Multi-laboratory validation study of a real-time PCR method for detection of Salmonella in frozen fish

A multi-laboratory validation (MLV) study was performed to validate a quantitative PCR (qPCR) method developed by the FDA for detecting Salmonella in frozen fish. Frozen fish represents foods that use blending sample preparation procedure in the BAM Salmonella method. This study was carried out after a successful MLV study with baby spinach which represents foods that use soaking sample preparation procedure. Fourteen laboratories participated in this study and they each analyzed twenty-four blind-coded frozen fish test portions using both the qPCR and BAM culture (reference) methods. Across collaborators, this study yielded a ∼39 % and a ∼40 % positive rate for the qPCR and culture methods, respectively, which were within the fractional range (25 %–75 %) that is required by the FDA's Microbiological Method Validation Guidelines. The difference (ND-PD) and sum (ND + PD) of the numbers of negative deviation (ND) and positive deviation (PD), which measure disagreement of results between the qPCR and culture methods, both did not exceed the Acceptability Limit required by ISO 16140–2:2016. The relative level of detection (RLOD), which compares the levels of detection between the two methods, was approximately 1. Together, these results demonstrate that the qPCR and BAM culture methods performed equally well for the detection of Salmonella. Furthermore, the study showed that the qPCR method yielded reproducible results among laboratories and was both sufficiently sensitive and specific for the detection of Salmonella in foods. Automatic DNA extraction improves qPCR sensitivity with a higher quality of DNA extract and provides a high throughput application for FDA laboratories.