Comparison of sampling methods used for MRSA-classification of herds with breeding pigs

Since the first report on methicillin resistant Staphylococcus aureus (MRSA) CC398 in pigs, several countries have determined the prevalence of MRSA-positive pig herds using different sampling and laboratory techniques. The objective of the study was to compare three sampling methods for MRSA-classification of herds. Therefore, nasal swabs of pigs and environmental wipes were collected from 147 herds with breeding pigs. Per herd, laboratory examination was done on 10 pools of 6 nasal swabs (NASAL), 5 single environmental wipes (ENVSINGLE) and one pool of 5 environmental wipes (ENVPOOL). Large differences in apparent prevalence of MRSA-positive herds between methods were found: 19.1% for ENVPOOL, 53.1% for ENVSINGLE, and 70.8% for NASAL. Pairwise comparisons of methods resulted in relative sensitivities of 26.9% (ENVPOOL vs. NASAL), 34.6% (ENVPOOL vs. ENVSINGLE), and 72.1% (ENVSINGLE vs. NASAL) with relative specificities of respectively 100%, 98.6% and 93.0%. Cohen's kappa was respectively 0.18, 0.32 and 0.55, thus varying between very poor and moderate agreement. Examination of environmental wipes is an easy and non-invasive method to classify herds for MRSA. The number of environmental wipes needed depends on e.g. required detection limits and within-herd prevalence. In low prevalent herds (e.g. herds with <3 positive pools of nasal swabs), 25 single environmental wipes are required to be 90% sure that MRSA is detected at a detection limit similar to analyzing 10 pools of nasal swabs. Individual analysis of environmental wipes is highly recommended, as pooling 5 environmental samples resulted in a substantial reduction of the apparent prevalence.