Abstract
Campylobacteriosis is the most frequently reported bacterial food-borne illness in the European Union and contaminated broiler meat is considered the most important source of infection in humans. The aim of the present study was to evaluate real-time PCR (qPCR) and digital PCR (dPCR) for quantification of Campylobacter jejuni in 75 broiler neck-skin samples collected from a poultry slaughterhouse, and to compare them with the ISO 10272-2 standard plate count method. For qPCR standard curve, C. jejuni-negative neck-skin samples were spiked with C. jejuni suspension with a known number of bacterial cells. The observed CFU/g values by qPCR correlated greatly with the expected values and qPCR showed good performance with the reliable limit of detection (rLOD) and limit of quantification (LOQ) of three and 31 target copies per reaction, respectively. However, both rLOD (1219 CFU/g) and LOQ (12,523 CFU/g) were beyond the EFSA-proposed critical limit of 500-1,000 CFU/g of neck skin. Although C. jejuni cell counts were ≤1,000 CFU/g in only 7/75 samples by plate counting, they were ≤LOQ in 60/75 and ≤rLOD in 26/75 (≤1,000 CFU/g in 24/75) samples by qPCR. A strong and statistically significant correlation was observed between qPCR and dPCR. Both PCR-based methods correlated significantly with the plate count method; however, the correlation was moderate. Using the Bland-Altman analysis, an average agreement was noted between all three methods, although with a large standard deviation. A significant bias toward overestimation in dPCR was observed, probably due to the relatively high number of false positive calls. The linear dynamic range was comparable in both PCR-based methods; however, qPCR proved to be more suitable for routine use. In the future, the establishment of a reliable molecular quantification of C. jejuni in poultry samples showing a wide range of contamination levels down to the proposed critical limit is needed to enable time- and cost-effective surveillance throughout all stages in the food production chain. As both rLOD and LOQ were beyond this limit, a modification of the procedure is suggested to include less sample dilution prior to DNA extraction to enable PCR-based quantification of C. jejuni at the proposed microbiological criteria.