Researchers at ETH Zurich have built a handheld breathalyzer that detects whether the body is burning fat by measuring acetone in exhaled breath, with accuracy that closely matches traditional laboratory ketosis testing. The device, developed under the supervision of Andreas Güntner, was validated across 12 participants in multiple diet and exercise scenarios. The study was published in Device, a Cell Press journal, and ETH Zurich issued its official press release on July 21, 2026. Engadget, ETH Zurich.

The clinical value proposition is straightforward: breath acetone is a well-established proxy for ketone body concentration, and ketosis correlates with fat metabolism. Existing methods for confirming ketosis, blood beta-hydroxybutyrate assays and mass spectrometry, require lab infrastructure or at minimum finger-prick blood draws. This device collapses the measurement chain into a single portable unit.

The core engineering problem the team solved is the signal integrity of exhaled breath. Prior attempts at portable acetone detection have been undermined by excess moisture and ambient contaminants that throw off readings. The ETH Zurich device filters out moisture and contaminants before the sample reaches the sensor, addressing what has historically been the primary accuracy bottleneck for real-world breath testing. Cell Press / Device, Engadget.

A companion smartphone app guides the user through the breathing protocol. Breath pressure and sample collection timing are the two variables most likely to degrade a reading, and the app detects when a user applies too much or too little breath pressure, adjusting sample collection timing to compensate. That closed feedback loop between app and hardware is what distinguishes the system from a passive sensor reading. Engadget.

In the study, acetone measurements from the handheld device closely matched lab results for detecting ketosis across the 12 participants, who were tested in various diet and exercise conditions. The match to lab-grade results, even in a small cohort, is the key finding that separates this work from earlier consumer acetone sensors that correlated loosely at best.

The device is already commercialized under the name Nutrion, sold by the company Alivion. Engadget. New Scientist also reported on the study, with coverage dated August 25, 2026.

Looking at the broader context, this sits at the intersection of two trends technology professionals will recognize: the miniaturization of analytical instrumentation and the shift toward app-mediated biological feedback. We have watched continuous glucose monitors move from clinical tools to consumer lifestyle devices over the past decade. Acetone breath analysis targets an adjacent metabolic signal, and if the accuracy holds up in larger studies, it could follow a similar adoption curve. The 12-person validation cohort is a meaningful limitation worth flagging. Demonstrating correlation in a controlled, small-N study is a necessary first step, but real-world variability in breath composition, diet adherence, and individual metabolic differences will test whether the device's filtering and app corrections hold at scale.

The pairing of a hardware sensor with a software layer that actively manages sample quality is architecturally sound for this class of device. It is the same pattern that made continuous glucose monitors viable: the sensor is necessary but insufficient without algorithmic correction for user behavior and biological noise. Whether Nutrion's commercial traction follows depends on factors the study does not address, but the engineering approach is credible.