Pharmacokinetics of 0.25% levobupivacaine with adrenaline following caudal epidural administration in children

Brief Synopsis Levobupivacaine, the S(-) enantiomer of racemic bupivacaine, has been shown to be less cardiotoxic than racemic bupivacaine and its R(+) enantiomer while retaining equipotent local anaesthetic properties, and is commonly used by paediatric anaesthetists. The pharmacokinetics of levobupivacaine in children under 2 years of age (1), and of children less than 3 months of age (2) after caudal epidural blockade have been published; pharmacokinetics in children after levobupivacaine administration via lumbar epidural catheter and ilioinguinal nerve block have also been examined (3, 4). Adrenaline is commonly added to local anaesthetic solutions, both to provide a marker of inadvertent intravascular injection of solution, and with the intention to induce vasoconstriction thereby reducing the rate of systemic absorption of local anaesthetic (resulting in both lower peak plasma levels and therefore potential for toxicity, and longer duration of local anaesthetic effect). The pharmacokinetics of levobupivacaine administered with adrenaline have not been examined. We propose a study of the pharmacokinetics of levobupivacaine when administered via the caudal epidural route, with the addition of adrenaline. We plan to enroll 50 subjects up to 18 years of age undergoing elective sub-umbilical surgical procedures for which caudal epidural analgesia is indicated. Subjects will receive 2mg/kg levobupivacaine, as a 0.25% solution with adrenaline 5 mcg/mL (1:200 000), via the caudal route, under general anaesthesia. Peripheral venous blood samples will be taken from a dedicated intravenous catheter inserted at the time of surgery. Up to 5 blood samples per subject will be taken in the period up to six hours post caudal injection. Samples will be assayed for plasma levobupivacaine. Analysis of raw data will use a population, rather than individual, pharmacokinetic model (NONMEM), allowing accurate estimation of population parameters from data taken from a small number of subjects, and allowing inclusion of incomplete sample data from individual subjects (5). Sample timing is not crucial where this form of analysis is used, so proposed sampling times need not be strictly adhered to.