Abstract
Background:
The CareSens Air (CSAir) continuous glucose monitoring system is approved for use in adults in South Korea and received adults-only CE marking in early 2024. The aim of this prospective, multicenter, pivotal study in children and adolescents with insulin-dependent diabetes mellitus was to determine if the performance and safety of CSAir in these populations are comparable to those observed in adults.
Method:
Sensors were worn for 15 days on the upper arm in children and adolescents aged 7 to 17 years and on the upper arm and/or upper buttock in children aged 2 to 6 years. Frequent capillary comparator measurements were performed during in-clinic sessions at the beginning, middle, and end of the sensor-wearing period. Point accuracy was assessed via agreement rates (AR) and mean absolute relative difference (MARD). Additional assessments included clinical point accuracy, technical reliability, and safety.
Results:
For subjects aged 7 to 17 years, %20/20 AR was 90.8%, and MARD was 8.9%. A total of 99.9% of values fell within Zones A and B of the Diabetes Technology Society error grid. Sensor survival probability reached 75.2%. In children aged 2 to 6 years, %20/20 ARs were 86.6% and 98.4%, and MARDs were 10.3% and 7.8% for arm- and buttock-worn sensors, respectively. No serious adverse events were reported. Mild and moderate adverse device effects comprised skin issues such as itching or erythema at the sensor application or insertion site.
Conclusions:
The observed performance and safety of CSAir in children and adolescents are similar to those in adults. These data may support an extension of regulatory approval of CSAir for children aged 2 years and older.
Keywords: CareSens Air, CGM accuracy, children and adolescents, type 1 diabetes
Introduction
Many people with type 1 diabetes (T1DM) use continuous glucose monitoring (CGM) systems to monitor and manage their blood glucose, which has been shown to improve their glycemic control. 1 Similar benefits of CGM use on glycemic control have been reported in children and adolescents with T1DM.2-4 Generally, the physiology of children cannot directly be projected to adults; thus, a CGM system that has been proven to be safe and accurate in adults might perform differently in children. Consequently, a designated evaluation of CGM performance in pediatric populations is necessary and required by regulatory bodies. 5
The first-generation CareSens Air CGM system (CSAir, i-SENS Inc., Seoul, Republic of Korea) has been approved by the South Korean Ministry of Food and Drug Safety 6 and received CE marking for adjunctive use in adults in February 2024. 7 A subsequent update to the calibration algorithm, implemented in the marketed product in 2025, reduced the warm-up time to 30 minutes, rendered calibrations optional, and changed the CE-marked intended use from adjunctive to non-adjunctive use. A recent study in adults with diabetes demonstrated robust performance and reliability, with improved accuracy following the algorithm update, confirming the system’s suitability for non-adjunctive use. 7 The aim of the present multicenter pivotal study was to evaluate the performance and safety of CSAir in children and adolescents (age 2-17 years) with T1DM. At the time of study execution, the current algorithm was not yet available; thus, it was retrospectively applied to the raw data, and the resulting CGM readings were analyzed.
Materials and Methods
Study Design
This study was conducted between March and December 2025 as a prospective, multicenter, single-arm, open-label, pivotal study in children and adolescents with T1DM at 4 different clinical study sites across Germany. The study was organized, managed, and evaluated by the Institute for Diabetes Technology Ulm GmbH (IfDT) and sponsored by the manufacturer of CSAir.
The recruitment goal of the study was to enroll 58 subjects aged 7 to 17 years and 25 subjects aged 2 to 6 years. The population size for subjects aged 7 to 17 years was based on a formal sample size calculation, which assumed, among other factors, expected and minimum %20/20 agreement rates (ARs) (see definition below) of 85.5% and 82.5%, respectively, as well as a power and significance level of 90% and 2.5%, respectively. The population size for subjects aged 2 to 6 years was chosen such that they would make up 30% of the total targeted subject population of 83.
The main inclusion criteria were a clinical diagnosis of T1DM and intensified insulin therapy for at least 3 months, willingness to comply with protocol requirements, and signing of informed consent by the subject and/or parent/legal guardian. The main exclusion criteria were hypoglycemia unawareness, severe hypoglycemia or diabetic ketoacidosis in the 3 months before screening, skin abnormalities or diseases at the CGM sensor application sites, a chronic medical, psychiatric, or developmental/behavioral condition that, at the discretion of the investigator, would affect their performance in the study, intake of substances or planned medical procedures known to affect the performance of the examined CGM system, pregnancy, the lactation period, or lack of a negative pregnancy test (for female subjects with sexual maturity).
Ethical approval was granted by the responsible ethics committee, the competent authority was notified of the study, and it was registered with the German Clinical Trials Register (DRKS00035064). The study was conducted in compliance with the Declaration of Helsinki.
CGM Use
In accordance with the 15-day wearing time of the CSAir sensors, the study duration was 16 calendar days for each study participant. Sensor insertion was performed by the subjects or their guardians under the supervision of the study staff, and sensors could optionally be fixed with an additional overpatch. All sensors originated from 2 manufacturing lots. In subjects aged 7 to 17 years, only the arm was assessed as a wearing location, and either 1 or 2 sensors could be worn simultaneously. If 2 sensors were worn, 1 sensor was inserted in each upper arm. In addition, 1 sensor was defined as the primary sensor prior to insertion. In subjects aged 2 to 6 years, both the arm and the upper buttock were designated as possible sensor-wearing sites. Here, 2 sensors were inserted per subject and worn simultaneously in 3 possible configurations: 1 sensor on each upper arm, 1 sensor on each side of the upper buttock, or 1 sensor on the arm, and 1 sensor on the upper buttock. Sensors failing within the first 24 hours after insertion could be replaced once.
Comparator Measurement Approach
Continuous glucose monitoring readings were compared to capillary blood glucose measurements using the Contour Next blood glucose monitoring system (Ascensia Diabetes Care Holdings AG, Basel, Switzerland). All capillary measurements were carried out in duplicate, and a third measurement was performed if the deviation between duplicate measurements exceeded 10% for measurements of 100 mg/dL or greater or 10 mg/dL for measurements of less than 100 mg/dL.
All measurements were performed with test strips from the same manufacturing lot, whose accuracy was validated in a study conducted with 32 subjects each before and toward the end of the CGM performance study. At both time points, 100% of measurements (n = 128) were within ±10 mg/dL or ±10% of a hexokinase-based reference measurement device (Cobas Integra 400 plus, Roche Diagnostics, Mannheim, Germany). Bias was −1.8% before and +0.8% at the end of the study, while imprecision in terms of coefficient of variation was 2.0% before and 1.6% at the end of the study. The accuracy of the test strip lot used was therefore sufficiently stable and complied with accuracy requirements for comparator devices proposed recently. 8
Testing Procedures
In-clinic visits for comparator data collection at the study sites were scheduled at the beginning, middle, and end of the study period. Subjects aged 2 to 6 years participated in 1 in-clinic session with frequent comparator measurements for 4 hours every 30 ± 5 minutes either on study days 1 to 2, 8 to 10, or 15 to 16. Older subjects participated in 2 in-clinic sessions with measurements every 15 ± 5 minutes on study days 1 to 2, 8 to 10, or 15 to 16, separated by at least 4 days. For subjects aged 7 to 13 years and 14 to 17 years, the measurement period had a duration of 4 and 6 hours, respectively. At the start of each session, subjects aged 2 to 13 years received a meal and were asked to manage their diabetes in the usual manner. For subjects aged 14 to 17 years, glucose levels were manipulated by adjusting insulin dosing and meal intake. Thus, a wide range of hyperglycemic (>250 mg/dL) and hypoglycemic (<70 mg/dL) comparator data were obtained, while still limiting the time during which glucose levels were less than 70 mg/dL or greater than 250 mg/dL. Serious or non-serious, anticipated or unanticipated adverse device effects ((S)ADEs) were recorded.
Data Analysis and Statistics
During the study, the first-generation device of CSAir, requiring manual calibrations, was used. For the analysis of the performance of the updated algorithm with optional calibrations presented in this article, the manufacturer retrospectively applied the algorithm without using any user-entered calibration values and provided the resulting CGM readings to the IfDT for analysis. These readings represent the real-time data displayed to the user by the updated algorithm in 5-minute intervals.
For analysis of point accuracy, comparator measurements were paired with CGM readings recorded simultaneously or up to 4 minutes after the comparator measurement, and CGM readings outside the measuring range (40-500 mg/dL) were not considered. This pairing methodology is similar to that used in other recent CGM performance studies,9,10 which allows for comparability between studies. Performance was analyzed separately for data from subjects aged 7 to 17 years (primary sensors only, all arm-worn sensors), as well as data from all sensors in subjects aged 2 to 6 years. Here, data from primary and secondary sensors were pooled to increase the number of available data points, and data from arm-worn and buttock-worn sensors were analyzed separately.
Analytical point accuracy was characterized by the %15/15, %20/20, and %40/40 ARs, defined as the percentage of CGM readings within ±15/20/40% (for comparator values ≥100 mg/dL) or within ±15/20/40 mg/dL (for comparator values <100 mg/dL) of the corresponding comparator measurements. In addition, mean absolute relative difference (MARD), relative bias (also known as mean relative difference), and mean paired absolute relative difference (PARD) between CGM readings of sensors worn by the same subject were calculated. Clinical point accuracy was determined by the Diabetes Technology Society (DTS) error grid analysis. 11 The stability of the CGM system was assessed by stratifying point accuracy metrics with respect to the sensor-wearing period. Alert reliability was evaluated for hypo- and hyperglycemia through true alert rates (percentage of CGM values <70 or >250 mg/dL concurrent with a corresponding comparator value below or above these thresholds within ±15 minutes) together with the true detection rate (percentage of comparator values <70 or >250 mg/dL that were concurrent/nonconcurrent with a CGM value below or above the same threshold within ±15 minutes). Time lag was estimated from primary sensors worn by subjects aged 7 to 17 years, using the method proposed by Kovatchev et al. 12 The technical reliability was characterized through a Kaplan-Meier sensor survival analysis including all sensors and examination of gaps in the CGM data.
For subjects aged 2 to 6 years, only aggregate results are shown due to the limited number of available data points.
Results
Study Population
A total of 79 children and adolescents with T1DM (2-6 years: N = 12, 7-13 years: N = 29, 14-17 years: N = 29) were enrolled in the study, compared with a pre-specified sample size of 83 subjects. In the 2- to 6-year age group, the target number could not be reached, as recruitment of young children is inherently challenging in clinical studies involving repeated capillary blood sampling. Furthermore, there was 1 dropout before sensor insertion in this age group. Baseline participant characteristics by age group (2-6 years and 7-17 years) are shown in Table 1. Body mass index (BMI) normalized with respect to age cohort (z-BMI) was calculated according to Rosario et al. 13
Table 1.
Subject Population Characteristics.
| Demographic | Age 2-6 years | Age 7-17 years |
|---|---|---|
| N | 21 | 58 |
| Sex, N (%) | ||
| Male | 9 (42.9%) | 33 (56.9%) |
| Female | 12 (57.1%) | 25 (43.1%) |
| Age (years) | ||
| Mean (SD) | 5.2 (1.0) | 13.1 (2.9) |
| Median (Min, Max) | 5.0 (3, 6) | 13.5 (7, 17) |
| z-BMI | ||
| Mean (SD) | 0.93 (0.72) | 0.46 (0.78) |
| Median (Min, Max) | 1.13 (−0.74, 2.31) | 0.46 (−1.09, 2.20) |
| HbA1c (%) | ||
| Mean (SD) | 7.1 (0.6) | 6.8 (0.7) |
| Median (Min, Max) | 7.1 (6.1, 8.7) | 6.9 (5.4, 8.3) |
| Diabetes duration (years) | ||
| Mean (SD) | 2.9 (1.1) | 5.6 (4.3) |
| Median (Min, Max) | 3.0 (1.0, 5.0) | 4.0 (0.3, 14.0) |
| Diabetes type, N (%) | ||
| Type 1 | 21 (100%) | 58 (100%) |
| Therapy regimen, N (%) | ||
| MDI | 1 (4.8%) | 3 (5.2%) |
| CSII | 20 (95.2%) | 55 (94.8%) |
| AID a | 20 (95.2%) | 53 (91.4%) |
Abbreviations: z-BMI, body mass index normalized with respect to age cohort; SD, standard deviation; MDI, multiple daily injections; CSII, continuous subcutaneous insulin infusion; AID, automated insulin delivery.
Included in CSII population.
CGM Performance
The analytical point accuracy by age groups and sensor-wearing sites is shown in Table 2. Analytical point accuracy was high in both age groups and both wearing sites in children aged 2 to 6 years. Buttock-worn sensors in subjects aged 2 to 6 years had the highest ARs and lowest MARDs and bias. Overall, there was a negative bias of CGM readings compared to capillary comparator measurements. Due to accumulating sensor failures, the number of data pairs was lower for the later in-clinic sessions.
Table 2.
Aggregate Analytical Point Accuracy of CSAir for Different Ages and Sensor-Wearing Locations.
| Age 2-6 (arm) | Age 2-6 (buttock) | Age 7-17 (arm) | |
|---|---|---|---|
| N (subjects) | 14 | 11 | 54 |
| N (sensors) | 17 | 14 | 54 |
| N (data points) | 149 | 123 | 2002 |
| %15/15 AR (%) | 81.9% | 93.5% | 84.1% |
| %20/20 AR (%) | 86.6% | 98.4% | 90.8% |
| %40/40 AR (%) | 98.7% | 100.0% | 98.9% |
| MARD (%) | 10.3% | 7.8% | 8.9% |
| Bias (%) | –6.9% | –2.7% | –5.4% |
Abbreviations: AR, agreement rate; MARD, mean absolute relative difference.
Table 3 displays point accuracy of sensors in subjects aged 7 to 17 years stratified by comparator blood glucose range. A total of 3.9% and 14.6% of comparator blood glucose values were less than 70 mg/dL and greater than 250 mg/dL, respectively. The %20/20 AR was high in all glucose ranges; MARD was highest during hypoglycemia and lowest during hyperglycemia.
Table 3.
Point Accuracy of CSAir in Subjects Aged 7 to 17 Years Stratified by Comparator Blood Glucose (BG) Range.
| BG range (mg/dL) | N | %15/15 AR | %20/20 AR | %40/40 AR | MARD | Bias |
|---|---|---|---|---|---|---|
| <54 | 7 | 100.0% | 100.0% | 100.0% | 10.0% | –0.6% |
| 54-69 | 71 | 93.0% | 98.6% | 100.0% | 10.8% | +0.8% |
| 70-180 | 1157 | 81.5% | 89.1% | 98.6% | 9.6% | –5.9% |
| 181-250 | 474 | 83.8% | 90.3% | 98.7% | 8.7% | –6.3% |
| >250 | 293 | 92.5% | 96.2% | 100.0% | 6.2% | –3.3% |
Abbreviations: AR, agreement rate; BG, blood glucose; MARD, mean absolute relative difference.
Table 4 displays point accuracy of sensors in subjects aged 7 to 17 years stratified by different CGM glucose rate of change (RoC). Accuracy was highest at low RoCs, with no substantial differences in accuracy observed across RoC brackets.
Table 4.
Point Accuracy of CSAir In Subjects Aged 7 to 17 Years Stratified by CGM Rate of Change (RoC).
| CGM RoC range (mg/dL/min) | N | %15/15 AR | %20/20 AR | %40/40 AR | MARD | Bias |
|---|---|---|---|---|---|---|
| <−2 | 139 | 84.2% | 89.9% | 100.0% | 9.5% | –6.3% |
| ≥−2 to <−1 | 334 | 84.1% | 91.0% | 98.8% | 9.3% | –5.4% |
| ≥−1 to ≤+1 | 1087 | 85.3% | 91.4% | 98.8% | 8.6% | –5.5% |
| >+1 to ≤+2 | 216 | 80.1% | 87.5% | 98.1% | 9.4% | –5.2% |
| >+2 | 210 | 83.3% | 91.4% | 100.0% | 8.8% | –3.3% |
Abbreviations: AR, agreement rate; CGM, continuous glucose monitoring; MARD, mean absolute relative difference.
Mean PARD between CGM readings of 2 sensors worn by the same subject at the same location was 8.0% in subjects aged 7 to 17 years (56 sensor pairs) and 6.4% (arm; 4 sensor pairs) and 5.8% (buttock; 4 sensor pairs) in subjects aged 2 to 6 years. These results for subjects aged 2 to 6 years should be interpreted with caution because of the low number of available sensor pairs.
For subjects aged 7 to 17 years, 99.9% of CGM-BG pairs were in Zones A and B of the DTS error grid. For subjects aged 2 to 6 years, 99.3% of CGM-BG pairs for arm-worn sensors and 100% of CGM-BG pairs for buttock-worn sensors were in Zones A and B of the DTS error grid. Value pairs found outside of Zones A and B were all located in the lower Zone C (Figure 1).
Figure 1.
Results of the Diabetes Technology Society Error Grid analysis of age group 2 to 6 years separated by sensor-wearing sites (panel A) and age group 7 to 17 years (panel B).
Abbreviations: CGM, continuous glucose monitoring; BG, blood glucose.
Table 5 displays accuracy metrics in subjects aged 7 to 17 years stratified by sensor-wearing periods. Sensor accuracy peaked during the middle wearing phase and was lowest during the early phase, with the late phase showing intermediate accuracy.
Table 5.
Stability of CSAir in Subjects Aged 7 to 17 Years.
| Sensor-wearing period (days) | N | %15/15 AR | %20/20 AR | %40/40 AR | MARD | Bias |
|---|---|---|---|---|---|---|
| Early 1-2 (day 1) | 654 (421) | 71.3% (62.2%) | 80.6% (72.9%) | 97.6% (96.2%) |
12.5% (14.7%) | –9.9% (−13.0%) |
| Middle (7-9) | 812 | 94.2% | 97.8% | 99.8% | 6.3% | –2.4% |
| Late (14-15) | 536 | 84.5% | 92.7% | 99.3% | 8.7% | –4.2% |
Abbreviations: AR, agreement rate; MARD, mean absolute relative difference.
Table 6 shows true alert rates, that is, the percentage of CGM alerts that coincided (±15 minutes) with comparator measurements below/above the same threshold, and the true detection rates, that is, the percentage of comparator measurements below/above the alert threshold that coincided (±15 minutes) with a CGM alert. The true alert rate for hypoglycemia was substantially lower than for hyperglycemia; true detection rates were similar for hypo- and hyperglycemia and similar to true hyperglycemic alert rates.
Table 6.
Blood Glucose Alert Reliability for Hypo- (<70 mg/dL) and Hyperglycemia (>250 mg/dL) in Subjects Aged 7 to 17 Years.
| N (alerts) | True alert rate | N (detections) | True detection rate | |
|---|---|---|---|---|
| <70 mg/dL | 376 | 68.9% | 90 | 90.0% |
| >250 mg/dL | 792 | 96.1% | 297 | 89.2% |
The estimated time lag based on sensors worn by subjects aged 7 to 17 years was 4.1 ± 2.5 minutes (range, 0-11 minutes). The availability of CGM data was 99.5% or greater in all groups, with most gaps in data having a duration of 15 minutes or less. Gaps with a duration of more than 60 minutes were extremely rare and transient, with normal data collection resuming without recurrence under similar conditions. Here, no consistent patterns or common failure modes were identified, and there was no evidence indicating a systematic malfunction of the devices.
Survival analysis of the 155 inserted sensors (primary and secondary sensors combined; 3 sensors were replaced) separated by age group and wearing location is shown in Figure 2. Overall, there were 45 sensor failures, of which 25 were caused by a technical malfunction; that is, the updated algorithm automatically terminated the sensor because its signal did not meet predefined operational or performance criteria. The remaining 20 failures were attributed to adhesive weakness. In addition, 1 sensor was inadvertently removed by the subject despite functioning normally and therefore censored from the survival analysis. The mean survival times and probability at the end of the study were 11.7 days and 45.0% for arm sensors in subjects aged 2 to 6 years, 13.4 days and 71.8% for buttock sensors in subjects aged 2 to 6 years, and 13.1 days and 75.2% for arm sensors in subjects aged 7 to 17 years.
Figure 2.
Kaplan-Meier survival curves by age group and wearing site. The vertical lines indicate censoring due to user error or scheduled sensor removals.
Safety
There were no serious adverse events in the study, and 20 adverse device effects (ADEs) that affected 15 (19%) subjects. A total of 18 of these ADEs were mild and involved bleeding and/or pain at the sensor application site (N = 2), erythema at the sensor application site caused by the sensor adhesive or overpatch (N = 2), erythema at the sensor insertion site (N = 3), itching at the sensor application site (N = 5), pressure marks (N = 3), and skin irritations at the sensor application site (N = 3). The remaining 2 ADEs were classified as moderate and described as suspected irritant contact dermatitis. In 1 case, xerosis persisted for several weeks; in the other, cutaneous alterations extended to skin areas not covered by the sensor adhesive or overpatch.
Discussion
The performance of the CSAir CGM system with the updated calibration algorithm in children and adolescents aged 2 to 17 years was evaluated for the first time and showed performance similar to that in adults. Specifically, the %20/20 AR of 90.8% in subjects aged 7 to 17 years is only slightly lower and thus comparable to the reported performance with the updated algorithm in adults (%20/20 AR, 93.9%). 7 Point accuracy was stable across the sensor wear time, with moderately lower accuracy in the first 2 days of sensor wear. This was also observed in adults and is common in CGM systems. 14 CSAir had a slight negative bias, consistent with a high frequency of hypoglycemia alerts, including a reduced rate of true alerts in hypoglycemia. However, the true hypoglycemia detection rate was 90%, suggesting that CSAir reliably detects hypoglycemic events, although the lower true alert rate may increase the risk of alarm fatigue. 15
Compared with the performance of other current-generation CGM systems (FreeStyle Libre 3+ [FSL3+] and Dexcom G7 10-day [DG7]) approved for use in children, the performance of CSAir observed in this study was similar. In children aged 2 to 6 years, MARD was 10.3% and 7.8% for arm- and buttock-worn CSAir, respectively, compared with 11.2% for FSL3+ 14 and 9.3% for DG7 16 in the corresponding age group. For comparison purposes, it has to be considered that FSL3+ is only approved to be worn on the arm, and the results for DG7 refer to a mix of insertion sites (arm, abdomen, and buttock), which were not reported separately. In arm-worn sensors of older cohorts, MARD for CSAir was 8.9%, which is only slightly higher compared with 8.1% for FSL3+ 14 and DG7. 16 This comparison is intended solely to provide descriptive context within the broader CGM literature. It must be noted that this comparison is limited by the fact that, in the studies with FSL3+ and DG7, venous blood samples measured with YSI 2300 STAT PLUS were used as comparator samples in subjects aged 7 and older, whereas the present study collected capillary comparator measurements in all subjects. For the present study, capillary comparator measurements were chosen for several reasons: first, generally, it is the reference for people with diabetes in daily life; second, it does not require an invasive venous catheter to be used; and third, the amount of blood withdrawn is lower in the particularly vulnerable pediatric population.
Due to the vulnerability of subjects aged 2 to 6 years, the number of sensors and paired data points was considerably lower compared with the older age cohort. This is common in CGM performance studies with minors and means that the derived results are more susceptible to the effects of outliers. With %20/20 ARs of 86.6% and 98.4% for sensors worn on the arm and buttock, respectively, arm-worn sensors showed slightly lower results compared with the older age cohort, while results for buttock-worn sensors were considerably higher. This observed site-related accuracy difference cannot be fully explained on the basis of available evidence and likely reflects individual sensor variability amplified by the small sample size rather than a systematic site effect. Published comparative studies in pediatric populations have not demonstrated consistent superior accuracy for buttock over arm placement. 17 Notably, both wearing sites demonstrated clinically acceptable accuracy.
The 15-day sensor survival rates were slightly decreased compared with those in adults, but similar to those of other CGM systems in a pediatric population.18,19 This was largely attributable to sensor detachment, which occurred more frequently than in adults. The reduced sensor wear stability and higher detachment rate are consistent with the behavioral characteristics of very young children, including frequent movement, physical play, bathing, and unintentional device contact, as well as limited body surface area for site selection.20,21
The safety evaluation of this study revealed no serious ADEs. However, in the present study, 19% of study participants experienced ADEs, which is more than the 6% reported in the pivotal study in adults. 6 This appeared to be caused by an increase in skin-related issues such as itching, pressure marks, or erythema, most of which were classified as mild. The increased incidence of cutaneous ADEs observed in children and adolescents is likely attributable to known age-related differences in pediatric skin morphology and physiology,22-28 which predispose children and adolescents to heightened susceptibility to cutaneous irritation. Still, the safety profile observed in this study does not appear to indicate any new or unexpected safety concerns regarding the use of the device in the pediatric population.
Conclusions
The performance and safety of CSAir in children and adolescents were sufficiently similar to those observed in adults, supporting the suitability of the system for use in children aged 2 years and older and providing evidence for regulatory approval in this population. In children aged 6 years or younger, both the arm and upper buttock were evaluated as wearing locations, with better overall performance observed for buttock placement.
Acknowledgments
The authors would like to thank i-SENS, the manufacturer of CareSens Air, for sponsoring the study and providing the investigational CGM systems and technical product information.
Footnotes
Abbreviations: ADE, adverse device effect; AID, automated insulin delivery; AR, agreement rate; BG, blood glucose; BMI, body mass index; CGM, continuous glucose monitoring; CSAir, CareSens Air; CSII, continuous subcutaneous insulin infusion; DTS, Diabetes Technology Society; MARD, mean absolute relative difference; MDI, multiple daily injections; PARD, paired absolute relative difference; RoC, rate of change; SD, standard deviation.
ORCID iDs: Torben Biester 
https://orcid.org/0000-0001-8051-5562
Jantje Weiskorn
https://orcid.org/0009-0006-8663-3723
Manuel Eichenlaub
https://orcid.org/0000-0003-2150-3160
Anne Beltzer
https://orcid.org/0000-0002-8308-8540
Guido Freckmann
https://orcid.org/0000-0002-0406-9529
Funding: The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study described in this article was sponsored by i-SENS, Inc. MW received funding from the Federal Ministry of Research, Technology and Space (Bundesministerium für Forschung, Technologie und Raumfahrt, BMFTR) as part of the German Center for Child and Adolescent Health (DZKJ) under the funding code 01GL2407A.
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: TBI reports study support from Dexcom and Ypsomed, lecture fees from Abbott, Dexcom, Insulet, Lilly, Medtronic, NovoNordisk, Sanofi, Synlab, Tandem, Vitalaire, and Ypsomed, and advisory board participation with Dexcom, Insulet, Medtronic, Sanofi, Tandem, and Ypsomed. GF/IfDT has received grants, honoraria, consulting or speaking fees during the past 3 years from Abbott, Ascensia, Bionime, Boydsense, Dexcom, Glucoset, Insulet, i-SENS, Lilly, Menarini, Perfood, Pharmasens, Roche, Sinocare, Vertex, and Ypsomed. JW received lecture fees from Sanofi, Ypsomed, and Chiesi and advisory board participation from Chiesi and Sanofi. ML, ME, and AB are employees of IfDT.
References
- 1. Maiorino MI, Signoriello S, Maio A, et al. Effects of continuous glucose monitoring on metrics of glycemic control in diabetes: a systematic review with meta-analysis of randomized controlled trials. Diabetes Care. 2020;43(5):1146-1156. [DOI] [PubMed] [Google Scholar]
- 2. Laffel LM, Kanapka LG, Beck RW, et al. Effect of continuous glucose monitoring on glycemic control in adolescents and young adults with type 1 diabetes: a randomized clinical trial. JAMA. 2020;323(23):2388-2396. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Strategies to Enhance New CGM Use in Early Childhood (SENCE) Study Group. A randomized clinical trial assessing continuous glucose monitoring (CGM) use with standardized education with or without a family behavioral intervention compared with fingerstick blood glucose monitoring in very young children with type 1 diabetes. Diabetes Care. 2021;44(2):464-472. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Thabit H, Prabhu JN, Mubita W, et al. Use of factory-calibrated real-time continuous glucose monitoring improves time in target and HbA(1c) in a multiethnic cohort of adolescents and young adults with type 1 diabetes: the MILLENNIALS study. Diabetes Care. 2020;43(10):2537-2543. [DOI] [PubMed] [Google Scholar]
- 5. Garg SK, Akturk HK. A new era in continuous glucose monitoring: food and drug administration creates a new category of factory-calibrated nonadjunctive, interoperable class II medical devices. Diabetes Technol Ther. 2018;20(6):391-394. [DOI] [PubMed] [Google Scholar]
- 6. Kim KS, Lee SH, Yoo WS, Park CY. Accuracy and safety of the 15-day CareSens Air continuous glucose monitoring system. Diabetes Technol Ther. 2024;26(4):222-228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Jendrike N, Eichenlaub M, Link M, et al. Comparative performance analysis of manual and updated optional calibration algorithms for the CareSens Air CGM system. J Diabetes Sci Technol. Published online July 3, 2025:19322968251351318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Pleus S, Eichenlaub M, Dabla PK, et al. Clinical assessment and acceptance criteria for continuous glucose monitoring (CGM) system performance: a proposed guideline by the IFCC working group on CGM. Clin Chim Acta. 2026;580:120728. [DOI] [PubMed] [Google Scholar]
- 9. Garg SK, Kipnes M, Castorino K, et al. Accuracy and safety of Dexcom G7 continuous glucose monitoring in adults with diabetes. Diabetes Technol Ther. 2022;24(6):373-380. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Garg SK, Bailey TS, Castorino K, et al. Accuracy of the 15.5-day G7 iCGM in adults with diabetes. Diabetes Technol Ther. 2025;27(6):430-438. [DOI] [PubMed] [Google Scholar]
- 11. Klonoff DC, Freckmann G, Pleus S, et al. The diabetes technology society error grid and trend accuracy matrix for glucose monitors. J Diabetes Sci Technol. 2024;18(6):1346-1361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Kovatchev BP, Shields D, Breton M. Graphical and numerical evaluation of continuous glucose sensing time lag. Diabetes Technol Ther. 2009;11(3):139-143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Rosario AS, Kurth BM, Stolzenberg H, Ellert U, Neuhauser H. Body mass index percentiles for children and adolescents in Germany based on a nationally representative sample (KiGGS 2003-2006). Eur J Clin Nutr. 2010;64(4):341-349. [DOI] [PubMed] [Google Scholar]
- 14. Alva S, Bhargava A, Bode B, et al. Accuracy of a 15-day factory-calibrated continuous glucose monitoring system with improved sensor design. J Diabetes Sci Technol [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Giza S, Kotanidou EP, Tsinopoulou VR, Poukoulidou T, Ntouma S, Galli-Tsinopoulou A. Can glucose alarm fatigue threaten the absolute clinical benefit of continuous glucose monitoring in optimal glucose management in children and adolescents with type 1 diabetes? a narrative review. Children. 2025;12(12):1668. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Laffel LM, Bailey TS, Christiansen MP, Reid JL, Beck SE. Accuracy of a seventh-generation continuous glucose monitoring system in children and adolescents with type 1 diabetes. J Diabetes Sci Technol. 2023;17(4):962-967. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Faccioli S, Del Favero S, Visentin R, et al. Accuracy of a CGM Sensor in pediatric subjects with type 1 diabetes. J Diabetes Sci Technol. 2017;11(6):1147-1154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Food and Drug Administration. 510(k) Substantial equivalence determination decision summary, K212132, FreeStyle Libre 3 continuous glucose monitoring system 2022. Accessed October 31, 2022. https://www.accessdata.fda.gov/cdrh_docs/reviews/K212132.pdf
- 19. Food and Drug Administration. 510(k) Substantial equivalence determination decision summary, K213919, Dexcom G7 continuous glucose monitoring (CGM) system 2022. Accessed November 20, 2024. https://www.accessdata.fda.gov/cdrh_docs/reviews/K213919.pdf
- 20. Hilliard ME, Levy W, Anderson BJ, et al. Benefits and barriers of continuous glucose monitoring in young children with type 1 diabetes. Diabetes Technol Ther. 2019;21(9):493-498. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Englert K, Ruedy K, Coffey J, et al. Skin and adhesive issues with continuous glucose monitors: a sticky situation. J Diabetes Sci Technol. 2014;8(4):745-751. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Stamatas GN, Nikolovski J, Luedtke MA, Kollias N, Wiegand BC. Infant skin microstructure assessed in vivo differs from adult skin in organization and at the cellular level. Pediatr Dermatol. 2010;27(2):125-131. [DOI] [PubMed] [Google Scholar]
- 23. Mack MC, Chu MR, Tierney NK, Ruvolo E, Jr, Stamatas GN, Kollias N, et al. Water-holding and transport properties of skin stratum corneum of infants and toddlers are different from those of adults: studies in three geographical regions and four ethnic groups. Pediatr Dermatol. 2016;33(3):275-282. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Nikolovski J, Stamatas GN, Kollias N, Wiegand BC. Barrier function and water-holding and transport properties of infant stratum corneum are different from adult and continue to develop through the first year of life. J Invest Dermatol. 2008;128(7):1728-1736. [DOI] [PubMed] [Google Scholar]
- 25. Walters RM, Khanna P, Chu M, Mack MC. Developmental changes in skin barrier and structure during the first 5 years of life. Skin Pharmacol Physiol. 2016;29(3):111-118. [DOI] [PubMed] [Google Scholar]
- 26. Stamatas GN, Morello AP, Mays DA. Early inflammatory processes in the skin. Curr Mol Med. 2013;13(8):1250-1269. [DOI] [PubMed] [Google Scholar]
- 27. Berg AK, Passanisi S, von dem Berge T, et al. SKIN-PEDIC: a worldwide assessment of skin problems in children and adolescents using diabetes devices. Horm Res Paediatr. 2026:99(5):666-679. [DOI] [PubMed] [Google Scholar]
- 28. Burgmann J, Biester T, Grothaus J, Kordonouri O, Ott H. Pediatric diabetes and skin disease (PeDiSkin): a cross-sectional study in 369 children, adolescents and young adults with type 1 diabetes. Pediatr Diabetes. 2020;21(8):1556-1565. [DOI] [PubMed] [Google Scholar]