Accuracy, feasibility, and clinical impact of prospective Bayesian pharmacokinetically guided dosing of cyclophosphamide, thiotepa, and carboplatin in high-dose chemotherapy.

PURPOSE Relationships between toxicity and pharmacokinetics have been shown for cyclophosphamide, thiotepa, and carboplatin (CTC) in high-dose chemotherapy. We prospectively evaluated whether variability in exposure to CTC and their activated metabolites can be decreased with pharmacokinetically guided dose administration and evaluated its clinical effect. EXPERIMENTAL DESIGN Patients received multiple 4-day courses of cyclophosphamide (1,000-1,500 mg/m2/d), thiotepa (80-120 mg/m2/d), and carbop latin (area under the plasma concentration-time curve 3.3-5 mg x min/mL/d). Doses were adapted on day 3 based on pharmacokinetic analyses of cyclophosphamide, 4-hydroxycyclophosphamide, thiotepa, tepa, and carboplatin done on day 1 using a Bayesian algorithm. Doses were also adjusted before and during second and third courses. Observed toxicity was compared with that in patients receiving standard dose CTC (n = 43). RESULTS A total of 46 patients (108 courses) were included. For cyclophosphamide, thiotepa, and carboplatin, a total of 39, 58, and 65 dose adaptations were done within courses and 17, 40, and 43 before courses. The precision within which the target exposure was reached improved compared with no adaptation, especially after within-course adaptations (precision for cyclophosphamide, thiotepa, and carboplatin is 19%, 16%, and 13%, respectively); >85% led to an exposure within +/-25% of the target compared with 60% without dose adjustments. Toxicity was similar to that in a reference population, although the incidence of veno-occlusive disease was reduced. CONCLUSIONS Bayesian pharmacokinetically guided dosing for CTC was feasible and led to a marked reduction in variability of exposure.