The purpose of our study was to determine the utility of the practice of culturing percutaneous liver biopsy specimens obtained from pediatric LT recipients for evaluation of fever and/or elevated serum aminotransferases. Accordingly, a retrospective analysis was done of the 101 liver biopsies obtained during an eight-year period which had been submitted for bacterial, fungal and/or viral culture (out of a total of 174 biopsies in 38 patients). The purpose of the analysis was to ask three questions. (1) What organisms were cultured? (2) Were there clinical profiles that were characteristic of the type of organism? (3) Was the practice cost-effective? The analysis indicated that 34/86 biopsy cultures were positive for bacteria, 4/75 for fungus and 2/81 for virus. Clinical and laboratory data for children with cultures positive for enteric flora (n = 9) were compared to those with cultures positive for Gram-positive organisms (n = 17), laboratory contaminants (n = 8), and those with negative cultures (n = 52). Children with biopsies positive for enteric flora had a 'cholestatic profile': mean direct bilirubin 7 mg/dl, ALT 78 IU/l, direct bilirubin/ALT 0.09, in comparison to children with biopsies positive for Gram-positive flora. These children had a 'hepatocellular profile': mean direct bilirubin 4 mg/dl, ALT 332 IU/l, direct bilirubin/ALT 0.01 (p = 0.04 versus the enteric flora values) and a high percentage of polymorphonuclear leukocytes (mean 69% versus 38% for those with negative cultures, p = 0.001.) The charge for performing each bacterial culture was $28 (total $28 x 86 = $2408: $268 per enteric flora-positive biopsy; $93 per biopsy positive for either enteric flora or Gram-positive flora). The charge for each fungal culture was also $28 (total $28 x 75 = $2100: $525 per positive culture), while the cost for each viral culture was $140 (total $140 x 81 = $11,340: $5670 per positive culture). Thus, discounting the eight cultures positive for laboratory contaminants, a total of $15,848 was spent for 32 positive cultures. Given the high cost of liver transplantation, this information suggests that discretion should be used in submission of liver biopsy samples for culture in pediatric transplant patients. We recommend that when liver biopsies are performed for evaluation of elevated serum aminotransferases and/or fever, culture of biopsy specimens for bacteria should be considered in children with a 'cholestatic profile': direct bilirubin > or = 7 mg/dl and direct bilirubin/ALT > 0.08, or a 'hepatocellular profile': direct bilirubin < or = 4 mg/dl and direct bilirubin/ALT < 0.05, together with polymorphonuclear leukocytes > 70%. Following these guidelines might provide valuable information pertinent to patient management (especially since Gram-negative organisms can sometimes be cultured from the liver and not from blood) while reducing costs. Fungal cultures should be restricted to critically ill children. However, our data suggest that the practice of obtaining fungal and viral cultures of the liver in most pediatric transplant patients has an unacceptably high cost/benefit ratio, particularly since recovery of the organism from the peripheral blood is likely.