Comparison of susceptibility to beta-lactam antimicrobial agents among bacteria isolated from intensive care units

Evaluating microbial isolation and susceptibility patterns from institutional settings is a well-established component of ongoing infection control activities. At the University of Michigan Hospitals, susceptibility profiles for selected bacteria tested from 1984 to 1989 were analyzed for major changes in the percentage of organisms susceptible to beta-lactam antimicrobials. Data on bacteria isolated from respiratory specimens obtained from 1433 patients in intensive care units (ICUs) during a 10-month interval were compared with like data obtained from 750 non-ICU patients. Antimicrobial agents studied were chosen based on hospital formulary availability and prevailing usage in the institution. Susceptible and moderately susceptible categories were combined for purposes of reporting, since empiric therapeutic doses would cover strains having both susceptibility levels. Antimicrobic susceptibilities were compared and differences analyzed among the ICUs. Major shifts in susceptibility were noted during the 5-year period. The incidence and susceptibility profiles of the microorganisms varied considerably between ICU and non-ICU patients. Pseudomonas aeruginosa isolates from individual ICUs showed large variations in prevailing susceptibilities, with the burn unit harboring the most resistant strains. However, the neurological, surgical, and critical care medicine units also showed large numbers of antimicrobial-resistant pseudomonads. Among Enterobacter cloacae isolates, only imipenem showed a high level of activity against both ICU and total hospital isolates. When examined by individual ICU, however, imipenem resistance was seen in the general medicine and burn units. The burn and pediatric ICUs showed increased rates of recovery of beta-lactam-resistant E. cloacae isolates, although significantly high resistance rates were seen throughout all ICUs. The surgical ICU was noted to have an abnormally high incidence of lower respiratory infections caused by P. aeruginosa. The antibiogram indicated that one possible epidemic strain was involved. However, when the isolates were subjected to fatty-acid profiling by gas-liquid chromatography, it was found that cross-contamination with five discernible strains had occurred among the ten patients tested. These preliminary data suggest that resistant pseudomonads can be harbored and spread within an ICU, and that the ICU can act as a reservoir of resistance that is spread to a "step-down" unit.