Cambridge hybrid closed-loop algorithm in children and adolescents with type 1 diabetes: a multicentre 6-month randomised controlled trial
Julia Ware, Charlotte K. Boughton, Janet M. Allen, Malgorzata E. Wilinska, Martin Tauschmann, Louise Denvir, Ajay Thankamony, Fiona Campbell, R. Paul Wadwa, Bruce A. Buckingham, Nikki Davis, Linda A. DiMeglio, Nelly Mauras, Rachel Besser, Atrayee Ghatak, Stuart A. Weinzimer, Korey K. Hood, D. Steven Fox, Lauren Kanapka, Craig Kollman, Judy Sibayan, Roy W. Beck, Roman Hovorka, Roman Hovorka, Carlo L. Acerini, Ajay Thankamony, Janet M. Allen, Charlotte K. Boughton, Klemen Dovč, David B. Dunger, Julia Ware, Gianluca Musolino, Martin Tauschmann, Malgorzata E. Wilinska, Josephine Hayes, Sara Hartnell, Sonja Slegtenhorst, Yue Ruan, M Haydock, J Mangat, Louise Denvir, SK Kanthagnany, Jennifer Law, Tabitha Randell, P Sachdev, M Saxton, A Coupe, S Stafford, A Ball, R Keeton, R Cresswell, Lindsay Crate, H Cripps, H Fazackerley, L Looby, H Navarra, C Saddington, Vincent P. Smith, V Verhoeven, S Bratt, N Khan, L Moyes, K Sandhu, C West, R. Paul Wadwa, G. Todd Alonso, Gregory P. Forlenza, Robert Slover, Lindsey Towers, Cari Berget, Alexander Coakley, Estella Escobar, Emily Jost, Samantha Lange, Laurel H. Messer, Katie Thivener, Fiona Campbell, James Yong, Emily Metcalfe, Matthew R. Allen, S Ambler, Syeda Sana Waheed, Jane Exall, J Tulip, Bruce A. Buckingham, Laya Ekhlaspour, David M. Maahs, Lisa Norlander, Tali Jacobson, M Twon, Christine Weir, Brianna Leverenz, Jordan Keller, Nikki Davis, A Kumaran, Nicola Trevelyan, Hilary Dewar, G Price, Gillian Crouch, R Ensom
Abstract
BackgroundClosed-loop insulin delivery systems have the potential to address suboptimal glucose control in children and adolescents with type 1 diabetes. We compared safety and efficacy of the Cambridge hybrid closed-loop algorithm with usual care over 6 months in this population.MethodsIn a multicentre, multinational, parallel randomised controlled trial, participants aged 6–18 years using insulin pump therapy were recruited at seven UK and five US paediatric diabetes centres. Key inclusion criteria were diagnosis of type 1 diabetes for at least 12 months, insulin pump therapy for at least 3 months, and screening HbA1c levels between 53 and 86 mmol/mol (7·0–10·0%). Using block randomisation and central randomisation software, we randomly assigned participants to either closed-loop insulin delivery (closed-loop group) or to usual care with insulin pump therapy (control group) for 6 months. Randomisation was stratified at each centre by local baseline HbA1c. The Cambridge closed-loop algorithm running on a smartphone was used with either (1) a modified Medtronic 640G pump, Medtronic Guardian 3 sensor, and Medtronic prototype phone enclosure (FlorenceM configuration), or (2) a Sooil Dana RS pump and Dexcom G6 sensor (CamAPS FX configuration). The primary endpoint was change in HbA1c at 6 months combining data from both configurations. The primary analysis was done in all randomised patients (intention to treat). Trial registration ClinicalTrials.gov, NCT02925299.FindingsOf 147 people initially screened, 133 participants (mean age 13·0 years [SD 2·8]; 57% female, 43% male) were randomly assigned to either the closed-loop group (n=65) or the control group (n=68). Mean baseline HbA1c was 8·2% (SD 0·7) in the closed-loop group and 8·3% (0·7) in the control group. At 6 months, HbA1c was lower in the closed-loop group than in the control group (between-group difference −3·5 mmol/mol (95% CI −6·5 to −0·5 [–0·32 percentage points, −0·59 to −0·04]; p=0·023). Closed-loop usage was low with FlorenceM due to failing phone enclosures (median 40% [IQR 26–53]), but consistently high with CamAPS FX (93% [88–96]), impacting efficacy. A total of 155 adverse events occurred after randomisation (67 in the closed-loop group, 88 in the control group), including seven severe hypoglycaemia events (four in the closed-loop group, three in the control group), two diabetic ketoacidosis events (both in the closed-loop group), and two non-treatment-related serious adverse events. There were 23 reportable hyperglycaemia events (11 in the closed-loop group, 12 in the control group), which did not meet criteria for diabetic ketoacidosis.InterpretationThe Cambridge hybrid closed-loop algorithm had an acceptable safety profile, and improved glycaemic control in children and adolescents with type 1 diabetes. To ensure optimal efficacy of the closed-loop system, usage needs to be consistently high, as demonstrated with CamAPS FX.FundingNational Institute of Diabetes and Digestive and Kidney Diseases.