New research reveals limits of continuous glucose monitors during severe diabetes episodes

While continuous glucose monitors have revolutionised diabetes management, recent studies highlight a critical blind spot: their inability to precisely measure extreme blood sugar levels, potentially affecting severe episode management.

Continuous glucose monitors have changed diabetes care for many people with type 1 diabetes, giving them near-continuous readings instead of relying on repeated finger-prick checks. But new research suggests the devices have a hidden weakness: when blood sugar rises or falls beyond their measurement range, they can show that levels are extreme without saying exactly how extreme. That matters most during the most severe highs and lows, when clinicians and patients may want the finest detail. According to The Conversation article, this does not alter the standard summary measures used in routine care, but it does mean some information is lost at the edges of the data.

The devices, usually worn on the upper arm, measure glucose every few minutes and send the results to a phone app. The US Centers for Disease Control and Prevention says CGMs can reduce the need for frequent finger-stick testing, while Johns Hopkins notes they typically provide hundreds of readings a day. The limitation described in the new research is similar to a thermometer that can tell you something is at least 100C, but not whether it is 120C or 180C.

To test how often this happens, the researchers examined nearly 47 million publicly available readings from 948 people with type 1 diabetes across four international clinical trials. They found that 93.5% to 100% of participants hit either the upper or lower limit at least once, and in three of the four studies more than a third of high-glucose episodes stayed at the top limit for at least an hour. In younger people and in those with higher HbA1c, a blood test that reflects average glucose over two to three months, the limit was reached more often.

The reassuring part is that standard clinical summaries still held up well. The researchers compared CGM data with simultaneous finger-prick measurements and found that measures such as average glucose, variability and time in range were largely unaffected. The main issue is not that the monitors become useless, but that they stop giving precise detail during the most severe episodes. The authors argue that software could easily flag how often and for how long a device hit its limit, giving doctors and patients a clearer picture without changing the basic technology.