Diabetes mellitus (DM) is a major global health challenge, affecting 537 million people worldwide and causing over 6.7 million deaths annually. If not properly managed, DM can lead to severe complications, including vision loss, vascular damage, kidney dysfunction, nerve damage, and ultimately death. The disease comprises two main types: type 1 diabetes (T1DM), characterized by insufficient insulin production, and type 2 diabetes (T2DM), associated with impaired insulin utilization.
Current diagnostic and monitoring methods rely predominantly on blood-based measurements, which are invasive, time-consuming, and costly. In this context, breath analysis has emerged as a highly promising alternative. It offers the potential to (i) enable large-scale, non-invasive screening of populations, (ii) monitor individuals at risk of developing DM, and (iii) provide patients with a painless and convenient tool for continuous self-monitoring. Specific volatile organic compounds (VOCs), such as acetone, propane, and ethanol, have been identified as a relevant biomarker cluster for DM diagnosis and monitoring.
However, translating breath analysis into practical diagnostic tools remains challenging. Target VOCs are typically present at low concentrations (ppm levels) and must be detected selectively in the presence of numerous interfering compounds, including other VOCs and high levels of humidity. In addition, natural variability in baseline VOC levels between individuals and over time necessitates frequent and reliable measurements using portable devices.
Current state-of-the-art technologies present a clear trade-off: highly selective analytical techniques (e.g. laboratory-based instruments) are expensive, bulky, and require skilled operators, while portable and low-cost sensors often lack the selectivity and stability required for reliable measurements, frequently suffering from issues such as baseline drift.
Within this context, there is a strong need to bridge this gap by developing a miniaturized and elegant sensing platform capable of accurate detection of VOC biomarkers in complex breath mixtures while