Abstract
The advantages of real-time continuous glucose monitoring (rtCGM) over intermittently scanned CGM (isCGM) reportedly include lower glycated hemoglobin (HbA1c) levels as well as reduced glycemic variability. However, there have been few studies of the effect of switching from isCGM to rtCGM on glycemic control, as well as the specific factors underlying any observed improvements. To that end, all patients with type 1 diabetes mellitus who used the DEXCOM rtCGM device (Terumo Corporation, Tokyo, Japan) at our institution were reviewed, and 16 individuals with type 1 diabetes who switched from isCGM to rtCGM were investigated. The patients’ HbA1c decreased in 75% of the cases (p = 0.02). On the other hand, GMI increased in 75% of the cases (p = 0.01). Intriguingly, the percentage of time below range and coefficient of variation were significantly improved with rtCGM compared to isCGM (2.9% vs. 7.6%, p = 0.016 and 35% vs. 40%, p = 0.0019, respectively). We also found that the discrepancy between HbA1c and GMI among users of isCGM was a key indicator that improved when switching to rtCGM. If discrepancies are observed between HbA1c and GMI when using isCGM, switching to rtCGM should be considered for improving glycemic control.
GLYCEMIC CONTROL TECHNOLOGIES continue to advance, especially for patients with type 1 diabetes with markedly reduced insulin secretion who require buffering of glycemic excursions. Glycemic control in such patients is harder to achieve than in patients with type 2 diabetes with relatively preserved insulin secretion [1]. Continuous glucose monitoring (CGM) devices are one of the major technological breakthroughs in recent decades, as they enable physicians and patients to detect unexpected glucose fluctuations. CGM has been shown both in randomized clinical trials [2, 3] and real-world studies [4, 5] to not only decrease glycated hemoglobin (HbA1c) levels but also reduce glycemic variability to greater extents than when patients perform self-monitoring of blood glucose. The International Consensus in Time in Range recently proposed that the percentages of time in range (TIR), time below range (TBR), and time above range (TAR), as well as the coefficient of variation (%CV), should be considered key CGM metrics for glycemic control over a short period [6]. Moreover, the glucose management indicator (GMI) that is calculated using sensor glucose (SG) values reflects the mean HbA1c and assists patients with diabetes in determining the current state of their glycemic profile [6].
CGM systems are classified into 2 types: professional and personal [7]. Professional CGM systems, also referred to as masked CGM, are typically owned by physicians; patients remain unaware of their monitoring results until their glucose values are downloaded [7]. Personal CGM systems are owned by patients who use them to continuously monitor their blood glucose levels; this allows them to make immediate and self-initiated therapeutic adjustments [8]. Personal CGM is further classified into 2 subtypes: intermittently scanned (isCGM) and real-time (rtCGM) [8]. While isCGM requires patients to scan a sensor attached to their skin to check their SG on demand, rtCGM can determine SG values via a wireless connection to the sensor without the need for manual scanning [9]. Moreover, because the isCGM sensor can save only 8 hours of data, information can be lost if patients do not perform scans regularly. Even though isCGM eliminates the need for calibration to match the SG and blood glucose values using a glucose meter, any inaccuracies cannot be corrected; however, rtCGM can use calibration to correct such discrepancies [9]. More importantly, and despite rtCGM devices generally being more expensive than isCGM counterparts, the former not only alerts the user to hypoglycemia and hyperglycemia but also predicts these conditions in advance, allowing patients to intervene against glucose excursions and thereby improve their quality of life [8, 9]. Such advantages of rtCGM over isCGM have been observed in terms of better HbA1c control as well as reduced glycemic variability in a randomized clinical trial [10] and a real-world study [11]; however, only a few trials to date have compared the quality of glycemic control when switching from isCGM to rtCGM [10, 12].
In this study, we examined the effect of switching from isCGM to rtCGM on glycemic control among patients with type 1 diabetes.
Materials and Methods
Ethical statement
This study was approved by the Gunma University Institutional Review Board (ID 150008) and conformed to the provisions of the Declaration of Helsinki (revised in Fortaleza, Brazil, in October 2013). Each patient provided written informed consent before undergoing any study-related procedures.
Subjects
All patients with type 1 diabetes who switched from isCGM (with the third-generation algorithm) to either a DEXCOM G4 with G6 algorithms (Terumo Corporation, Tokyo, Japan) or G6 (Terumo, Japan) rtCGM device at the Department of Internal Medicine, Division of Endocrinology and Diabetes, Gunma University Hospital between 2021 and 2022 were reviewed. Each patient was informed that the rtCGM device includes features such as an emergency low glucose alarm, urgent low glucose risk alert (only for G6), low glucose alert, and high glucose alert; the settings of these alerts are shown in Table 1. The patients were advised that rtCGM allows for blood glucose calibration in case of discrepancies with actual blood glucose measurements. They were also instructed to avoid calibrating during periods of significant blood glucose fluctuations. Furthermore, the cost per month was estimated to be approximately 3,000 Japanese yen, with the possibility of price increases depending on the frequency of blood glucose measurements. Individuals who wore the device for at least 3 months were included, as previous studies showed a correlation between HbA1c and mean SG values [13]. Patients who were pregnant were excluded since target CGM metrics are different during pregnancy.
Table 1
Profile of patients included in the study
| N (male/female) |
16 (3/13) |
| Age (years) |
46 (35–59) |
| Duration of diabetes (years) |
21 (8–32) |
| Body mass index (kg/m2) |
23.5 (20–24.7) |
| HbA1c (%) |
7.2 (6.8–7.7) |
| Retinopathy (%) |
13 |
| Nephropathy (%) |
19 |
| Total daily dose (units) |
31.5 (27–34.4) |
| %Bolus (%) |
67.1 (59.4–76) |
| CSII (%) |
37.5 |
| Urgent low glucose risk alerts on (only for G6) (%) |
81.3 |
| Low glucose alerts |
|
| Off (%) |
25 |
| <60 (%) |
18.8 |
| <65 (%) |
43.8 |
| <70 (%) |
75 |
CSII, continuous subcutaneous insulin infusion; HbA1c, glycated hemoglobin.
The patients’ glucose levels were measured using rtCGM and downloaded to the DEXCOM Clarity Clinic Portal web-based software. The proportions of time in which glucose values were between 70 and 180 mg/dL (i.e., the TIR), below 70 mg/dL (the TBR), and above 180 mg/dL (the TAR) were calculated. The %CV was also calculated as the index of glucose fluctuation. We examined the changes in each of these parameters before and after switching the CGM method based on the average values at the time of consultation; we also assessed the mean HbA1c and CGM values at least twice during the time of consultation as well as during the period before switching (median, 70 days; range: 63–97 days) and after (median, 91 days; range; 77–97 days).
Statistical analysis
Data are presented as the median (interquartile range) or percentage (%) for categorical variables. Matched pairs were subjected to a nonparametric Wilcoxon signed-rank test. Group comparisons were performed using analysis of variance and the Wilcoxon rank-sum test for continuous variables. Associations between continuous variables were examined using Pearson’s correlation coefficient analysis. All tests of significance and the resulting p-values were 2-sided; the level of significance was set at 5%. Statistical analyses were performed using the JMP Pro software version 15.2.0 (SAS Institute, Cary, NC, USA).
Results
The profiles of all 16 included patients are shown in Table 1. The median age and duration of diabetes were 46 and 21 years, respectively. The median HbA1c was 7.2%, which indicated that the diabetes was moderately controlled. The main reason for requesting a change from isCGM to rtCGM was to alleviate concerns regarding hypoglycemia (82% of the responses), followed by inadequate blood glucose control (59%; data not shown). On the other hand, 47% of the patients expressed concerns about the financial burden associated with the adoption of rtCGM, and 41% worried about potential skin issues caused by such devices (data not shown).
Next, we retrospectively compared the glycemic indices of patients who switched from isCGM to rtCGM (Table 2). The HbA1c significantly decreased in 75% of the cases after switching (p = 0.02) (Fig. 1A), while the GMI (calculated using the SG values to predict HbA1c) increased in 75% of the cases (p = 0.01) (Fig. 1B). Intriguingly, the TIR did not change significantly (64% vs. 61%, p = 0.481), but the TBR and %CV were significantly improved with rtCGM compared to isCGM (7.6% vs. 2.9%, p = 0.016 and 40% vs. 35%, p = 0.0019, respectively). Before switching, patients using isCGM had scanned SG regularly (median, 8.7 times) and the percentage of “time CGM active” was already very high (90.0%); however, this value improved further to 95% after switching (p = 0.028). Moreover, patients experienced a median of 0.79 hypoglycemic alerts per day after switching to rtCGM, and performed a median of 0.65 calibrations per day.
Table 2
Comparison of the parameters of subjects who switched from isCGM to rtCGM
|
isCGM |
rtCGM |
p-value |
| HbA1c (%) |
7.2 (6.8–7.7) |
7.3 (6.3–7.5) |
0.02 |
| Glucose management indicator (%) |
6.6 (6.1–7.3) |
7.3 (6.9–7.4) |
0.01 |
| Time in range (%) |
64 (56–74) |
61 (55–71) |
0.481 |
| Time below range (%) |
7.6 (4.9–13.7) |
2.9 (2.1–4.2) |
0.016 |
| Time above range (%) |
25 (14–35) |
37 (26–42) |
0.0066 |
| Coefficient of variation (%) |
40 (38–43) |
35 (32–38) |
0.0019 |
| %Time CGM active |
90 (63–97) |
95 (91–98) |
0.028 |
| Scan/day |
8.7 (4.3–17.9) |
— |
|
| Frequency of hypoglycemic alert (/day) |
— |
0.79 (0.43–0.93) |
|
| Frequency of calibration (/day) |
— |
0.65 (0.33–3.2) |
|
Bold values denote statistical significance at the p < 0.05 level.
CGM, continuous glucose monitoring; HbA1c, glycated hemoglobin; isCGM, intermittently scanned CGM; rtCGM, real-time CGM.
We then investigated the pre-switching parameters that contributed to HbA1c improvement after switching; 12 patients had a decreased median HbA1c (ranging from –0.10% to –0.8%) while 4 had worsened values (ranging from +0.2 to +0.6%) (Table 3). Although no significant difference was observed, there was a tendency for higher GMI values in the improved group than in the non-improved group before switching (6.9% vs. 6.2%, p = 0.089) (Table 3). Moreover, when we examined the correlation between the extent of change in HbA1c before and after switching (ΔHbA1c) as well as the extents of changes in other parameters, only the ΔGMI was found to be significantly correlated with ΔHbA1c (r = 0.546, p < 0.05) (Fig. 2A).
Table 3
Factors of improved versus non-improved subjects who switched from isCGM to rtCGM
|
Improved |
Not improved |
p-value |
| N (male/female) |
12 (1/11) |
4 (2/2) |
0.065 |
| Age (years) |
44 (35–64) |
49 (32–8) |
1 |
| Duration of diabetes (years) |
22 (14–33) |
13 (2–30) |
0.275 |
| CSII (%) |
33 |
50 |
0.55 |
| Total daily dose (units) |
32 (28–34) |
28 (18–38) |
0.467 |
| %Bolus (%) |
65 (59–73) |
77 (53–85) |
0.271 |
| Retinopathy (n) |
1 |
1 |
0.38 |
| Nephropathy (n) |
1 |
2 |
0.066 |
| Body mass index (kg/m2) |
24.0 (20.3–24.8) |
21.0 (16.9–24.0) |
0.182 |
| HbA1c (%) |
7.5 (6.8–7.9) |
7.1 (6.6–7.2) |
0.302 |
| GMI (%) |
6.9 (6.2–7.5) |
6.2 (5.9–6.5) |
0.089 |
| HbA1c-GMI (%) |
–0.50 (–0.93 to –0.20) |
–0.85 (–1.20 to –0.13) |
0.543 |
| Time in range (%) |
60 (51–70) |
73 (67–78) |
0.039 |
| Time below range (%) |
6 (4–13) |
10 (8–14) |
0.163 |
| Time above range (%) |
29 (15–40) |
15 (11–22) |
0.078 |
| Coefficient of variation (%) |
42 (36–45) |
40 (38–41) |
0.493 |
| %Time CGM active |
89 (75–98) |
95.2 (89.2–96.3) |
0.865 |
| Scan/day |
9 (5–17) |
13 (3–24) |
0.903 |
Bold values denote statistical significance at the p < 0.05 level.
CGM, continuous glucose monitoring; CSII, continuous subcutaneous insulin infusion; GMI, glucose management indicator; HbA1c, glycated hemoglobin; isCGM, intermittently scanned CGM; rtCGM, real-time CGM.
Since GMI but not HbA1c was a potential factor in determining HbA1c improvement after switching, we speculated that addressing discrepancies between GMI and HbA1c could contribute to improving glycemic control. Accordingly, we assessed the correlations of GMI and HbA1c with both isCGM and rtCGM. As shown in Fig. 2B and C, HbA1c and GMI were strongly correlated with both isCGM and rtCGM; however, greater discrepancies were found with respect to isCGM than rtCGM. Finally, we found that rtCGM was more accurate than isCGM, as the GMI in patients using the latter was much lower than the HbA1c (Fig. 2D).
Discussion
In this study, we demonstrated that switching from isCGM to rtCGM improved not only HbA1c levels but also the TBR and %CV. Moreover, we demonstrated that a key indicator in improving the HbA1c was the GMI of patients undergoing isCGM. To our knowledge, ours is the first study to identify factors linked to improving HbA1c when switching from isCGM to rtCGM.
Intriguingly, the discrepancy between HbA1c and the GMI in the isCGM group was higher than that in the rtCGM group (Fig. 2B–D). Indeed, the HbA1c was higher than the GMI in the isCGM group, indicating that patients who depended on their SG readings may have encountered difficulties maintaining strict glycemic control while trying to avoid hypoglycemia given that the SG reading may have been lower than the actual glucose levels. In this respect, using rtCGM could produce more accurate assessment of hypoglycemia than using isCGM. More importantly, patients can calibrate their rtCGM devices when their SG readings are inaccurate, whereas this is not possible for isCGM devices [9, 14]. Patients using rtCGM performed a median of 0.65 calibrations per day when there were discrepancies between their actual blood sugar measurements and the device’s SG values, even during periods of stable blood sugar (Table 2).
Patients using isCGM in our study scanned their glucose levels regularly (8 times per day), which was presumably sufficient not only for obtaining accurate CGM data but also for stabilizing glycemic variability [6]. Yet, blood sugar was even better controlled when switching from isCGM to rtCGM, indicating that the devices’ alerts for present or imminent hyperglycemia/hypoglycemia were critical for glycemic control and factored into improving the HbA1c. These patients received a median of 0.79 hypoglycemic alerts per day to avoid hypoglycemia, contributing to glycemic stability (Table 2).
A previous study found that the rate of HbA1c improvement among patients using rtCGM was approximately 0.3% compared to those using isCGM; these findings contrasted sharply with our data [10]. Furthermore, other studies showed improvements in TIR or TBR but not in HbA1c among patients introduced to isCGM [3, 5], while patients using CSII are also reportedly unable to benefit from rtCGM [14]. The HbA1c values among our patients (including a substantial proportion of CSII users) decreased significantly. The median HbA1c of patients in our isCGM group was 7.2%, which was already relatively good; therefore, the fact that patients using rtCGM could improve HbA1c even further is noteworthy given the advantages of doing so in terms of avoiding hypoglycemia [9, 15].
Patients in our study had type 1 diabetes for a relatively long time (a median of 20 years). Recently the American Diabetes Association recommended that patients use CGM devices from the onset of disease given that glycemic control is more effective the sooner patients start doing so [8, 16, 17]. Interestingly, however, our data indicated that switching from isCGM to rtCGM maintains its advantages regardless of the duration of diabetes.
Several limitations need to be considered when interpreting our findings. This study had a cross-sectional retrospective design with a small sample size, and we only evaluated patients treated at our hospital. For this reason, our patient parameters might be different from those of subjects at other hospitals in Japan, especially with respect to sex ratio and treatment method. Additionally, while a cause-and-effect relationship could not be discerned, our longitudinal follow-up was only for 3 months; i.e., a very short time.
In summary, we demonstrated that rtCGM is superior to isCGM not only for maintaining better HbA1c values but also for better detecting or predicting hypoglycemia. More importantly, we found that the key factor that led to improvement when switching from isCGM to rtCGM was the discrepancy between HbA1c and GMI when using the former. Further long-term studies will be necessary to confirm our findings. Moreover, the gaps between isCGM and rtCGM are narrowing, and next-generation isCGM devices may have alert systems for hypoglycemia [18]. As such, CGM technologies and devices ought to continue being improved in the future.
Disclosure
None of the authors have any potential conflicts of interest associated with this research.
Authorship Contribution
Eijiro Yamada: Study design, data analysis, and drafting the article.
Yasuyo Nakajima: Data analysis and advising on the drafting of the article.
Kazuhiko Horiguchi and Shuichi Okada: Advising on the drafting of the article.
Masanobu Yamada: Final approval of the article.
Funding
No external funding was received for this study.
References
- 1 Holt RIG, DeVries JH, Hess-Fischl A, Hirsch IB, Kirkman MS, et al. (2021) The management of type 1 diabetes in adults. A consensus report by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetologia 64: 2609–2652.
- 2 Lind M, Polonsky W, Hirsch IB, Heise T, Bolinder J, et al. (2017) Continuous glucose monitoring vs. conventional therapy for glycemic control in adults with type 1 diabetes treated with multiple daily insulin injections: The GOLD Randomized Clinical Trial. JAMA 317: 379–387. Erratum in: JAMA 317: 1912.
- 3 Bolinder J, Antuna R, Geelhoed-Duijvestijn P, Kröger J, Weitgasser R (2016) Novel glucose-sensing technology and hypoglycaemia in type 1 diabetes: a multicentre, non-masked, randomised controlled trial. Lancet 388: 2254–2263.
- 4 Parkin CG, Graham C, Smolskis J (2017) Continuous glucose monitoring use in type 1 diabetes: longitudinal analysis demonstrates meaningful improvements in HbA1c and reductions in health care utilization. J Diabetes Sci Technol 11: 522–528.
- 5 Charleer S, De Block C, Van Huffel L, Broos B, Fieuws S, et al. (2020) Quality of life and glucose control after 1 year of nationwide reimbursement of intermittently scanned continuous glucose monitoring in adults living with type 1 diabetes (FUTURE): a prospective observational real-world cohort study. Diabetes Care 43: 389–397.
- 6 Battelino T, Danne T, Bergenstal RM, Amiel SA, Beck R, et al. (2019) Clinical targets for continuous glucose monitoring data interpretation: Recommendations from the International Consensus on Time in Range. Diabetes Care 42: 1593–1603.
- 7 Vigersky R, Shrivastav M (2017) Role of continuous glucose monitoring for type 2 in diabetes management and research. J Diabetes Complications 31: 280–287.
- 8 American Diabetes Association Professional Practice Committee (2022) 7. Diabetes technology: standards of medical care in diabetes-2022. Diabetes Care 45(Suppl 1): S97–S112.
- 9 Adolfsson P, Parkin CG, Thomas A, Krinelke LG (2018) Selecting the appropriate continuous glucose monitoring system—a practical approach. Eur Endocrinol 14: 24–29.
- 10 Visser MM, Charleer S, Fieuws S, De Block C, Hilbrands R, et al. (2021) Comparing real-time and intermittently scanned continuous glucose monitoring in adults with type 1 diabetes (ALERTT1): a 6-month, prospective, multicentre, randomised controlled trial. Lancet 397: 2275–2283.
- 11 Brown RE, Chu L, Norman GJ, Abitbol A (2022) Real-world glycaemic outcomes in adult persons with type 1 diabetes using a real-time continuous glucose monitor compared to an intermittently scanned glucose monitor: a retrospective observational study from the Canadian LMC diabetes registry (REAL-CGM-T1D). Diabet Med 39: e14937.
- 12 Préau Y, Armand M, Galie S, Schaepelynck P, Raccah D (2021) Impact of switching from intermittently scanned to real-time continuous glucose monitoring systems in a type 1 diabetes patient French cohort: an observational study of clinical practices. Diabetes Technol Ther 23: 259–267.
- 13 Xing D, Kollman C, Beck RW, Tamborlane WV, Laffel L, et al. (2011) Optimal sampling intervals to assess long-term glycemic control using continuous glucose monitoring. Diabetes Technol Ther 13: 351–358.
- 14 Bailey T, Bode BW, Christiansen MP, Klaff LJ, Alva S (2015) The performance and usability of a factory-calibrated flash glucose monitoring system. Diabetes Technol Ther 17: 787–794.
- 15 Foster NC, Miller KM, Tamborlane WV, Bergenstal RM, Beck RW, et al. (2016) Continuous glucose monitoring in patients with type 1 diabetes using insulin injections. Diabetes Care 39: e81–e82.
- 16 Champakanath A, Akturk HK, Alonso GT, Snell-Bergeon JK, Shah VN (2022) Continuous glucose monitoring initiation within first year of type 1 diabetes diagnosis is associated with improved glycemic outcomes: 7-year follow-up study. Diabetes Care 45: 750–753.
- 17 Mulinacci G, Alonso GT, Snell-Bergeon JK, Shah VN (2019) Glycemic outcomes with early initiation of continuous glucose monitoring system in recently diagnosed patients with type 1 diabetes. Diabetes Technol Ther 21: 6–10.
- 18 Nguyen A, White JR (2022) FreeStyle Libre 3. Clin Diabetes 41: 127–128.
