Avian pathogenic Escherichia coli (APEC), uropathogenic E. coli (UPEC), and newborn meningitis-causing E. coli (NMEC) establish infections in extraintestinal habitats (extraintestinal pathogenic E. coli; ExPEC) of different hosts. As diversity, epidemiological sources, and evolutionary origins of ExPEC are so far only partially defined, we screened a collection of 526 strains of medical and veterinary origin of various O-types for assignment to E. coli reference collection (ECOR) group and virulence gene patterns. Results of ECOR typing confirmed that human ExPEC strains mostly belong to groups B2, followed by group D. Although a considerable portion of APEC strains did also fell into ECOR group B2 (35.1%), a higher amount (46.1%) belonged to group A, which has previously been described to also harbour strains with a high pathogenic potential for humans. The number of virulence-associated genes of single strains ranged from 5 to 26 among 33 genes tested and high numbers were rather related to K1-positive and ECOR B2 strains than to a certain pathotype. With a few exceptions (iha, afa/draB, sfa/foc, and hlyA), which were rarely present in APEC strains, most chromosomally located genes were widely distributed among all ExPEC strains irrespective of host and pathotype. However, prevalence of invasion genes (ibeA and gimB) and K1 capsule-encoding gene neuC indicated a closer relationship between APEC and NMEC strains. Genes associated with ColV plasmids (tsh, iss, and the episomal sit locus) were in general more prevalent in APEC than in UPEC and NMEC strains, indicating that APEC could be a source of ColV-located genes or complete plasmids for other ExPEC strains. Our data support the hypothesis that (a) poultry may be a vehicle or even a reservoir for human ExPEC strains, (b) APEC potentially serve as a reservoir of virulence-associated genes for UPEC and NMEC, (c) some ExPEC strains, although of different pathotypes, may share common ancestors, and (d) as a conclusion certain APEC subgroups have to be considered potential zoonotic agents. The finding of different evolutionary clusters within these three pathotypes implicates an independently and parallel evolution, which should be resolved in the future by thorough phylogenetic typing.