Sage Journals: Discover world-class research

Abstract

Aims:

To systematically assess the impact of automated insulin delivery (AID) systems on body mass index (BMI), weight, and body composition and to analyze potential associations with the total daily, basal, and % of autocorrection insulin doses.

Materials and Methods:

We performed this review through systematic searches of Pubmed, EMBASE, The Cochrane Library, Web of Science, Clinicaltrials.gov, and the International Clinical Trials Registry Platform between September 1, 2015, and September 1, 2024. PICOS framework was used in the selection process, and summary outcome data were provided for each study. A meta-analysis determined the impact of BMI z-score and total daily insulin (TDI) pro-Kg increase. We used the GRADE approach to rank the quality of evidence. This study was registered with PROSPERO, CRD42024613557.

Results:

A total of 161 studies were identified after duplicate removal, and 18 were included. According to studies with moderate quality evidence (No. 13), the meta-analysis revealed an unchanged BMI z-score after 1-year follow-up with AID: 0.0174 (95% confidence interval [CI]: −0.095 to 0.130), despite the increase in TDI: 0.54 U/Kg per day (95% CI: 0.34 to 0.75), mainly due to the autocorrection boluses.

Conclusions:

Evidence confirms the importance of educating youths with type 1 diabetes using AIDs and their families about healthy diet composition and carbohydrate intake. However, these systems also improve glucose control in subjects who are less accurate with bolus at meals without an impact on BMI z-score.

Introduction

The prevalence of overweight and obesity among youth with type 1 diabetes (T1D) is increasing, as recently reported.^1–4 Excessive body weight (BW) is often related to higher daily and basal insulin doses,⁵ associated with worse glucose control⁶ and increased risk of macro- and microvascular complications.⁷

The advent of continuous subcutaneous insulin infusion (CSII) allowed the possibility of efficiently administering extra doses for snacking, which could theoretically contribute to unhealthy nutritional choices. Studies that observed body mass index (BMI) in children with T1D using insulin pumps showed no unequivocal results: in some people with CSII, a positive association between higher basal insulin dose and excessive weight gain has been reported,⁸ whereas a recent meta-analysis suggested no change in body mass and weight gain in children treated with either insulin pumps or multiple daily injections.^9,10

Regarding hybrid closed loop (HCL) systems, higher BMI z-score values than CSII users have been found in some studies^11–13 but not in others.^14,15 Since 2019, automated insulin delivery (AID) systems have been commercially available to manage T1D in youths. These systems automatically adjust basal insulin infusion to the current glucose level together with autocorrection dosing, and they are more efficient in improving glucose control^16,17 and psychological outcomes.¹⁸ Indeed, automated correction boluses are delivered to cover up missed, delayed, or inaccurate insulin boluses. In case of insufficient carbohydrate counts, increased portion sizes, or fatty foods, which are difficult to manage with conventional therapy, AID systems help by delivering an autocorrection dose without the patient’s intervention,¹⁹ allowing a more liberal and flexible approach to dietary aspects of the therapy.^19,20 In a few studies, a similar basal rate and an increased percentage of autocorrection in total daily insulin (TDI) have been reported,^21,22 but the impact of AIDs on insulin need and, in turn, on BMI z-score is still debated. A systematic review on this topic is not yet available in the literature; therefore, this study aims to provide an up-to-date summary of the impact of AID on weight, BMI, and body composition in youths with T1D, and potential associations with the percentage of auto mode, and autocorrection insulin doses.

Materials and Methods

Search strategy

We searched electronic databases (Pubmed, EMBASE, The Cochrane Library, Web of Science, Clinicaltrial.gov, International Clinical Trials Registry Platform) for studies published between September 1, 2015, and September 1, 2024. Search terms, or “MeSH” (Medical Subject Headings) for this systematic review included different combinations: “automated insulin delivery” or “AID” or “advanced hybrid closed loop” or “closed loop” or “AHCL” AND “body mass index” or BMI or weight or “body composition” or “fat mass” AND “type 1 diabetes” or T1D.

Criteria for study selection

We conducted a systematic search of the literature according to the PICOS model (Population, Intervention, Comparison, Results, Study design):

Population	Pediatric study participants (1–18 years old) with T1D
Intervention	Use of AIDs (during follow-up)
Comparison	Pre-AID treatment: multiple daily injections (MDI), continuous subcutaneous insulin infusion (CSII), sensor augmented pump (SAP) with low glucose suspend (LGS) or predictive low glucose monitoring (PLGM) system, or hybrid closed loop (HCL) or Baseline AID use
Results	Changes in BMI, BMI z-score, weight, body composition, fat mass Predictors of BMI values
Study design	Randomized clinical trials (RCTs), observational studies, prospective studies, cross-sectional studies, exploratory studies, and a mix of qualitative and quantitative studies

Inclusion criteria were: i) study population: children and adolescents (aged 1–18 years) with T1D; ii) study type: prospective observational studies, exploratory studies, a mix of qualitative and quantitative studies; review articles were excluded, but their reference lists were screened to identify potentially eligible studies. Cross-sectional studies were included if appropriate for outcome, as the authors analyzed at least predictors of BMI values in different groups of subjects; only published full papers were included, whereas abstracts only were not included; iii) data on intervention: use of AIDs and at least one outcome analyzed; iv) publication date: last 10 years (2015–2024).

Exclusion criteria: i) data available only for adults ≥18 years or not separable from the adults one; ii) case reports; iii) studies with less than 10 pediatric participants; iv) full article not available; v) study not yet published; vi) studies not reporting data on BMI or weight or body composition; vii) languages other than English were not “a priori” exclusion criteria.

Data extraction and management

Two independent investigators (F.D.C. and S.F.) screened, for inclusion, the title and abstract of all the studies identified using the search strategy. Any discrepancies were resolved by consensus or consultation with a third investigator (R.F.). After abstract selection, four investigators conducted the full paper analysis (C.G., A.D., E.P., E.M.A.).

The following characteristics were evaluated for each study in the full paper: i) reference details: authorship(s); published or unpublished; year of publication; year in which the study was conducted; other relevant cited papers; ii) study characteristics: design, comparator, setting, follow-up duration; iii) population characteristics and study design: number of participants using AID, age, diabetes duration, type of insulin therapy pre-AID and/or in the comparator; iv) % in auto mode; v) daily insulin doses: TDI, percentage of basal, % of autocorrection, carbohydrate intake; vi) main results: BMI, BMI z-scores, weight, body composition, HbA1c, glucose metrics.

Assessment of the certainty of the evidence

We used the GRADE approach (Grading of Recommendations Assessment, Development, and Evaluation) to rank the quality of evidence (www.gradeworkinggroup.org) for the included studies. Two authors (E.M. and R.F.) independently assessed the certainty of the evidence for each of the outcomes. In the case of risk bias in the study design, imprecision of estimates, inconsistency across studies, indirectness of the evidence, and publication bias, the recommended option of decreasing the level of certainty by one or two levels according to the GRADE guidelines was applied.²³ In the grading process of the selected studies, three main criteria were used to assess the precision of the reported outcomes, using median values as cut-off: i) at least 50 study participants (less than 50 was reported as “small cohort”); ii) follow-up of at least 12 months: study duration less than 12 months was considered as “short follow up”; iii) value of carbohydrate (CHO) intake per day reported.

The GRADE approach results in an assessment of the certainty of a body of evidence and allocation to one of four grades:

High:	Further research is very unlikely to change confidence in the estimate of the effect
Moderate	Further research is likely to have an important impact on confidence in the estimate of the effect and may change the estimate
Low	Further research is very likely to have an important impact
Very low	Any estimate of effect is very uncertain.

Meta-analysis

A meta-analysis was conducted on the impact of AIDs on the change in BMI z-score and TDI/Kg from baseline to 12 months after. The analysis included studies that reported BMI z-score or TDI/Kg at T0 and T12 months. The analysis used the standardized mean difference as an outcome measure. A random-effects model was fitted to the data. The amount of heterogeneity (i.e., tau²) was estimated using the restricted maximum-likelihood estimator.²⁴ In addition to the estimate of tau², the Q-test for heterogeneity and the I² statistic are reported.²⁵ In case any amount of heterogeneity is detected (i.e., tau² > 0, regardless of the results of the Q-test), a prediction interval for the true outcomes is also provided.

Results

After duplicates were removed, 161 studies were identified following the literature review. After reviewing titles and abstracts, 130 additional records were excluded: 8 review articles, 11 studies including only participants older than 18 years or pediatric data were not separable from the adults one, 3 including less than 10 subjects, 51 regarding subjects not wearing AIDs, 40 studies reporting outcomes different from those of interest, 16 studies not available as full paper, 1 study with publication period before 2015.

A total of 31 full-text manuscripts were assessed for eligibility; after full-text examination, 18 studies were excluded (Supplementary Appendix). The PRISMA flow diagram (Fig. 1) summarizes the publications screening process. A detailed description of outcomes and related measures used in the 13 selected studies is reported in Table 1.^{20–22,26–35} A summary of the studies included in this systematic review and the evidence grading are reported in Table 2.^{20–22,26–35}

FIG. 1.

Publication selection process summarized by the PRISMA flowchart.

Table 1.

Literature Analysis After PICOS Selection: Summary of the 13 Studies Included in Each Survey’s Systematic Review and Evidence Grading

Reference	Population and Study Design	Daily Insulin Doses	BMI, Weight, CHO, And Glucose Values	Strengths, Limitations, Risk Of Bias, Level Of Evidence
Tornese G et al. 2021²¹	44 children- adolescents Age: 13.3 years (range 2–21 years) (27 females) Pre-AID treatment: n.a. Design: Retrospective. MiniMed 780G (No. 24) vs. HCL MiniMed 670G at T0, 1, 3, 6 m f/up: 6 m Region: Italy	% in Auto mode: no increase during f/up (P > 0.05) TDI/Kg: increase from baseline (manual mode) 0.74 (0.56–0.84) to auto mode 0.83 (0.70–1.03), P < 0.01, and increase during f/up up to 3 m: 0.94 (0.76–1.09): P < 0.01 Basal insulin%: in AHCL users from baseline to 6 m: 43 (38; 54) to 42 (35; 48), P = 0.60 Bolus insulin: from baseline 57 (47; 62) to 6 m: 58 (52; 65), P = 0.60 % of autocorrection/TDI for AHCL: increased from auto start 26 (21; 37) to 1 m: 38 (28, 50) P < 0.01 to 6 m: 33 (21; 39) (n.s.)	BMI SDS not significantly change throughout the study period in AHCL (P > 0.05) CHO: Meals, CHO entered per day, and CHO intake did not change significantly in AHCL (increased from Manual mode to auto mode use in HCL users only, P < 0.01)	Limitations: retrospective, small sample size, short f/up, no pubertal assessment ⨁⨁⊖⊖ LOW
Tornese G et al. 2021²¹			GMI, TIR and TAR: improved significantly from manual to all time points in both HCL and AHCL users (P < 0.01) HbA1c: AHCL users had a significantly lower HbA1c (P < 0.01) compared to HCL after six months of HCL use (7.1 vs. 7.7%) No significant differences in coefficient of variation (CV) were observed throughout the study period
Seget S et al.²⁰	50 children-adolescents Age: 5.4–16.8 years (48% boys) Pre-AID treatment: SAP-LGS/PLGS Design: Prospective. Subjects on Minimed 780G Comparator: T0, 6, 12 m f/up: 12 m Region: Poland	% in auto mode: increased significantly from baseline 92.32 ± 14.72 to 12 m: 97.56 ± 2.79 (P < 0.05) TDI/Kg: from 0.67 ± 0.21 U/kg at baseline to 0.80 ± 0.21 U/kg (P < 0.001) at 12 m Basal insulin (%): from baseline to 12 m; 34.88 ± 6.91 to 35.08 ± 6.36 (P > 0.05) Bolus insulin: from baseline to 12 m, 65.02 ± 6.9 to 64.92 ± 6.36 (P > 0.05) Autocorrection (U): increased significantly from baseline 2.85 ± 2.42 to 4.09 ± 2.60 after 1 year (P < 0.05)	BMI z-score: no change: from 0.57 ± 2.85 at baseline to 0.51 ± 2.79 after 1 year (P > 0.05)	Limitations: no value of CHO/day, no HbA1c data lack of control group, no pubertal assessment ⨁⨁⨁⊖ MODERATE
Seget S et al.²⁰			Time in range (70–180 mg/dL): 79.28 ± 8.12% vs. 80.40 ± 8.25%, (P > 0.05).
Piccini B et al.²⁶	83 children-adolescent Age: 13 ± 4.5 y (46% males) Pre-AID treatment: 37% on MDI therapy, 63% on CSII Design: Retrospective. MiniMed 780G Comparator: manual mode (T0), and 3 and 6 m after auto-mode f/up: 6 m Region: Italy	% in auto mode was stable: 95.7% ± 8.8% at 3 m and 92.9 ± 11.1% at 6 m TDI/Kg increased from baseline 0.7 ± 0.2 to 0.8 ± 0.2 (6 m, P < 0.001) Basal insulin decreased from 43.6% ± 10.9% to 39.2% at 6 m (<0.001) Bolus insulin increased from 56.4% ± 10.9% to 60.8 ± 7.7% at 6 m (P < 0.001) % of autocorrection in TDI: 24.1% ± 9.9% at 3 m and remained stable	Weight: from 43.2 ± 18.2 to 45.1 ± 18.2 kg (<0.01) BMI and BMI z-score: no change (% = 0.050) Number of meals increased at 3 and 6 m (4.4 ± 1.2 vs. 5.0 ± 1.5, % = 0.002 and 5.1 ± 1.7, P < 0.001) No changes in CHO consumed, meal composition, and calories intake	Limitations: retrospective; dietary evaluation before and after the switch was available only for a subgroup of patients; no pubertal assessment ⨁⨁⨁⊖ MODERATE
Piccini B et al.²⁶			TIR increased after 3 m (76.6% ± 8.5% P < 0.001) and was maintained at 6 m (75.7% ± 7.8%, P < 0.001). In manual mode it was 66.7% ± 13.2% CV did not change GMI decreased in auto mode: 7.1% ± 0.5% in manual mode vs. 6.7% ± 0.3% at 3 m and 6.7% ± 0.5% at 6 m HbA1c decreased in auto mode: 7.2% ± 0.7% in manual mode vs 6.9% ± 0.7% (P = 0.002) at 3 m and 6.9 ± 0.6% (P = 0.003) at 6 m
Criego AB et al.²⁷	112 children (not considered 128 adolescents and adults) Age children 6–13.9 y Pre-AID treatment: CSII 89.3%, MDI 11.6% Design: Prospective; Omnipod 5 Comparator: pre- 15 m post f/up: 15 m Region: USA	% in Auto mode: 96.1% (IQR 92.7, 97.6) and remained stable during f/up TDI/kg: increased, from 0.85 ± 0.24 U/kg with standard therapy to 1.04 ± 0.28 U/kg (P < 0.0001 from baseline) at 15 m Number of boluses/day: 7.1 ± 2.7 at baseline and 5.9 ± 2.0 at 15 m (P < 0.0001) Bolus %: from 56.9% ± 10.1% baseline to 46.5% ± 7.7% (P < 0.0001) % autocorrection: n.a.	BMI z-score: increased slightly in children from 0.42 ± 0.79 at baseline to 0.52 ± 0.83 (P = 0.0248) at 15 m.	Strengths: large sample size Limitations: no value of CHO/day, the lack of a control group, no pubertal assessment ⨁⨁⨁⊖ MODERATE
Criego AB et al.²⁷			TIR: increased by 13.5% ± 12% (52.4% ± 15.6% to 65.9% ± 8.9%, P < 0.0001) TAR: significant decreased from 45.5% ± 16.6% with standard therapy to 32.3% ± 9.5% (P < 0.0001) TBR: remained unchanged HbA1c was reduced by 0.5% ± 0.7% (5.5–7.7 mmol/mol, P < 0.0001): from 7.7% ± 0.9% (61–9.8 mmol/mol) to 7.2% ± 0.7% (55–7.7 mmol/mol)
Rachmiel M et al.²⁸	22 participants Age: median age 13.9 (range 7.7–24 years) Pre-AID treatment: CSII 86.4%, MDI Design: Retrospective, Minimed 780G Comparator: First 2 week, 1, 3, 6 m f/up: 6 m Region: Israel	% in Auto mode: not reported TDI (U): increased significantly from baseline 36 (22, 59) to 6 m: 49 (42, 57) U (p = 0.002) Basal insulin per day: decreased from baseline 44 (37, 51) to 40 (35, 43) % at 6 m (P = 0.028) Bolus of insulin per day: from 56 (49, 53) to 60% (P = 0.11) at 6 m % of autocorrection: at the end of the study 24 (18, 33) %	BMI SDS from 0.53 (−0.37, 0.82) to 0.53 (−0.24, 1.02), P = 0.33 WT SDS 0.74 (−0.35, 1.12) to 0.71 (−0.1, 0.23), P = 0.30 Mean daily CHO g: from 140.5 (94, 183) to 149 (113, 192), P = 0.31	Strengths: data on pubertal status Limitations: retrospective study, small sample size, short f/up ⨁⨁⊖⊖ LOW
Rachmiel M et al.²⁸			TIR: increased significantly from 65% to 75%(P > 0.001) TAR: decreased significantly from 28% to 22% (P = 0.047) TBR: decreased significantly from 4.0% to 2.5% (P = 0.10) GMI: not changed significantly during the study
DeSalvo DJ et al.²⁹	80 participants Age: 2–5.9 years Pre-AID treatment: CSII 85%, MDI 15% Design: Prospective; effectiveness of the Omnipod® 5 Comparator: 14 d, 3 m, 15 m f/up: 24 m Region: USA	% in Auto mode: no change during f/up TDI/Kg: no significant difference from baseline 0.69 ± 0.18 U/kg with standard therapy to 0.71 ± 0.15 U/kg (P > 0.05) during the pivotal phase at 3 m, but there was an increase to 0.80 ± 0.17 U/kg (P < 0.0001) during the extension phase (16.7 ± 3.2 m) Bolus %: from 58.7% ± 13.6% during standard therapy to 53.5% ± 8.1% during the extension phase (P = 0.0001) % of autocorrection: n.a.	BMI or BMI z-score: 0.74 ± 0.95 at baseline, 0.76 ± 0.89 at 3 m, 0.04 ± 6.41 at 16.7 ± 3.2 m (P = 0.40) CHO/day: n.a.	Strength: long f/up Limitations: single-arm design, no value of CHO/day ⨁⨁⨁⊖ MODERATE
DeSalvo DJ et al.²⁹			TIR: increased by 10.0% ± 10.1% (57.2% ± 15.3% to 67.2% ± 9.3%, P < 0.0001) TAR: decreased by 9.2% ± 10.9% from 39.4% ± 16.7% with standard therapy to 30.2% ± 10.1% (P < 0.0001) during the extension (16.7 ± 3.2 m) TBR: decreased from 1.66% (0.40, 4.21) with standard therapy to 1.46% (0.84, 2.91) during the extension (P = 0.03) HbA1c: reduced by 0.4% ± 0.7% (4.4 ± 7.7 mmol/mol, P < 0.0001), decreasing from 7.4% ± 1.0% (57 ± 10.9 mmol/mol) at baseline to 7.0% ± 0.7% (53 ± 7.7 mmol/mol) after 15 m
Lombardo F et al.³⁰	111 children and adolescents Age: 13.1± 3.1 y (54.1% female) Pre-AID treatment: CSII (66.7%) or MDI (33.3%) Design: Prospective; MiniMed 780G Comparator: T0 (baseline), 3, 6 m f/up: 6 m	% in Auto mode: increased from 92.8 ± 9.7 at 0–3 m to 93.1%± 10.3% at 6 m TDI/Kg: baseline 0.82 ± 0.38 IU/kg, at 3 m 0.85 ± 0.30, at 6 m 0.84 ± 0.22 (n.s.) Basal insulin delivery (%): baseline: 51.3 ± 7.1, at 3 m: 41.9 ± 6.6, at 6 m: 41.8 ± 6.9 Bolus (%): baseline: 48.7 ± 6.8, at 3 m: 58.1 ± 6.6, at 6 m: 57.6% ± 8.1% (P not available) % of autocorrection: n.a.	BMI z-score: at 3 m 0.76 ± 0.95, at 6 m 0.81 ± 1.03; no significant changes (no P value reported) WT z-score: at 3 m 0.53 ± 1.16, at 6 m 0.60 ± 1.20 CHO/Kg/day intake: remained unchanged (at 3 m 4.2 ± 1.8, at 6 m 4.0 ± 1.5)	Strength: large sample size Limitations: short f/up, no pubertal assessment ⨁⨁⨁⊖ MODERATE
Lombardo F et al.³⁰			TIR: increased significantly from 63.5± 13.1 to 74.8± 9.0 after 6 m (P < 0.001) TAR: decreased significantly from 34.1 ± 14.1to 22.5 ± 9.6 after 6 m (P < 0.001) TBR: increased not significantly from 2.4 ± 2.8 to 2.7 ± 2.2 after 6 m (P = 0.095) HbA1c: was 6.9% ± 0.7% corresponding to a relative reduction of 4.2% from baseline (T0) GMI: decreased significantly from 7.3 ± 0.7 to 6.8% ± 0.4% after 6 m (P < 0.001)
Seget S et al.²²	50 children-adolescents Age: 9.9 ± 2.4 y (boys 48%) Pre-AID treatment: PLGM Design: Prospective. From PLGM to AHCL Minimed 780G. Comparator: T0, 6, 12, 24 m f/up: 24 m Region: Poland	% in Auto mode: 94.26 ± 6.28 to 97.58 ± 3.21 (P = 0.03) TDI/kg: 0.67 ± 0.20 at baseline to 0.84 ± 0.22 (P < 0.001) after 24 m Basal insulin (%): 34.88 ± 6.91 to 34.82 ± 6.21 (P = 0.49) Bolus insulin (%): 65.06 ± 6.90 to 65.17 ± 6.21 (P = 0.46) % of autocorrection in TDI: increased significantly (11.24 ± 6.08 at baseline to 14.61 ± 7.00 after 24 m, P < 0.005)	BMI z-score: no change (P > 0.05) but data not reported The average value of CHO/day increased (from 184.74 to 215.20) P = 0.0288)	Strengths: f/up duration Limitations: no HbA1c data lack of control group ⨁⨁⨁⊖ MODERATE
Seget S et al.²²			TIR (14 days): no change during f/up
Henry Z et al.³¹	48 adolescents Age: 10–17 years (55% females) Pre-AID treatment: CSII (100%) Design: Retrospective. MiniMed 780G or Tandem t:slim X2 control-IQ Comparator: at baseline and 1–3-6–9-12 m post initiation f/up:12 m Region: France	% in Auto mode: not reported TDI/kg: 46.57 (54.6–58.4) at baseline to 49.51 (45.9–53.1) after 1 year (p < 0.01). These data were derived from 780G users: 45.66 (54.6–58.4) at baseline to 49.32 (44.9–53.9) after 1 year Bolus insulin: baseline 26.61 (24.4–28.8), after 1 y: 28.02 (26.9–32.6), P < 0.01, due to 780G group (P < 0.001) % of autocorrection in TDI increased not significantly during f/up from 23.23 (21.3–25.4) to 1 year: 23.76 (20.9–26,6), P > 0.05 but was higher in the 780G group: 30.45 (27.38–33.5) at 1 year vs TCIQ at 1 year: 8.58 (6.9–10.3), P < 0.01	BMI baseline 24.74 (24.1–25.4), in 780G group: 25.14 (24.4–25.9) and increase after 3 m Median WT increased significantly after 1 year with +2.21 kg (1.07; 3.33) (n = 74, p < 0.001) for 780G users and –0.85 kg (–5.5; 3.8) for CIQ users (n = 36). The increase in WT on 780G users was correlated to initial WT (P < 0.05), baseline TAR >250 (P = 0.016), baseline TIR (P < 0.01) and baseline GMI (P < 0.05) Baseline BMI did not influence the performance of AID systems CHO/day (g) announcements decreased during f/up from 169 (159–179) to 151 (139–163) after 1 year	Limitations: single center, only half users reached 1 y f/up, no pubertal assessment. Data reported in supplementary material divided by BMI classes, without the overall ones ⨁⨁⊖⊖ LOW
Henry Z et al.³¹			TIR after 1 year was higher than baseline (+14%). HbA1c decreased 0.7% and GMI by 0.6% 780G users had better CGM results than control-IQ users but similar HbA1c levels GMI decreased by 0.6% after 360 days. At baseline 16.9% had GMI <7% vs. 40.3 after 360 days HbA1c decreased by 0.7% after 360 days. At baseline 20.3% had HbA1c <7% vs. 42.5 % after 360 days TAR >180 mg/dL was significantly reduced from 40.05% (38.03–42.06) to 26.92% (24.95–28.88) after 360 days (n = 116, P < 0.001) No significant changes in TBR CV was reduced from 37.26% (36.4–38.1) at baseline (n = 231) to 35.07% (6.9–7.1) after 360 days (n = 122, p < .001)
Atik-Altinok Y et al.³²	22 children-adolescents <18 years Age: 12.5 ± 4.8 years (48.3% female) Pre-AID treatment: CSII (79%) or MDI Design: Prospective; baseline vs 6 m in MiniMed 780G 3 Day food intake diary at T0, 3 and 6 m Comparator T0, 3, 6 m f/up: 6 m Region: Turkey	% in auto mode: 100 at 3 m, 98 at 6 m TDI/Kg: available only at baseline, 0.7 ± 0.3 Bolus from 60.0 % (17.5) in manual mode to 62.0 (11.0) in auto mode after 6 m (P > 0.05) % of autocorrection bolus: 18% (11.0) at 3 m, 19% (7.5) at 6 m (P < 0.05)	BMI: no change during the f/up (p = 0.25) CHO/day (g): not reported No statistically significant differences in macronutrient distribution during the f/up. Fat intake increased during the f/up (not significant); saturated fat intake was above, and fiber intake was lower than the recommendations for both baseline and f/up. The mean number of meals and snacks did not change during the follow-up	Limitations: small sample size, short f/up, no value of CHO/day, no pubertal assessment ⨁⨁⊖⊖ LOW
Atik-Altinok Y et al.³²			TIR: increased from 79.0 (15.5)% to 81.0 (6.5)% (P < 0.05) TAR: significantly decreased GMI: decreased from 6.6 (0.4) to 6.5 (0.3) (p < 0.05) HbA1: from 6.9% (1.2) to 6.7% (1.3) after 6 months
Bismuth E et al.³³	117 children in the Free-Life Kid AP study Age: 10 ± 1.1 years (6.8–14) Pre-AID treatment: CSII (100%) Design: RCT; Control IQ 24/7 or only evening and night Data pre and 24 m post f/up: 24 m (extension of the study) Region: France	% in Auto mode: n.a. TDI/Kg: 0.81 ± 0.15 in the pre-pubertal group from 9 to 24 m, and 0.87 ± 0.22 U/kg/day in the pubertal group from 9 to 24 m. Basal/bolus ratio: n.a. % autocorrection: n.a.	BMI Z scores were statistically noninferior at 24 m compared with baseline, with a mean difference of 0.07 ± 0.49 (95% CI 0.03 to 0.23, P = 0.0004 for noninferiority, using margin of 0.2 as clinically relevant difference) CHO/day (g): not reported	Strengths: large sample size, long f/up Limitations: mainly prepuberal children and 18% of participants entered puberty after 9 m and 54% during the 2 years of f-up, no value of CHO/day ⨁⨁⨁⊖ MODERATE
Bismuth E et al.³³			TIR improvement was maintained over the 2 y of f/up (P < 0.0001 for noninferiority test, with margin of 5% as clinically relevant difference) TBR remained stable (P < 0.0001 for noninferiority test, with margin of 4% as clinically relevant difference) HbA1c improvement (7%) was maintained during the extension period
Rapini N et al.³⁴	19 children 11 boys; 8 girls Age: 4.8 (4.4, 5.5) y Pre-AID treatment: MDI or CSII (% not reported) Design: Prospective;780G and TCIQ vs MDI and CSII Comparator: 1, 3, 6 m f/up: 6 m Region: Italy	% in Auto mode: n.a. TDI/kg: from baseline 0.6 (0.5, 0.8) increment at 1 and 3 m (0.7U p = 0.02 and 0.8 P = 0.004), without a significant change at the end of f/up period compared to baseline (0.6 IU/kg/day, P = 0.4) Bolus (%): from 47.1 (45, 55.3) to 54.7 (50.4, 58.8) (P = 0.22) % of autocorrection: n.a.	BMI: no change (P = 0.27) CHO/day: remained stable during the study period (p = 0.12, 0.43, and 0.62, respectively)	Strengths: real life setting Limitations: small sample size, short f/up, no value of CHO/day ⨁⨁⊖⊖ LOW
Rapini N et al.³⁴			TIR: increased significantly during the 6 m of follow-up, reaching a difference from baseline of +10% (p = 0.03) TAR: decreased significantly from 25% (21, 27.5) to 21% (18.3, 23.8) (P = 0.02) TBR: increased from 2% (1, 3) to 2.5% (2, 4) (P = 0.04) HbA1c: from 56.3 (52, 62.5) to 55 (44.8, 58.7) mmol/mol (P = 0.0007) GMI: from 7.2 (6.9, 7.8) to 7 (6.8, 7.2) % (P = 0.03)
Passanisi S et al.³⁵	368 children and adolescents Age: median 12.9 years Pre-AID treatment: CSII or MDI (33.4%) Design: Prospective; Minimed 780G baseline vs 12 m f/up: 12 m Region: Italy	% in Auto mode: from 95.0 ± 9.6 at +15 d to 95.7 ± 8.0 (n.s.) TDI/kg: baseline 0.82 ± 0.22, at 12 m: 0.93 ± 0.40 (P = 0.036) Basal insulin doses: 52.2% ± 5.3% at baseline, 40.7 ± 7.3 (P < 0.001) Bolus insulin doses: 47.8% ± 4.9% at baseline, 59.3 ± 7.4 (P < 0.001) % of autocorrection: 28.0 ± 11.6 to 30.8 ± 12.2 (P = 0.625)	BMI SDS: no change between baseline and 12 m: 0.54 ± 1.1 vs. 0.57 ± 1.1, P = 0.162) Daily boluses (No.): from 4.3 ± 1.3 to 4.7 ± 1.6 (P = 0.601) Daily CHO (g/kg): from 4.2 ± 2.0 to 3.9 ± 1.7 (P = 0.948)	Strengths: large sample size Limitations: no pubertal assessment ⨁⨁⨁⊖ MODERATE
Passanisi S et al.³⁵			TIR: increased significantly from 62.8 ± 14.1% to 75.3 ± 8.7% (P < 0.001) TAR: decreased significantly from 34.7 ± 15.0 to 22.4 ± 9.3 after 12 m (P < 0.001) TBR: increased not significantly from 2.5 ± 2.7 to 2.4 ± 2.3 after 12 m (P = 0.342) HbA1c: from 7.4% ± 0.9% to 6.9% ± 0.6% (P < 0.001) GMI: decreased significantly from 7.2 ± 0.5 to 6.8% ± 0.4% after 12 m (P < 0.001)

High quality of evidence: ⨁⨁⨁⨁, Moderate: ⨁⨁⨁⊖, Low: ⨁⨁⊖⊖, Very Low: ⨁⊖⊖⊖.

AHCL, advanced hybrid closed loop; AID, automated insulin delivery; BMI, body mass index; CHO, carbohydrate; CI, confidence interval; CSII, continuous subcutaneous insulin infusion; CV, coefficient of variation; d, day; f/up, follow up; GMI, glucose management indicator; HCL, hybrid closed loop; y, year/s, m, month; MDI, multiple daily injections; n, night; n.s., not significant; PLGM, predictive low glucose management; RCT, randomized controlled trial; SDS, standard deviation score; Tandem CIQ, Tandem Control IQ; TAR, time above range; TBR, time below range; TDI, total daily insulin; TIR, time in range; T1D, type 1 diabetes.

Table 2.

Summary of the Evidence for AID Impact on BMI and TDI During Follow-Up Compared with Baseline or Pre-AID Treatment

Reference	Population and study design	% in auto mode	TDI/Kg	% Basal	% Bolus (user initiated + autocorrection)	% Autocorrection bolus / trend	BMI (WT)	BMI z-score	Weight	CHO intake	Level of evidence
Moderate level of evidence
Seget S et al.²⁰	No. 50, age 5.4–16.8 y, prospective (12 m) 780G baseline vs. 12m	↑	↑	=	=	n.a./↑	n.a.	=	n.a.	=	Moderate
Piccini B et al.²⁶	No. Eighty-three, age 13 ± 4.5 y, Retrospective (6 m) 780G manual mode vs. auto mode	=	↑	↓	↑	24.1/=	=	=	↑	=	Moderate
Criego AB et al.²⁷	No. 112, Age 6–13.9 y, Prospective (15 m) Pre-AID vs Omnipod 5 baseline vs 24 m	=	↑	↑	(↓ user-initiated)	n.a.	n.a.	↑	n.a.	n.a.	Moderate
DeSalvo DJ et al.²⁹	No. 80, Age 2–5.9 y, Prospective (24 m) Pre-AID vs. Omnipod 5 baseline vs. 15 m	=	↑	↑	(↓ user-initiated)	n.a	=	=	n.a.	n.a.	Moderate
Lombardo F et al.³⁰	No. 111, age 13.1± 3.1 y, Prospective (6 m) 780G baseline vs. 6 m	↑	=	↓	↑	n.a	n.a.	=	= (SDS)	=	Moderate
Seget S et al.²²	No. 50, age 9.9 ± 2.4 y, Prospective (24 m) From PLGM to 780G	↑	↑	=	=	11.2–14.6/↑	n.a.	=	n.a.	↑	Moderate
Bismuth E et al.³³	No. 117, age 10 ± 1.1 y, RCT (24 m) TCIQ 24/7 vs. only evening and night	n.a.	n.a	n.a.	n.a.	n.a.	n.a.	=	n.a.	n.a.	Moderate
Passanisi S et al.³⁵	No. 368, age median 12.9 y, Prospective 780G baseline vs. 12 m	=	↑	↓	↑	28.0–30.8/=	n.a.	=	n.a.	=	Moderate
Low level of evidence
Tornese G et al.²¹	No. 44, age 13.3 y (range 2–21 y), Retrospective (6 m) 780G vs. 670G	=	↑	=	=	26–33/=	n.a.	=	n.a.	=	Low
Rachmiel M et al.²⁸	No. 22, age 13.9 y (range 7.7–24), Retrospective (6 m) 780G baseline vs. 6 m	n.a.	↑	↓	=	24/n.a.	n.a.	=	= (SDS)	=	Low
Henry Z et al.³¹	No. 48, Age 13 ± 4.5 y, Retrospective (12 m) 780G and TCIQ baseline vs. 6 m	n.a.	↑ (only 780G)	↓	↑ (only 780G)	23.2–23.7/↑	↑ (only 780G)	n.a.	↑(only 780G)	↓	Low
Atik-Altinok Y et al.³²	No. 22, Age 12.5 ± 4.8 y, Prospective (6 m) 780G baseline vs. 6 m	=	n.a.	=	=	18–19/↑	n.a.	=	n.a.	n.a.	Low
Rapini N et al.³⁴	No. 19, Age 4.8 (4.4, 5.5) y, Prospective (6 m) 780G and TCIQ vs. 6 MDI or CSII	n.a.	↑/-	=	=	n.a.	=	n.a.	n.a.	=	Low

AHCL, advanced hybrid closed loop; BMI, body mass index; CHO, carbohydrate; CI, confidence interval; CSII, continuous subcutaneous insulin infusion; d, day; f/up, follow up; HCL, hybrid closed loop; RCT, randomized controlled trial; y, year/s; m, month; n.s., not significant; MDI, multiple daily injections; n, night; PLGM, predictive low glucose management; SDS, standard deviation score; Tandem CIQ, Tandem Control IQ; TDI, total daily insulin; T1D, type 1 diabetes; WT, weight; ↓, reduced; ↑, increased; =, no significant change; ↑/-, heterogeneity in the response, mixed results.

Included studies were one RCT,³³ four retrospective,^21,26,28,31 and eight prospective in design. The setting was Europe in nine studies, Israel in one, and the United States in two. Eleven studies were conducted on the AID model Minimed 780G® and Tandem Control IQ®, and two were conducted using Omnipod 5®.^27,29 Sample size ranged from 19 to 368 subjects with AIDs (median 50 individuals), and study duration varied from 6 to 24 months (median 12 months); study participants were younger than 6 years in two studies^29,34 and children adolescents older than 6 years in the other studies. The pubertal stage was considered a variable related to TDI in two 11 studies that included subjects older than 6 years.^28,33

BMI measure was included as a secondary outcome in all but three studies.^20,26,32 A comparison between pre-AID treatment and AID in auto mode was reported in three studies,^27,29,32 whereas a comparison during the follow-up period in the others. Eight studies were classified according to the GRADE system as moderate evidence level, five as low, and none as high. Tables 1 and 2 report the main results of these studies, divided by the outcome, and here we report the main results according to moderate level of evidence studies.

BMI, weight, and CHO intake

BMI z-score or BMI was found unchanged in all but one of the studies²⁷ during the 6–24 months of follow-up. In the study of Criego et al.,²⁷ the BMI z-score increased from 0.42 ± 0.79 at baseline to 0.52 ± 0.83 (P = 0.0248) at 15 months, but data on CHO intake was not reported. Weight was reported in only one study²⁶ and increased from 43.2 ± 18.2 kg to 45.1 ± 18.2 kg (P < 0.01) during the 6-month follow-up without increased CHO intake.

Four studies were included in a metanalysis to assess the impact of AID on BMI z-score change after 12 months of usage.^20,27,29,35 The observed differences ranged from −0.15 to 0.12 z-score. Figure 2 shows the meta-analysis results for all the studies, and the mean difference estimated from the random-effects model was 0.0174 (95% confidence interval [CI]: −0.095 to 0.130). Therefore, the average outcome did not differ significantly from zero (z = 0.303, P = 0.762). According to the Q-test, the outcomes were not heterogeneous [Q(3) = 1.823, P = 0.610, tau² = 0.01, I² = 0%].

FIG. 2.

Forest plot showing the pooled difference in BMI z-score in youth from baseline to 12 months of AID use. Data are reported as median and 95% CI. Measures of heterogeneity: Q(3) = 1.823, P = 0.610, tau² = 0.0, I² = 0%. BMI, body mass index; AID, automated insulin delivery.

After the switch to AIDs, the number of meals did not increase in three studies that considered this outcome,^21,32,35 while in one study, it increased at 3 and 6 months.²⁶

CHO intake was unchanged in seven out of eight studies, while in the study of Seget et al., 2024,²² the average value of CHO/day increased from 184.74 to 215.20 g (P = 0.0288) during the 2 years follow-up, but a control group was not available, and BMI remained unchanged (P > 0.05). No differences in meal composition (% of fat, saturated fat) or calories intake or snacks were highlighted.^26,32

TDI, user-initiated, and autocorrection bolus

TDI pro-kg increased from baseline to the end of the follow-up in six out of eight studies, whereas it was unchanged in another and not reported in the other one.³⁰ The basal/bolus ratio was reduced or unchanged after AID start in five studies,^{20,22,26,30,35} as the bolus component was increased due to the rate of autocorrection bolus volume.^26,30,35 The rate of autocorrection bolus was about 20%–30% of total bolus insulin and increased during follow-up in two studies,^20,22 whereas it remained unchanged during follow-up in others.^26,35 The rate was related to the high use of the auto-mode, which increased in three studies^20,22,30 but not in the other four.^26,27,29,35

The % of basal was increased (and % of bolus decreased) in two studies after Omnipod 5®^27,29 associated with reduced user-initiated bolus.

Four studies were included in a meta-analysis to assess the impact of AID on TDI/kg change after 12 months.^22,27,29,35 The observed differences ranged from 0.34 to 0.73 U/kg. Figure 3 shows the meta-analysis results for all the studies, and the mean difference estimated from the random-effects model was 0.54 (95% CI: 0.34 to 0.75). Therefore, the average outcome differed significantly from zero (z = 5.17, P < 0.001). According to the Q-test, the true outcomes were not heterogeneous [Q(3) = 7.785, p = 0.051, tau² = 0.0245, I² = 57.91%].

FIG. 3.

Forest plot showing the pooled difference in TDI/kg in youth from baseline to 12 months of AID use. Data are reported as median and 95% CI. Measures of heterogeneity: Q(3) =7.785, P = 0.051, tau² = 0.0245, I² = 57.91%.

Discussion

This systematic review and meta-analysis summarize current evidence, graded with the GRADE approach, about the impact of AIDs on BMI z-score in the pediatric population.

BMI z-score is a relative measure of BMI that takes into account age- and sex-specific differences in BMI growth, and BMI curves have been created for children; this measure is a simple and widely utilized screening tool for overweight and obesity in children and adolescents, even it presents limited accuracy as surrogate of adiposity, compared with triceps skinfold, waist circumference, electrical bioimpedance, and DXA-calculated fat mass.³⁶

Only 13 studies respected the inclusion criteria, and BMI z-score was a secondary outcome in 10 studies, while glucose metrics were the main aim in most studies. The main result of this review is the stability of BMI z-score in children and adolescents throughout the follow-up period, which ranged from 6 to 24 months. This finding challenges our original hypothesis of a progressive increase in BMI over time, as reported on CSII therapy,⁸ due to the possibility of efficiently administering insulin several times per day, leading to a more flexible diet, with the risk of unhealthy nutritional choices and weight gain. The metanalysis reported a non-significant mean change of 0.02 in the BMI z-score from baseline to 12 months after AID start.

The AIDs can cover unannounced CHO with a tolerable glucose excursion in children and adolescents,²¹ as auto-correction intervenes when the user inaccurately counts CHO at meals or when meal boluses are delayed or skipped/forgotten. The last one is a widespread problem, especially among adolescents with T1D. However, even if AIDs allow more freedom in food choices, our data confirm that in real life with AIDs, CHO intake was not increased, and there was no deterioration to less healthy eating choices or more frequent snacking. No differences in meal composition (% of fat, saturated fat) or calories intake or snacks were highlighted, excluding qualitative food choices worsening.^26,32 A contribution to BMI stability could also derive from the lower intake of glucose, due to significant reduction in hypoglycemic episodes associated with AIDs, and in case of hypoglycemia <70 mg/dL less glucose (e.g., 5 to 10 g or 0.1 g/kg) is required for AID users.^37,38 In parallel, the meta-analysis found a significant increase of 0.54 U/kg of BW in the TDI dose. According to this review, it can be due mainly to two reasons: i) the increase in the proportion of automated corrections for inaccuracies in CHO estimation or missed boluses;²⁶ ii) the physiological increase in insulin need in the studied population during the 1–2 years of follow-up, as part of the adolescents might have entered puberty, as reported in a few studies included in this review,^28,33 while the data were unknown for others. In this review, increased TDI was related to an increased rate of autocorrection boluses during follow-up, but the authors found no change in BMI z-score.^22,26,30,35 These data agree with previous findings that subjects on CSII associated the risk of weight gain with the rising percentage of basal insulin in the TDI;⁸ one mechanism that leads to increased weight gain is related to increased anabolic activity of insulin, which reduces lipolysis and protein catabolism, promotes lipogenesis and protein formation, leading to increased fat accumulation.³⁹

The strengths of this systematic review are: i) it is the first one presenting a meta-analysis of BMI z-scores in subjects with AID; ii) we were able to report TDI, basal/bolus ratio, and % of autocorrection that could predict change in BMI z-scores, for most studies.

Some limitations of the studies included in this review need to be acknowledged: i) first, the sample size of most studies was small, between 20 to 50 subjects, primarily including Caucasian people, and therefore, the results might not be representative of the whole population of youths with AIDs; ii) some studies did not evaluate CHO intake and the meal composition, others did not consider pubertal stage: all these are determinant factors for insulin requirement evaluation; iii) among the 13 studies, only one compared AIDs with the best currently available contemporary therapy (HCL),²¹ while the others did not enroll a control group, and the study design was a comparison between baseline and follow-up; iv) BMI measure was included as secondary outcome in all but three studies^20,26,32 and some studies can be underpowered to demonstrate significant differences on this outcome.

Conclusions

AIDs are not fully automated, and patients must receive retraining in CHO counting periodically, as well as health care professionals verify that meals are announced and if the insulin/CHO settings are still reliable. However, AIDs may be beneficial not only for people carefully following diabetes self-care but also for those who less accurately assess food or forget meal bolus. AIDs can also improve glucose control in these subjects, as autocorrection bolus covers delayed or skipped/forgotten meal bolus, increasing TDI/kg, which has no impact on BMI z-score. More studies are needed on the pediatric population, setting BMI z-score as the primary outcome and considering a control group, carbohydrate intake, pubertal stage, and all the pump settings.

Footnotes

Authors’ Contributions

R.F. and E.M. made a substantial contribution to the design of this literature review, in the acquisition of data, their interpretation and analysis as well as in the writing of the article. S.F., D.R.A., and F.D.C. selected the articles for this literary review. C.G., M.M., R.F., and E.M. contributed to the analysis of studies included in the review. R.P. performed the statistical analysis. E.M. performed a critical revision of the article and performed a thorough proofreading of the article. All the authors have approved publishing the version.

Data Availability Statement

All databases generated for this study are included in the article.

Author Disclosure Statement

The authors have no conflicts of interest to declare.

Funding Information

The authors have no conflicts of interest to declare. This research was not supported by funding.

Supplementary Material

Abbreviations Used

References

Baskaran

, Volkening

, Diaz

, et al. A decade of temporal trends in overweight/obesity in youth with type 1 diabetes after the Diabetes Control and Complications Trial. Pediatr Diabetes, 2015; 16(4):263–270; doi: 10.1111/pedi.12166

Ciężki

, Kurpiewska

, Bossowski

, et al. Multi-faceted influence of obesity on type 1 diabetes in children—from disease pathogenesis to complications. Front Endocrinol (Lausanne), 2022; 13:890833; doi: 10.3389/fendo.2022.890833

Liu

, Lawrence

, Davis

, et al.; SEARCH for Diabetes in Youth Study Group. Prevalence of overweight and obesity in youth with diabetes in USA: The SEARCH for Diabetes in Youth study. Pediatr Diabetes, 2010; 11(1):4–11; doi: 10.1111/j.1399-5448.2009.00519.x

Maffeis

, Birkebaek

, Konstantinova

, et al.; SWEET Study Group. Prevalence of underweight, overweight, and obesity in children and adolescents with type 1 diabetes: Data from the international SWEET registry. Pediatr Diabetes, 2018; 19(7):1211–1220; doi: 10.1111/pedi.12730

The DCCT Research Group. Weight gain associated with intensive therapy in the diabetes control and complications trial. Diabetes Care, 1988; 11(7):567–573; doi: 10.2337/diacare.11.7.567

DuBose

, Hermann

, Tamborlane

, et al.; Type 1 Diabetes Exchange Clinic Network and Diabetes Prospective Follow-up Registry. Obesity in youth with type 1 diabetes in Germany, Austria, and the United States. J Pediatr, 2015; 167(3):627–632.e1-4; doi: 10.1016/j.jpeds.2015.05.046

Purnell

, Zinman

, Brunzell

, DCCT/EDIC Research Group. The effect of excess weight gain with intensive diabetes mellitus treatment on cardiovascular disease risk factors and atherosclerosis in type 1 diabetes mellitus: Results from the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications Study (DCCT/EDIC) study. Circulation, 2013; 127(2):180–187; doi: 10.1161/CIRCULATIONAHA.111.077487

Boucher-Berry

, Parton

, Alemzadeh

. Excess weight gain during insulin pump therapy is associated with higher basal insulin doses. J Diabetes Metab Disord, 2016; 15:47; doi: 10.1186/s40200-016-0271-5

Yeh

, Brown

, Maruthur

, et al. Comparative effectiveness and safety of methods of insulin delivery and glucose monitoring for diabetes mellitus: A systematic review and meta-analysis. Ann Intern Med, 2012; 157(5):336–347; doi: 10.7326/0003-4819-157-5-201209040-00508

10.

REPOSE Study Group. Relative effectiveness of insulin pump treatment over multiple daily injections and structured education during flexible intensive insulin treatment for type 1 diabetes: Cluster randomised trial (REPOSE). Bmj, 2017; 356:j1285; doi: 10.1136/bmj.j1285

11.

Propper-Lewinsohn

, Shalitin

, Gillon-Keren

, et al. Glycemic variability and disordered eating among adolescents and young adults with type 1 diabetes: The role of disinhibited eating. Diabetes Technol Ther, 2025; 27(2):113–120; doi: 10.1089/dia.2024.0267

12.

Garg

, Weinzimer

, Tamborlane

, et al. Glucose outcomes with the in-home use of a hybrid closed-loop insulin delivery system in adolescents and adults with type 1 diabetes. Diabetes Technol Ther, 2017; 19(3):155–163; doi: 10.1089/dia.2016.0421

13.

Tauschmann

, Thabit

, Bally

, et al.; APCam11 Consortium. Closed-loop insulin delivery in suboptimally controlled type 1 diabetes: A multicentre, 12-week randomised trial. Lancet, 2018; 392(10155):1321–1329; doi: 10.1016/S0140-6736(18)31947-0

14.

Brown

, Forlenza

, Bode

, et al.; Omnipod 5 Research Group. Multicenter trial of a tubeless, on-body automated insulin delivery system with customizable glycemic targets in pediatric and adult participants with type 1 diabetes. Diabetes Care, 2021; 44(7):1630–1640; doi: 10.2337/dc21-0172

15.

Brown

, Kovatchev

, Raghinaru

, et al.; iDCL Trial Research Group. Six-month randomized, multicenter trial of closed-loop control in type 1 diabetes. N Engl J Med, 2019; 381(18):1707–1717; doi: 10.1056/NEJMoa1907863

16.

Seget

, Rusak

, Polanska

, et al. Prospective open-label, single-arm, single-center follow-up study of the application of the advanced hybrid closed loop system in well-controlled children and adolescents with type 1 diabetes. Diabetes Technol Ther, 2022; 24(11):824–831; doi: 10.1089/dia.2022.0148

17.

Phillip

, Nimri

, Bergenstal

, et al. Consensus recommendations for the use of automated insulin delivery technologies in clinical practice. Endocr Rev, 2023; 44(2):254–280; doi: 10.1210/endrev/bnac022

18.

Franceschi

, Mozzillo

, Di Candia

, et al. A systematic review on the impact of commercially available hybrid closed loop systems on psychological outcomes in youths with type 1 diabetes and their parents. Diabet Med, 2023; 40(9):e15099; doi: 10.1111/dme.15099

19.

Tornese

, Carletti

, Giangreco

, et al. Carbohydrate tolerance threshold for unannounced snacks in children and adolescents with type 1 diabetes using an advanced hybrid closed-loop system. Diabetes Care, 2022; 45(6):1486–1488; doi: 10.2337/dc21-2643

20.

Seget

, Jarosz-Chobot

, Ochab

, et al. Body mass index, basal insulin and glycemic control in children with type 1 diabetes treated with the advanced hybrid closed loop system remain stable - 1-year prospective, observational, two-center study. Front Endocrinol (Lausanne), 2022; 13:1036808; doi: 10.3389/fendo.2022.1036808

21.

Tornese

, Buzzurro

, Carletti

, et al. Six-month effectiveness of advanced vs. standard hybrid closed-loop system in children and adolescents with type 1 diabetes mellitus. Front Endocrinol (Lausanne), 2021; 12:766314; doi: 10.3389/fendo.2021.766314

22.

Seget

, Chobot

, Tarasiewicz

, et al. Glycemic control in children with type 1 diabetes treated with the advanced hybrid closed loop system 2-year prospective, observational, two-center study. Front Endocrinol (Lausanne), 2024; 15:1332418; doi: 10.3389/fendo.2024.1332418

23.

Schünemann

, Cuello

, Akl

, et al.; GRADE Working Group. How ROBINS-I and other tools to assess risk of bias in nonrandomized studies should be used to rate the certainty of a body of evidence. J Clin Epidemiol, 2019; 111:105–114; doi: 10.1016/j.jclinepi.2018.01.012

24.

Viechtbauer

. Bias and efficiency of meta-analytic variance estimators in the random-effects model. Journal of Educational and Behavioral Statistics, 2005; 30(3):261–293; doi: 10.3102/10769986030003261

25.

Cochran

. The combination of estimates from different experiments. Biometrics, 1954; 10(1):101–129; doi: 10.2307/3001666

26.

Piccini

, Felicioni

, Pessina

, et al. Glycemic control, basal/bolus distribution, bmi and meal management in children and adolescents with type 1 diabetes and advanced hybrid closed loop. Nutrients, 2023; 15(23):4875; doi: 10.3390/nu15234875

27.

Criego

, Carlson

, Brown

, et al. Two years with a tubeless automated insulin delivery system: A single-arm multicenter trial in children, adolescents, and adults with type 1 diabetes. Diabetes Technol Ther, 2024; 26(1):11–23; doi: 10.1089/dia.2023.0364

28.

Rachmiel

, Lebenthal

, Mazor-Aronovitch

, et al. MiniMed 780G advanced hybrid closed-loop system outcomes according to pubertal status: Awesome Study Group Real-Life Experience. Diabetes Technol Ther, 2023; 25(9):643–651; doi: 10.1089/dia.2023.0081

29.

DeSalvo

, Bode

, Forlenza

, et al. Glycemic outcomes persist for up to 2 years in very young children with the Omnipod® 5 automated insulin delivery system. Diabetes Technol Ther, 2024; 26(6):383–393; doi: 10.1089/dia.2023.0506

30.

Lombardo

, Passanisi

, Alibrandi

, et al. MiniMed 780G Six-month use in children and adolescents with type 1 diabetes: Clinical targets and predictors of optimal glucose control. Diabetes Technol Ther, 2023; 25(6):404–413; doi: 10.1089/dia.2022.0491

31.

Henry

, Villar Fimbel

, Bendelac

, et al. Beneficial effects of automated insulin delivery over one-year follow-up in real life for youths and adults with type 1 diabetes irrespective of patient characteristics. Diabetes Obes Metab, 2024; 26(2):557–566; doi: 10.1111/dom.15344

32.

Atik-Altinok

, Mansuroglu

, Demir

, et al. Does minimed 780GTM insulin pump system affect energy and nutrient intake?: Long-term follow-up study. Eur J Clin Nutr, 2024; 78(7):615–621; doi: 10.1038/s41430-024-01422-y

33.

Bismuth

, Tubiana-Rufi

, Rynders

, et al. Sustained 3-year improvement of glucose control with hybrid closed loop in children with type 1 diabetes while going through puberty. Diabetes Care, 2024; 47(9):1696–1703; doi: 10.2337/dc24-0916

34.

Rapini

, Martino

, Arnaldi

, et al. Efficacy and safety of advanced hybrid closed loop systems in children with type 1 diabetes younger than 6 years. Front Endocrinol (Lausanne), 2024; 15:1382920; doi: 10.3389/fendo.2024.1382920

35.

Passanisi

, Salzano

, Bombaci

, et al.; ISPED Diabetes Study Group Collaborators. Sustained effectiveness of an advanced hybrid closed-loop system in a cohort of children and adolescents with type 1 diabetes: A 1-year real-world study. Diabetes Care, 2024; 47(6):1084–1091; doi: 10.2337/dc23-2311

36.

Monasor-Ortolá

, Quesada-Rico

, Nso-Roca

, et al. Degree of accuracy of the BMI Z-score to determine excess fat mass using DXA in children and adolescents. Int J Environ Res Public Health, 2021; 18(22):12114; doi: 10.3390/ijerph182212114

37.

Zucchini

, Tumini

, Scaramuzza

, et al. Recommendations for recognizing, risk stratifying, treating, and managing children and adolescents with hypoglycemia. Front Endocrinol (Lausanne), 2024; 15:1387537; doi: 10.1111/j.1463-1326.2006.00686.x

38.

Abraham

, Karges

, Dovc

, et al. ISPAD Clinical Practice Consensus Guidelines 2022: Assessment and management of hypoglycemia in children and adolescents with diabetes. Pediatr Diabetes, 2022; 23(8):1322–1340; doi: 10.1111/pedi.13443

39.

Russell-Jones

, Khan

. Insulin-associated weight gain in diabetes–causes, effects and coping strategies. Diabetes Obes Metab, 2007; 9(6):799–812; doi: 10.1111/j.1463-1326.2006.00686.x

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.09 MB

A Systematic Review and Metanalysis on the Impact of Automated Insulin Delivery Systems on Body Mass Index in Youths with Type 1 Diabetes and Potential Predictors

Abstract

Aims:

Materials and Methods:

Results:

Conclusions:

Introduction

Materials and Methods

Search strategy

Criteria for study selection

Data extraction and management

Assessment of the certainty of the evidence

Meta-analysis

Results

BMI, weight, and CHO intake

TDI, user-initiated, and autocorrection bolus

Discussion

Conclusions

Footnotes

Authors’ Contributions

Data Availability Statement

Author Disclosure Statement

Funding Information

Supplementary Material

Abbreviations Used

References

Supplementary Material