Hybrid Closed Loop in Adolescents and Young Adults Living with Type 1 Diabetes: A Real-World Study

Abstract

To evaluate the long-term metabolic outcomes and safety of hybrid closed loop (HCL) in youth living with type 1 diabetes, including people with very poor metabolic control, this observational, retrospective, monocentric study included 101 patients aged 15–25 years. After a mean (±standard deviation; range) follow-up of 16 (±6; 3–31) months, HbA1c decreased from 9.6% ± 2.3% (82 ± 25 mmol/mol) at baseline to 8.0% ± 1.4% (64 ± 15 mmol/mol) (P < 0.001). Continuous glucose monitoring parameters showed the same trend (P < 0.001): the glucose management indicator decreased from 8.8% to 7.9%, time in range increased from 39% to 57%, and time below range decreased from 3.0% to 1.6%. Ten patients discontinued HCL. There was no more severe hypoglycemia or diabetes ketoacidosis under HCL compared with the year before HCL initiation. Five cases of new onset and four cases of worsening retinopathy were reported. Youth with type 1 diabetes can improve metabolic control thanks to HCL without increasing complication risk. Vigilance is required for preexisting retinopathy monitoring.

Introduction

Type 1 diabetes (T1D) and difficulties in managing treatment can be responsible for acute or chronic complications, requiring repeated hospitalizations, leading to damaged quality of life, and reducing life expectancy. This can be prevented by maintaining optimal metabolic control with glycemic targets established by international guidelines,^1,2 but adolescence and early adulthood are challenging times with HbA1c reaching the highest level between ages 15 and 18.³ The last international consensus⁴ confirms hybrid closed-loop (HCL) superiority in glycemic improvement, but studies involving young people with very poor metabolic control are lacking, which limits practitioners to offer them these systems. To provide results concerning this particular population regarding metabolic control and safety, we conducted an observational real-world study in patients aged 15–25⁵ living with T1D equipped with an HCL, regardless of their previous HbA1c and treatment.

Materials and Methods

Study design

This retrospective study included HCL users treated in the diabetology department of a tertiary care center (Sud-Francilien hospital, Corbeil-Essonnes, France) between October 1, 2021, and September 1, 2023.

Data were collected between September 1, 2023, and February 29, 2024.

Patients could choose between three HCL systems: MiniMed™ 780G SmartGuard system (Medtronic), Tandem t:slimX2™ (Tandem Diabetes Care) with Control-IQ technology (CIQ), and DBLG1 (Diabeloop), regardless of the previous treatment.

All patients received initial training on HCL devices (continuous glucose monitoring [CGM], catheter, insulin pump, algorithm functioning, etc.) during 1–4 days at a hospital center with an appropriate educational program. Follow-up consisted of a phone call a week after HCL initiation, and then, medical appointments at months 1, 3, 6, and 12. Most of them had retinopathy assessment at HCL initiation and at month 3 if retinopathy was found. In that case, the system was initially set to a highest glycemic target, through specific HCL modes, so that metabolic control could be improved gradually.

Participants

We included patients living with type 1 diabetes for at least a year, aged 15–25, having used HCL for at least 3 months. All patients treated at our center with eligible criteria were included after they had given their consent to participate.

Data collection

Patient’s data were collected from their medical record on the intrahospital software. Glycemic outcomes were collected on the platform relative to each HCL system.

Data were deidentified before the study. The study received approval from the hospital’s research ethics committee and was registered under NCT06253351 (https://www.clinicaltrials.gov/).

Outcomes

The primary outcome was the comparison of HbA1c levels before starting HCL and at the time of data collection. Secondary outcomes included CGM parameters of the last 3 months for each participant (CGM use, time in range [TIR], time below range [TBR], glucose coefficient of variation [CV] and management indicator [GMI]) before and after HCL initiation, as well as time spent in closed-loop mode, number of HCL discontinuations, ketoacidosis (DKA) and severe hypoglycemia (SH) events, and new onset or worsening of retinopathy cases. We collected the number of acute metabolic events the year before baseline compared with the year before data collection.

Statistical methods

We described the distribution of continuous variables using the mean, median, standard deviation, and range, and described categorical variables with sample sizes and percentages. To evaluate the evolution of HbA1c and CGM parameters, we used linear mixed-effect models, considering participants as a random effect. We also tested interactions between time and baseline HbA1c, sex, and age.

Finally, we performed a sensitivity analysis of CGM-related outcomes, including only observations when sensor usage was ≥70%.

All analyses were performed with R version 4.2.1.

Results

One hundred and one patients living with type 1 diabetes (60% women) were included with a mean (±SD; range) age of 21 years (±3; 15–25) and a mean diabetes duration of 12 (±5; 1–21) years. Fourteen patients were previously on multiple daily injections. Ninety-two patients initiated CIQ, seven initiated Medtronic, and two initiated DBLG1 system.

At the time of data collection, the mean duration since HCL initiation was 16 months (±6; 3–31) and nearly two-thirds of the cohort had been using their HCL device for more than a year. Only two patients were lost to follow-up without information on metabolic outcomes (HbA1c and CGM) on HCL therapy, but were not excluded from the study as their baseline metabolic characteristics were still informative for the analyses.

We observed a large increase in CGM use from 57% to 83% (P < 0.001), aligning with the mean time spent in closed-loop mode (80% [±21; 16–99]).

Mean baseline HbA1c was 9.6% (±2.3; 6.4–18.7) (82 ± 25; 46–181 mmol/mol) and significantly decreased to 8.0% (±1.4; 5.5–13.8) (64 ± 15; 37–127 mmol/mol), resulting in a significant (P < 0.001) decrease of −1.6% (−20 mmol/mol) (Table 1).

Table 1.

Results of Primary (HbA1c) and Secondary (CGM-Related) Metabolic Outcomes

	Before hybrid closed-loop initiation	Under hybrid closed-loop therapy	P value^a
Main analysis (n = 101)
Primary outcome
HbA1c, mean (SD)	9.6 (2.3)	8.0 (1.4)	<0.001
Secondary outcomes
Time in range, mean (SD)	39 (16)	57 (12)	<0.001
Time below range, mean (SD)	3.0 (4.3)	1.6 (1.4)	<0.01
CV, mean (SD)	39 (7)	38 (4)	0.14
GMI, mean (SD)	8.8 (1.4)	7.9 (0.8)	<0.001
Sensitivity analysis ^b	n = 40	n = 72
Time in range, mean (SD)	44 (14)	58 (11)	<0.001
Time below range, mean (SD)	3.6 (4.2)	1.6 (1.4)	<0.001
CV, mean (SD)	41 (6)	37 (4)	<0.001
GMI, mean (SD)	8.4 (1.2)	7.8 (0.8)	<0.001

Time below range accounts for low (54–70 mg/dL) and very low (<54 mg/dL).

Time effect is evaluated using linear mixed-effect models, considering participants as the random effect.

Accounting for CGM parameters while sensor usage ≥70%.

CGM, continuous glucose monitoring; CV, coefficient of variation; GMI, glucose management indicator; SD, standard deviation.

CGM parameters also showed significant improvement: GMI decreased by 0.9%, TIR increased from 39% to 57%, and TBR almost halved (from 3.0% to 1.6%) (Table 1). CGM parameters are detailed in Supplementary Figure S1 for all patients with available data (regardless of the level of CGM use).

These findings were consistent with the results of the sensitivity analysis (considering only CGM parameters while sensor usage ≥70%) that additionally showed significant improvement of CV decreasing from 41% to 37% (Table 1).

Results of subgroup analysis were homogenous regarding age groups (Fig. 1). Regarding sex, the difference was close to significance (P = 0.06) with higher HbA1c improvement in men. Only baseline HbA1c showed significant interaction with metabolic improvement. The higher the baseline HbA1c, the greater the reduction under HCL therapy (reduction by 0.41%, 1.02%, and 3.1% for the subgroups ≤8%, 8%–10%, and ≥10%, respectively).

FIG. 1.

Association between HbA1c reduction and baseline characteristics (sex, age, HbA1c).

Ten patients discontinued the HCL system after a mean of 9 (range: 3–17) months. The main reasons were skin reactions, pump alarms, and technical issues.

Concerning safety, over 12 months prior data collection at baseline and at the end, there were fewer acute metabolic events for at least the 16 months of HCL treatment up to a year before data collection. Before HCL, incidence of DKA was about 0.66 for 100 person-month (p-m.) and 0.58 SH for 100 p-m. Under HCL, the incidence is 0.45 DKA for 100 p-m. and 0.36 SH for 100 p-m.

Despite using a preventive wider glycemic target through temporary goal (Medtronic) or exercise mode (Control IQ, Diabeloop), five cases of new-onset mild retinopathy were reported, and four patients worsened their retinopathy, mostly by one stage.

Discussion

This real-world study is the first to involve a substantial sample of adolescents and young adults, including participants with very poor metabolic control, with a follow-up of more than 12 months. Our study showed a significant improvement of all metabolic outcomes, without major adverse events. These results support the benefit of HCL, including this challenging population that often struggles with diabetes management, without reversal of the risk–benefit balance.

Previous real-world studies involving young populations typically reported baseline HbA1c ranging from 6.8% (51 mmol/mol) to 8.6% (70 mmol/mol), with a follow-up generally limited to 12 months.^6

–11 Only one randomized trial and four real-world studies included young patients with poor metabolic control, all with limited sample size. First, in the only randomized trial including patients with high HbA1c at baseline, Boucsein and al.¹² found an HbA1c improvement of 2.6% after 13 weeks for 80 patients using MiniMed 780G whereas stability in the control group. Concerning real-world studies, Michaels et al.¹³ showed more than 20 patients, with a mean age of 18.8 years and an HbA1c baseline of 10.5%, a decrease of HbA1c by 2.5% (4 mmol/mol), and an increase of TIR by 34.9% after 12 months of HCL use. While Castorani et al. observed a decrease from 10% (86 mmol/mol) to 7% (53 mmol/mol) leading to a TIR improvement of 41.5% after 6 months in 20 nonadherent patients.¹⁴ Moreover, in 13 young American historically minoritized (median 14.8 years, baseline HbA1c11.7%), HbA1c decreased by 2.7% (6 mmol/mol) and TIR increased by 23.7% after 6 months of HCL use.¹⁵ Finally, in Deepchand et al.’s¹⁶ real-world study, mean HbA1c decreased from 12.3% (111 mmol/mol) to 9.4% (79 mmol/mol) and TIR increased by 18.7% after 12 months in 32 young adults.

Compared with these studies, we observed a weaker HbA1c reduction, which might be explained by lower baseline HbA1c, longer follow-up, larger sample size, or specific population characteristics. Even if the mean HbA1c of our cohort did not reach the target recommended by international guidelines,^2,17 it is well demonstrated¹ that HbA1c reduction, at these high levels, can markedly reduce microangiopathy risk. Moreover, we observed a substantial improvement in CGM use by 26% during follow-up, as well as a high percentage of HCL use, even 12 months after its initiation, which are very positive findings for such a challenging population. This evolution was made possible by optimal baseline training and an appropriate and tight follow-up by specialized health care professionals.

Regarding the risk of acute complications with HCL in patients with poor metabolic control, it is important to note that all ketoacidosis events occurred when patients were not using HCL mode. Moreover, all these patients had a history of ketoacidosis and only one was newly equipped with an insulin pump. Severe hypoglycemia happened mostly in one patient who misused the system, declaring “false” meals to make glycemia decrease faster. Regarding retinal risk, studies reporting retinopathy evolution after HCL initiation are scarce. In our study, seven of the nine patients with new-onset or worsening retinopathy crossed only one retinopathy stage, mostly to mild retinopathy (Supplementary Fig. S2). This effect on retinopathy may reflect both too rapid glycemia improvement because of HCL and natural progression of diabetes’ complications. We observed one case of severe worsening retinopathy that required laser treatment in a patient whose GMI decreased from 11.6% (103 mmol/mol) to 8.5% (69 mmol/mol) within 2 months after HCL initiation, despite using exercise mode. Thus, clinicians must pay attention to retinopathy status evolution after HCL initiation, especially for those with poor metabolic control and preexisting retinopathy. Being able to program higher glycemic targets would be of great interest for this high-risk population.

As expected, we observed the greatest HbA1c improvement for the highest baseline HbA1c levels, suggesting that young people who are the most in trouble with diabetes are those who can benefit most from HCL for metabolic control. This is true, despite their lack of rigor in managing technical issues and announcing carbohydrate intakes. Indeed, Petrovski and al.¹⁸ showed that young people can reach international targets of glycemic control after 12 months of HCL use with a preset of three personalized fixed carbohydrate amounts. However, to get these results, it involves the availability of a trained team for youth, needing a close and tailored follow-up, at least until the age of 25.

The main limitation of this study is its retrospective and monocentric design, without a control group, resulting in missing data (although limited to two patients for the primary outcome) and limited external validity. Moreover, this study lacks an evaluation of the evolution of patient’s quality of life, an important determinant of patients’ adherence to treatment.^11,13,15

Technological advances with HCL are in expansion, suggesting that patients who are still struggling with their diabetes will have more and more options to improve their metabolic control. New generation of HCL with dual-hormone pumps that deliver both insulin and glucagon is eagerly awaited.¹⁹ However, it seems challenging at this time to envision youth using such devices in their current form, given their size and the strict usage requirements they entail.

Conclusions

To conclude, this real-life study involved adolescents and young adults with a broad level of metabolic control at baseline, over an extended observation period, and offered valuable outcomes regarding metabolic control and safety with HCL treatment: HbA1c and CGM parameters showed significant and sustained improvement without an increased risk of acute complications. Thus, specialists should propose these systems to all patients living with T1D, regardless of their treatment or their metabolic control at baseline. However, vigilance is warranted to ensure patient safety for preexisting retinopathy monitoring in patients with poor metabolic control. It suggests that HCL access conditions should be enlarged to include patients who need help the most.

Footnotes

Acknowledgments

The sponsor was Centre Hospitalier Sud Francilien.

Authors’ Contributions

J.E. coordinated the study. M.U., A.P., C.V., S.F., and J.E. researched the data. C.A. contributed to data and statistical analyses. M.U., A.P., C.A., and J.E. wrote the first draft. All authors reviewed the article. M.U. and J.E. are the guarantors of this work and, as such, have full access to all data in the study and take responsibility for the integrity of the data and the accuracy of data analysis.

Author Disclosure Statement

J.E. and M.U. have no conflict of interest linked to this study. C.A. received speaker honoraria from Diabeloop (Inc) in 2020. S.F. has received congress invitations, speaker honoraria, and consultant fees from Novo Nordisk, Roche Diagnostics, LifeScan, Sanofi, Diabeloop, and Eli Lilly and is a shareholder of Diabeloop. A.P. has received congress invitations, speaker honoraria, and consultancy fees from Abbott, Diabeloop, Dexcom, Eli Lilly, Medtronic, Novo Nordisk, Sanofi, and has served on advisory board panels for Abbott, Insulet, Medtronic, Novo Nordisk, and Sanofi.

Funding Information

This study received no funding.

Supplementary Material

Supplementary Figure S1

Supplementary Figure S2

References

ElSayed

, Aleppo

, Aroda

, et al. 6. Glycemic targets: Standards of Care in Diabetes—2023. Diabetes Care, 2023; 46(Suppl 1):S97–S110.

Gregory

, Cameron

, Joshi

, et al. ISPAD clinical practice consensus guidelines 2022: Diabetes in adolescence. Pediatr Diabetes, 2022; 23(7):857–871.

Foster

, Beck

, Miller

, et al. State of Type 1 diabetes management and outcomes from the T1D exchange in 2016-2018. Diabetes Technol Ther, 2019; 21(2):66–72.

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.

Sawyer

, Azzopardi

, Wickremarathne

, Patton

. The age of adolescence. Lancet Child Adolesc Health, 2018; 2(3):223–228.

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.

Messer

, Berget

, Pyle

, et al. Real-world use of a new hybrid closed loop improves glycemic control in youth with type 1 diabetes. Diabetes Technol Ther, 2021; 23(12):837–843.

Forlenza

, Carlson

, Galindo

, et al. Real-World evidence supporting tandem Control-IQ hybrid closed-loop success in the Medicare and Medicaid type 1 and type 2 diabetes populations. Diabetes Technol Ther, 2022; 24(11):814–823.

Petrovski

, Al Khalaf

, Campbell

, et al. Glycemic outcomes of advanced hybrid closed loop system in children and adolescents with type 1 diabetes, previously treated with Multiple Daily Injections (MiniMed 780G system in T1D individuals, previously treated with MDI). BMC Endocr Disord, 2022; 22(1):80.

10.

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.

11.

Reznik

, Bonnemaison

, Fagherazzi

, et al. The use of an automated insulin delivery system is associated with a reduction in diabetes distress and improvement in quality of life in people with type 1 diabetes. Diabetes Obes Metab, 2024; 26(5):1962–1966.

12.

Boucsein

, Zhou

, Michaels

, et al. Automated insulin delivery for young people with type 1 diabetes and elevated A1c. NEJM Evid, 2024; 3(10):EVIDoa2400185.

13.

Michaels

, Boucsein

, Watson

, et al. Glucose and psychosocial outcomes 12 months following transition from multiple daily injections to advanced hybrid closed loop in youth with type 1 diabetes and suboptimal glycemia. Diabetes Technol Ther, 2024; 26(1):40–48.

14.

Castorani

, Rigamonti

, Frontino

, et al. Turning the tides: Achieving rapid and safe glucose control in adolescents with suboptimally controlled type 1 diabetes using advanced hybrid closed loop systems. Front Endocrinol (Lausanne), 2024; 15:1243565.

15.

Marks

, Grundman

, Meighan

, et al. Hybrid closed loop systems improve glycemic control and quality of life in historically minoritized youth with diabetes. Diabetes Technol Ther, 2024; 26(3):167–175.

16.

Deepchand

, Balla

, Eroukhmanoff

, et al. Closed-loop insulin delivery systems for people with type 1 diabetes and chronic very poor metabolic control: It works and is safe!. Diabetes Technol Ther, 2024; 26(2):125–129.

17.

Holt

RIG

, DeVries

, Hess-Fischl

, et al. The management of type 1 diabetes in adults. A consensus report by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetologia, 2021; 64(12):2609–2652.

18.

Petrovski

, Campbell

, Pasha

, et al. Twelve-month follow-up from a randomized controlled trial of simplified meal announcement versus precise carbohydrate counting in adolescents with type 1 diabetes using the MiniMed^TM 780G advanced hybrid closed-loop system. Diabetes Technol Ther, 2024; 26(S3):76–83.

19.

van Bon

, Blauw

, Jansen

TJP

, et al. Bihormonal fully closed-loop system for the treatment of type 1 diabetes: A real-world multicentre, prospective, single-arm trial in the Netherlands. The Lancet Digital Health, 2024; 6(4):e272–e280.