A Portable Gluten Sensor for Celiac Disease Patients May Not Always Be Reliable Depending on the Food and the User.

A strict lifelong gluten-free (GF) diet is currently the only known effective treatment for celiac disease (CD), an inflammatory disorder of the small intestine with a worldwide prevalence of about 1%. CD patients need to avoid wheat, rye, and barley and consume GF foods containing <20 mg/kg of gluten. However, strict adherence to a GF diet tends to reduce the quality of life of CD patients compared to the general population and may lead to fear of inadvertent gluten consumption, especially when eating out. To help alleviate risk of gluten exposure, a portable gluten sensor was developed by Nima Labs that allows CD patients to test foods on site prior to consumption. With very limited independent information on the analytical performance of the Nima sensor available so far, our aim was to evaluate the reliability of the sensor using a variety of different foods with defined gluten content. All samples were tested with the sensor and analyzed by enzyme-linked immunosorbent assay as reference method. Of the 119 samples with gluten content ranging from 2 to 101,888 mg/kg tested in total, the sensor showed 80 positive (67.2%), 37 negative (31.1%) and 2 invalid results at the first of three consecutive measurements. The detection rate for samples containing ≥20 mg/kg of gluten was 90%. Samples containing 2 mg/kg of gluten or below consistently tested negative, but samples with a gluten content between 2 to 20 mg/kg of gluten may either test positive or negative. Overall, the performance of the sensor was acceptable in our study, but we observed systematic variation between different users that also appeared to depend on the sample being tested. This highlights the need to improve user education especially regarding the effect of sampling, testing limitations in case of partially hydrolyzed, fractionated or fermented gluten and training users on how to perform the test in a way that gluten will be reliably detected.