We read with interest the article by Sadow et al.1 We found the data to be thought provoking, but we disagree with their recommendations regarding the choice of empirical antibiotics for febrile infants younger than 29 days. A recommendation to base decisions on certain tests, such as cerebrospinal fluid and urine studies, must take into account the accuracy of these tests. Regarding cerebrospinal fluid studies in neonates, several authors have already shown that initial cerebrospinal fluid findings can be quite misleading.2- 4 Of special interest are 2 cases of neonatal meningitis in which the initial cerebrospinal fluid findings were normal but grew group B streptococci.2,3 In these cases the treatment of choice would be to administer ampicillin and gentamicin. Regarding microscopic urinalysis in neonates, our experience has been that the analysis can be quite insensitive. Of additional concern are the data recently presented by Baker and Bell.5 In their study they did have patients with Listeria infections and, as expected, they had patients with Enterococcus infections,5 2 pathogens not adequately covered by treatment with cephalosporins or aminoglycosides. Last, there is a growing body of data that indicates that gentamicin is both safer and more effective when given on a 24-hour dosing schedule, which is also more cost-effective than more frequent dosing.6 In conclusion, we believe it is still prudent to use ampicillin with gentamicin or ampicillin with a third-generation cephalosporin in febrile infants younger than 29 days.
Del Vecchio MT, Sundel ER, Benstock MA. Choice of Antibiotics in Febrile Neonates. Arch Pediatr Adolesc Med. 2000;154(2):205-206. doi:10.1001/archpedi.154.2.205