Neonates in neonatal intensive care units (NICUs) are particularly vulnerable to infection. The widespread use of broad-spectrum antibiotics in this situation places neonates at risk of becoming colonized by resistant strains of bacilli. De Man and colleagues studied the effect of changing antibiotic policy in NICUs on the emergence of resistant gram-negative bacilli.

The study compared different empiric antibiotic regimens in two identical NICUs. The standard regimen for suspected sepsis in one NICU was amoxicillin and cefotaxime therapy. In the second unit, the standard regimen was penicillin G and tobramycin. After six months, the regimens were switched for a further six months. Data collected for each neonate for each day of admission included measures of growth and development, bacterial cultures, treatment and type of ventilation.

During the year-long study, 436 infants spent at least one day in the units. The 218 infants admitted to each unit were comparable in all significant variables. In both units, the risk of colonization with resistant gram-negative bacilli was significantly higher during the amoxicillin-cefotaxime regimen. The relative risk of colonization with a bacterial strain resistant to the prevailing antibiotic therapy was 18 times higher in the amoxicillin-cefotaxime regimen compared to the penicillin-tobramycin regimen. Forty-one neonates were colonized during amoxicillin-cefotaxime therapy compared with three during use of the penicillin-tobramycin regimen. Enterobacter cloacae was the most common colonizing organism during the amoxicillin-cefotaxime regimen, and Escherichia coli predominated during the penicillin-tobramycin regimens. One death was attributed to sepsis during the study, and this occurred in a child receiving chemotherapy for neuroblastoma. Related outcome measures showed a significant reduction of 1.5 days in average NICU stay and a decreased proportion of days with a central venous line during penicillintobramycin use.

The authors conclude that policies governing the choice of empiric antibiotics make a significant difference to the development of antibiotic resistance in neonates in NICUs. This, in turn, results in significant clinical benefits to these patients and has implications for use in other hospital settings.