O-29: Stable glucose, adaptive insulin: menstrual cycle modulation in adolescent girls with type 1 diabetes using automated insulin delivery
C Morais Antas, M Afonso, C Bastião de Almeida, J Vinhais, C Grácio, T Teixeira, C Matos, F Caetano, L Lopes, C Limbert, J GalhardoAbstract
Introduction
Hormonal fluctuations across the menstrual cycle may alter insulin sensitivity in girls with type 1 diabetes (T1D). In adolescents using automated insulin delivery (AID), these effects may be underestimated if glucose metrics remain stable because the system compensates through insulin delivery.
Objectives
To assess glycaemic outcomes across the menstrual, follicular, and luteal phases in menstruating girls with T1D using AID, and to determine whether phase-related variation is more evident in insulin delivery than in glucose metrics.
Methods
We conducted a retrospective real-world study at a tertiary paediatric centre. Complete menstrual cycles were divided into menstrual, follicular, and luteal phases. Phase-specific CGM and AID metrics were analysed using mixed-effects models. The primary outcome was time in range (TIR 70-180 mg/dL); secondary outcomes included mean glucose, total daily insulin dose, automated correction boluses, and time in automated mode.
Results
32 girls (mean age 15.9±2.1 years) contributed 64 complete cycles (192 phase observations); 6 used combined hormonal contraception. TIR was similar across phases (72.3% follicular, 71.9% luteal, and 73.2% menstrual), as was mean glucose (150.1, 151.5, and 149.8 mg/dL). Time in automated mode remained >92% throughout. By contrast, insulin delivery tended to rise in the luteal phase, with a higher total daily insulin and more automated correction boluses than in the follicular phase. Exploratory subgroup analyses suggested that this luteal increase was more apparent in girls not using combined hormonal contraception, whereas phase-related differences were less evident in users.
Conclusions
In adolescent girls with T1D using AID, menstrual cycle-related variation may be more apparent in insulin delivery than in glucose metrics, suggesting an algorithmic buffering phenotype. Combined hormonal contraception may modify this pattern, with less marked phase-dependent insulin adaptation in users than in non-users.