Intravenous busulfan in adults prior to haematopoietic stem cell transplantation: a population pharmacokinetic study.

An IV form of busulfan (IV Bu) has recently become available for high dose conditioning regimen before haematopoietic stem cell transplantation (HSCT). This IV form is expected to reduce the high pharmacokinetic variability exhibited with oral busulfan and as a result, to better target the plasma area under the curve (AUC). Pharmacokinetics (PK) of IV Bu was investigated on 127 adult patients (333 PK administrations) who received 0.8 mg.kg-1 of Bu as a 2-h infusion every 6 h over 4 days, followed by cyclophosphamide (60 mg.kg-1 day-1x2). A retrospective population PK analysis was carried out to search for important predictive factors of IV Bu PK and to develop a limited sampling strategy (LSS) through Bayesian methodology. The analysis was conducted using the Non Linear Mixed Effect methodology and included a validation process on an independent data set. Adjusted Ideal Body Weight (AIBW) and Body Surface Area (BSA) were the best covariates to explain the inter-patient variability. The final inter-patient variability (CV=16%) in IV Bu clearance (Cltot) was estimated close to the intra-patient variability (CV=13%). There was neither age-dependency nor gender effect. IV Bu Cltot was not affected by elevated hepatic enzymes or by co-administration of either fluconazole or acetaminophen, and was not altered in heavily pre-treated or pre-transplanted patients. Normalised Cltot based on either AIBW or BSA was comparable between normal and obese patients (BMI=18-26.9 kg.m-2, >26.9 kg.m-2, respectively) whereas significant differences existed when based on either actual (ABW) or ideal body weight (IBW). As a consequence, no dose adjustment is required in obese patients when using a AIBW- or BSA-based dose calculation. A fixed dose of 0.80 mg.kg-1 of AIBW or 29 mg.m-2 of BSA yielded an average AUC of 1,200 microM.min, with 80% of patients within the "therapeutic" AUC range of 900-1,500 microM.min. Alternatively, 0.80 mg.kg-1 based on either ABW or IBW for normal patients and on AIBW for obese patients would achieve the same performance. A limited sampling strategy based on a Bayesian methodology was developed and validated on an independent dataset: AUCs obtained from one to two samplings were demonstrated to be reliably estimated.