The choice of empirical antibiotic therapy for nosocomial pneumonia.

Nosocomial pneumonia has the highest crude mortality rate of all hospital-acquired infections. The choice of empirical antibiotic therapy can be based on in vitro sensitivities of sputum bacterial isolates. Organisms recovered from purulent sputum of 500 intensive care unit (ICU) patients from five hospitals in Louisville, Kentucky, had antibiotic sensitivities measured by microbroth dilution. The most common isolates were Staphylococcus aureus (26.6%), Pseudomonas species (24.8%), Enterobacter species (11.8%), Escherichia coli (7.2%), Klebsiella species (7.0%), Streptococcus species (4.4%) and Serratia species (4.0%). These organisms were tested for susceptibility to amikacin, ciprofloxacin, imipenem/cilastatin, ticarcillin/clavulanate, piperacillin, ceftriaxone and ceftazidime. The antibiotics effective against more than 80% of the seven most common isolates included amikacin (97%), imipenem (94%), ciprofloxacin (92%) and ticarcillin/clavulanate (84%). These sensitivity patterns may serve as a guide to choosing empirical antibiotics for ICU-acquired pneumonia in Louisville. When the use of a beta-lactam antibiotic alone or in combination with an aminoglycoside is preferred, imipenem or ticarcillin-clavulanate may be considered the antibiotics of choice. Studies are in progress to delineate the clinical use of ciprofloxacin alone or in combination with beta-lactam antibiotics for treating nosocomial pneumonia.