Results
A central question was whether insulin delivered next to the glucose sensor disturbs the reading. The following figures are manufacturer statements from the SMART02 feasibility study. Transient signal disturbances appeared mainly with boluses of 10 units or more. In that dose range they occurred with about 20 to 25 percent of boluses. On average the sensor signal settled again after about 20 minutes.
The approach is interesting because it could reduce the number of devices worn on the body. Clinical relevance will depend on how reliably glucose sensing and automated insulin delivery work during those disturbances.
Limits
For the endpoints time in range, hypoglycaemia and longer-term reliability, the following applies: The cited manufacturer statement contains no results on these endpoints. Whether this becomes a clinical advantage over established systems has to be shown in further studies.