Limited Sampling Strategies for Investigating Paclitaxel Pharmacokinetics in Patients Receiving 175 mg/m² as a 3-Hour Infusion

Summary

Results from a randomised European-Canadian trial showed that paclitaxel 175 mg/m² infused as a 3-hour infusion is recommended in patients with platinum-refractory ovarian cancer. We assessed limited sampling models for relevant paclitaxel pharmacokinetic parameters using stepwise forward multiregression analysis for both area under the plasma concentration versus time curve from time 0 to 27 hours (AUC) and the time above a plasma threshold concentration of 0.1 µmol/L (T≥0.1 µmol/L) for the 3-hour infusion of 175 mg/m². These strategies will be incorporated into future studies investigating pharmacokinetic/pharmacodynamic relationships of paclitaxel, using this dose and schedule. 24 pharmacokinetic curves were investigated from 24 patients with ovarian cancer treated with paclitaxel. 13 curves were selected by computer randomisation for the development of the model and the other 11 curves were used for the validation of the developed strategy. All patients received paclitaxel during a 3-hour infusion at a dosage of 175 mg/m². Both pharmacokinetic parameters of interest, AUC and T≥0.1 µmol/L, could be predicted adequately with a 2-sample point model. This 2-point model, selected as optimal for the determination of the T≥0.1 µmol/L, was:

$${\rm{T}} \ge 0.1\;\mu {\rm{mol}}/{\rm{L}}\;({\rm{h}}) = - 4.11\;({\rm{h}} \bullet \mu {\rm{mol}}/{\rm{L}}) \bullet {{\rm{C}}_{1{\rm{h}}}}(\mu {\rm{mol}}/{\rm{L}}) + 68.53\;({\rm{h}} \bullet \mu {\rm{mol}}/{\rm{L}}) \bullet {{\rm{C}}_{8{\rm{h}}}}(\mu {\rm{mol}}/{\rm{L}}) + 8.579\;({\rm{h}})$$

Using the best 2-point model for T≥0.1 µmol/L, a 2-point model was developed for estimating the AUC:

$${\rm{AUC}}\;(\mu {\rm{mol}}/{\rm{L}} \bullet {\rm{h}}) = 5.49\;({\rm{h}}) \bullet {{\rm{C}}_{{\rm{lh}}}}\;(\mu {\rm{mol}}/{\rm{L}}) + 12.88({\rm{h}}) \bullet {{\rm{C}}_{8{\rm{h}}}}(\mu {\rm{mol}}/{\rm{L}}) + 6.312xxx(\mu {\rm{mol}}/{\rm{L}} \bullet {\rm{h}})$$

where C_1h is the plasma concentration of paclitaxel at 1 hour after the end of a 3-hour infusion and C_8h is the paclitaxel concentration at 8 hours after the end of the infusion.

Both models are predictive for T≥0.1 µmol/L (r² = 0.91) and AUC (r² = 0.92), and proved to be unbiased and precise for T≥0.1 µmol/L with a relative mean predictive error (MPE%) of −0.08 ± 1.94%, and a relative root mean square error (RMSE%) of 6.2% and for the estimation of the AUC (MPE% = −0.94 ± 2.49%, RMSE% = 7.9%).

For patients treated with paclitaxel at a dose of 175 mg/m² given by a 3-hour infusion, the T≥0.1 µmol/L and the AUC of paclitaxel can be adequately estimated from two of the same timed plasma concentrations (C_1h and C_8h). These limited sampling strategies may greatly accelerate future studies on pharmacokinetic/pharmacodynamic relationships of paclitaxel in patients treated with 175 mg/m² of the drug given as a 3-hour infusion.