Van Houten M A, Uiterwaal C S, Heesen G J, Arends J P, Kimpen J L
Wilhelmina Children's Hospital, University Medical Center Utrecht Julius Center for Patient Oriented Research, The Netherlands.
Pediatr Infect Dis J. 2001 Feb;20(2):171-7. doi: 10.1097/00006454-200102000-00011.
Gram-negative bacteremia in children, a major cause of morbidity and mortality, may in part be induced by intensive treatment procedures and nonspecific use of antibiotics. Our primary objective was to study the causal relationship between the use of vancomycin and Gram-negative bacteremia, for which this antibiotic is not specifically indicated.
The study was conducted in a 105-bed tertiary care children's hospital in the period of 1994 to 1997. The study pertains to a cohort of children with suspected bacteremia, in whom a blood culture was performed during hospital stay. Using the bacteriologic laboratory registration system, we selected all pediatric cases with bacteriologically proved Gram-negative bacteremia (n = 105) and a random sample of 225 pediatric controls with negative blood cultures. Using logistic regression analysis we examined associations between Gram-negative bacteremia and the following factors: preceding use of antibiotics, antacids, corticosteroids, surgery, mechanical ventilation, parenteral nutrition, and invasive instrumentation; and the intensity of care assessed with the Therapeutic Intensity Scoring System (TISS 28).
Gram-negative bacteremia was positively associated with the use of aminoglycosides, cephalosporins, surgical interventions, central venous catheters, parenteral nutrition, antacids and dexamethasone. The strongest association was with the use of vancomycin (odds ratio, 8.1; 95% confidence interval, 3.1 to 20.9). In a multiple logistic regression model containing all above-mentioned variables, the use of vancomycin remained positively and strongly associated with Gram-negative bacteremia (odds ratio, 3.88; 95% confidence interval, 1.34 to 11.21). Further adjustments and restrictions in the analysis did not materially change these findings concerning vancomycin.
Among children suspected of bacteremia there are several drugs and clinical procedures influencing the risk for Gram-negative bacteremia. Empiric use of vancomycin is strongly and independently associated with Gram-negative bacteremia. The safety of using vancomycin solely on the basis of suspicion of bacteremia in children may not be warranted.
