Objective: Challenges with infusion sets and skin sites may hinder optimal glycemic levels in automated insulin delivery (AID). This study investigated how infusion set wear time affects glycemic control, insulin doses, and infusion site tissue characteristics and evaluated the impact of ultrasound-guided site rotation. Research Design and Methods: Children and adolescents (6–18 years) using AID completed a 28-day prospective study. Skin assessments and ultrasound imaging were performed at baseline, day 14, and day 28. Hyperechogenicity was interpreted as early lipohypertrophy. Analyses were intention-to-treat. Time in range (TIR) and insulin dose/kg/day were modeled as piecewise functions by days since infusion set change. Results: Of 40 participants, 31 had sufficient continuous glucose monitoring data (mean age 11.3 years, diabetes duration 5.6 years, 55% male, HbA1c 54 mmol/L 7.1%). After guiding, TIR increased by 2.5% points/day for 2 days post set change ( P < 0.0001) and then decreased by 4.7% points/day ( P < 0.0001). Insulin doses rose by 0.036 U/kg/day after day 2 ( P = 0.009). Hyperechogenicity was associated with lower TIR after day 2 and correlated with clinical lipohypertrophy ( P < 0.001, risk ratio = 3.63). Participants with ultrasound-only tissue changes had higher TIR than those with both ultrasound and clinical signs (73% vs. 67%, P < 0.0001). Conclusions: TIR declines from day 2 of infusion set wear despite rising insulin doses, indicating reduced absorption not fully compensated by AID. Ultrasound-guided site selection improves glycemic outcomes, likely by supporting more consistent and effective site rotation.
Bjerrekær et al. (2026) studied this question.