Automated insulin delivery during the first 6 months postpartum (AiDAPT): a prespecified extension study

Abstract

Background: Clinical guidelines in the UK and elsewhere do not specifically address hybrid closed loop (HCL) use in the postpartum period when the demands of caring for a newborn are paramount. Our aim was to evaluate the safety and efficacy of HCL use during the first 6 months postpartum compared with standard care.

Methods: In this prespecified extension to a multicentre, randomised controlled trial, pregnant women with type 1 diabetes at nine UK sites were followed up for 6 months postpartum. Eligible participants (AiDAPT participants recruited after the implementation of the postpartum protocol amendment approval, those still pregnant or within six months of delivery at the time of amendment implementation and still using HCL or continuous glucose monitoring [CGM] therapy) continued their randomly assigned treatment, either standard insulin therapy with CGM or HCL therapy (CamAPS FX system version 0.3.1, CamDiab, Cambridge, UK). Participants were randomised in a 1:1 ratio with stratification by clinical site using randomly permuted block sizes of 2 or 4. The primary outcome was the between-group difference in percentage time in range ([TIR] 3·9-10·0 mmol/L [70-180mg/dL]), measured during the periods of month 0 up to 3, months 3 to 6, and over 6 months postpartum. The study is registered at ClinicalTrials.gov (ISRCTN56898625) and is complete.

Findings: Of the 124 AiDAPT trial participants, 66 (53%) were ineligible for inclusion in the postpartum extension, and 57 participants consented to continue their treatment per original random allocation. The mean age was 31 years (SD 4), and all participants had early pregnancy HbA_1c 59·4 mmol/mol (SD 10·5 [7·6% SD 1·0%]). In the 6 months postpartum, mean time with glucose levels within the target range was higher in the HCL group compared with the standard care group (72% [SD 12%] vs 54% [17%]), with an adjusted treatment difference of 15% (95% CI 7 to 22). Results for hyperglycaemia (>10·0 mmol/L) and mean CGM glucose also favoured HCL (-14% [95% CI -23% to -6%] and -1·3 mmol/L [-2·3 to -0·3], respectively). Hypoglycaemia rates were low, with no between-group differences (2·4% vs 2·6%). There were no treatment effect changes depending on postpartum period (0 up to 3 months vs 3 to 6 months) and no unanticipated safety problems.

Interpretation: Participants in the HCL group maintained 70% TIR during the first 6 months postpartum, supporting continued use of HCL rather than standard insulin therapy for people with diabetes once they have given birth.

Funding: National Institute for Health Research, Juvenile Diabetes Research Foundation, and Diabetes Research & Wellness Foundation. CGM devices were provided by Dexcom at a discounted price.

Figure 1. Consort Diagram

a Reasons for not meeting trial eligibility criteria (N=66) were: Participants more than 6 months postpartum (N=60) Completed the study and returned to NHS care prior to extension being implemented and so these participants were no longer using hybrid closed-loop (CamAPS FX) or continuous glucose monitoring (Dexcom G6). (N=6) *This participant had a neonatal death and it was felt by both trial team and local site investigators inappropriate to approach to continue into postnatal extension. d In the HCL group, three participants had missing data in the 0-<3 months period and two had missing data in the 3-6 months period as assessed by continuous glucose monitoring. In the standard care group, five participants had missing data in the 0-<3 months period and two had missing data in the 3-6 months period as assessed by continuous glucose monitoring.

Figure 2. Time in target range 3.9-10.0mmol/L (70-180mg/dL) during the six months postpartum*

*This figure shows TIR by 4-week period, and the TIR for the first 4-week period from day of delivery were 70% for HCL and 50% for standard care Dots are means, and the boxes are medians and quartiles. The whiskers are the 10^th and 90^th percentiles.