CO09 | RETHINKING REFERENCE METHODS: THE CASE FOR THE POUR-PLATE TECHNIQUE IN ROUTINE BIVALVE TESTING

Rosa Luisa Ambrosio, Valeria Vuoso, Beatrice Testa, Giusi Fiore, Sara Varcasia, Aniello Anastasio | DMVPA Università "Federico II", Napoli, Italy.

Objective. This study aimed to evaluate the analytical performance of the alternative pour-plate method (ISO 16649-2) in comparison with the reference Most Probable Number (MPN) method (ISO 16649-3) for Escherichia coli quantification in bivalve molluscan shellfish (BMS), assessing its accuracy profile, relative trueness, precision and detection sensitivity at critical low thresholds in a complex matrix.

Methods. Mussels (Mytilus galloprovincialis) harvested from Class A areas were homogenized (flesh and intra-valvular liquid-FIL). For each test (each replicate per level), 100 g of homogenised FIL was sampled, independently artificially contaminated with E. coli (ATCC 25922), and re-homogenised to ensure even distribution of the inoculum. Eleven concentration levels were tested (ranging from 2.0 to 7.0 Log CFU/100 g; inoculum estimated via McFarland densitometry and confirmed by plate count). For each concentration level, five replicates were analysed in parallel using both the reference MPN and the alternative pour-plate methods. To maximize analytical sensitivity and lower the limit of detection to match the MPN reference standard, the alternative method was implemented by distributing a total volume of 10 mL of the 1:10 dilution across three plates. Statistical evaluation used the Accuracy Profile calculation tool (MCS; EN ISO 16140) to assess compliance with predefined Acceptability Limits (AL = ±0.5 Log), and relative trueness was visualized via a Bland-Altman plot.

Results. The Accuracy Profile demonstrated highly satisfactory performance for the alternative method across the entire range. For all tested levels, lower and upper β-expectation tolerance intervals calculated through the ISO comparison tool were fully contained within the established AL (±0.5 Log), mathematically confirming the method equivalence. Notably, at extremely low concentrations, the alternative pour-plate method demonstrated superior sensitivity, yielding more positive replicates than the reference MPN, which failed to detect E. coli in several trials due to its probabilistic limits. Inoculating 10 mL over three plates enhanced analytical sensitivity, proving robust detection capability even at low contamination levels. Furthermore, the Bland-Altman plot showed a negligible mean bias of 0.01333±0.1650 Log, with 95% Limits of Agreement ranging from −0.3101 to 0.3368 Log. All data points fell well within these limits, indicating excellent relative trueness and no significant systematic methodological bias across the entire 2.0 to 7.0 Log spectrum.

Conclusions. Building on preliminary studies carried out on artificially contaminated water and naturally contaminated BMS, this research supported a preliminary validation of the alternative pour-plate method by demonstrating its robust accuracy profile and relative trueness within a complex challenge model. By distributing 10 mL of the first tenfold dilution across three plates, this protocol circumvents traditional sensitivity limits of the pour-plate technique, exhibiting superior recovery rates compared to the reference MPN at the lowest contamination thresholds. According to our findings, the validated method represents a high-throughput, reproducible, and resource-efficient alternative for routine regulatory testing of mussels, capable of streamlining laboratory workflows without compromising food safety assessments. Nevertheless, further multi-laboratory studies with larger sample sizes, a wider range of bivalve species, and diverse E. coli strains are warranted to comprehensively corroborate these findings.