Evaluation of the biomedical and psychosocial impact of automated Complete-Loop (Artificial Pancreas) insulin delivery in women with type 1 diabetes during pregnancy

Aim: To compare the biomedical and psychosocial effectiveness outcomes of STANDARD patient-directed and AUTOMATED Complete-Loop insulin delivery in pregnant women with type 1 diabetes. Design: Open label two-arm multicentre randomised controlled trial (RCT). Target Population: n=124 pregnant women with type 1 diabetes (T1D), age 18-45 years, 7-14 weeks gestation, standard patient-directed insulin delivery (pen or pump), HbA1c 48-86 mmol/mol. Setting: 10 NHS antenatal diabetes clinics. Each will recruit 6-7 women per year (12-13 over 2 yrs) based on CONCEPTT recruitment. Treatments being assessed: Two forms of insulin delivery each aiming to optimise the time spent within the NICE recommended target glucose range (3.5-7.8mmol/L). Intervention: An automated closed-loop insulin delivery (AiD) system comprising a Continuous Glucose Monitor (CGM), an insulin PUMP, and a mobile PHONE hosting the computer-based model predictive control (MPC) algorithm. Control: A standard insulin delivery system which is either insulin pump (Continuous Subcutaneous Insulin Infusion - CSII) or multiple daily injections (MDI) without closed-loop Planned outcome measures: Maternal CGM measures from trial entry to delivery. Maternal assessments at baseline, 24 & 34 weeks gestation & hospital discharge. Neonatal assessment at delivery & hospital discharge. PRIMARY OUTCOME: The time spent with glucose levels between 3.5-7.8 mmol/L based on CGM measures (Time In Range TIR 3.5-7.8mmol/L) from 16 weeks gestation until delivery. SECONDARY OUTCOMES: Maternal Glycaemic Outcomes 1. Time spent with CGM glucose levels above and below target range (TAR>7.8mmol/L, TBR<3.5mmol/L), mean CGM glucose and CGM glucose variability measures (CV, SD). 2. Frequency and severity of hypoglycaemia episodes defined as CGM glucose levels TBR <3.5 mmol/L (level 1 hypoglycaemia) and TBR <3.0 mmol/L (level 2 hypoglycaemia) for at least 15 minutes 3. International consensus targets for glycaemic assessment; TIR 3.5-7.8mmol/L >70% (16hr 48 min), TAR >7.8mmol/L <25% (6hr), TBR <3.5mmol/L <4% (1hr), and TBR <3.0mmol/L <1% (15min) 4. Low Blood Glucose Index (LBGI) and High Blood Glucose Index (HBGI) measures 5. Where possible, blood samples will be collected at baseline, 24-26 weeks, 34-36 weeks for HbA1c testing to assess the change in the maternal level. Samples will be stored for further metabolic studies (optional). 6. CGM glucose levels during the first (<12 weeks 6 days gestation), second (13-27 weeks 6 days gestation) and third trimesters (28 weeks until delivery). 7. CGM glucose levels during the 24 hours (midnight to midnight) and overnight time 23.00-07.00hr Maternal Obstetric Outcomes 1. Gestational weight gain (weight gain from booking visit to 36 weeks) 2. Maternal hypertensive disorders (Gestational, worsening of pre-existing hypertension or preeclampsia) 3. Fetal growth (ultrasound estimated fetal weight, head and abdominal circumference measurements) 4. Mode of delivery (vaginal, instrumental, elective caesarean section and emergency caesarean section) 5. Gestational age at delivery and indication for any preterm delivery (<37 weeks) 6. Adverse events including pregnancy loss <24 weeks, stillbirth, neonatal death 7. Maternal hospital admissions (all admissions including the delivery admission) 8. Hospital length of stay (all admissions including the delivery admission) Infant Outcomes 1. Neonatal morbidity including treatment for neonatal hypoglycaemia, neonatal jaundice and respiratory distress between the time of infant delivery and discharge from hospital. 2. Infant birth weight (customised birth weight percentile, incidence of large for gestational age (LGA), and small for gestational age (SGA) 3. Neonatal intensive care unit (NICU) admission >24 hours 4. Infant feeding at hospital discharge, 8-12 weeks postpartum, and 24 weeks postpartum (breast, bottle, both) 5. Hospital length of stay (from delivery until hospital discharge), including re-admissions >24h within the first seven days from birth Safety Outcomes 1. Frequency and severity of diabetic ketoacidosis 2. Number and severity of episodes of severe hypoglycaemia 3. Number and severity of episodes of adverse device effect Psychosocial Outcomes The following questionnaires will be completed at Baseline and 34-36 weeks: Insulin Delivery Systems: Perspectives, Ideas, Reflections and Expectations (INSPIRE), EQ-5D Health-Related Quality of Life Questionnaire; Diabetes Distress Scale (DDS), Hypoglycaemia Fear Survey Questionnaire II (HFSQ II) (Worry scale only), Pittsburgh Sleep Quality Index (PSQI) Qualitative interviews will be conducted with a smaller sample of trial participants who provide additional consent for a qualitative interview: 1. 25 women randomized to the AiD arm will be interviewed post randomization and again at 34-36 weeks 2. Up to 25 staff from the trial sites will be interviewed Health economic outcomes 1. Cost of the AiD closed-loop system 2. Cost of the control-arm glucose monitoring and insulin delivery 3. Training costs for AiD and control arms 4. Maternity health care use for AiD and control arms, including NHS antenatal clinic visits, between visit contacts, antenatal hospital admissions, length of hospital stay 5. Neonatal health care use for AiD and control arms, including costs of delivery vaginally or by caesarean section, costs associated with any complication of delivery, and neonatal complications, neonatal intensive care unit admissions, total neonatal length of hospital stay. 6. The EQ-5D Health-Related Quality of Life Questionnaire Analyses: Intention to treat (ITT), with secondary per protocol analyses. Primary endpoint is the comparison between the two treatment arms based on a linear mixed effects model. Randomisation is stratified for centre and insulin delivery (pen or pump). Sample size information: It is accepted that a difference of 10% in CGM time-in-target (2.4hrs/day) is clinically worthwhile. To detect a clinically relevant difference of 10% with a SD of 15% at 90% power and 5% two sided significance requires a sample size of 98 or 124 total to allow for up to 10% drop outs and 10% pregnancy loss before 34 weeks gestation. Team expertise: diabetes, device trials, engineering, qualitative methodology, obstetrics, psychology, service users, statistics.