they can go home.

Dr. DeVries: That being said, considering pump penetration around the world, for many countries it would be a huge step for a lot of patients. There are countries with pump penetration below 5%, so from that perspective it’s not for everyone.

Dr. Jeffrey Joseph (Thomas Jefferson University, Philadelphia, PA): Is the improved control with intraperitoneal insulin due to suppressed hepatic glucose production?

Dr. Renard: Animal studies suggest that this occurs. In these studies, the main difference compared to subcutaneous insulin is with the speed that hepatic glucose production is suppressed.

Comment: It is evident that for any algorithm, whether PID or MPC, one size does not fit all. We need to do more physiology studies so we can individualize the algorithm to the patients. From a sensor standpoint, that is one of the weakest links in closed-loop systems. We need to understand the physiology of the delay in glucose sensing. That will hopefully improve sensor algorithms.

Dr. Ken Ward (Oregon Health & Science University, Portland, OR): Dr. Renard, I agree with the examples you gave of likely candidates. However, I would note that our closed- loop studies have enrolled patients who are keep themselves in good glycemic control but with great personal cost – 12-to-13 fingersticks per day and constant vigilance. These patients say that they greatly appreciate being in closed-loop studies and having a vacation from diabetes management, even for a day-and-a-half. They might be good candidates for an artificial pancreas as well.

Dr. Buckingham: I would add that the typical adolescent, who is on a “vacation” most days, is also a good candidate.

Dr. DeVries: I would agree that a one-day break can be of huge value to patients.

Dr. Renard: At the same time, in these patients it would probably be hardest to show the benefit of AP. The patient with whom we did outpatient research said that the only benefit was that he didn’t have to care about diabetes. But this is difficult to explain scientifically.

Dr. DeVries: Let alone from a reimbursement point of view.

Q: To Hans, in the study you just did, was the meal bolus a full calculated meal bolus? And then Eric, you showed a sensor profile where it overestimated by a 1/3. We are concerned about a sensor that is 1/3 high, but we routinely acknowledge that a patient can’t count carbohydrates. How big is meal bolus and how important is it to get it right?

Dr. DeVries: The bolus was determined by the algorithm and the input differed by the two algorithms. I’ll defer to Roman.

Dr. Hovorka: The Cambridge algorithm gave 80% of the calculated meal bolus.

Comment: We routinely acknowledge that in open-loop parameters patients will have the wrong carbohydrates. It strikes me that the closed loop should be more robust to these errors.

Dr. Renard: If you look at current methods, you just look at what’s on the plate and compute carbohydrates in your brain. We need some tools for patients that will be easy to manage. We have developed better tools now, but in this we are still very conservative. There is room for technology to improve this. We can use components of the closed loop to help patients count carbohydrates.

Comment: Carbohydrates are difficult to determine if patients go to a restaurant and have something they didn’t prepare. Dieticians in California were asked to rate meals on the amount of carbohydrates and they pretty much all got it wrong. Counting carbohydrates is difficult.

Dr. Renard: You cannot control everything in life.

Dr. DeVries: If I go to US restaurants, the number of carbohydrates is often on the menu.

Dr. David Klonoff (Mills-Peninsula Health Services, San Mateo, CA; University of California, San Francisco, San Francisco, CA): Every one of these systems uses a specific combination of hardware. How do you choose a combination? Also, when you consider all of these possibilities, not to mention the variety of control algorithms and ways of handling meals, what do you think this means for establishing performance standards?

Dr. DeVries: In the CAT Trail we used the products that were made available to us. Companies like Dexcom invest a huge amount of time, and also money, making devices for closed-loop research. The same goes for Insulet – these companies should be commended.

Dr. Renard: We should have the companies working on platforms that work with any pump and any sensor. As a patient you should not be constrained in your choice of pump because you want to have an artificial pancreas. This will be a new world of diabetes management; I think that the environment should be open. The platform will be validated on pumps X, Y, and Z, not limited to a single manufacturer.

Dr. Buckingham: I think that we all use what is made available to us and appreciate the companies that make their systems available. I think that if you take a car and it gets you from point A to point B – i.e., a good glycemic result – it is a good automobile. However much it costs, I think that it’s worth it.

Dr. Hovorka: I think that instrumentation has been the limiting factor thus far.

Q: For the predictive system, did you impose any constraints on the suspend time?

Dr. Buckingham: I didn’t state that the threshold was 80 mg/dl and a single episode couldn’t be longer than two hours. There was a maximum of three hours overnight for total suspension time. That was protection against a sensor failing overnight.

Q: In the future we may have ultrafast acting analog insulin. What will be the impact of having that insulin?

Dr. Hovorka: I think it will have a tremendous affect on efficacy and safety. If you have something in your sleeve, as soon as possible would be good.

Dr. DeVries: It is always difficult to predict future and if we would have these faster do we need to go the intraperitoneal route?

Dr. Renard: It depends on whether it is variable in absorption.

Dr. Stuart Weinzimer (Yale University, New Haven, CT): To address Dr. Klonoff’s question: as primarily a clinician, I could envision picking out a closed-loop system that I think is best for individual patients, just as I now do when choosing among diabetes products now. There won’t be one system ideal for every scenario.

Dr. Hovorka: Agreed.

Q: To repeat a question asked to the panel of engineers, what do you think has to happen to get an artificial pancreas on the market? I was surprised that none of the engineers mentioned the sensor as being a problem, but I noticed that two of you did.

Dr. DeVries: I think the engineers did a very good job answering the question. One thing that they perhaps did not mention is the legal ramifications of closed-loop control. We will probably see severe hypoglycemia with a closed-loop device at some point in the future. Who is responsible? What are the consequences? My view is that if the trials are decisive and the labeling is clear, we will just have to accept these risks.

Dr. Buckingham: The Helmsley Charitable Trust data indicates that 7.5% of patients are having seizures or loss of consciousness each year; people are dying now from the dead-in-bed syndrome. I think that if you can reduce these risks with an artificial pancreas, then it is good to use one. The sensors are getting better, and this will lead to better system performance.

Dr. Renard: When you use closed-loop algorithms, you reduce time in hypoglycemia and thereby in a way improve sensor performance, since accuracy is worst when glucose is low.

Dr. Renard: Hypoglycemia reduction isn’t seen in every case when mean glucose goes up, though; using a closed-loop algorithm is preferable to simply relaxing control.

Dr. Hovorka: How are we going to evaluate home studies? With CGM? Is CGM good enough?

Dr. Buckingham: I think it is. You don’t have another choice other than someone living in the house with a YSI machine. We’re going to have to rely on continuous glucose.

Dr. DeVries: We see now in translation trials that we completely rely on CGM, but groups apply telemonitoring too. Your group, Roman, is going out of hospital without telemonitoring. I’ll turn the question back to you, is it safe?

Dr. Hovorka: My question was different. Once you collect CGM, we found in our analysis that there was a bias introduced in using a CGM.

Dr. Renard: Safety authorities make it so difficult for outpatient trials. People with diabetes have nothing when they inject insulin before going to bed. There is no CGM, no measurement...what should be forbidden is to use insulin or a pump without CGM or frequent measurements. It’s curious to see how many hurdles there are to outpatient trials, when it is a safer way than how we use insulin now.

2. Glucose Monitoring and Glycemic Variability

JDRF Student Research Award


Yoeri M. Luijf, MD, MSc (Academic Medical Centre, Amsterdam, The Netherlands)

This year’s JDRF Student Research Award Winner, PhD-candidate Dr. Yoeri Luijf, presented a 20- patient comparison of three CGM sensors: the Abbott FreeStyle Navigator I, the Medtronic Enlite, and the Dexcom G4 Version A (i.e., the version used with the Animas Vibe in Europe, not the more advanced G4 Platinum available in the US). During the in-clinic portion of the study on day one, the G4A’s accuracy was significantly worse than that of the Navigator and Enlite (as measured by mean absolute relative difference, MARD). However, during the home-use period that followed, the Navigator and G4A had statistically similar accuracy, and the Enlite’s was significantly worse. (Dr. Luijf thus hypothesized that the G4A’s warm-up time may be longer than indicated by the manufacturer.) Sensor longevity was also assessed; all three sensors had median lifetimes longer than their respective indicated wear times, but accuracy for sensors that outlasted their indicated wear-time was significantly better for the Dexcom sensor than Abbott’s or Medtronic’s. During Q&A Dr. Luijf said that the researchers will conduct a follow-up study using the G4 Platinum, with the in-clinic portion of the study conducted on day three rather than day one to ensure that all three sensors have fully warmed up.

  • The study enrolled 20 patients with type 1 diabetes, who were simultaneously fitted with three different CGM sensors: the Abbott FreeStyle Navigator I, the Medtronic Enlite, and the Dexcom G4 version A – i.e., the sensor approved for use with the Animas Vibe in Europe, not the more-advanced G4 Platinum. On the first day of sensor wear, patients stayed in the clinical research center for a glycemic challenge (breakfast with an insulin bolus that was delayed and then increased). Reference blood glucose values in the clinical research center were taken with YSI. After this in-clinic portion, one of the three sensors was randomly removed so that patients would need to wear only two sensors for the rest of the study. Each patient thenwore those two sensors at home, taking fingerstick blood glucose measurements for reference. To assess sensor longevity, patients wore each sensor for as long as they could until apparent technical failure or two consecutive days of mean absolute relative difference (MARD) greater than >25%.

  • During the clinical research center (CRC) portion of the study on day one, the Dexcom G4 A had significantly worse YSI-matched accuracy than either the Abbott FreeStyle Navigator I or the Medtronic Enlite (which were not statistically different from each other). This same pattern was seen in sub-analyses of glucose values below 100 mg/dl and between 100 and 200 mg/dl. For glucose values above 200 mg/dl, however, accuracy did not significantly differ between the sensors.





Navigator I









MARD = Mean absolute relative difference; SD = Standard deviation

  • The median ± interquartile range of sensor longevity were as follows for the Navigator (8.5 ± 3.5 days), G4 A (10.0 ± 1.0 days), and Enlite (8.0 ± 1.5 days). Maximum observed sensor lifetime was 26 days for the Navigator, 82 days for the G4A, and 15 days for the Enlite. (Dr. Luijf noted that three of the Dexcom sensors lasted for over 40 days, though he emphasized that these were outliers.) As a reminder, the indicated wear time is five days for the Navigator I, seven days for the G4A, and six days for the Enlite.
  • During the study’s home phase, accuracy within labeled wear time was statistically significantly worse for the Enlite than the Navigator or G4A; beyond labeled wear time, the G4A’s accuracy was best by a significant margin. Accuracy during home use was assessed by comparison to self-monitoring of blood glucose (SMBG) fingerstick values.



MARD during specified lifetime

MARD after specified lifetime

Navigator I









MARD = Mean absolute relative difference

Questions and Answers

Q: You were comparing error rates from the CRC and in the home. What was the home standard?

Dr. Luijf: SMBG fingersticks values.

Q: Can you comment how you decided the site of implantation for the different sensors because there are differences in dimensions? And did you do any analysis in males versus females?

Dr. Luijf: Our procedures were straightforward. To be as honest as possible to the CGM manufactures, we placed the device according to manufacturer specification. So they were all in the abdomen. And we did not look at differences between males and females, but remember we only had 20 participants so there would probably be a power problem. As you are aware, female are less likely to participate in scientific research.

Q: Have you characterized BMI or fat content?

Dr. Luijf: Yes, from top of my head, BMI was 23 kg/m2, but we don’t think it would influence the performance of the sensor.

Q: For the last speaker, you implanted three sensors on the same patients. Were the wounds different? Was one wound much worse than another?

Dr. Luijf: You are absolutely right. There are patients that are sensor-genic and others that are not sensor- genic. There are individual parameters that influence sensor performance. That’s why all sensors were worn on the same patients – the individual differences cancel each other out.

Q: You talked about the outliers – 82 days of wear in one patient for the Dexcom. There were 20 sensors – what were the significant outliers?

Dr. Luijf: Of course we wondered about 82 days of wear. I triple checked all the data that it lasted for 82 days with good accuracy. This study shows that it’s possible to use certain sensors beyond a manufacturer’s lifetime. It’s okay to reactivate the sensor. But it’s a whole different thing when you look at the Kaplan-Meier analysis: 50% of sensors fail before 10 days.

Q: Why did you do CRC assessment on the first day?

Dr. Luijf: The goal was to assess the sensors in a real-life setting and in a setting which manufacturers use themselves. Manufactures do it in the CRC, and in the CRC we induced aggravated glucose excursions on purpose. Our two comparisons are assessments in extremely tough conditions and real life conditions.

Q: Do you think the Dexcom performance was not related to warm up but using extreme conditions?

Dr. Luijf: The MARD on all three sensors was higher in the CRC. For Dexcom, we did the CRC assessment on day one and we think it just needed a longer time to warm up. We are going to repeat the study with the G4 Platinum and do the assessment on day three. For the other two sensors, I do believe the MARD CRC was higher because they were in tougher conditions.

Arleen Pinkos (FDA, Silver Spring, MD): Another possibility why the CGM comparison data from the CRC was poorer than the home data – it may not necessarily be that the CRC was more challenging. From the data we’ve seen, most home blood glucose values are collected before meals. At that time, patients tend to be more steady state and the sensors are more accurate.

Dr. Luijf: You are absolutely right. We published an article on this last year on the differences between the CRC and home environments. Your hypothesis is completely right about fingersticks just before meals.


The Sensor-Tissue Interface: Industry Panel


Joseph Lucisano, PhD (President and CEO, GlySens Inc.)

Dr. Joseph Lucisano presented a promising first-in-human feasibility study of the first-generation GlySens implantable glucose sensor, building on positive 18-month data in pigs (Sci Med Transl 2010). Implanted for six months in insulin-using patients with diabetes (n=6), the glucose-oxidase (GOx)-based sensor was generally well tolerated. Sensor performance was stable over time with calibrations performed once per month (during in-clinic glycemic clamp studies). However, the time offset of the sensor from plasma glucose varied from day-to-day, patient-to-patient, and based on whether glucose was rising or falling. Dr. Lucisano said that GlySens’ second-gen sensor features a longer range for data transmission, more optimal sensitivity to oxygen, and no contact between tissue and the sensor’s enzymes (the source of the only adverse event in the first-gen sensor’s feasibility study). This next-gen sensor is in “the final stage” of development and is anticipated for clinical evaluation in early 2013. The company’s goal for its first commercialized product is yearlong implantation with quarterly calibration.

  • The GlySens sensor is based on the research of Dr. David Gough at the University of California, San Diego and uses a “dual-detector” method. In addition to glucose oxidase, the system uses another enzyme called catalase that breaks down hydrogen peroxide (a byproduct of the glucose oxidase reaction). Both enzymes’ reactions are monitored, which enables sensing to be more specific and also extends the lifetime of the sensor (because hydrogen peroxide accumulation can be deleterious).

  • Tolerability, the feasibility study’s primary endpoint, was generally favorable for six months. The only adverse event was an increase in non-neutralizing antibodies to glucose oxidase in one patient, which led the investigators to explant the sensor at five months. Dr. Lucisano said that this would likely not be an issue in future versions of the system, since tissue- enzyme contact has been precluded in the second-generation sensor. The implant site was initially somewhat painful, but patient-reported data were consistent with better tolerability over time. No incidents of swelling, erythema, itching, or significant discomfort were reported.

  • Dr. Lucisano presented an analysis of time offset (lag or lead) as measured during the monthly glucose clamp studies. When glucose was rising, the mean ±SD lag from plasma glucose values was 9 ± 6 minutes (maximum lag 17 minutes). When glucose was falling, lag was -3 ± 12 minutes (maximum lag 9 minutes). Dr. Lucisano noted that the mean ‘lag’ was negative during falling fronts, meaning that sensor glucose changes actually preceded blood glucose changes. During Q&A he proposed that this likely reflects true physiology rather than a software error, though we have not often heard this phenomenon described for current CGM sensors. (He also acknowledged that the high levels of insulin during the clamp study could have had confounding effects and that this analysis was still important to conduct.) When glucose was falling and in the hypoglycemic range, lag was 7 ± 4 minutes (maximum lag 13 minutes). Dr. Lucisano noted that the time lag did not appear to change over time in a systematic way, but rather to show high inter-day variability throughout the six months. Calibration drift averaged 4% per week; the worst weekly drift observed was 9%.

  • The feasibility study’s overall accuracy analysis included 1,097 sensor glucose values paired with four-times-daily home fingerstick glucose tests. Nearly 30% of fingersticks were excluded because no sensor data were available within 10 minutes of the measurement; Dr. Lucisano indicated that the communication range between the first-generationsensor and receiver was too limited but that this had been corrected in the gen two. The overall Clarke Error Grid Analysis placed 88% of points in either the A or B zones, 11% in the C zone, and 12% in the D zone; overall MARD was roughly 20%. Dr. Lucisano also showed Clarke Error Grid results for a single subject (presumably the one with the best results): for 125 data pairs, that patient had 62.4% in the A zone, 31.2% in the B zone, and 5.6% in the C zone; MARD for this patient was roughly 16%.

  • Dr. Lucisano summarized the status of GlySens’ next-generation sensor, which is being targeted for clinical evaluation in 2013. A smaller sensor is in production, the telemetry signal has been improved, and the sensor’s response range had been broadened.


Fotios Papadimitrakopoulos, PhD (Scientific Advisor, Biorasis Inc.)

Dr. Fotios Papadimitrakopoulos presented several approaches to mitigating the foreign body response to implanted sensors. His company, Biorasis, is studying sensors coated with microspheres that gradually degrade to release dexamethasone (to reduce inflammation) and VEGF (to increase vascularization). An indirect advantage of this approach is that as the microspheres degrade, they create macro-sized pores around the membrane; this counterbalances the negative effects of fibrous capsule formation. During the question and answer period Dr. Papadimitrakopoulos said that the major limitation of this approach is the amount of drug that can be loaded around the sensor and that Biorasis is targeting sensors with a three-to-six-month lifetime.



Robert Boock, PhD (Director R&D, Dexcom)

Dr. Robert Boock presented data from the pivotal trial of the Dexcom G4 Platinum CGM, which was approved by the FDA on October 5, 2012 (see https://closeconcerns.box.com/s/mzv18lhhd8hmtevwubqc). Compared to the Dexcom Seven Plus, the G4 Platinum showed better overall accuracy (with improvements seen from day one to day seven in the same sensor), less inter-individual variability, better accuracy in the hypoglycemic range, less sensitivity to low oxygen concentration, and less “background” interference. Dr. Boock attributed much of this improvement to the G4 Platinum’s new sensor membrane, designed for improved biocompatibility. However, he said that concurrent benefits came from the G4 Platinum’s improved algorithm and that the membrane and algorithm effects were hard to separate. To conclude Dr. Boock said that Dexcom’s experience with its latest device “gives us a good feeling that we have not yet reached the optimum performance of a glucose-oxidase-based sensor.” For greater detail on the G4 Platinum’s clinical accuracy, see our coverage of Dr. Tom Peyser’s similar presentation at EASD 2012, available at https://closeconcerns.box.com/s/ajac8dtnvl26v8v8z0ik.



Achim Mueller, PhD (EyeSense GmbH, Grossostheim, Germany)

Dr. Achim Mueller described results from two clinical studies of EyeSense’s percutaneous fluorescent glucose sensor. The first (n=6; two-week wear time) was completed in October 2011, and the second (n=10; four-week wear time) concluded in October 2012. Measuring glucose with this prototype sensor required interrogating the implanted sensor with a photometer in the clinical research center. Calibrations lasted for as long as 12 hours, the longest in-clinic sessions in the study. Accuracy results were better for sensors placed in the upper arm compared to the abdomen due to less movement of the tissue (for the four-week study, Clarke A zone scores were 92% vs. 88.5%). Overall mean absolute relative difference (MARD) results looked good (~10% in the upper arm and ~11% in the abdomen), though accuracy worsened somewhat over time (four-week MARD of ~11% in the upper arm and ~14% in the abdomen).


Dorian Liepmann, PhD (University of California, Berkeley, Berkeley, CA), Mike McShane, PhD (Texas A&M University, College Station, TX), Fotios Papadimitrakopoulos, PhD (Biorasis Inc, Storrs/Mansfield, CT), Joseph Lucisano, PhD (GlySens Incorporated, San Diego, CA), Robert Boock, PhD (Dexcom, San Diego, CA), Achim Mueller, PhD (EyeSense GmbH, Grossostheim, Germany)

Q: Dr. Lucisano, did your data include correction for time lag or time lead?

Dr. Lucisano: There was no time-lag correction in any of the data you saw. At GlySens we have seen no systematic change in the time lag over time; the capsule per se does not seem to affect this. However, there is day-to-day variability in time offset.

Q: What was the MARD?

Dr. Lucisano: About 20% overall and 16% in that one subject that I highlighted.

Q: Are data taken from one sensor or multiple sensors at a time?

Dr. Lucisano: There is an array of detectors and an averaging scheme.

Q: You said that most of the time the sensor estimate lagged blood glucose but that sometimes on the downward trends the sensor glucose trend led. Do you think that this true and physiological or an issue of miscalibration?

Dr. Lucisano: We believe that it’s physiologic. The belief is that when glucose is falling, glucose is being consumed by the peripheral tissues first. So, you might expect the interstitial fluid glucose to lead the blood glucose.

Q: Can you comment on where the sensors were implanted and if there was any restriction to patient movement?

Dr. Lucisano: The sensors were implanted in the abdomen, essentially right below the belt line. We had no issues with movement.

Q: Did you take measurements of capsule thickness?

Dr. Lucisano: We did not do tissue analysis.

Q: Dr. Mueller, how often did you calibrate in 28 days?

Dr. Mueller: We calibrated each time that patients came in.

Q: How long is that calibration good for?

Dr. Mueller: We measured for about 12 hours in a row, so we know that it lasts for at least that long. Because of the sensor properties we think that it would last longer.

Q: Dr. Lucisano, you explanted a sensor because of non-neutralizing antibodies to glucose oxidase. But every type 1 diabetes patient makes antibodies to insulin. Have animal studies seen whether anti-GOX antibodies cause ill effects?

Dr. Lucisano: In animal studies we have seen neutralizing antibodies to GOx. This was not the case in our study, however.

Q: In the animals generating those antibodies, does something happen to them clinically? Dr. Lucisano: I am not aware of any ill effects to the animals.

Q: There is very little data on the effects of changes in the membrane interface. Dr. Boock, you showed data that included changes in both the membrane and the algorithm. Did you ever study the new membrane with the old algorithm?

Dr. Boock: In most of our experiments we designed the algorithm wedded to the sensor. In some early studies we got an interesting understanding of what’s going on. Surprisingly, the interface is one of the most misunderstood and misrepresented aspects of CGM. I think we can’t underestimate what’s going on there. Controlling the materials properties is essential for getting good performance.

Dr. Yogish Kudva (Mayo Clinic, Rochester, MN): Dr. Lucisano, what was the longest period as far as the implantable glucose sensor was concerned?

Dr. Lucisano: Six months in this study, except for the patient whose sensor was explanted at five months. We have gone out to 18 months in animals.

Dr. Jessica Castle (Oregon Health & Science University, Portland, OR): Dr. Mueller, I noticed that the number of subjects analyzed fell from nine down to five in day 28.

Dr. Mueller: That was a problem with the photometer such that no measurements could be taken in some patients on that time.

Dr. Castle: Dr. Lucisano, can you elaborate on how someone might achieve similar calibration at home that mimics your in-clinic glucose-clamp method?

Dr. Lucisano: The details of implementation of calibration have not been worked out. However, the goal would be once-quarterly calibration with reference plasma values, in the clinic.

Q: Dr. Papadimitrakopoulos, do you have data on the tissue change around your sensor over time?

Dr. Papadimitrakopoulos: Those studies have been reported. We have done histology up to three months; there was no inflammation in the rat model we used. We have repeated these studies over again and optimized the dose of DXM that we gradually release. We have never seen divergence. Based on that we assume that things are the same, and we continue to tackle biofouling. We are working also with mini-pigs now and have seen similar results to the rat studies.

Q: Dr. Boock, when you talk about improving biocompatibility, I assume you are talking about the interface rather than traditional biocompatibility.

Dr. Boock: Yes, we are talking about sensor-tissue interactions that affect response. This is different than a traditional biocompatibility analysis.

Q: What endpoints are you looking at – the tissue? Macrophage response?

Dr. Boock: We are looking at noise phenomena and other factors more relevant for our purposes. We are trying to reduce artifacts and improve device performance, not necessarily to look at specific tissue effects.

Dr. Papadimitrakopoulos: Dr. Lucisano, these are fabulous results. The ability to have such a long-lasting device is amazing to me. The way I understand the mechanics is that the device should have mechanical properties whereby it slowly ruptures the fibrous tissue around it, so that you are constantly exposing the device to fresh plasma. In the Sci Transl Med paper about the sensors in pigs at 18 months, you must have formed fibrous capsules. How can the glucose and oxygen penetrate those capsules? If you are using mechanical force to loosen the capsule, when you shrink the device to a smaller size will you compromise these effects?

Dr. Lucisano: There is no indication that the fibrous capsule is being continuously ruptured; it seems intact in animal studies. The raw signals do decrease at six-to-eights weeks and then stabilize at a lower level. A capsule forms, but it still seems to have enough permeability to allow sufficient glucose and oxygen to cross. Using the differential sensor technique that I described takes out the effect of the capsule, which seems to affect glucose and oxygen to a similar degree.

Dr. Liepmann (University of California, Berkeley, Berkeley, CA): I was fascinated that the response is better in the upper arm than the abdomen. Could you discuss?

Dr. Mueller: We see less movement in the upper arm. From the first to second trial we tried to improve the fixation in abdomen, which improved the data quality.

Dr. McShane: We saw some fully implanted sensors and some percutaneous ones. For the developers of each, what do you see as the longest operating time they can achieve, and what is the limiting factor?

Dr. Papadimitrakopoulos: For Biorasis, there is no problem for the device per se to last beyond three months and up to a year. The limit is the amount or real estate that we have around the coating – how much dexamethasone we can load to have a sustained release. Once the DXM runs out, the inflammation sets back in, the fibrous capsule sets back in, and the device stops working. You don’t lose 100% of the signal, as Dr. Lucisano noted, but you might lose 90%. When you have such a tiny device as ours, 90% is bad. Our goal is t0 build a system that lasts for three-to-six months.

Dr. Lucisano: Our goal at GlySens is a one-year lifetime for the first-generation commercial system. Limiting factors include battery life and enzyme activity. Until enzyme activity drops below a critical activity level, though, there is no loss of calibration. We think the enzyme activity is up to two years at this point; we are targeting only one year at first.

Dr. Boock: Most of our sensors’ lives are dictated by adhesive; limits here must be overcome to get beyond Dexcom’s seven-day limit currently.

Dr. Mueller: The EyeSense sensor per se is very stable; it was implanted in eyes for up to two years. Maintaining the insertion of a percutaneous sensor was a challenge for us, too. Work remains in improving the signal-to-noise ratio.

Dr. Tessa Lebinger (U.S. Food and Drug Administration, Silver Spring, MD): Dr. Lucisano, some literature suggests that in a hyperinsulinemic clamp study, you may have increased glucose uptake by the cells, so that the normal uptake in the peripheral tissue is raised and the glucose gradient changes. Did you see whether this affected the sensor lag and lead times?

Dr. Lucisano: We recognize that the clamp studies may not have been representative of daily life. We have not yet looked at correlation of insulin infusion rate with lag and lead times, but we intend to do so.


The Sensor-Tissue Interface: Academic Panel


Kristen Helton, PhD (University of Washington, Seattle, WA)

Dr. Kristen Helton presented an overview of best practices and experimental considerations for those developing and testing new continuous glucose sensors. She recommended that to minimize displacing forces, sensors should be stabilized deep in the subcutaneous space and/or made of a flexible material that resembles the local tissue. Sensor size and porosity are important, but so is the site of sensor placement (on this note, she said that we do not yet have enough comparative data on different sensor insertion sites in different animal models). Dr. Helton also mentioned that some sensing technologies are sensitive to a lack of oxygen. One way to address this problem is by including oxygen sensors on the CGM electrode, so that these artifacts can be canceled out. In light of some of the challenges that she mentioned, for her own work Dr. Helton uses an optical glucose sensor (which can be made of a flexible hydrogel and is stable during oxygen shortages).



Julia Mader, MD (Medical University of Graz, Graz, Austria)

Dr. Julia Mader reviewed the preliminary clinical studies that have been conducted on infusing insulin and measuring glucose at a single subcutaneous insertion site. In people without diabetes (n=5), the tissue-to-plasma glucose ratio remained constant during all but the start of six-hour insulin infusion regimen using either microdialysis or microperfusion cannulas (Lindpointner et al., Diabetes Care 2010). In otherwise healthy people with C-peptide-negative type 1 diabetes (n=10), insulin infusion from a microperfusion catheter resulted in a median absolute relative difference of 10.9% for tissue glucose vs. plasma glucose (Lindpointner et al., Diabetes Care 2010). Most recently, another single-port study in C-peptide-negative type 1 diabetes (n=13) showed better results still: glucose measurements in the tissue around a perforated cannula had a median absolute relative difference of 8.0% and a Clarke A zone score of 83.5% compared to plasma glucose (Regittnig et al., Diabetes Technol Ther e-pub 2012). Upcoming studies in the European AP@home project will evaluate existing commercial CGM sensors either inserted in vivo or placed just outside the body (with glucose drawn up to contact the sensors). Dr. Mader also mentioned that AP@home research has recently begun on a glucose-sensing “tattoo” technology called SPIDIMAN, which could also be used in a single-port approach.



Jeffrey Joseph, DO (Thomas Jefferson University, Philadelphia, PA)

In a presentation jam-packed with physiological insights and engineering recommendations, Dr. Jeffrey Joseph discussed a variety of factors that affect the body’s inflammatory response to subcutaneous sensors. For future CGM products, Dr. Joseph suggested the use of redundant sensor arrays, optimized membranes, standardized depth and insertion site (for more predictable performance), a lubricious insertion needle that pushes cells out of the way rather than slicing through them, a soft electrode (like “limp spaghetti”) that will not move around inside the body and re-aggravate the initial wound, and a method to inhibit blood clots around the sensor (because “the most important stimulus for inflammation is clotted blood”). Dr. Joseph also presented some interesting data on insulin infusion (simulated by infusing dye in a pig model). His group found that if a fibrin layer had completely enclosed the catheter, dye would actually drifted up onto the skin rather than going into the body. This result seems to imply that more traumatic insertions are detrimental not only to CGM performance but also insulin pharmacodynamics and pharmacokinetics.



Kaiming Ye, PhD (National Science Foundation, Arlington, VA)

Head of the National Science Foundation’s Biomedical Engineering program, Dr. Kaiming Ye had little to say about glucose sensing. Instead he discussed research on turning stem cells into insulin-producing beta cells. Dr. Ye noted that cellular differentiation is affected not just by the soluble molecules used in traditional two-dimensional cell cultures but also by time- and space-dependent factors like pressure, oxygen exposure, and insoluble proteins in the extracellular matrix. Thus his group is studying a three- dimensional synthetic scaffold on which glucose-responsive insulin-secreting cells can be developed with 40-60% yield (as compared to 7% with 2-D cultures). This work remains in very early stages, as highlighted during Q&A, but we are certainly glad to see cutting-edge type 1 diabetes research getting attention at the highest federal levels.



Christine Kelley, PhD (National Institutes of Health, Bethesda, MD), Yogish C. Kudva, MBBS (Mayo Clinic, Rochester, MN), Kristen Helton, PhD (University of Washington, Seattle, WA), Julia Mader, MD (Medical University of Graz, Graz, Austria), Jeffrey Joseph, DO (Thomas Jefferson University, Philadelphia, Pennsylvania)

Q: In addition to their inflammatory effects, neutrophils secrete hyaluronanidase. Dissolving the hyaluronan has a profound effect on the diffusion of large molecules like insulin as well as even glucose and water. Have there been studies of hyaluronidase on insertion of CGM sensors? It is a drug that is available and could be used.

Dr. Joseph: It is a good idea, and I suggest you look into it. There is a dynamic local environment, and these large structural proteins and macromolecules affect the wound and the diffusion of relevant molecules. If you could use this enzyme and make it a more fluid environment, I think that is a good area but it is an active area of research. To my knowledge it has not been used in a sensor surface, though it is being studied for insulin delivery.

Q: A great part of ability to recognize tissues have to do with integrins and other fixed proteins in ECM. Have you tried a decellularized matrix to see if you get better growth and differentiation? Cardiac tissue has been completely regenerated using such a cell-free matrix.

Dr. Ye: We have up until now focused on using a synthetic material, to see if we can mimic the in vivo environment. We used an atomic force microscope to characterize the mechanical stiffness we made and compared it to pancreas from the mouse. We found that the two tissues were very similar in response to force. If we inoculate human stem cells in extracellular matrix, we are not sure whether they would differentiate into more mature cells. Another group has tried this, but without too much success based on our reading of their paper.

Dr. Horacio Rilo (University of Arizona, Tucson, AZ): In doing clinical islet and experimental transplantation, I hear people discuss benefits of human stem cells but never drawbacks. Have you looked at teratomas?

Dr. Ye: The big problem I have seen is that those cells differentiate from adult cells and function well in the lab, they are responsive to glucose, but in transplanting them to mice they have problems.

Dr. Rilo: I have data on how beta-cell differentiation occurs in humans. They come from a common precursor and so co-express insulin and glucagon. We are using the wrong models – perhaps we should implant the precursors in vivo to finalize maturation. Also, you showed a response to high glucose, but what about low glucose? A problem with stem cells is that they don’t attenuate their response in low glucose.

Dr. Ye: We tried 5 mmol/l (90 mg/dl) to 20 mmol/l (360 mg/dl) to challenge the developing cells; we haven’t tried a concentration lower than 5 mm (90 mg/dl). That is an excellent comment.

Dr. Ken Ward (Oregon Health & Science University, Portland, OR): Dr. Helton, I saw you had a slide on membrane pore size. Recommendations on the appropriate pore size range from 5 to 50 um. It looked like you had a systematic inquiry into pore size and foreign body response. Does pore size matter?

A: If you separate out the data in our study, we do see a difference between smaller and larger pores. From the Ratner lab we have seen data that in the mouse model 40 um was optimal; that was not reproduced in a rat model. Also, the size of the implant is critical. It is difficult to do a systematic comparison; often the methods used by other researchers have varied quite a bit, and so have the sizes of the implants. There are a lot of confounds there.

Q: You mentioned that insertion trauma could be minimized with an automated inserter. Can this be combined with the slow insertion you recommended, such as is used in inserting brain electrodes?

Dr. Joseph: If you go in faster, there tends to be less patient discomfort but more cellular injury. It is a tradeoff; the answer is that instead of the needle being the initial site that penetrates the tissue, we are looking at some other way in material science: perhaps a jet of water.

Q: You said that the gradient is stable. If we had smaller devices and could pick up more local points around the infusion site, would we see a difference? What is the variability between locations or from one person to another?

Dr. Mader: We have to test in other patients to see if gradients stable, but so far this is what we have seen. If someone can provide advice for higher resolution of monitoring, this would be great. Currently we can get needles only so close together.

Dr. Bruce Buckingham (Stanford University, Stanford, CA): We know that insulin pharmacokinetics and pharmacodynamics change over time. Did you try aging the infusion site catheters in your pig studies?

Dr. Joseph: Unfortunately the projects we’ve done have been only on day 3 and day 5. We want to insert the pig with sensors on 1, 2, and 3 days pre-euthanization. We did not do this in the current studies. But if the fibrin layer was complete, dye tended to track up to the skin; this was surprising to use.

Dr. Buckingham: I think that the skin gets thicker as pigs age, and as you travel farther from the midsection on any given pig. Did you have that problem as you went laterally?

Dr. Joseph: we didn’t notice the skin thickness depending on location. We tended not to be up in the rib area because the muscle is close to the skin. When you went down to soft belly though, the epidermis and dermis tended to tent, and you had difficulty getting the sensor in. if the tissue underneath was firm you got better results, but we tended to be in soft belly to try to approximate where the sensors are placed in humans.

Dr. Tessa Lebinger (U.S. Food and Drug Administration, Silver Spring, MD): Are any of you aware of any data on patients on anti-clotting medications and whether that could potentially affect sensor performance?

Dr. Joseph: We have done some work placing sensors in hospitalized patients. When you insert a sensor you get clot formation; once you form it, heparin should not cause active bleeding. But in some patients we have had active bleeding. The timing of when a patient gets anticoagulated seems to be an issue. The bottom line is that we don’t know the answer to your question. In the hospital, sensors we have had increase bleeding in patients on blood thinners. You asked whether that affects sensor performance; typically what happens is that blood comes up the tissue and wets the sensor, and the adhesive falls off.

Q: Instead of developing a weeklong sensor, we are trying to develop one that can last for a month or year. We found that the reactive oxygen species (ROS) can be a huge problem for the sensor. Do you worry about ROS generated by traditional electrochemistry?

Dr. Helton: I am also in sensor development; I think that having strategies to monitor ROS and potentially scavenge is absolutely the way to go.

Comment: It seems worst in the first week. So far I don’t see many papers on ROS interference.

Dr. Helton: Certainly percutaneous sensors are trying to sense during the most dynamic time in wound- healing response. If you can get past the first week or so and don’t have ongoing inflammation, and your sensing chemistry can make it through the first week, I think you are on the path toward long-term sensing. I think a lot of people are thinking about ROS and how to monitor it so that we can better characterize the body’s response to sensor insertion.

Dr. Kudva: Dr. Helton, you briefly talked about flexible sensor materials. What is exciting in development?

Dr. Helton: I am not really focused on electrochemical sensors. We are using softer tissue-like materials like hydrogel and using optical signals through the skin to measure glucose.

Dr. Kudva: Could you comment on developing good animal models?

Dr. Helton: I think the best you can hope to do is track multiple sensors in an animal; try to track them all, and do the follow-up histology. I wish there was more literature on the right implantation site and animal model to use.

Poster Presentations


J Tamada, A Jina, M Tierney, S McGill, S Desai, and B Chua

In a clinical trial assessing the efficacy of a MicroTip Array-Based continuous glucose monitor (developed by ArKal Medical), the sensor recorded 15% mean absolute relative difference and 75% of measurements in zone A during a Clark Error Grid analysis. This non-invasive system relies on passive diffusion of glucose from the interstitial fluid through MicroTip array chips (MACs), consisting of ~200 MicroTips each, that penetrate the outermost layer of dead skin. MACs are attached to a glucose diffusion chamber filled with ArKal’s proprietary buffer solution to support glucose diffusion; the chamber in turn is fixed to the sensor, which measures glucose via immobilized glucose oxidase on the sensor’s electrode. While the system’s accuracy is not quite on par with commercially available subcutaneous CGMs (for comparison, in a Clarke Error Grid analysis Dexcom’s G4 Platinum and Abbott’s FreeStyle Navigator recorded an ever so slightly higher 80% and 83% of readings in the A- Zone, respectively) we found this initial data quite impressive and look forward to following the device’s development as the non-invasive factor would undoubtedly be greatly appreciated by patients.

  • The continuous glucose monitor contains MicroTip array chips (MACs), which are each comprised of ~200 MicroTips in a six mm2 array. The MicroTips penetrate the stratum corneum (the outermost layer of dead skin) to allow for the passive diffusion of glucose from interstitial fluid into the lumens of the MicroTips. MACs are secured onto a glucose diffusion chamber, in which the ArKal buffer is added to prompt glucose diffusion through the MACs and across the chamber for sensor reading. The glucose sensor, attached to the chamber, measures glucose via immobilized glucose oxidase on an electrode. Glucose values were determined by a prospective algorithm, previously “trained” on data from individuals without diabetes.

    • The ArKal buffer was validated in glucose flux experiments. MACS were fixed onto sampling cells (with no sensor) to determine: 1) whether the diffusion of glucose through the MicroTip lumens was proportional to blood glucose levels; 2) whether lumens occluded during use; and 3) whether the buffer formulation provided superior glucose correlation (i.e., diffused glucose amounts better corresponded to blood glucose amounts) compared to a saline buffer solution (control). The investigators concluded that blood glucose correlated with glucose flux over three days of continuous wear and that ArKal’s buffer resulted in higher correlations between blood glucose and glucose flux compared to control solution, likely because occlusion in the lumens only occurred with control solution use.

  • The study assessed sensor accuracy in 10 adult subjects with type 1 or type 2 diabetes. Each individual wore four devices on the upper arm or forearms; six participants wore the CGM for 48 hours and four participants wore the CGM for 72 hours. The CGM was calibrated against a once daily fingerstick (following a two-hour warm-up period upon initial application). CGM values were compared to fingerstick measurements. Fingersticks were taken every 20 minutes during the day and included periods of hypoglycemic and hyperglycemic glucose excursions to result in 1,396 paired data points from 37 sensors, representing a 7.5% sensor failure rate during the study.

  • Overall MARD was 15% and m ARD was 11%; Clark Error Grid analysis showed 75% of readings in zone A. Ninety-eight percent of readings fell in zone A + B.



Clack Error Grid Analysis



























Percent A + B


Technologies for Metabolic Monitoring


Wayne Menzie (Director of Technology and Clinical Development, Echo Therapeutics, Philadelphia, PA)

In a product-focused presentation, Mr. Wayne Menzie reviewed the design specifics of Echo Therapeutics’ hospital transdermal continuous glucose monitor, the Symphony tCGM system. He began by contrasting a transdermal approach versus an intravenous approach, stating, “Every CGM is an exercise in tradeoff.” While a transdermal system is noninvasive, the lag time (due to measuring interstitial fluid [ISF]) and resulting lower accuracy are drawbacks. An intravenous system has improved rapidity and accuracy; however, the need for an additional line introduces the chance of infection. Further, Mr. Menzie remarked that most intravenous systems measure blood glucose at longer intervals (every 15 minutes) than to transdermal (every minute). Delving into the specifics of the Symphony, the sensor, which sits under the transmitter, contains a three-electrode system and a hydrogel surface (with immobilized glucose oxidase) that serves as the skin interface. To gain transdermal access to the ISF, the system uses Prelude SkinPrep to remove the stratum corneum (outer layer of dead skin) allowing ISF to passively diffuse into the sensor. In order to monitor the skin preparation process and ensure that ISF can reach the sensor, Echo measures transepidermal water loss (TEWL). Normal skin, explained Mr. Menzie, has water loss between five and ten g/m2 per hour. When the stratum corneum is removed, TEWL should be in the range of 10-100 g/m2 per hour (Echo targets 20-40 g/m2 per hour). While Mr. Menzie noted that skin characteristics vary by age, he did not believe this posed a problem for the device. Mr. Menzie closed with a brief review of recent Symphony studies (see below) and a look towards the future. Echo will be conducting additional clinical trials, scaling up manufacturing, submitting for CE Mark in mid-2013, and working towards developing the system for home use.

  • Mr. Menzie briefly presented the high level findings from recent clinical trials.​ Across six studies (n=82 patients with mixed disease status) the Symphony had an average MARD of 11.8%. While the average MARD approaches that of OptiScan’s OptiScanner (10.9%), which is impressive considering the Symphony is transdermal (the OptiScanner is intravenous), we note that the range of glucose values included in a study can greatly affect how challenged a system is. For example, the second critical care Symphony study did not include hypoglycemic points, where CGM tends to be the least accurate. For our discussion of Echo’s ICU trials, please see our Echo Therapeutics 2Q12 report at http://www.closeconcerns.com/knowledgebase/r/5e66c830.




Patient Type



Data Points



A + B (Clark Error Grid Analysis)



Type 1 &

Type 2











Type 1 &

Type 2

















Type 1 &

Type 2










Critical Care






Critical Care









Avg: 11.8%

Avg: 98.3%


James Causey, BS (OptiScan Biomedical, Inc., Hayward, CA)

Reviewing the potential benefits of improving glycemic control and free up nursing time in the intensive care unit (ICU), James Causey described the specifications and performance of OptiScan’s OptiScanner CGM. The OptiScanner uses an optical sensor that can to any standard venous line. Plasma glucose is measured every 15 minutes using mid-infrared spectroscopy, and no calibration is required throughout the 72 consecutive hours that each sensor can be in place. Demonstrating the accuracy that OptiScan believes to be the highest achieved for a continuous glucose monitor, Mr. Causey presented results from the CONTROL I study in people with type 1 diabetes (n=1,155, Clarke A zone 98.8%), the CONTROL II study in people with type 1 diabetes (n=832, Clarke A zone), and the MANAGE I study in ICU patients (n=65 patients, Clarke A zone = 94.2%). The OptiScanner received CE mark in September 2011 and is launching in Italy (“dozens of orders” have been placed); on the home front, OptiScan is preparing for its device’s US pivotal trial.

  • Mr. Causey reviewed the specifications of the OptiScanner, which can connect to any standard venous line (central venous catheter (CVC), peripherally inserted central catheter (PICC), or peripheral intravenous). Every 15 minutes the device’s fluidic system extracts ~0.1 ml of blood, spins it in a centrifuge, and measures the plasma glucose concentration using a miniature mid-infrared spectrometer; during the flush process no heparin is returned to the patient. The color, touchscreen interface displays current and recent glucose values as well as trending and time in range. The console is topped with two poles to hold IV bags so that it adds minimal space at the bedside, and the entire unit is mobile, with a battery life of four hours so that it can travel when a patient who needs to be moved.

  • The OptiScanner does not require any calibration; rather, each spectroscopic measurement is passed through a series of filters to account for potential interferents. OptiScan’s filtering algorithm currently includes over 2,000 known spectroscopic signatures and can be updated to include more. (As we understand it, any potential modifications are validated against OptiScan’s database of all past results to ensure that the new filter makes all these past results either identical or more accurate.) Mr. Causey emphasized that eliminating the requirement for calibration saved both time and, more importantly, the potential for user or mechanical error. An accompanying slide quoted Boston Children’s Hospital’s Dr. Michael Agus’ comments from the FDA’s June 25, 2012 Public Meeting on Hospital Glucose Sensors: “If we’re going to be doing calibration on a (bleepy) blood value then you’re (bleeping) the patient for the next 12 hours” [sic].

  • Mr. Causey summarized two clinical trials comparing the OptiScanner to YSI reference values in patients with type 1 diabetes (glucose range <40-600 mg/dl in each). In the CONTROL I study (n=1,155 data pairs), Clarke Error Grid Analysis put 98.8% of data pairs in the A zone and 1.2% in the B zone. Coefficient of variation was 6.43% (Jax et al., J Diabetes Sci Technol 2011). Blood clots occurred in the lines of two patients, but Mr. Causey cited the study investigator as saying that clots are relatively normal in glucose clamp studies. In the CONTROL II study, which used an updated algorithm, patients were maintained under relatively tight glycemic control for the first six-to-nine hours of the study (during which reference measurements were taken every 15 minutes) and managed their own glycemic control for the rest of the study (measurements taken every two hours for the rest of the study). No measurements were excluded, and Clarke Error Grid analysis of the data pairs (n=832) placed 99.7% in the A zone and 0.3% in the B zone.

  • During the nine-month MANAGE I trial, the OptiScanner was tested in over 60 ICU patients who were – to Mr. Causeys’ knowledge – the sickest population ever enrolled in CGM research. Over 90% of these patients were on mechanical ventilation, 78% used vasoactive drugs, and 8% had diabetes. More than one fifth of the patients died while in the hospital, and 16% died in the ICU. Mr. Causey explained that for the first three months of the study, accuracy was affected by an interferent that had not previously been included in the OptiScanner’s filtering algorithms. However, after three months OptiScan’s engineers had implemented a new filtering algorithm that dramatically improved the data (retrospectively and going forward). With these updated algorithms, the accuracy was quite high: Clarke Error Grid Analysis placed 94.2% of data pairs in the A zone and 5.8% in the B zone. Coefficient of variation was 9.4%, and bias was -0.6%. (Note: these results were originally presented at ISICEM 2012).


Yi Lu, PhD (University of Illinois at Urbana-Champaign, Urbana, IL)

Dr. Yi Lu delved into the physiologic mechanism behind GlucoSentient, a sensing technology for point-of-care (POC) meters designed to detect multiple analytes relevant to a patient with diabetes. Dr. Lu emphasized that given the complexity of diabetes, there are numerous biomarkers that require monitoring; however, within the POC field, each class of targets requires its own dedicated device (impressively blood glucose meters account for 80% of the $20 billion/year global POC market). To this end, Dr. Lu and his colleagues worked to redesign blood glucose meter strips to make the meter more versatile and extensible, with the objective of increasing patient access to testing for other markers and decreasing cost of testing. His team faced two major challenges: 1) converting other markers into glucose without modifying the meter and 2) converting low concentrations of markers into relatively high concentrations of glucose. By using an invertase enzyme to hydrolyze sucrose into glucose (which includes a conjugation step that only occurs in the presence of the target enzyme) and by using enzymatic turnovers to amplify signals, Dr. Lu believes GlucoSentient has overcome both hurdles. By employing antibodies, aptamers, or DNA in combination with invertase, he explained that the technology could detect cancers and toxins, Hepatitis B DNA, and drugs, respectively. Currently, the strip is an “extension” that is connected to meters’ strips in order to obtain a blood sample. Because different meters use different strips, creating an extension allowed Dr. Lu’s team to use GlucoSentient across multiple meter platforms. However, in the future Dr. Lu hopes to work with manufacturers to integrate multiple analyte testing and blood glucose testing into a single strip. We note, however, as raised during the Q&A discussion, that there are questions concerning the clinical relevance of testing other biomarkers as frequently as blood glucose/A1c. To this end, Dr. Lu suggested that GlucoSentient could also be used in the research environment to validate whether certain biomarkers did indeed need to be tested with such frequency.



Rich Williams, MD (NASA Headquarters, Washington, DC), Ananda Basu, MD, FRCP, Mayo College of Medicine (Mayo Clinic, Rochester, MN), Yi Lu, PhD (University of Illinois at Urbana-Champaign, Urbana, IL), James Causey, BS (OptiScan Biomedical, Inc., Hayward, CA), Wayne Menzie, BS (Echo Therapeutics, Inc., Franklin, MA)

Dr. DeVries: Mr. Menzie, you used the word “generally painless.” You have 82 patients, did any drop out because he or she said “generally painless” is too painless?

Mr. Menzie: We did not have any patients who dropped out because of that. The reason I said that is because you do feel it; there is a sensation.

Dr. DeVries: Would you compare it to electrical shaving?

Mr. Menzie: Yes, it’s pretty similar.

Q: This is for Mr. Causey. You spoke about a variability alarm. Can you expand a little on that? What measure do you use to trigger the alarm and what do you do if it sounds?

Mr. Causey: It is not implemented in current version; it’s still in the R&D phase. We use rate of change versus unit time detection.

Q: Dr. Lu, not every patient or institution has the same target. How do you choose which to monitor?

Dr. Lu: All we have is a tool to monitor analytes. We wouldn’t validate each analyte ourselves; we just provide the tools for others to do so.

Q: Mr. Menzie, did you look at the lag time compared to venous glucose for your transdermal system?

Mr. Menzie: Yes, we did. We found lag time was consistent with what was reported with subcutaneous CGM, which is on the order of 10 minutes. It was very comparable to lag time of other CGMs.

Q: Dr. Lu, glucometers are basically potentiostats. All of the assays you mentioned could be performed electrochemically. Why not just test each analyte straightaway, rather than converting to glucose?

Dr. Lu: A lot of people have been developing devices to measure each device, but this requires much time and effort. With our approach we bypass the manufacturing and development costs, and we accelerate validation since glucose meters are well characterized. We are not saying that ours is the only way, but we think it is a good way. That said, we are not married to a particular modality of glucose measurement. As long as the input is glucose, we can use it.

Q: A glucose meter is actually measuring hydrogen peroxide; you could just convert A1c to hydrogen peroxide. I don’t think that developing the chemistry to do this directly is all that difficult. I like your idea, but I think that your engineering challenges are equivalent to those of designing electrochemical strips specific to each analyte.

Dr. Lu: I agree that there are additional challenges for us. But our chemistry is done before the meter’s hydrogen peroxide measurement, so we can correct for interferences.

Dr. Basu: What you are saying is that you use the glucose meter with a different strip for each analyte?

Dr. Lu: Our strip is coupled to whatever strip the manufacturer uses. In the end, we want to integrate our strip with the manufactures’. You cannot interchange strips with one another, so right now our output is glucose, but eventually we’d like to work with manufactures of meter so we can integrate everything into one strip.

Mr. Menzie: So you would integrate multiple analytes into a single strip?

Dr. Lu: That’s the future direction. Right now it’s meter limited, we only do one at a time because we don’t want to become a manufacturer of meters.

Peter Picton (University of Toronto, Toronto, Canada): Wayne, you have some great technology with great potential. You said that you are working toward further development for home use; what do you see as the technology hurdles here?

Mr. Menzie: I think it is really just a question of focusing our small company’s resources – the effort and the money to develop a home-use monitor, since the requirements in each setting are different. But our preliminary type 1 diabetes studies have shown that the sensing technology performs well even in wider glucose ranges than seen in the ICU.

Dr. Basu: Mr. Menzie, do you monitor how perspiration or fever affects the patient?

Mr. Menzie: I know we measure temperature at the site, so I wouldn’t expect fever to cause any problems. We have induced sweating in patients and we have not seen an impact on performance of the device.

Dr. Williams: Can you comment on the shelf life of your models?

Mr. Menzie: At a minimum, we are looking at 18-month shelf life.

Mr. Causey: We have no reagents.

Dr. Lu: We are in the R&D stage. A limitation for our reagent’s shelf life is the antibody use – the most vulnerable part of our whole reagent is the antibody component.

Dr. Lu: Mr. Menzie, you showed only a minor indentation on the skin after 24 hours of sensor wear. When you place the next sensor, do you put it on the same site?

Mr. Menzie: We have to cycle through different sites; each site needs to heal before it can be re-used. Generally you need five-to-seven days before the site can be re-used. This isn’t because it isn’t amenable to measurement but because the impedance measurements during the skin preparation process are affected.

Q: Dr. Lu, you are using blood glucose meters to measure analytes, but have you looked from the clinical perspective as to whether you need to measure these on regular basis?

Dr. Lu: We are just beginning to do that – maybe therapeutic drugs need monitoring on a daily basis and also CV markers may need to be monitored more frequently. We are working with experts. Hopefully we can use to it to work with people in the audience to go look and search and validate those markers.

Q: Will you have to take into account the patient’s blood sugar before performing the measurement?

Dr. Lu: Yes, we have to perform two tests – a first one for glucose, and then a second one for the assay of interest. This is why we need novel apps either on phones or built into the meters themselves, so that the patient doesn’t have to enter the glucose number manually.


Continuous Glucose Monitoring


John Mastrototaro, PhD (VP, Research and Technology, Medtronic Diabetes)

The highly-regarded Dr. John Mastrototaro discussed the present and future of Medtronic’s research on closing the loop of glucose control. The first step is Low Threshold Suspend (LTS) functionality, which has been implemented in the sensor-augmented pump products Veo (available in Europe) and the MiniMed 530G (under FDA review). Dr. Mastrototaro interactively reviewed data from the ASPIRE I study of exercise-induced hypoglycemia, in which the use of LTS significantly shortened hypoglycemia duration. The next step beyond LTS will be the Predictive Low Glucose Module (PLGM), which cuts off insulin delivery for 30-120 minutes when sensor values are predicted to be heading below 70 mg/dl. After that will come closed-loop overnight control, with or without semi-automated treat-to-target insulin delivery during the day. (Dr. Mastrototaro alluded to many failsafe algorithms in the overnight system “that we are excited about,” and he looked forward to reporting data on these algorithms in 2013.) To reach the ultimate goal – a fully automated artificial pancreas – Dr. Mastrototaro believes that redundant glucose sensors will be needed. To this end Medtronic is studying a combination of optical and electroenzymatic sensing.

  • As the answer to one of his audience-response questions, Dr. Mastrototaro said that the Enlite’s sensitivity is only 75% for detecting glucose values that are truly <70 mg/dl. Sensitivity can be improved by raising the threshold (as we understand it, this is because sensors are more accurate closer to normoglycemia). However, setting a higher threshold might lead to undesirably frequent suspension. An alternative way to improve sensitivity is by incorporating predictive algorithms, as is being done with Medtronic’s investigational predictive low glucose module (PLGM). The combined use of threshold and predictive alerts, both set at 70 mg/dl, increases sensitivity to 92.4%.


Sean Doherty, PhD (Wilfrid Laurier University, Waterloo, Canada)

Dr. Doherty gave a unique, well-received presentation on combining CGM data with geographic information (i.e., GPS data collected from a smartphone). He conducted a study in 40 patients in Toronto with diabetes that concurrently used a smartphone (GPS and accelerometer), the Medtronic MiniMed CGMS, a heart rate monitor, and a food/activity diary (the one manual tool) over three days. Dr. Doherty cleverly displayed colored time-in-range pie charts at various geographic locations on a map (e.g., things are under control at home, but glucose levels seems to be very high at work and very low at the gym) – we thought this was a unique way of looking at the data. Using time series models individualized to each patient, Dr. Doherty also tried to predict glucose averages against many other factors (e.g., exercise in last five minutes, automobile time in last five minutes, previous blood glucose readings, etc.). He had a number of interesting conclusions: 1) patients were successfully managing the effects of food and exercise, but they were leaving out other factors; 2) every patient was completely different (“advice must be personalized”); 3) data is not very expensive to get; and 4) maps provide a different view of the data and can be used as a complementary tool to stimulate discussion with patients.

  • As CGM continues to be adopted, we suspect new ways of looking at and combining the data will continue to emerge. Furthermore, as glucose monitoring data moves to the smart phone and the cloud (e.g., Sanofi’s iBGStar, LifeScan’s OneTouch VerioSync, Telcare’s meter, Dexcom’s Gen 5, Senseonics’ implantable sensor, C8 MediSensors), we suspect it will be easier to combine geographic and glycemic data sources.


Geoff Chase, PhD (University of Canterbury, Christchurch, New Zealand)

Dr. Geoff Chase presented preliminary results from five patients participating in an observational study assessing whether CGM could be used in conjunction with STAR, a computer based, tight glycemic control protocol for the ICU. The study’s primary aim was to assess the reliability of CGMs in ICU patients. Two CGMs were tested in the study: Medtronic’s iPro2 CGM and Medtronic’s Guardian Real- Time CGM; each patient wore three sensors (an iPro2 and Guardian on either side of the abdomen, with another iPro2 on the thigh). The CGMs were calibrated using arterial blood gas glucose measurements (radiometer ABL90 Flex). Comparing the abdominally placed CGMs, the iPro2 performed better than the Guardian, likely due to the retrospective versus real-time calibration, respectively. Regarding the between site comparisons, the iPro2 located on the thigh tended to report lower glucose readings than the iPro2 located on the abdomen. Dr. Chase hypothesized that edema (which tends to be greater in the abdomen) or movement (which tends to be greater in the thigh) could affect CGM readings and lead to inter-site variability. The secondary aim of the study was to assess whether “off the shelf glucose meters” were accurate enough for ICU use. The monitor assessed in the study was Abbott’s Optium Xceed; each time blood was drawn from the arterial line for blood gas analysis, Optium Xceed also used blood samples to measure glucose. Dr. Chase suggested that thus far, glucose meter readings have been sufficiently accurate to use in the STAR protocol. The study is ongoing and Dr. Chase hopes to enroll additional patients, as he has received ethics approval for up to 50 participants.

  • Dr. Chase opened his presentation with a series of audience response questions to prime the audience for his discussion.

    • Does edema “matter” and will it affect CGM performance in ICU patients?

      • Yes: 62%

      • No: 12%

      • Maybe: 15%

      • I’m too afraid to answer: 12%

    • Does sensor location matter in an ICU patient?

      • Yes: 77%

      • No: 4%

      • I’d guess yes, but don’t know why (polite maybe): 13%

      • Is it dinner yet? (polite no): 7%

    • Are glucometers that use blood from a central line a good measure of blood glucose in ICU patients? (i.e. good enough to be used in control)

      • Yes: 37%

      • Maybe: 38%

      • Never! 18%

      • Glucometer? 7%

      • Where is CGM technology currently at for the ICU?

        • It doesn’t work, stop wasting your time: 2%

        • A useful monitoring tool, but that’s it: 52%

        • Good enough to make clinical decisions using CGM data: 40%

        • Close the loop already, they are perfect: 6%


Steven Russell, MD, PhD (MGH Diabetes Associates, Boston, MA), John Mastrototaro, PhD (VP, Research and Technology, Medtronic Diabetes), Geoff Chase, PhD (University of Canterbury, Christchurch, New Zealand), Sean Doherty, PhD (Wilfrid Laurier University, Waterloo, Canada)Dr. Russell: It might be appropriate to have contextual alarms for CGM. For example, driving is a good time not to be low and you may want to have a more sensitive threshed at that time?

Dr. Doherty: Driving is an unusual situation. I think I saw it in multiple patients that it’s a dangerous situation to be in. We see different patterns in each patient, so I wouldn’t automate anything, but we need to improve the situation.

Dr. Gary Steil (Children’s Hospital, Boston, MA): At Children’s Hospital, we just published a trial in the New England Journal of Medicine. We gave insulin with a CGM and a titration algorithm. We had a lower hypoglycemia rate than the background rate in the ICU. That data address the question of whether you can use CGM in the ICU. The answer is very definitely yes. This was in just under 1,000 kids, some younger than three years old. That was the Medtronic Guardian Real-Time. Now, Medtronic has moved on from the Guardian to the Enlite. I’m bothered by seeing the sensor tracings of the effect of edema. We certainly didn’t have that observation. I don’t think we saw that at all.

Dr. Chase: In patients with severe edema, about one-fourth can go a bit wonky. With about three quarters, it works as planned. In edema patients, there is a significant inflammatory response. We are reporting observational data. Our colleagues in Europe have seen a similar thing. The exact cause is undetermined.

Q: John, I know the ASPIRE I study is published. But just to clarify, with regard to exercise, what was the relationship of the last meal intake to the start of exercise and was there any decrease in insulin at the start of exercise – how much was the basal decreased by?

Dr. Mastrototaro: It was without any meal preceding it, so it had been several hours. We did not touch the preprogrammed basal rate in the system. Patients would exercise slowly in 15-minute intervals until they hit 85 mg/dl and were falling. Then we would watch as their blood glucose continued to drift down below 70 mg/dl.

Q: Did you look at Vo2 max target or just glucose drop?

Dr. Mastrototaro: Just glucose drop. The goal was to cause a hypoglycemic event and look at two situations – when the threshold was on and when it was off.

Q: As for the redundant sensing strategy, is there a timeline to when the product will be ready?

Dr. Mastrototaro: We received a JDRF grant in June to help support a study of redundant sensing. But we had acquired an optical sensor as well a while back. So JDRF is helping accelerate the mating up of electroenzymatic and optical sensing. We think this will provide us a more robust measure of glucose. The other thing, even with the existing electroenzymatic sensor, we are developing redundant sensing within that sensor.

Dr. Riccardo Bellazi (University of Pavia, Pavia, Italy): Your presentation was really exciting. I think that it starts the idea of using big data – data coming from heterogeneous sources. It’s new ways to combine the data and provide support and decision-making. It’s very important and exciting. You used the time series analysis – an auto-regressive moving average integrated model. That’s a good idea. A regression cannot work for sure since the data changes over time – it’s not stationary. In the time series you used, some of data were probably qualitative. Others were real signals. I’d like to know how you pre-processed the data. Did you take raw data from the accelerometer and code it yes or no?

Dr. Doherty: With accelerometer, we used the basic Vmag, a combination of all three axes in a five-minute period. We took the average over the previous five-minute period. Time use is the amount of time in the last five minutes spent driving or walking. Those were technically continuous variables. They were derived from GPS and accelerometer data and get confirmed through the diary. The only yes/no variables were day and night and whether they took medications or not.

Q: Dr. Doherty, you put up a slide showing lags. What are the intervals for the lags?

Dr. Doherty: Five minute lags.

Q: Do you have a sense of the quality of nutrition data as it came in?

Dr. Doherty: I took a look at total calories they reported and the number of meals and snacks. Of all the methods, I trust the food diaries the most. They also used a digital pen that sent what they were writing to my server right away. Sometimes they left out time and only occasionally did they forget the amount.

Q: I have found food photos to work well. It’s labor intensive, but good data. How did you judge whether people were around others?

Dr. Doherty: It was from the online diary. There was a column asking who was with them at the time.

Dr. Roman Hovorka (University of Cambridge, Cambridge, UK): Fantastic original research on tracking. What about the privacy issues and ethics committees? Were there issues around the study for following people?

Dr. Doherty: Privacy is big. I consulted someone in Ontario before the study. In the particular maps I showed, I used some masking techniques – I had to hide things like where people lived. In a study, we’re less concerned about privacy because we have consent forms and ethics approval. We can take some risks. This is less invasive than some other things I’ve done – for instance, someone wearing a video camera on their body pointing outwards and filming all day. Talk about getting ethics approval for that. I was using a smartphone with a lot of security. It has world-class leading encryption and security. In the long term, we must protect this data. I did a market research study and asked people, “Would you share this data with caregivers if there was a good reason for it?” Over 90% said they would. People are more than anxious to share this data. I don’t think this will be as big of an issue.

Dr. Hovorka: Dr. Mastrototaro, we know little of the predictive low glucose suspend studies being done. What are you planning to do to get approval?

Dr. Mastrototaro: We’re talking to FDA about a pivotal trial for the predictive system now. We are looking at running some studies outside of the US and inside of the US and we are looking to do experiments somewhat similar to ASPIRE, but use two forms glucose lowering: exercise and increased basal rates. It has not been finalized.

Dr. Hovorka: I’m talking about EU approval.

Dr. Mastrototaro: We are going to use something similar to what we’ve done with ASPIRE I and collect data to demonstrate how effective it is at minimizing the lows.

Oliver Chan (Senseonics, Germantown, MD): A question for Medtronic on prediction. Due to CGM lag, many algorithms would result in a delay if applied in real time. Did you observe any delay? How long of a delay is there for a 30-minute prediction?

Dr. Mastrototaro: We’re trying to have an algorithm predict 30 minutes in advance. But in reality, it’s 20 minutes-ish in advance. We have also have another predictor that we cannot comment on yet. One of the fortunate things is the wealth of data from CGM and known YSI glucose values. We know true low glucose events. We can do a lot of retrospective algorithm testing. We did millions of computations of simulated patients using a variety of algorithms for detection. We had a high detection rate goal. We didn’t want to have a lot of false alarms. For a time horizon further out, it’s more likely you are going to be getting a lot of nuisance alerts. There is a tradeoff there.


New Metrics to Quantify Glycemic Variability


Thomas Peyser, PhD (VP, Science and Technology, Dexcom)

In an elegant talk, Dr. Thomas Peyser presented two unconventional measurements of glycemic variability designed to be simple to understand, describe, and calculate. The first, glycemic variability index (GVI), is simply a measure of the relative length of the line that would be shown in a CGM trace. For example, someone with a perfectly flat CGM trace (never found in nature) would have a GVI of 1.00, and someone with moderate glucose variability would have a GVI of 1.2-1.5. The other metric, patient glycemic status (PGS), is designed to capture both variability and overall glycemic control; it is the multiplicative product of GVI, mean sensor glucose, and percent time spent outside of target range. Dr. Peyser proposed that with further study, GVI and PGS might be found useful as a clinical flag for when patients need help -- a method readily understandable even for non-specialists (an important concern as CGM becomes more widely used outside of endocrinology practices). At the other end of the spectrum, he suggested that the metrics could be used to characterize how well different drugs reduce GV and improve overall glycemic control.

  • Dr. Peyser half-jokingly asked whether anyone in the audience would feel comfortable publicly writing out the equations for glucose standard deviation, MAGE, and CONGA. (No one raised a hand.) He noted that most metrics for glycemicvariability are mathematically complicated, whereas most people involved in diabetes seem to feel about glycemic variability the way that Supreme Court Justice Potter Stewart described hard-core pornography: “I know it when I see it.” Glycemic variability has been found to be associated with lower mood and quality of life (Penckofer et al., Diab Technol Ther 2012); Dexcom’s perspective is that wearing CGM should reduce GV (and that patients who cannot reduce their GV using CGM tend to be dissatisfied with the technology).

  • As a relatively simple and intuitive way to quantify glycemic variability, Dr. Peyser proposed adding up the diagonal distance between each pair of consecutive sensor values and then dividing this by the horizontal distance to standardize for any given amount of time. He called this method glucose variability index (GVI) and listed some example values: 1.0 would be a perfectly flat line (never seen in nature), 1.0-1.2 would be found in people without diabetes, 1.2-1.5 would indicate modest variability, and >1.5 would correspond to high GV. Dr. Peyser said that GVI correlates well with MAGE, though it does not penalize single large glucose spikes as much as MAGE does.

  • For a basic topological measurement of overall glycemic control, Dr. Peyser proposed that GVI could be multiplied by mean sensor glucose and the percentage of time spent outside glycemic range: i.e., GVI x MG x (1-PTIR), where MG = mean glucose and PTIR = percent time in range. The resulting method, patient glycemic status (PGS), is conceptually similar to area under the curve but somewhat more precise in the contributions of each aspect of glucose control, Dr. Peyser said. PGS would be at or below 35 for most people without diabetes, people with diabetes and good glycemic status might have PGS 35-100, 100-150 would be a poor PGS, and >150 would indicate “very poor” glycemia. Dr. Peyser said that retrospective data haven’t yet been analyzed to compare PGS and A1c but that PGS > 150 would probably correlate with an A1c above 9.0%.

  • Dr. Peyser acknowledged that GVI and PGS were similar to metrics proposed by Drs. Cynthia Marling, Jay Shubrook, and Frank Schwartz and colleagues at Ohio University (Marling et al., J Diabetes Sci Tech 2011).


Lane Desborough, MSc, PEng (Product Strategist, Medtronic Diabetes)

Taking a relatively sophisticated approach to measuring glycemic control, Mr. Lane Desborough described a metric that accounts for the non-normal distribution of glucose values (i.e., the graph is not a perfect “bell curve” because hyperglycemic values extend so high and hypoglycemic values can go only so low). Rather than using the arithmetic mean and standard deviation (which would be appropriate for a normal distribution), Mr. Desborough prefers using the geometric mean and standard deviation (which correspond to a log-normal distribution). He then presented an equation that incorporated geometric mean, geometric standard deviation, and glycemic loss (which penalizes hypo- and hyperglycemic values based roughly on their clinical implications rather than just their arithmetic distance from a glycemic target). That equation, while perhaps too complex for many clinicians to understand at first glance, can be a useful metric for glycemic control in research applications. Indeed, the metric has been used by Medtronic’s artificial pancreas to evaluate over three million simulated patient-days and refine closed-loop control algorithms.



Francesco Valgimigli, PhD (Director, Technology and Scientific Affairs, A. Menarini Diagnostics)

Rather than discussing glycemic variability (GV) as a measure of retrospective glucose control as is typically done, Dr. Francesco Valgimigli proposed that real-time GV metrics could be useful for achieving tight glycemic control in the intensive care unit (ICU). In particular he suggested an inpatient CGM display with four real-time data traces. The first was simply sensor glucose. The second was a new metric called variability index, the multiplicative product of coefficient of variation (“an index of variability”) with M-value (“an index of control”), to give a fairly comprehensive picture of current glucose. The third and fourth metrics were high- and low blood glucose index (HBGI and LBGI), which serve as up-to-the-minute predictions of risk for hypo- and hyperglycemia, respectively. In closing he noted that the utility of these measurements will depend on the accuracy of the CGM sensor itself. To this end he looked forward to results from the ongoing pivotal trial of A. Menarini’s inpatient microdialysis sensor.



Holly Schachner, MD (Senior Medical Director, Diabetes Care, Bayer Health Care)

Reminding the audience that most people with diabetes did not attend Stanford (“or even MIT”), Dr. Holly Schachner advocated that glycemic variability metrics should be clear to patients who use self- monitoring of blood glucose (SMBG) and who do not download their data (roughly 80% of SMBG users do not download, she said). She presented screenshots from two generations of Contour blood glucose meters, both of which can display how many meter readings in the past week (or 14 or 28 days) were outside of the patient- and provider-specified target zone. Dr. Schachner said that patients can readily understand what these numbers mean, so that they “ultimately can manage their diabetes in a more meaningful way.” She also noted that this approach is relatively non-controversial: while the importance and definition of “glycemic variability” remain contested, nearly everyone agrees that both hypo- and hyperglycemia are important to avoid.



Guillermo Arreaza-Rubin, MD (National Institutes of Health, Bethesda, MD), Marc Breton, PhD (University of Virginia, Charlottesville, VA), Thomas Peyser, PhD (Dexcom, San Diego, CA), Lane Desborough, MSc, PEng (Medtronic, Northridge, CA), Francesco Valgimigli, PhD (A. Menarini Diagnostics, Florence, Italy), Holly Schachner, MD (Bayer, Tarrytown, NY)

Dr. Breton: I loved the measure of using the measure of line length. But I am concerned that this method might over-emphasize hyperglycemic excursions, which have longer lines than hypoglycemic dips. Could Dr. Peyser and Mr. Desborough discuss the utility of log- transformation?

Dr. Peyser: That is a good point. Glucose is not normally distributed, and there are different consequences for being low and high. I think that this would be a good idea to consider [the logarithmic adjustments that Mr. Desborough described] in improving the GVI.

Mr. Desborough: We are not all math or statistics majors, and many of us see plus-or-minus-one standard deviation and think that this captures 68% of all the data. It is unfortunate that David Rodbard isn’t here and could discuss even how we visualize data: he would say that too much of the screen on CGM is devoted to hyperglycemia relative to hypoglycemia.

John Mastrototaro (VP, Research and Technology, Medtronic): If the sensor signal is noisy, that will add increased length of line even if the actual glucose values don’t have much variability.

Dr. Peyser: This metric was based on the latest Dexcom CGM, which has fairly low noise. I would have to look at datasets with earlier CGMs to see how noise affects these metrics; I think that such artifacts highlight the need for improved CGM.

Dr. Natalie Wisniewski (Medical Device Consultancy, San Francisco, CA): I think there are high hopes for promise of glycemic variability, but still there is much thinking to be done. For Dr. Valgimigli, I have a multi-part question about the micro-dialysis system. Is it inserted intravascularly? Is it safe for home use? Do you see a difference in GV at home vs. in the clinic? Do you see a change in GV over the duration of insertion?

Dr. Valgimigli: At Menarini, we developed a platform that can be used both for home use and for intravascular applications in the hospital. The two uses have different purposes. Interstitial monitoring has some limitations in terms of lag time between blood glucose and the subcutaneous compartment. Applying this technology in the hospital intravascularly enables overcoming many traditional limitations of CGM. We are conducting clinical trials now; clearly we overcome issues of lag time. We can achieve a level of accuracy that can be hopefully comparable to blood gas analyzers but without the problem of intermittent measurement. For patients at home, you can evaluate glycemic control in the long term; in the hospital you need to address short-term control. I am not sure if this addressed your question.

Dr. Horacio Rilo (University of Arizona, Tucson, AZ): For safety, we have been thinking of using CGM in the ICU for patients getting islet transplants. There is a concept that if you maintain glycemic control in the perioperative period, you get better results with the transplant. If I were going to use the wider target that you described, it would be bad for the islets. Do you think it is possible to have tighter control without the risk for hypo- or hyperglycemia?

Dr. Valgimigli: Of course the issue of tight glycemic control in the ICU is very debated worldwide. At the moment the risk to perform too-tight glycemic control and provoke hypoglycemia is high. In my view, it is possible to safely attain tight control, but this depends on the accuracy of the CGM used. The higher the accuracy, the lower the risk of making the wrong decision in insulin treatment. My answer is “yes,” we can achieve tight glycemic control in a safe way in the future.

Dr. Rilo: Mr. Desborough, I have done initial studies with the iPro 1 and then switched to iPro 2. My understanding is that they differ in the timing of monitoring. Do you see differences in these two devices?

Mr. Desborough: I am not familiar with the differences between the two iPros, but I can point you to some people in the audience who are.

Dr. Rilo: Do you see an advantage of one vs. the other?

Mr. Desborough: I would use the more recent one.

Dr. David Simmons (Chief Medical Officer, Bayer Diabetes Care): I love the notion of the length of the line. To me A1c is useful for time-weighted area under the curve, but the length of the line seems a good next analysis. To have clinical utility, you might have to show that the measure of variability can travel distinctly from the A1c – otherwise what value does it add? You also need to show that it tracks with known markers of better control. To the earlier point about noise, it may not matter what the CGM changes are, exactly, if you could draw a sort of envelope line around the CGM trace and measure that.

Dr. Peyser: I think that if we want to use GVI we would indeed need to assess larger datasets. The goal was not just simplicity but to give clinicians a tool. With regard to noise, the truth is that all the CGMs are getting better. Certainly this is true of the new Dexcom product; I am sure it is true as well of the new Medtronic sensor.

Mr. Desborough: There seems to be consensus that both mean and standard deviation contribute to outcomes. As a control engineer we would first address the dispersion and then try to move the mean. I don’t think that it is “and/or”; I think that it is “both.” Having a small variability with a mean of 400 mg/dl is not safe either.

Dr. Arreaza-Rubin: How do you decide which metrics are more useful for clinical practice relative to research purposes – for example, for artificial pancreas design?

Mr. Desborough: I was at an artificial pancreas meeting a few years ago when Dr. Aaron Kowalski said that time in range is the best measurement. Initially I disputed this, but having looked at more data I agree with him. It is very important to look at the fraction of time spent in range and, when not in range, to know where you are.


Glycemic Variability: Clinical Applications


Fergus Cameron, MD (Royal Children’s Hospital-Parkville, Melbourne, Australia)

Dr. Fergus Cameron noted that although metrics for glycemic variability (GV) have been around since 1965, we still lack consensus on how much any of these metrics matters in a clinical setting (let alone which is best). Drawing on a review by Nalysnyk and colleagues (Diabetes Obes Metab 2010), he said that mounting evidence seems to link GV in type 2 diabetes to micro- and macrovascular complications, but the same relationship has not been shown in type 1 diabetes. However, one reason for this lack of evidence in type 1 diabetes may be that the data points were too far apart from each other or that the sampling period was not long enough. Dr. Cameron’s group has conducted small retrospective analyses of CGM data in children with type 1 diabetes (n=20-48) in which the accuracy of standard deviation, CONGA, and MAGE would have declined dramatically if glucose had been taken less frequently than every one-to-two hours and/or if data had been collected for fewer than six consecutive days. All eight of the type 1 diabetes studies in the Nalysnyk review failed to meet these thresholds, which might mean that the world is still waiting for meaningful data on GV and outcomes in type 1 diabetes – in our view only time (and long-term studies with CGM) will tell.



Horacio Rilo, MD (University of Arizona, Tucson, AZ)

A specialist in islet transplantation, Dr. Horacio Rilo explained that post-transplant outcomes are often quite different even in patients with similar, low A1c values. By comparison, metrics based on CGM data (e.g., mean glucose and time spent in hypo- and hyperglycemia) seem to be more useful predictors of whether islet transplant recipients will reach insulin independence or more modest glycemic benefits. Dr. Rilo suggested that CGM can therefore be a good tool for patients and clinicians to make accurate forecasts after surgery. During subsequent discussion he also suggested that CGM could be useful to maintain tight glycemic control in the period immediately before, during, and after surgery; this would likely prolong the survival of the newly transplanted islets.



James Krinsley, MD (Stamford Hospital, Stamford, CT)

Unfortunately, Dr. James Krinsley was unable to attend the session, and no presentation was given on glycemic variability in the hospital. For a recent review of the topic by Dr. Robert Rushakoff (University of California, San Francisco, San Francisco, CA), please see page 6 of our International Hospital Diabetes Meeting 2012 coverage at https://closeconcerns.box.com/s/6no7s4mazxvtptfjj7av.


Stephanie Fonda, PhD (Walter Reed Army Medical Center, Washington, DC)

In this intriguing overview, Dr. Stephanie Fonda discussed evidence that glycemic variability (GV) in obese patients precedes the diagnosis of diabetes (and even the traditional diagnosis of prediabetes). She also reviewed data on interventions designed to reduce GV, most of which have been conducted in patients with type 2 diabetes. Significant, favorable glycemic effects on have been seen with relatively minor changes in diet (e.g., eating vegetables before carbohydrates rather than vice versa) and exercise (hyperglycemia was dramatically reduced with either an hour of exercise every other day or a half-hour every day). Dr. Fonda also presented a detailed analysis of the groundbreaking Walter Reed study of CGM in type 2 diabetes patients not using mealtime insulin (Vigersky et al., Diabetes Care 2012). Notably, nearly all of these patients appeared to benefit immediately from using real-time CGM. Dr. Fonda therefore hypothesized that for these patients, CGM was not so much an educational tool as a reminder of how much behavior affects glucose control.


Howard Zisser, MD (Sansum Diabetes Research Institute, Santa Barbara, CA), Judith Fradkin, MD (National Institutes of Health, Bethesda, MD), Fergus Cameron, MD (Royal Children’s Hospital-Parkville, Melbourne, Australia), Horacio Rilo, MD (University of Arizona, Tucson, AZ), Stephanie Fonda, PhD (Walter Reed Army Medical Center, Washington, DC)

Q: We should remember a patient-centered approach. The higher the GV, the lower the quality of life. This is perhaps one of the main reasons to measure GV and improve it. Secondly, I don’t know whether GV per se or free fatty acids, cytokines, etc. contribute to increased risk of complications. Finally, the activity in improving glycemic variability: we have a paper coming out in diabetes car that even low-grade physical activity of 1.7-2.2 METs significantly improves glycemic variability and postprandial glucose – something like washing dishes. You don’t have to run a marathon.

Dr. Darrell Wilson (Stanford University, Stanford, CA): Dr. Fonda, I was intrigued by your graphs of estimated average glucose. These seemed to be very far from what I would have guessed based on the means in the glucose profile. Why do you think this was?

Dr. Fonda: There were some things about the study that we would have done differently. The mean glucose, as of the first cycle of wearing CGM, was lower than we would expect based on A1c at enrollment. My hypothesis is that the act of wearing the CGM, from the very beginning, was self-reinforcing enough to encourage the behavioral modifications that we wanted. One might imagine that wearing CGM would be educational. But I think that this population already knew what to do, for the most part, and that this was more a reminder than an educational tool. Hence the results on glycemia as early as they were.

Dr. Robert Vigersky (Walter Reed Army Medical Center, Washington, DC): Let me reinforce that with an anecdote. Our so-called poster child for this study had an A1c in the mid-eights. A year-and-a-half later he came back to see me as a patient (trial subjects had not actually been our patients at the time of enrollment). He said, “I was in your study.” I asked him how he did; he said, “I am Italian and I love bread. But every time I ate bread I saw my sugar go sky-high, so I cut it out.” He lost over 30 pounds, and his A1c fell into the sixes. I think that he made this change early in the study. We didn’t have a blinded period to confirm that this was the case, but that is an anecdote to confirm what we thought was true.

Dr. Zisser: I always see the laundry list of ways to measure glycemic variability. Has anyone tried a compound metric?

Dr. Cameron: I think that papers are increasingly recording multiple measures. Some measures that incorporate GV and mean are being used; so far I don’t think one superior system has come to the fore. There may be things in play other than glucose variation. My interest is in neural regulation. Perhaps insulin levels or counterregulatory hormones are really what is important. It’s like all things: the more you understand it, the more complicated it becomes.

Dr. Fradkin: Dr. Rilo, I thought that the data on CGM as a proxy for islet number was very interesting. Did you measure beta-cell function in addition to number?

Dr. Rilo: I looked at digestion time and other factors to determine islet effective number. I find CGM fascinating. I embrace the technology. CGM has a predictive value. A lot of the patients that come from other programs or other parts of the country said that their primary care physician said A1c is normal and so they would have normal islet function after transplantation. I show these data to patients, and I think that [appropriately] decreases expectations of insulin independence.


Reference Methods for Assessing Performance of Blood Glucose Monitors


Alberto Gutierrez, PhD (U.S. Food and Drug Administration, Silver Spring, MD)

Dr. Alberto Gutierrez summarized the FDA’s philosophy and methodology in reviewing the performance of medical diagnostic devices, specifically point-of-care blood glucose meters. Quoting Barak Obama, he acknowledged that some aspects of the current review process (such as the wide ISO accuracy standards) are outdated and imperfect – more like “horses and bayonets” than aircraft carriers. He agreed with Dr. Mitchell Scott that YSI is not truly a “reference method” even though FDA classifies it as such. Nonetheless, he considers YSI to be an important “silver standard” that offers a fairly good balance of performance and ease-of-use compared to other diagnostic methods. Dr. Gutierrez concluded with hopes that the regulatory review process would get better in the future: e.g., with the upcoming adoption of a tighter ISO standard.



Mitchell Scott, PhD (Washington University School of Medicine, St. Louis, MO)

Dr. Mitchell Scott gave a semantics-focused presentation, emphasizing a few key points: 1) there are only four true reference methods for glucose; 2) YSI is NOT a reference method, it’s traceable to a reference method; and 3) glucose meters have wide variability in accuracy. In terms of true reference methods, the catalog on reference methods published by the JCTLM has defined four higher order reference methods for glucose: three are mass spectrometry and one is enzymatic (these have very high precision and errors <1%). Interestingly, he noted that YSI is not a “reference method” even though the FDA often uses that language when referring to it in guidance documents. Rather, YSI should technically be called a “comparison method” or “lab analyzer” because it is not as accurate as true reference methods. Dr. Scott concluded with a very brief review of glucose meter accuracy. For the single sample shown, meters had “tremendous” variability from a low of 78 mg/dl to a high of 124 mg/dl.



Jay Johnson, PhD and Jamie Lussier, MS (YSI Life Sciences, Yellow Springs, OH)

Dr. Jay Johnson and Mr. Jamie Lussier gave an overview of YSI past and potential future development. Dr. Johnson began with a look back at 1975, when the YSI 23A came to market for diagnostic whole blood glucose testing. In 1992, YSI Life Sciences introduced the YSI 2300 Stat Plus, which Dr. Johnson said focused primarily on improving user friendliness. The company has since developed a new platform, the YSI 2900, a general-purpose analyzer primarily used in bioprocess applications. YSI Life Sciences is considering whether to build the next-generation YSI 2300 glucose analyzer on this platform. In particular, Mr. Lussier highlighted that the YSI 2900 platform has a slim, modular design and easily exportable data (via Ethernet or USB). He asked the audience and the field to play an active role in shaping the next-generation glucose analyzer by informing the company what kinds of features and functionality are desired.



Darrell Wilson, MD (Stanford University and the Lucile Packard Children’s Hospital at Stanford, Stanford, CA), George Cembrowski, MD, PhD (University of Alberta Hospital, Edmonton, Alberta, Canada), Mitchell Scott, PhD (Washington University School of Medicine, St. Louis, MO), Alberto Gutierrez, PhD (U.S. Food and Drug Administration, Silver Spring, MD), Jay Johnson, PhD (YSI Life Sciences, Yellow Springs, OH), Jamie Lussier, MS (YSI Life Sciences, Yellow Springs, OH), Guido Freckmann, MD (Institute for Diabetes Technology, Ulm, Germany)

Dr. John Mastrototaro (VP, Research and Technology, Medtronic Diabetes): I think that when most of us have used YSI or similar methods, the accuracy is phenomenal at measuring the sample. But sometimes that sample can itself be compromised based on the way it’s collected. Are there best practices on drawing blood?


Dr. Scott: Pre-analytic issues in hospitals are huge – the major source of measurement error. There are standard practices for drawing the volume from the intravascular line, getting rid of it, and then measuring from a fresh draw.

Dr. Freckmann: I agree that pre-analytic procedures are very important to evaluate and standardize.

Dr. Johnson: I don’t think that we at YSI get close to recommending exactly how samples should be drawn; we typically just recommend following whatever the standard practices are. I guess we are conservative in this regard.

Dr. Cembrowski: In the intensive care unit, for example, many samples are drawn from venous blood and many are drawn from capillary blood. This makes comparisons problematic. I think that it takes someone who is very tough in an ICU to manage one standard. Capillary blood draws are probably preferable to venous line draws, but these require much more work.

Q: When we do patient blood tests in the hospital, how reliable are they compared to outpatient testing? And how do you calibrate those point-of-care instruments?

Dr. Scott: A variety of sample types are used in any hospital setting. The literature is clear in the critically ill that capillary testing versus venous laboratory testing does not compare as well as does venous versus venous. It has to do with hyperperfusion of critically ill patients and contamination of fingerstick tests. The literature suggests either venous or arterial will agree with central laboratory testing in critically ill patients.

Dr. Wilson: But that’s often an hour delay

Dr. Scott: Oh no, I’m talking about a venous sample on a meter.

Q: You had that the YSI works down to 5 mg/dl. I’ve taken it down to 1 mg/dl and it still gives good results. With low glucose levels can you take one enzyme channel and one non- enzyme channel and subtract the signals?

Dr. Johnson: That might work, but I would be cautious. In fact that’s where we started in this business. On the original analyzer we used a second electrode to compensate and it doesn’t work if the electrode is totally unprotected and in those days we weren’t protecting the electrode. If it is partially protected with a membrane, we think that removes the interferences. My point is, maybe you can do compensation if you have the electrode partially protected with a membrane, but I don’t think it will help that much.

Q: So the ideal filter would be denatured glucose oxidase?

Dr. Johnson: Yeah, that would be ideal. It might work, but I’d be cautious.

Gary Steil (Children’s Hospital, Boston, Boston, MA): I may have heard [Dr.] Darrell [Wilson] just say it: in the clinic when we evaluate POC results vs. lab measures, it takes significantly longer to get the lab measurement. For example, if you have a lab method that is accurate within 3%, but it takes 10 or 20 minutes to get the sample, then a real-time point-of-care measurement might actually be more clinically useful, even if it is only accurate within 10%.

Dr. Cembrowski: The problem in the US is that you are so profit-oriented, you keep track of every blood gas analysis from the central lab. You could have blood gas analyzers bedside.

Dr. Wilson: We do, and we bill for that too. [Laughter.] I agree with you about the delay – sometimes we see a time delay of up to an hour.

Dr. Scott: Yes – if you are getting central lab results within 10 minutes than you have a much better lab than I do. The time delay absolutely negates the value of central lab measures in units. But the problem with meters is that they have such inaccuracy and bias, there is concern that incorrect decisions are being made with them. This is a question both for FDA and for hospitals like us that currently use meters widely in inpatient care.

Q: Nobody dared to touch on CGM and how to assess accuracy of those. And should we call it something different besides accuracy? Do our sensors have to be as good as blood glucose meters or are they in a separate category?

Dr. Freckmann: The sensor can only be as good as the blood glucose meter used for calibration. So you have to be very cautious what blood glucose meter you use for calibration.

Dr. Gutierrez: I do think that the idea of how you actually determine performance of glucose sensors is difficult. In point-to-point comparisons to the meter or YSI, they don’t perform as well, but we know there is some value in having frequent measurements and knowing whether glucose is going up or down. How you quantify that is difficult. We had a lot of challenges on this with the first CGM submission. I am still not sure, but we have a better handle now. There is a big amount of learning that a patient goes though with CGM and how do you quantify that. When you have a new, disruptive technology that changes how you look at things, how do you evaluate whether it is better or worse or how good it is? We have to figure that out as we move toward using glucose sensors for replacement instead of tracking and trending – when is data good enough and how do you address accumulative data.

Dr. Andreas Pfützner (IKFE GmbH, Mainz, Germany): I think that the YSI is great device; I use it on a daily basis. However, sometimes it has some challenges. In particular we see agreement issues in multicenter studies. Are there any attempts to accept other methods than YSI in clinical trials, Dr. Gutierrez?

Dr. Gutierrez: When we talk about a reference method, we say “for example” YSI. If you were to use a true gold standard method, in some ways that would be preferable. Is it worth the effort to switch to a different, more accurate method? That is to a large extent up to you. We will accept a measure with equivalent or better accuracy and precision. We have used hexokinase, for example.

Comment: When you are looking at precision and accuracy, certainly methods should not have any cross sensitivities, but I think for rating analytical methods we have seen many studies showing cross sensitivities, say from drugs within blood. This is a very important parameter to be looked at when you rate different analytical methods.

Dr. Gutierrez: I would second that particularly when we have evidence of interferences that could result in patient deaths. There were some cases like that with maltose, so the agency pushed hard to solve interference issues so we wouldn’t see that kind of problem.

Dr. Scott: How many of you are aware that every meter – except for the last two most recent ones – overestimate glucose, by definition, in anemic patients? The reason for that is the difference between red cell water content and plasma content. Meters use a factor of 1.22 to translate the whole blood measurement to a plasma value. This correction is exactlyright if hematocrit is 45, but if you have a hematocrit of 25 then your glucose will be overestimated by about 20%. The newer meters measure for hematocrit and correct for it, much as is done with the YSI. But there aren’t yet many available meters doing this.

Dr. Gutierrez: One thing we make meter manufacturers do is test a range of hematocrit. No, we don’t think that the labels are often read, but we do put a lot of effort into putting information on them for those that look.

Dr. Cembrowski: Do you ever use patient blood glucose meters in the hospital?

Dr. Freckmann: We frequently using patient meters for studies but then we do quality controls and we train them to use it. If you use good and well-running meters, there is no big difference. But if you have hematocrit problems and other things, then you can run into problems. But in people with type 1 diabetes, we had good experience using patient meters with some controls.

Live Demonstrations


Larry Katz, PhD (LifeScan, Inc., Wayne, PA)

 We had hoped Dr. Larry Katz might perhaps discuss a next-gen Verio pattern algorithm, but this presentation focused instead on the currently available OneTouch Verio IQ – it was similar to the product theater we saw at ADA 2012 (see pages 69-70 of our report at https://closeconcerns.box.com/s/phnv2z6hpe8x4r81v1kw; for our test drive on the OneTouch Verio IQ, please see http://diatribe.us/issues/41/test- drive). As a reminder, the new LifeScan meter automatically searches for recent high and low blood glucose patterns and alerts the user on the screen. Dr. Katz emphasized how the Verio IQ meter in combination with the included “Tools for Life Pattern Guide” (a small paper guidebook with pullout tabs suggesting actions based on the pattern alerts) helps overcome several challenges patients face. Most new to us was Dr. Katz’s review of the HCP impact study, an evaluation of 64 HCPs in 10 different US cities. Each physician reviewed six paper logbooks and six Verio IQ meters containing 30 days of data. Researchers measured the time it took HCPs to conduct assessments of the logbook and Verio IQ data, the accuracy at which they interpreted the data, their estimates of the 30-day average glucose, and responses to a survey. Notably, it took HCPs about eight minutes to do the evaluations and find all the low and high patterns on the paper logbooks, which declined to less than one minute when using the features in the Verio IQ meter – we found this quite compelling. We look forward to seeing the complete results once they are published.

  •  In the HCP Impact Study, 94% of physicians agreed/strongly agreed that they would recommend the OneTouch Verio IQ (the other 6% neither agreed/disagreed). Additionally, 86% agreed/strongly agreed that the meter would allow for rapid pattern identification (13% neither agreed/disagreed, 2% strongly disagreed), and 92% were either satisfied or strongly satisfied with the meter (8% neither agreed/disagreed). While these percentages might not be immediately accepted as “absolute” since this wasn’t an independent study, they are certainly encouraging to see. We believe that HCPs really value saving time, and this seems like a meter that can help them do it.


3. Insulins and Insulin Delivery

Novel Insulins


Steve Prestrelski, PhD (Chief Science Officer, Xeris Pharmaceuticals, Austin, TX)

Dr. Steve Prestrelski gave a comprehensive overview of Xeris’ non-aqueous stabilized glucagon formulation for use in an auto-injector pen for severe hypoglycemia (G-Pen), a mini dose pen for mild/moderate hypoglycemia (G-Pen Mini), and a formulation for the bi-hormonal artificial pancreas. To stabilize glucagon, Xeris is using non-aqueous solutions with biocompatible solvents (DMSO) that   are already FDA approved. According to Dr. Prestrelski, the company’s glucagon formulation has stability for two years at room temperature. The compound is currently preclinical, though an IND enabling program has been agreed upon with the FDA and the quicker 505(b)(2) regulatory pathway will be used. Xeris has scaled for clinical production (2,000 syringes), completed the bridging pharmacological/toxicological study, and will begin a phase 2a clinical trial in 1Q13 under PI Dr. Ralph DeFronzo (Texas Diabetes Institute). Data is hopefully expected early next year. The company’s ultimate hope is license the technology – during Q&A, Dr. Prestrelski explained that it could come on the market as soon as 2014 if all goes smoothly. (This would be ahead of Biodel’s stable glucagon formulation,   which is expected to be filed with the FDA in mid-2014 per comments at the October 12 Analyst Day.) There is a much less clear timeline and regulatory pathway for Xeris’ G-Pen Mini and the bi-hormonal AP indication, though it is notable that the company is working with the two major insulin-glucagon AP research teams: Drs. Ed Damiano and Steven Russell in Boston and Dr. Ken Ward in Oregon. A clinical trial for the bi-hormonal AP indication is expected to start in 2014 (n=14) under PI Dr. Ken Ward

  • Instead of a water-based formulation, Xeris is mixing Lilly’s glucagon powder with an FDA approved, biocompatible, non-aqueous solvent. The solvent, DMSO, is already in FDA approved parenteral products and is approved at volumes ~200 times greater than what Xeris is using. According to the company, the novel formulation remains stable and free of fibrillation after incubation at 104 degrees Fahrenheit (40 degrees Celsius) for at least two months, and stability data at room temperature predict a shelf life of at least two years. A 505(b)(2) regulatory pathway will be pursued using the Lilly glucagon kit as the reference product. As a reminder, the company had a late-breaking poster at ADA 2012 (see our extensive report, including an update on the field, at http://www.closeconcerns.com/knowledgebase/r/cc525978). Xeris’ ADA poster can be found at http://xerispharma.com/ADA_Poster_FINAL_5_30.pdf.

  • The G-Pen, an emergency glucagon treatment for hypoglycemia, would reduce the current nine-step glucagon administration process to two steps. Dr. Prestrelski demonstrated use of the EpiPen-like auto-injector on stage: 1) remove the safety cap and 2) press the device against the body. The device will reduce the volume of injection through a five-fold increase in glucagon concentration – the benefit is a one-second injection time and a 200 micro- liter injection volume. The pen will have a two-year expiration date at room temperature.

  • The Xeris glucagon formulation is currently preclinical; a rodent model demonstrated identical PK/PD compared to traditional aqueous formulations. Dr. Prestrelski explained that Cmax, Tmax, and AUC were all equivalent to aqueous control. The glucagon formulation demonstrated rapid absorption (Tmax ~5 minutes) and elevation of blood glucose levels within 15 minutes.

  • The G-Pen Mini for mild-to-moderate hypoglycemia is in very early development and there is no final design as of now. It would be just like an insulin pen (e.g., Novo Nordisk’s FlexPen) but would contain glucagon for self-injection for mild or moderate hypoglycemia. The recommended dose would be ten micrograms per year of age. The G-Pen Mini would hold enough glucagon for 30 or more injections (i.e., enough for one month if injecting once per day). Like the auto-injector, it will have a two-year expiration date at room temperature. Conceptually, the G-Pen Mini is a major win in our view – the ability to precisely dose glucagon could help many patients spend more time in zone and less time on the “roller coaster” pattern of hypoglycemia, carb overtreatment, hyperglycemia, overcorrecting with insulin, hypoglycemia, etc. Development risk certainly exists, given regulatory barriers for new products, but we believe interest will be very high from multiple parties.

  • Xeris is working with “various pump manufacturers,” Drs. Edward Damiano, Steven Russell, and Ken Ward to develop a glucagon formulation for use in a bi- hormonal artificial pancreas. The slide specifically displayed the Tandem t:slim pump, corroborating what we’ve heard at recent conferences about Tandem’s development of a dual- chamber pump. Xeris’ target product profile for a pump is a concentration of 5 mg/ml sold in a 2 ml per vial of glucagon with a two-year expiration date at room temperature. The goal is compatibility with most or all available pumps and stability at 37 degrees Celsius (98.6 degrees Fahrenheit) for up to four weeks. The one challenge is that Xeris will need to demonstrate compatibility with pumps due to the non-traditional solvents (“not always easy”).

    • Dr. Prestrelski very briefly mentioned Xeris’ recent NIH/NIDDK Small Business Innovation Research (SBIR) grant to develop its glucagon for use in a bi-hormonal artificial pancreas. The phase 1 grant is for $336,793 specifically for the AP indication. The phase 1 grant is the initial installment of a phase 1-2 fast track SBIR grant, with the potential for a total award of $1.05 million. Phase 2 funding will support IND-enabling preclinical studies and a foundational clinical trial to be conducted at the Oregon Health and Science University under PI Dr. Ken Ward. It is expected to include 14 patients and start in 2014.



Tim Heise, MD (Profil Institute for Metabolic Research, Neuss, Germany)

Dr. Tim Heise’s presentation arguing the need for ultra-long-acting insulins was particularly timely in light of the November 8 Endocrinology and Metabolic Drug Advisory Committee’s 8-4 vote in favor of approving Novo Nordisk’s insulin degludec (as a reminder, discussion centered around hypoglycemia benefits and an uncertain CV signal from degludec’s phase 3 clinical program; see our discussion of the advisory committee meeting at https://closeconcerns.box.com/s/xzy2i90tlpj06o8z5cr3). Dr. Heise’s data-drive presentation was bereft of CV risk data – “I didn’t include CV safety because I myself am not convinced we have an issue here.” After reviewing data on degludec (Novo Nordisk) and LY2605541 (Lilly’s PEGylated lispro), Dr. Heise argued that ultra-long acting insulins resulted in a flatter PK/PD profile, less within-patient variability1, less nocturnal hypoglycemia, more flexible dosing2, and greater body weight loss (with LY2605541). Dr. Heise stated, “It would be wonderful, I think, to have better basal insulins to reduce hypoglycemic episodes.”

  • Dr. Heise began his presentation with a question for the audience: Do you think better basal insulins are needed?

    • Yes definitely! 22%

    • Yes, the majority of patients would benefit from better basal insulins: 42%

    • Yes, but only for a few patients: 17%

    • No, the basal insulins we have are good enough: 19%

  • Insulin degludec and LY2605541 have markedly longer half-lives (25 and 45-76 hours, respectively) than currently available basal insulins, translating to a flatter profile during steady state plasma concentrations. However, Dr. Heise noted that it does take a longer time to reach the steady state with longer acting basal insulins. For comparison, NPH has a half-life between five and 10 hours, glargine has a half-life of ~12 hours, and detemir has a half- life of five to seven hours. Furthermore, Dr. Heise explained that when treated with longer acting basals, the overlapping of insulin doses helps to buffer against changes in any one dose’s absorption pattern leading to less within-patient profile variability.

  • Dr. Heise reviewed data from recent trials comparing demonstrating that longer- acting basals can reduce the risk of nocturnal hypoglycemia. Several recent trials comparing degludec to glargine have confirmed this conclusion (see table below). Additionally, in a phase 2 study comparing LY2605541 to glargine in patients with type 2 diabetes, LY2605541 conferred a 48% reduction in the risk of nocturnal hypoglycemia (p=0.02; Bergenstal et al., Diabetes Care 2012). Dr. Heise also drew attention to the finding in this study that patients had significant weight loss from baseline with LY2605441 treatment over 12 weeks (0.58 kg loss [~1.3 lb]; p=0.01), as well as significant weight loss compared to glargine treatment (p=0.01).

Degludec vs. Glargine


Relative Risk Reduction in Hypoglycemia





Pre-treatment Therapy

All Confirmed



Type 2 Diabetes

Zinman 2012









Garber 2012









Zinman 2011








Only 1 event

Type 1 Diabetes

Heller 2012









Birkeland 2011









Significant risk reductions are indicated with a (*)

1Data was only presented on degludec, but we would expect LY2605541 to confer similar benefits given that the benefit is based on the longer half-­‐life.

2Again, data was only presented on degludec, but we would expect LY2605541 to confer similar benefitsgiven that the benefit is based on the longer half-­‐life.



Lutz Heinemann, PhD (Science & Co., Duesseldorf, Germany)

Dr. Lutz Heinemann prefaced his talk by saying that all of the takeaways could be conveyed by a simple equation, ≈≠= (“similar is not the same”). Indeed, he emphasized that small differences in the complex insulin-manufacturing process will inevitably cause biosimilars to differ from the products they are designed to mimic: the only question is whether those differences are clinically relevant. Fortunately the regulatory guidance on biosimilars is clear in both the EU and US, Dr. Heinemann said, albeit perhaps arbitrary or imperfect in some respects. (For example, he noted that in the US, follow-on insulins will not be regarded as biosimilars until April 2019, at which time they will shift to the regulatory office responsible for biosimilars. He also expressed doubts about the usefulness of pharmacovigilance to assess concerns about immunogenicity – the “most important safety issues” facing biosimilar insulin.) Dr. Heinemann’s personal forecast is that “the insulin world will look different some years from now.” That is, even though the predicted price reductions for biosimilars are only 30-50%, he anticipates that the market landscape will be re-shaped by the offerings of large, high-tech, ambitious companies that remain relatively unfamiliar in the US and EU – such as Wockhardt, Biocon, Bioton, and Gan & Lee.

  • Dr. Heinemann reminded the audience that for biologic molecules, “the process is the product.” Not only are between-manufacturer differences important, but each manufacturer must maintain strict quality assurance to guard against batch-to-batch variability (which could translate to unpredictable or inadequate glycemic control). To illustrate the many ways in which manufacturing can differ for biologics with the same amino acid structure, Dr. Heinemann contrasted the manufacturing processes of the recombinant human insulins Novolin and Humulin to illustrate the vast number of possible differences. He noted that to his knowledge, Novolin and Humulin have never been compared to see whether they would meet strict definitions of biosimilarity.


G. Alexander Fleming, MD (Kinexum, Harpers Ferry, West Virginia), Poul Strange, MD, PhD (Integrated Medical Development, Princeton Junction, New Jersey), Lutz Heinemann, PhD (Science & Co., Duesseldorf, Germany), Tim Heise, MD (Profil Institute for Metabolic Research, Neuss, Germany), Steve Prestrelski, PhD (Xeris Pharmaceuticals, Austin, TX), Yoeri M. Luijf, MD, MSc (Academic Medical Centre, Amsterdam, The Netherlands)

Q: I’m from Becton Dickinson. Can you comment on the regulatory approval pathway for your glucagon formulation and devices? And what are the projected commercialization dates?

Dr. Prestrelski: For the glucagon rescue pen, we have agreed with FDA that we can follow the 505(b)(2) regulatory pathway. We have chosen the Lilly product as a reference. Our business model is to license it, so we cannot predict when it will be commercialized. In a fully funded scenario, where someone takes it smoothly, it could be on the market in 2014. For the mini pen and artificial pancreas, those are new indications. We don’t have as clear a pathway to the market. All will need to be discussed with the FDA to get those products approved.

Q: Dr. Heise, you showed data with basal lispro where you had weight loss. Can you speculate as to the mechanism, is it because of less hypoglycemia?

Dr. Heise: I think it hasn’t been fully understood. Lilly presented animal data that showed the insulin was working primary on liver. Because it is a PEGylated lispro, it might have difficulty crossing tight junctions in peripheral tissues. I think evidence in humans is still outstanding and currently it is not fully clear if this is the only mechanism. From detemir data, I think less hypoglycemia is not a very likely explanation.

Dr. Fleming: Can you comment on the challenges of making comparison between insulin products? This is the tie that binds biosimilar insulin and innovative insulin development. We’ve seen comparisons between degludec and glargine in terms of cardiovascular safety. And then we go back to biosimilars. The issue is confined to looking at PK/PD, and then a kind of readout on antigenicity. Can you comment on the challenges?

Dr. Heinemann: I’m not sure how many hours you will have to discuss this topic [Laughter]. With respect to the cardiovascular effects of degludec, I acknowledge that I’m a bit puzzled. I have a fear that if each new insulin must fulfill cardiovascular studies, research for insulins in general will be hampered in the future. The question is what is the pathophysiological mechanism for insulin to have an impact.

When it comes to comparison, I think this is one of reasons the EMA is looking at revised guidelines. The original guidelines were for soluble insulin only. It’s easy to demonstrate no significant difference. When you talk about long acting insulins and there is no clear Cmax, then you can end up in hundreds of numbers and clamp studies. Is this the appropriate approach? The FDA and EMA have been approached by companies for approval of biosimilar insulins. These are tricky questions. Imagine someone wants to develop a biosimilar of degludec. With that flat time action profile, how do you compare Cmax? There is some impossibility of it. I’m not sure of the equivalence for such long-acting analogs. We need to have open scientific discussions. We should acknowledge that up until now, scientific and academic groups have largely the ignored topic of biosimilar insulin. I hope that this event will stimulate more interest in this topic.

Q: I have a question on anhydrous products, have you thought about doing it for insulin? Would you get same advantages of stability as opposed to having all these additives?

Dr. Prestrelski: We’ve been paying some attention to that. We are working in a commercial environment so we needed to protect IP first and we have done that now. We can isolate and stabilize the insulin monomer and we are seeing if that will translate this into an ultra-rapid version.

Q: Great talk. It was quite impressive to see how steady the values were for days six or seven and looking at the titration period to get there. How long does it take to get to steady state?

Dr. Heise: We have published data on the time it takes to reach steady state – within two to three days with degludec. Not a single patient in the clinical trials took longer than three days. This suggests you could adjust doses every third day. There’s not much sense doing it earlier. There’s not data for basal insulin lispro. It would certainly take longer – something like six or seven days to reach steady state. This is one of the issues we really have to keep in mind.

Q: For nocturnal hypoglycemia, you showed impressive percentage rates. Can you tell us about the absolute numbers? What is the clinical relevance?

Dr. Heise: That’s an excellent question and all companies will look into that. The absolute numbers are for the numbers needed to treat. It seems to range and be very different for population. The number needed to treat to avoid one episode of severe hypoglycemia is between seven and sixty patients. It’s better than I would have expected.

Dr. Strange: A couple of days ago, the advisory committee showed that the analysis of nocturnal hypo was sensitive to the time period chosen. Benefits from 0-6 hours disappeared when analyzed from 0-8 hours. [Editors note: for our discussion on the advisory committee meeting, please see our report at http://www.closeconcerns.com/knowledgebase/r/9e56681d]. And can you comment or speculate on once a week ultra-long acting insulins?

Dr. Heise: Some companies are working on a weekly insulin. It might be feasible pharmacologically to achieve such a long half-life. Whether they are associated with clinically relevant advantages, I don’t know - then we are approaching convenience issues. If they will ever come to market, you have to pick the patients using once weekly insulin very carefully.

Q: Hypoglycemia seems to be less common with longer acting insulins. What about in extreme situations when you have gastroenteritis and have to wait several days for the insulin to work its way out?

Dr. Heise: That is a common concern, which is also related to exercise and it is too late to reverse the dose. Phase 3 trials don’t indicate higher hypoglycemia during illness or exercise or that hypoglycemia last longer. You have to keep in mind that this is still just a basal so if it is titrated correctly, all it will do is prevent increases in blood glucose when fasting. You could easily correct this with other insulins or by eating more carbohydrates.


Insulin Delivery Technology


Andreas Pfützner, MD, PhD (IKFE GmbH, Mainz, Germany)

Expanding on interim results presented at EASD 2012, Dr. Andreas Pfützner reviewed full results from    a trial assessing the effect of the InsuPad device on glucose excursion and insulin dose requirement (for our discussion of interim findings, please see page 12 of our EASD Day #4 Highlights at https://closeconcerns.box.com/s/bs9g1oa0a1nrdkf5bvbh). As a reminder, InsuPad provides localized heating in an effort to improve microcirculation and the pharmacodynamic profile of injected insulin. In a randomized, open-label crossover study, Dr. Pfützner and colleagues found that when the same  amount of fast-acting insulin was injected with the device, maximal glucose excursion was lower than when insulin was injected without the device. Furthermore, when 0.16 U/kg of fast-acting insulin  (insulin aspart) was injected with the InsuPad, it seemed to result in a similar glucose excursion pattern as when 0.20 U/kg was injected without the device, implying that the device could reduce insulin dose requirement. While we wonder whether all MDI patients would be receptive to wearing the product (which adheres like a patch pump), Dr. Pfützner believes that the initial patient reactions in these studies suggest InsuPad uptake will be good.

  • For background, the InsuPad device applies localized heating after insulin injection. It is comprised of a disposable pad (containing the insulin injection window, which has space for six injections; intended for one-day use) and a reusable control unit (contains the heating block, electronics, and a rechargeable battery). The device delivers heat in a wave-like fashion such that three 10-minute cycles of heat (~40 degrees Celsius/103 degrees Fahrenheit) are applied over 50 minutes. The product is being developed by InsuLine Medical and has already received CE marking.

  • Dr. Pfützner and colleagues tested the effect of the device on glucose excursion in an open-label, randomized, four-period, one-way crossover study in which patients with type 2 diabetes (n=16, mean A1c 8.53%) were challenged by liquid meal tolerance tests (MTTs). Each subject went through four MTT protocols: 1) control – 0.20 U/kg insulin aspart (Novorapid) injected immediately before the MTT without the InsuPad; 2) 0.20 U/kg aspart injected immediately before the MTT with the InsuPad; 3) 0.16 U/kg aspart injected immediately before the MTT with the InsuPad (i.e., 20% insulin reduction); and 4) 0.20 U/kg aspart injected 30 minutes post-MTT with the InsuPad.

  • Glucose excursion trends showed that the InsuPad device reduced glucose excursion compared to control when the same amount of insulin was injected (0.20 U/kg). Maximal glucose excursion was significantly less using the InsuPad device. Interim results at EASD 2012 provided greater detail on this portion of the study, showing that maximal glucose excursion was 23% lower with the device.

  • While the result was not quantified, a graph of glucose excursion post-MTT showed that qualitatively, 0.16 U/kg of insulin aspart injected with the InsuPad seemed to result in a similar excursion pattern as the control test (0.20 U/kg pre-meal without device). Dr. Pfützner suggested that the findings demonstrate that the InsuPad device could reduce the quantity of insulin that an individual needs. While he noted that of course there is inter-individual variability, he believes this finding is indeed generalizable.

  • While injecting insulin aspart 30 minutes post-MTT with the InsuPad device led to a faster postprandial glucose excursion, maximal glucose excursion was lower than in the control test (in which insulin was injected before the MTT). Mean maximum glucose excursion was 142 mg/dl in the control arm vs. 129 mg/dl in the post-meal injection group (p<0.013).

  • InsuLine has completed enrollment for the Barmer Study, which will test the device in patients with type 1 diabetes. This study will be an open-label, randomized, parallel group study in over 160 patients. Forty-eight individuals have already completed the trial. While Dr. Pfützner recognized that patients may or may not like adding another device to their routine, he believes patient response will be quite positive: of the 11 Barmer Study patients at Dr. Pfützner’s site who have completed the study, 10 opted to continue into another clinical trial with the device; one wanted to continue using the device, but outside of the clinical domain. Trial continuation was incentivized by a “few years” of InsuPad supply.


Scott Blackman, MD, PhD (Johns Hopkins Children Center, Baltimore, MD)

To study the relationship between pump use and A1c data in children under six years of age, Dr. Scott Blackman mined the Helmsley Charitable Trust Type 1 Diabetes Exchange’s trove of data on pediatric diabetes. Using both cross-sectional (n~700) and longitudinal (n~1,800) analyses of the T1D Exchange data, he found that the pump-using population tended to have higher parental income and education, as well as a higher percentage of Caucasian race (though the most dramatic predictor was simply the center where patients were treated: pump penetration in young children ranged from 5% to 93% in the 16 largest T1D Exchange centers). Young children typically experienced an A1c benefit after starting on an insulin pump (also of note, the young patients who eventually used pumps tended to have baseline lower A1c than other injection users). On the safety side, a higher percentage of pump-using children than injection users reported one or more DKA events in the past year (9.7% vs. 8.2%); this difference bordered on statistical significance after being adjusted for demographics and other factors (p=0.053). Dr. Blackman concluded that pumps seem to be a safe and effective way to meet the challenges of managing type 1 diabetes in young childhood, and he called for more research on the barriers that limit some populations’ use of pumps.

  • Based on a cross-sectional analysis of the Helmsley Charitable Trust’s 67-center, 25,833-patient Type 1 Diabetes Exchange registry, Dr. Blackman suggested several ways that young children who use pumps (n=322) tend to differ from those who useinsulin injections (n=344). (The analysis included children who had been diagnosed with type 1 diabetes for at least one year and who were younger than six years of age at the time of enrollment in the T1D Exchange). Pump use was found to correlate with Caucasian race, higher parental education, higher annual household income, ownership of private insurance, lower BMI, slightly longer duration of diabetes, and lower A1c at enrollment (7.9% for pumpers vs. 8.5% for injection users).
  • Dr. Blackman also presented a longitudinal comparison of A1c data from over 1,000 young children divided into three groups: those who used insulin injections and never switched to pumps, those who were using insulin injections at the time of A1c measurement but switched to pumps later on, and those who had used pumps long enough to become familiar with the technology. At every age from two-to-five years old, mean A1c was lower among pump users than injection users. Notably, however, mean A1c was also lower among the injection users who later switched to pumps than the injection users who never switched to pumps. This suggests that young children who use pumps tend to be at a double advantage relative to their peers: overall they have lower baseline A1c while still on injections, and they see additional A1c benefits after switching to pump use. Given the apparent benefits of pump use (and the relatively low uptake among certain groups of patients), Dr. Blackman called for more research
  • The strongest predictor of pump use seemed to be the center where a patient was treated, rather than any individual-specific factors. The 16 largest pediatric centers in the T1D Exchange varied from 5% to 93% pump use among young patients (with fairly even distribution of the percentages in between). Dr. Blackman said that the researchers had not yet found any confounding variables that could explain these between-center differences. We did not find this surprising; we think that the outlook and culture of the clinical staff has a huge effect on patients’ adoption of new technologies and therapies.


Ronald Pettis, PhD (Becton Dickinson, Research Triangle Park, NC)

Dr. Ronald Pettis provided a quick review of BD’s intradermal insulin delivery technology, which is based on 1.5 mm x 31-34 gauge stainless steel microneedles. He summarized the results of two previously published studies: Pettis et al., Diabetes Technology and Therapeutics 2011 and McVey et al., JDST 2012. The first study showed faster pharmacokinetic (PK) and pharmacodynamic (PD) profiles for microneedles relative to subcutaneous delivery of human insulin (PK: -75 minute improvement in Tmax; PD: -30 mg/dl improvement in BGmax, 0-90 min); for lispro, the PK/PD differences were much less impressive (PK: -28 minute improvement in Tmax; PD: a non-significant improvement). The second study demonstrated no detectable time in range differences between intradermal and subcutaneous delivery, though several secondary PD endpoints were significantly in favor of intradermal delivery (a 12 mg/dl improvement in BG0-90min, a 7 mg/dl improvement in BGmax). Intradermal delivery also appears to reduce intra-subject variability from dose to dose, which is also encouraging to see. BD just finished a 72-hour PK/PD study of microneedles and will be defining the regulatory pathway going forward – the critical aspect is that commercial insulins are not currently approved for intradermal administration. Given the sincere clinical need for faster insulin (and the closer-to-market nature of intradermal delivery vs. a new drug), we are hope that pathway is straightforward and could potentially bring intradermal technology to the market near-term.

  • BD’s stainless microneedles are 1.5 mm long, 31-34 gauge, and already on the market for a pre-filled flu vaccine. BD has also incorporated them into infusion sets forinsulin pumps. The keys to the technology are that the fluid is delivered to the same space each and every time with the same accuracy and reliability.
  • Dr. Pettis summarized the results from two previously published studies of microneedles in people with diabetes: Pettis et al., Diabetes Technology and Therapeutics 2011 and McVey et al., JDST 2012. The table below summarizes the pharmacokinetic and pharmacodynamic results from both studies for lispro only – we note that in the first study, the improvement with intradermal delivery was more impressive for human insulin than for lispro. We were disappointed not to see a time in range difference between intradermal lispro and subcutaneous lispro in the second study, though Dr. Pettis explained this may have been due to the study design (a low number of excursions and high overall control).
  McVey et al., JDST 2012 Pettis et al., DT&T 2011
Pharmacokinetics (PK) Improvement with Intradermal Microneedles
  • Tmax: -16 minutes* and -25% intra-subject variability
  • Tmax: -28 minutes*
  • Cmax: 125%*
  • Ins-AUC 0-90 min: 130%*
Pharmacodynamics (PD) Improvement with Intradermal Microneedles
  • "No detectable time in range differences due to high level of overall control"
  • BG0-90min: -12 mg/dl
  • BGmean: -7 mg/dl
  • BGAUC90: -5%
  • Intra-subject variability (BGAUC and BGmax): -15%
  • Lower BG0-90min and BGmax but not statistically significant**


* p <0.05

** "May have been affected by variable subject insulin sensitivity and weight normalized dosing strategy.”

The safety and adverse event data for intradermal delivery raises no major concerns. Intradermal (ID) delivery is associated with lower perceived insertion pain as measured by 10-point VAS scores (0.3 for intradermal vs. 0.6 for subcutaneous), though pain was increased after boluses (2.4 for intradermal vs. 1.1 for subcutaneous). There were no route differences in adverse events, though slightly higher observed Draize dermal erythema and edema was observed with intradermal delivery. The new delivery method has shown feasibility in insulin pumps without leakage or pump alarms and no route differences in hypoglycemia.


Thomas Frei, BS (Roche Diabetes Care AG, Burgdorf, Switzerland)

In a technical presentation, Mr. Thomas Frei reviewed the design changes to the next-generation DiaPort (Roche’s system for intraperitoneal insulin delivery), which will launch in Europe and Australia in January. As a reminder, the current DiaPort is comprised of an infusion set with a ball cannula, fixation disc (to stabilize the port), membrane, port body (implanted in the body), polyester field, and catheter (which goes directly into the peritoneal cavity). The device creates a direct line between the port and the peritoneal cavity such that 75% of insulin enters the body through the portal vein in the liver and 25% of insulin goes through other vessels in the body. The hope is that this makes insulin delivery more physiologic. The next-generation device has: 1) a smaller membrane that only needs to be changed every six months (as opposed to every three months with the current DiaPort); 2) improved tools to exchange the membrane such that it can be done by a diabetologist (assuming sterility can be achieved); 3) a polyester felt on the port body to help prevent infection, hold the port in place, and improve ingrowing of implantation; 4) a redesigned catheter that is flatter (to reduce torque in the body) and has a larger diameter (to prevent occlusion); and 5) accessories to make it easier to connect the infusion set. Mr. Frei closed his presentation with a look towards the future, announcing that clinical trial results of the new DiaPort system will be presented at ATTD 2013, which is being held February 27-March 2 in Paris, France. He also reminded the audience of the ongoing trial in France (supported by JDRF) that will test the DiaPort as part of the artificial pancreas. According to ClinicalTrials.gov (Identifier: NCT01555788), the DiaPort AP trial is still recruiting; the trial is slated to complete December 2013.



Gerold Grodsky, PhD (University of California, San Francisco, San Francisco, CA), Andreas Pfützner, MD, PhD (IKFE GmbH, Mainz, Germany), Graham B.I. Scott, PhD (National Space Biomedical Research Institute, Houston, TX), Thomas Frei, BS (Roche Diabetes Care AG, Burgdorf, Switzerland); Scott Blackman, MD, PhD (Johns Hopkins Children Center, Baltimore, MD)

Q: With the enormous variation in pump usage between centers, were A1c outcomes at centers varied? Were centers using more pumps likely to have better outcomes?

Dr. Blackman: We are in the process of looking at exactly that question. I cannot answer for you today.

Dr. David Klonoff (Mills-Peninsula Health Services, San Mateo, CA): How long does the DiaPort catheter stay in the abdomen? Must it be changed at certain time? And for everyone, do these new delivery technologies mean lower amounts of insulin?

Dr. Frei: You only exchange the catheter when there is a problem – occlusion or growing of tissue. There’s no reason to exchange the catheter when there is no problem. On the insulin question, normally when you change patients to DiaPort, you reduce the amount of insulin needed by about 10%. Of course, if patients have insulin resistance or bigger amount of insulin – 400 units or 1000 units – you reduce it much more.

Dr. Pettis: There is some differential in postprandial usage. The total bioavailability of intradermal insulin is equivalent to subcutaneous insulin.

Dr. Pfützner: For the subcutaneous space, the faster insulin is removed from the injection site, the lower the amounts that are needed. There is less degradation. It’s true for all three delivery methods presented. You also have the metabolic impacts – shutting off hepatic gluconeogenesis. For the InsuPad device – heating the injection site after injection – patients need approximately 20% less insulin. In one patient – and this anecdotal – we saw reduced prandial insulin requirements from 50 units per day to 15 units per day without losing control. There was huge variability between patient populations. Some don’t change dose at all.

Dr. Yogish Kudva (Mayo Clinic, Rochester, MN): Dr. Pettis, what is the maximal intradermal volume you can use in a bolus?

Dr. Pettis: Most of the studies I presented on are dose-ranging – we used about 3-14 units to cover meals. However, in other studies we have shown linear PK with 20-30 units, and good PD response as well.

Dr. Lutz Heinemann (Science & Co., Duesseldorf, Germany): Dr. Pfützner, you show figures with local changes in blood flow. Is this in the skin or subcutaneous tissue? Also, you mention inter-individual differences. Could you teach me why we see these differences?

Dr. Pfützner: The heat does not differentiate by type of tissue: there is impact on capillary blood flow both directly under the skin and in the subcutaneous tissue. (The effect is less pronounced in the subcutaneous space, which is further from the heat source.) As for inter-individual variability, you and I wrote the papers on insulin absorption together: obviously the composition and thickness of tissues change among individuals. There is no general rule, but we can say that when you improve microcirculation – other means, like massage, also work – you see enhanced insulin uptake and absorption. To me the InsuPad’s beauty is that it is a fairly well understood physical system, not a new formulation or analog.

Dr. Heinemann: What about intra-patient variability with the InsuPad?

Dr. Pfützner: That’s one of the next studies the company will run.

Dr. Steven Russell (MGH Diabetes Associates, Boston, MA): There was very rapid uptake of the dye that you showed in the imaging. But on the time scale of the movie, there was not any movement draining out of the lymph nodes.

Dr. Pettis: This is more of a limitation of the technology. This was peripheral imaging technology and the light penetration isn’t enough to go beyond a few centimeters down into tissue.

Dr. Russell: What about the residence time in the lymph nodes? Looking at the speed, wouldn’t the PK be faster?

Dr. Pettis: We were not specifically trying to measure residence time in the nodes. We think there is two- phase absorption process. You have rapid uptake that goes deeper into the nodal chain, where you get trafficking into systemic circulation and distribution. Then, there is a later phase of absorption of some of depot left at the site. It’s slightly slower.

Q: I want to turn back to the InsuPad. Is there any similar device for pump patients?

Dr. Pfützner: Yes. The InsuPatch has been studied in the US, and FDA approval is on its way.

Q: Do you have any idea if heating the subcutaneous tissue causes any long-term damage to the skin?

Dr. Pfützner: In none of the studies, for the InsuPad or the InsuPatch, have we had evidence or clinically relevant descriptions of adverse site effects or skin effects. I should have mentioned that the heat goes up to approximately 45˚ C (123˚ F) and that it is intended to be placed at a different site on a daily basis; it was designed to be used for up to six injections of rapid-acting insulin.

Q: What should the price be for such a device?

Dr. Pfützner: I would refer you to the people from the company who are here at the meeting.

Q: For Dr. Pettis, since you are injecting into an immunologically competent region [with microneedles], is there any antibody formation?

Dr. Pettis: I can give various perspectives on that. Lymphatic absorption is probably not well studied or understood. It probably plays a larger role in subcutaneous processes than people give it credit for. We’re pushing the lymphatic absorption pathway by the routes we've shown. Is there a differential in terms of lymphatic exposure? Without getting good quantification of subcutaneous lymphatic exposure, it’s hard to say. I don’t think there is substantial residence time, but obviously there is transit time involved. We’ve done some preclinical animal studies with multiple exposures over time. There is a slight increase in overall antibody exposure. We are not able to tell any differences in neutralizing antibodies to insulin. The assay was not optimal. We are beginning to monitor total exposures in human studies. The disadvantage is that we have not gone to multi-week, multi-month exposures at this time.

Q: How many subjects have received the new DiaPort? What is the longest that any has gone without needing to change the catheter?

Dr. Frei: There are 12 patients in our trial in Germany; they started using DiaPort last November and December, so for them it’s been about a year. So far we have seven compassionate-use patients, one of whom has now used the DiaPort for 13.5 months. There are also five patients in Paris, bringing the total with the new system to 29. With the older system, the longest use was seven years. However, I am confident that with the new system the duration will be longer. After launch we will have a registry of DiaPort users to keep track of outcomes with the system.

Dr. Klonoff: Dr. Blackman, you said that pump users have lower A1cs. The also came from families that had higher education and higher income. Can you speculate on whether the improved A1c is due to the technology or the environment of children?

Dr. Blackman: Those differences persisted even after adjusting for the demographic factors in model. A better way is to look at insulin pump users before vs. after they started using pump. Either way there is a difference.

Dr. Grodsky: Have any of you tried adding each other’s approaches together? For example, a small needle with site warming?

Dr. Pfützner: We don’t know if there will be additive or exponential defects. I would also mention that we have rapid-acting insulins already under investigation. It would be interesting to see how much Biodel’s insulin, for instance, can be further accelerated. But if we get faster than 15 minutes, than that is faster than physiology – we will see if that makes any sense.

Dr. Pettis: I think that the potential of combining approaches is high but that each one needs to be validated on its own first.

Comment: Dr. Grodsky: This is a question for anyone in audience. Are there any insulin manufacturers out there? We presented some of the approaches here. What hasn’t been discussed is formulations of insulins themselves. Are any of the companies working to change the formulations to provide closed-loop insulins? [no response]

Q: The PK and PD of insulin are driven by the metabolic needs inherent to the organism. What is more important for pure speed is the need to balance physiological needs. We need some way to match insulin to the needs of organism and with the sugar.

Dr. Pfützner: It’s a good question. We are injecting the right hormone, at the wrong site, in order to achieve something called physiology. To do that, we need insulin to have a first pass effect at liver. Insulin connects to metabolic activities and those on the endothelial vasculature. When faster insulins get into the body, we see more beneficial effects. If you give insulin in the wrong dose at the wrong time, it has an atherogenic effect. Having insulin in a closer physiologic range is certainly beneficial.

Dr. Pettis: The key is that we are doing what we can with what we have. We have ways to make it more physiologic, at least in the time course. That doesn’t account for the other physiologic issues that others have brought up. There is potential for a lot of these and how you can put them into advanced systems such as the AP.

Q: In our pharmacodynamic studies, maybe we could look at not just glucose, but postprandial fatty acids after a meal?

Dr. Pfützner: Absolutely. We can also measure the direct effect on the vasculature. The technologies are getting better.


Regulatory Science, Technology Adoption, Software, and Other

Keynote Address


Jeffrey Shuren, MD, JD (Director, CDRH, FDA, Silver Spring, Maryland)

We were glad to see the FDA’s CDRH Director Dr. Jeffrey Shuren front and center at DTM in the Friday morning keynote speech. Dr. Shuren emphasized CDRH’s vision: “to give (diabetic) patients access to high quality, safe, and effective devices of public health importance first in the world” – admittedly, we are pretty far from this today. He explained that the division is taking this vision “very seriously” through a number of key steps: collaborating with companies (CGMs, pumps, the artificial pancreas, data management), developing a public-private partnership to focus on regulatory science for medical devices (this sounds encouraging though few details were shared), developing better tools and software, and use of the new innovation pathway and entrepreneur in residence program. Overall, we found Dr. Shuren’s words encouraging to hear, though they were fairly general in scope and light on details (e.g., “we are working with companies”). However, he also had some very promising and frank comments about FDA’s challenges and how it can better encourage innovation and help companies get devices to market faster – this perspective was really great to hear. In terms of specific diabetes technology, he only mentioned the Medtronic Veo and was very non-committal on its status: “we’ll see where that goes and whether we’ll have that technology for US patients in the near future.”

  • Dr. Shuren emphasized the major challenges facing regulators and regulatory science: communication, infrastructure, and funding. He explained that regulatory science (the tools, standards, and approaches needed to evaluate safe and effective medical devices) is not well understood or appreciated in the medical device ecosystem. Additionally, most of the evaluation work is scattered throughout the country – it is inefficient (one expert here, one project there) and there is very little investment by the federal government. For comparison, NIH’s FY12 budget for research was $30.7 billion, including $575 million for the NIH’s new Center for Advancing Translational Sciences (NCATS). By contrast, FDA has $15 million for medical device regulatory science (excluding staff). “The disparity is huge,” he noted.

  • Dr. Shuren highlighted classes of diabetes technology as examples of how CDRH is working to get better devices approved sooner. His discussion was fairly general for the most part, more about broad approaches than specific companies or devices.

    • 1) Data management: Dr. Shuren explained that we have meters, CGMs, and pumps collecting data, but clinicians have limited time to look at it and must deal with lots of cables and downloading hassle. For patients, this makes it challenging to manage diabetes. FDA is currently working with companies to bring data management technologies into a single stream of information. Additionally, the Agency is working todevelop analytical tools that make interpretation easier. Dr. Shuren’s slide highlighted that “medical cell phone apps [are] coming soon.” We hope this means the finalized mobile medical applications guidance is on the horizon.

    • 2) CGM: Dr. Shuren emphasized that “CGMs are not as accurate as we need them to be” for the closed loop. FDA is working with companies to develop better, more reliable, and more accurate sensors. He stated that companies “are taking on that challenge” and it’s “encouraging to see some of the advances that are hitting the market” – this may have been an indirect reference to Dexcom’s recently approved G4 Platinum.

      • How do we better assess CGM? FDA is working on better in vitro screening of substances that can interfere with CGM readings. Work is also ongoing to better understand the effects of biofouling and how to manipulate a sensor’s surface. The idea is to understand if there are important physiological differences that lead to differences in long-term sensor function. FDA would then provide this feedback to companies to make better technology.

      • Emerging technologies in hospital glucose sensors. FDA is working to understand how changes in physiological pH affect sensor accuracy in the hospital setting. The Agency is also focusing on optical glucose biosensing and minimally invasive sensing.

    • 3) Insulin pumps. Dr. Shuren reminded the audience of the FDA’s 2010 initiative on infusion pumps. Previously, there were thousands of adverse events being reported for infusion pumps, many related to insulin pumps. Interestingly, Dr. Shuren believes the new initiative has resulted in higher quality regulatory submissions. In the two-year period prior to initiative, the Agency cleared 51% of pumps. Now, the FDA is clearing about 70% of them. The FDA has also developed better tools for manufacturers to use and development of a generic insulin infusion pump safety model is ongoing.

    • 4) Artificial pancreas. Dr. Shuren noted, “We’re not there yet because we need better components, but we’re well on our way to getting there.” He emphasized that the Agency is committed to this technology and would “love to see it come to the US first.” The FDA has consolidated the review team in CDRH to help improve oversight over the AP. In the past year, Dr. Shuren highlighted that the FDA has approved four or five clinical trials every single month devoted just to the AP. The Agency has also approved the first outpatient closed loop AP study in the US.

      • On the FDA status of the Medtronic Veo, Dr. Shuren was disappointingly non-committal and fairly vague: “We have an in-house application and we’ll see where that goes and whether we’ll have that technology for US patients in the near future.” He mentioned that it has LGS technology and “is already CE Marked in Europe” – we would of course add that it’s been a three- plus year delay…

      • Dr. Shuren explained that the AP draft guidance documents (subsequently finalized and posted a few hours after his talk) were somewhat unique – usually, the regulatory pathway must catch up with the science. For the AP, it was the other way around: the regulatory pathway needed to get ahead of the science (we would note that this was really not the case with low glucose suspend).

    • 5) Bioartificial pancreas. FDA is working to better understand combination products and is focusing on different ways to have successful encapsulation of pancreatic islet cells.

  • To overcome the challenges of funding and inefficiency, FDA is setting up a public- private partnership with LifeScience Alley. The Agency is working to set up a 501(c)(3) organization that will be separate and only focused on advancing the regulatory science for medical devices. One of the areas will be diabetes. The hope is to get this off the ground in the “near future.” The partnership will allow sharing of resources, dollars, expertise, data, and allow for companies to come together and not run into legal challenges. We hope this could allow for independent testing of devices, especially blood glucose meters, which would jointly be supported by money from all companies.

  • Dr. Shuren closed with a review of the FDA’s innovation pathway, a new route to market for breakthrough technologies. It serves as an “incubator cell” for new approaches and tools to reduce the time and cost of development, assessment, and review of breakthrough (and other) devices. It also transforms how the FDA and innovators work together. Part of the program includes the entrepreneurs in residence program, which invites experts from the medical device industry (VCs, patients, experts, and companies) to the FDA for some of the Agency’s day- to-day work. The pathway includes an application process, a collaboration phase, a clinical trials phase, and market approval. A new version of the pathway was recently launched and Dr. Shuren specifically mentioned that there are three products for end-stage renal disease. He did not specifically address diabetes, though we certainly hope industry experts are taking part in the process.


Questions and Answers

Dr. Yogish Kudva (Mayo Clinic, Rochester, MN): It’s very important to all of us that the US be ahead of the curve and have the first device approvals. What are the lessons we can all learn from how the Medtronic Veo process has gone? Why are we lagging behind?

A: From an FDA perspective, in the past few years, we’ve had challenges on a variety of fronts. Our programs have not been sufficiently predictable, transparent, or timely. We’ve been very public about those challenges. We recognized those problems when I came and we put out two reports on it. We were very frank about the challenges. In early 2011, we developed a plan of action and steps to start fixing the program. Not only has there been lots of progress, but we are now for the first time seeing changes in our performance that we have not seen in some cases for a decade. We will put this data out in the coming weeks on what it’s been like before and where we’re headed now. We’re focused on making our programs better.

We need to tackle this issue on science. Most other countries are not safety and effectiveness. The bar is different – other countries don’t need to be effective. We believe that’s important for patients. If that bar becomes irrationally too high, it becomes a barrier. You need to find a sweet spot to have safety and effectiveness, but do it in a way that is rational, timely, reasonable, and not costly. Regulatory science is the linchpin for getting there. Through things like computer models, we can test drive technology without animal studies – that is a game changer. Those are the advances we’re talking about. Better pathways, putting out guidance, and advancing the very science itself.

Dr. Robert Vigersky (Walter Reed National Military Medical Center, Washington, DC): One of the barriers in getting devices into hands of patients is the IRBs at the various institutions. What is the FDA doing to work with IRBs on an individual basis or to put out guidance to get protocols to IRBs?

A: We don’t have authority over IRBs. This is one of the issues on the table for the entrepreneur in residence program. How can we streamline clinical trials? The scope isn’t about what’s solely within the jurisdiction of FDA. It’s anything that we can influence in any way, shape, or form. IRBs are something we’ll look at. Do you move to a central IRB model? It could be a time saver. How much time does it take contracting if you’re doing a 70-site study? Why not have set templates with contracts? That can lead to big efficiencies and not big costs to do.

Peter Rule (OptiScan, Hayward, CA). If we could envision the ideal collaboration between industry and the FDA, with certain endpoints and certain outcomes, and jointly meet them, there would be a high probability of approval. But the current risk benefit standard makes that difficult. It’s hard to power trials as a manufacturer. Do you see the day where there is more refinement around the notion that risk-benefit is a subjective state?

A: At the end of the day, there will always be a little bit of subjectivity. Science is often gray. I think we make a lot of smart decisions, but not as consistently as we could. What is the right kind of assessment – can you get out ahead and work with industry and academia. That’s what we try to do with the AP. Just this past April, we released a final framework on benefit-risk for those seeking PMA or de novo decisions. It’s very patient centric. When technologies come on the market, they’re not coming to be used on you. They are used on patients. We need to be focused on patients’ perception of risk-benefit. With a new technology, you cannot expect it to be a home run on the first iteration. You must take that into account. If you don’t, you will never let it on the market and it will never get better. My staff is applying that to every single PMA and de novo submission. For some technologies, there is a risk that some people would not take. But some would. In that circumstance, if you’re explicit about that risk, let’s let patients and practitioners make that call.

Q: For technologies like glucose sensing, what about an implantable vs. a non-invasive sensor – is there a useful domain for both?

A: You go where the technology takes you. If the technology is good enough that you didn’t need implantable, you wouldn’t use it. If the answer is no, you would still have implantable. It would be a wonderful world if we didn’t have to use as many invasive technologies on patients – we’re a long way from that, but it’s a terrific goal that we should be shooting for.

Dr. David Klonoff (Mills-Peninsula Health Services, San Mateo, CA): Thank you for a really good presentation. It’s nice for us to hear that you really get it and understand the issues that are going on. On behalf of everyone, we want to thank you for the hard work.


Increasing Adoption of Diabetes Technology: Better Science


Natalie Wisniewski, PhD (Medical Device Consultancy, San Francisco, CA)

In order to allow a broader set of diabetes technologies to come to market, Dr. Natalie Wisniewski proposed that the diabetes technology field must adopt a wider array of accepted clinical trial endpoints. Drawing from what she has seen in cardiology, Dr. Wisniewski explored the possibility of expanding accepted diabetes trial endpoints by considering composite outcomes and additional surrogate outcomes. In the same way that major adverse cardiac events (MACE) are used to evaluate cardiovascular devices, she suggested that a composite outcome for safety of diabetes devices (i.e., one that adds up retinopathy, neuropathy, nephropathy) could address the low incidence rate of each individual outcome and add statistical power to trials. Similarly, a composite outcome could be used for a more complete picture of quality of life (i.e., one that measures number of hospitalizations, ER visits, medications taken, severe hypoglycemic events). Turning to surrogate outcomes, she emphasized the distinction between true clinical, patient-important endpoints (e.g., mortality, quality of life, macro- and microvascular complications) versus surrogate endpoints (e.g., A1c, postprandial glycemia, fasting plasma glucose). While selecting surrogates needs to be a very measured process (“a correlate does not make a surrogate”), she argued that it is in the field’s interest to find new surrogates, because true clinical outcomes necessitate longer, larger trials. In order to evaluate new endpoints, Dr. Wisniewski proposed a working group made up of industry, academic, government, and foundation leaders to standardize and validate endpoints.

  • We admired Dr. David Klonoff’s (University of California, San Francisco, San Francisco, CA) quick response to Dr. Wisniewski’s proposal for a working group on endpoints. To open the next panel, Dr. Klonoff announced that the Diabetes Technology Society would organize such a working group and encouraged the audience to reach out to him about how to have a viable meeting to discuss endpoints and what the goals of the meeting should be.


Bruce Quinn, MD, PhD (Foley Hoag LLP, Boston, MA)

When we talk about regulation, said Dr. Bruce Quinn, we have “regulatory science,” but when we talk about reimbursement, we are far from “reimbursement science.” Dr. Quinn underscored the complexity of the US health care system composed of multiple organizations – Medicare, Medicaid, private insurers that all act differently. Even within individual organizations, the system is dynamic with varying priorities vying for the top spot. For example, Dr. Quinn explained that one Medicare policy-maker might consider the biggest issue in diabetes to be fraud, while another might think that the biggest issue is the cost of basic diabetes supplies. Dr. Quinn drew attention to some of the negative characteristics of the system. First, he explained that non-traditional products, like OmniPod, can “fall between the cracks,” and are not easily reimbursed by the system because they do not neatly fall into a category. (OmniPod is currently classified neither as insulin nor durable medical equipment.) Next, he pointed out that the health assessment world can be very toxic – “any innovation can be picked apart by someone else.” To this end, Dr. Quinn pointed to selected endpoints (particularly, surrogates) and baseline characteristics as components of randomized clinical trials that are often used by cost-conscious payers to debase findings. Dr. Quinn closed his presentation by adding that there are positive features of the US healthcare system as well: 1) he believes that the growth in integrated care systems will be beneficial for the device industry because providers and payors will become more invested in what happens to patients years later; and 2) he believes that the people at CMS really will listen – while there may not be a “reimbursement science,” there is a science to getting reimbursed and it includes getting credible publications, stakeholders, patient group involvement, and importantly, advocates within the Agency.



Naomi Hamburg, MD (Boston University Medical Center, Boston, MA)

Dr. Naomi Hamburg applied lessons from the field of peripheral artery disease (PAD) to diabetes care. She reviewed how technology has changed the management of PAD, specifically, how the explosion of new devices available to treat PAD has led to a dramatic rise in the use of endovascular interventions such that more people with PAD are being treated. She explained that procedures are being offered at a higher rate to those with moderate and more severe cases of PAD than in the past. Dr. Hamburg remarked that though there are more medical treatments for PAD to chose from, patients and providers should remember that medical therapy itself is a choice; she gave the example of how exercise has successfully reduced symptoms for many with PAD – certainly, the option of lifestyle intervention applies in diabetes care as well). Dr. Hamburg concluded that, similar to the PAD field, as the diabetes field sees more and more medical interventions become available, evidence will need to guide therapy optimization.



Naomi Hamburg, MD (Boston University Medical Center, Boston, MA); Natalie Wisniewski, PhD (Medical Device Consultancy, San Francisco, CA); and Bruce Quinn, MD, PhD (Foley Hoag LLP, Boston, MA)

Q: I have a question for Bruce and Natalie. What is the trend in the US regarding the acceptance of surrogate markers? In Europe I see it moving in the opposite direction. NICE in the UK is ranking publications according to evidence level. The highest evidence level is only achieved with patient related outcomes not disease related, like A1c. Patient-related outcomes are increased survival, reduction in cost, or quality of life for example. I think that the acceptance of surrogate markers will become more difficult. What is your opinion?

Dr. Quinn: I totally agree with that. There is a lot of skepticism. A lot of what I see I can pin a psychological affect to it. If you have a few surrogates that blow up, that is very vivid in the mind of policy makers. They’ll talk for years about it. The outlier studies that show up in JAMA or NEJM or Lancet that show real tight glucose control doesn’t help – that’s what they remember, not the other 700 studies. You’re right, surrogates have more skepticism.

Dr. Wisniewski: I agree. There is a higher level of scrutiny. But I also believe that because of new tools like CGM and there’s a kind of understanding that maintaining glucose in a certain range is very important – I think there is room to explore. I don’t think we have to run a massive trial to validate it, though that would be optimal. Endpoint validations have been pieced together by meta-analyses before. There is certainly an abundance of CGM-type trials and data from smaller trials can pieced together.

Dr. Lutz Heinemann: We are very skeptical that the DCCT can be repeated. Do we need a DCCT with CGM in order to validate that CGM is a different world, which it is in my opinion. The payor might want to avoid CGM because it is more costly. We have a dilemma, but what is the way out.

Dr. Klonoff: Is there anyone participating in a working group to identify mutually acceptable endpoints? How could this be done?

Dr. Wisniewski: I haven’t seen it. The closest I got was a Clarke Error Grid workshop, but that was a one- time session. I think there is a place to gather clinicians and foundation leaders and people with experience in a variety of areas to really put a framework around what endpoints should be and why and which endpoints can be easily validated or obtained in the way trials are structured.

Dr. Klonoff: The meeting you went to has moved along. There is a group working on the Clarke Error Grid.

Dr. Heinemann: I think there is a need to have an end point discussion.

Dr. Klonoff: I think these additional endpoints are needed. Dr. Hamburg, how do you compare our more limited number of endpoints with what you have in cardiology and do you see same trends in cardiology in terms of patient endpoints over disease endpoints?

Dr. Hamberg: The same questions are in play in cardiology. Major adverse cardiovascular events (MACE) has been well accepted, which includes a combination of patient and disease specific endpoints. The difficulty in diabetes is whether it is about managing the glucose or managing the complications so it’s harder to know where the true clinical endpoint is. In the cardiovascular field there’s been attempts to think about surrogates, like carotid IMT, and that is appealing from and industry standpoint but I don’t think it is okay from an approval standpoint.

Dr. Klonoff: We have surrogates like decline in renal function and we don’t even try to get those accepted as endpoints right now. Is there anyone from industry who wanted to do a trial but didn’t have the right endpoint?

Comment: Sure, David. If you have a device designed for a patient who is in good control and wants to stay there, then change from baseline A1c doesn’t do much, it simply misses a lot of the point. It’s a different question from what endpoint will actually get you reimbursed for innovation – that’s different from what endpoint has meaning.

Dr. Klonoff: Is there any endpoint you wanted to use and decided not to bother because it wouldn’t lead to reimbursement?

Comment: Urinary microalbumin, markers of inflammation, you can do changes in retinopathy…I’d ask the cardiologist, is LDL a valid baseline anymore and what’s going on with MACE? In my opinion, in the last major trials because we calculated powers based on event predictions before everyone was on statins, we have underpowered studies published in journals that love to poke in the eye rather move the field along.

Dr. Hamburg: We accept markers like blood pressure and A1c until there are exceptions. And then people play up the exceptions. As we improve all the therapies that we deliver, the major adverse events are going to be less common. If you did DCCT now given the baseline medical therapy now, would you see the same kinds of benefits given everyone on aspirins and statins?

Comment: Science and methodology around surrogates is often more robust than around patient outcomes. Compare A1c versus quality of life. How do you measure quality of life? Even something as seemingly simple as hospital admissions is related to thresholds of different clinicians. Even death can be difficult. The science around whole organisms outcomes is softer than for surrogates. So I think there is a reverse tension there.

Dr. Hamburg: It’s interesting that in other fields like pulmonology or spine reconstruction that everything is centered around quality of life because there aren’t great biomarkers. Really, it’s accepted that quality of life endpoints are soft endpoints that are based patient questionnaires, and in some fields that’s what the gold standards are. It just hasn’t come to bear in the diabetes field. Maybe we need to have a numeric value along with a quality of life metric

Comment: There is a delay process from the end of a randomized clinical trial to peer review to publication to health assessment. With medical technology that innovates every 18 months whether incremental or big, it’s a minimum three years for health technology assessment. I think yes, surrogate markers do something, but we need to address lag time in medical device innovation.

Dr. Heinemann: We need studies addressing not a given product, but a class of devices. In JDRF CGM trials, the devices were not presented according to product but to class of innovation. We need more class- related studies and not so many device-related studies. I would ask here for larger studies, but a smaller number of studies.

Comment: I do think though there are differences in classes of products.

Dr. Wisniewski: To run the studies necessary – that is where a role comes in for a working group. We need support from foundations to power a study appropriately and define in advance what we think the best clinical endpoints would be in order to design a larger trial to validate those clinical endpoints. We need a working group because no industry or academic player is going to be able to come up and do it on their own.

Dr. Heinemann: With glycemic variability, I’m confused. What is the outcome of the trials we have? I have no clear understanding about the clinical impact of glycemic variability.

Dr. Wisniewski: That’s perfect for the working group to address. There is a session later today where they’ll be debating that.


Increasing Adoption of Diabetes Technology: Better Products


Quynh-Nhu Nguyen, BS (Injections Systems Human Factors Specialist, FDA)

In her presentation, Lieutenant Quynh-Nhu Nguyen gave an overview of the FDA’s human factors expectations. She opened with a definition: human factors (HF) relate to the application of knowledge about human capabilities to the design and development of tools and devices. In 2011, the FDA issued an HF guidance; while the guidance has not been finalized, Lieutenant Nguyen commented that it reflects the Agency’s current thinking. In the human-machine interface, the machine delivers an output, which the human perceives, processes, and acts on. The human’s action serves as input for the machine, which in turn processes the input and provides the output to complete the cycle. Errors and failures, said Lieutenant Nguyen, can occur at any point in the cycle; thus, the FDA asks manufacturers to minimize the potential for error in the design of their product. She outlined the FDA’s expectations for HF data submission: 1) conduct a comprehensive risk assessment; 2) identify and mitigate use-related risks; 3) conduct human factors/usability validation testing on any strategies implemented to mitigate significant use-related risks; and 4) document the process in the Design History File. The extended review process for Insulet’s next-generation OmniPod (~18 months to date) underscores the importance that the FDA has placed on HF analysis and testing. As a reminder, OmniPod approval was delayed mainly due to the FDA’s concerns about potential areas of user error (e.g., incorrectly setting basal rates or the pump clock). Insulet modified the device’s software and submitted additional HF testing data to the FDA in late September; the company awaits feedback from the Agency. For more details on this ongoing example of HF considerations in product development, please see our Insulet 3Q12 report at https://closeconcerns.box.com/s/dvwhxbhw2i2ovmdkx6la.



Amy Tenderich, MA (Alliance Health Networks, San Francisco, CA)

Ms. Amy Tenderich’s presentation spoke to the value of social media in diabetes care, as she stated, “social media is a powerful shared experience,” that helps to fight against the isolation felt by people with diabetes. Individuals need support and information outside of the clinic, something that social media provides by enabling closer relationships and fostering community among people with diabetes, as well as encouraging patient led social advocacy. Directly addressing the company representatives in attendance, she made it clear that “even if you think you don’t have anything to do with social media, you do. We are all expressing opinions about products everyday…people are talking about the products you are creating and the studies you are doing, whether you are part of the conversation or not.” Ms. Tenderich continued to explain that social media has revolutionized communication between consumers and companies. Previously, companies had one-way communication, delivering a packaged message to consumers. Now, markets are interactive dialogues – “conversation is ongoing, spontaneous, and authentic.” She likened the online space created by social media to a return to the old market place, in which consumers and sellers were constantly and actively interacting. Recent statistics suggest that there is indeed a diabetes market online – major online diabetes communities (e.g., TuDiabetes) have more than 850,000 members, and 3,000 people on Facebook mention diabetes as one of their interests. People with diabetes are not the only ones who can benefit from the explosion of social media and the dynamic market place, companies can learn from consumers’ reviews of products to improve next- generation designs and better address people with diabetes’ needs. Ms. Tenderich believes that diabetes care stands to be improved by social media and the open dialogue between the users of diabetes products and the companies that manufacture them.



Quynh-Nhu Nguyen, BS (Injections Systems Human Factors Specialist, FDA) and Amy Tenderich, MA (Alliance Health Networks, San Francisco, CA)

Q: In human factors, is there an accommodation for any kind of learning curve? I see a trend towards wanting to have a patient open a box read instructions in short time and do a test perfectly with no chance to do it again. How does FDA feel about a learning curve?

Ms. Nguyen: From the perspective of users, it depends on the frequency of the injection. We ask sponsors to evaluate frequency of use and take that into consideration with regards to training. There is a lot of work that needs to be done before a validation test. We think a learning curve can be shown in a validation test. You can delay a validation test to see if users have retained training. So, if a sponsor claims there is a learning curve and a patient performs an injection twice a day, we ask them to demonstrate not just through first injection, but also through subsequent injections.

Ms. Tenderich: That’s where social media can help as well. My doctor wanted me to try to Bydureon. I went to YouTube and followed video instructions. A week later I forgot how to do it, and watched the video again. That’s so helpful. People talk about how Apple doesn’t come with user manual because products are so intuitive. You can watch instructions again in a two-minute video.

Q: I totally agree with the importance of human behavior. In this discussion about measuring validation, isn’t the ultimate measure simply patient adherence? Isn’t that more simple than a qualitative assessment?

Ms. Nguyen: It’s subjective to measure patient adherence. That’s why we’ve developed guidance for performance data and subjective data. A device could be easy to use, but in that process a user may commit an error so there needs to be adequate measure in play to address these errors. We need the sponsor to demonstrate that you can use the device and have a positive subjective experience.

Q: I’m a pediatric diabeteologist and one of my concerns about patient support groups is how representative they are. Is there a silent majority we don’t hear from? The patients who are getting on with life reasonably happy are quite silent. The people who talk may have an axe to grind.

Ms. Tenderich: I hope I haven’t given you the impression that this is about people wanting to gripe. That’s not what it is. Living with diabetes takes a huge emotional toll on people. You have to think about it all day long and it’s a burden. Following instructions and taking a pill is just this much of it, and then there’s everything else. It’s important for people to connect with others who are living with this. I think that helps us feel more motivated and wear this piece of equipment. You know you need to do it because it helps in the long run. The group of people active online is not a disgruntled minority, and it’s broadening to be more mainstream. This kind of support needs to be part of the prescription. People suffer with this in silence alone and then they try to reach out and so many hit rock bottom before they do. The Diabetes Advocates are working to get educators to realize these groups and these communities are there for patients. It’s about so much more than the griping. There is something still to saying we are unhappy about X, that’s important too. Before you were in a vacuum. You had no way of knowing for example if someone was able to get reimbursement. We’re human beings, there are people who are just grumpy and going to be negative. But there are many aspects to this. One of them is pushing for improvements where we see problem, but a lot of it is creating community.

Comment: I am in human factors from FDA. What we want to do is find the clinical context of the use and how patients are given instructions. If it is over the counter, they are going to use the information to do the next dose of insulin. There is not a lot of leeway for them to be wrong. So we have the expectation that the information is clearly presented. And if the device fails we need the user to be able to realize that. But if it is going to give a number that is wrong and the patient won’t realize that, we need to think about whether there needs to be a device or education redesign.

Q: I work at IDEO (a design and innovation consulting firm). Part of our job is to work very hard to design those user experiences, but what are the most effective ways we can make an impact in this process as we explore mobile, the cloud etc., and what kind of impact can we make? Should we be communicating with FDA?

Ms. Nguyen: I think its helpful to get back to the drawing board and see how mobile can benefit patients in disease management. You need to evaluate use scenarios and see what applications can provide and give information to FDA. If any questions are raised in the process they can submit to the agency via a presubmission application. We are working closely with the medical device industry to collaborate to put out safe products.

Ms. Tenderich: Patients should be communicating with FDA. We always encourage patients to submit comments when it’s called for, but if you’ve ever tried to find the link on FDA’s site, it’s so user-unfriendly. Making that smoother that would be great. On other end with health applications, we need to connect dots between Silicon Valley and health/mobile via smartphones. So many health applications are downloaded and then discarded; they don’t actually add value to patient’s lives. A lot of pharmaceutical companies think they need to have an application so they do, but it doesn’t really do something. An application with a lot glucose data points to send to the physician – physicians don’t have the time or reimbursement incentive to analyze that. What we’re working towards is algorithms so people can make use of data themselves. How can we provide more value with the technology we have?

Q: What is the FDA perspective on how we can implement best practices like more agile developments? And Amy, are you aware of any best practices or cases where products developed in a different way and therefore met patient needs better?

Ms. Nguyen: We anticipate that device manufacturers work closely with constituents to develop safe and effective products. We recommended to manufacturers to test with regular users so they can fix issues easily and less expensively. From my standpoint, I don’t work closely with consumer groups. But there are some constituents in FDA that do.

Ms. Tenderich: Products are the best when there was someone in the organization who had diabetes or a child with diabetes. I wear OmniPod; it’s tubeless. Glooko is another example, the founder wanted to create an easy way to take his glucose numbers to the doctor. The technology is not the solution but it can enable the solution.

Dr. Robert Vigersky (Walter Reed Army Medical Center, Washington, DC): I googled Victoza and got on your site. What I saw was a discussion about off-label use of Victoza and people having side effects or problems. What kind of medical oversight do you have? What liability do you engender for hosting these sites and what kind of warnings do posters get?

Ms. Tenderich: At Diabetic Connect we do have moderators. At Diabetes Mine we will post a correction if something is incorrect. We can’t argue with a person’s experience though. We don’t police it unless people say something offensive or clearly erroneous, then we will post a correction. People will shout each other down if people say something off base.

Dr. Vigersky: But who is saying what’s erroneous? Is it you?

Ms. Tenderich: This is not deep science; it’s a user forum talking about experiences. It’s a support forum. We’re not asking for deep insights that we report back to the pharmaceutical industry. Our site is more geared towards insulin delivery and monitoring devices. Again, clearly it’s a patient blog. When we post something published by us, it is vetted. If people post comments we may or may not agree with, that’s something they do and would do on any publication.

Dr. Heinemann: This is a discussion we need to have. My understanding is that I was ignorant about social media and that it can have huge impact nowadays. We need to have more of this world integrated into our meetings.

Q: How is the funding done?

Ms. Tenderich: Early patient community sites withered away because they couldn’t maintain the business. Manny Hernandez, who started TuDiabetes, went the nonprofit route. In my case, I joined the Alliance Health Networks and they acquired Diabetes Mine. I consider Diabetes Mine a publication, so there is banner advertising – it is an advertising model.

Q: I love what I see in online communities. What is the best way for those of us in industry to engage with a site like yours and do it in an appropriate way that lets us hear the early voice of customer and helps us do better? How do we bring that together?

Ms. Tenderich: Pay attention to social media. When something new comes out people who have it are posting reviews with feedback I think that is invaluable. Just paying attention is great. We’ve had guest posts by industry asking the community to comment on things – as long as they are open and authentic that they are from industry. I think just being a player and active participant makes a difference. In terms of actual processes, it’s a work in progress.


Increasing Adoption of Diabetes Technology: Better Marketing


Ronald Goodstein, PhD (Georgetown University, Washington, DC)

“You talk more about yourselves and your attributes than your benefits. Start talking about benefits… customers buy benefits!” exclaimed Dr. Ronald Goodstein during his crash course on product marketing. Further, Dr. Goodstein believes that by selling benefits, companies can get out of “price wars” with providers and payors. The easiest way to begin to do this, explained Dr. Goodstein, is to “change the language of your communications.” To illustrate that the field has room for improvement in this respect, Dr. Goodstein displayed several company product websites and offered his very honest opinions on how the sites could be improved (for which, Dr. Goodstein apologized with a smile). In particular, he told the audience that their website should not be set up to just be efficient at transactions (“you are medical partners in health after all”), not look like Dell advertising, and not mistake more text for better marketing. To drive the latter point home, he asked the room of ~60 people how many had read all of their respective companies’ marketing communications. Only three raised their hands. “If it’s not interesting to you, why would anyone else do it?” In order to make marketing (and products) better, he implored companies to face the customer, and view problems from each customer’s perspective. Your customers have more at risk by adopting your product than you do by selling it to them, he said. If they adopt your technology, their care is in your hands. He explained that companies need to build “better” not just different, meaning that products should be distinctive in a way that really matters for customers. We agree! By communicating benefits and by building “better,” Dr. Goodstein believes that companies can start to get real advocates (e.g., patients, providers, medical administrators) who will help the product sell itself.


Anton Petkov, MD (Sanofi, Frankfurt, Germany)

Echoing Dr. Ronald Goodstein’s message in the preceding presentation, Dr. Anton Petkov stressed the importance of marketing benefits, not just product features. While Dr. Petkov acknowledged that companies have made progress on this front in the past three years, he believes that companies have not adequately transitioned their marketing messages to focus on outcome-based benefits. One of the challenges in the medical field, according to Dr. Petkov, is that companies have to market to multiple target groups (i.e., patients, payors, providers) and each constituency has a different expectation for the technology. By identifying these expectations and appropriately targeting each group, Dr. Petkov thinks that companies can greatly improve their marketing strategy.



Kelly Close, MBA (Close Concerns, San Francisco, CA)

With the patient’s perspective in mind, Kelly suggested that industry engagement in social media should be motivated by the potential to fill unmet patient needs, rather than the potential to increase product adoption (though, the former does have the ability to accelerate the latter). With a growing presence of an online diabetes community (including people with diabetes, families, friends, and providers), companies have opportunity and reason to enter this market space in a meaningful way. By taking a patient-centric approach, Kelly believes that companies can utilize social media to: 1) make it easier for patients to give product feedback to companies (which in turn can improve next-generation products, and thus, improve both patient care and product adoption); 2) facilitate communication between patients and providers via telemedicine and data interpretation tools; 3) create applications that drive outcomes; 4) incentivize education and patient participation in online diabetes communities (and perhaps improve product reimbursement by incentivizing patient engagement in a way that fosters better adherence and outcomes); 5) work towards helping the neediest of patients; and 6) relay company views through widespread, active participation in social media. Kelly concluded that diabetes- related companies should engage in social media, but only after carefully considering the best way to become involved and always with the end-goal of sustaining and strengthening patient engagement in diabetes care. For a deeper delve into Kelly’s presentation beyond the summary provided here, you can view her slide deck at https://closeconcerns.box.com/s/zjh1ubve6o06avtwb08j.

  • Exponential growth in social media platforms has created the means for diabetes and healthcare companies to expand their online presence. The statistics are telling: Facebook touts over one billion monthly active users, a 26% year-over-year increase from 2011; the proportion of online adults using Twitter has quadrupled since late 2010; Pinterest gained over 10 million monthly active users in two years; and Instagram reached more than 80 million registered users in two years. While many companies have penetrated social media in significant ways (e.g. Starbucks touts ~3,000,000 Twitter followers), diabetes companies have a comparatively small presence (e.g., Novartis has 30,000 Twitter followers, the most of any diabetes company). While the market for diabetes products is smaller than for the leading consumer good companies on social media, Kelly maintained that diabetes companies still have substantial potential to both increase their social media presence and make their current presence more impactful.

  • Growing presence of an online diabetes community has provided the motivation for diabetes and healthcare companies to engage in the online marketplace. Kelly explained that the explosion of social media does not alone necessitate that companies expand into social media; however, given that the diabetes community is indeed turning to these platforms for diabetes information, there is reason for companies to explore how best to become a part of this space.

    • Data from a dQ&A patient panel showed that 14% of type 1 respondents (n=1,050) and 7% of type 2 respondents (n=2,985) rate online communities or chats as an important source of information about managing and living with diabetes. Eleven percent and 7% of type 1 and 2 respondents, respectively, rated blogs as an important source and seven percent of both people with type 1 and type 2 rated company websites as an important source. Kelly noted of course that doctors and CDEs were still the most frequently used source for information, underscoring their incredible importance is patients’ lives.

    • Further, Kelly showed that 14% of people with type 1 diabetes (of 1,322 survey respondents) and 5% of people with type 2 diabetes (of 3,504 survey respondents) have accessed diabetes content on Facebook. The caveat however, explained Kelly, is that people won’t engage with social media and online sites unless they “trust” it. Just 7% of people with type 1 diabetes and 8% of people with type 2 diabetes in the panel trusted diabetes content on Facebook; 5% and 8%, respectively, trusted diabetes content on Twitter. For more detail on and data from the dQ&A market research panel, please write richard.wood@d-qa.com.

  • Social media already benefits people with diabetes…Online communities (like tudiabetes, Children with Diabetes, Six Until Me, and Diabetes Daily) help prevent isolation by connecting people with diabetes (PWD) and their families with other PWD and their families. Diabetes- related blogs have helped to share best tips and practices, offer product reviews, and enable individuals to follow specific interests within the field (e.g., technology). Further, Kelly commented that social media has helped to raise awareness and support off-line meetings and events (like Walk to Cure Diabetes) as well as promote exercise and wellness in the community (the Big Blue Test was a prime example).

  • …and companies should make sure to interact in this space in a way that protects and increases these benefits. Kelly encouraged companies to carefully consider how they engage and market over social media, stressing that they should take a patient-centric approach.

    • What are patients’ and HCPs’ largest needs? Instead of asking how to best promote product adoption, Kelly proposed that companies ask how best to fill unmet patient needs. Upon recognizing a need, she emphasized that companies should address it in a way that protects and respects the patient. Further, she asked companies to evaluate whether they have filled the need effectively, a determination that is complicated by the multiplicity of constituencies affected (e.g., can the company evaluate value from the perspective of the patient? the company? payors? providers?).

    • Can social media expand communication between patients and providers? And how can companies facilitate this communication? Kelly suggested that expert involvement in social media can help ensure that information disseminated over this space is safe and valuable for PWD; while she noted that objections to HCP engagement include patient privacy concerns, time pressures, and lack of reimbursement, Kelly believes that HCPs who do participate and communicate with patients online are lauded. Perhaps, she suggested, HCPs could act as “curators” of online information, directing PWD to important and relevant sources of care information.

    • Can social media incentivize and motivate treatment adherence? Similarly, can companies use social media to sustain and strengthen patients’ engagement in their care? As Kelly said during the panel discussion, “we need to think creatively about how to increase adherence.” Certainly, companies also stand to benefit from increased patient adherence (as increased adherence often translates to increased product use).

    • What are healthcare companies’ obligations to patients? Kelly thought this might be best captured by a quote she shared from Ms. Dana Lewis (Founder, #HSCM [and former Close Concerns Summer Associate]): “It’s not about you, it’s about the patient.”



Ronald Goodstein, PhD (Georgetown University, Washington, DC); Anton Petkov, MD (Sanofi, Frankfurt, Germany); and Kelly Close, MBA (Close Concerns, San Francisco, CA)

Q: Maybe one company has evidence that their blood glucose meter is more accurate than others. What is the value of talking about accuracy?

Dr. Goodstein: There are different kinds of attributes in the market. There are must haves – having more of it doesn’t make the product better, but not having it would kill you. I would not make my value proposition strictly around accuracy because that will not make you different. If you’re not accurate, you’re out of the market, but being accurate is a commodity. Everyone is accurate so I wouldn’t start there. It’s important, but everyone offers it.

Comment: The FDA has a huge role in this. If you make a claim that is not in your label – even saying the grass is green – you’ll get in trouble.

Dr. Goodstein: You can’t do anything that’s off label. But you can do non-branded advertising. You can bring more people into the category and then it just so happens you’re the best within that category. There is more room to talk about research and ongoing studies. If you’re the leader and you’re bringing more people into the category, you’re going to win.

Q: But what about your benefits?

Dr. Goodstein: The only benefits you can talk about are the ones you’re approved for. Talking about what the attributes provide that are approved – that’s okay. Think about on-label discussions for broad based communications and off-label discussions for HCP initiated discussions. That would be the benefit of unbranded advertising.

Dr. Lutz Heinemann (Science & Co., Duesseldorf, Germany): Let me challenge you about blood glucose meters. I believe you have to explain to patients and physicians what the benefit is – for example, higher accuracy of blood glucose meters. So how should you advertise that?

Dr. Goodstein: The biggest issue I see for you guys is that your language and the customer’s language are completely different things. You talk about ease or flexibility, but what does that mean to a patient or healthcare provider? Flexibility for a patient is different than flexibility for a provider. What does accuracy mean? In terms of accuracy, consumers aren’t just looking for a number. How are the numbers trending? Can I get a monitor that shows trending patterns? There are other things they’re looking for in terms of accuracy. Look at their definition, not just yours.

Kelly: It is important to recognize that when you have 100 people with diabetes in a room, they have a diagnosis in common. Some might worry about accuracy, some might worry about time in zone, and some might be on GLP-1 and not have to worry about hypoglycemia as much. It is a constantly shifting sand box.


Bolus Calculators for Insulin Dosing: Physiology


Howard Wolpert, MD (Joslin Diabetes Center, Boston, MA)

Dr. Howard Wolpert’s presentation drew attention to attention to the limitations and assumptions that go into current carbohydrate counting-based insulin dose calculations. First, he questioned whether carbohydrate counting was even a realistic goal for patients considering the difficulty of the process. Second, he observed that carbohydrates are not the only dietary ingredient affecting insulin requirements. Third, he believes that too much focus is put on matching the insulin bolus with the quantity of carbohydrate intake; instead he suggested that the focus should be on matching insulin action with the rate of carbohydrate absorption. Dr. Wolpert then zeroed in on the effects of fat on insulin requirement. He presented a small, closed-loop study to demonstrate that dietary fat intake 1) increases insulin requirements; and 2) changes the appropriate carbohydrate-to-insulin ratio. He posed that higher-fat meals require alternative dosing algorithms with altered insulin delivery pattern and dose. While this may make a fixed dosing increase for higher fat meals seem attractive, he suggested that given the high inter-individual variability between patients, a fixed dosing increase would not be safe or effective. He also emphasized that different kinds of fats will likely have different effects on insulin sensitivity. Carbohydrate counting, proposed Dr. Wolpert, needs to be combined with considerations of whether fat and whether high glycemic index foods are consumed. Further, he believes that the approach for many patients really should be embedded in meal planning and nutrition education that will steer people with diabetes to meal and food choices with less glycemic impact.

  • In a closed-loop experiment, Dr. Wolpert explored whether differing fat intake in a meal affected glucose control and insulin requirements in patients with type 1 diabetes. Seven patients (7.2% average A1c; 0.50 units average total daily insulin dose) were randomized to receive either a dinner high (60g) or low (10g) in saturated fat with identical carbohydrate content. Protein content varied between meals slightly to match overall caloricintake. The study was conducted in the clinical research center; patients came into the center at 12:00 noon, had identical lunches, then regulated glucose by open-loop control until a 6:00 pm high- or low-fat dinner. From 6:00 pm until 12:00 noon the next day, patients used proportional integral derivative (PID) closed-loop control. Study arms received identical breakfasts at 8:00 am. In the crossover design, patients then went through the same protocol until 12:00 noon the second day, but received the second dinner option.

  • After consuming the high-fat dinner, patients’ postprandial glucose was higher (p < 0.0001). Notably, patients also had a different carbohydrate-to-insulin ratio following the high-fat meal than the low-fat meal (p < 0.01), determined by the number of carbohydrates consumed divided by total insulin delivered (which was higher following the high-fat dinner). There also appeared to be significant inter-individual variability for how the ratio changed; however, given that only seven patients were included in the study, the degree of variability needs to be further explored.



Carbohydrate-to-Insulin Ratio



Low-Fat Dinner

High-Fat Dinner

Percent Increase






























Ewa Pankowska, MD, PhD (Medical Academy in Warszawa, Warsaw, Poland)

In Dr. Ewa Pankowska’s presentation, she proposed an alternative method to calculating insulin dosing that incorporated carbohydrate, fat, and protein counting. She briefly discussed the shortcomings of normal boluses and suggested glycemic control could be improved without inducing hypoglycemia by 1) using a dual bolus comprised of a normal and extended bolus; and 2) increasing insulin quantity of the extended bolus by adjusting for fat protein units (1 FPU = 100 kcal from fat and protein). Normally, she explained, dual boluses are calculated based on carbohydrate counting and then divided by two such that half of the insulin is given as a normal bolus and half as an extended bolus. In Dr. Pankowska’s “Warsaw” protocol, the normal bolus is calculated based on carbohydrate counting (insulin ratio x number of carbohydrate units), while the extended bolus is based on fat and protein (insulin ratio x number of FPUs). She explained that fat affects late postprandial in a number of ways, potentially some yet discovered, that include: 1) slowing gastric emptying; 2) slowing carbohydrate absorption; 3) raising insulin resistance (from high concentration of free fatty acids); and 4) stimulating glucagon secretion. Certainly, insulin calculation is a complex area as persons’ glycemic response to foods is complex (i.e., different fats can have different effects) and likely replete with inter-individual variability as Dr. Wolpert suggested. While Dr. Pankowska’s methodology is a likely improvement above the current carbohydrate-counting used, we impress that there is no one-size-fits all protocol.


Bruce Bode, MD (Atlanta Diabetes Associates, Atlanta, GA)

Dr. Bruce Bode gave a comprehensive comparison of available bolus calculators with the underlying message that no two bolus calculators are the same. While there is widespread agreement that the safest way to reduce insulin stacking from correction boluses is to subtract the insulin on board (IOB) from what would have been given without IOB consideration, Dr. Bode described two differing camps that characterize how companies calculate food boluses with excess IOB: 1) subtracting IOB from a bolus; or 2) never subtracting IOB from a food bolus, but using negative correction factors if blood glucose isbelow a target. A comprehensive comparison among calculation strategies of five commercially available pumps’ bolus calculators (Medtronic Paradigm, Animas OneTouch Ping, Insulet OmniPod, Roche Accu-Chek Spirit, and Tandem t:slim) illuminated just how different calculation approaches could be, raising important questions of whether bolus calculators should be standardized in some manner and whether FDA has a role in doing so.

  • Dr. Bode began with a brief review of the history of bolus calculators. In 2002, Deltec Cozmo (from Smiths Medical) was the first insulin pump to come to the market with a bolus calculator. Medtronic shortly followed with its own bolus calculator and drove the Deltec Cozmo off the market after successfully suing Smiths for patent infringement. Today, said Dr. Bode, all insulin pumps have some form of a bolus calculator.

  • Importantly, bolus calculators are able to incorporate insulin on board (IOB) into bolus calculations to help mitigate the danger of insulin stacking, which occurs when insulin from multiple boluses is active at the same time. IOB is determined using either a curvilinear or linear method based on the pharmacodynamics data from insulin aspart (Novolog). (The calculators that use linear methods do so due to intellectual property concerns, he said.) While the data from aspart show that the insulin lasts on average for six hours, with minimal effect after 4.5 hours, he noted that on most pumps, patients can set the duration of insulin action to be anywhere from 1.5 to eight hours. Dr. Bode seemed to disagree with patients’ having this degree of control over the setting, as it would often mean that patients select durations that no longer correspond to the physiologic data.

  • The safest way to minimize insulin stacking for correction boluses is to subtract IOB from the bolus which would otherwise have been given; however, for food boluses when there is extra IOB, there are two differing approaches: 1) subtract IOB from a food bolus (always or if blood glucose is below a threshold value); or 2) never subtract IOB from a food bolus, but use negative (reverse) correction factors if blood glucose is below a target. Dr. Bode polled the audience for their take on the appropriate method: What should happen with extra IOB in bolus calculators for a food bolus?

    • Subtract from a food bolus if there is IOB remaining: 25%

    • Only subtract from a food bolus if blood glucose is below 70 mg/dl: 14%

    • Always cover a food bolus, but if below target, use a negative or reverse correction (i.e., [70-100]/3 = minus one unit from meal bolus): 33%

    • I do not believe in bolus calculators: 28%

  • Dr. Bode compared five commercially available pumps’ bolus calculators to drive home his takeaway message that no two bolus calculators are the same.

    • Medtronic Paradigm pumps never subtract IOB from a food bolus, and they use a target range rather than a single glucose value, such that hypoglycemic values will be corrected to the lower end of the range, and hyperglycemic values will be corrected to the upper end of the range. The pumps use reverse correction if blood glucose is below the lower target.

    • J&J (Animas)’s OneTouch Ping does not subtract IOB from a food bolus unless the blood glucose value is below target (which can be either a single value or a range); the Ping has reverse correction.

    • Insulet’s OmniPod only counts insulin from correction boluses toward IOB, (not from food boluses). The OmniPod subtracts IOB once correction bolus is zero, and corrects to a single target (with an option for a correction threshold such that the OmniPod will not calculate any boluses unless blood glucose exceeds the bounds of the threshold. Dr. Bode noted that the next-generation OmniPod would employ calculation methods more similar to the calculators in other available pumps. [Editor’s note: As of this writing, Insulet is still awaiting feedback from the FDA on the additional data submitted in late September related to its next-generation OmniPod.]

    • Roche’s Accu-Chek Spirit lets patient set an acceptable post-meal glucose by setting an allowable meal excursion amount and an offset time period when blood glucose should not fall due to food being absorbed. The meter corrects to a single target and calculates active insulin from both correction and meal boluses, but not snack boluses. Further, the meter recommends carbohydrates when blood glucose is below target.

    • Tandem’s t:slim does not subtract IOB from food bolus, unless blood glucose is below 70 mg/dl and does not use reverse corrections unless blood glucose is below 70 mg/dl.


Comparison of Pump Calculations


IOB Calculated from this Bolus

IOB Subtracted from this Bolus

Reverse Correction










Animas One Touch Ping
























+/- range)


Insulet OmniPod2













Medtronic Paradigm













Roche Spirit







Tandem t:slim







1.Unless blood glucose is below target; 2. Will not calculate any bolus below the set minimum blood glucose; 3. Corrects low BG to low end, high BG to high end; 4. Uses predictive dose curve; 5. Unless below 70 mg/dl

  • Dr. Bode closed his presentation with three audience response questions.

    • Which pump appears to have the most logical bolus calculator?

      • Animas OneTouch Ping: 3%

      • Insulet OmniPod: 6%

      • Medtronic Paradigm: 29%

      • Roche Accu-Chek Spirit: 11%

      • Tandem t:slim: 3%

      • None: 49%

    • Should all bolus calculators use the same rules and methodology for calculations?

      • Yes: 65%

      • No: 27%

      • I do not care: 8%

    • Should bolus calculators (rules and methodology) be regulated by the FDA?

      • Yes: 54%

      • No: 36%

      • I do not care: 10%



Arthur Sherman, PhD (National Institutes of Health, Bethesda, MD)

Dr. Arthur Sherman reviewed the pancreas’ physiological “strategy” for regulating postprandial glucose in the body, proposing that the islet cell’s mantra could be “get ahead of the curve, stay ahead of the curve, and don’t go it alone.” First, he explained that islets get ahead of the glucose curve with a first phase of insulin release wherein a calcium influx into the cell releases docked granules (the readily releasable pool of insulin [RRP]). Second, islets stay ahead of the curve with a second phase of insulin secretion: glucose metabolism resupplies the RRP for sustained insulin release. Third, Dr. Sherman emphasized that islet cells don’t work in isolation; they collect information from other cells in the body. He pointed to a number of important modulators that can enhance the process, highlighting protein kinase C (PKC) and cyclic AMP (cAMP). Furthermore, acetylcholine (from the vagus nerve) and free fatty acids (from adipocytes) act as upstream effectors of PKC. Dr. Sherman hypothesized that preabsorptive acetylcholine (acetylcholine released immediately following meal ingestion) acts through PKC to build up the RRP of insulin and enable a bigger first phase of insulin release. Meanwhile, GLP-1 from L cells, GIP from K cells, and GLP-1 from taste buds act upstream of cAMP release.



Arthur Sherman, PhD (National Institutes of Health, Bethesda, MD); Howard Wolpert, MD (Joslin Diabetes Center, Boston, MA); Ewa Pankowska, MD, PhD (Medical Academy in Warszawa, Warsaw, Poland); and Bruce Bode, MD (Atlanta Diabetes Associates, Atlanta, GA)

Q: Dr. Bode, do you think these calculators should be standardized or do you think it is an advantage for there to be differences?

Dr. Bode: I think the terminology should be standardized and the rules explored. It’s too confusing for physicians and then for patients who switch from one pump to another – they can do harm to themselves. So I think you need the same terminology.

Comment: We always say patients should get choices, but it’s a major confusion factor with these calculators.

Dr. Bode: And every doctor thinks they work the same.

Dr. Wolpert: While standardization is good, the assumption is that there are rules that always work and that’s not true. I don’t think we should blur subtleties by coming up with something that is too rigid a standard.

Dr. Bode: The terminology needs to be standard. But whether you cover fat or not, that’s a different story.

Q: Dr. Sherman, how do you address the interaction between islets?

Dr. Sherman: How do they communicate with each other? There are two ideas. The first is control from the outside from neural input. The second is communication through common glucose. If one islet secretes insulin, it affects glucose and thereby other islets. Both probably play some role.

Q: It is clear you need extra insulin for fat and protein. What is an easy way for a person to calculate that? Most people can’t tell you the number of calories. One unit of insulin per 100 calories seems too much.

Dr. Pankowska: It’s a huge obstacle. I have used this system in my clinic for a long time and we completely changed the education training. These are skills. When you want a driver’s license, you have to focus on new issues and you have to train. It’s a comparable situation. When one person gets type 1 diabetes, we focus on food estimation. I think that everyday we don’t eat too many different meals. We prepare similar meats and similar dinners, so the list is not that large. We train patients to create individual lists and then train them in the calculation. For the second comment, we haven’t had a problem with hypoglycemia.

Dr. Wolpert: The challenge from our data is that there is inter-individual variability. It’s difficult to say there is a standard increment to increase insulin by for fat and protein. I am concerned practically with the risk of hypoglycemia. The other big message is that given so much inter-individual variation and that it’s impossibly difficult to count carbohydrates and fat, I think the big lesson here on an individual basis is to look at food intake and glucose and see if they are related to fat. When you illustrate that for people, you can get people to change their eating behavior. It’s different from getting people to cut back on fat for cardiovascular risk reduction because that is more of an abstract goal.

Q: How should we compare and validate different algorithms?

Dr. Bode: The problem, as Howard says, is that people can’t count carbohydrates. We need people to use their calculators and then you can identify when the calculation works and when it doesn’t. Validation is hard for the FDA. If you put a target on the calculator that is set by the user, is the FDA going to say there is a difference when you set the target to 80 mg/dl versus 90 mg/dl versus 100 mg/dl? Then you would have to enroll tens of thousands of people. This went though the FDA without them paying attention to it in the past. That’s why we have differences right now in the market. Now they’re paying attention to everything. So how do you validate it? All you can do is validate the calculation from the algorithm.

Comment: That’s validating the algorithm, not the clinical outcome.

Dr. Bode: Is a massive study to compare carbohydrate counting versus calorie counting worth doing? What Dr. Pankowska did in her study was give more insulin and it worked. You have to set up a true study and it can be done. Right now Helmsley is the only one who would sponsor it. We won’t be able to convince industry to change these calculators unless we can get a bigger market and unless we have scientific proof that counting fat matters. We have data that the less calories you eat the better you do, the more you check glucose the better you do, and the more you bolus the better you do.

Dr. Wolpert: Very few people use bolus calculators when it comes to meals. They do it manually because they’ve learned it over time. We’re pretending that there’s a technological solution when it comes down to patients developing insight, using data, using CGM data, and figuring out how to dose.

Q: If you take people eating unsaturated fats, will that make a difference in your calculations?

Dr. Wolpert: That’s an excellent question – whether different free fatty acids have different impacts on insulin sensitivity. The other issue relates to carbohydrates. Carbohydrates from pizza are a challenge when you already have maximal insulin resistance from free fatty acids. In some of the meals we gave to people, if they ate carbohydrates earlier you didn’t see as profound an impact because carbohydrates were absorbed before free fatty acids were affecting insulin resistance. We’re fooling ourselves if any formula is going to give us the answer. I don’t think there is a simple solution here.

Dr. Pankowska: In our protocol, we use the same calculation whether the fat is saturated or unsaturated. Probably, we can expect lower insulin requirements when food includes unsaturated fat, but it’s important to consider the whole calorie count from the meal in insulin dosing.

Dr. Bode: That needs to be study. I don’t think it will ever get into a bolus calculator, but it will help clinicians advise people.

Q: What is the effect of insulin on insulin secretion?

Dr. Sherman: Certainly, insulin inhibits beta cell secretion. That’s a direct effect through insulin receptors.

Comment: We have to be careful about insulin inhibiting insulin secretion. We have never showed a direct effect of insulin on insulin secretion. Clearly in vivo it’s a different story, because as soon as you have insulin you are decreasing glucose and getting other effects. But I think it is still challenging to say there’s an immediate effect. Long term though is different than acute.


Bolus Calculators for Insulin Dosing: Clinical Applications


Patricia Beaston, MD, PhD (Medical Officer, Center for Devices and Radiological Health, FDA)

Dr. Patricia Beaston gave the FDA perspective on regulatory considerations behind bolus calculators, highlighting four important considerations. First, she explored clinical considerations including for whom the device was intended (physicians, CDEs, patients…) and what the source of glucose values would be. If values come from CGM, she explained, it would make the entire system a class three medical device, which would require a PMA as opposed to the 510(k) that bolus dose calculators currently require. She highlighted the importance of the user’s being able to identify when an error has occurred, and whether the user has the power to accept, reject, or change proposed values. Next, Dr. Beaston explored hardware considerations. Everyone wants a mobile platform, she said. However, not all platforms have the same operating systems; thus they require different coding. Moreover, she explained that the security on these platforms is important to evaluate as well (a recent focus of hers upon learning that insulin pumps could be hacked – “we are taking security seriously now”). Third, Dr. Beaston explained software considerations. Specifically, she emphasized that manufacturers must provide a rationale for their approach to the calculation. Once such support has been shown and the necessary information is in the user manual, Dr. Beaston believes it is up to the prescribing physician as to whether a certain pump’s calculator is good for the patient. Lastly, Dr. Beaston spoke to human factors and the importance of the device being tested in the population who is expected to use it.

  • “We’re not insensitive that patients don’t want to carry a lot of devices, but we want these devices to be used safely.” Dr. Beaston emphasized that incorporating systems on to one platform like a smart phone introduces error that is not present in a stand-alone system. Because mobile devices each have different operating systems, she explained that they could require different coding. Thus, in order to submit to the FDA, companies need to address how algorithms will run on specific platforms. Further, she said incorporating diabetes management systems onto smart phones introduces the potential for new software downloads on the phone to corrupt the diabetes software. Dr. Beaston seemed wary of putting diabetes management tools onto a new platform “until we get a better handle as to how softwares merge.”


Signe Schmidt, MD (Hvidovre Hospital, Copenhagen, Denmark)

Dr. Signe Schmidt gave a literature review of six randomized clinical trials that assessed the benefits of bolus calculators in terms of A1c and treatment satisfaction. The bolus calculators included in the studies ranged from the simple cardboard InsuCalc Wheel to the more advanced Accu-Chek Aviva Expert (a blood glucose meter with integrated bolus advisor). The study results were mixed, leading Dr. Schmidt to conclude that we don’t have solid evidence to either confirm or reject that bolus calculators have a positive effect. She noted, however, that the review had several limitations, including: 1) the broad span of calculators; 2) varying study populations; 3) differing treatment regimens; 4) differing carbohydrate counting training programs; and 5) lack of power. Furthermore, Dr. Schmidt commented that bolus calculators are limited by the accuracy of the information that the user provides. Bolus calculators, said Dr. Schmidt, may indeed improve metabolic control and may improve treatment satisfaction, but successful use is dependent on user skills. This underscores the need to find interventions that work in patients with lower levels of skill and diabetes-related education.

  • The InsuCalc Wheel is a pocket-sized cardboard wheel. The user dials the wheel to reflect the pre-meal blood glucose level and estimated carbohydrate amount to be consumed to see the recommended prandial insulin dose.

  • Calsulin is a handheld calculator that prompts the user to input the blood glucose level, carbohydrate ratio, number of carbohydrate grams to be consumed, and post-injection exercise level in order to calculate the required insulin dose.

  • Diabetes Interactive Diary (DID) is a telemedicine system that features a bolus calculator component and facilitates communication between the HCP and patient by text messages; DID is currently available in Italy and Spain.

  • The Diabeo Software can be uploaded onto smart phones and calculates the prandial insulin dose based on anticipated carbohydrate intake, anticipated exercise, and pre-meal blood glucose levels. While the system is not currently available, Dr. Schmidt noted that the company expects to launch the product in France in 2014.

  • Roche’s Accu-Chek Aviva Expert integrates a bolus advisor into the blood glucose meter. The bolus advisor function is only available in the EU.

  • Deltec Cozmo is Smiths Medical’s insulin pump with an integrated bolus calculator. Dr. Schmidt remarked, however, that the Cozmo was no longer commercially available.

  • Given the studies’ mixed results, Dr. Schmidt was hesitant to draw firm conclusions on the efficacy of bolus calculators. She believes that studies have yet to confirm or disprove that bolus calculators have an added positive effect in diabetes care.


Age Group


Effect of Calculator on A1c

Effect of Calculator on Treatment Satisfaction

InsuCalc Wheel









Not assessed

Diabetes Interactive Diary





Diabeo Software




Not assessed

Roche Accu-Chek Aviva Expert





Deltec Cozmo

Children + Adolescents



Not assessed


Frank Flacke, PhD (Medical Director, Devices, Sanofi)

Dr. Frank Flacke argued that there is a need for automated decision support systems, both for basal and bolus insulin dose calculations. He reviewed current commercially available systems; decision support systems for basal dose calculation were noticeably absent. (He commented that Hygieia has developed a basal/bolus insulin titration system; however, it is not yet available in the US. For greater detail on the company’s device, please see our September 7 Closer Look at https://closeconcerns.box.com/s/67xvtmbnmkv3qqum56my.) For bolus dose calculation, on the other hand, there are several available systems, including: 1) applications for smartphones; 2) Internet resources; 3) blood glucose meters with bolus calculators; and 4) insulin pumps. Dr. Flacke asserted that insulin dose calculators recommend more appropriate insulin doses (and therefore, enable better self care) than doses that are calculated manually. In a multicenter study, 205 patients with MDI manually calculated two prandial insulin doses based on one high and one normal blood glucose test result; bolus doses were also calculated by Abbott’s FreeStyle InsuLinx meter (note: the bolus calculator function is only approved in the EU). Sixty-three percent of doses manually calculated by the participants were incorrect versus 6% of doses calculated by the InsuLinx’s bolus calculator. While the discrepancy was telling, Dr. Flacke still believes that there is room for improvement in bolus calculator design. He proposed that future calculators should take into account exercise and sickness, give carbohydrate estimations, and allow users to give feedback on doses, which would be taken into consideration in future calculations.

  • Dr. Flacke polled the audience: which of the following parameters would you like to see incorporated as an automated feature into a bolus calculator?

    • Carbohydrate content estimation: 43%

    • Dose feedback: 30%

    • Exercise adjustment: 19%

    • Sick adjustment: 8%



Matthias Axel Schweitzer, MD, MBA (Head of Medical and Scientific Affairs, Roche Diagnostics GmbH)

Dr. Matthias Schweitzer explained that dosing insulin takes three steps: 1) testing blood glucose; 2) calculating insulin dosing; and 3) injecting the insulin. Dr. Schweitzer believes that industry stands to improve this process by automating the calculation step via bolus calculators. To this end, Roche currently offers the Accu-Chek Aviva Expert, which features a bolus advisor integrated into the meter (available in Europe only). The bolus advisor’s algorithm separates meal-centric insulin (i.e., insulin that is used to cover carbohydrates for a meal) from glucose-centric insulin (i.e., active insulin that is used to lower blood glucose levels), which Dr. Schweitzer explained allows for a more aggressive calculation of a correction bolus. He reviewed first results from the Automated Bolus Advisor Control and Usability Study (ABACUS), which assessed the value of the bolus calculator feature on the Aviva Expert. The study randomized patients with poorly controlled type 1 or type 2 diabetes on MDI to receive either standard MDI therapy or bolus advisor (BA) supported MDI therapy (i.e., the Accu-Check Aviva Expert). At six months, a greater percentage of patients in the BA arm achieved the A1c reduction target (>0.5% change from baseline) compared to those in the standard MDI group (p <0.01). The results were originally presented at EASD 2012. For more details, please see our discussion on page 141 of our EASD 2012 full report at https://closeconcerns.box.com/s/kt7rf3v6uy09x6t9ldke. We think that built-in bolus calculators certainly stand to add value by encouraging more appropriate bolus dosing; however, these calculators may have different value propositions for people with different levels of MDI and carbohydrate counting knowledge, or for people with different levels of motivation.

  • In the panel discussion that followed, Dr. Schweitzer posited that for those who are highly skilled, the bolus calculator may improve care by providing better convenience; for those who have the least skill, the calculator may improve care by giving patients confidence in their calculations. While not included in the slide deck during this ABACUS presentation, we note that when subjects were segmented by baseline MDI and carbohydrate competencies, A1c benefits were statistically significant only among patients with perfect baseline competency scores in MDI and carbohydrate counting. Benefits trendedtoward significance in patients who did not have perfect scores in either category, and were non- significant in patients with only one perfect score.


Frank Flacke, PhD (Medical Director, Devices, Sanofi); Patricia Beaston, MD, PhD (FDA, Silver Spring, MD); Matthias Axel Schweitzer, MD, MBA (Roche Diagnostics GmbH, Mannheim, Germany); and Signe Schmidt, MD (Hvidovre Hospital, Copenhagen, Denmark)

Dr. Robert Vigersky (Walter Reed National Military Medical Center, Bethesda, MD): Did you look at hypoglycemic episodes?

Dr. Schmidt: In our own study, there was no difference in the number of hypoglycemic episodes between groups. Few of the other studies actually reported hypoglycemic events but there is a study I did not mention in this review by Satish Garg – my reasons for not mentioning it was because the bolus calculator is not commercially available. The patients using the bolus calculator did experience more hypoglycemia; the investigators concluded it was because that bolus calculator did not have an insulin on board function.

Q: When a new drug is approved, an important part is the package labeling, which provides a lot of information with regard to the mechanism of action. With regard to clinical decision support algorithms, is there anything equivalent in terms of an algorithm description?

Dr. Beaston: The standard is that they are supposed to put it in user manual and it is up to the physician to look at what they report. If a company wants to come up with something complicated and keep it confidential, my expectation is that they would provide additional data for our review for what they want to advertise and make claims based on. Right now bolus calculators are straightforward. You can hand calculate and see if the bolus calculator does what you want. We don’t regulate it, and I don’t know that we would ever regulate what the bolus calculation should be. The community can certainly petition the FDA and we can listen to those kinds of requests and act on them if it’s within our regulatory roles.

Q: When you use the integrated meter, are you able to switch off the bolus calculator function and just use it as a meter, or is it mandatory for patients to look at carbohydrate ratios?

Dr. Schmidt: In the study we did?

Q: Yes. I was surprised you didn’t get good results. Were patients just using the meter rather than the bolus calculator portion of the meter?

Dr. Schmidt: Patients used the calculation function and used it at about 80% of meals. No, they did not just use the meter for glucose measurements. I’m surprised too that we did not get better results. It was a pilot study, so power may have been a problem.

Dr. Schweitzer: You have to be careful in interpreting studies. Sample size calculation may not be optimal. This is a slide I did not show, but if we compare the three populations in ABACUS – those who were skilled in MDI and carbohydrate counting, those with moderate skills, and those who were really bad – it looks like those who were really good and those who were really bad benefit from a bolus calculator. The moderately skilled individuals – not so much. Bolus calculators may not be for everyone. With people with really good skills it may be a convenience. For those with really bad skills, maybe they can do it for the first time with confidence. We need to assess this with more and larger studies.

Dr. Beaston: When we look at studies, sometimes it is just the study effect. If you enroll patients who never had optimal education to begin with and give them new information or additional attention, sometimes the control group performs really well. Sometimes we recommend they do a run-in period to get everyone to the same level and then do the intervention.

Dr. Lutz Heinemann (Science & Co., Duesseldorf, Germany): There is a risk of confusing patients with so many different bolus calculators with differing terminology. I see a certain risk to glucose-centric insulin as an additional feature.

Dr. Schweitzer: The principle behind it is the important thing – separating insulin given for carbohydrates and insulin given to correct glucose level. These are two different things. We have indications that this is easy to learn, and to an extent intuitive, and we plan to test that. This separation makes more sense than one bolus to cover everything.

Dr. Heinemann: With all these different terminologies or aspects of bolus calculators, I have to confess I am confused here now. How can we progress, how can we have a definite study? Can lots companies set up one study?

Dr. Beaston: Just so you understand, companies come in with their calculators and unless it’s a novel calculation, our role is to make sure the information provided to patients and providers is clear and that the algorithm is correct. We don’t regulate the practice of medicine or ideas. It’s up to the community to say what they want and what their expectations are.

Q: What is the effect of how well clinicians set up initial bolus calculator parameters. And for longer studies, how often do you look at and adjust these parameters?

Dr. Schmidt: The healthcare provider involved in a patient’s treatment is important and it is important they continuously update the devices. It is also important to train the patients to be critical about their own therapy.

Dr. Schweitzer: I personally think we have to go much more into detail in those studies and have studies with enough patients. There is not one study providing the truth. We need many well-performed studies, even with different designs, to bring us closer to truth.

Dr. Schmidt: Also, if you don’t set up the bolus calculator correctly, then patients may lose confidence in it and stop using it.

Comment: Dr. Heinemann, I don’t think we can come to a conclusion on whether bolus calculators work. I agree that we need to dig much deeper. We need to say that we need this option for those who need it. I think we cannot come to a clear conclusion on whether it works, but we need to have this option because I am not at all doubting that it works for some patients.

Dr. Heinemann: I’m asking for the class of products to be tested in different patient groups and to avoid having a lot of small inconclusive studies. I am saying we aim for a better quality of evidence.

Dr. David Klonoff (University of California, San Francisco, San Francisco, CA): I am concerned these studies have flaws in them and that they may not be as valuable as we think. This happened with SMBG in type 2 diabetes. Many concluded that there was no benefit, but when the studies were analyzed, there were a lot of wholes in them. There are only three good studies and they showed benefit. I am concerned that some of the studies with bolus calculators have flaws. I think hypoglycemia needs to be looked at from the beginning, and what blood glucose monitor was used to base decisions off of matters – if it’s an inaccurate monitor, there is going to be some scatter. I think it is useful to set up the hypothesis in advance to see which groups benefit the most. I am not convinced by post- hoc analysis. We can’t do that other than to suggest a future hypothesis. I just can’t believe that bolus calculators are useful for nobody.

Dr. Vigersky: Pumps have automatic insulin on board of course, where as with MDI you don’t. Sanofi is the only company that makes insulin and has a meter. Can we communicate between a pen and a meter and have insulin on board recorded as accurately as for a pump?

Dr. Flacke: Great question. Right now, the pens that are out there have limited capacities in terms of data capturing. It would be a good to deal with that and come up with a pen that can close the gap for MDI patients.


Live Demonstrations


Kong Chen, PhD (National Institutes of Health, Bethesda, MD)

After a brief, whirlwind overview of modern technologies for monitoring activity, Dr. Kong Chen demonstrated a new accelerometer that he thinks especially promising: the ActiGraph, manufactured by a company of the same name. The ActiGraph captures motion in three dimensions, and it also senses light and sleep efficiency. ActiGraph’s software can display retrospective or real-time data; it can also integrate information from other sensors (e.g., heart rate monitors, GPS). So far the major application for such technologies seem to be in research settings, but Dr. Chen forecasted that combined physiological sensors is the “wave of the future” – hopefully this wave will promote smoother sailing for diabetes patients trying to maintain tight glucose control.

  • Although Dr. Chen demonstrated the ActiGraph interface on a computer, he said that mobile phones could be a good platform too. (He noted that the accelerometers built into modern phones work well, but he cautioned that the results could easily be confounded – if people shake their cell phones while playing video games, for example. Thus he favors using a standalone accelerometer like the ActiGraph, with the mobile phone used only for data display and/or storage.)



Breanne Everett, MD (Orpyx Medical Technologies Inc., Calgary, Alberta, Canada)

The statistics on diabetic nephropathy that Dr. Breanne Everett presented were incredibly depressing: over 60% of people with diabetes will suffer from diabetic nephropathy; 100,000 limbs are lost to diabetes each year in the US; 25% of people with diabetes will have a minor foot injury develop into an ulcer; over half of foot ulcers will require hospitalization; and one in five foot ulcers will require amputation. Believing that “awareness is the best medicine,” Dr. Everett founded Orpyx Medical Technologies and developed pressure-sensing insoles, called The SurroSense Rx, to reduce the number of diabetic nephropathy cases that turn into foot ulcer cases. The insoles provide real-time feedback to the wearer via a mobile application. During a simulation of the device, Dr. Everett showed how the insoles assess pressure overtime and indicate (by means of color changes on a foot map in the mobile application and an accompanying alert) when pressure has exceeded safe levels. Then, the application provides instructions to alleviate the pressure and notifies the user when he or she has successfullyoffloaded that area. A pilot study assessing The SurroSense Rx is expected to commence January 2013; a patent is pending for the device.

  • Dr. Everett outlined 12 steps to prevent diabetic foot ulcers: 1) ask your doctor to test you for diabetic neuropathy; 2) check your feet daily for cuts, blisters, redness, and swelling; 3) wash your feet daily, and dry carefully between your toes; 4) cut your toenails straight across; 5) keep your feet warm and dry; 6) wear loose fitting socks to bed, and never use heating pads or hot water bottles; 7) wear comfortable shoes; 8) never walk barefoot; 9) don’t smoke (because it reduces blood circulation); 10) ask your podiatrist about orthotics and insole options; 11) never perform “bathroom surgery” (i.e., removing calluses); and 12) work with your doctor to control your diabetes.


Larry Katz, PhD (LifeScan, Inc., Wayne, PA); David C. Klonoff, MD (Mills-Peninsula Health Services, San Mateo, CA); Breanne Everett, MD (Orpyx Medical Technologies Inc., Calgary, Alberta, Canada); Kong Chen, PhD (National Institutes of Health, Bethesda, MD)

Dr. Cynthia Marling (Ohio University, Athens, OH): Your alerts appear when a pattern is detected, but in order to take action the patient has to consult the Tools for Life Pattern Guide. That guide looks paper-based. Why not just put the recommendations on the meter?

Dr. Katz: We are thinking about this. There isn’t a one-size-fits-all. There is a lot of information packed into this guide, and only so much space on the meter. But maybe we should have more information on the meter, for patients that don’t want to carry the book around with them.

Dr. Marling: The clinicians I work with don’t use paper logbooks; electronic logging seems much better. What is the prevalence of paper logbook use?

Dr. Katz: Although uploading is wonderful, LifeScan is working with software and tools to help utilize information. There are a low number of offices and people that download data. Really, the IQ meter is about how we can get the patient to have some sort of self-awareness between the office visits. You don’t see the doctor for very much time during the year. This allows patients to be more engaged and more self- aware between visits.

Q: I’m from Kaiser. What happens if patients don’t tag the readings?

Dr. Katz: They don’t have to tag if they feel that it’s not before or after a meal. For the algorithm that alerts them to patterns, the low pattern can happen anytime – it does not need to be tagged. The high pattern has to be tagged pre-meal. We decided not to incorporate the pattern alert technology for post-meal highs; there are lots of things that could cause post-meal highs. We wanted to instead focus on pre-meal highs.

Dr. Klonoff: How easy or difficult will it be to transform the activity information you described into a form that can be applied to an artificial pancreas?

Dr. Chen: How accurate do you want the energy expenditure to be? I am an engineer and can make it happen; it’s a question of how sophisticated the system needs to be and what sorts of data are required. I think that is the conversation we ready to have now.

Dr. Klonoff: I will put you in touch with the algorithm engineers. Are there endpoints you expect to prevent?

Dr. Everett: what we can do in three months is probably limited in terms of clinical monitoring. We monitoring quality-of-life metrics and might see something there. We are measuring new incidence of ulcers; we are not enrolling patients who have active ulceration. We will also look at hospitalizations, mortality, and overall use of healthcare resources. One of the main things is just to get some pilot data. I doubt that we will see much in terms of clinical outcomes, but it could lead to our next study, in which we plan to enroll 220 patients and collect clinical data over a year.

Q: Hello, I’m from Dexcom. For Dr. Everett, how often do patients get an alarm to offload pressure on their feet? How did you decide on that?

Dr. Everett: We built the device so time and pressure thresholds can be changed. The device could be annoying to patients. When we actually do the study, we’ll be able to see all of this and track it over time. We will be able to see what action patients took, how frequent the feedback is, and how that makes a difference to the behavioral patterns of patients.

Dr. Patricia Beaston (FDA, Silver Spring, MD): For the accelerometer, does it require training? What about if you’re typing or riding in a car? And for the shoe inserts, do patients have to do any kind of training?

Dr. Chen: There is no training for the subjects. What we do on the model side is pattern recognition – is this a car ride vs. walking. All we have to do is program it and they wear it. Accelerometers are being used in NHANES studies. They are collecting data from over 8,000 people. These devices collect the raw data. We do the modeling post-hoc.

Dr. Everett: For the insoles, there is no calibration on the patient level. All of the output and alerts are pressure and time thresholds. The goal is to measure capillary pressure between the foot and the insole itself. For the second tactile device, there will be calibration – they will feel a sensation on the back and it will be calibrated.


-- by Adam Brown, Kira Maker, Joseph Shivers, and Kelly Close