Alison Boast

Voriconazole dosing and therapeutic drug monitoring in children: experience from a paediatric tertiary care centre.

OBJECTIVES Therapeutic drug monitoring (TDM) of voriconazole is recommended to achieve trough concentrations of 1-5 mg/L. In children, this is challenging due to age-related variability in voriconazole pharmacokinetics. This study describes our experience with voriconazole, focusing on dosing regimens, dose adjustment and TDM. METHODS We reviewed the medical records of immunocompromised children who received voriconazole from July 2009 to January 2015 and had TDM. Demographic, clinical and voriconazole dosing and monitoring data were collected. RESULTS Fifty-five children received 62 courses of voriconazole and had TDM, with a total of 256 samples taken. Only 71.0% of courses (44/62) had TDM at the correct time, and at least one therapeutic level was achieved in only 52.3% (23/44) of these. Twenty-six courses had at least one sub-therapeutic level and in only 61.5% was the dose adjusted. Patients aged <6, 6-12 and >12 years required median intravenous doses of 8.8, 7.5 and 4.0 mg/kg twice daily, respectively (P < 0.001). With oral administration, patients aged 6-12 and >12 years required median doses of 4.7 and 4.3 mg/kg twice daily, respectively (P = 0.307). Levels within the target range were observed to fall below 1 mg/L in 36.4% of unchanged dosing regimens. Photosensitive skin reactions (20.0%) and hepatotoxicity (12.7%) were the most frequent adverse events and occurred in children with voriconazole levels <5 mg/L. CONCLUSIONS There is significant intra- and inter-individual variability in voriconazole concentrations in children, particularly in children <6 years of age. This warrants repeated TDM throughout treatment. Standardized guidelines for TDM and dose adjustment are required in children.

Effective CSF concentrations achieved with continuous-infusion flucloxacillin in a child.

A 3-month-old girl presented with irritability and fever 6 weeks after ventriculoperitoneal (VP) shunt insertion for obstructive hydrocephalus. On examination, she had a full fontanelle and shunt site erythema. The shunt was replaced with an external ventricular drain (EVD) from which cerebrospinal fluid (CSF) was sampled daily. CSF analysis showed the following: white cell count (WCC) 876×10/L (neutrophilia), glucose <1.1 mmol/L, protein 6.2 g/L, and methicillin-sensitive Staphylococcus aureus (MSSA) was cultured (flucloxacillin MIC 0.38 mg/L, vancomycin 1.0 mg/L, rifampicin <0.5 mg/L, linezolid 2.0 mg/L). Despite intravenous (IV) vancomycin 20 mg/kg 8 hourly and serum concentrations of 10– 16 mg/L, the CSF culture remained positive (day 3) and oral rifampicin 15 mg/kg/day was added with subsequent CSF sterilisation. Despite initial symptomatic improvement, on day 10, the CSF WCC increased (1299×10/L) and MSSA was recultured. The EVD was replaced, and IV linezolid 10 mg/kg 8 hourly was substituted for rifampicin. Intraventricular (IVT) vancomycin (12.5 mg/day) was also commenced with a CSF trough concentration of 39 mg/L. Although this initially led to CSF sterilisation and a reduction in WCC, the WCC increased to 1182×10/L (day 16). The patient was therefore changed to continuous-infusion (CI) flucloxacillin 300 mg/kg/day and rifampicin. The CSF concentration 37 h after commencement of CI flucloxacillin was 9.2 mg/L. After 96 h, the CSF and serum concentrations were 5.4 and 36 mg/L respectively (taken 20 min apart). After 3 days of CI flucloxacillin, the CSF WCC fell to 42×10/L (Fig. 1). A new VP shunt was inserted on day 31, and antibiotics ceased 7 days later. No treatment adverse effects were observed. Therapeutic CSF antibiotic concentrations are required for effective treatment of CNS infections. CSF penetrance is determined by patient factors (blood-brain barrier and blood-CSF permeability, age, CSF flow rate) and drug factors (lipophilicity, plasma protein binding, molecular size) [1]. Many antistaphylococcal agents are used to treat CNS infections; however, each treatment option is limited by specific physicochemical drug factors. Vancomycin is recommended as empiric therapy for shunt infections [2]. It is hydrophilic and has a high molecular weight (MW) resulting in variable CSF penetrance (CSF/serum ratio 0–0.66 in children) [1, 3]. The high doses recommended (60 mg/kg/day) to attain target serum concentrations (15–20 mg/L) [2] raises concerns regarding toxicity [4]. In our case, clinical failure with IV vancomycin was likely due to low serum concentrations. Linezolid is amphiphilic with low MW and low protein binding enabling excellent CNS penetration (CSF/serum ratio 1.2–2.3), even in the absence of inflamed meninges [1, 5]. However, the bacteriostatic activity of this drug may limit clinical response. Furthermore, in vitro studies have demonstrated antagonism * Amanda Gwee amanda.gwee@rch.org.au