Serious bacterial infections in febrile outpatient pediatric heart transplant recipients.

OBJECTIVES

The purpose of this study was to describe the incidence of serious bacterial infections (SBIs) in febrile outpatient pediatric heart transplant recipients and to assess the utility of using white blood cell (WBC) indices to identify patients at low risk for bacteremia.

METHODS

A retrospective study was conducted on all heart transplant recipients followed at a single children's hospital. All outpatient visits from January 1, 1995, to June 1, 2007, in which fever was evaluated were reviewed. Patients with history of a primary immunodeficiency, receiving concurrent chemotherapy, or having had a stem cell or small bowel transplant were excluded. Demographic, historical, physical examination, laboratory, and radiographic data were then recorded.

RESULTS

Sixty-nine patients had 238 individual episodes of fever evaluation; of these, 217 (91.2%) had blood cultures drawn with results available in their initial evaluation. There were six (2.8%) true-positive blood cultures and eight (3.7%) false-positive cultures. Chest radiography was done in 185 evaluations (77.8%), and 44 episodes of pneumonia (23.8%) were diagnosed. Of 112 urine cultures done, one (0.9%) was positive. Neither of two lumbar punctures performed were positive. In non-ill-appearing children without indwelling central lines or focal bacterial infections (pneumonia, cellulitis), the incidence of bacteremia was 1.2%. In children with a focal bacterial infection, the rate of bacteremia was 6.3%. WBC indices were not significantly different between bacteremic and nonbacteremic patients. A band-to-neutrophil ratio (BNR) of > or =0.25 and a published guideline for identifying low-risk infants using WBC indices identified all bacteremic patients, each with a sensitivity of 100% (95% confidence interval [CI] = 48% to 100% and 54% to 100%, respectively).

CONCLUSIONS

The incidence of bacteremia was low in febrile, outpatient pediatric heart transplant patients, especially in those who were not ill-appearing and did not have a focus of serious infection. Two different low-risk criteria performed well in identifying the bacteremic patients, although given the low number of true-positive cultures, the CIs for the sensitivities of these tests were extremely wide, and neither test could be reliably used at present. A prospective multicenter study is required to confirm the low incidence of bacteremia and low-risk criteria in this population.