Real-World Evaluation of a Near-Infrared Spectroscopy–based Noninvasive Glucose Monitoring Device in Adults With Type 2 Diabetes
Sandra Amuedo, Mónica Enríquez Macías, José Manuel Fornés, Arturo Moreno Arteaga, Jesús Riesgo Barrau, Virginia Bellido, Fernando Losada Viñau, Luz María Hidalgo Pichardo, Alfonso Soto Moreno, Antonio Puppo Moreno, Alfonso Pumar LópezBackground:
Self-monitoring of blood glucose remains widely used in many individuals with type 2 diabetes (T2D), although the burden of capillary testing often limits the frequency of measurements. Noninvasive technologies may improve glucose monitoring. This study evaluated the analytical performance and applicability of a near-infrared spectroscopy–based noninvasive glucose monitoring device.
Methods:
In this prospective, open-label clinical investigation, 52 adults with T2D performed 4 daily paired glucose measurements for more than 21 days using the investigational device and a reference capillary glucometer (Ascensia Contour Next). Following individualized calibration (10 paired measurements), post-calibration measurements within the 70 to 400 mg/dL range were included in the primary analysis. Accuracy was assessed using a combined mean absolute difference/mean absolute relative difference (MAD/MARD) metric, Parkes consensus error grid (CEG), and DTS error grid analyses.
Results:
Forty-eight participants were included in the primary analysis, contributing 2,616 paired post-calibration measurements. The mean age was 58.0 ± 9.9 years, and 31.3% were female. The overall combined MAD/MARD value was 16.3, calculated by applying MAD to paired measurements with reference glucose values <100 mg/dL and MARD to paired measurements with reference glucose values ≥100 mg/dL. Overall, 69.8%, 86.3%, and 94.5% of measurements were within ±20%, ±30%, and ±40% of reference values, respectively. Parkes CEG analysis classified 69.9% of measurements in zone A, 29.1% in zone B, 1.0% in zone C, and none in zones D or E. Accuracy improved after the first week. No device-related adverse events were reported.
Conclusion:
The device demonstrated a favorable clinical risk profile, stable analytical performance, and good safety under real-world conditions. Although analytical precision remains lower than current continuous glucose monitoring systems, its fully noninvasive design supports further development as a complementary monitoring strategy for individuals with T2D unwilling or unable to use invasive monitoring technologies.
Trial Registration:
ClinicalTrials.gov (NCT07075796).