Objective: Blood cultures are commonly included in the evaluation of febrile children younger than 3 years without focal source of infection. Clinicians treat patients with a positive blood culture before final identification of the organism. Their treatment might include reevaluation in the emergency department (ED), additional tests, parenteral antibiotics, and hospital admission even for children who ultimately have false-positive (FP) blood cultures. The advent of pneumococcal conjugate vaccine (PCV) has made occult bacteremia less common, decreasing the likelihood that a positive blood culture result before final organism identification will be a true pathogen. This study will identify the characteristics of patients with FP blood cultures in the post-PCV era.
Methods: Charts were reviewed of all children ages 1 to 36 months with a temperature of at least 38.08 degrees C who had a blood culture obtained in our community hospital ED from January 1997 to January 2005.
Results: Bacteria grew in 106 (3.5%) out of 2971 blood cultures. True positives (TPs), defined as true pathogens, had a prevalence of 0.7%, representing 19.8% of positives. FPs, defined as contaminants, occurred in 2.8% of cultures, representing 80.2% of positives. Patients with FP cultures had lower mean white blood cell (WBC) counts (10.51 x 10(9)/L vs. 16.95 x 10(9)/L; P = 0.0001) and lower mean presenting temperatures (38.8 degrees C vs. 39.4 degrees C; P = 0.005). FPs had longer time to positivity (34.6 vs. 17.7 hours; P = 0.001) than TPs. A culture with a Gram stain suggestive of a contaminant, time to positivity greater than 24 hours and a WBC of less than 15 x 10(9)/L had a PPV for an FP of 97%. When analysis was restricted to well-appearing children age 2 to 36 months with temperature of more than 39 degrees C without focal source of infection who were discharged from the ED, these three criteria had a PPV for an FP of 100%. In these highly febrile children, the FPs had significantly lower WBCs (9.14 x 10(9)/L vs. 22.84 x 10(9)/L; P = 0.0001) and longer time topositivity (33.4 vs. 19.8 hours; P = 0.007) than TPs. The likelihood of obtaining FP cultures increased after the introduction of PCV from 62.5% to 87.8% odds ratio, 4.3; 95%confidence internal, 1.44-13.38).
Conclusions: In the post-PCV era, the majority of blood culture results will be FPs. FP cultures are predictable in febrile children with WBC counts less than 15.00 x 10(9)/L, time to positivity of more than 24 hours, and a Gram stain result suggestive of a contaminant. Prospective studies applying these criteria to the at-risk population for occult bacteremia are indicated.