The aim of this investigation was to evaluate gentamicin dosing schedules and TDM in patients on ECMO at Glenfield Hospital following the introduction of a new ECMO circuit. Specifically the objectives were:

• ▶ to collect peak and trough plasma gentamicin levels and associated dosing regimens

• ▶ to assess the proportion of patients that achieve target peak and trough levels

• ▶ to suggest improvements to current dosing/TDM approaches.

Methods Patients treated on ECMO at Glenfield Hospital from April 2007 to May 2009 were identified from an ECMO database. Study data were extracted from medical notes, prescription charts and an online microbiology/biochemistry database. Data collected included, patient demography, doses, plasma gentamicin levels, urea and creatinine and if relevant duration of continuous veno-venous hemofiltration (CVVH) support. Gentamicin dosing schedules vary between ages and therefore data were stratified into age groups (neonate (<4 weeks), infant (1–12 months), children (1–18 years) and adults (>18 years)) to enable comparisons between groups.

For the purposes of this investigation, therapeutic peak and toxic trough levels were defined as 7–10 and >1.5 mg/l, respectively for all multiple dosing schedules of <4 mg/kg/dose. Extended interval doses of 5–7 mg/kg require trough levels or at least 6–12 h post dose to assess clearance. In UHL we use the Hartford nomogram to assess the clearance of 7 mg/kg doses, which has only been validated in adults. Descriptive statistics were used to analyse the data.

Results 43 patients were included in the audit (21 neonates, 3 infants, 7 children and 12 adults) and 167 plasma gentamicin levels were assessed (82 neonates, 17 infants, 22 children and 41 adults). 9 patients received CVVH support.

Gentamicin doses ranged from 1.6 to 7.0 mg/kg, reflecting the variety of dosing schedules. Preliminary results show that for neonates 50% of troughs were toxic and 50% peaks were subtherapeutic. In infants 21% of troughs were toxic and 11% of peaks were subtherapeutic whereas in children 80% of troughs were toxic and 0% subtherapeutic peaks, although the numbers of patients in these two groups were limited. In adults there were 51% toxic troughs and 11% subtherapeutic peaks.

The analysis is ongoing and it is hoped that a pharmacokinetic modelling approach will provide further insight.

Conclusion The preliminary results show that gentamicin levels in a large proportion of ECMO patients were outside the therapeutic range. There are complex schedules for patients treated with gentamicin and while some of these inappropriate doses/schedules are due to the incorrect selection by the prescriber, others are because our recommended doses no longer seem to be appropriate. From this it seems that the schedules currently used tend to lead to toxic troughs and subtherapeutic peaks (particularly in neonates) and therefore risk ineffective treatment of sepsis and nephrotoxicity in vulnerable patients.