We are focusing our gifts in two broad areas:
Transform Day-to-Day Treatments. Fund technologies that improve the ease and accuracy of care.
Develop A “Practical Cure”. Fund innovations offering a medium to long-term respite from insulin injections.
Our goal is to provide capital to exceptionally talented leaders who treat financial stakeholders as true partners, share openly about both progress and challenges, and have a clear vision of how their work will successfully improve the lives of Juvenile Diabetics.
Grant Strategy #1: Transform Day-to-Day Treatment.
We are funding innovations that offer a technological leap forward in day-to-day diabetes insulin therapy. These projects have a higher probability of success and a quicker path to commercial viability than other projects we might fund. Most projects within this portfolio utilize computer intelligence to simplify care and will hopefully fully automate diabetes care entirely one day. We hope to see these innovations commercially available within five years.
A. Closed Loop control
“Closed loop control” insulin management is an automated system to manage all aspects of a diabetic’s insulin therapy using computer intelligence. The most common design incorporates two existing technologies, an insulin infusion pump and a continuous glucose monitor (CGM), both of which have been FDA approved for more than ten years. The newest innovations use advanced computer algorithms that have the analytics and “intelligence” to make more accurate insulin therapy decisions than a human can. When complete, these new software solutions will automatically receive the patient’s blood glucose reading from his CGM, will analyze the information and automatically communicate any necessary insulin infusion adjustments to the patient’s insulin pump. Most of this closed loop control software currently runs on a smartphone (such as an Android or iPhone) that serves as the “brains” to precisely, continuously and automatically control insulin delivery based upon real-time CGM blood glucose readings. If successful, a truly closed loop control system has the potential to:
- free the user from the day-to-day insulin therapy decision making,
- eliminate user related mistakes and errors,
- provide more accurate insulin therapy management,
- keep the user within desired blood glucose ranges longer,
- reduce the occurrence, duration and severity of hypoglycemia and hyperglycemia,
- dramatically lower the possibility of devastating complications, such as blindness and kidney failure, with long term use because of tighter glycemic control, and
- reduce the time and costs associated with physicians managing their patients’ day-to-day insulin therapy.
A fully-developed closed loop control system will function like an autopilot – automatically receiving CGM blood glucose readings, interpreting the information and adjusting insulin dosing accordingly. The ultimate goal is for this system to be completely automated with minimal human interaction. The theory is, and to some degree the clinical evidence so far supports the contention that, these computer algorithms make better insulin management decisions than humans do. And while the closed loop control prototypes today represent a transformative breakthrough, we believe, in the long run, that an implantable closed loop control device would be one of the largest innovations since the discovery of insulin. At this point, no one is seriously discussing the prospects of such an implantable device. Nevertheless, Medtronic did, at one point, have an implantable insulin pump in clinical trials in the U.S. – therefore the basic proof of concept groundwork has already been laid.
There are a variety of closed loop systems in development. Some projects are more ambitious than others. There are three basic approaches:
- Low Glucose Suspend (LGS) which manages insulin delivery during the nighttime (when most diabetics are prone to severe hypoglycemia);
- The Artificial Pancreas which manages insulin delivery automatically and continuously both day and night; and
- The Bionic Pancreas a dual hormone system which uses both insulin (to reduce blood sugar) and glucagon (to increase blood sugar) to provide a fully automated insulin management system.
B. Low Glucose Suspend
The Low Glucose Suspend (LGS) system uses software algorithms that automatically shut off insulin delivery if the user’s glycemic concentrations reach low levels (or, when such a system is fully optimized, to shut off when the patient’s blood glucose levels are forecasted to go low). The LGS system is likely to be the first automated insulin management system to reach the U.S. market. The LGS is, in our opinion, more of a safety feature than a comprehensive insulin management system. LGS merely stops the flow of insulin when its software system detects or predicts the patient is prone to hypoglycemia. This is clearly an important and useful function because most diabetics have severe hypoglycemia at night while they are sleeping – approximately 60% to 75% of hypoglycemia episodes occur during this time.
Large companies already have LGS products on the market in Europe. Medtronic, the largest supplier of insulin pumps in the U.S., has the only LGS system available in the market, the MiniMed 530G. The system was available in Europe four years before final U.S. approval in October 2013.
CGM limitations. As discussed in more detail below, most CGM’s on the market today are not accurate enough to serve as the patient’s primary means of measuring blood glucose; nor have been approved by the FDA for this use.
Insulin delivery lags. Insulin delivery to the bloodstream has a delay of between 30 and 90 minutes. With any “insulin on board” the typical patient’s blood glucose will continue to trend low even after the initial LGS indications because of the delay in the insulin absorption. As a result, until faster acting insulins are developed, an LGS system is likely to have limited functionality for patients with much insulin on board or those with long absorption delays.
Because of the limited functional use (controlling low blood glucose) and because large medical equipment companies already have a product on the market in Europe, we have decided not to fund LGS projects.
C. Artificial Pancreas
When commercially available we envision the Artificial Pancreas as a fully automated insulin management system. The system will consist of an insulin infusion pump controlled by sophisticated software algorithms using real-time CGM blood glucose readings. In hospital-based clinical trials, these algorithms have been demonstrated to provide improved glycemic control with substantially more time within desired blood glucose control ranges and fewer instances of hypoglycemia and hyperglycemia. Like an autopilot, the Artificial Pancreas will be capable of interpreting vast amounts of data and making informed decisions.
Most of the current Artificial Pancreas systems in clinical trials are not fully automated. Various funding sponsors favor a more limited and incremental approach to overcoming perceived regulatory barriers. An example of this incrementalism is a platform that makes recommendations for the user to accept or modify (an “advisory system”), but does not automatically infuse insulin during the day. At night these systems operate as an LGS system. We think this solution offers too few benefits to be widely adopted in the marketplace. We prefer to support a bolder development plan seeking FDA regulatory approval of a fully automated system.
The main charitable sponsors believe that the FDA will be reluctant to approve a fully automated Artificial Pancreas system without a series of clinical trials with each demonstrating small incremental improvements of the previous trials – for the following reasons.
First, insulin is one of the most toxic drugs on the market. An inadvertent overdose can lead to severe hypoglycemia, coma and even death. Nevertheless, users can already inadvertently overdose using an infusion pump incorrectly.
Second, allowing a computer to automatically and independently infuse a toxic drug would set a unique precedent by the FDA. On the other hand, the use of computers in infusion pumps and the surgical suite are well established in hospitals and use of autopilot features in commercial aircraft is also widely adopted.
Third, the use of a smartphone as a medical device would set an important precedent for the FDA. We believe that it will be a considerable challenge to obtain FDA approval of a fully functioning consumer smartphone as a medical device. It will be impossible to prove that no smartphone application today or in the future will interfere with the Artificial Pancreas decision-making – after all one can’t disprove a negative. We believe there are technological alternatives to overcome the smartphone concerns.
Despite these challenges we believe it is more cost effective to seek comprehensive approval today rather than embarking on a series of independent clinical trials to prove smaller incremental improvements where each trial is likely to cost tens of millions of dollars. And therefore we favor a more ambitious approach of seeking FDA approval for a fully automated Artificial Pancreas. It will have the collateral benefit of allowing a fully functional Artificial Pancreas device to reach the marketplace faster.
Some of the original inpatient pilot-feasibility trials were conducted in 2010 and 2011 using a laptop computer to run the algorithms controlling insulin therapy. Patients were carefully monitored and controlled in the hospital as the system managed their insulin. These original tests demonstrated safety and efficacy of the Artificial Pancreas concept (at least under tightly controlled hospital environment) and, as a result, the FDA approved further clinical trials. In late 2011 researchers began developing a mobile Artificial Pancreas system. The original concept was for the system to run on a smartphone, such as an Android, which now has the computing and memory to support the Artificial Pancreas software system. In order to move through FDA approval faster, the Android phones were stripped of all consumer features except basic communications and the Artificial Pancreas software system. In late 2011 and 2012 the first outpatient pilot-feasibility studies took place using this platform.
The Android driven closed-loop system is headed into the third leg of its out-patient transitional studies in the second half of 2013. These transitional studies are designed to test the system under a “real-life” environment that most closely resembles a patient’s day-to-day life. Pivotal trials, the last step before market availability, are expected to start sometime in 2014 and go through 2016.
Artificial Pancreas “app”. The latest infusion pumps have better computing power and larger memory storage capacity. We believe it is now possible to have the Artificial Pancreas “ported” to run directly on these more powerful pumps. Since an infusion pump is already an approved medical device, the path to commercialization should be quicker than attempting to approve a smartphone, a consumer electronic device, as a medical device.
Full automation. The various Artificial Pancreas systems require user input to operate; full automation is not yet available.
CGM Accuracy: Most CGMs are still not as accurate as finger stick blood glucose meters. Since the CGM will generate the glycemic information used for Artificial Pancreas system decision-making, there are risks and uncertainties including whether regulators will allow the expanded reliance on a CGM or whether extensive finger stick measurements must be incorporated. Other technical issues with CGM’s are error readings, attenuation of the signal, and for some systems interaction with common over-the-counter drugs (especially Tylenol).
Faster Insulin Delivery: Since insulin is being injected subcutaneously in the skin, there is a time delay between injection and the activation of lowering blood glucose. To address this problem a number of studies have been launched to find faster insulin absorption methods. This includes implantable insulin delivery devices, inhalable insulin, heating patch at the site of infusion and new insulin formulation. Each insulin user has a different insulin absorption rate – some as quick as 30 minutes after injection and some as long as 90 minutes. This absorption variability causes obvious technical challenges for the Artificial Pancreas system.
JDRF and NIH made a joint multi-year commitment of $5 million to fund clinical trials at the ten universities called the Artificial Pancreas consortium (which is lead by the University of Virginia team).
D. The Bionic Pancreas
The Bionic Pancreas is the creation of Ed Damiano, Associate Professor of Biomedical Engineering at Boston University. Dr. Damiano is the father of a teenage son with Juvenile Diabetes. Dr. Damiano has a distinctive, if not bolder, vision than other researchers. In fact, to distinguish his device he has branded it the Bionic Pancreas.
The Bionic Pancreas is similar in many ways to the other Artificial Pancreas systems because it uses sophisticated control algorithms using automated inputs to continuously and automatically control an insulin infusion pump. But Damiano’s system is different in many important respects including the following:
Dual hormone control. The Bionic Pancreas dispenses both insulin (to lower blood glucose) and glucagon (to increase it). This is truly unique. With the ability to inject glucagon, the Bionic Pancreas has the ability to cause a user to “recover” from rapidly falling blood sugar levels and, as a result we believe, Damiano’s device could be safer than an insulin-only Artificial Pancreas. This rescue feature could, we believe, result in the Bionic Pancreas moving faster through FDA approval, at least in comparison to another fully automated single-hormone device. The dual hormone system currently provides tighter glycemic control than the insulin-only systems. Users also have fewer and less severe hypoglycemia events.
iPhone and Separate hardware. Damiano’s software currently runs on an iPhone for clinical trials but Ed has plans to develop a separate piece of hardware that will contain the Bionic Pancreas software. This is to overcome the FDA’s concerns of having a consumer grade smartphone as a medical device.
Fully automated. The Bionic Pancreas requires the user to only input his weight; no other user settings or user input is necessary. This is truly unique from other Artificial Pancreas systems that require the user to input several diabetes care statistics in order to initialize the system. Similarly, no “meal announcements” (or boluses as they are known) are required; the Bionic Pancreas system figures out when you have eaten and adjusts insulin accordingly. In fact, after entering the user’s weight, no other user interaction is required for the system to function properly. Glycemic control is remarkably tight with no user interaction, but is somewhat (but not statistically significant) improved if the user makes “meal announcements” which are completely optional in the Bionic Pancreas system.
Open loop mode. As a backup, if there is a technical problem with the device or if the user is sick, then the Bionic Pancreas can run in “open loop” mode, essentially offering the same manual control that is currently available on a standard insulin infusion pump.
Tight glycemic control. The Bionic Pancreas provides the tightest blood glycemic control available, so called control-to-target. In many respects this is a direct result of Dr. Damiano’s dual hormone design. Because the system has an imbedded “rescue” through glucagon, the administration of insulin therapy can be more aggressive.
Dr. Damiano received FDA approval for a series of more extensive clinical trials using a fully mobile and wireless Bionic Pancreas system during 2013. The first of these trials began in February and continued for approximately six months. The patients stayed overnight in a Beacon Hill hotel but were be free to pursue normal day-to-day activities during the daytime. Each patient’s trial will last for approximately two weeks. During the summer Dr. Damiano ran further clinical trials at a diabetes summer camp, sponsored by the Helmsley Trust. During 2014, Dr. Damiano will conduct another summer camp study in the pediatric population and a larger multi-center hospital staff study in the adult population. This last study will serve as the last transitional study before moving to pivotal trials.
Common Artificial Pancreas limitations. Damiano’s system has many of the same limitations imposed on other researchers: limited accuracy of current CGM’s, limited ability of CGM’s to communicate directly with insulin pumps and the possible regulatory issues surrounding the use of a smartphone as a controller device. Ed is working on all of these challenges but each represents a formidable barrier.
Glucagon limitations. No stable, extended-life liquid glucagon exists today. The current liquid glucagon is usable for less than a day unless refrigerated. In order for the Bionic Pancreas to work, Dr. Damiano will need a liquid glucagon formulation lasting for at least one week at room temperature. At least three small companies are working to develop such an extended-life liquid glucagon formulation. These new glucagon formulations will require clinical trials and FDA approval prior to their incorporation into Dr. Damiano’s human clinical trials.
Dual chamber infusion pump. A dual chamber insulin infusion pump does not exist today. Tandem Diabetes has plans to develop such a device but no prototype yet exists. JDRF announced in 2013 that they would fund Tandem’s development of the dual chamber pump. Such a development is subject to all of the usual development and regulatory approval risks.