Clinical microbiologists are often asked to determine the relatedness of a group of bacterial isolates, that is, to type them. During the last decade, traditional methods of strain typing, such as bacteriophage typing and serotyping, have been supplemented or replaced in many laboratories with newer molecular methods, such as plasmid fingerprinting (43), ribotyping (40), PCR-based methods (45), and analysis of chromosomal DNA restriction patterns by pulsed-field gel electrophoresis (PFGE)(4, 14, 27). Although bacteriophage typing is still used in a number of large reference laboratories around the world for epidemiologic studies of Staphylococcus aureus (36) and serotyping continues to be a useful tool for epidemiologic surveillance of Salmonella species (30), there is a need for a method of strain typing that can be used to type a broader array of bacterial species. At present, PFGE comes closest to satisfying that need (3, 42).

PFGE involves embedding organisms in agarose, lysing the organisms in situ, and digesting the chromosomal DNA with restriction endonucleases that cleave infrequently (14, 27). Slices of agarose containing the chromosomal DNA fragments are inserted into the wells of an agarose gel, and the restriction fragments are resolved into a pattern of discrete bands in the gel by an apparatus that switches the direction of current according to a predetermined pattern. The DNA restriction patterns of the isolates are then compared with one another to determine their relatedness. Currently, there are no standardized criteria for analyzing the fragment patterns. Consequently, different investigators viewing the same PFGE results may come to quite different conclusions as to which isolates should be designated as outbreak related and which should be designated as non-outbreak related. This guest commentary proposes a set of guidelines for interpreting DNA restriction patterns generated by PFGE. The authors are investigators from the United States who, over the last several years, have correlated epidemiologic data from dozens of outbreaks with strain typing results produced by PFGE. These guidelines are intended to be used by clinical microbiologists in hospital laboratories to examine relatively