Key points are not available for this paper at this time.
BACKGROUND: We assessed whether fully closed-loop insulin delivery (the so-called artificial pancreas) is safe and effective compared with standard subcutaneous insulin therapy in patients with type 2 diabetes in the general ward. METHODS: For this single-centre, open-label, parallel-group, randomised controlled trial, we enrolled patients aged 18 years or older with type 2 diabetes who were receiving insulin therapy. Patients were recruited from general wards at Addenbrooke's Hospital, Cambridge, UK. Participants were randomly assigned (1:1) by a computer-generated minimisation method to receive closed-loop insulin delivery (using a model-predictive control algorithm to direct subcutaneous delivery of rapid-acting insulin analogue without meal-time insulin boluses) or conventional subcutaneous insulin delivery according to local clinical guidelines. The primary outcome was time spent in the target glucose concentration range of 5·6-10·0 mmol/L during the 72 h study period. Analyses were by intention to treat. This trial is registered with ClinicalTrials.gov, number NCT01774565. FINDINGS: Between Feb 20, 2015, and March 24, 2016, we enrolled 40 participants, of whom 20 were randomly assigned to the closed-loop intervention group and 20 to the control group. The proportion of time spent in the target glucose range was 59·8% (SD 18·7) in the closed-loop group and 38·1% (16·7) in the control group (difference 21·8% 95% CI 10·4-33·1; p=0·0004). No episodes of severe hypoglycaemia or hyperglycaemia with ketonaemia occurred in either group. One adverse event unrelated to study devices occurred during the study (gastrointestinal bleed). INTERPRETATION: Closed-loop insulin delivery without meal-time boluses is effective and safe in insulin-treated adults with type 2 diabetes in the general ward. FUNDING: Diabetes UK; European Foundation for the Study of Diabetes; JDRF; National Institute for Health Research Cambridge Biomedical Research Centre; Wellcome Trust.
Thabit et al. (Wed,) studied this question.