Does Minimed 780G™ Insulin Pump System Affect Energy and Nutrient Intake?: Long-Term Follow-up Study

Automated insulin delivery (AID) systems improve glycemic control and quality of life and are recommended for youth with diabetes. These systems algorithmically adjust insulin delivery using continuous glucose monitoring data and other inputs, but users must still announce carbohydrate intake for adequate postprandial coverage. Nutritional management remains central to optimal glycemic control, with guideline targets for macronutrient distribution (approximately 40–50% carbohydrate, <35–40% fat with saturated fat <10%, and 15–25% protein). While clinical trials and real-world data show AID improves glycemic outcomes across demographics and baseline characteristics, little is known about its impact on energy and nutrient intake. Observing expectations among youth and caregivers that auto-correction might obviate strict carbohydrate counting and permit consumption of high-fat foods without glycemic excursions, the authors hypothesized that switching to AID might affect diet, potentially increasing fat intake. The study evaluated macronutrient and fiber intake at baseline and at 3 and 6 months after initiating AID to determine whether dietary patterns change with a-HCLS use.

Prior research shows AID systems are safe and improve glycemic control compared with prior therapy in children, adolescents, and adults, with robust real-world performance data. However, existing AID studies have not assessed effects on food intake. Independent literature on diet in youth with T1D indicates frequent deviations from guidelines, including high saturated fat and low fiber intake, and that better diet quality is associated with more optimal glycemic control. These dietary patterns raise concern for elevated LDL cholesterol and atherosclerosis risk, underscoring the importance of limiting saturated fats and promoting polyunsaturated/monounsaturated fats and fiber-rich foods in pediatric T1D care.

Design: Prospective, real-world, 6-month follow-up study. Participants: Twenty-nine children, adolescents, and young adults with T1D who switched to the MiniMed 780G™ advanced hybrid closed-loop system between November 2021 and May 2022. Inclusion comprised those adhering to protocol; exclusions were comorbidities affecting diet (e.g., celiac disease, cystic fibrosis). Training and device initiation: All participants/caregivers received training and began in manual mode. Auto mode was initiated after 3 days for prior sensor-augmented pump users (MiniMed 640G) and after 10 days for multiple daily injection users. Food diaries: Participants completed 3-day weighed food diaries (two weekdays, one weekend day) at baseline (before auto mode) and at 3 and 6 months (total 9 days per participant). Diaries were reviewed by dietitians for accuracy, with follow-up queries as needed. Dietary analysis: Using Ebispro for Windows; Turkish Version (BeBiS 8.2), the following were calculated: total energy (kcal), carbohydrate/protein/fat energy percentages, saturated fat (energy %), dietary cholesterol (mg), and dietary fiber (g/1000 kcal). Anthropometrics: Height and weight measured with standard equipment; BMI calculated and SDS determined using Turkish references. Normal weight was defined per age-specific SDS or adult BMI 18.5–24.9 kg/m2. Glycemic data: HbA1c measured by turbidimetric inhibition immunoassay. Device data were uploaded to CareLink and analyzed for TIR (70–180 mg/dL), TBR (<70 mg/dL), TAR (>180 mg/dL), coefficient of variation (CV), glucose management indicator (GMI), sensor wear, time in closed loop, auto-bolus %, and sensor glucose. Initial a-HCLS settings: active insulin time 2.5 h, glucose target 100 mg/dL, adjusted as needed. Pre-transition therapy: Participants previously on MDI (glargine plus rapid-acting analogs) or sensor-augmented pump (MiniMed 640G). Carbohydrate counting accuracy/consistency was assessed before a-HCLS initiation. Statistical analysis: SPSS v25 used. Significance p<0.05. Data presented as mean±SD or median (IQR). Between-group comparisons (MDI vs pump at baseline) used independent t test or Mann–Whitney U. Repeated measures analyzed by repeated-measures ANOVA or Friedman test; pairwise by Bonferroni/Dunn or paired t as appropriate. Correlations assessed between macronutrient/fiber intake and glycemic metrics (TIR, TAR, TBR, CV, GMI). Post-hoc power analysis: for repeated measures (3 time points), η2=0.046 (Cohen’s f≈0.2125), yielding power≈0.71 at alpha 0.05.

• Participants: n=29; mean age 12.7±4.3 years; 48.3% female; diabetes duration median 2.2 (4.1) years; baseline HbA1c 6.9±1.2%. Eighty percent were on sensor-augmented pump and 20% on MDI before transition. BMI remained normal and unchanged over follow-up.
• Macronutrient intake: Baseline mean energy %: carbohydrate 49.1±4.5, protein 17.8±2.3, fat 33.0±3.9. No statistically significant changes in carbohydrate or fat energy % over 6 months (carbohydrate p=0.416; fat p=0.264). Protein showed a small fluctuation across time points (ANOVA p=0.041) with pairwise differences 0–3 months (p=0.031) and 3–6 months (p=0.02), but overall macronutrient distribution remained within guideline ranges.
• Fiber and saturated fat: Fiber intake stayed below recommendations (≈11.2–11.4 g/1000 kcal vs recommended 14), and saturated fat intake remained above recommendations (median ≈11.6–12.8% energy; target <10%) at all time points.
• Meals/snacks: Number of meals (median 3/day) and snacks (≈1/day) did not change over time.
• Glycemic outcomes: TIR increased from 79.0 (15.5)% at baseline to 81.0 (6.5)% post-transition (p<0.05). GMI decreased from 6.6 (0.4) to 6.5 (0.3) (p<0.05). Mean glucose decreased from 143.9 to 136.0 mg/dL at 3 months (p=0.038 pairwise), with CV showing a small overall change (Friedman p=0.039) but non-significant pairwise differences.
• Auto-correction boluses: Median auto-correction % increased from 14.0 (9.5)% after 14 days in auto mode to 18.0 (11.0)% at 3 months and 19.0 (7.5)% at 6 months (p<0.05). Auto-correction correlated positively with TAR and GMI at 3 months (TAR r=0.775; GMI r=0.691; both p<0.01) and with TAR (r=0.440, p<0.05) and GMI (r=0.529, p<0.01) at 6 months. It correlated negatively with TIR at 3 months (r=-0.747, p<0.01) and 6 months (r=-0.395, p<0.05).
• Announced vs recorded carbohydrates: At baseline, announced carbohydrate to the pump did not differ from food diary carbohydrate. At 3 and 6 months, announced carbohydrate exceeded diary amounts (3 months 220.9±104.1 vs 184.7±44.2 g/day, p=0.036; 6 months 230.1±112.7 vs 184.3±59.3 g/day, p=0.011).
• Additional correlations: At baseline, protein energy % correlated positively with HbA1c (r=0.412, p<0.05). At 6 months, carbohydrate energy % correlated negatively with TBR (r=-0.403, p<0.05). Carbohydrate and fat energy % showed strong inverse correlations across time points.

Switching to the MiniMed 780G™ advanced hybrid closed-loop system improved glycemic control (higher TIR, lower GMI) without altering overall energy intake or macronutrient distribution over 6 months. Persistent high saturated fat and low fiber intakes align with prior reports in youth with T1D and have implications for long-term cardiovascular risk. The approximately 20% contribution of auto-correction boluses and their negative association with TIR suggest that overreliance on auto-correction, inaccurate carbohydrate counting, or skipped meal boluses may impair optimal glycemic outcomes. The finding that announced carbohydrates exceeded recorded intake at follow-up indicates possible attempts to manipulate glycemia by announcing carbohydrates without consuming them, which may adversely affect control. Educational interventions emphasizing accurate carbohydrate announcement and discouraging attempts to “trick” the system are warranted. The modest TIR improvement relative to other studies likely reflects already high baseline TIR in this cohort, yet the system sustained or slightly improved glycemic metrics while dietary patterns remained stable, indicating that AID benefits were independent of changes in diet composition.

In this 6-month real-world study of children, adolescents, and young adults with T1D initiating the MiniMed 780G™ system, glycemic control improved while energy intake and macronutrient distribution did not change. Despite AID use, participants consistently consumed saturated fat above and fiber below recommended levels. Ongoing monitoring of dietary choices and targeted nutrition education to reduce saturated fat and increase fiber are recommended to mitigate cardiovascular risk. Future work with larger samples and additional behavioral/physical activity data could further clarify diet–technology interactions and identify thresholds for auto-correction use that prompt tailored education.

Main limitations include the small sample size and lack of physical activity diaries, which may affect interpretation and generalizability of dietary and glycemic associations. Although conducted in real-life conditions, the study was single-cohort without a control group, and participants had relatively high baseline TIR, potentially limiting observable improvements.