DL: Double Intraperitoneal Artificial Pancreas

The project aimed to create a fully automatic control of insulin delivery in people with diabetes type 1 (DM1), i.e. a so-called artificial pancreas (AP). Patients with DM1 are completely dependent on correct insulin doses to get acceptable glucose control. Too high glucose lead to late complications such as cardiovascular disease, kidney disease, changes in the fundus of the eyes and effects on the nervous system. To inject correct insulin doses, it is also necessary to know the patient's glucose level. In the past this was done with a small drop of blood from the finger many times a day, but in recent years continuous glucose monitoring (CGM) have become the most common. CGMs measure glucose continuously via a thin sensor in the subcutaneous tissue. Many patients today use a combination of CGMs and insulin pump. By connecting these two units using an advanced mathematical formula, a so-called algorithm, information from the CGM can be used to control insulin delivery. This requires robust and reliable glucose measurements. Internationally, there are now several commercial solutions for such an AP, but all these solutions have their limitations. Among other things, the patient still has to report how much carbohydrates are consumed so that the system can calculate the insulin dose to be given. This means that the patient must still keep a daily focus on his illness and its treatment. The ultimate goal of our project was and is a completely self-managed system, i.e. a fully automatic AP without the need of daily involvement of the patients. This is very demanding since our ambition is also a completely normal or close to normal glucose level without serious episodes of too low glucose, so-called hypoglycaemia. The main challenge is that it takes a long time, at least three quarters of an hour, from when mealtime insulin is injected under the skin until the main effect on glucose occurs. In addition, the glucose change in the skin is slow; from blood to subcutaneous there is a delay of at least 6-7 minutes. In an automatic steering mechanism, this is problematic - it is almost like driving a car where the wheels turn an hour after you have turned the steering wheel and the windscreen is replaced by a camera that shows the road in front of the car with a 10 minute delay. Our original solution was to put insulin and measure glucose in the abdominal cavity, i.e. between the intestines. From here, insulin is absorbed twice as quickly as from the subcutaneous tissue. We, and others, have found that sometimes the time delay in glucose level from blood to the abdominal cavity is only a few seconds, i.e. much faster glucose changes compared to subcutaneous. Computer simulations show that faster insulin action and glucose measurements in the abdominal cavity will theoretically give approximately the same glucose level as in people without diabetes. Therefore, we worked with what we called a double intraperitoneal solution for an AP. We have now carried out experiments in pigs, and with a slightly different glucose measurement technology, we have found that the time delay for glucose measurement in the abdominal cavity is probably significantly longer than what we - and others - have previously observed. At the same time, we have studied the effect of glucagon, a hormone that increases glucose levels, injected into the abdominal cavity in rats and pigs. We then found that the effect on blood glucose was started a little earlier than if glucagon was injected subcutaneously and that the effect of the same small amount of glucagon was somewhat greater when glucagon was injected into the abdominal cavity. We then carried out our first acute trials (<24 hours) on pigs under anesthesia which suggested that with an AP with continuous glucose measurement in the subcutaneous tissue and delivery of both insulin and glucagon into the abdominal cavity, we may achieve a glucose control comparable to people without diabetes. In 2022, we performed a long-term trial (1 week) in awake pigs where glucose was measured in the skin, insulin was given in the abdominal cavity and glucagon in the abdominal cavity or in the skin. We also got very good blood sugar control using our AP in this setting. We then observed that glucagon glucagon gives a strong but entirely local increased in subcutaneous blood flow. The absorption of insulin is very dependent on local blood flow. We now assume that micro-doses of glucagon placed in the same place as insulin will result in significantly faster uptake of insulin and insulin effect on the glucose level. We now hypothesize that an AP with CGM in the skin and micro-doses of glucagon to speed up the insulin absorption and effect on glucose levels will achieve superior glucose control. We also hold this as a better option for patients due to the much lesser invasiveness of the solution. We are now exploring, and have a pending patent application, for this use of micro-doses of glucagon in an AP.

Vi har ikke oppnådd det primære mål med prosjektet som var en sikker og robust kunstig pankreas (AP) ved en såkalt intraperitoneal (IP) tilnærming med IP glukosemåling og IP injeksjoner av insulin. Vi har imidlertid oppnådd å sannsynliggjøre at en enklere og vesentlig mindre invasiv løsning for en AP kan gi den glukosekontrollen som vi opprinnelig mente at vi måtte bruke en IP-tilgang for å oppnå. Vi har heller ikke oppnådd sekundærmål A, B eller E. Derimot har vi oppnådd sekundærmål C og D. Det viktige er at vi har innlemmet glukagon i vår AP løsning, noe som ikke var en del av de opprinnelige prosjektplanene. Glukagon skal først og fremst brukes både til å fremme absorpsjonen av subkutant (SC) tilført insulin, men også til å heve blodsukkeret når det er for lavt. Og ved at vi også benyttes SC glukosemåling så ender vi opp med en løsning for AP som ikke er mer invasiv en dagens løsning for hybride AP`er. Den eneste forskjellen vil være en noe større pumpeenhet siden den skal inneholde utstyr for infusjon av to hormoner og ikke bare et. Vår løsning med mikrodoser glukagon for raskere absorpsjon av insulin er patentsøkt. Innvilges patentet som eies og forvaltes av NTNU TTO, så kan et resultat bli betydelige inntekter og mulighet for vårt forskningsmiljø å delta i utviklingen av en AP-løsning basert på dette prinsippet.

This project combines expertise from life sciences, technology and mathematics in order to address a major medical challenge: The artificial pancreas. In diabetic patients it is difficult to achieve a stable, normal glucose level without risk of serious hypoglycaemia. Achieving this will decrease both the short- and the long-term burden of disease, which will benefit many patients and reduce the economical burden for the society. Current commercial systems and the majority of research on artificial pancreas (AP) utilize subcutaneous (SC) sensors and SC insulin delivery, in which substantial, inherent time delays jeopardize robust glucose control. We will minimize these delays by combining continuous intraperitoneal (IP) glucose sensing and IP insulin delivery, facilitating significant improvements in glucose control; maybe even normalizing the glucose levels without risk of serious hypoglycaemia. Our pilot animal experiments have already confirmed one of our hypotheses; that IP glucose sensing sometimes is substantially faster than SC sensing. It remains to elucidate the cause of major variability in the delays in IP glucose sensing. Since our sensor technology is at the in-vitro stage and not yet ready for in-vivo intraperitoneal use, we will use SC glucose sensors in our preliminary closed-loop studies. In those studies we will add IP glucagon, as a way of improving the robustness and safety of the system, since glucagon is able to increase the glucose level (opposite effect compared to insulin). The Artificial Pancreas Trondheim (APT, http://www.apt-norway.com ) research group was established in September 2013. At present (Oct 2020) we have 11 research positions (2 postdocs, 6 PhD candidates, 2 researchers, 1 engineer) funded by The Research Council of Norway and by other sources. Three of our PhD candidates have completed and defended their degree successfully (in Dec 2018, May 2019, Aug 2020). We will succeed because: 1) our group is transdisciplinary; 2) we have experience in the field; 3) we have access to novel Norwegian glucose sensor technology; 4) our double IP approach will eliminate the major obstacles in today's AP research.