The concern about vitamin C is not hypothetical. Ascorbic acid, one of the most widely consumed over-the-counter supplements in the world and a natural component of fruits, vegetables, and orange juice, has been shown in previous studies to interfere with some electrochemical glucose sensors. Devices such as the Abbott FreeStyle Libre 2 and Libre 3 have been reported to produce falsely elevated glucose readings in the presence of ascorbic acid, a failure mode that could prompt a user to inject more insulin than needed and slide into severe hypoglycemia. Exercise poses a different challenge: many CGM sensors rely on glucose oxidase, an enzyme that requires oxygen as an electron acceptor, and because oxygen levels in tissue fluctuate during physical activity, readings can drift precisely when people with diabetes need reliable numbers most.

The device at the center of the new study, the iCan i3 CGM (also marketed in Europe as GlucoMen iCan), takes a different biochemical route. Instead of glucose oxidase, its sensor uses flavin adenine dinucleotide-dependent glucose dehydrogenase, or GDH-FAD. Because this enzyme does not use oxygen as an electron acceptor, its catalytic activity is not perturbed by the oxygen swings that accompany exercise. The sensor and its external transmitter are built as a single unit and stream readings to a smartphone app every three minutes over Bluetooth. The device is approved for people aged two years and older with type 1 or type 2 diabetes and is available in roughly one hundred countries across Asia, the Middle East, Europe, and Latin America.

To test whether this chemistry translates into real-world robustness, researchers led by Ronald Brazg at the Rainier Clinical Research Center in Washington State ran a single-center, double-blinded, randomized trial between May and June 2022, registered as NCT05348928. Sixteen adults with type 1 diabetes, aged 18 to 65 and managed with multiple daily injections or an insulin pump, each wore three sensors for the full fifteen-day wear period, two on the abdomen and one on the back of the upper arm. Participants attended three in-clinic visits of roughly eight hours each, at the beginning, middle, and end of the sensor life, during which venous blood was drawn every fifteen minutes and analyzed with a Yellow Springs Instrument 2300 analyzer, the laboratory reference endorsed by the US Food and Drug Administration.

The interference protocol was deliberately demanding. At each visit participants ate two meals designed to produce pronounced glucose excursions, dosing their own insulin as they normally would. At the second and third visits they also performed thirty minutes of moderate exercise on a treadmill or reclining bicycle between the meals. Before the second visit, participants were randomized to swallow 1000 milligrams of ascorbic acid, the standard dose found in most vitamin supplements, before the first meal and exercise at one of those two visits in a cross-over design. Throughout the clinic days the researchers compared hundreds of paired sensor and reference values, calculating agreement rates and the mean absolute and relative differences between the CGM and the YSI reference, both with and without vitamin C and both with and without exercise.

The headline result is one of quiet reassurance. Across the full measurement range of 35 to 450 milligrams per deciliter, 92.0 percent of CGM readings fell within 20 percent of the reference value for glucose above 100 milligrams per deciliter, or within 20 milligrams per deciliter for values at or below that threshold, a performance consistent with integrated CGM boundaries and with data later reported for the same device in a German study. More striking were the interference findings. The overall change in mean absolute difference between the exercise and non-exercise conditions was minus 1.3 milligrams per deciliter, and between the ascorbic acid and non-ascorbic acid conditions minus 1.1 milligrams per deciliter. The corresponding changes in mean absolute relative difference were minus 1.5 percent and minus 1.6 percent. In other words, if anything, accuracy was marginally better under the interference conditions, and none of the shifts were considered clinically meaningful.

The detailed breakdowns reinforced that conclusion. In the hypoglycemic range below 70 milligrams per deciliter, the ascorbic acid condition produced a change in mean absolute difference of minus 3.1 milligrams per deciliter, while in the target range of 70 to 180 and the hyperglycemic range above 180 the mean absolute relative difference changed by minus 1.1 percent and minus 1.7 percent respectively. Surveillance error grid analysis, which maps each paired reading onto a risk surface for clinical decision-making, showed that the vast majority of measurements fell within the no-to-mild risk zones both with and without vitamin C, and no measurements landed in the moderate-to-extreme risk regions. Analyses stratified by dosing sequence and by sex told the same story. Trend accuracy, measured by how well the sensor’s rate-of-change categories matched the reference, was also reasonable, with 61.8 percent concordance in the mildly falling range of minus 1 to 0 milligrams per deciliter per minute.

The authors attribute the resilience to the GDH-FAD chemistry combined with the device’s third-generation sensor architecture. Because the enzyme does not depend on oxygen, the tissue oxygen fluctuations that accompany moderate exercise simply do not enter the electrochemical equation. And because ascorbic acid interferes at the electrode surface in a chemistry-specific way, the different redox mediator system of the iCan sensor appears to sidestep the vulnerability documented for glucose oxidase-based competitors. The contrast with earlier literature is stark: a 2021 review reported a 3.8 percent change in MARD for the Dexcom G6 after thirty minutes of aerobic exercise and a 13.1 percent change for the FreeStyle Libre after forty-five minutes of cycling, whereas the present trial found changes in the opposite direction and of negligible magnitude.

Safety and tolerability were also favorable. All sixteen enrolled participants completed the study. Four adverse events were recorded among two participants, three of them insertion-site reactions such as bleeding and bruising and one a mild gastrointestinal complaint; there were no serious or severe events, insertion pain on a visual analog scale was mild, and only four sensors had to be replaced, three for insertion failure and one for a mid-wear malfunction. Mean sensor wear time was 14.2 days, close to the device’s rated lifespan, and participants lived their normal lives between clinic visits, managing their diabetes with their own usual tools while blinded to the study sensor’s output.

The trial’s limitations are worth keeping in view. Sixteen participants is a small sample, and although recruitment was blind to race, gender, ethnicity, height, weight, and body mass index, a larger and more demographically diverse cohort would strengthen the evidence. The 1000-milligram vitamin C dose matches standard supplements, but total body ascorbic acid from diet varies between individuals, and the study could not account for that background. Interference was also assessed only at defined time points during the laboratory visits rather than continuously in free-living conditions, and the exercise was limited to thirty minutes of moderate intensity on a treadmill or bicycle, leaving open questions about prolonged or vigorous activity. Even so, the study was designed to stress the sensor across the full glycemic range with real meals, real insulin dosing, and a full fifteen-day wear, and it passed those tests. For people with type 1 diabetes who take their vitamin C with breakfast and head to the gym afterward, the message is that this new generation of GDH-FAD-based sensors, at least, can be trusted to keep counting honestly.

Subject of Research: Effects of ascorbic acid and exercise on the accuracy of a GDH-FAD continuous glucose monitoring system in adults with type 1 diabetes

Article Title: Impact of Ascorbic Acid and Exercise on Accuracy of a Glucose Monitoring System with GDH-FAD Technology in Adult Patients with Type 1 Diabetes Mellitus: A Double-Blinded, Randomized Trial

Article References: Brazg, R., Mitchell, T., Fei, J., Zheng, J., Gao, F., Gao, A., Zhu, S., Flacke, F., Shi, L., & Strange, P. (2026). Impact of Ascorbic Acid and Exercise on Accuracy of a Glucose Monitoring System with GDH-FAD Technology in Adult Patients with Type 1 Diabetes Mellitus: A Double-Blinded, Randomized Trial. Diabetes Therapy. https://doi.org/10.1007/s13300-026-01918-9

Image Credits: AI Generated

DOI: 10.1007/s13300-026-01918-9

Keywords: continuous glucose monitoring, type 1 diabetes, ascorbic acid, vitamin C interference, exercise, GDH-FAD, glucose sensor accuracy, MARD, randomized trial, diabetes technology, interstitial glucose, hypoglycemia