One day of glucose data can flag hidden diabetes risk

Prediabetes is a crossroads: some people drift back towards normal glucose control while others progress to type 2 diabetes, and the standard measures do not always tell the two apart. Continuous glucose monitors produce a rich stream of data, but different devices can read slightly differently, which raises doubts about whether any single metric can be trusted across platforms. There is also a practical question of how much wear time is needed before the data mean anything. This study asked whether the High Blood Glucose Index, a metric that emphasises how far readings climb above the healthy range, is robust across devices and predictive from just one day of monitoring.

Researchers analysed 365 valid continuous glucose monitoring records collected with three different devices across four independent groups of people. Statistical modelling partitioned how much of the variation in the High Blood Glucose Index came from the device used, the dataset it came from and the person's metabolic status. Full monitoring periods were used to compare the index across healthy, prediabetic and diabetic stages. Separately, the team tested whether a single day's baseline recording could predict who later progressed to overt type 2 diabetes.

Device type explained only 3.4% of the variation in the index, while metabolic status explained 31.8%, suggesting the metric behaves consistently whichever monitor is worn. The index rose stepwise from healthy through prediabetes to diabetes. A single day's baseline value separated people who later progressed from those who did not, with moderate predictive accuracy (an area under the curve of 0.74), matching mean glucose and beating time in range, the M-value and glucose variability measures. Most strikingly, among prediabetic people whose established glucose indices sat within recommended ranges, a value above 0.174 identified a hidden subgroup with a 5.39- to 6.13-fold higher risk of progression.

Established monitoring metrics such as time in range are widely used, but this work suggests they can look reassuring in people who are nonetheless moving towards diabetes. The High Blood Glucose Index appears to add complementary information rather than replace existing measures, and its consistency across devices is a genuine practical advantage. The findings extend the case for making richer use of monitoring data in prediabetes. The key open question is whether the 0.174 threshold holds up, because the number of people who actually progressed was limited and the authors themselves call for validation in larger, independent groups.

If you have prediabetes and use, or are considering, a continuous glucose monitor, this is a reason to think that even a single day of data may say more about your trajectory than a simple average, provided the right metric is applied. It is also a reason not to be fully reassured by readings that sit within recommended ranges, since some risk appears to hide beneath them. That said, this remains an early finding built on limited progression events, so it is a development to watch rather than a test to seek out today.