Suspected infection in children with cancer

Abstract

The microbiology and severity of suspected infections in children with cancer admitted to the John Hunter Children’s Hospital was determined in order to assess whether any alteration to the infection treatment protocol was required. All children with cancer aged 1–17 years who had an episode of suspected or proven infection that required parenteral treatment at John Hunter Children’s Hospital (JHCH) during 1994/95 were reviewed. Thirty-seven children were treated for cancer at JHCH; 62 admissions for suspected infection which involved 26 children were reviewed. Sixteen of these children had a permanent central line. Children with a central line had an increase in the number of days of inpatient treatment required for the treatment of suspected infection, and they had more episodes of infection. A pathogen was isolated more frequently with blood cultures being positive more often and Gram- positive species were methicillin resistant more often. These differences were not statistically significant. A pathogen was isolated in 52% of admissions. Sixteen pathogens were Gram positive; 12 were Gram negative, two were fungal and two were viral. Blood cultures were positive in 21 of 62 admissions, skin swabs in four admissions, urine cultures in three admissions, stool in two admissions and one species was isolated from an epidural catheter tip and from the sputum. In 16% of admissions, the identified organism was resistant to the initial empirical therapy of tobramycin and piperacillin. In a further 13%, flucloxacillin was added to the empirical regimen when a sensitive Staphylococcus was identified. No significant differences between the culture-negative and culture-positive groups were observed in admission pulse, fever or admission neutrophil count. However, those patients with a central line had a higher incidence of having a pathogen isolated if their temperature was > 39.5°C. The median length of stay was longer for patients with a pathogen isolated on blood culture. Admission blood cultures were positive in 53% of admissions with an initial neutrophil count > 1000 × 10⁹/mL. Each of these children had a central line. Only one child died of infection during the 2-year study period. This review supports the observations that Gram-positive infection is now more common than Gram-negative infection in children with cancer. Despite the management advantages a permanent central line affords it is clear those children with a central line have an increased rate of infection and there needs to be caution in their use. The most important is the observation that any fever > 39.5°C in a child with a central line is likely to be associated with a documented infection irrespective of the neutrophil count. The clinical outcomes observed in the present study indicate that tobramycin and piperacillin are effective empirical treatments for suspected infection in children with cancer.