Self-Monitoring of Blood Glucose Requirements with the Use of Intermittently Scanned Continuous Glucose Monitoring: A Follow-Up Analysis Using Real-Life Data

Abstract

Hypoglycemia-prone individuals with type 1 diabetes (T1D) who use intermittently scanned continuous glucose monitoring (isCGM) systems spend about 5 h/day in states where self-monitoring of blood glucose (SMBG) is indicated. Here we present estimates of the need for SMBG testing by retrospectively analyzing isCGM data from a cohort of real-world isCGM users. Data from 67 individuals were included in the analysis. Mean (SD) 3.18 (1.63) h/day was spent in an SMBG-indicated state and the number of transitions to an SMBG-indicated state was 3.86 (1.46)/day. Frequency of clinically important hypoglycemia [<3.0 mmol/L (<54 mmol/dL)] was median (IQR) 1.5 (0.6–3.4) episodes/week, of which only 50% were associated with a scan during the episode, and the average duration was 75.2 (63.9–91.8) min/episode. The need for continued SMBG testing remains important for all isCGM users and may affect the overall cost-effectiveness of isCGM. Impaired awareness of hypoglycemia and incidence of asymptomatic hypoglycemia may be underreported among real-life isCGM users in clinical practice.

Introduction

Intermittently scanned continuous glucose monitoring (isCGM) has become a popular technology tool for replacing finger-prick capillary blood glucose testing in diabetes self-management. The only commercially available isCGM device is the Freestyle Libre (Abbott Diabetes Care). Access to reimbursement of isCGM varies within and between countries, but still mainly applies to people with type 1 diabetes (T1D).¹

A recent systematic review of randomized-controlled trials (RCTs) evaluating isCGM in individuals with diabetes concluded that isCGM may lead to improvement in HbA1c in certain subgroups only and that the evidence is mixed regarding isCGM impact on improving time in range (TIR), glycemic variability, and hypoglycemia.² isCGM does not appear to reduce the risk of severe hypoglycemia (requiring third-party assistance to treat),³ and recent real-word data suggest that emergency admissions for hypoglycemia may increase with isCGM use.⁴

A U.K. based cost-effectiveness analysis for isCGM in T1D showed that the additional cost of isCGM compared with self-monitoring of blood glucose (SMBG) may be offset by the reduction in hypoglycemia observed with isCGM.⁵ However, the model assumed SMBG 0.5 times/day for isCGM use as observed in the IMPACT study,⁶ and the cost improvement was for severe hypoglycemia.

The isCGM system does not allow calibration and does not provide alerts and alarms for impending hypo- and hyperglycemia, in contrast to real-time CGM devices. The Freestyle Libre is licensed for nonadjunctive use, but the manufacturer's regulatory-approved U.K. label clearly defines indications for when a confirmatory capillary blood glucose test is required.⁷

We have previously shown, using 8 weeks of isCGM data from 20 individuals with T1D randomized to isCGM in an RCT comparing isCGM with standard care (the I HART CGM study), that people with T1D at high risk of hypoglycemia spent an average of 5 h per day in a state during which SMBG should be performed (namely a “SMBG-indicated state”—see definition below) in addition to the use of isCGM, and on average more than five SMBG tests per day were required.⁸ We concluded that our findings were important for recalculating cost-effectiveness of isCGM and may have important implications for decision-making around insulin dose calculations and driving.⁸ We recognize that our previous retrospective analysis was limited by the high propensity toward hypoglycemia in our population, and the results may therefore not be applicable to users of isCGM in real life. Here we present an analysis of the time spent in an SMBG-indicated state and the number of transitions into an SMBG-indicated state, using retrospective isCGM data derived from individuals with T1D commenced on isCGM in line with the U.K. reimbursement criteria for isCGM¹ in clinical practice. In this brief report, we additionally present the frequency and duration of hypoglycemic episodes and timing of scanning around hypoglycemia.

Methods

This was a retrospective data analysis. The study population included all adults with diabetes who either met the reimbursement criteria¹ or were self-funding isCGM at a large teaching hospital in London, United Kingdom.

All individuals had consented to sharing their isCGM data with the diabetes clinic through the LibreView platform. isCGM data were extracted from the LibreView clinic account and all data were anonymized before analysis. The study was not registered as a trial.

The percentage of values meeting one or any of the criteria for SMBG testing was calculated for each person, expressed as hours per day. The number of SMBG-indicated events per day was calculated for each individual based on the number of transitions to an SMBG-indicated state, assuming 96 readings per day. Given a time instant t, we define the isCGM to be in “SMBG-indicated state” at time t if any of the following criteria is met:

The isCGM reading at time t is <3.9 mmol/L (<70 mg/dL)

The isCGM rate-of-change at time t is >0.11 mmol/L/min (>2 mg/dL/min)

The estimated glucose value at t + 15 min is predicted to be <3.9 mmol/L (<70 mg/dL)

We define “transition to an SMBG-indicated state” when the isCGM is in “SMBG-indicated state” and the previous isCGM reading was not in “SMBG-indicated state.” For example, a hypoglycemic event [isCGM readings <3.9 mmol/L (70 mg/dL)] lasting for 45 min would have three isCGM readings in “SMBG-indicated state” (assuming 15-min sampling frequency) and only one “transition to an SMBG-indicated state.”

In addition, we calculated mean glucose, glucose variability represented by the coefficient of variation (CV), percentage TIR [3.9–10 mmol/L (70–180 mg/dL)], hypoglycemia [<3.0 mmol/L (<54 mmol/L)], hyperglycemia [>10 mmol/L (>180 mg/dL)], the frequency of daily scans, and the frequency and duration of episodes of hypoglycemia. Hypoglycemic events were defined as Level 1 [<3.9 mmol/L (<70 mg/dL)] and Level 2 [clinically important hypoglycemia, <3.0 mmol/L (<54 mg/dL)] for ≥20 min in line with the International Hypoglycaemia Study Group consensus.⁹ The results are presented as mean (SD) or median (IQR) depending on the data distribution.

Results

Data sets from 68 adults were extracted from the LibreView platform, of which 57 had more than 30 days of data. Data from one individual were excluded (<3 days of isCGM data), leaving data from 67 individuals available for analysis. Fifty-three percent were female. One individual had mitochondrial diabetes (on insulin) and the rest had T1D.

Median (IQR) age was 35 (29–50) years, HbA1c 62 (54–72) mmol/mol [7.8 (7.1–8.7%)], and diabetes duration 20 (11–31) years. In the study cohort, 28.7% were on pump therapy and 92.3% had intact awareness of hypoglycemia (documented as Gold score <4 or as free-text in the medical notes). The indications for Freestyle Libre were as follows: >8 capillary blood glucose tests per day (39.4%), occupational/psychosocial (24.2%), requiring third-party assistance to test (7.6%), alternative technology unsuitable (9.1%), and other (self-funders or initiated elsewhere with no clear indication, 19.7%).

The median (IQR) duration of data used in the analysis was 118 (50–190) days from a total time span of 136 (73–264) days. The median (IQR) number of scans was 9.26 (7.30–14.50) during the day and night (24 h), 7.68 (5.59–11.38) during the day (07:00–22:00 h), and 2.13 (1.51–3.43) during the night (22:00–07:00 h). Median (IQR) glucose was 9.3 (8.1–11.0) mmol/L [168 (146–198) mg/dL], % CV 39.9 (35.8–43.1)%, TIR [3.9–10 mmol/L (70–180 mg/dL)] 52.4 (39.6–65.5)%; time in hypoglycemia 4.0 (2.1–8.2)% and 1.2 (0.4–3.1)% [<3.9 mmol/L (<70 mg/dL) and 3.0 mmol/L (<54 mg/dL), respectively] and time in hyperglycemia [>10 mmol/L (>180 mg/dL) 40.4 (25.7–56.5)%].

The frequency (events/week) and duration (min) of Level 1 and 2 hypoglycemia overnight are summarized in Table 1. Overall (day and night), 54% of Level 1 hypoglycemic episodes and 50% of Level 2 hypoglycemic episodes were associated with a scan during the episode. If a scan was not done during the hypoglycemic event, then the mean (SD) timing of the nearest scan was 128 (48) min before or 105 (36) min after the onset of Level 1 hypoglycemic event. Similarly, scanning was done 165 (78) min before or 115 (42) min after a clinically important (Level 2) hypoglycemic event. In a subanalysis of waking hours, only (07:00–22:00 h) 61% of Level 1 and 58% of Level 2 hypoglycemic episodes were associated with a scan. Again, if a scan was not done during the event, the mean (SD) timing of the nearest scan was 98 (53) min before or 81 (38) min after the onset of Level 1 hypoglycemic event. Similarly, scanning was done 141 (98) min before or 80 (45) min after a clinically important (Level 2) hypoglycemic event.

Table 1.

Frequency and Duration of Hypoglycemic Events and Time Taken for Scan After Onset of Hypoglycemia

Metric	Day and night (24 h)	Overnight (22:00–07:00 h)
Metric	Mean (SD), median (IQR), range (min, max)	Mean (SD), median (IQR), range (min, max)
Level 1 hypoglycemia, <3.9 mmol/L (<70 mg/dL)
No. of events/week	6.0 (4.3), 4.8 (3.0–7.7), 23.0 (0.0, 23.0)	2.1 (1.3), 2.0 (1.2–2.9), 5.8 (0.0, 5.8)
Duration per event (min)	86.6 (27.9), 82.0 (70.1–104), 128.3 (40.2, 168.5)	121.3 (46.3), 112.5 (93.9–145.6), 220.4 (35.1, 255.6)
Level 2 hypoglycemia, <3.0 mmol/L (<54 mg/dL)
No. of events/week	2.5 (2.8), 1.5 (0.6–3.4), 13.8 (0.0, 13.8)	1.1 (1.1), 0.8 (0.3–1.3), 5.3 (0.0, 5.3)
Duration per event (min)	79.9 (26.5), 75.2 (63.9–91.8), 122.9 (30.0, 152.9)	105.1 (39.5), 103.9 (72.6–120.0), 225.8 (30.0, 255.8)

The times spent in an SMBG-indicated state based on the individual meeting one or any criteria and the total number of SMBG-indicated events (equates to number of transitions to an SMBG-indicated state) are summarized in Table 2. The mean total number of SMBG-indicated events per day, excluding the rapidly rising criteria, equates to 2.3 (1.0) events suggesting that over half of the SMBG-indicated events are associated with rapidly falling glucose, hypoglycemia [3.9 mmol/L (<70 mg/dL)], or impending hypoglycemia.

Table 2.

Mean (SD) Time Spent in an SMBG-Indicated State and Mean (SD) Number of Transitions to an SMBG-Indicated State

Metric	Mean (SD)
Time rapidly changing, h/day	1.43 (0.75)
Time rapidly rising	0.87 (0.43)
Time rapidly falling	0.55 (0.36)
Time impending hypoglycemia, h/day	1.67 (1.36)
Time <3.9 mmol/L (70 mg/dL), h/day	1.41 (1.34)
Total time in SMBG-INDICATED state (i.e., time any criterion was met), h/day	3.18 (1.63)
No. of SMBG-indicated events per day	3.86 (1.46)

SMBG, self-monitoring of blood glucose.

There were no statistically significant correlations between scanning frequency during the day with total number of Level 1 (r = −0.04, P = 0.77) and Level 2 hypoglycemic (r = −0.22, P = 0.08) events/day (Supplementary Fig. S1). However, data distribution suggests that when there are 12 or more scans during the day, the rate of Level 2 hypoglycemia is reduced by at least a factor of 3 relative to the rate when there are <12 scans per day.

There was positive correlation between daily scan frequency and transition to an SMBG-indicated state (events/day), P < 0.001 (Supplementary Fig. S2), but due to lack of SMBG data in this data set, we cannot confirm whether those who scan more actually do more SMBG tests.

Discussion

The outcomes from this retrospective analysis have shown that real-life users of isCGM experience on average 4.8 episodes of hypoglycemia [<3.9 mmol/L (<70 mg/dL)] per week and more importantly have 1.5 episodes per week of clinically important hypoglycemia [<3.0 mmol/L (<54 mg/dL)]. Interestingly, we observed that no scan was undertaken during half of all episodes of hypoglycemia. The average duration of clinically important hypoglycemic episodes lasted 80 min, and even longer during the night (105 min), suggesting a degree of impaired awareness of hypoglycemia (IAH) and frequent episodes of asymptomatic hypoglycemia among those who meet the U.K. reimbursement criteria for isCGM. A recent prospective observational study showed that the time to scanning from reaching hypoglycemia is significantly delayed in people with IAH compared with those without IAH.¹⁰

Compared with our previous post-hoc analysis in participants at high risk of hypoglycemia,⁸ isCGM users in this cohort spent less, but still a significant, time in an SMBG-indicated state [3.18 (1.63) h/day vs. 4.93 (1.81) h/day] and similarly, on average, required fewer SMBG tests per day [3.86 (1.46) vs. 5.43 (1.42)], which was unsurprising taking into account the differences in the cohorts analyzed. However, the results support our previous findings seen in the high-risk cohort, and therefore, our previous conclusions remain applicable to real-life users of isCGM. There is a need for existing cost-effectiveness models for isCGM to be revised, continued SMBG testing is required despite using isCGM, and omitting SMBG testing may have important implications for decision-making around insulin dosing and driving.

Accuracy of isCGM is thought to be reduced in the hypoglycemic range¹¹ and it is possible that our findings represent an overestimation of time spent in hypoglycemia. Recent data from a large observational study showed an improvement in HbA1c in individuals with an HbA1c > 58 mmol/mol at baseline following initiation of isCGM, however, an increase in HbA1c was observed in individuals with a baseline HbA1c < 58 mmol/mol.¹⁰ The reason for the deterioration seen in glycemic control in well-controlled individuals remains unclear, but may reflect that the occurrence of false-positive hypoglycemia measured by isCGM, without confirmatory SMBG, may trigger unnecessary ingestion of quick-acting carbohydrates resulting in glucose excursions above target. Current accuracy issues with glucose sensing technologies and the potential detrimental impact they may have on actual blood glucose may partly be overcome by following instructions in the device-specific user guidance. While it is recommended that all isCGM users should follow the manufacturer's label for safety, including the indications for SMBG testing, we also recognize that many isCGM users opt not to do so. The isCGM education provided in different centers may influence how diligently users follow the manufacturer's guidance. It has previously been shown that greater frequency of scanning with isCGM is associated with fewer episodes of hypoglycemia,¹² emphasising the importance of providing appropriate education to isCGM users.

The limitations of our study are the retrospective nature of the data collection and analysis, the inclusion bias (we only included isCGM data from individuals who had shared their data with the clinic), and no records of SMBG frequency as this is not uploaded routinely.

It is estimated that a quarter of individuals with T1D have IAH.¹³ However, in our study, only a 7.7% had IAH and may therefore not represent a typical cohort of isCGM users elsewhere in the world. Self-reporting of IAH may be underestimated in clinic due to the potential consequences IAH may have on driving and work. Furthermore, individuals with T1D at high risk of hypoglycemia (such as those with hypoglycemia unawareness) who meet the NICE criteria for insulin pump therapy or real-time CGM are offered either or both of these technologies as first-line therapy in conjunction with education. If pump therapy or real-time CGM is not suitable for some reason in these individuals, then isCGM is offered in line with the NHS England reimbursement criteria for isCGM. This approach may vary at different diabetes centers due to local reimbursement policies and resources available.

Although the evidence base for impact of isCGM on glycemic outcomes is mixed, it has been shown to have a positive impact on treatment satisfaction in RCTs² and from prospective observational real-life data.¹⁴ Our findings highlight the importance of reviewing isCGM data with the user during and in-between clinic consultations, with an emphasis on recognizing patterns suggestive of IAH to enable individualized support with diabetes self-management, and consideration of other sensing and insulin delivery modalities in line with NICE guidance.¹⁵

In conclusion, the need for continued SMBG testing remains important for all isCGM users and may effect the overall cost-effectiveness of isCGM. IAH and the incidence of asymptomatic hypoglycemia may be underreported among real-life isCGM users in clinical practice assuming the data reflect true hypoglycemia.

Footnotes

Acknowledgments

The authors thank the type 1 diabetes educators who run the Freestyle Libre initiation clinics at Imperial College Healthcare NHS Trust. The views expressed are those of the authors and are not necessarily those of the NHS, the NIHR, or the Department of Health. Imperial College London is supported by the NIHR Local Clinical Research Network and the Imperial NIHR Biomedical Research Centre. The authors also thank Dr Giada Acciaroli, PhD (Senior Data Scientist, Dexcom), for her contribution toward the data analysis.

Authors' Contributions

M.R. interpreted the results and wrote the article. N.O. reviewed and edited the article. M.R. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Author Disclosure Statement

M.R. has received research funding toward an investigator-initiated study from Dexcom and has participated in advisory boards for Roche Diabetes. N.O. has received research funding toward investigator-initiated studies from Dexcom, and has participated in advisory boards for Roche Diabetes, Dexcom, and Medtronic Diabetes.

Funding Information

No funding was received for this article.

Supplementary Material

Supplementary Figure S1

Supplementary Figure S2

References

NHS England. Flash glucose monitoring: national arrangements for funding of relevant diabetes patients. 2019. https://www.england.nhs.uk/wp-content/uploads/2020/02/national-arrangements-for-funding-of-relevant-diabetes-patients-v1.2.pdf. (accessed August 10, 2020 ).

Cowart

, Updike

, Bullers

: Systematic review of randomized controlled trials evaluating glycemic efficacy and patient satisfaction of intermittent-scanned continuous glucose monitoring in patients with diabetes. Diabetes Technol Ther, 2020; 22:337–345

Davis

TME

, Dwyer

, England

, et al.: Efficacy of intermittently scanned continuous glucose monitoring in the prevention of recurrent severe hypoglycemia. Diabetes Technol Ther, 2020; 22:367–373.

Stimson

, Dover

, Ritchie SA, et al.: HbA1c response and hospital admissions following commencement of flash glucose monitoring in adults with type 1 diabetes. BMJ Open Diabetes Res Care, 2020; 8:e001292.

Hellmund

, Weitgasser

, Blissett

: Cost calculation for a flash glucose monitoring system for UK adults with type 1 diabetes mellitus receiving intensive insulin treatment. Diabetes Res Clin Prac, 2018; 138:193–200.

Bolinder

, Antuna

, Geelhoed-Duijvestijn

, et al.: Novel glucose-sensing technology and hypoglycaemia in type 1 diabetes: a multicentre, nonmasked, randomised controlled trial. Lancet, 2016; 388:2254–2263.

Freestyle Libre manufacturer UK label. 2019. https://www.myfreestyle.com/provider/sites/all/themes/provider20/pdf/FreeStyle-Libre-In-Service-Guide.pdf. (accessed August 10, 2020 ).

Reddy

, Jugnee

, El Laboudi

, et al.: A randomized controlled pilot study of continuous glucose monitoring and flash glucose monitoring in people with type 1 diabetes and impaired awareness of hypoglycaemia. Diabet Med, 2018; 35:483–490.

International Hypoglycaemia Study Group. Glucose concentrations of less than 3.0 mmol/L (54 mg/dL) should be reported in clinical trials: a joint position statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care, 2017; 40:155–157.

10.

Moser

, Ziko

, Hochfellner

, et al.: People with type 1 diabetes and impaired awareness of hypoglycaemia have a delayed reaction to performing a glucose scan during hypoglycaemia: a prospective observational study. Diabet Med 2020; https://doi.org/10.1111/dme.14362

11.

Edelman

, Argento

, Pettus

, Hirsch

: Clinical implications of real-time and intermittently scanned continuous glucose monitoring. Diabetes Care, 2018; 41:2265–2274.

12.

Dunn

, Xu

, Hayter

, Ajja

: Real-world flash glucose monitoring patterns and associations between self-monitoring frequency and glycaemic measures: a European analysis of over 60 million glucose tests. Diabetes Res Clin Pract, 2018; 137:37–46.

13.

Geddes

, Schopman

, Zammitt

, Frier

: Prevalence of impaired awareness of hypoglycaemia in adults with type 1 diabetes. Diabet Med, 2008; 25:501–504.

14.

Charleer

, De Block

, Van Huffel

, et al.: 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, 2020; 43:389–397.

15.

National Institute for Health and Care Excellence. Type 1 diabetes in adults: diagnosis and management. August 2015; https://www.nice.org.uk/guidance/ng17 (accessed August 10, 2020 ).

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.14 MB

0.11 MB