Identification of Activity, Insulin, and Dietary Thresholds Associated With Exercise‐Related Hypoglycaemia and Hyperglycaemia in Children With Type 1 Diabetes: A Free‐Living Study Using Explainable Machine Learning
Abstract
Aims: Some studies have explored associations between physical activity (PA) and hypoglycaemia in real‐life in type 1 diabetes (T1D) but without fully accounting for two major confounders, diet and insulin. We aimed to identify thresholds of PA characteristics, insulin, and carbohydrates associated with dysglycaemia across the exercise‐recovery cycle in children with T1D. Materials and Methods: Continuous glucose monitoring and accelerometry data, self‐reported PA sessions (timing, duration and perceived intensity), diet (timing, type and quantity; optional photographs) and insulin data (doses and timing; corrective/meal boluses and basal insulin) were collected in 36 children with T1D (11.9 ± 3.3 years, injections or open‐loop pumps) over seven free‐living days. Accelerometer data were analysed for periods corresponding to self‐reported PA sessions. Ensemble machine‐learning models classified hypoglycaemia (< 70 mg/dL) and hyperglycaemia (> 180 mg/dL) during three phases: PA, 2‐h post‐exercise (early recovery), overnight. Shapley analysis identified risk and protection thresholds of features with high importance. Results: Models achieved moderate to strong performance (AUC: 0.65–0.99; F1‐score: 0.62–0.95) across outcomes and phases. Insulin boluses > 11% of total daily dose within 4‐h pre‐exercise and > 17% during early recovery were associated with hypoglycaemia risk during PA and early recovery respectively. Carbohydrate intake showed collinearity with insulin, resulting in complex associations with glycaemia. Self‐reported PA > 80 min was associated with hypoglycaemia risk during PA, while accumulating > 15 min of accelerometer‐derived vigorous PA was actually protective against early recovery hypoglycaemia. Multiple daily sessions were associated with nocturnal hyperglycaemia protection. Conclusions: Phase‐specific thresholds across exercise and insulin domains associated with exercise‐related dysglycaemia were identified.
Related
- Neurological manifestations in metabolic dysfunction syndrome are associated with cumulative metabolic burden beyond hyperglycaemia: a translational study
- Sensor‐Detected Hypoglycaemia Is Associated With Inflammatory Protein and Immune Cell Alterations in People With Diabetes: Findings From the Hypo‐METRICS Study
- Revisiting Identification Thresholds of Specific Learning Disorders ( SLD ) in Gifted Children ( GC ). A Comparison of Three Groups: Twice‐Exceptional ( TE ) Children, GC Without SLD and Non‐ GC With SLD
- The application of explainable machine learning for the identification of MASLD in non-obese populations
- Temporal patterns of tuberculosis-associated hyperglycaemia and its effect on treatment outcomes in individuals not previously diabetic: A prospective cohort study
BibTeX
@article{iqbal2026identification,
title = {{Identification of Activity, Insulin, and Dietary Thresholds Associated With Exercise‐Related Hypoglycaemia and Hyperglycaemia in Children With Type 1 Diabetes: A Free‐Living Study Using Explainable Machine Learning}},
author = {Asif Iqbal and Haithem Dahimi and Philippe Preux and Elodie Lespagnol and Angéline Melin and Pierre Morel and Joris Heyman and Rémi Rabasa-Lhoret and Elsa Heyman},
journal = {Diabetes, Obesity and Metabolism},
year = {2026},
month = oct,
doi = {10.1111/dom.71165},
url = {https://hal.science/hal-05733768v1},
note = {CC BY-NC 4.0},
}
RIS
TY - JOUR
TI - Identification of Activity, Insulin, and Dietary Thresholds Associated With Exercise‐Related Hypoglycaemia and Hyperglycaemia in Children With Type 1 Diabetes: A Free‐Living Study Using Explainable Machine Learning
AU - Asif Iqbal
AU - Haithem Dahimi
AU - Philippe Preux
AU - Elodie Lespagnol
AU - Angéline Melin
AU - Pierre Morel
AU - Joris Heyman
AU - Rémi Rabasa-Lhoret
AU - Elsa Heyman
PY - 2026
DA - 2026/10/16
JO - Diabetes, Obesity and Metabolism
DO - 10.1111/dom.71165
UR - https://hal.science/hal-05733768v1
AB - Aims: Some studies have explored associations between physical activity (PA) and hypoglycaemia in real‐life in type 1 diabetes (T1D) but without fully accounting for two major confounders, diet and insulin. We aimed to identify thresholds of PA characteristics, insulin, and carbohydrates associated with dysglycaemia across the exercise‐recovery cycle in children with T1D. Materials and Methods: Continuous glucose monitoring and accelerometry data, self‐reported PA sessions (timing, duration and perceived intensity), diet (timing, type and quantity; optional photographs) and insulin data (doses and timing; corrective/meal boluses and basal insulin) were collected in 36 children with T1D (11.9 ± 3.3 years, injections or open‐loop pumps) over seven free‐living days. Accelerometer data were analysed for periods corresponding to self‐reported PA sessions. Ensemble machine‐learning models classified hypoglycaemia (< 70 mg/dL) and hyperglycaemia (> 180 mg/dL) during three phases: PA, 2‐h post‐exercise (early recovery), overnight. Shapley analysis identified risk and protection thresholds of features with high importance. Results: Models achieved moderate to strong performance (AUC: 0.65–0.99; F1‐score: 0.62–0.95) across outcomes and phases. Insulin boluses > 11% of total daily dose within 4‐h pre‐exercise and > 17% during early recovery were associated with hypoglycaemia risk during PA and early recovery respectively. Carbohydrate intake showed collinearity with insulin, resulting in complex associations with glycaemia. Self‐reported PA > 80 min was associated with hypoglycaemia risk during PA, while accumulating > 15 min of accelerometer‐derived vigorous PA was actually protective against early recovery hypoglycaemia. Multiple daily sessions were associated with nocturnal hyperglycaemia protection. Conclusions: Phase‐specific thresholds across exercise and insulin domains associated with exercise‐related dysglycaemia were identified.
SN - 1462-8902
SN - 1463-1326
ER -