During 2007-2010, 13 545 confirmed human VTEC infections and 777 haemolytic uraemic syndrome (HUS) cases were reported in the EU; isolates from 85 % of cases were not fully serotyped and therefore could not be classified using the Karmali seropathotype concept. Seropathotype group D covered 5 % of isolates from fully serotyped cases; 14 cases (0.7 %) belonged to seropathotype group E, defined by Karmali et al. (2003) as non-human only. Isolates from around 27 % of cases could not be assigned. There were no HUS cases reported for the serotypes in groups D and E but 17 HUS cases could not be assigned. The health outcome was reported for only a fraction of confirmed cases. About 64 % of patients presented with only diarrhoea; VTEC infection resulted in HUS in around 10 % of cases. The new ISO/TS 13136:2012 standard improves the detection of VTEC in food. An alternative concept based on the detection of verocytotoxins alone or genes encoding such verocytotoxins does not provide a sound scientific basis on which to assess risk to the consumer because there is no single or combination of marker(s) that fully define a ‘pathogenic’ VTEC. Strains positive for verocytotoxin 2 gene(vtx2)- and eae (intimin production)- or [aaiC (secreted protein of EAEC) plus aggR (plasmid-encoded regulator)] genes are associated with higher risk of more severe illness than other virulence gene combinations. The 2011 O104:H4 outbreak demonstrated the difficulty of predicting the emergence of ‘new’ pathogenic VTEC types by screening only for the eae gene or by focusing on a restricted panel of serogroups. A molecular approach utilising genes encoding virulence characteristics additional to the presence of vtx genes has been proposed.