A recent publication in the Journal of Diabetes Science Technology from Avari and colleagues reports additional outcomes from the I HART study,1 focusing on measures of glycemic variability (GV), with an emphasis on the risk of hypoglycemia. Study endpoints include a direct comparison between real-time continuous glucose monitoring (rtCGM) and flash glucose monitoring (FGM) at 8 weeks and 16 weeks, when 20 of the subjects initially using FGM were switched to rtCGM for a second 8-week period.
A major concern is the flawed design of the I HART study, which does not exclude the possibility of bias in making comparisons between the two systems.2 This has been acknowledged by the I HART study authors3 and raises identical concerns for interpretation of the current GV measurements. All baseline measurements in the rtCGM arm used Dexcom devices, the accuracy of which required twice-daily fingerstick calibration. Baseline measurements in the FGM arm were also taken with Dexcom devices, even though the FreeStyleLibre FGM system is factory calibrated and has different sensitivity and specificity at low glucose compared with the Dexcom systems.4 Therefore, the sensor-derived GV outcomes at 8 or 16 weeks reported by Avari cannot be compared between the rtCGM and FGM systems.
Notwithstanding the flawed study protocol, the overall conclusion that most, but not all, GV measures improve with rtCGM compared with FGM is not supported by the data reported in Tables 1 and 2. Significant differences in measures of GV are shown only for standard deviation (SD), coefficient of variation (CV), mean absolute glucose (MAG) and GV percentage (GVP), two of which favor rtCGM (SD, CV) at 16 weeks and two of which favor FGM (MAG, GVP). Interestingly, from a dynamic perspective, MAG and GVP are both measures that are responsive to amplitude and frequency of glucose oscillations, whereas CV and SD are responsive to amplitude but not frequency.5 Thus, the comparative sensitivities of rtCGM versus FGM cannot be deduced in the context of this analysis.
The use of median and interquartile range values to define the study outcomes is also problematic since this relies heavily on the most consistent 50% of observations and ignores the outlying readings which might substantially change the outcome in such small study groups. Notably, expert opinion now highlights the use of extended 5%-95% percentiles in assessing problematic hypoglycemia.6 Consequently, it would be more valuable to know the means and SDs of each outcome measure in judging the differential impact of rtCGM or FGM on GV or hypoglycemia.
In summary, because of acknowledged concerns with the I HART study design,2-4 the only supportable conclusion from the analysis by Avari is that there was no difference at 8 and 16 weeks in the frequency of severe hypoglycemia or the overall Gold scores for impaired awareness of hypoglycemia (IAH), and no change in HbA1c between rtCGM and FGM.3 The study does not provide evidence for the superiority of either rtCGM or FGM in assessing or managing GV in subjects with or without IAH, nor in the propensity to hypoglycemia.
Footnotes
Declaration of Conflicting Interests: The author declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: AS is an employee and shareholder of Abbott.
ORCID iD: Alexander Seibold 
https://orcid.org/0000-0002-2008-4593
References
- 1. Avari P, Moscardo V, Jugnee N, Oliver N, Reddy M.Glycemic variability and hypoglycemic excursions with continuous glucose monitoring compared to intermittently scanned continuous glucose monitoring in adults with highest risk type 1 diabetes. J Diabetes Sci Technol. 2020; 14(3): 567-574. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Seibold A. Regarding:switching from flash glucose monitoring to continuous glucose monitoring on hypoglycemia in adults with type 1 diabetes at high hypoglycemia risk: the extension phase of the I HART CGM Study – Concerns that reduce confidence in conclusions regarding hypoglycemia. Diabetes Technol Ther. 2019;21(2):97-98. [DOI] [PubMed] [Google Scholar]
- 3. Oliver N, Reddy M. Reply to Seibold and Schlaegeri HART: methodological inconsistency precludes hypoglycaemia conclusions. Diabet Med. 2018; 35(11):1619-1620. [DOI] [PubMed] [Google Scholar]
- 4. Bailey T, Bode BW, Christiansen MP, Klaff LJ, Alva S. The performance and usability of a factory-calibrated flash glucose monitoring system. Diabetes Technol Ther. 2015;17(11):787-794. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Peyser TA, Balo AK, Buckingham BA, Hirsch IB, Garcia A. Glycemic variability percentage: a novel method for assessing glycemic variability from continuous glucose monitor data. Diabetes Technol Ther. 2018; 20(1):6-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Johnson ML, Martens TW, Criego AB, Carlson AL, Simonson GD, Bergenstal RM. Utilizing the ambulatory glucose profile to standardize and implement continuous glucose monitoring in clinical practice. Diabetes Technol Ther. 2019; 21(suppl2):S217-S225. [DOI] [PubMed] [Google Scholar]