Continuous Glucose Monitoring-Derived Glycemic Profiles in Adolescents With Obesity Without Diabetes

Abstract

Background:

Continuous glucose monitoring (CGM) is increasingly used in individuals without diabetes. This study aimed to characterize CGM metrics in youth with obesity without diabetes and examined variation by sex, age, excess percent of the 95th percentile for body mass index (%BMI_p95), physical activity, and clinical biochemical fasting samples.

Methods:

The analytic sample comprised youth aged 12 to 18 years with obesity, defined as BMI > 95th percentile for age and sex, without diabetes or obesity-related complications. Participants were recruited for a 24-week behavioral weight loss intervention. Prior to intervention, participants wore an unmasked CGM for up to five days under free-living conditions. Glycemic metrics were calculated using the iglu R package and summarized as median (interquartile range [IQR]). Associations with sex, age, %BMI_p95, self-reported physical activity (PROMIS [Patient-Reported Outcomes Measurement Information System] Physical Activity questionnaire), and fasting samples were examined using Wilcoxon rank-sum tests, Spearman’s correlations, and quantile regression.

Results:

Fifty-four youth (median age = 16 years [14-18]; BMI = 39.8 kg/m² [36.0-45.4]; 54% female; 85% Hispanic/Latino) wore CGM devices for up to five days. Median mean glucose was 114.8 mg/dL (107.0-128.3), estimated hemoglobin A1c (eHbA1c) was 5.6% (5.4-6.1), and glucose management indicator (GMI) 6.1 (5.9-6.4). Time above range 140 mg/dL was 10.1% (4.1-24.4). While no significant sex differences were observed for mean glucose, eHbA1c, or GMI, fasting plasma glucose was positively correlated with mean CGM glucose (Spearman’s ρ = 0.36, P = .01), and baseline hemoglobin A1c (HbA1c) was positively associated with mean CGM glucose, with each +1% in HbA1c corresponding to +19 mg/dL higher mean CGM glucose (P = .01). Quantile regression revealed significant associations between %BMI_p95 and coefficient of variation when adjusting for sex, age, and PROMIS score (P = .01).

Conclusion:

Among youth with obesity without diabetes, CGM revealed modest BMI-related glycemic variability and no sex differences, providing normative CGM data for this population.