Bacterial cross-transmission was investigated during a 12-month period in an adult intensive care unit (ICU) by the generation of random amplified polymorphic DNA (RAPD) fingerprinting profiles, combined with automated laser fluorescence (ALF) analysis. The potential episodes of cross-transmission identified, were compared with those detected by the conventional first-line screen of antibiogram typing. Over the year, 215 primary gram-negative bacterial isolates were obtained from 160 patients. In total, 22 possible episodes of cross-transmission, involving 70 (44%) of the 160 patients, were identified by RAPD-ALF analysis, and 19 of these were substantiated with epidemiological evidence. Conversely, 31 possible episodes were identified on the basis of antibiogram data, but only three of these episodes, two involving Acinetobacter baumannii and one involving Serratia marcescens, correlated with those identified by RAPD-ALF analysis. It was concluded that analysis of antibiogram data alone is an unreliable method for assessing bacterial cross-transmission, unless the organism involved has a particularly stable or unusual resistance pattern. In contrast, the technique of RAPD-ALF analysis may provide a rapid and simple technique for obtaining an insight into the population dynamics of gram-negative bacteria in adult ICUs.
Copyright 2000 The Hospital Infection Society.