Summary Background Invasive fungal disease is a significant cause of morbidity and mortality in immunosuppressed children. The recognition of patients at risk for candidaemia is paramount to a better prognosis. Objectives To characterize Candida spp bloodstream infections (BSI) in a reference centre for paediatric oncology and to describe the most prevalent risk factors associated with candida infections. Patients/Methods This is a retrospective cohort study carried out with paediatric patients followed up with at the Institute of Pediatric Oncology, Brazil, who presented positive blood culture for Candida spp from January 2004 to December 2016. Results Ninety episodes of candidaemia were analysed; patients had a median age of 4.5 years, and 57.8% were males, with a diagnosis of solid tumours in 54.5% of cases. The most common Candida species were C albicans (35.6%), C parapsilosis (30.0%) and C tropicalis (16.7%). C tropicalis BSI was associated with neutropenia and skin lesions. Therapy was successful in 67.1% of the episodes. Older age and thrombocytopenia were associated with therapeutic failure. Death within 30 days occurred in 24.4% of patients; predictive factors were older age and admission to an ICU C parapsilosis candidaemia was a protective factor for death when compared to C albicans. Conclusion The main species isolated were C albicans, C parapsilosis and C tropicalis. C tropicalis BSI was associated with neutropenia and skin lesions. The death rate was significant, and a worse prognosis was associated with older age, thrombocytopenia and admission to an ICU C parapsilosis infection proved to be a protective factor against mortality.

Summary Background The wide pharmacokinetic variability of voriconazole leads to uncertainty regarding adequate exposure. Objectives To create a pharmacokinetic model that could help to explain the variability. Methods Retrospective review of paediatric patients with cancer. Models were built using Pmetrics. Results We analysed 158 trough measurements in 55 patients; in 41.8%, the serum levels were between 1 and 6 mg/L on initial measurement. After the measurements, dosage adjustments were made in 42 (76.3%) patients, and the percentage of adequate levels rose to 54.5%. Fourteen deaths (25.4%) were attributed to invasive fungal diseases. The mean serum levels were higher in deceased patients (mean ± SD: 3.1 ± 3.2 mg/L vs 2.5 ± 3.6 mg/L in survivors; P = 0.018), but the median doses per kg were higher in survivors. Drug exposure was also higher in deceased patients (mean ± SD of AUC: 19.2 ± 8.1 vs 9.5 ± 19.1 in survivors; P = 0.005). No correlation was found between serum concentrations <1 mg/L and death attributable to fungal disease. Bioavailability was estimated in 50%. The maximum velocity of clearance was reduced in deceased patients. Conclusions Extremely ill patients can be poor metabolizers of voriconazole. Therapeutic monitoring promotes only a limited improvement in drug management.

This a preprint and has not been peer reviewed. Data may be preliminary.