Strategies are urgently needed for the rising incidence of type 1 diabetes (T1D), including overcoming the challenge of early detection. Only 15 % of cases occur in individuals with a first-degree relative affected by T1D. The remainder can only realistically be identified through thorough screening. According to a review published in Diabetologia, continuous glucose monitoring (CGM) could play a vital role as an addition to established markers such as oral glucose tolerance testing (OGTT), HbA1c, and islet autoantibodies.

“The current review by Alice L. J. Carr et al. notes that 10 of 14 studies identified time above 7.8 mmol/L as supportive of progression prediction. So, current data show that CGM can reveal metabolic instability alongside standard markers, positioning it as a useful complement to OGTT and HbA1c,” summarises Pieter Gillard, co-author and professor at UZ Leuven in Belgium. However, he adds, “At this moment, no current CGM system is considered sufficiently accurate by major professional or regulatory bodies to diagnose stage 2 T1D on its own, which is an important limitation for present-day clinical implementation.”

Looking out for glucose

OGTT or HbA1c, although important markers for glucose metabolism, are not suited as continuous surrogate markers of beta cell health, nor do they capture the timing of functional decline. This is, in part, because individuals tend to alternate between normoglycaemia (stage 1) and dysglycaemia (stage 2).

CGM could help to bridge this gap by providing longitudinal datasets.

“The major strength of CGM lies in its ability to deliver continuous, real-world insight into glycaemic patterns, capturing transient dysglycaemia that OGTT or HbA1c may not reliably detect,” says Jurgen Vercauteren, co-author and associate professor at KU Leuven. Vercauteren himself has been living with T1D for 11 years.

Across multiple cohorts, the review found that the proportion of time spent above 7.8 mmol/l, even at levels as low as 5 %, consistently predicted a higher risk of progression to stage 3 T1D.

“This pattern appears early, often before OGTT abnormalities emerge, and reflects subtle glucose instability that intermittent testing easily misses. As someone living with T1D myself, it resonates strongly how these ‘small drifts’ in glycaemia already carry meaningful biological information,” Vercauteren explains.

Looking out for people

Beyond physiological markers, the balance between informational value and patient burden is crucial when considering screening and staging approaches. HbA1c is minimally invasive, relatively inexpensive, and easy to implement, but it offers limited information. In contrast, OGTT and related tests, such as the mixed-meal tolerance test (MMTT) or intravenous glucose tolerance test (IVGTT), provide detailed information at the cost of substantial burden, particularly for children.

“I know first-hand how taxing repeated OGTTs can be for individuals and families,” says Vercauteren.

CGM may offer an improved balance with high informational value and low burden. In addition, it could potentially even be perceived positively, as it provides additional data and helps to contextualise glucose patterns and progression risk. However, this is currently supported by limited evidence, with only one of the studies in the review (Roberts et al., 2024) directly addressing this question.

More than a tool, but not yet a solution

While these are considerable benefits, CGM is currently not accepted as a primary endpoint for early-stage T1D prevention or preservation trials, although both the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA) recognise its clinical value.

“Scientifically, there is still no consensus on what constitutes a normal CGM profile across age groups. Older adults may spend around 10 % of time above 7.8 mmol/l even without diabetes, making age-specific thresholds essential. Psychosocial burden, wearability, data overload, and the possibility of behavioural change during monitoring all require much more structured investigation,” says Gillard. He adds that, importantly, CGM is also not a direct measure of beta cell function and should therefore not be interpreted in isolation from other markers. Lastly, there is also a lack of consensus on the use of CGM to guide insulin initiation.

Unlocking CGM’s full potential requires technological standardisation, and both authors agree that large prospective multicentre datasets must be provided.

“We need harmonised accuracy standards, large normative datasets, clearer guidance for blinded versus unblinded use, integration into public health strategies, and better education and support frameworks. Incorporating user-centred design and lived-experience insights will be essential to ensure the technology empowers rather than frustrates,” says Vercauteren.

He shares experiences from everyday challenges with CGM, from losing the Bluetooth connection to skin irritations.

“These real-world issues remind us that, while CGM is powerful, it is not yet flawless, and these practical limitations must be addressed before broad early-stage implementation.” Vercauteren’s message to manufacturers is clear: “Get it done!”

Making sense of glucose data

Finally, a frequently overlooked challenge is not capturing glucose data but analysing it meaningfully. Combining CGM and AI could help with this.

“An important strength is that CGM captures glucose as a time series rather than a single snapshot, which opens the door to more sophisticated pattern recognition, including AI and machine-learning approaches that may outperform simple threshold metrics in the future,” says Gillard.

Vercauteren highlights preliminary findings from the EDENT1FI study, coordinated by KU Leuven, and Helmholtz Munich, and involving a global consortium of 27 partners from academia, industry, and patient organisations across 13 countries, assessing early-stage type 1 diabetes in 200,000 children.

“We clearly see that children in early stages of T1D exhibit considerable glucose fluctuation, with dynamic patterns that classical screening tools fail to capture. This heterogeneity highlights the need for AI-driven pattern discovery to understand which trajectories predict faster transitions from stage 1 to stage 2, and eventually symptomatic stage 3. This will be crucial to ensure insulin is started early enough, while still avoiding unnecessary overtreatment,” Vercauteren says.

Looking ahead

As a first step, CGM may be particularly valuable for individuals at high immunological risk, where it can provide actionable insights. In other groups, such as people prone to data-related anxiety or very young children with sensory sensitivities, blinded CGM approaches might offer a more appropriate solution.

“The best candidates are those in whom CGM meaningfully changes follow-up intensity or treatment timing,” says Gillard.

Early detection is critical for several reasons. It can help to prevent cases of diabetic ketoacidosis (DKA). It could also pave the way for timely interventions to slow the autoimmune destruction of beta cells in stage 2 T1D using novel disease-modifying therapies. Now that teplizumab has become the first such therapy to be approved by the EMA at the beginning of 2026, strategies are needed to deploy these therapies effectively.

Much remains to be done before CGM can be used as a standalone marker for early detection of T1D. However, the direction is clear. With further validation, standardisation, and user-centred implementation, CGM has the potential to move from a supportive tool to a central component in screening strategies.

Key points:

• In 10 of 14 studies, time above 7.8 mmol/l emerged as a predictor of type 1 diabetes (T1D) progression. This supports the role of CGM alongside established markers such as OGTT and HbA1c.

• However, continuous glucose monitoring (CGM) use remains limited by the lack of scientific consensus on normal glucose profiles, missing technical standardisation, and practical challenges in everyday implementation.

• The authors highlight the need for harmonised standards, large prospective datasets, and clearer guidance on blinded vs unblinded CGM use to unlock CGM’s full potential.

• Broader implementation will require integration into public health strategies, improved education and support systems, and a stronger focus on user-centred device design.

Read the original publication here.

Author: Hanna Gabriel, BA MSc. Any opinions expressed in this article are the responsibility of EASD e-Learning.