Comparison of epidemiological methods for differentiation of Proteus mirabilis

During the study of suspected cross-infection within a hospital, epidemiological typing techniques are often employed to aid in the identification of transmission patterns of nosocomially-acquired bacteria. In this study, five bacteriological typing methods were assessed for their ability to type and differentiate 100 clinical isolates of Proteus mirabilis. The methods evaluated include biotyping, bacteriophage typing with the method of Hickman and Farmer and the scheme of Schmidt and Jeffries, and typing by both bacteriocin production and bacteriocin sensitivity. Each of the phage and bacteriocin methods were further evaluated by typing 14 isolates of P mirabilis derived from an outbreak of infection on a surgical intensive care unit. All 100 isolates were typed by biochemical characteristics and were divided into seven distinct biotypes. Ninety-five percent of these strains were separated into 61 lysis patterns with the phage set of Hickman and Farmer whereas the Schmidt and Jeffries phage scheme was able to type 67% of the cultures into nine unique lysis patterns. Bacteriocin sensitivity typing resulted in 41 distinct patterns with 71% of the strains typable and typing by bacteriocin production yielded 29 distinct bacteriocin types among the 80% typable strains. Typing by bacteriocin production demonstrated that 13 of the 14 epidemic isolates were the same strain. All epidemic-related cultures were untypable with the Hickman and Farmer phage set, whereas typing by bacteriocin sensitivity and the phage set of Schmidt and Jeffries resulted in two and four isolates typable, respectively. Based on the results of typing random and epidemic isolates of P mirabilis, bacteriocin production typing is recommended as a sensitive and specific typing method for future epidemiological surveys.