Bacillus cereus food poisoning: a provisional serotyping scheme.

A provisional serotyping scheme for Bacillus cereus has been developed. Eighteen selected cultures of B. cereus isolated from foods and clinical specimens were used to prepare agglutinating sera against the flagellar antigen. There were no significant cross-reactions between the sera. The sera were used to type 137 isolates of B. cereus from 34 British and Australian incidents of food poisoning associated with the consumption of cooked rice. The majority of the isolates from samples of food (56 of 84; 66-7%) and from clinical specimens (43 of 53; 81-1%) were of H-serotype 1. Types 1, 3 and 5 were isolated from foods, clinical specimens or both, obtained from 31 of the 34 episodes. Sixty-six isolates of B. cereus were obtained from 63 samples of uncooked long-grain rice. The majority of the isolates (38 of 66; 57-6%) were untypable and types 1, 3 and 5 made up only 5 of 66 (7-5%) of the total. The results support the theory that B. cereus may cause two distinct forms of foodborne illness. Strains from incidents of food poisoning characterised by a longer incubation period and with diarrhoea as the main symptom have provided serotypes 2, 6, 8, 9 and 10, whereas those causing the rice-associated emetic illness are usually serotypes 1, 3 or 5.