Hypoglycemia incidence and behavioural adjustments during free-living unstructured physical activity in adults with type 1 diabetes using AID systems: Results from the RAPPID study.

Aims: To assess the frequency and management of hypoglycaemia during unstructured physical activity (PA) in adults with type 1 diabetes (T1D) using automated insulin delivery (AID) systems in real-life settings.

Materials and methods: RAPPID is a prospective, multicenter, observational study conducted over 1 month in four French tertiary care centres. Adults with T1D using one of three AID systems (MiniMed 780G, Tandem t:slim X2 with Control-IQ, or Ypsopump with CamAPS FX) and performing ≥2 unstructured PA sessions per week were included. Participants completed paper logbooks documenting each PA session (type, intensity, hypoglycemia, adjustments). Continuous glucose monitoring (CGM) and pump data were downloaded. Glycaemic control was assessed using CGM-derived metrics within predefined peri-exercise time windows.

Results: Eighty-six participants (mean age 42.5 ± 14.3 years; 43% women; diabetes duration 23.6 ± 13.1 years; mean HbA1c 52 ± 7 mmol/mol (6.9% ± 0.6%)) reported 954 PA sessions (73% aerobic; 61% moderate intensity). TBR (<70 mg/dL) increased from 1% pre-exercise to 6% during and 5% post-exercise (early recovery phase). Clinical hypoglycaemia occurred in 20% of sessions (one-third of episodes were asymptomatic); 38% of participants experienced at least 1 level 2 event (<54 mg/dL). Anaerobic or high-intensity sessions were associated with lower hypoglycaemia risk. Temporary targets were used in 73% of sessions but initiated ≥1 h before PA in only 27%. Carbohydrate intake before and during PA was frequent but often suboptimally timed or dosed.

Conclusions: Hypoglycemia remains common during and after PA in AID users. Suboptimal adjustment strategies and impaired symptom awareness contribute to risk. Individualised education remains essential to enhance safety.