Co-detection of virulent Escherichia coli genes in surface water sources

Abstract

McNemar’s test and the Pearson Chi-square were used to assess the co-detection and observed frequency, respectively, for potentially virulent E. coli genes in river water. Con- ventional multiplex Polymerase Chain Reaction (PCR) assays confirmed the presence of the aggR gene (69%), ipaH gene (23%) and the stx gene (15%) carried by Enteroaggrega- tive E. coli (EAEC), Enteroinvasive E. coli (EIEC) and Enterohermorrhagic E. coli (EHEC), respectively, in river water samples collected from the Berg River (Paarl, South Africa). Only the aggR gene was present in 23% of samples collected from the Plankenburg River system (Stellenbosch, South Africa). In a comparative study, real-time multiplex PCR as-says confirmed the presence of aggR (EAEC) in 69%, stx (EHEC) in 15%, ipaH (EIEC) in 31% and eae (EPEC) in 8% of the river water samples collected from the Berg River. In the Plankenburg River, aggR (EAEC) was detected in 46% of the samples, while eae (EPEC) was present in 15% of the water samples analyzed using real-time multiplex PCR in the Plankenburg River. Pearson Chi-square showed that there was no statistical differ-ence (p > 0.05) between the conventional and real-time multiplex PCRs for the detection of virulent E. coli genes in water samples. However, the McNemar’s test showed some variation in the co-detection of virulent E. coli genes, for example, there was no statistical difference in the misclassification of the discordant results for stx versus ipaH, which im-plies that the ipaH gene was frequently detected with the stx gene. This study thus high- lights the presence of virulent E. coli genes in river water and while early detection is crucial, quantitative microbial risk analysis has to be performed to identify and estimate the risk to human health.