Several studies have indicated a possible causative role of toxigenic bacteria in sudden infant death syndrome (SIDS). This study examined the effect of toxigenic E. coli on pregnant and infant mice to determine if these animals could be used as a model for SIDS pathogenesis. Strains of E. coli from the intestinal contents of infants who have died of SIDS or other causes and from the faeces of healthy infants were collected over a broad time scale. The isolates were tested for their ability to produce then known toxins of E. coli and were serotyped (O and H antigens). Certain serotypes (e.g. O1:H- and O25:H1) emerged significantly more frequently from cases of SIDS than from healthy infants and isolates of these types were generally toxigenic in Vero-cell cultures but whose verotoxicity was not related to classical Shiga or other known toxins. This mouse model was developed to test the effects of these toxigenic and also non-toxigenic strains. Four apparently healthy pups aged between 17 and 21 days died unobserved overnight but no pups of the 54 control mice died suddenly (P = 0.0247, Fisher's exact test). These were considered to represent sudden unexpected deaths. Pathological effects compatible with those in SIDS were observed in mouse pups exposed to toxigenic strains indicating this model may be suitable for further study into the pathogenesis of unexpected deaths in infancy. Providing an animal model of SIDS would promote a much better avenue for studying the pathogenesis of this enigmatic condition.