Performance of a New Continuous Glucose Monitoring System in German Adults Living with Diabetes

Abstract

Introduction

The True Vie I3 continuous glucose monitoring system (i3 CGM, Sinocare Meditech Inc., also approved in Europe as GlucoMen® iCan or iCan CGM system) is a new real-time continuous glucose monitoring system (CGM) intended for the management of diabetes mellitus. This pivotal study evaluated the performance of the factory-calibrated CGM system.

Methods

In this center-specific dataset, 35 adults with type 1 diabetes (T1D) and type 2 diabetes (T2D) wore sensors on the abdomen and arm for 15 days. Four in-clinic visits were scheduled, during which frequent comparator sampling of venous blood was performed every 5–15 min for up to 10 h, and a glucose manipulation was performed. CGM performance compared to Yellow Springs Instrument 2300 Stat Plus Glucose and Lactate Analyzer (Yellow Springs, OH) glucose analyzer was evaluated for abdomen and arm sensors separately, regarding mean absolute relative difference (MARD) and agreement rates (AR) stratified by glucose range and rate of change (RoC). Additionally, clinical accuracy, sensor attachment rate, pain, and safety were assessed. This single-center analysis was developed with the intention to provide European—and particularly German—data. The presented site was the highest-enrolling center in the study and, as such, can be considered representative of the overall study population—an assumption that the analysis confirmed.

Results

20/20 AR and MARD were 95.5% and 9.4% for abdomen sensors, and 95.3% and 9.8% for arm sensors, respectively. Consensus error grid (CEG) analyses revealed that 100% of CGM–comparator pairs fell in zones A and B for abdomen and arm sensors. Accuracy of sensors remained stable throughout the wearing time. Adhesion rate was 100% for abdomen sensors and 97.1% for arm-worn sensors, without the use of any over-tape during the 15-day study period. Pain during insertion and removal was reported as minimal, and no unexpected safety issues were identified.

Conclusions

Data from a single study center showed that the performance of the i3 CGM is comparable to that published for other established CGM devices, and accuracy results were within limits specified for integrated continuous glucose monitoring systems (iCGM). The i3 CGM showed reliability, and its safety was validated during the 15 study days.

Trial Registration

The study was registered under ClinicalTrials.gov (ID: NCT05806554).

Key Summary Points

Why carry out this study?

Continuous glucose monitoring (CGM) systems have become the standard of care in the therapy of people with insulin-treated diabetes mellitus.

This study evaluated the performance of a new factory-calibrated CGM system in adults with diabetes during a 15-day study period in a mixed inpatient and outpatient setting.

What was learned from this study?

Performance was characterized by 20/20 agreement rates (AR) and mean absolute relative difference (MARD) of 95.5% and 9.4% for abdomen sensors, and 95.3% and 9.8% for arm sensors, respectively. Data confirm clinical performance data previously reported, with 100% of readings in consensus error grid (CEG) zones A + B and a high wear time in the abdominal region. Performance results were consistent and complied with US Food and Drug Administration integrated continuous glucose monitoring systems (iCGM) special controls; additionally, no unexpected safety issues were reported.

Performance of the investigational device suggests the device to be comparable to that published for established CGM systems.

Introduction

Continuous glucose monitoring (CGM) systems have become the standard of care in the therapy of people with insulin-treated diabetes mellitus. In the ever-changing market for CGM systems, constant advancements are being introduced, and new devices are seeking to establish themselves. Despite the lack of internationally binding requirements, guidelines such as the POCT05 from the Clinical and Laboratory Standards Institute [1] and the more rigorous approval criteria for the integrated continuous glucose monitoring systems (iCGM, only in the USA) [2] are applied to ensure that CGM systems are safe and reliable. Therefore, reaching the best possible performance has become a primary focus for CGM manufacturers to ensure a high-quality diabetes therapy.

The True Vie I3 continuous glucose monitoring system (i3 CGM, Sinocare Meditech Inc., also approved in Europe as GlucoMen^® iCan or iCan CGM system) is a new real-time CGM system for the management of diabetes mellitus in individuals from the age of 2 years. It is factory-calibrated and intended to replace the need for capillary blood confirmation to make therapy decisions in the western markets. The sensor has a warm-up period of 2 h, after which it provides glucose values every 3 min over a measuring range from 36 to 450 mg/dL. Its dimensions are 32 mm in diameter and 5.7 mm in thickness, and it has a total wear time of 15 days. This single-center data from a pivotal multicenter study evaluated the performance of the i3 CGM system.

Methods

Study Design

This performance evaluation was conducted as a randomized, multicenter study in adults with type 1 diabetes (T1D) or type 2 diabetes (T2D) who were under intensified insulin therapy. Six study centers were involved in the performance evaluation. The data presented in this publication includes the participants from one study site in Ulm (Germany), which conducted the study between February and May 2024 and was the center with the highest number of enrolled subjects. Main exclusion criteria for participants were an HbA1c > 9%, having a hematocrit below 10% under the limit of the normal range, inadequate intravenous access on the arms, having experienced a severe hypoglycemic event requiring third-party assistance or an episode of diabetes ketoacidosis within the last 6 months at the time of screening, having conditions that predispose to hypoglycemia, and having a history of skin adhesive tolerance issues. Sample size calculation based on the fulfillment of the iCGM special controls yielded a required number of 140 participants for the full study. The study was registered under ClinicalTrials.gov (ID: NCT05806554). Ethical approval for this study was obtained by the Ethics Committee of the State Medical Association of Baden-Württemberg (Ethik-Kommission der Landesärztekammer Baden-Württemberg, Stuttgart, Germany). The study was performed in accordance to the declaration of Helsinki. All study participants provided written informed consent.

This analysis was developed with the intention to provide European—and particularly German—data. Since the complete dataset from the pivotal study was under regulatory review and therefore not available for publication, we decided to conduct a center-specific analysis in accordance with the statistical analysis plan used for the full study. The presented site represented the highest-enrolling center in the study and, as such, can be considered representative of the overall study population—an assumption that our analysis confirmed.

Study Procedures

The study duration was 15 study days, in line with the sensor lifetime, during which CGM values were blinded, and participants spent time in both free-living and in-clinic settings. Participants signed informed consent forms at the beginning of the screening visit. After being included, sensor insertion was scheduled on the evening before the first in-clinic session or on the same day. Participants were randomized on the basis of two strata (T1D or T2D) to either wear two sensors on the abdomen and one on the arm, or one sensor on the abdomen and two on the arms. Participants inserted a total of three sensors themselves, with the first sensor at each insertion site serving as the primary sensor.

Four in-clinic visits were scheduled for each participant at the beginning, middle (two visits), and end of the sensor lifetime. During in-clinic visits, frequent comparator sampling was performed every 15 ± 3 min for up to 10 h through an indwelling venous catheter, while keeping participants’ arms warm with heating pads. For cases where blood glucose (BG) was < 70 mg/dL or > 300 mg/dL, venous blood was sampled every 5 ± 2 min. Comparator values were generated by measuring centrifuged plasma samples with the Yellow Springs Instrument 2300 Stat Plus Glucose and Lactate Analyzer (Yellow Springs, OH). Moreover, blood ketones were monitored before and after performing the frequent sampling measurements. Glucose manipulation was performed under medical supervision during in-clinic visits for each participant, depending on their fasting BG level at the beginning of the session. If BG levels were below 180 mg/dL, they were manipulated to < 70 mg/dL for 60 ± 30 min by delivering an insulin dose based on the individual correction factor, with time spent < 54 mg/dL being limited to 15 ± 10 min. If BG levels were ≥ 180 mg/dL, they were manipulated to > 300 mg/dL for 90 ± 30 min, with a standardized meal. At the end of each in-clinic session, participants left the study site once their BG concentration was between 100 and 300 mg/dL and blood ketones remained < 0.6 mmol/L. During free-living settings of the study, participants were instructed to continue with their everyday routines and follow their usual diabetes therapies. Sensor removal occurred on day 15 at the study site, after the sensor expired.

Data Analysis

CGM readings were provided by the manufacturer. Performance was separately evaluated for the two application sites, abdomen and arm, for each of which T1D and T2D data were pooled. Comparator measurements and CGM data outside the measuring range of 36–450 mg/dL were excluded. For each CGM reading, a corresponding CGM rate of change (RoC) was calculated as the linear, least-squares regression slope of CGM readings within the previous 15 min [1]. If fewer than six readings were available, no CGM RoC was calculated. Comparator RoCs were calculated analogously, by considering measurements within the previous 18 min (maximum sampling interval) and calculating no RoC if fewer than two measurements were available.

Performance analysis was based on CGM data from the primary sensors. Comparator measurements were paired with CGM values recorded simultaneously or up to 2 min afterwards. Comparator and CGM RoCs were paired analogously. Agreement rates (AR), mean absolute relative difference (MARD), and mean relative difference (also known as bias) were calculated and stratified by BG range and CGM RoC to assess analytical point accuracy. Special iCGM controls were applied, the lower one-sided 95% confidence bound [3] was calculated for the shares of CGM measurements within ± 15 mg/dL of the corresponding BG value for CGM measurements < 70 mg/dL (A) or within 15% for CGM measurements at 70–180 mg/dL (B) and > 180 mg/dL (C), as well as for the shares of CGM measurements within ± 40 mg/dL of the corresponding BG value for CGM measurements < 70 mg/dL (D) or within 40% for measurements at 70–180 mg/dL I and > 180 mg/dL (F), and, lastly, for CGM measurements throughout the measuring range within ± 20% of the corresponding BG value (G). Special CGM safety controls regarding CGM values < 70 mg/dL and corresponding BG values > 180 mg/dL (H) and vice versa (I) were also included. Controls regarding trend accuracy requiring that, when the comparator RoC is < − 2 mg/dL/min, ≤ 1% of corresponding CGM RoCs shall be > + 1 mg/dL/min (J) and vice versa (K), were also analyzed [2]. Moreover, the expected distribution of deviations between the CGM and comparator data along different glucose ranges was studied with the continuous glucose deviation interval and variability analysis (CG-DIVA) analysis [4]. Sensor stability was characterized by stratifying MARD for different sensor-wearing periods (beginning, early middle, late middle, end), and clinical accuracy was assessed using the Consensus Error Grid (CEG). In addition, sensor adhesion rate was described, as well as the sensor functioning rate, i.e., provision of readings, along the 15 study days. Safety was surveyed through adverse event (AE) documentation, and pain was assessed after insertion and removal of the sensors through the visual analogue scale (VAS), where respondents marked their pain on a 100-mm line, where a score of 0 means no pain and a score of 100 means the worst imaginable pain [5].

Results

Study Population

A total of 41 potential participants were screened, of whom five fulfilled at least one exclusion criterion. Of the 36 included participants, one dropped out before the sensor insertion for reasons unrelated to the study; therefore, 35 participants were included in this analysis. Detailed participant demographics are summarized in Table 1.

Table 1.

Participant characteristics (N = 35)

Category	Value
Age [years]
Mean (± SD)	54.9 (12.7)
Min, median, max	27.0, 58.0, 71.0
Sex, N (%)
Female	12 (34.3%)
Male	23 (65.7%)
Diagnosis, N (%)
T1D	31 (88.6%)
T2D	4 (11.4%)
Ethnicity, N (%)
White	35 (100.0%)
BMI [kg/m2]
Mean (± SD)	26.8 (4.6)
Min, median, max	20.1, 25.0, 44.5
HbA1c [%]
Mean (± SD)	6.9 (0.7)
Min, median, max	5.6, 7.0, 8.4
Duration of diabetes [years]
Mean (± SD)	27.1 (14.0)
Min, median, max	1.0, 27.0, 52.0
Therapy regimen, n (%)
MDI	14 (40.0%)
CSII	21 (60.0%)
AID (part of CSII)	12 (34.3%)

T1D type 1 diabetes, T2D type 2 diabetes, SD standard deviation, BMI body mass index, MDI multiple daily injections, CSII continuous subcutaneous insulin infusion, AID automatic insulin delivery

Accuracy

A total of 107 sensors were used, two of which were replaced prior to insertion because of mechanical or handling issues. A total of 5459 and 4872 CGM–comparator pairs from the primary sensors inserted in the abdomen and in the arm (35 sensors each) were available from the in-clinic visits, respectively. 15/15, 20/20, and 40/40 AR were 87.9%, 95.5%, and 99.8% for sensors inserted on the abdomen, and 86.7%, 95.3%, and 99.9% for sensors inserted on the arm. MARDs were 9.4% and 9.8% for abdomen sensors and arm sensors, respectively. Point accuracy results, stratified by comparator BG measurements, are shown in Tables 2 and 3.

Table 2.

Point accuracy stratified by glucose range for abdomen sensors

Comparator range [mg/dL]	N	% 15/15a	% 20/20a	% 40/40a	MARD	Bias
Total	5459 (100%)	87.9%	95.5%	99.8%	9.4%	− 0.6%
< 54	205 (3.8%)	84.4%	96.1%	100.0%	19.5%	+ 18.4%
54 to < 70	860 (15.8%)	95.1%	99.4%	100.0%	12.2%	+ 8.7%
70 to 180	2656 (48.7%)	83.9%	92.7%	99.6%	8.7%	− 2.4%
> 180 to 250	551 (10.1%)	83.5%	95.3%	99.6%	9.4%	− 6.9%
> 250	1187 (21.7%)	94.0%	99.1%	100.0%	7.4%	− 3.8%

MARD mean absolute relative difference.

^aFor comparator values < 70 mg/dL, absolute differences between comparator and CGM values were calculated; for comparator values ≥ 70 mg/dL, relative differences were calculated

Table 3.

Point accuracy stratified by glucose range for arm sensors

Comparator range [mg/dL]	N	% 15/15a	% 20/20a	% 40/40a	MARD	Bias
Total	4872 (100%)	86.7%	95.3%	99.9%	9.8%	− 0.3%
< 54	183 (3.8%)	90.2%	97.8%	100.0%	17.8%	+ 15.5%
54 to < 70	762 (15.6%)	88.7%	97.5%	100.0%	13.7%	+ 12.1%
70 to 180	2392 (49.1%)	84.4%	93.5%	99.7%	8.7%	− 1.2%
> 180 to 250	481 (9.9%)	84.4%	93.8%	100.0%	9.4%	− 7.7%
> 250	1054 (21.6%)	91.1%	97.9%	100.0%	8.1%	− 6.7%

MARD mean absolute relative difference

^aFor comparator values < 70 mg/dL, absolute differences between comparator and CGM values were calculated; for comparator values ≥ 70 mg/dL, relative differences were calculated

Accuracy results stratified by the rates of change are shown in Tables 4 and 5 for sensors inserted on the abdomen and arm, respectively. The mean absolute RoC of comparator BG values, when paired with a CGM reading from either arm or abdomen sensors, was 0.78 mg/dL/min.

Table 4.

Point accuracy stratified by rate of change for abdomen sensors

CGM RoC range [mg/dL/min]	N	% 15/15a	% 20/20a	% 40/40a	MARD	Bias
< − 3	54 (1.0%)	83.3%	92.6%	96.3%	10.1%	− 6.7%
− 3 to < − 2	153 (2.8%)	83.0%	90.8%	98.0%	9.7%	− 5.4%
− 2 to < − 1	469 (8.6%)	86.4%	94.5%	100.0%	9.0%	− 4.9%
− 1 to 1	4053 (74.4%)	89.1%	96.3%	99.9%	9.6%	+ 0.4%
> 1 to 2	360 (6.6%)	81.1%	91.9%	99.7%	9.4%	− 3.6%
> 2 to 3	191 (3.5%)	83.2%	94.2%	100.0%	8.6%	− 3.2%
> 3	171 (3.1%)	87.1%	94.2%	100.0%	7.8%	+ 0.8%

CGM continuous glucose monitoring, RoC rate of change, MARD mean absolute relative difference

^aFor Yellow Springs Instrument 2300 (YSI) values < 70 mg/dL, absolute differences between YSI and CGM values were calculated; for YSI values ≥ 70 mg/dL, relative differences were calculated

Table 5.

Point accuracy stratified by rate of change for arm sensors

CGM RoC range [mg/dL/min]	N	% 15/15a	% 20/20a	% 40/40a	MARD	Bias
< − 3	37 (0.8%)	75.7%	89.2%	100.0%	9.2%	− 5.6%
− 3 to < −2	109 (2.2%)	80.7%	93.6%	97.2%	10.4%	− 8.3%
− 2 to < −1	384 (7.9%)	92.4%	96.9%	100.0%	8.5%	− 4.3%
− 1 to 1	3691 (75.8%)	87.4%	95.9%	99.9%	9.9%	+ 1.3%
> 1 to 2	357 (7.3%)	78.2%	88.5%	100.0%	10.3%	− 5.9%
> 2 to 3	166 (3.4%)	84.3%	95.8%	100.0%	8.5%	− 4.2%
> 3	125 (2.6%)	84.0%	94.4%	100.0%	8.7%	− 4.5%

CGM continuous glucose monitoring, RoC rate of change, MARD mean absolute relative difference

^aFor Yellow Springs Instrument 2300 (YSI) values < 70 mg/dL, absolute differences between YSI and CGM values were calculated; for YSI values ≥ 70 mg/dL, relative differences were calculated

Using the analysis methodology described in this article, the i3 CGM passed all iCGM special controls regarding accuracy and safety (Table 6). The expected range of deviations of the sensors stratified according the comparator BG range for the abdomen and arm sensors are depicted in Fig. 1. For both insertion sites, deviation across the comparator BG concentrations remained within acceptable thresholds. Sensor stability over the wearing time was characterized in terms of the MARD during the four in-clinic visits (Fig. 2). A slight increase in accuracy was observed for both application sites during wear time.

Table 6.

iCGM special controls for abdomen and arm sensors

Requirement	Abdomen			Arm
	N	Point estimate	CI	N	Point estimate	CI
(A) CGM < 70 mg/dL within ± 15 mg/dL Goal: CI > 85%	802	91.9%	87.7%	613	92.8%	89.6%
(B) CGM 70–180 mg/dL within ± 15% Goal: CI > 70%	3060	79.1%	76.3%	2867	77.7%	74.9%
(C) CGM > 180 mg/dL within ± 15% Goal: > 80%	1597	93.6%	91.3%	1392	91.1%	87.1%
(D) CGM < 70 mg/dL within ± 40 mg/dL Goal: CI > 98%	802	99.3%	98.3%	613	99.5%	98.8%
(E) CGM 70–180 mg/dL within ± 40% Goal: CI > 99%	3060	99.5%	99.0%a	2867	99.8%	99.7%
(F) CGM > 180 mg/dL within ± 40% Goal: CI > 99%	1597	100.0%	99.8%	1392	100.0%	99.8%
(G) Overall within ± 20% Goal: CI > 87%	5459	91.5%	90.0%	4872	90.6%	89.0%
(H) CGM < 70 mg/dL Goal: No Comp > 180 mg/dL	802	0	–	613	0	–
(I) CGM > 180 mg/dL Goal: No Comp < 70 mg/dL	1597	0	–	1392	0	–
(J) Comp RoC < −2 mg/dL/min Goal: ≤ 1% of CGM RoC > + 1 mg/dL/min	141	0.0%	–	128	0.8%	–
(K) Comp RoC > + 2 mg/dL/min Goal: ≤ 1% of CGM RoC < −1 mg/dL/min	329	0.0%	–	300	0.0%	–

CI lower, one-sided 95% confidence interval, CGM continuous glucose monitoring reading, RoC rate of change, Comp comparator blood glucose measurement

^aCI 99.03% (> 99%)

Fig. 1.

Results of the continuous glucose deviation interval and variability analysis (CG-DIVA) for the sensors inserted on the abdomen and arm

Fig. 2.

Sensor stability according to MARD at different sensor wearing periods

CEG analysis for sensors on both application sites showed that 100% of CGM–comparator pairs fell within zones A and B, indicating that the sensor performance was clinically acceptable (Fig. 3).

Fig. 3.

Consensus error grid for CGM–comparator pairs of abdomen and arm sensors

The sensor adhesion rate was 100% for abdomen-worn sensors and 97.1% for arm-worn sensors, without the use of any over-tape during the 15-day period. Data were provided until the end of the study by 94.3% and 85.3% of abdomen- and arm-worn sensors, respectively.

Safety and Pain

The safety assessment was performed by pooling reported AEs for all inserted sensors. A total of 16 adverse device effects (ADEs) were recorded in 14 participants. These involved erythema, hematoma, transient bleeding, and itching at the sensor insertion site. All 16 ADEs were classified as mild.

Pain assessment only included ratings for the primary sensor. Pain rating scores upon sensor insertion on the abdomen were approximately half of that reported for the arm, as reflected by pain scores of mean (min, max) 4.8 (0.0, 27.0) vs. 9.4 (0.0, 71.0) on a scale from 0 to 100, respectively. Reported pain upon removal for both the abdomen and arm sensors was rated lower, with a score of mean (min, max) of 2.5 (0.0, 22.0) and 3.7 (0.0, 28.0), respectively (Table 7). Overall, little to no pain was reported, as reflected by mean ± SD scores ranging from 0 to 10 for 70 sensors (7.1 ± 10.9) at insertion and 69 sensors (3.1 ± 5.7) at removal.

Table 7.

Summary of VAS pain score

	Insertion	Removal
Overall
N assessments	70	69
Mean (SD)	7.1 (10.9)	3.1 (5.7)
Min, median, max	0.0, 3.0, 71.0	0.0, 1.0, 28.0
Abdomen
N assessments	35	35
Mean (SD)	4.8 (6.2)	2.5 (4.7)
Min, median, max	0.0, 2.0, 27.0	0.0, 1.0, 22.0
Arm
N assessments	35	34
Mean (SD)	9.4 (13.9)	3.7 (6.6)
Min, median, max	0.0, 5.0, 71.0	0.0, 1.0, 28.0

VAS visual analogue scale, SD standard deviation

Discussion

This study evaluated the performance and safety of the i3 CGM. Point accuracy assessment revealed 20/20 AR of 95.5% and 95.3% for abdomen- and arm-inserted sensors, respectively. MARD was 9.4% and 9.8% for sensors inserted in the abdomen and arm, respectively. Reliability of the system was further investigated through CG-DIVA analysis, as well as by MARD stability over the study period. In addition, the CGM system is clinically accurate as indicated by 94.9% and 95.5% of data pairs falling in zone A of the CEG, for abdomen and arm, respectively, and the lack of data pairs in risk zones C, D, and E, which is consistent with previously reported results [6]. The mean absolute RoC, i.e., the glucose dynamics induced with the study procedures, of 0.78 mg/dL reported in this study is comparable to the published values for other approved systems [7–9]. A comprehensive validation of the device’s safety was performed by pooling data from all sensors. Moreover, a high rate of adhesive attachment without the need to use any sensor over-tape was also reported. Attachment issues often lead to premature sensor failure and the need for replacement [10]; hence, they are a central factor in using CGM systems.

iCGM criteria are a series of rigorous requirements that CGM systems need to fulfill in order to be approved by the US Food and Drug Administration [11]. This study showed that the single-center data of the investigational device satisfy all iCGM special controls, indicating its high quality. Therefore, the i3 CGM also complies with the recently proposed minimum performance requirements for regulatory compliance in Europe, known as eCGM [12]. As shown in this study, the performance of the i3 CGM can be compared to the published performance of established CGM devices assessed under comparable procedures such as the Dexcom G7 (8.0% MARD and 94.2% 20/20 AR, 15.5-day use) [13] and the FreeStyle Libre 3 (8.2% MARD and 94.2% 20/20 AR, 15-day use) [14].

The fact that this performance evaluation only concerns data from one of the six study sites, as well as its focus on the accuracy of the primary sensor, may be seen as a limitation. Subject population size addressed in the literature widely varies (i.e., depending on the study objective), and there is no standardized requirement for CGM performance studies. Nevertheless, this study’s population size aligns with the reported median size of 30, as determined by Freckmann et al. [7] from over 129 performance evaluation studies. This single-center data included 35 participants, whereas the sample size for the full study required a total of 140 participants. The complete dataset from the pivotal study was under regulatory review during the publication process and therefore was not available. However, this analysis was intended to provide European data, particularly German. This data corresponds to the highest-enrolling site; hence it can be considered representative for the overall study population.

Conclusions

The single-center data of this pivotal study investigated the performance of the new True Vie i3 CGM system. Data collected from 35 participants showed robust point accuracy results, which simultaneously fulfilled the iCGM special controls required by the Food and Drug Administration. Together with the clinical accuracy results and its safety validation, these results indicate the potential of the i3 CGM for the management of diabetes.

Author Contributions

Nina Jendrike, Manuela Link, and Sükrü Öter contributed as investigators in the study and in the review of the manuscript. Manuel Eichenlaub performed the data evaluation and contributed to data interpretation and the review of the manuscript. Guido Freckmann contributed to conceptualization, project administration, data interpretation and to the review of the manuscript. Leon Shi and Poul Strange contributed to planning, execution, analysis of the data as well as review of the manuscript. Frank Flacke contributed to data interpretation, conceptualization and review of the manuscript. Jiangfeng Fei, Jiyun Zheng, Fei Gao, Ao Gao, and Siting Zhu contributed to the development of the study protocol, interpretation of the data, as well as the review of the manuscript.

Funding

This study was funded by Sinocare. The journal’s Rapid Service Fee, as well as medical writing assistance was funded by Sinocare.

Data Availability

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

Declarations

Conflict of Interest

Nina Jendrike, Manuela Link, Sükrü Öter, and Manuel Eichenlaub are employees of the Institut für Diabetes-Technologie Forschungs- und Entwicklungsgesellschaft mbH an der Universität Ulm (IfDT), which carries out clinical studies, e.g., with medical devices for diabetes therapy on its own initiative and on behalf of various companies. Guido Freckmann is the general manager and medical director of IfDT. Guido Freckmann/IfDT have received research support, speakers’ honoraria or consulting fees in the last 3 years from Abbott, Ascensia, Bionime, Boydsense, Dexcom, Insulet, I-Sens, Lilly, Menarini, Novo Nordisk, Perfood, Pharmasens, Roche, Sinocare, Terumo, Vertex, Ypsomed. Leon Shi and Poul Strange are employees of Integrated Medical Development, which plans and executes clinical studies and is a contractor of Sinocare. Frank Flacke is employee and owner of Flacke Consulting GmbH providing strategy advise, medical affairs and clinical affairs services and is or was within the past 3 years a contractor of Sinocare, BOYDSense, embecta, Sanofi, Glucoset. Jiangfeng Fei, Jiyun Zheng, Fei Gao, Ao Gao, and Siting Zhu are all employees of Sinocare.

Ethical Approval

Study was registered under ClinicalTrials.gov (ID: NCT05806554). Ethical approval for this study was obtained by the Ethics Committee of the State Medical Association of Baden-Württemberg (Ethik-Kommission der Landesärztekammer Baden-Württemberg, Stuttgart, Germany). The study was performed in accordance to the declaration of Helsinki. All study participants provided written informed consent.