This study focused on assessing Nima’s sensitivity at the lower levels of detection. As in the past, we used cornbread as a representative matrix, since it is easier to prepare accurately and in large quantities. The cornbread samples were prepared with various amount of gluten concentrations.
The samples were prepared by using Bob’s Red Mill Gluten Free Cornbread Mix as the base, and spiking in gluten to achieve the desired concentrations after baking. The desired concentrations were, 2, 4, 10, 15, 20, and 30ppm. However, it was determined that base cornbread contained 5.6 ppm of gluten so the final concentrations were 5.6, 7.6, 9.6, 15.6, 20.6, 25.6 and 35.6ppm respectively.
The samples were also prepared at different final weights, 232mg and 465mg. This weight variation allows us to cover varied sampling sizes of different users. These weights represent the range of sizes for what could be interpreted as ‘pea sized amount’. All the testing was conducted at Nima, by Nima’s team of scientists.
In total, 277 tests were performed, approximately 20 replicates at each concentration and weight.
The test distribution was as follows: 38 replicates at 5.6ppm, 39 replicates at 7.6 ppm, and 40 replicates at the remaining concentrations. Of the 277 capsules and samples tested, two resulted in errors (0.72%), while the remaining 275 capsules completed with valid results. Below are the results for each sample size and concentration, as well as the aggregate data.
Table 1: Results for 232 mg samples
Table 2: Results for 465 mg samples
Table 3: Aggregate Results
In aggregate, detection of gluten at 20.6 ppm and above was 100%. At 15.6 ppm, one sample out of 40 yielded a smile, while gluten was detected in all 40 samples at 9.6 ppm. The sensitivity of Nima decreased to 87% at 7.6 ppm and 16% at 5.6 ppm.
When comparing sample sizes, 232 mg vs 465 mg, there was a decrease in sensitivity at the lower concentrations, particularly at 5.6 and 7.6 ppm. All samples were positive for gluten at 9.6 ppm, and one result out of 20 was negative for gluten at 15.6 ppm for the 232 mg samples while all samples were positive for gluten at 465 mg. Above 20 ppm, all samples were positive for gluten at both weight sizes.
Nima sensitivity above 20 ppm was 100%. The error rate was 0.72% (2 out of 277 capsules).
Sensitivity increased as the amount of gluten present in each sample increased. The goal of Nima is to detect 20ppm with high certainty, which requires the sensor to be able to detect lower amounts as well.
Some community members using Nima report results of Gluten Found for samples of packaged food items labeled gluten free. An item may be legally labeled as gluten found if (among other things) the level is below 20 ppm but Nima may still detect gluten if present, however, the sensitivity is more variable below 20 ppm as can be seen by this study.