The differential impact of automated insulin delivery systems on body mass index in children with type 1 diabetes.

OBJECTIVES

Automated insulin delivery (AID) improves glycaemic outcomes in children and young people (CYP) with type 1 diabetes (T1D), but effects on body weight are less clear. We evaluated longitudinal changes in BMI z-score following AID initiation and examined predictors using mixed-effects modelling.

METHODS

We retrospectively studied 170 CYP with T1D starting AID at a tertiary centre. BMI z-score was assessed at baseline, six, and twelve months. Predictors of six-month BMI change (ΔBMI) included demographic, socioeconomic, and glycaemic variables with sensor-weighted glucose metrics.

RESULTS

Baseline BMI z-score was 0.75 (SD 1.27), rising to 0.90 at six months (p<0.001) and stable at twelve (p=0.851 vs. six). Increases occurred in underweight (+1.12, 95 % CI 0.61-1.62, p<0.001) and below-average healthy weight (+0.38, 95 % CI 0.20-0.57, p<0.001), with minimal change in overweight or obese groups. Mixed-effects modelling identified baseline BMI z-score (B = -0.14, p<0.001), baseline mean blood glucose (MBG; B = +0.07, p<0.001), and their interaction (B = -0.04, p=0.025) as independent predictors. Age, sex, socioeconomic status, ethnicity, therapy type, HbA1c, and sensor wear were not significant. Across groups, AID increased time in range by 15-17 % and reduced HbA1c by 6 mmol/mol without increasing hypoglycaemia.

CONCLUSIONS

AID supports weight restoration in underweight CYP while maintaining stability in overweight or obese groups. BMI change is determined by baseline BMI and MBG, independent of demographic or treatment factors.