Loading ...

Elevated and prolonged fluctuations in blood sugar were consistently associated with discordance in glucose management indictor (GMI) measures — based on a mean of continuous glucose monitoring (CGM) readings — and laboratory-measured A1c, in adults with type 1 diabetes (T1D), according to the results of a new study.

“Traditionally, A1c has been considered the primary marker of long-term glycemic exposure, while CGM was often viewed mainly as a tool for assessing glycemic variability,” said corresponding author Sang-Man Jin, MD, PhD, a professor of endocrinology and metabolism at Samsung Medical Center, Sungkyunkwan University, Seoul, Korea.

“However, several recent studies using large-scale real-time CGM datasets have shown that not only glycemic variability metrics but even CGM mean glucose and GMI themselves, can predict microvascular complications, cardiovascular outcomes, and mortality independently of A1c,” Jin told Medscape Medical News.

The widespread adoption of CGM has driven the use of the GMI, but discordance between GMI and A1c is common, and the clinical implications remain unclear, Jin and colleagues noted in their study published in Diabetes Care.

Glucose Excursion Patterns

The researchers examined whether the discordance itself could reflect distinct glucose excursion patterns with physiologic and clinical implications. Using data from 611 adults with T1D, they paired 90-day CGM traces with A1c results obtained within 15 days before or after the CGM data. The mean age of the participants was 45.8 years, and 42.6% were men.

The glucose rate increase detector (GRID) algorithm was used to quantify glucose excursions with varied peak glucose and time-to-peak thresholds. Discordance was defined using both GMI-to-A1c ratios and updated GMI (uGMI)-to-A1c ratios. The researchers compared associations with GRID-derived excursion metrics and conventional CGM-derived variability metrics.

Overall, higher excursions (peak glucose, ≥ 250 mg/dL) and longer excursions (time to peak, ≥ 90 minutes) were significantly associated with greater discordance based on the uGMI-to-A1c ratios. These findings were consistent between two CGM devices, described as sensor type 1 (Dexcom G5, G6, or G7) or sensor type 2 (FreeStyle Libre 1 or 2) with beta values of 0.174 (95% CI, 0.147-0.201) and 0.102 (95% CI, 0.068-0.136), respectively (P < .001 for both).

In a restricted cubic spline analysis, adjusting for glucose excursion metrics increased the informative value of the GMI and uGMI with regard to albuminuria and elevated triglyceride-glucose index, although no significant increase in clinical informativeness was noted for A1c.

“What impressed us most was how consistently major glucose excursions explained GMI-A1c discordance across different CGM platforms, A1c strata, and analytic conditions,” Jin told Medscape Medical News.

“Historically, much of the discussion around GMI-A1c discordance focused on nonglycemic factors such as red blood cell lifespan, renal function, or glycation biology,” he said. “In contrast, our findings showed that large and prolonged glucose excursions themselves appear to be an important determinant of the discordance.”

The association between higher GMI-A1c ratios and several conventional glycemic variability metrics influenced by mean glucose — including time above range, mean amplitude of glycemic excursion, and continuous overall net glycemic action — was interesting, said Jin. The consistent associations of the high level and longer duration excursions with discordance across patient subgroups was striking as well, he added.

The current findings suggest that discordance between GMI and A1c should not always be interpreted simply as measurement error. “In some patients, especially those with frequent prolonged postprandial hyperglycemia, elevated GMI relative to A1c may reflect a distinct glycemic phenotype characterized by substantial glucose excursions,” he explained.

Consequently, clinicians may find benefits in evaluating excursion patterns in addition to standard CGM summary metrics, he noted.

Seeking New Angles Supports Practice

“There is an innate difference in how the CGM GMI and A1c are derived; the former from interstitial fluid glucose, and the latter from glycation of red blood cells and therefore dependent on red blood cell life span, genetic, and environmental factors, so discordance is not unexpected,” said M. Cecilia Lansang, MD, MPH, professor of medicine and director of diabetes technology in endocrinology at the Cleveland Clinic in Ohio.

“However, instead of accepting this as the only explanation, the authors sought other explanations for this difference,” said Lansang, who was not involved in the study.

The current study findings offer a more solid basis upon which clinicians can attribute clinical observations and suppositions regarding the discordant GMI and A1c in practice and supplement their assessment of patients’ overall glycemic control using these two sources, she said.

“For practitioners who are not able to automatically include looking at the weekly or daily ambulatory glucose profiles into their practice to see the glucose excursions, the discordance between GMI and A1c may be used as a screen to delve further into more detailed review of the CGM downloads,” she explained.

Prospective studies are needed on the supplementary use of GMI with A1c in proactively changing patients’ diabetes treatment plans, said Lansang.

“Teasing out its predictive ability for complications can further elucidate the role of CGM, even for patients who are not on insulin,” she added.

The study was supported by the Bio & Medical Technology Development Program of the National Research Foundation funded by the Korean government. The researchers reported having no financial conflicts of interest related to the current study. Lansang disclosed current or recent investigator-initiated research funded by Dexcom and Insulet and serving on the scientific advisory board of Willow Laboratories.