Association of human enteric pathogenicity and mouse lung toxicity of Escherichia coli.

Mouse lung toxicity of 439 strains (431 Escherichia coli, 1 Shigella dysenteriae 1, 1 Enterobacter cloacae, 5 Vibrio sp., 1 Klebsiella) was compared to other pathogenicity tests (mouse virulence, enterotoxicity, guinea pig eye test), to serogroup distribution, loss of virulence on storage, origin and haemolytic activity. Mouse lethality was highest in serogroup O4 (p < 0.001), O18a,c (p < 0.001); serogroups O6, O20, O75, O115, O147 were next in order. E. coli serogroups O19, O26, O28a,b, O32, O51, O53, O55, O73, O78, O79, O83, O105, O111, O112, O114, O117, O119, O124, O129, O136, O142 failed to show lung toxicity. Strains O4 and O18 isolated at different periods of time did not differ significantly in the lung test (p = 0.05, p = 0.01, p > 0.1, p = 0.05, p > 0.1). There was no significant difference between strains isolated from the stools of patients with enteritis and of healthy individuals (p = 0.1, p > 0.99) and between those isolated from all faecal specimens and from extraintestinal samples (p = 0.05, p > 0.3). There was no correlation between lung toxicity and other pathogenicity tests. Since strains isolated from healthy individuals were also toxic for mice, a positive lung test cannot be considered a criterion of the aetiological role of the agent.