A single plasmid clone was predominantly responsible for gentamicin resistance at the Minneapolis Veterans Administration Medical Center (MVAMC) for 9 years, although two unrelated R plasmids were found.Epidemiological data and restriction endonuclease analysis of 25 plasmid isolates suggested that the clonally derived plasmid population had persisted at the hospital. However, one case of reintroduction of the original epidemic plasmid from a hospital in another state was documented 4 years after the introduction of the original plasmid to the MVAMC. Resistance by the clonally derived plasmid population was not localized geographically within the MVAMC but rather was a hospital-wide problem. Furthermore, previously described classes of DNA rearrangement of the original plasmid were also widely disseminated in the hospital, implying that spread of strains bearing index-related plasmids was relatively unimpeded within the MVAMC despite extensive barrier isolation and cohorting measures. Potential environmental reservoirs of the plasmids were identified in hospital sinks and drains, but their relation to continued patient infection is not known.Plasmid-encoded resistance to aminoglycosides in gramnegative bacilli has been a significant clinical problem in many hospitals since the middle 1970s. Epidemiological studies of gentamicin resistance plasmids have proven that individual plasmids can be responsible for large nosocomial outbreaks of resistant organisms (1-3, 7, 9, 11, 13, 14, 16, 17). A single index plasmid species was responsible for an epidemic of gentamicin-resistant enteric gram-negative bacilli between 1975 and 1976 at the Minneapolis Veterans Administration Medical Center (MVAMC) (5, 13, 14). The index plasmid first isolated during the 1975 to 1976 outbreak had a mass of about 60 megadaltons (13,14) and encoded resistance to ampicillin, cephalothin, chloramphenicol (most strains), sulfathiazole, kanamycin, neomycin, and tobramycin in addition to gentamicin (5). Since the initial reports, we have shown that plasmids very similar to the index plasmid that was responsible for the 1975 to 1976 outbreak were present at the MVAMC as late as 1983 and that the population of R plasmids related to the index was stably polymorphic with respect to several DNA rearrangements (8).We have previously focused on the molecular evolution rather than the hospital epidemiology of the index-related plasmid population of the MVAMC (8). The epidemiological questions which we address here include the following. (i) During the study period have index-related plasmids been reintroduced to the MVAMC from external sources? (ii) Were index-related plasmids disseminated in the MVAMC or were any of the index-related plasmid classes limited in distribution to particular regions of the hospital? (iii) Were there any index-unrelated gentamicin resistance plasmids at the MVAMC? While interesting from the standpoint of nosocomial R plasmid ecology, these questions are also of practical import since gentamicin resistance in gramnegative ba...