Use of Freestyle Libre 2.0 in children with type 1 diabetes mellitus and elevated HbA_1c : Extension phase results after a 12‐week randomized controlled trial

Abstract

Aim

To investigate extension phase outcomes with intermittently scanned continuous glucose monitoring (isCGM 2.0) in children with type 1 diabetes mellitus (T1DM) and elevated HbA_1c (7.5–12.2% [58–110 mmol/mol]).

Methods

One hundred children with T1DM aged 4–13 years were initially in a 12‐week randomised controlled trial (RCT) comparing glycaemic outcomes with isCGM 2.0 (intervention group, n = 49) with self‐monitored blood glucose (Control group, n = 51). After the 12‐week RCT both groups were offered an extension phase with isCGM 2.0 for another 12 weeks. HbA_1c, CGM metrics, psychological outcomes and device utilization attitudes were measured.

Results

After the initial 12‐week RCT, 66 participants completed this 12‐week extension: 36/49 (73%) and 30/51 (58.8%) from the isCGM/isCGM and Control/isCGM groups, respectively. In the isCGM/isCGM group, time below range 70 mg/dL (3.9 mmol/L) (TBR70) reduced from 10.7 ± 11.3% at baseline to 2.8 ± 2.8% and 2.1 ± 2.4% at 12 and 24 weeks, respectively (p < 0.01 for both 12 and 24 weeks). Glucose test frequency increased from 4.7 (2.7) at baseline to 10.7 (4.6) and 9.2 (4.7) at 12 and 24 weeks, respectively (p < 0.01 for both 12 and 24 weeks). The Control/isCGM group decreased TBR70 from 10.7 ± 7.4% at 12 weeks to 2.9 ± 2.8% at 24 weeks and increased daily glucose test frequency from 3.2 (1.6) to 10.7 (5.4) from 12 to 24 weeks (both p < 0.01). However, HbA_1c and time in range (TIR) were non‐significant at 24 weeks in both groups.

Conclusions

Extension phase outcomes with intermittently scanned continuous glucose monitoring (isCGM 2.0) in children with T1DM and elevated HbA_1c showed a sustained reduction in hypoglycaemia and increased testing frequency at 24 weeks.

What's new?

What is already known?

Our previous 12‐week RCT of isCGM in children with elevated HbA_1c improved testing frequency and reduced hypoglycaemia.

What this study has found

This extension study found a sustained reduction in hypoglycaemia and increased glucose testing in the intervention isCGM/isCGM group. The Control SMBG/isCGM group following crossover to isCGM for 12 weeks have showed a similar reduction in hypoglycaemia and increased testing frequency.

What are the implications of the study?

At 6 months, isCGM 2.0 continues to have beneficial effects on reducing hypoglycaemia and increasing glucose testing in children with elevated HbA_1c.

1. INTRODUCTION

Type 1 diabetes mellitus (T1DM) is challenging to manage, especially for school‐age children who are required to monitor their glucose levels at least 6–10 times per day. ¹ , ² A range of factors contribute to management challenges in children, including difficulties articulating hypoglycaemia symptoms; the potential for greater glycaemic variability than adults and more needle phobia for fingersticks and stigma during routine management. ³ All of these result in real‐world outcomes of elevated HbA1c for the majority, which subsequently increases the risk for acute and chronic diabetes complications. ⁴

Continuous glucose monitoring systems (real‐time systems [rtCGM] or intermittently scanned systems [isCGM]), can offer advantages over fingerstick testing and are currently recommended as a part of standard care in T1DM. ¹ , ⁵ , ⁶ This technology is rapidly evolving and whilst Libre 2.0 now offers rtCGM functions, at the time this initial trial ⁷ was conducted isCGM 2.0 was novel in much of the world and came with only optional threshold‐based alarms resembling rtCGM, notifying out‐of‐range glycaemia for timely intervention.

In general, CGM (of all types) is superior in improving HbA_1c and % time‐in‐range 70–180 mg/dL (3.9–10.0 mmol/L, TIR) and time below range when compared to fingerstick capillary glucose. ⁸ , ⁹ , ¹⁰ , ¹¹ However, for youth aged 13–20 years with baseline HbA_1c ≥ 7.5% (58 mmol/mol), using isCGM 1.0 for 24 weeks did not significantly improve glycaemic outcomes, despite diabetes satisfaction and monitoring frequency increased. ¹² Longer‐term observations also showed benefits were not sustained for 12 months. ¹³ We have shown a reduction in hypoglycaemia in our recent 12‐week randomized controlled trial (RCT) in children aged 4–13 years after 12 weeks with elevated HbA_1c, but without a reduction in HbA_1c. ¹⁴ This is the only RCT to evaluate isCGM 2.0 in children to date. Whether these changes are sustained or whether any further improvement could be achieved with longer isCGM 2.0 use is not known. Thus, the current study aimed to describe the ongoing impact of isCGM 2.0 on glycaemic outcomes, psychological outcomes and to investigate the attitudes on glucose monitoring performance amongst high‐risk young T1DM participants receiving an additional 12‐week free‐living extension following the original RCT.

2. METHODS

2.1. Study design and participants

The study flowchart is shown in Figure 1. In brief, 100 T1DM children (aged 4–13 years) with elevated HbA_1c between 7.5% and 12.2% (58–110 mmol/mol) were recruited from 5 diabetes centres that cover approximately one‐third of New Zealand's total population. After the 2‐week run‐in phase, participants were randomly allocated to either the intervention group, using FreeStyle Libre 2.0 system (Abbott Diabetes Care, Witney, UK) plus scheduled education, or the waitlist control group who continued using self‐monitored blood glucose (SMBG) for the 12 weeks RCT. Following this, both groups were then provided with 12 weeks of isCGM 2.0 supplies, and the Control group also received scheduled education when starting isCGM 2.0. Protocol and primary outcomes of the original RCT have been previously published. ¹⁴ , ¹⁵

FIGURE 1.

Study design. isCGM, intermittently scanned continuous glucose monitoring (isCGM); SMBG, self‐monitoring capillary blood glucose; FSL‐pro, Freestyle Libre Pro.

2.2. Ethics

The study was approved by the Northern A Health and Disability Ethics Committee (ethics reference: 20/NTA/12), underwent Māori (indigenous New Zealanders) consultation, and locality approvals at all study institutions. The trial is registered with the Australian New Zealand Trial Registry (ACTRN12620000190909p; ANZCTR.org.au) and the World Health Organization International Clinical Trials Registry Platform (Universal Trial Number U1111‐1237‐0090). The isCGM manufacturer was not involved with the planning, funding or conduct of the study.

2.3. Data collection

2.3.1. Demographics and glycaemic outcomes

Study visits took place at baseline, 12 and 24 weeks for collection of demographic and clinical data. HbA_1c was measured at all visit time points using a calibrated point of care device (DCA Vantage analyser; Siemens Healthcare Diagnostic Ireland) at local centres. Participants in the intervention group (isCGM/isCGM) and those who crossed over to use isCGM (Control/isCGM) were both encouraged to upload data every month if applicable.

Glycaemic data at up to 14 days post‐baseline and prior to the 12‐week RCT completion (Control group only) were collected from blinded sensors (FreeStyle Libre Pro, Abbott Diabetes Care, Witney, UK). For the intervention group at 12 and 24 weeks, and for the Control group at 24 weeks, data were extracted from the FreeStyle Libre 2.0 isCGM system 2 weeks prior. Interstitial glucose profiles were interpreted using recommended clinical targets, ¹⁶ with TIR, TBR70 (time below range 70 mg/dL [3.9 mmol/L]), TBR54 (time below range 54 mg/dL [3.0 mmol/L]), TAR (time above range 180 mg/dL [10.0 mmol/L]) and CV (coefficient of variability) calculated. Glucose test frequency was recorded from device downloads for the prior 14 days to capture the number of interstitial (scanned) and capillary blood glucose (finger‐prick) tests per day, using Freestyle Libre software (Abbott Diabetes Care, UK, version 1.0) and Smartlog Diabetes Management software (i‐SENS, Inc., Korea, version 2.4.4), respectively.

2.3.2. Psychological outcomes

Psychological outcomes and overall diabetes treatment acceptance were collected through validated self‐report questionnaires completed online using REDCap (Research Electronic Data Capture) software. At baseline, 12 and 24 weeks visits, children and parents were asked to complete age‐appropriate psychological questionnaires examining quality of life (PedsQL Diabetes Module 3.2 young child/child/teen version, as appropriate), ¹⁷ fear of hypoglycaemia (Hypoglycaemia Fear Survey. HFS), ¹⁸ and participants aged 10–14 years (inclusive) were asked to complete the Self Efficacy for Diabetes Self‐Management questionnaire (SESDM). ¹⁹ The isCGM acceptability was evaluated for all participants 12 weeks after using isCGM 2.0 via a non‐validated instrument ²⁰ that rates the self‐reported opinion regarding the acceptability of sensor application, device wear/use and comparison to SMBG.

2.4. Statistical analysis

All analyses in the extension phase were conducted based on randomized treatment group assignment and on available data only, referring to those who completed the 24‐week visit. Demographic and clinical data descriptions were presented with mean and standard deviation (SD) for continuous variables, number, and percent of participants for categorical variables. Compared to the analysis of 12‐week RCT data, the extension data analyses are within‐group, not between‐group and the Control group data are provided to highlight similar trends in glycaemic outcomes. Changes were tested using paired t‐tests for normally distributed continuous variables and non‐parametric tests for non‐normally distributed variables.

3. RESULTS

3.1. Demographic and clinical characteristics

A total of 100 participants were initially randomly assigned to intervention (n = 49) or Control (n = 51) groups; from this, 92 participants completed the 12‐week RCT phase. ¹⁴ Subsequently, 66 participants completed this extension study: 36 of 49 (71.4%) in the isCGM/isCGM cohort, and 30 of 51 (60.8%) in the Control/isCGM group. Children completing this study were aged 11.1 ± 2.3 years, with 54.5% female, with a diabetes duration of 4.4 ± 3.1 years and mean HbA_1c 8.9 ± 1.2% (74 ± 13 mmol/mol). Despite nearly a 30% dropout in both groups, the demographics between baseline and 24‐week visits were similar (p all >0.05) (see Table 1).

TABLE 1.

Demographic and clinical characteristics of the participants.

	Baseline (n = 100)	12 weeks (n = 92)	24 weeks (n = 66)
Age (years), mean (SD)	10.9 (2.3)	10.9 (2.3)	11.1 (2.3)
Female (%)	58 (58.0)	52 (57.1)	36 (54.5)
Prioritized ethnicity, n (%)
Māori a	25 (25.0)	22 (24.2)	14 (21.2)
Pacific youth	22 (22.0)	18 (19.8)	10 (15.2)
New Zealand European	37 (37.0)	36 (39.6)	29 (43.9)
Asian/other	15 (15.0)	15 (16.5)	13 (19.7)
NZ deprivation index, n (%) b
Quintiles 1–3 (low deprivation)	26 (28.9)	25 (30.1)	19 (32.8)
Quintiles 4–7 (medium deprivation)	31 (34.4)	29 (34.9)	21 (36.2)
Quintiles 8–10 (high deprivation)	33 (36.7)	29 (34.9)	18 (31.0)
BMI (z‐score) c , mean (SD)	0.9 (0.9)	0.9 (0.9)	0.9 (0.9)
Diabetes duration (years), mean (SD)	4.2 (3.0)	4.2 (2.9)	4.4 (3.1)
Insulin therapy d , n (%)
MDI	83 (83.8)	75 (82.4)	53 (80.3)
CSII	16 (16.2)	16 (17.6)	13 (19.7)
Insulin estimated total daily dose (units/kg/day), mean (SD)	1.0 (0.4)	1.0 (0.4)	1.0 (0.4)
HbA1c (mmol/mol), mean (SD)	75 (14)	75 (14)	74 (13)
HbA1c (%), mean (SD)	9.1 (1.3)	9.0 (1.2)	8.9 (1.2)
SMBG checks/day, mean (SD)	4.3 (2.5)	4.3 (2.5)	4.4 (2.5)

Abbreviations: BMI, body mass index; CSII, continuous subcutaneous insulin infusion; HbA_1c, haemoglobin A1c; MDI, multiple daily injections; SD, standard deviation; SMBG, self‐monitoring capillary blood glucose.

^a Māori are the indigenous people of New Zealand, (note that 1/100 (1%) missing data as 1 participant did not complete ethnicity at baseline).

^b NZDep18, The New Zealand Deprivation Index is an area‐based measure of socioeconomic deprivation (in which 1 represents the least socioeconomic deprivation and 10 the most deprived). Post office boxes and some rural addresses cannot be derived from this index.

^c BMI (z‐score), calculated using the Centre for Disease Control Guidelines.

^d 1/100 (1%) missing data in insulin therapy.

Post‐hoc analysis for the two groups completing the extension phase, and comparisons between those lost to follow‐up and those completing 24 weeks are shown in Supplementary Tables S1 and S2. Amongst 34 participants lost to follow‐up, 67.6% identified as Māori (n = 11/34, 32.3%) or Pacific (n = 12/34, 35.3%), and 46.9% (15/34) lived in areas of high deprivation. Additionally, 90.9% (30/34) used MDI therapy prior to the study and those lost to follow‐up had higher baseline HbA_1c, 9.4 ± 1.3% (79 ± 15 mmol/mol) compared with those completing (8.9 ± 1.2% [74 ± 13 mmol/mol]), which did not reach significance (p = 0.09).

3.2. Device usage

The 49/65 (75.4%) participants set low alerts ranging from 67 to 81 mg/dL (3.7–4.5 mmol/L) and 46/65 (70.7%) participants set high alerts ranging from 162 to 288 mg/dL (9–16 mmol/L) at isCGM 2.0 initiation. Mean daily glucose test frequency in the isCGM/isCGM group significantly improved after 12 weeks and remained stable at 24 weeks end, with values increasing from 4.7 ± 2.7 at baseline to 10.7 ± 4.6 at 12 weeks and 9.2 ± 4.7 times/day in study end. For the Control/isCGM group, after crossing over, glucose test frequency also increased from 3.2 ± 1.6 to 10.7 ± 5.4 times/day.

3.3. Glycaemic outcomes

HbA_1c did not significantly change throughout 24 weeks in the isCGM/isCGM group (Table 2) with a mean difference of 0.11% (95% CI: −0.27 to 0.47%, p = 0.56) between baseline and 12 weeks and of 0.13% (95% CI: −0.25 to 0.51%, p = 0.50) between 12 and 24 weeks. HbA_1c in the Control/isCGM group was also similar, with a mean difference: 0.32% [95% CI −0.04 to 0.67%]; p = 0.08) from 12 to 24 weeks.

TABLE 2.

Glycaemic outcomes (HbA_1c and CGM‐related metrics) at each study time point for the isCGM/isCGM and Control/isCGM groups.

Glycaemic measures, mean (SD)	Baseline (N = 100)	12 weeks (N = 92)	24 weeks (N = 66)
isCGM/isCGM (intervention group)	n = 49	n = 45	n = 36
SMBG+ccan frequency (times/d) a	4.7 (2.7)	10.7 (4.6)	9.2 (4.7)
HbA1c, mmol/mol	75 (13)	75 (14)	77 (13)
HbA1c, %	9.0 (1.2)	9.0 (1.2)	9.2 (1.2)
CGM metrics b
%TAR	60.6 (18.8)	61.1 (13.3)	65.8 (14.7)
%TIR	28.7 (16.6)	36.4 (13.2)	32.1 (13.0)
%TBR70	10.7 (11.3)	2.9 (2.8) c	2.2 (2.4) c
%TBR54	3.7 (4.8)	0.9 (2.7) c	0.2 (0.3) b
%CV	46.1 (8.5)	40.8 (4.7)	41.4 (4.5)
Control/isCGM (Control group)	n = 51	n = 47	n = 30
SMBG+scan frequency (times/d) a	3.9 (2.2)	3.2 (1.6)	10.7 (5.4)
HbA1c, mmol/mol	76 (15)	75 (14)	77 (13)
HbA1c, %	9.1 (1.4)	9.0 (1.3)	9.2 (1.3)
CGM metrics b
%TAR	62.2 (20.9)	64.3 (16.6)	61.4 (16.9)
%TIR	27.3 (16)	25.1 (12.9)	35.5 (15.5)
%TBR70	10.0 (8.1)	10.7 (7.2)	3.2 (3.0) d
%TBR54	4.3 (4.9)	3.7 (4.0)	0.3 (0.5) d
%CV	46.5 (9.9)	46.9 (8.5)	42.0 (5.5)

Note: p‐value from a paired t‐test or signed rank test, as appropriate. Only analysed for participants who had values at both time points.

Abbreviations: CGM, continuous glucose monitoring; CV, coefficient of variance; HbA_1c, haemoglobin A1c; isCGM, intermittently scanned continuous glucose monitoring; SD, standard deviations; SMBG, self‐monitoring capillary blood glucose, TAR, time above range 180 mg/dL (10.0 mmol/L); TBR54, time below range 54 mg/dL (3.0 mmol/L); TBR70, time below range 70 mg/dL (3.9 mmol/L); TIR, time in range 70–‐180 mg/dL (3.9–‐10.0 mmol/L).

^a Glucose frequency tests at baseline for both groups and for the Control/isCGM group at 12 weeks were calculated from self‐reported SMBG frequency. For isCGM/isCGM at 12 weeks and both groups at 24 weeks, glucose frequency tests were the combined numbers of sensor scans and capillary blood glucose tests.

^b Available CGM metrics in the isCGM/isCGM group were calculated from 44 participants at baseline, 37 participants at 12 weeks and 29 at 24 weeks. For the Control/isCGM group, those were from 43 participants at baseline, 28 at 12 weeks and 16 at 24 weeks.

^c Nonparametric test (n = 23) showed that compared with baseline, p‐values are <0.01.

^d Nonparametric test (n = 10) showed that compared with 12 weeks, p‐value for 24 weeks is significant (p < 0.01).

There were 45 (68.2%) participants bringing the reader and/or uploading data at 24 weeks end. In the isCGM/isCGM group, TBR70 significantly decreased from 10.7 ± 11.3% to 2.9 ± 2.8% at 12 weeks, which maintained at 24 weeks (mean difference from baseline was 6.5% [95%CI: 4.1 to 9.4%]; p < 0.01). TBR54 also decreased by 3.3% (95%CI: 1.2 to 5.4%; p < 0.01). In the Control/isCGM group, TBR70 decreased by 8.2%, 95%CI: [3.2 to 13.1] % from 10.0 ± 8.1% to 3.2 ± 3.0% (p = 0.004); TBR54 decreased from 3.7 ± 4.0% to 0.3 ± 0.5% yet did not reach statistical significance (p = 0.09). No significant changes in other metrics were observed in both groups (p > 0.05). Similar findings were observed in the intention‐to‐treat analysis (Supplementary Table S3).

3.4. Psychological outcomes

In the isCGM/isCGM group, improvement in quality of life after 24‐week isCGM use was observed with scores of 57.1 ± 12.9 improving to 61.1 ± 14.0 (mean difference: 4.51 [−0.04, 9.07]; p = 0.05) (Table 3); this improvement was particularly significant in the symptoms sub‐section, with scores of 52.7 ± 13.0 increasing to 56.9 ± 15.0 (mean difference: 5.86 [0.16, 11.57]; p = 0.04). No significant change in the HFS and SEDSM scores was observed. In the Control/isCGM group, differences after isCGM use were not significant (p all >0.05) in all psychological outcomes assessed amongst both participants and parents.

TABLE 3.

Changes in psychological outcomes for children and parents from baseline to 12 and 24 weeks.

Children's Scores	Control/isCGM group (Control group)		isCGM/isCGM (intervention group)
	∆ 3‐month change (95% [CI]) a	p‐value	∆ 6‐month change (95% [CI]) a	p‐value
	(at 24 weeks from 12 weeks)		(at 24 weeks from baseline)
PedsQL b
Total quality of life	4.5 (−0.0, 9.1)	0.052	−1.7 (−5.3, 1.9)	0.337
Diabetes quality of life	5.9 (0.2, 11.6)	0.044	−2.2 (−6.2, 1.8)	0.268
Treatment	5.3 (−2.4, 13.1)	0.170	3.8 (−1.9, 9.5)	0.179
Adherence	4.8 (−2.6, 12.2)	0.195	−1.3 (−6.9, 4.3)	0.628
Worry	7.1 (−2.6, 16.8)	0.145	−1.7 (−9.4, 6.0)	0.659
Communication	7.7 (−2.2, 17.5)	0.120	−5.5 (−13.2, 2.2)	0.154
HFS c
Behaviour subscale	−0.1 (−0.4, 0.2)	0.640	0.0 (−0.3, 0.2)	0.865
Worry subscale	0.1 (−0.4, 0.5)	0.685	0.1 (−0.1, 0.3)	0.311
SEDSM d
Total scores	0.4 (−0.5, 1.2)	0.371	0.1 (−0.6, 0.8)	0.819
Parents' scores
PedsQL b
Total quality of life	2.4 (−2.4, 7.2)	0.310	−1.9 (−6.2, 2.5)	0.388
Diabetes quality of life	2.7 (−3.5, 8.8)	0.379	−2.2 (−6.7, 5.4)	0.341
Treatment	1.1 (−5.0, 7.1)	0.719	1.1 (−5.8, 8.0)	0.743
Adherence	0.5 (−6.4, 7.3)	0.894	−3.1 (−9.8, 3.6)	0.354
Worry	1.3 (−11.9, 14.6)	0.837	−1.5 (−13.0, 9.9)	0.783
Communication	6.0 (−2.0, 14.1)	0.136	−3.5 (−14.2, 7.2)	0.510
HFS e
Behaviour subscale	−0.1 (−0.4, 0.2)	0.400	−0.1 (−0.3, 0.2)	0.519
Worry subscale	0.2 (−0.2, 0.6)	0.332	0.0 (−0.3, 0.3)	0.899

Abbreviations: CI, confidence interval; HFS, hypoglycaemia fear survey; isCGM, intermittently scanned continuous glucose monitoring; PEDsQL, paediatric quality of life inventory; SEDSM, self‐efficiency of diabetes self‐management; SMBG, self‐management capillary blood glucose.

^a Within‐group changes using paired t‐test or nonparametric rank test as appropriate. Only included participants who had values at both time points.

^b PedsQL total scores and sub‐scales scores range from 0 to 100. A higher total score indicates better diabetic‐specific quality of life and higher sub‐scales scores means lower problems of diabetes‐specific symptoms/barriers/adherence/worry/communications, respectively. Available data for children and parent's version is n = 27 in the isCGM/isCGM group, and n = 24 in the Control/isCGM group. Demographics between the two groups were non‐significant.

^c HFS behaviour subscale means item scores range from 0 to 4; higher scores indicate a greater tendency to avoid hypoglycaemia. HFS‐Worry subscale scores range from 0 to 4; higher scores indicate more worry concerning episodes of hypoglycaemia and its consequences. Available data for analysis is n = 27 in the isCGM/isCGM group and n = 24 in the Control/isCGM group. Demographics between the two groups were non‐significant.

^d For children aged 10 years or older. Self‐Efficacy for Diabetes Self‐Management (SEDM) is a 10‐item self‐report questionnaire for youth aged 10–16 years that examines confidence to carry out self‐care behaviours and covers all the key areas of diabetes self‐management. Participants are asked ‘How sure are you that you can do each of the following, almost all the time’ and items are rated from 1 (not at all sure) to 10 (completely sure) and averaged. Available data for analysis are n = 19 in the isCGM/isCGM group and n = 22 in the Control/isCGM group. Demographics between the two groups were non‐significant.

^e HFS‐parents version is only for children aged up to 14 years. Available data are n = 18 for the isCGM/isCGM group and n = 26 for the Control/isCGM group.

3.5. Attitudes for isCGM after at least 3‐month use

Most (>80%) participants and their parents were satisfied with practical sensor use (Table S4), agreeing that it is comfortable to wear (Q4), easy to put on (Q2)/scan (Q5) particularly when compared with fingerpick testing (Q6–Q8), and the easier access to more glucose information (Q13 and Q14). However, 32% (21/66) of participants did not agree with the statement ‘it does not hurt when the sensor is put on’ (Q1). When asked about whether the sensor makes participants feel more motivated to improve their (or their child's) diabetes management, 30.3% (20/66) of participants were undecided.

4. DISCUSSION

Our extension study findings following 24 weeks of isCGM 2.0, largely re‐iterate the 12‐week RCT results. These highlight sustained improvements in hypoglycaemia and glucose test frequency. In addition, as with the original RCT, ¹⁴ there were also no significant differences in HbA_1c nor TIR outcomes in either group changed from baseline to 24 weeks of isCGM use, despite trends to improved TIR (largely by reduced time in hypoglycemia) as well as improved quality of life observed in isCGM/isCGM group.

These results contrast with the FLASH‐UK isCGM 2.0 trial in adults where significant improvements were seen in HbA_1c over 6 months. Differences here are the much younger age and more diverse ethnicity in our cohort. However, findings from the only RCT conducted in children and adolescents using first‐generation isCGM are remarkably similar to these second‐generation findings. ¹² One reason for the difference in findings between the adult and paediatric studies may be related to being more ethnically diverse backgrounds of participants we included, with more than 40% being Māori and/or Pacific people who are at higher risk of less optimal glycaemia and unfortunately higher rates of complications than Europeans for complex reasons. ²¹ , ²² Another consideration is the relatively lower glucose monitoring frequency than other studies ¹⁰ , ¹¹ (9.4 times/day in the isCGM/isCGM group and 6.0 times/day in the Control/isCGM group). Not all achieved the recommended six times/day (or put in another way approximately 25% scan less frequently). ²³ All of these again highlighted the necessary comprehensive management for this challenging multi‐ethnic group with a history of high glycemia, as gaining access to more glucose values is the important first step. What is then needed is safe treatment and accurate insulin administration which requires greater skills/knowledge, more third‐party assistance (from caregivers or professionals) and higher self‐motivation for healthier glycemia to be achieved. ²⁴ To this end, this group may need a technology that does more than simply inform on glucose levels to have the impact on reducing HbA_1c; automated insulin delivery technology, which has been recently reported to achieve unprecedented glycaemic improvements amongst these high‐risk youth is the next logical step. ²⁵ , ²⁶ , ²⁷

Improvements in hypoglycaemia (equal to around 1.6 h/day decrease) were evident and were sustained for 24 weeks, which is broadly similar to most RCTs of CGMs with alarms. ²⁸ This benefit was not observed in RCTs focusing on high‐risk T1DM groups amongst children and adults using isCGM (1.0) without alarms despite 12‐month use, ¹⁰ , ¹² , ¹³ suggesting alarms are a key benefit of this next‐generation technology. Unfortunately, information on how often alarms are used was not able to be accessed, a similar limitation to the adult RCT. ¹¹ The trend to improved quality of life could also be attributed to the reduction in hypoglycaemia, even though they did not reach statistical significance in some psychosocial measures. The real‐time availability and function of predictive low/high alerts equipped in most rtCGMs was not available in isCGM 2.0 when our study was initiated. Recently, the manufacturers have moved forward with compatible sensors/software enabling isCGM (2.0) to receive real‐time glucose readings every minute and the Freestyle Libre 3.0 is also real‐time.

To the best of our knowledge, this is the only extension trial to date of isCGM 2.0 in children, capturing the post‐RCT data to explore longer period efficacy in free‐living environments and aligning with the study duration of the only RCT in adults with T1DM. ¹¹ Additionally, the study was independent of the manufacturer. We also intentionally included the groups from multiple ethnic and indigenous backgrounds thus providing broader evidence for global clinical practice. Another challenge for the study was that the extension phase was conducted during the COVID‐19 pandemic, therefore some participants could not complete all in‐person visits and data were excluded, so bias from missing data cannot be ruled out.

In conclusion, the present study confirms the improvements in hypoglycaemia and glucose test frequency after 12 weeks of isCGM use, now sustained out to 24 weeks. Additionally, a trend towards improved quality of life and positive acceptance was observed although these did not translate into HbA_1c or statistically significant TIR improvement. Further investigations are required to explore the optimal combination of diabetes technology and behavioural interventions to improve glycaemic outcomes in children with type 1 diabetes.

AUTHOR CONTRIBUTIONS

C.J., B.J.W., S.E.S., H.R.C. and E.J.W. were involved with the design of the study. B.J.W., A.B., S.E.S., B.C., H.R.C., A.L., V.C., E.J.W., A.S.S. and C.J. were involved with recruitment and data collection. Y.W.Z. and V.R.M. performed the statistical analyses. Y.W.Z., E.J.W., B.J.W. and C.J. wrote the first draft of the manuscript and all authors worked collaboratively to review and prepare the final manuscript. C.J. is the guarantor of this work and, as such, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest to declare that are relevant to the content of this article.

Supporting information

Data S1. Supporting Information.

Zhou Y, Wheeler BJ, Boucsein A, et al. Use of Freestyle Libre 2.0 in children with type 1 diabetes mellitus and elevated HbA_1c : Extension phase results after a 12‐week randomized controlled trial. Diabet Med. 2025;42:e15494. doi: 10.1111/dme.15494