Surveillance of Shiga toxin–producing E. coli (STEC) in the Netherlands is performed in 1 or both of 2 ways: by mandatory case notification from medical laboratories or physicians to regional public health services, which then report the cases to the Netherlands National Institute for Public Health and the Environment (RIVM), or by microbiological surveillance, when medical laboratories send suspected STEC isolates to RIVM for confirmation and further molecular characterization. Since 2016, STEC isolates have been subjected to whole-genome sequencing for high-resolution detection of clusters and outbreaks nationwide. During detailed retrospective analysis of the whole-genome sequencing data, we identified isolates from 2 clinical cases of STEC infection from 2019 and 2020 and characterized them as stx2a-encoding EAEC O104:H4 (isolates NL1981 and NL2076). Both cases involved middle-aged nonhospitalized women with abdominal cramps and bloody diarrhea. Neither of the 2 case-patients reported recent travel. The 2019 case-patient reported consuming cooked minced beef, and the 2020 case-patient reported consuming beef, hamburger meat, and vegetables and fruits from her own garden. In 2021, the Austrian Agency for Health and Food Safety obtained a sample of similar stx2a-encoding EAEC O104:H4 from a 10-year-old girl with hemolytic uremic syndrome. This case-patient also reported no recent travel, but she had consumed raw veal. Finally, we retrospectively identified stx2a-encoding EAEC O104:H4 in a sample obtained from pig meat (pork) in 2017 in the Netherlands.

Figure 1

Figure 1. Single-linkage hierarchical clustering tree of enteroaggregative Shiga toxin–producing Escherichia coliO104:H4 from the Netherlands and reference sequences. Tree results from core- and accessory-genome multilocus sequence typing with a heatmap...

Figure 2

Figure 2. Minimum-spanning tree from cgMLST (Enterobase STEC scheme) of enteroaggregative Shiga toxin–producing Escherichia coliO104:H4 from the Netherlands and reference sequences. Colored isolates are those added during this study: green...

We performed a genomic comparison among the recently identified isolates, 2 Shiga toxin–producing EAEC O104:H4 isolates in cases from the Netherlands from 2013 described elsewhere (8), and a set of representative isolates from a 2011 outbreak in Germany and France, including isolates from case-patients in Germany who shed the outbreak strain well into 2012 (long-term shedders). To achieve our full set of isolates (Table), we supplemented the dataset with O104:H4 isolates retrieved from public repositories to gain more insights into the relatedness among investigated isolates and track possible evolutionary events. We assessed genomic relations between isolates by core- and accessory-genome multilocus sequence typing using the Enterobase STEC schema implemented in Ridom Seqsphere version 8.0.1 (https://www.ridom.de). The resulting dendrogram was rooted with strain 1060_13, a Clermont type A isolate. The 2020 patient isolate from the Netherlands (NL2076) and the pork isolate (ESBL3427) clearly clustered with the representative outbreak isolates (Figures 1,2). Conversely, patient isolates NL1981 from the Netherlands in 2019 and MRV-21-00239 from Austria in 2021 appeared more similar to each other and the 2 patient isolates (338 and 381_1) from the Netherlands in 2013 than to isolates from the outbreak cluster (Figures 1,2).

We determined the content of the plasmid, virulence, and antimicrobial resistance genes using the PlasmidFinder, VirulenceFinder, and ResFinder databases (http://www.genomicepidemiology.org). Within the outbreak cluster, including the patient (NL2076) and pork (ESBL3427) strains from the Netherlands, isolates showed virulence profiles typical of the 2011 outbreak strain, including STEC virulence markers stx2a, lpfA, and iha; the EAEC pAA plasmid (with aggR, aar, aap, sepA, the aatPABCD operon, and the aggABCD operon); and some chromosomal EAEC markers such as pic and sigA, but they lacked the STEC hallmark intimin eae. In addition, the isolates within the outbreak cluster showed similar antimicrobial resistance profiles, including bla_CTX-M-15 and bla_TEM-1B. Exceptions were the 2012 isolates from long-term shedders in Germany, which lost the pAA plasmid but belong to the outbreak cluster on the basis of their core genome (9) (Figure 1). Ancestral trait reconstruction as implemented in the R package ape v.5.5 (10), indicated a high likelihood that the most recent common ancestor of the isolates identified in this study were lysogenized with an stx-encoding phage. Among the isolates in the outbreak cluster, the core-genome allelic distance was 0–15 alleles (Figure 2).

After the 2011 outbreak in Europe, only a few sporadic cases of infection with stx-producing EAEC O104:H4 strains, most related to travel to Turkey or North Africa (8,11), were reported. The fact that the particular EAEC O104:H4 strains have never convincingly been isolated from food, animal, or environmental sources, and that EAEC in general are primarily human-adapted (12), support the hypothesis of a human reservoir and potential multiple events of import by travelers from an area where this pathogen is endemic. On the other hand, the loss of essential virulence markers in long-term shedders might be a preview of common trends of bacterial genome reduction in long-term carriage in humans which, in turn, might indicate that the newly signaled isolates are maintained in niches other than the human one. Genomic analysis of several post-outbreak EAEC O104:H4 isolates suggests that they are not derived from the 2011 outbreak but share a recent common ancestor (13). Our analysis indicates that distinct more- or less-distant variants of Shiga toxin–producing EAEC O104:H4 are circulating worldwide. These variants are more likely to represent independent evolutionary events than continuous diversification of a single clade established and circulating in Europe after the large 2011 outbreak (14). In the absence of any previous indication of Shiga toxin–producing EAEC O104:H4 in animals, it was surprising to retrieve such an isolate from a pork product. However, this finding does not necessarily imply that pigs are a reservoir, because the contamination could also have originated from a food handler or contaminated feed rather than the pigs.

In conclusion, we show that Shiga toxin–producing EAEC O104:H4 isolates highly related to the 2011 outbreak strain are sporadically occurring in Europe. We emphasize the need to optimize safeguarding vulnerable chains of food production to prevent future outbreaks.