The incidence of type 1 diabetes (T1D) is increasing, particularly among children.1 The number of new cases of T1D in European children younger than 5 years has been predicted to double between 2005 and 2020, with a 70% rise in children younger than 15 years.2 At present, management of T1D involves multiple daily injections of insulin or the use of an insulin pump. These have a number of disadvantages including the need to calculate glucose doses and the time lag between dose administration and effect. Children with type 1 diabetes have increased insulin sensitivity compared to adolescents and adults, and are at increased risk of severe hypoglycaemia,3 particularly overnight.4,5 The risk of this potentially fatal condition is also a constant concern for patients and their caregivers.6 The fear of hypoglycaemia also prevents people with T1D from exercising.7 As a result, many children do not meet guideline recommended treatment goals.8
One of the most active areas in type 1 diabetes research is the development of a closed loop system that aims to automate insulin delivery based on continuous glucose monitoring (CGM). Such systems are becoming commercially viable; Medtronic’s MiniMed 670G closed loop system has recently been launched in the US.9 The latest progress in closed loop technology was highlighted in two small studies that were presented on 12 June 2017 at the American Diabetes Association (ADA) 2017 Scientific Sessions in San Diego, California
The first study investigated a hybrid closed loop system comprising an Omnipod® patch pump, Dexcom, G4 CGM sensor with Bluetooth technology built into the receiver, and a personalized predictive control algorithm.10 At mealtimes, the patient inputs the amount of carbohydrates ingested to determine the meal dose of insulin. The study included 12 children aged 6-12 years and involved an inpatient assessment with meals ranging from 30-90 grams of carbohydrates and limited physical activity. At 36 hours, 69% of overall glucose values were within prespecified targets of 70-180 mg/dl and 82% of overnight readings were within this range. The participants' average glucose level was 157 mg/dl, and only 2% of readings were <70 mg/dl with the closed loop system compared to 4% when at home and not using the system. The mean fasting glucose level following overnight use of the closed-loop system was 136±24 mg/dL. These findings indicate that the Omnipod® automated glucose control algorithm performed well and was safe during day and night use in children with type 1 diabetes.
Chief investigator Bruce Buckingham, from Lucile Salter Packard Children’s Hospital at Stanford University, said: “Hybrid closed-loop systems do a great job improving glucose control overnight, significantly lowering the risk of hypoglycaemia.” He also highlighted the convenience of use of this system, saying: "Many patients prefer wearing an untethered patch pump, which provides more flexibility to enjoy physical activities without worrying about infusion set detachments." Future work investigating this system will involve longer outpatient studies under home living conditions and a broader range of age groups.
Another study investigated whether exercise-related hypoglycaemia can be reduced through a closed-loop system that responds automatically to physical activity.11 The study assessed a range of systems, including a single-hormone system dosing insulin only, a dual-hormone system dosing both insulin and glucagon, a predictive system that shut off insulin if glucose was predicted to be too low, and standard of care where patients used their usual methods of glucose control. Glucose values from a Dexcom G5 sensor were transmitted every 5 minutes to a Google Nexus phone running the closed loop algorithm. In both closed-loop systems, a Zephyr heart rate monitor and accelerometer detected exercise and signalled insulin cut-off for 30 minutes followed by 50% insulin reduction for 60 minutes. In the dual-hormone system, exercise detection also signalled an increase of glucagon by 200% for 1.5 hours.
Participants (20 adults) underwent four separate outpatient visits, which entailed exercising for 45 minutes at 60% peak oxygen uptake (VO2max) on day one and day four in a human performance lab, and at least one home-based exercise session to determine the amount of time in hypoglycaemia. Participants inputted their estimated carbohydrate intake into the insulin pump, which automatically delivered a portion of the estimated pre-meal insulin dose. Blood glucose levels were measured four times daily.
Preliminary data from a subgroup of 17 visits showed that the dual hormone system reduced exercise-induced hypoglycaemic events from 6.3% to 1% when compared with insulin only, and was more effective than the predictive system and standard care. Time spent in hypoglycaemia (<70 mg/dl) was 11.0% for the dual-hormone system; 3.4% for single-hormone; 1.2% for the predictive system; and 2.1% for standard care. However, participants undergoing standard care prevented exercise-induced hypoglycaemia by keeping their blood sugar levels significantly higher before the onset of exercise. The mean glucose level after exercise was significantly lower for single-hormone compared with dual-hormone systems: 67±6 mg/dL and 100±9 mg/dL, respectively.
Lead investigator Peter G. Jacobs, PhD, of Oregon Health & Science University in Portland, said, "Our findings show that fully automated insulin and glucagon delivery, combined with wearable physical activity sensors that detect exercise, effectively controlled glucose levels, reduced exercise-induced hypoglycaemia and can safely be used in a home environment," Future studies aim to use the system on a smart watch rather than a phone, as this will enable people to monitor their exercise more easily.
These studies represent important steps in the development of closed loop systems that have great potential in the years ahead to further improve the lives of people with T1D.