MiniMed 780G Advanced Hybrid Closed-Loop System Outcomes According to Pubertal Status: Awesome Study Group Real-Life Experience

Abstract

Background and Aims:

Achieving good glycemic control is a major challenge for adolescents with type 1 diabetes (TID). The introduction of the MiniMed 780G system, an advanced hybrid closed-loop (AHCL) that enables an automatic correction of insulin, gave hope for improved glycemic outcomes in adolescents. We assessed specific characteristics associated with glycemic measures in youth with T1D switching to Minimed 780G.

Methods:

This retrospective observational real-life multicenter study from the AWeSoMe Group assessed continuous glucose monitoring (CGM) metrics of 22 patients (59% females, median age 13.9 interquartile range [IQR 11,18] years), from a high socioeconomic background. CGM metrics were recorded for 2-week periods before AHCL, after 1, 3, 6 months, and at the end of follow-up (median 10.9 [IQR 5.4, 17.4] months). Delta-variables (Δ) were calculated as the difference between the end of follow-up and baseline.

Results:

Time in range (TIR)_{70–180mg/dL} increased from 65% [52, 72] to 75% [63, 80], P = 0.008, from baseline to end of follow-up. Time above range_>180mg/dL decreased from 28% [20, 46] to 22% [14, 35], P = 0.047. Advanced pubertal stage was correlated with less improvement in ΔTAR_>180mg/dL, r = 0.47, P = 0.05, and less CGM usage r = −0.57, P = 0.05. A longer disease duration was associated with less improvement in ΔTAR_{180–250mg/dL}, r = 0.48, P = 0.05. Lower pump site change frequency was associated with higher glucose management indicator, r = 0.5, P = 0.03, and lower TIR_{70–180mg/dL} r = −0.52, P = 0.08.

Conclusion:

The use of AHCL enabled improvements in TIR_{70–180mg/dL} in youth with T1D. More advanced pubertal stages, longer disease duration, and less compliance were associated with less improvement, stressing the need for continuous support, and re-education in this age group.

Introduction

Type 1 diabetes (T1D) is a chronic condition requiring continuous intensive care for glucose management. Achieving good glycemic outcomes is a major challenge at any age, but specifically during adolescence.¹ Incorrect insulin dosing, forgetfulness, disruptions to the daily routine, issues with peers, fear of hypoglycemia, and insulin omission for weight management are only some of the issues adolescents face, all of which have been demonstrated to worsen glycemic control.¹

The introduction of the advanced hybrid closed-loop (AHCL) system gave hope for improved glycemic outcomes in adolescents. The MiniMed 780G system contains an AHCL algorithm that can automatically adjust the user's basal insulin delivery based on their continuous glucose monitoring (CGM) data, adjusting blood sugar levels to be stable according to a chosen target glucose level, throughout the day and night.^2,3 Users customize their settings based on their individual needs, and set their target glucose range to adjustable levels of 100 (5.5), 110 (6.1), and 120 (6.7) mg/dL (mmol/L).

Importantly, users of the 780G pump still require to enter their carbohydrate intake for meals, and may need to manually adjust their insulin delivery settings during exercise. However, missed or inaccurate boluses are mitigated by the 780G system due to its features by giving automated correction boluses. Although timely pump and reservoir changes and additional proper handling in accordance with system instructions are required, the system decreases disease burden in a seamless way.

There is mounting evidence on the impact of the MiniMed 780G system on improving glycemic outcomes in adults with T1D,^3
–5 as well as on its safety.² Studies on the impact of the 780G on glycemic outcomes in children and adolescents focused on sleep quality,⁶ and the transition from multiple daily injections to the MiniMed 780G system.⁶

Randomized controlled studies demonstrated significant improvements in mean time in range (TIR) and HbA1c.⁷ However, real-life data on the effect of the 780G on glycemic parameters among youth are still scarce and not consistent, some demonstrate significant improvement⁸ while others do not show any significant change.⁹

We found it intriguing to understand the cause for that inconsistency of data in the pediatric population. Thus, we aimed to assess the relative impact of the 780G system on glycemic outcomes in children and adolescents in a real-world setting, among a specific, highly capable, and motivated population of those who could afford purchasing the system out of pocket. We sought to identify characteristics that constitute a barrier to maximal glycemic outcomes.

Research Design and Methods

Research design

This retrospective, real-life, multicenter, observational study is based on data retrieved from medical files, CGM systems, and continuous subcutaneous insulin infusion sets (pumps), acquired from Dexcom Clarity, CareLink, and Tidepool softwares at baseline, and from the CareLink platform after 780G initiation. Data reported were retrieved from those systems for the 2 weeks period preceding the baseline, 1, 3, 6 months, and the time of the last follow-up visits. The study was performed in accordance with the guidelines of the 2013 Declaration of Helsinki on human experimentation.

Data confidentiality and anonymity were always maintained and the individuals are not identifiable either in the database or in this article. Due to the anonymous nature of Israeli patients' data coding in both Carelink and health files, and mandatory collection of the information included in the dataset, informed consent was waived. Personally identifiable information was pseudonymized. The study was approved by the Institutional Review Board of each of the participating medical centers.

Study population and setup

The study population comprised all individuals with T1D, aged <25 years, who purchased the 780G pump out of pocket (disposables were supplied by health insurance companies) and who were treated at the AWeSoMe Study Group, that is, pediatric diabetes multidisciplinary, university-affiliated medical centers: Edith Wolfson Medical Center, Dana-Dwek Children's Hospital, Shamir (Assaf Harofeh) Medical Center, and Edmond and Lily Safra Children's Hospital. Individuals with T1D who received other glucose-mediating medications were excluded. In all centers, the educational process for transitioning to the AHCL system was performed in an outpatient setting by the multidisciplinary team and a technical expert from Medtronic. The manual mode “run-in” period was between 7 and 14 days.

Data collection

The information retrieved from medical files included demographic characteristics (age, sex), Israeli socioeconomic position (SEP) cluster (range 1–10) and index (range −2.797 to 2.590) by home address, based on the Israel Central Bureau of Statistics' Socio-Economic Level of the Population 2015,¹⁰ clinical data of other illnesses, diabetes duration, and HbA1c, and anthropometric data (weight, height, body mass index [BMI]), (BMI-standard deviation score (SDS) and height SDS were calculated based on CDC 2000 growth charts).¹¹ Tanner stage of puberty¹² was defined according to breast development in females and testicular volume in males measured using Prader beads by board-certificated experts in pediatric endocrinology. Tanner stages were analyzed as an ordinary variable, although they are presented in the graph as three categories: Tanner I and II, Tanner III and IV, and Tanner V.

The information retrieved from the 2-week ambulatory glucose profile reports included percent time-spent in various glycemic ranges: TIR (70–180 mg/dL; 3.9–10 mmol/L), time below range (TBR), including the hypoglycemic range (<70 mg/dL; <3.8 mmol/L), mild hypoglycemic range (TBR 54–70 mg/dL; 3–3.8 mmol/L), and severe hypoglycemic range (TBR <54 mg/dL; <3 mmol/L); time above range (TAR), including the hyperglycemic range (>180 mg/dL; >10 mmol/L), moderate hyperglycemic range (TAR 180–250 mg/dL; 10–13.3 mmol/L), and severe hyperglycemia (TAR >250 mg/dL; >13.3 mmol/L); and coefficient of variation (CV), mean glucose levels (mg/dL), and glucose management indicator (GMI).¹³ In addition, we assessed an optional additional parameter of the change in TIR after the initiation of 780G per individual.

As the clinical relevance of the effect of a change in TIR is dependent on the initial TIR, we used a parameter defined as the improvement in TIR relative to the maximum potential improvement (rTIR). The rTIR was calculated by the ratio between the difference between TIR at the end of follow-up compared to TIR at baseline (TIR after-TIR before), and the potential improvement for that patient (100-TIR before).¹⁴ This parameter may distinguish those who had the best chance of improving (lower initial TIR) from those who were already in the target TIR and had little room for improvement. A cutoff of 10% in the rTIR parameter was considered clinically significant.

The reported amount of carbohydrates, the mean total daily dose (TDD) of insulin, the mean percent of baseline basal or long-acting insulin, the pump settings, and the handling parameters of the 780G System, including percent time CGM active, frequency of escaping to manual mode, frequency of pump and reservoir changes, pump disabling, and CGM expiration events, were recorded.

Outcome measures

The primary outcome measures were (1) the changes (Δ) in TIR_{70–180mg/dL}, TBR_<70mg/dL, and TAR_>180mg/dL at the end of study compared to baseline. (2) Baseline characteristics associated with the Δ in TIR_{70–180mg/dL}, TBR_<70mg/dL, and TAR_>180mg/dL, including pubertal stage, age at baseline, and diabetes duration (as continuous variables). The secondary outcome measures were the associations between glycemic outcomes, pump settings, and pump handling parameters.

Statistical analysis

Categorical variables were described using frequency and percentage. Continuous variables were expressed as the median and interquartile range (IQR). The Wilcoxon test was used to compare continuous variables between each 2-week period to baseline. Spearman's correlation coefficient was used to study associations between continuous variables not consistent. r > 0.36 was considered a moderate correlation, while r > 0.67 was considered a strong correlation.¹⁵ All statistical tests were two-sided. Statistical significance was defined at P < 0.05. Bonferroni correction was used as control for multiple comparisons. To note, due to the small study population, multivariable model could not be applied. The statistical analyses were performed by SPSS software (IBM SPSS Statistics version 28, IBM Corp., Armonk, NY, 2021).

Results

The study cohort comprised 23 individuals who purchased the MiniMed 780G system out of pocket; 1 was excluded from the analysis as metformin treatment was initiated during the study period. The final study population included 22 participants (59% females), with a median age of 13.9 [IQR 11.0, 18.0] years (range 7.7–24.0 years), a median age at T1D diagnosis of 8.3 [IQR 5.8, 11.3] years, and a median diabetes duration of 5.5 [1.7, 10.1] years. Five (21.7%) were prepubertal, 4 were in Tanner stages II–III, and 13 were in Tanner stages IV–V. Ninety-one percent belonged to high SEP; median SEP cluster 8 [IQR 7, 9.25], and median SEP index 1.4 [IQR 0.7, 2.4].

At baseline, 19 (86.4%) were on pump therapy, and 21 (95.5%) were using a CGM, full baseline percent of time spent in ranges was available for 18 participants (82%). The median baseline HbA1c level was 7.4% [IQR 7.0, 8.3]. CGM parameters demonstrated a GMI of 7.0% [IQR 6.4, 8.1]. The median follow-up duration of the study was 10.9 months [IQR 5.4, 17.4].

Changes (Δ) over time in glycemic measures and anthropometric parameters

Glycemic parameters over time are presented in Figure 1, and the comparisons between baseline and the end of follow-up are presented in Table 1. The percent TIR_{70–180mg/dL} increased significantly within the first month (from 65% [IQR 52, 72] to 72% [IQR 65, 78], P < 0.001), and this significant improvement persisted until the end of follow-up to 75% [IQR 63, 80], P = 0.008). The median rTIR was 26.6% (11.4, 40.9). The percent time spent in the hyperglycemic range of >180 mg/dL decreased significantly over time, from 28% (20, 46) at baseline to 22% (17, 31), (P = 0.047) at 3 months, and was stable until the end of the follow-up at 22% (14, 35), (P = 0.047).

FIG. 1.

The change in the percent of time spent at each glycemic range over time, according to pubertal category at baseline. The horizontal axis expresses the time from baseline, at which this measurement was assessed. The vertical line represents the percent of time spent at each glycemic range. The bars represent the median and 25th and 75th percentiles. The X represents the average. The whiskers represent the minimum and maximum. Each parameter at each time point of data collection was compared to baseline. Statistically significant differences compared to baseline are presented by:, *P = 0.02–0.05, ^^p = 0.008, ^# P = 0.003–0.005, **P = 0.001.

Table 1.

Anthropometric, Clinical, Diabetes-Related, and Pump Characteristics of Study Population (n = 22), at Baseline and at End of Study, After a Median of 10.9 Months (5.4, 17.4)

Characteristic		Baseline	At end	P
Glycemic parameters	% TIR (70–180 mg/dL)	65 (52, 72)	75 (63, 80)	0.008
	% TAR (>180 mg/dL)	28 (20, 46)	22 (14, 35)	0.047
	% Time in 180–250 mg/dL	21 (15, 31)	18 (12, 25)	0.11
	% Time >250 mg/dL	5.0 (2.0, 12.7)	5.0 (2.0, 9.5)	0.06
	% TBR (<70 mg/dL)	4.0 (0.9, 9.0)	2.5 (0.3, 5.8)	0.10
	% Time <54 mg/dL	0.03 (0, 1)	0 (0, 1)	0.18
	% Time in 54–70 mg/dL	3.0 (1.0, 7.0)	2.5 (0.3, 4.0)	0.08
	Coefficient of variation, %	32 (29, 40)	36 (35, 38)	0.48
	Mean glucose, mg/dL	154 (130, 178)	145 (133, 166)	0.21
	Glucose SD mg/dL	59 (43, 66)	52 (44, 61)	0.53
	Laboratory HbA1c, %	7.4 (7.0, 8.3)	7.3 (6.8, 8)	0.23
	Laboratory HbA1c, mmol/mol	57.4 (53, 67.2)	56.3 (50.8, 63.9)	0.23
	GMI, %	7 (6.4, 7.8)	6.9 (6.4, 7.2)	0.34
	GMI, mmol/mol	53 (46.4, 61.7)	51.9 (46.4, 55.2)	0.34
Anthropometric data	Weight SDS	0.74 (−0.35, 1.12)	0.71 (−0.1, 0.23)	0.30
	Height SDS	0.66 (0.2, 1.2)	0.44 (−0.44, 1.09)	0.09
	BMI SDS	0.53 (−0.37, 0.82)	0.53 (−0.24, 1.02)	0.33
Insulin, pump, and CGMs characteristics	Insulin TDD	36 (22, 52)	49 (42, 57)	0.002
	% Basal insulin per day	44 (37, 51)	40 (35, 43)	0.028
	% Bolus insulin per day	56 (49, 63)	60 (58, 65)	0.11
	% Autocorrection per day	NA	24 (18, 33)	NA
	Mean daily carbohydrate (g)	140.5 (94, 183)	149 (113, 192)	0.31
	No. days between pump change	3.3 (3, 4.2)	3.7 (2.95, 4.5)	0.53
	% Time CGM Active	65.5 (53, 94)	84.5 (80, 90)	0.023
Physician dependent factors	Active insulin duration (h)	3 (2.5, 3)	2 (2, 2)	0.009
	I:C ratio	9.0 (7.3, 11)	7.7 (6.6, 9.7)	0.012
	Correction factor	45 (31, 75)	40 (34, 51)	0.043

Bold-statistically significant characteristics.

Data are presented as median and interquartile range, and percent for categorical data.

BMI, body mass index; CGM, continuous glucose monitoring; GMI, glucose management indicator; I:C, insulin carbohydrate ratio; SD, standard deviation; SDS, standard deviation score; SDSCGMSHbA1NA SDSDSTAR, time above range; TBR, time below range; TDD, total daily dose; TIR, time in range.

There were no statistically significant changes when TAR was divided into subranges of TAR_{180–250mg/dL} and TAR_>250mg/dL at the end of follow-up. As depicted in Figure 1, there was a trend of decrease in both TAR ranges, 180–250 and >250 mg/dL, but the difference reached a significance only at 3 months.

The percent time spent in the hypoglycemic range below 70 mg/dL decreased significantly within the first 3 months from 4.0% (0.85, 9) to 3.0% (2.0, 3.75), (P = 0.04), and to 2.5% (0.25, 4.0), P = 0.1, at the end of follow-up. TBR_54–70mg/dL decreased significantly from 3.5% (1, 7) at baseline to 2% (1.25, 3), (P = 0.043) after 3 months, and remained similar at the end of the follow-up.

The median glucose level, its standard deviation, CV, HbA1c, and GMI did not change significantly over time and at the end of the study (Table 1). The number of individuals with an HbA1c level lower than 7% remained unchanged, six individuals at baseline and at the end of the study. There was no significant unexpected change in anthropometric parameters, including BMI-SDS, which remained not significantly different (Table 1). However, as expected a higher delta BMI-SDS was correlated with higher pubertal state, r = 0.71, P = 0.014.

Changes over time in insulin pump delivery

The mean TDD of insulin gradually increased from 36 (22, 52) units per day to 39 (28, 64) units per day at the first month P = 0.005, and continued to increase significantly in comparison to baseline, with a median of 49 (42, 57) units, (P = 0.002) at the end of the follow-up. Basal insulin decreased significantly from 44% of TDD (37, 51) at baseline within a month to 39% (34, 44), P = 0.048, and remained stable, and significantly lower than baseline at the end of the study, 39% (35, 42) units, P = 0.028, respectively. There was no significant change in the percent of bolus insulin per day administered by the patient according to carbohydrates; however, a significant amount of insulin was delivered by pump as autocorrection boluses. At the end of the study, up to 24% of the delivered insulin as a bolus was delivered as autocorrection boluses.

Description of pump and CGM handling habits at the end of the follow-up

The median time between changing the pump and the reservoir was 3.7 (2.9, 4.5) and 3 days (2.7, 3.9), respectively. Manual mode events were recorded at a median frequency of 11.5 (3.5, 16.8) per 2 weeks, and 0–1 CGM expiration events. Calibration alerts were detected among six to eight participants throughout follow-up. There was a significant increase in active CGM use from baseline 65% (53, 94) to 85% (80, 90), at end of follow-up, P = 0.023.

Significant changes in pump settings, performed by the health professionals, were observed, including a stronger insulin carbohydrate ratio, a shorter active insulin duration time, and increased correction factors (although used in cases of manual mode only), as shown in Table 1.

Association between clinical characteristics and pump handling pitfalls and glycemic measures

Tanner stage of puberty was moderately and significantly correlated with the ΔTAR_>180mg/dL between end of study and baseline, r = 0.47, (P = 0.05), particularly the ΔTAR_{180–250mg/dL}, r = 0.52, (P = 0.03), and diabetes duration, r = 0.52, P = 0.01, as demonstrated in Table 2.

Table 2.

Correlations Between Clinical Characteristics, Pump Handling Habits, and Glycemic Control Outcome Parameters

Glycemic parameter	Tanner stage		Diabetes duration		Days between reservoir change		Days between pump changes
Glycemic parameter	r-Correlation coefficient	P	r-Correlation coefficient	P	r-Correlation coefficient	P	r-Correlation coefficient	P
Last visit TIR (70–180 mg/dL)	0.03	0.89	−0.19	0.39	−0.52	0.08	−0.39	0.07
ΔTIR (70–180 mg/dL)	−0.36	0.14	−0.15	0.56	NA	NA	NA	NA
Last visit TAR (>180 mg/dL)	0.03	0.9	0.18	0.08	0.44	0.15	0.39	0.07
ΔTAR (>180 mg/dL)	0.47	0.05	0.3	0.23	NA	NA	NA	NA
Last visit TAR (180–250 mg/dL)	0.018	0.94	0.16	0.47	0.45	0.15	0.35	0.11
ΔTAR (180–250 mg/dL)	0.52	0.03	0.48	0.05	NA	NA	NA	NA
Last visit TBR (<70 mg/dL)	0.04	0.86	0.08	0.72	−0.33	0.30	−0.13	0.57
ΔTBR (<70 mg/dL)	−0.27	0.39	−0.56	0.06	NA	NA	NA	NA
Last visit HbA1c	0.22	0.35	0.3	0.19	0.27	0.40	0.28	0.22
ΔHbA1c	−0.13	0.62	−0.19	0.44	NA	NA	NA	NA
Last visit GMI	0.15	0.53	0.19	0.44	0.43	0.17	0.50	0.03
ΔGMI	0.44	0.07	−0.02	0.96	0.44	0.13	0.53	0.115
rTIR	−0.51	0.03	−0.21	0.41	−0.28	0.38	−0.52	0.03
Last visit % autocorrection	0.10	0.66	0.16	0.22	0.44	0.15	0.36	0.09
No. days between pump changes	0.40	0.06	0.13	0.55	0.92	<0.0001	NA	NA
No. days between reservoir	0.49	0.1	0.24	0.45	NA	NA	0.92 ^*	<0.0001
Last visit % CGM usage	−0.57	0.05	−0.54	0.07	−0.25	0.43	−0.36	0.24

Bold-statistically significant characteristics.

Δ denotes the difference between the end of follow-up and baseline.

rTIR, the improvement in TIR relative to the maximum potential improvement; TIR, time in range.

The delta for glucose measures according to tanner stage groups is depicted in Figure 2. Tanner stage group I + II and III+IV were significantly different from zero for delta GMI, TIR_{70–180mg/dL}, rTIR, TAR_>180mg/dL, and TAR_{180–250mg/dL} meaning a response was observed, while the delta for stage 5 was not significantly different from zero meaning no improvement in response to AHCL.

FIG. 2.

The association between the delta in various glycemic measures and pubertal stages. (a) Depicts delta GMI for stages I + II, III+IV, and V. The response for stage V, is not substantially different from zero, indicating no improvement in response to AHCL whereas delta in stages I + II and III+IV is below zero indicating lower GMI at the end of follow-up. (b) ΔTIR_{70–180mg/dL} is above zero in stages I + II and III+IV, indicating more time spent in TIR_{70–180mg/dL}. (c) Lower pubertal stages were associated with a higher percent of improvement in rTIR. (d): ΔTAR_>180mg/dL is below zero indicating in stages I + II and III+IV indicating less time in glucose levels above 180 mg/dL at the end of the study. (e) ΔTAR_{>180–250mg/dL} is below zero in stages I + II and III+IV, indicating less time in glucose levels above at the end of the study. (f) Glucose levels above 250 mg/dL remained elevated in all Tanner stage subgroups at the end of the study. The bars represent the median and 25th and 75th percentiles. The whiskers represent the minimum and maximum. AHCL, advanced hybrid closed loop; GMI, glucose management indicator; TAR, time above range; TIR, time in range; rTIR, relative improvement in TIR.

Tanner stage of puberty was also moderately correlated with the number of days between pump site changes, r = 0.4, (P = 0.06), and with the increase in GMI r = 0.44, (P = 0.07). Pubertal stage was negatively correlated with the percent of CGM time, r = −0.57, (P = 0.05). No correlations were found between glycemic parameters and age at baseline. After adjustment for multiple comparisons, of the correlations of pubertal stage only diabetes duration, ΔTAR_{180–250mg/dL} and rTIR remained significantly correlated with pubertal stage (r = 0.52, P = 0.047, r = 0.53, P = 0.05, and r = −0.51, P = 0.047).

Longer disease duration was positively correlated with ΔTAR_{180–250mg/dL} r = 0.048. (P = 0.05), and negatively correlated with the TBR_<70mg/dL, and the percent CGM time. Longer time between pump site changes was positively correlated with GMI r = 0.5, P = 0.03.

Parameters associated with the rTIR

The rTIR and TIR are highly comparable, r = 0.93, P = <0.001, as expected. The rTIR in our cohort seems clinically significant at 26.6% (11.4, 40.9). Significant rTIR improvement was associated with GMI at end of the study and delta GMI r = −0.52, P = 0.03, and r = −0.56, P = 0.01, respectively. rTIR improvement was also associated with worse initial glycemic parameters TAR, r = 0.702, P = 0.001, including higher TAR_{180–250mg/dL} r = 0.642, P = 0.004, TAR_>250mg/dL r = 0.501, P = 0.034. rTIR improvement was negatively associated with TBR r = −0.567, P = 0.014, as well as pubertal stage r = −0.508, P = 0.031, age r = −0.474, P = 0.047 and days between pump site change r = −0.542, P = 0.045.

After adjustment for multiple comparisons, only baseline TAR_>180mg/dL and TAR_{180–250mg/dL} remained significantly associated with rTIR (P = 0.033 and P = 0.009, respectively).

Discussion

Our unique cohort of children and adolescents with T1D, from high socioeconomic backgrounds, and good glycemic control, who purchased the 780G pump out of pocket, allowed us to pinpoint the still existing challenges for better glycemic outcomes. These challenges are intensified during puberty, with a longer diabetes duration, and related to human behavior.

The use of the 780G system in a group of children and youth with good glycemic control as expressed by median HbA1c level of 7.4% enabled a significant improvement in percent TIR from 66% to 75% from the first month of use and up to nearly a year of follow-up. This improvement was a result of less time spent in the high glucose ranges of levels above 180 mg/dL. The improvement in TIR was in parallel with a 36% increase in total daily insulin. Of note, about a quarter of the bolus insulin was given as autocorrection boluses, reflecting the apparent inaccuracy in prandial bolus insulin administration. There was no change in reported carbohydrate intake or BMI-SDS, thus, the AHCL algorithm assisted in overcoming inaccurate carbohydrate counting or skipping insulin administration.

There are only a few studies on the MiniMed 780G system in children and adolescents followed for 1–12 months, and very limited data on real-life settings (Table 3). Petrovski et al.⁶ reported a significant improvement in TIR in a group of children with poor glycemic control, from 42% to 79% after 3 months, and Piccini et al.⁸ showed an increase from 69% to 77% in a group of adolescents in a real-life setting. Seget et al.⁹ on the contrary could not show a significant improvement in TIR in their study group, as baseline TIR was 79% and after 6 months 81%. In line with these reports, by using the rTIR, we were able to show that improvement is relative and the degree of improvement is dependent on baseline meters. Thus, those with poor glycemic control improve more than those who already reach the recommended TIR target.

Table 3.

Summary of Reports of Real-Life Data Regarding the Effectiveness of Advanced Hybrid Closed Loop After At Least 3 Months of Use, Compared with Standard Pump or Multiple Daily Injections

Country, author	N, design	Follow-up duration (months)	Age (years)	Disease duration (years)	Treatment modality at baseline	Glycemic control at baseline		Glycemic control at end		Conclusion/added information
Country, author	N, design	Follow-up duration (months)	Age (years)	Disease duration (years)	Treatment modality at baseline	TIR (70–180 mg/dL)	GMI%/HbA1c%	%TIR (70–180mg/dL)	GMI/HbA1c %	Conclusion/added information
Italy, Tornese et al.¹⁶	24 Retrospective	6	14.1 (9.2–15.9)	3.1 (1.1, 7.6)	MiniMed™ 670G	60 (46, 72)	7.8 (7.2, 8.5)	69 (63, 71)	7.1 (6.8, 7.2)	Younger age (<14 years), higher rate of autocorrection, lower amount of CHO inserted were positive predictors
Qatar, Petrovski et al.⁶	34 Prospective clinical trial	3	7–17	>1	MDI	42.1 ± 18.7	8.6 ± 1.7	78.8 ± 6.1	6.5 ± 0.7	High CGM, AHCL use, and team supervision, and support, and specific protocol might be additional positive predictor.
Carelink data, Arrieta et al.¹⁷	790 Retrospective	6	<15	No data	No data	62.1	7.2 ± 0.6	73.9	6.8 ± 0.4	Age ≤15 years was a positive predictor
Italy, Piccini et al.⁸	44 Retrospective	6	14.2 ± 4	7.5 (5.4–10.9)	MDI/CSII	69.3 ± 12.6	7.2 ± 0.7	76.3 ± 9.6	6.6 ± 0.5	TIR was higher in individuals >13 years at all time periods- including at baseline. Younger ones improved more with AHCL.
Poland, Seget et al.⁹	50 Prospective	12	9.9 ± 2.4 (5.4–16.8)	3.9 ± 2.6	No data	79.3 ± 8.1	6.5 ± 0.3	80.4 ± 8.2	6.5 ± 0.3	There was light increase in the total daily insulin dose, the percent of basal insulin was unchanged, amount of insulin in autocorrective boluses increased.
Our study	22 Retrospective	10.9	13.9 (1.0, 18.0)	5.5 (1.7, 10.1)	MDI/CSII	65% (52, 72)	7.4 (7.0, 8.3)	72 (65, 78)	7.3 (6.8, 8)	Pubertal stage, diabetes duration, older age, delayed pump and reservoir change are negative predictors

AHCL, advanced hybrid closed loop; CHO, carbohydrates; CSII, continuous subcutaneous insulin infusion; MDI, multiple daily injections.

Previous studies examined the effects of 780G system on glycemic outcomes in different age groups. However, the age cutoff for comparing the groups was not consistent and ranged between 13,⁸ 14,¹⁶ and 15 years.¹⁷ Our study is the first to analyze the data according to the Tanner stage of puberty. We demonstrate that advancement in Tanner stage was correlated with less improvement and more time spent at elevated glucose levels, particularly between 180 and 250 mg/dL, even with the AHCL. This was reflected by less GMI improvement and was associated with less CGM time usage and more days between pump site changes. When age was taken into account, Piccini et al.,⁸ reported that teenagers >13 years old demonstrated good technology adherence with optimal TIR and maintained greater TIR than children <13 years together with better glycemic variability indices.

The discrepancy with our findings may result from the fact that sexual maturity is not always correlated with age. Puberty is characterized by insulin resistance and behavioral changes. Euglycemic insulin-clamp studies in prepubertal and pubertal children with T1D showed that, at each stage of pubertal development, the stimulating effect of insulin on glucose metabolism was decreased.¹⁸ Furthermore, studies of the pharmacokinetic and pharmacodynamic properties of rapid acting insulin analogs showed a lower ability of an insulin bolus to stimulate glucose metabolism in pubertal compared to prepubertal subjects.¹⁹ On top of the insulin resistance that characterizes puberty, adolescence is characterized by a lack of compliance and less parental supervision, which may explain the better improvement in pre- and early pubertal children. We believe that switching to AHCL is an opportunity for re-educating adolescents.

Parallel to puberty, prolonged diabetes duration, and older age at baseline, were also correlated with less improvement in both TAR and TBR, indicating that disease burden is an accumulating parameter, as well as the ever-needed support and on-going education for sustained improvement.

We demonstrate an increase in CGM use from 67% (51, 94) to 84% (78, 91). The improved TIR in our cohort is in concordance with the findings of Carlson et al.²⁰ who showed an association between system usability and impact on glycemia. Our data show that poor handling of the pump and CGM were correlated with glycemic outcomes. More days between pump and reservoir changes were correlated with less TIR, more time spent above range at the end of the study, and increased GMI. Of importance is the fact that the children and adolescents either continued with their under-bolusing behavior, or, worse, learned that they could “trust” the technology and thus, were less cautious with insulin administration. A significant association was demonstrated between poorer glycemic outcomes and missing meal boluses or giving delayed boluses among adolescents with T1D.^21,22

As demonstrated in our study, up to one-third of bolus insulin was given by autocorrection boluses, the technological attempt to overcome those obstacles. Still, it was not enough, and this may explain why, despite improvement of the TIR throughout the follow-up period, we were not able to show an improvement in HbA1c, in contrast to other studies that found a significant improvement after 3 and 6 months.⁸ We therefore conclude that ongoing education emphasizing the importance of proper AHCL handling is required, and the most important instruction of premeal bolus should be reinforced.

It is noteworthy that the use of the AHCL system gave the medical team the confidence to change the indices of the insulin pump, as previously described.^6,9 The significant changes included shortening of the duration of the insulin activity, enhancement of the insulin-to-carbohydrate ratio, affecting both auto-mode and manual mode, and strengthening the correction factor, affecting the insulin bolus during the manual mode.

Our study has some limitations; first, it is a retrospective study. Second, it involves a population that could afford to purchase the 780G system, therefore, it may not represent all socioeconomic groups. Socioeconomic factors such as low income and low parental educational attainment are associated with a significantly higher HbA1c.^23,24 This limitation, however, became a strength in our cohort, it enabled us to demonstrate that a favorable SEP does not necessarily translate into good glycemic outcomes, and furthermore, it enabled us to pinpoint remaining barriers to diabetes management. Third, we report on a small study population, which did not allow us to perform multivariate analysis of the data. Nevertheless, we feel that the findings we report are valid and important.

Fourth, we allowed ourselves to use again¹⁴ the variable of rTIR, although it is not yet a validated parameter, indicating the ratio of improving of TIR to the possible improvement for that individual. It is not a rigorous parameter, its specificity and sensitivity according to delta HbA1c or GMI have not been tested; nevertheless, we believe that it may have additional insight in a partially well-controlled study-population. Finally, we report a cohort with CGMs requiring calibration, which was reported to be a significant factor in low closed-loop usage, and increased device burden. The strengths of this study are its real-world data analyzed according to pubertal stages, length of disease burden, and human behavior. Furthermore, we had no selection bias, our follow-up duration was longer than the previously reported 3–6 months, and everyone was his own comparative.

Our findings are important since the AHCL is on the market and knowing the limitations of the technology is of major importance. In conclusion, we show that the use of AHCL improved TIR, and decreased the time spent in hyper- and hypoglycemic ranges among children and adolescents. However, about one quarter of the bolus insulin was given as autocorrection. A more advanced pubertal stage, longer disease duration, and suboptimal engagement with management (longer time between pump site changes) were correlated with less improvement, and a worse outcome at the end of follow-up. These are the barriers to optimal glycemic outcomes that need further emphasis in care and patient re-education, as well as more research taking into consideration pubertal stage data.

Footnotes

Acknowledgment

We thank Dr. Ziv Tomer from the Department of Epidemiology and Preventive Medicine, School of Public Health, Tel Aviv University, for the statistical analysis and consultation.

Authors' Contributions

M.R. and O.P.-H. made substantial contributions to the conception and design of the study, acquisition, analysis, and interpretation of data, and drafting the initial article and revising it critically for important intellectual content. K.M.-A., A.B., N.L., T.J.P., T.B.-A., and S.A. contributed to the data used in this study, contributed to the discussion, and reviewed and edited the article. Z.L., and Y.L. made substantial contributions to the conception and design of the study, the interpretation of data, contributed to the discussion, reviewed, and edited the article for important intellectual content. All authors gave final approval of the version to be published. M.R. is the guarantor of this work, and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Availability of Data and Material

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

No funding was received for this article.

References

AlBurno

, Mercken

, de Vries

, et al. Socio-cognitive determinants affecting insulin adherence/non-adherence in late adolescents and young adults with type 1 diabetes: A systematic review protocol. J Diabetes Metab Disord, 2022; 21(1):1207–1215.

McVean

, Miller

. MiniMed^TM780G Insulin pump system with smartphone connectivity for the treatment of type 1 diabetes: Overview of its safety and efficacy. Expert Rev Med Devices, 2021; 18(6):499–504.

Silva

, Lepore

, Battelino

, et al. Real-world performance of the MiniMed™ 780G System: First report of outcomes from 4120 users. Diabetes Technol Ther, 2022; 24(2):113–119.

Beato-Víbora

, Gallego-Gamero

, Ambrojo-López

, et al. Rapid improvement in time in range after the implementation of an advanced hybrid closed-loop system in adolescents and adults with type 1 diabetes. Diabetes Technol Ther, 2021; 23(9):609–615.

Matejko

, Juza

, Kieć-Wilk

, et al. Transitioning of people with type 1 diabetes from multiple daily injections and self-monitoring of blood glucose directly to MiniMed 780G advanced hybrid closed-loop system: A two-center, randomized, controlled study. Diabetes Care, 2022; 45(11):2628–2635.

Petrovski

, Al Khalaf

, Campbell

, et al. Successful transitioning children and adolescents with type 1 diabetes from multiple daily injections to advanced hybrid closed-loop system in 10 days: A prospective intervention study on MiniMed 780G system. Acta Diabetol, 2022; 59(5):743–746.

Collyns

, Meier

, Betts

, et al. Improved glycemic outcomes with Medtronic MiniMed advanced hybrid closed-loop delivery: Results from a randomized crossover trial comparing automated insulin delivery with predictive low glucose suspend in people with type 1 diabetes. Diabetes Care. 2021; 44(4):969–975.

Piccini

, Pessina

, Casalini

, et al. Long-term effectiveness of advanced hybrid closed loop in children and adolescents with type 1 diabetes. Pediatr Diabetes, 2022; 23(8):1647–1655.

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.

10.

Characterization and Classification of Statistical Areas Within Municipalities and Local Councils by the Socio-Economic Level of the Population 2015. Available from: https://cbs.gov.il/mediarelease/pages/2019.

11.

Goldstein

, Haelyon

, Krolik

, et al. Comparison of body weight and height of Israeli schoolchildren with the Tanner and Centers for Disease Control and Prevention growth charts. Pediatrics, 2001; 108(6):E108.

12.

Tanner

. Growth and maturation during adolescence. Nutr Rev, 1981; 39(2):43–55.

13.

Bergenstal

, Beck

, Close

, et al. Glucose Management Indicator (GMI): A new term for estimating A1C from continuous glucose monitoring. Diabetes Care, 2018; 41(11):2275–2280.

14.

Rachmiel

, Lebenthal

, Mazor-Aronovitch

, et al. Glycaemic control in the paediatric and young adult population with type 1 diabetes following a single telehealth visit—What have we learned from the COVID-19 lockdown?. Acta Diabetol, 2021; 58(6):697–705.

15.

Taylor

. Interpretation of the correlation coefficient: A basic review. J Diagn Med Sonogr, 1990; 6:35–39.

16.

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.

17.

Arrieta

, Battelino

, Scaramuzza

, et al. Comparison of MiniMed 780G system performance in users aged younger and older than 15 years: Evidence from 12 870 real-world users. Diabetes Obes Metab, 2022; 24(7):1370–1379.

18.

Amiel

, Sherwin

, Simonson

, et al. Impaired insulin action in puberty. A contributing factor to poor glycemic control in adolescents with diabetes. N Engl J Med 24 1986; 315(4):215–219.

19.

Swan

, Weinzimer

, Dziura

, et al. Effect of puberty on the pharmacodynamic and pharmacokinetic properties of insulin pump therapy in youth with type 1 diabetes. Diabetes Care, 2008; 31(1):44–46.

20.

Carlson

, Sherr

, Shulman

, et al. Safety and glycemic outcomes during the MiniMed advanced hybrid closed-loop system pivotal trial in adolescents and adults with type 1 diabetes. Diabetes Technol Ther, 2022; 24(3):178–189.

21.

Datye

, Boyle

, Simmons

, et al. Timing of meal insulin and its relation to adherence to therapy in type 1 diabetes. J Diabetes Sci Technol, 2018; 12(2):349–355.

22.

Hilgard

, Konrad

, Meusers

, et al. Comorbidity of attention deficit hyperactivity disorder and type 1 diabetes in children and adolescents: Analysis based on the multicentre DPV registry. Pediatr Diabetes, 2017; 18(8):706–713.

23.

Alvarez-Ramos

, Jimenez-Carmona

, Alemany-Marquez

, et al. Socioeconomic deprivation and development of diabetic retinopathy in patients with type 1 diabetes mellitus. BMJ Open Diabetes Res Care, 2020; 8(2):e001387.

24.

Gallegos-Macias

, Macias

, Kaufman

, et al. Relationship between glycemic control, ethnicity and socioeconomic status in Hispanic and white non-Hispanic youths with type 1 diabetes mellitus. Pediatr Diabetes, 2003; 4(1):19–23.