A key uncertainty around the evidence is that the randomised controlled trial of people with type 1 diabetes included only adults whose diabetes was well controlled.