| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
QUANTITATIVE CLINICAL PHARMACOLOGY |
From Bristol-Myers Squibb, Research and Development, Princeton, New Jersey. Dr Blackwood-Chirchir is a former employee of Bristol-Myers Squibb.
Address for reprints: Amit Roy, PhD, Strategic Modeling & Simulation Group, Discovery Medicine & Clinical Pharmacology, Route 206 & Province Line Rd, Bristol-Myers Squibb R&D, Princeton, NJ 08543; e-mail: amit.roy{at}bms.com.
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Key Words: Dasatinib • modeling and simulation • model-based analysis • absorption modeling • chronic myeloid leukemia • SRC • BCR-ABL
Dasatinib (BMS-354825) is a novel, potent, orally active, multitargeted inhibitor of several critical oncogenic kinases. These kinases include BCR-ABL, SRC family, c-KIT, platelet-derived growth factor receptor, and ephrin receptor kinases, all of which are associated with multiple forms of human malignancy. In particular, the BCR-ABL tyrosine kinase is associated with CML, and several clinical studies have demonstrated the efficacy of dasatinib in patients with CML and Ph+ ALL.13-17 BCR-ABL expression is the hallmark of CML, as more than 95% of CML cases are associated with BCR-ABL expression. CML is a progressive disease, which progresses through the chronic phase, accelerated phase, and blast crisis phase. As the most advanced stage of CML, the blast phase is highly refractory to therapy. The blast phase phenotype is myeloblastic in two thirds of patients and lymphoblastic in the remaining one third.
Dasatinib has been shown to be efficacious and well tolerated in subjects with all phases of CML and Ph+ ALL that are resistant or intolerant to imatinib.13-17 Dasatinib is approved in the United States, European Union, and several other countries for the treatment of CML and Ph+ ALL in patients resistant or intolerant to imatinib. Dasatinib has been shown to be 325-fold more potent than imatinib and binds to both inactive and active conformational forms of the ABL kinase domain.18-20 Furthermore, it is active against 18 of 19 imatinib-resistant mutations.21,22
Several clinical pharmacology studies have been conducted to investigate the pharmacokinetics (PK) of dasatinib, including the effect of food and the potential for drug-drug interactions in healthy volunteers or subjects with CML. Dasatinib exhibits variability in absorption following oral administration with a median (range) tmax of approximately 1 hour (0.5-6 hours) in subjects with CML. Dasatinib exposure was marginally decreased following a high-fat meal in healthy subjects. The maximum concentration (Cmax) and the area under the concentration versus time curve (AUC) of dasatinib decreased by 63% and 61%, respectively, when dasatinib was administered 10 hours following a 40-mg dose of famotidine, possibly due to the pH-dependent solubility of dasatinib.23 Dasatinib is extensively metabolized by CYP3A4 in humans. Coadministration of dasatinib with ketoconazole increased exposure of dasatinib by approximately 4- to 5-fold. Conversely, the exposure of dasatinib was decreased by approximately 80%, when it was coadministered with rifampin.23
The objectives of this modeling and simulation analysis were to identify key determinants of dasatinib exposure and its variability in CML and Ph+ ALL patients. In contrast with noncompartmental analysis, the model-based approach employed in this analysis not only permits characterization of exposure with only sparsely sampled data but also offers the potential to partition variability in exposure into variability in bioavailability and elimination. The characterization of exposure included quantification of the contribution of interoccasion (dose-to-dose) variability in relative bioavailability to overall exposure variability, which is particularly relevant for a drug such as dasatinib, given its pH-dependent bioavailability. A simulation study was performed to demonstrate the feasibility of quantifying interindividual and interoccasion variability in relative bioavailability, as well as the interindividual variability in apparent clearance.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Analysis Dataset
The data analyzed were derived from 1 phase I and 5 phase II studies, details of which have been previously reported.13-17 A summary description of the study designs and sampling schemes is presented in Table I. All studies were approved by an institutional review board (IRB) or independent ethics committee (IEC) and were carried out in accordance with the ethical principles of the Declaration of Helsinki (a list of the IRB/IECs is provided in Appendix A). Dasatinib doses in the phase I study ranged from 15 to 180 mg, and dosing schedules included qd and bid (continuous or 5 days on/2 days off) schedules, whereas the dosing regimen in the phase II studies was 70 mg bid.
|
Comedication variables. Concomitant medications that could potentially affect dasatinib exposure were classified into 4 broad categories: pH modifiers, CYP3A4 substrates, CYP3A4 inducers, and CYP3A4 inhibitors. pH modifiers were further categorized to antacids, H2-receptor antagonists, and proton-pump inhibitors. A concomitant medication was associated with dasatinib concentration samples collected on the day that the comedication was taken or within 3 days of the last comedication dose. Start and end dates of concomitant medications were captured in case report form (CRF), but dose amount and adherence were not available.
Bioanalytical Assay
Dasatinib was quantified in human plasma samples using a validated high-performance liquid chromatography/tandem mass spectrometry (HPLC/MS/MS) assay that had a calibration range of 1 to 1000 ng/mL. All values below the lower limit of quantification (LLOQ) of 1 ng/mL were excluded from the analysis. A stable isotope-labeled [13C4, 15N2] dasatinib was used as an internal standard. Human plasma containing dasatinib and the internal standard was extracted by solid-phase extraction, and the resulting extract was injected onto a SCIEX API 4000 mass spectrometer in positive ion mode with HPLC column. The bioanalytical method performance was precise and accurate for the analysis of dasatinib in human plasma in all 6 studies. The between-run variability and within-run variability for the analytical quality controls (QCs) of dasatinib were no greater than 8.2% coefficient of variation (CV) and 8.1% CV, respectively. The deviations of the mean observed concentrations from the nominal concentrations were no more than ±9.7%.
Characterization of Variability in Dasatinib Exposure
Dasatinib exposure was characterized by a nonlinear mixed-effects ("population") compartmental pharmacokinetic model. A linear 2-compartment structural model parameterized in terms of apparent clearance (CL/F), apparent intercompartmental clearance (Q/F), apparent volume of central compartment (Vc/F), and apparent volume of peripheral compartment (Vp/F) with additional parameters for absorption was chosen based on graphical analyses of dasatinib concentration-time profiles that revealed that disposition of dasatinib appeared to follow biexponential decay kinetics. The identification and quantification of sources of variability in dasatinib exposure were achieved by determining the following component models using previously described model selection criteria24: (1) absorption models, (2) interindividual and interoccasion variability models, and (3) covariates models (including the effect of concomitant mediations). The residual error was described by a lognormal model. All analyses were performed using the NONMEM computer program in Linux (Version V, Level 1.1, GloboMax, Hanover, Maryland). Diagnostic graphics were obtained using S-Plus software (Version 7.0 for Linux, Insightful, Seattle, Washington). The parameters in all exposure models were estimated with the NONMEM first-order conditional estimation with interaction (FOCEI) method.
Absorption Models
The absorption model was determined by identifying the most parsimonious description of the absorption process that adequately described dasatinib concentrations during the absorption phase. The absorption models tested included conventional zero-order absorption (estimate either duration or rate of infusion), first-order absorption kinetics, Weibull absorption model, and time-dependent models modified from the work of Higaki et al.25 Brief descriptions of the absorption models tested are given in Appendix B.
Interindividual and Interoccasion Variability Models
Interindividual variability (IIV) in structural model parameters was described by lognormal distributions. Furthermore, IIV on relative bioavailability (FR) was introduced to account for correlations between the structural model parameters other than those of the absorption model, as the values of all these parameters are expected to increase with a decrease in bioavailability. The population average value of FR was fixed at unity to ensure that the other model parameters were identifiable. This approach is more parsimonious than estimating covariances between all possible combinations of structural model parameters. The plasma concentrations of dasatinib were measured on more than 1 occasion in 134 subjects (34%) in the population exposure dataset. These data provide an opportunity to characterize the interoccasion variability in the kinetics of dasatinib and identify the model parameters that contribute to this variability. Model development included assessment of interoccasion variability (IOV) on CL/F and FR, using the methods described in the seminal article on the estimation of IOV by Karlsson and Sheiner.26
Covariate Models (Including Concomitant Medications)
The influence of the following time-invariant covariates on dasatinib exposure was investigated: body weight, age, gender, race, smoking status, disease status, prior imatinib treatment, dose, baseline ALT and AST, and creatinine clearance (CrCL). The effect of hemoglobin and white blood cell count was tested as time-varying covariates on dasatinib exposure because these covariates can vary markedly in patients with CML or Ph+ ALL. Furthermore, the potential influence of concomitant medications was modeled as dichotomous categorical time-varying covariates. The functional forms of model parameters–covariate relationships examined and covariate selection methods are provided in Appendix C.
Model Evaluation
The accuracy of the final model was evaluated by a quantitative predictive check,27,28 and the stability of the covariate was evaluated by nonparametric bootstrap as described elsewhere.29,30 The predictive performance of the model was assessed by statistics of trough concentrations (Cmin) observed between 11 and 12 hours postdose from days 5 to 29. The 25th percentile, median, and 75th percentile were calculated, and the probability that the observed statistics is less than the simulated statistics (Ppc) was calculated from 1000 simulated trials.
Identifiability of Variability in Relative Bioavailability
A simulation study was performed to confirm the identifiability of both IIV and IOV in FR, with data simulated from sampling design, sample size, and population models similar to that for dasatinib. The population model parameter values used in the simulations were identical to those of the dasatinib final model (see Table III), except for IIV/IOV in FR and IIV of the proportionality rate constant (KA), which were modified to facilitate the simulation analysis. Three different scenarios were simulated by changing the IIV and IOV in FR (scenario 1: 50% IOV and no IIV in FR; scenario 2: 50% IIV and no IOV in FR; and scenario 3: 30% IIV and IOV in FR). Furthermore, the IIV of KA was set to 0.49 (CV = 70%), which represents a reasonably high variability for an orally administered drug. For each scenario, 200 data sets were simulated, and each dataset contains concentration values from 400 subjects receiving 70 mg bid of dasatinib. Samples were simulated from each subject as follows: 3 samples were taken on each of 3 different occasions (days 1, 8, and 26), and at each occasion, 1 sample was taken within each of the following 3 time windows (random uniform distribution within a time window): 0.5 to 3 hours, 5 to 8 hours, and 11 to 12 hours. The simulated data sets were then analyzed using 4 different population models: (1) a model without IIV and IOV in FR, (2) a model with IIV in FR, (3) a model with IOV in FR, and (4) a model containing both IIV and IOV in FR. The bias and precision of parameter estimates were calculated as described previously.30
|
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Subject Characteristics
Summary statistics of the demographic and other covariate variables for the final exposure dataset are provided in Table II. The analysis dataset contained 4044 measurements of dasatinib plasma concentrations from 399 patients, with dasatinib concentration observations on 1 to 3 occasions in a given subject (data were available for 137, 387, and 50 subjects on day 1, day 5/8, and day 26, respectively). Approximately 4.7% of the available dasatinib concentration measurements were below LLOQ and were excluded from the analysis dataset.
|
Absorption Models
A constrained first-order absorption model was selected over the other absorption models as some of those models terminated; others gave unreasonable parameter estimates with large uncertainty. The absorption model with individual KA constrained to be greater than population KE (CL/VC) was found to be the most stable (consistently achieved convergence with bootstrapped data sets) and parsimonious model among the absorption models examined (Appendix B), including the model with individual KA constrained to be greater than individual KE. The variance of IIV on KA was fixed to 1.0 as this parameter could not be reliably estimated from the data, and a sensitivity analysis showed that values between 0.5 and 3.0 resulted in reasonable estimates of IIV on Vc/F and did not have a marked effect on the fixed-effect parameters.
|
Evaluation of the diagnostic plots supports the validity of the model to describe the variability of dasatinib exposure in the target population. The observed data were randomly scattered about the line of unity in diagnostic plots of both observed versus typical predicted and individual-predicted concentration plots. The distribution of IOV in FR at different occasions did not appear to have any trend with respect to time, and 25th to 75th percentile fell within 80% to 125% of the population average (median values: 97.6%, 104.0%, and 86.5% at Day 1, Day 5/8, and Day 26, respectively). The conditional weighted residual versus time (after the first dose) diagnostic plot indicates that determinants of dasatinib exposure are time invariant and that there is no evidence of autoinhibition or autoinduction of elimination. The effect of time on CL/F was assessed by a model in which CL/F was estimated separately for each visit (day 1, day 5/8, and day 26). There was a small increase of fixed-effect estimates of apparent clearance on day 5/8 and day 26 (+6.5% and +9.0% compared with day 1, respectively), indicating that clearance is time invariant. Neither apparent clearance nor relative bioavailability was related to dose, suggesting that dasatinib exposure is dose linear for doses examined in this analysis (15-180 mg).
Covariate Models (Including Concomitant Medications)
Results of the initial covariate analysis indicated that CL/F was dependent on Hb and disease status, and Vc/F was dependent on time-varying Hb concentration. In addition, FR appeared to be reduced by concomitant use of proton pump inhibitors. These covariates were all significant at a 1% level and therefore were included in the full model. The results of the backward deletion of covariates in the full model showed that only Hb on CL/F was significant at the 0.1% level. Therefore, time-varying Hb was the only covariate that was retained in the final model, the parameter values of which are presented in Table III. However, Hb does not appear to be a clinically relevant covariate, as the CL/F at the 5th and 95th percentiles of Hb were within ±20% of the CL/F at the reference Hb value of 10 g/L. No other demographic or clinical covariates were found to be significant. Assessment of dose on either FR or CL/F as a power function did not achieve statistical significance, suggesting that neither CL/F nor FR was related to dose.
Effect of Clinical Covariates on Dasatinib Exposure
Summary results of the effect of categorical and continuous covariates on CL/F are presented in Figure 2A and Figure 2B, respectively. For all covariates on CL/F, the estimated effect and 95% confidence intervals are contained within 80% to 120% of the population average estimate of CL/F. Hb was found to be a statistically significant covariate. The highest value of Hb resulted in the lowest estimates of CL/F. There was no clinically relevant relationship between Vc/F and all clinical covariates.
Effect of Concomitant Medications on Dasatinib Exposure
Concomitant administration of proton-pump inhibitors (PPIs; n = 92) was associated with a median decrease of 17% (95% confidence interval [CI]: 4%-30%) in FR. Although the point estimate falls within 80% to 120% of the population average estimate, the 95% CI of the effect exceeds 80% of the average population estimate and does not contains 100%, suggesting the effect may be clinically relevant. The effect of the other class of concomitant medications that could potentially modulate CYP3A4 activity on CL/F and FR was also examined but was found to be either not statistically significant or not clinically relevant. Coadministration of a CYP3A4 substrate resulted in a 12% (95% CI: 0%-21%) decrease in relative bioavailability, and coadministration of the CYP3A4 inhibitor resulted in 3% increase in relative bioavailability. The effect of concomitant CYP3A4 inducers was not examined as only 3 subjects received CYP3A4 inducers.
|
|
Identifiability of Variability in Relative Bioavailability
The fixed-effect parameters were estimated with high precision and low bias for all scenarios (median absolute error was less than 15%, and interquartile range of bias was contained within ±20% for all parameters; fixed-effect parameters not shown). The distribution of bias (percentage of true value) of the parameter estimates is shown in Figure 4 for random-effect parameter estimates. Simulation scenarios 1 and 2 (ie, IOV in FR and IIV in FR) were analyzed with 3 different analytical models (a model with IIV in FR, a model with IOV in FR, and a model including both IIV and IOV in FR). For simulation scenarios 1 and 2 (IOV in FR, IIV in FR), the IOV or IIV was well estimated with <20% of bias when models were correctly specified (data not shown).
Scenario 3 was the motivation of this simulation study and may be applicable to some orally administered anticancer agents that exhibit absorption-related variability in exposure. When both IIV and IOV were included (scenario 3), distributions of bias of all fixed-effect parameter estimates with the exception of KA were within ±20% and were independent of the presence of IIV or IOV in the model (data not shown). However, bias in the random-effect estimates appears to be a more sensitive indicator compared with that in the fixed-effect estimates. As shown in Figure 4, the model accounting for both IIV and IOV in FR produced unbiased estimates for all random-effect parameters with exception of IIV for KA. Models that ignore IIV and IOV random effects or IOV when IOV is present produced inflated estimates of residual error with 82% and 77% of median bias, respectively. Furthermore, estimates of IIV for CL/F had large bias when IIV or IOV was not taken into consideration. The model with the inclusion of only IOV in FR performed slightly worse than the model accounting for both IIV and IOV in that bias of IIV for CL/F obtained from the model with only IOV was larger than 30%, and the CV (precision) was larger than 25% (precision data not shown) in contrast to less than ±30% bias and 25% precision obtained from the model with inclusion of both IIV and IOV. But all other estimates of fixed- and random-effect parameters were relatively comparable to that obtained from the model incorporating IIV and IOV.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Variability in dasatinib exposure appears to be mainly due to variability in bioavailability. A major goal of the population exposure analysis was to characterize dasatinib exposure and to gain an understanding of the sources of variability in dasatinib exposure. IOV and IIV in FR were included in the final model to explain the dose-to-dose and interindividual variability in exposure. Inclusion of IOV on FR provided a better description of the data than inclusion of IOV on CL/F or on Vc/F. Although it is likely that IOV may be present in all of these parameters, only IOV on FR was retained in the model in the interest of parsimony. This is similar to the principle whereby IIV is only included on some parameters in the model, although it is reasonable to expect that some degree of IIV is present in all parameters. Inclusion of IIV and IOV on relative bioavailability (FR) was recently reported. In some cases, FR was introduced to account for differences in different formulations or explain food effect or diurnal variability in absorption.31-35 In other cases, variability in FR was introduced to quantify the sources of variability in exposure, such as race,36 concomitant medication,37 or other extrinsic factors.11,38
Previous noncompartmental analyses suggested that interindividual variability in dasatinib exposure could be high. The variability of AUC0-12 h for bid continuous dosing regimens in the phase I study ranged approximately from 32% to 118% (unpublished results). In principle, the 2 sources of variability in AUC0-12 h and apparent clearance are variability in total clearance and variability in bioavailability. Although neither total clearance nor absolute bioavailability can be estimated without data from intravenously administered drug, it is possible to estimate the extent to which variability in bioavailability contributes to the total variability in apparent clearance.
In the present analysis, the interindividual variability in CL/F (adjusted for FR) was 25%, and the IOV and IIV in FR were 44% and 32%, respectively. The model did not include IOV on CL/F or Vc/F as the model of IOV on FR provided the best description of data. This suggests that the variability in CL/F may be mainly attributed to variability in relative bioavailability and that the variability in FR is greater within subjects than between subjects. It is worthwhile to note that interoccasion variability in dasatinib exposure is not expected to be as clinically relevant as interindividual variability in exposure, as prolonged exposure to dasatinib is generally needed to achieve clinically efficacious effects, and dose-to-dose variability in exposure averages out over repeated twice-daily dosing. A possible mechanistic reason for interoccasion variability in FR could be variable dissolution of dasatinib due to interoccasion variability in gastrointestinal pH, which could be caused by extrinsic factors such as concomitant medication. Another extrinsic factor that may explain the interoccasion variability in FR is variable adherence to the prescribed dose at the designated time.
Previously, Karlsson and Sheiner26 and Lalonde et al39 have reported on the importance of modeling IOV with pharmacokinetic data and pharmacodynamic data, respectively. However, previous studies do not address whether IIV and IOV in FR can be reliably identified in the absence of intravenous data. Here, we performed a simulation study to demonstrate that it is feasible to obtain reliable estimates for both IIV and IOV in FR given an adequate sparse sampling scheme. It is evident from our simulation study that ignoring the presence of IIV and IOV may lead to inflated residual error and result in overestimates of the IIV in CL/F. Furthermore, ignoring the presence of IIV and IOV may overestimate IIV for CL/F. By accounting for both IIV and IOV in FR, the final model provided more accurate estimates of IIV for CL/F. In situations where IOV is much larger than IIV, it might be reasonable to estimate IOV in FR as an alternative to the model with both IIV and IOV in FR as estimates of all fixed- and random-effect parameters, except IIV for CL/F was comparable to the model with IIV and IOV.
None of the tested clinical covariates contributed to variability in dasatinib exposure. In this analysis, Hb was the only covariate with a statistically significant effect on CL/F. A physiologically plausible reason for this finding could be that higher blood Hb is associated with a smaller volume of plasma per unit volume of blood, thus resulting in a smaller amount of drug that is extractable by the liver for a given hepatic blood flow rate. The estimated magnitude of the Hb effect was 8.5% and –17.7% at the 5th to 95th percentiles of Hb values in the analysis dataset, respectively. Although the magnitude of this effect does not appear to be clinically relevant, an effect greater than 20% cannot be ruled out because of the uncertainty in the parameter estimate. Investigation of other covariates in this analysis revealed no trends with respect to CL/F or Vc/F. As shown in Figure 2, the effect of all covariates tested on CL/F or Vc/F is within ±20% of the population average parameter estimate without the covariate when based on the point estimate of the covariate parameter. Based on this population exposure analysis, dasatinib can be administered without dose adjustment for age, body weight, gender, race, or smoking status.
Concomitant medications did not explain the variability in relative bioavailability. The effect of comedications was examined to quantify the extent to which concomitant medication can explain the variability in relative bioavailability because marked effects of pH modulators (H2-receptor antagonist), as well as CYP3A4 inhibition and induction, were observed. In the population exposure dataset, only 3 subjects (<1% of total subjects) were on antacids and 10 subjects (2.5%) on H2-receptor antagonist, so it is not surprising that no relationship between either antacids or H2-receptor blockers and relative bioavailability was discerned. Coadministration with PPIs decreased relative bioavailability of dasatinib by 17%, but this effect was not statistically significant. However there is a 30% decrease in bioavailability at the lower bound of the 95% confidence interval of the effect of PPIs, suggesting that the effect of PPIs could be clinically relevant.
The absence of a pronounced effect of concomitant medications on either apparent clearance or relative bioavailability suggests that a population-based assessment of drug-drug interactions may not provide a definitive assessment of drug-drug interactions due to inadequate sample size, interactions among multiple comedications, and imprecise dosing data for the concomitant medication. Nonetheless, a population analysis could provide an assessment of the magnitude of the effect under real-world conditions, with inhibitors and inducers that are less potent than those employed in a classical drug-drug interaction study, by prospectively designing the study, defining drug-drug interactions of interest, and proper recording of interacting drug dosage, frequency, and time.40 In the case of dasatinib, there are several possible reasons the pronounced effect of comedications on exposure was not discerned. First, drug-drug interaction studies of dasatinib were conducted with potent interacting probes given at high doses. In contrast, lower doses and shorter administration of either inhibitor or substrate might be common in phase II outpatient studies. Second, the use of multiple counteracting medications in this population reduces the power of the population-based analyses to detect the effect of an individual class of concomitant medications. Approximately 50% of subjects on 1 of the counteracting medications (such as CYP3A4 inhibitors and pH modifiers) are on 2 counteracting medications, and approximately 20% of subjects are on a combination of 3 medications. Finally, missing information on the actual time and dose regimen of administration of concomitant medications could have rendered this population-based analysis insensitive to drug-drug interactions. For instance, 1 drug-drug interaction study investigating the effect of pH modulators has clearly demonstrated the importance of time of administration of concomitant medications relative to that of dasatinib. In this study, administration of 30 mL of aluminum hydroxide/magnesium hydroxide (Maalox) 2 hours prior to a single dose of dasatinib resulted in no relevant change in dasatinib AUC and higher Cmax (26%). In contrast, the reduction of dasatinib exposure was pronounced (approximately 55%) when Maalox was administered simultaneously with dasatinib.23 This result indicated that dasatinib exposure would not be decreased when coadministered with antacids if the doses are temporally separated by 2 hours. Therefore, it is not uncommon that drug-drug interaction effects observed in clinical pharmacology studies are not fully characterized in population-based studies. Similar findings have been reported for imatinib and Tarceva.41,42
In conclusion, our analysis shows that a substantial portion of variability in dasatinib exposure can be explained by IOV in FR. Interoccasion variability may be less clinically relevant than IIV for chronically administered drugs such as dasatinib, as the random IOV in exposure is not expected to affect the cumulative exposure within a subject.
|
|
|
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Financial disclosure: None declared.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
1. Sheiner LB. Learning versus confirming in clinical drug development. Clin Pharmacol Ther. 1997;61: 275-291.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
2. Bhattaram VA, Bonapace C, Chilukuri DM, et al. Impact of pharmacometric reviews on new drug approval and labeling decisions: a survey of 31 new drug applications submitted between 2005 and 2006. Clin Pharmacol Ther. 2007;81: 213-221.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
3. Lesko LJ. Paving the critical path: how can clinical pharmacology help achieve the vision? Clin Pharmacol Ther. 2007;81: 170-177.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
4. Barrett JS, Labbe L, Pfister M. Application and impact of population pharmacokinetics in the assessment of antiretroviral pharmacotherapy. Clin Pharmacokinet. 2005;44: 591-625.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
5. Friberg LE, Henningsson A, Maas H, Nguyen L, Karlsson MO. Model of chemotherapy-induced myelosuppression with parameter consistency across drugs. J Clin Oncol. 2002;20: 4713-4721.
6. Latz JE, Rusthoven JJ, Karlsson MO, Ghosh A, Johnson RD. Clinical application of a semimechanistic-physiologic population PK/PD model for neutropenia following pemetrexed therapy. Cancer Chemother Pharmacol. 2006;57: 427-435.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
7. Gurney H. Dose calculation of anticancer drugs: a review of the current practice and introduction of an alternative. J Clin Oncol. 1996;14: 2590-2611.[Abstract]
8. Miller AA, Tolley EA, Niell HB. Therapeutic drug monitoring of 21-day oral etoposide in patients with advanced non-small cell lung cancer. Clin Cancer Res. 1998;4: 1705-1710.[Abstract]
9. Chatelut E, Canal P, Brunner V, et al. Prediction of carboplatin clearance from standard morphological and biological patient characteristics. J Natl Cancer Inst. 1995;87: 573-580.
10. Reigner B, Blesch K, Weidekamm E. Clinical pharmacokinetics of capecitabine. Clin Pharmacokinet. 2001;40: 85-104.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
11. Leger F, Loos WJ, Fourcade J, et al. Factors affecting pharmacokinetic variability of oral topotecan: a population analysis. Br J Cancer. 2004;90: 343-347.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
12. de Jonge ME, Huitema AD, Schellens JH, Rodenhuis S, Beijnen JH. Individualised cancer chemotherapy: strategies and performance of prospective studies on therapeutic drug monitoring with dose adaptation: a review. Clin Pharmacokinet. 2005;44: 147-173.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
13. Talpaz M, Shah NP, Kantarjian H, et al. Dasatinib in imatinib-resistant Philadelphia chromosome-positive leukemias. N Engl J Med. 2006;354: 2531-2541.
14. Cortes J, Rousselot P, Kim DW, et al. Dasatinib induces complete hematologic and cytogenetic responses in patients with imatinib-resistant or -intolerant chronic myeloid leukemia in blast crisis. Blood. 2007;109: 3207-3213.
15. Hochhaus A, Kantarjian HM, Baccarani M, et al. Dasatinib induces notable hematologic and cytogenetic responses in chronic phase chronic myeloid leukemia after failure of imatinib therapy. Blood. 2007;109: 2303-2309.
16. Guilhot F, Apperley J, Kim DW, et al. Dasatinib induces significant hematologic and cytogenetic responses in patients with imatinib-resistant or -intolerant chronic myeloid leukemia in accelerated phase. Blood. 2007;109: 4143-4150.
17. Quintas-Cardama A, Kantarjian H, Jones D, et al. Dasatinib (BMS-354825) is active in Philadelphia chromosome-positive chronic myelogenous leukemia after imatinib and nilotinib (AMN107) therapy failure. Blood. 2007;109: 497-499.
18. Shah NP, Tran C, Lee FY, Chen P, Norris D, Sawyers CL. Overriding imatinib resistance with a novel ABL kinase inhibitor. Science. 2004;305: 399-401.
19. Tokarski JS, Newitt JA, Chang CY, et al. The structure of dasatinib (BMS-354825) bound to activated ABL kinase domain elucidates its inhibitory activity against imatinib-resistant ABL mutants. Cancer Res. 2006;66: 5790-5797.
20. Lombardo LJ, Lee FY, Chen P, et al. Discovery of N-(2-chloro-6-methyl-phenyl)-2-(6-(4-(2-hydroxyethyl)-piperazin-1-yl)-2-methylpyrimidin-4-ylamino)thiazole-5-carboxamide (BMS-354825), a dual Src/Abl kinase inhibitor with potent antitumor activity in preclinical assays. J Med Chem. 2004;47: 6658-6661.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
21. Donato NJ, Wu JY, Stapley J, et al. Imatinib mesylate resistance through BCR-ABL independence in chronic myelogenous leukemia. Cancer Res. 2004;64: 672-677.
22. O'Hare T, Walters DK, Stoffregen EP, et al. In vitro activity of Bcr-Abl inhibitors AMN107 and BMS-354825 against clinically relevant imatinib-resistant Abl kinase domain mutants. Cancer Res. 2005;65: 4500-4505.
23. SPRYCEL® (dasatinib) [package insert]. Princeton, NJ: Bristol-Myers Squibb Company; July 2006.
24. Pfister M, Martin NE, Haskell LP, Barrett JS. Optimizing dose selection with modeling and simulation: application to the vasopeptidase inhibitor M100240. J Clin Pharmacol. 2004;44: 621-631.
25. Higaki K, Yamashita S, Amidon GL. Time-dependent oral absorption models. J Pharmacokinet Pharmacodyn. 2001;28: 109-128.[Web of Science][Medline] [Order article via Infotrieve]
26. Karlsson MO, Sheiner LB. The importance of modeling interoccasion variability in population pharmacokinetic analyses. J Pharmacokinet Biopharm. 1993;21: 735-750.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
27. Yano Y, Beal SL, Sheiner LB. Evaluating pharmacokinetic/pharmacodynamic models using the posterior predictive check. J Pharmacokinet Pharmacodyn. 2001;28: 171-192.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
28. Food and Drug Administration (FDA). Guidance for Industry: Population Pharmacokinetics. Rockville, MD: Center for Drug Evaluation and Research (CDER) and Center for Biologics Evaluation and Research (CBER), FDA; 1999.
29. Ette EI, Williams PJ, Kim YH, Lane JR, Liu MJ, Capparelli EV. Model appropriateness and population pharmacokinetic modeling. J Clin Pharmacol. 2003;43: 610-623.
30. Roy A, Ette EI. A pragmatic approach to the design of population pharmacokinetic studies. AAPS J. 2005;7: E408-E420.[CrossRef][Medline] [Order article via Infotrieve]
31. Hossain M, Wright E, Baweja R, Ludden T, Miller R. Nonlinear mixed effects modeling of single dose and multiple dose data for an immediate release (IR) and a controlled release (CR) dosage form of alprazolam. Pharm Res. 1997;14: 309-315.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
32. Hennig S, Wainwright CE, Bell SC, Miller H, Friberg LE, Charles BG. Population pharmacokinetics of itraconazole and its active metabolite hydroxy-itraconazole in paediatric cystic fibrosis and bone marrow transplant patients. Clin Pharmacokinet. 2006;45: 1099-1114.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
33. Wade JR, Snoeck E, Duff F, Lamb M, Jorga K. Pharmacokinetics of ribavirin in patients with hepatitis C virus. Br J Clin Pharmacol. 2006;62: 710-714.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
34. Hennig S, Waterhouse TH, Bell SC, et al. A d-optimal designed population pharmacokinetic study of oral itraconazole in adult cystic fibrosis patients. Br J Clin Pharmacol. 2007;63: 438-450.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
35. Othman AA, Tenero DM, Boyle DA, Eddington ND, Fossler MJ. Population pharmacokinetics of S(–)-carvedilol in healthy volunteers after administration of the immediate-release (IR) and the new controlled-release (CR) dosage forms of the racemate. AAPS J. 2007;9: E208-E218.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
36. Kappelhoff BS, Huitema AD, Yalvac Z, et al. Population pharmacokinetics of efavirenz in an unselected cohort of HIV-1-infected individuals. Clin Pharmacokinet. 2005;44: 849-861.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
37. Kappelhoff BS, Huitema AD, Crommentuyn KM, et al. Development and validation of a population pharmacokinetic model for ritonavir used as a booster or as an antiviral agent in HIV-1-infected patients. Br J Clin Pharmacol. 2005;59: 174-182.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
38. Crommentuyn KM, Kappelhoff BS, Mulder JW, et al. Population pharmacokinetics of lopinavir in combination with ritonavir in HIV-1-infected patients. Br J Clin Pharmacol. 2005;60: 378-389.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
39. Lalonde RL, Ouellet D, Kimanani EK, Potvin D, Vaughan LM, Hill MR. Comparison of different methods to evaluate population dose-response and relative potency: importance of interoccasion variability. J Pharmacokinet Biopharm. 1999;27: 67-83.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
40. Zhou H. Population-based assessments of clinical drug-drug interactions: qualitative indices or quantitative measures? J Clin Pharmacol. 2006;46: 1268-1289.
41. Schmidli H, Peng B, Riviere GJ, et al. Population pharmacokinetics of imatinib mesylate in patients with chronic-phase chronic myeloid leukaemia: results of a phase III study. Br J Clin Pharmacol. 2005;60: 35-44.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
42. Lu JF, Eppler SM, Wolf J, et al. Clinical pharmacokinetics of erlotinib in patients with solid tumors and exposure-safety relationship in patients with non-small cell lung cancer. Clin Pharmacol Ther. 2006;80: 136-145.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
43. Zhou H, Khalilieh S, Lau H, et al. Effect of meal timing not critical for the pharmacokinetics of tegaserod (HTF 919). J Clin Pharmacol. 1999;39: 911-919.[Abstract]
44. Wahlby U, Thomson AH, Milligan PA, Karlsson MO. Models for time-varying covariates in population pharmacokinetic-pharmacodynamic analysis. Br J Clin Pharmacol. 2004;58: 367-377.[CrossRef][Web of Science][Medline]
[Order article via Infotrieve]
CiteULike 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
