| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
PHARMACOKINETICS AND PHARMACODYNAMICS |
From Pharmacyclics Inc, Sunnyvale, California (Dr Miles, Dr Smith, Dr Phan, Dr Renschler, Dr Boswell); Quintiles Inc, Kansas City, Missouri (Dr Hutcheson); and UCLA Medical Center, Radiation Oncology, Los Angeles, California (Dr Ford).
Address for reprints: Dale R. Miles, Pharmacyclics Inc, 995 E. Arques Avenue, Sunnyvale, CA 94085.
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXREFERENCES |
|---|
13% from the population mean CL or V1 estimate. It was concluded that a 3-compartment, open, POP-PK model predicts small but significant effects of age, sex, alkaline phosphatase, hemoglobin, serum creatinine, and phenytoin on MGd pharmacokinetics.
Key Words: Motexafin gadolinium • population pharmacokinetics • brain metastases • glioblastoma multiforme
Motexafin gadolinium is detectable by magnetic resonance imaging (MRI) due to the presence of Gd(III) in the molecule. MRI studies in patients have shown that motexafin gadolinium selectively localizes in brain metastases compared with normal surrounding brain tissue.3 Motexafin gadolinium is postulated to modulate tumor cell resistance to radiation therapy by depleting intracellular-reducing antioxidants and forming reactive oxygen species.4
Motexafin gadolinium pharmacokinetics have been evaluated in early phase I and phase II clinical trials. Inductively coupled plasma atomic emission spectroscopy (ICP-AES) was used to measure the total amount of gadolinium in plasma. Plasma gadolinium concentrations were converted to µg-equivalents of motexafin gadolinium prior to pharmacokinetic analysis. In these studies, motexafin gadolinium pharmacokinetics exhibited multicompartmental behavior. In a phase I dose escalation trial, a linear increase in area under the curve from 1 to 24 hours (AUC1-24 h) with dose was observed over a dose range of 5.4 to 29.6 mg/kg.5 In a phase II multidose trial, the plasma pharmacokinetics of motexafin gadolinium was evaluated in a group of 20 patients receiving a median dose (range) of 5.5 mg/kg (5.0-6.5 mg/kg). The value of the maximum observed plasma concentration was 62.5 µg-equivalents/mL. For the dose range studied, there was no statistically significant relationship between maximum concentration or area under the curve and dose in the phase II study.6 When 14C-labeled motexafin gadolinium (1 mg/kg) was administered to healthy human subjects in a mass balance study, approximately 44% of the administered radioactivity was recovered in the feces, and approximately 26% was recovered in the urine (unpublished data).
To better understand the effect of patient characteristics and concomitant medications on the plasma pharmacokinetics of motexafin gadolinium, a population pharmacokinetic (POP-PK) analysis was performed on pooled data from a recent clinical phase III trial2 in patients with metastatic brain tumors and in 2 other trials enrolling patients with primary brain tumors (glioblastoma multiforme, GBM). In these studies, a liquid chromatography/mass spectroscopy/mass spectroscopy (LC/MS/MS) assay specific for motexafin gadolinium was used to measure plasma concentrations. The objectives of this study were to characterize the POP-PK behavior of motexafin gadolinium in patients with brain metastases or GBM; evaluate the potential effects of demographic variables, disease state, concomitant medications, and clinical laboratory values on the plasma pharmacokinetics of motexafin gadolinium; and characterize the performance of the POP-PK model in the prediction of motexafin gadolinium plasma concentrations.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXREFERENCES |
|---|
|
The patient exclusion criteria employed in each clinical trial are shown in Table II. The median infusion time for all motexafin gadolinium infusions was 20 minutes (5th-95th percentiles: 10-35 minutes). Some patients were hydrated with intravenous fluids prior to administration of motexafin gadolinium. Whole or partial brain radiation therapy (1.8-3.0 Gy/d) was administered 2 to 5 hours after each motexafin gadolinium infusion (see Table I) and on nondosing days (excluding weekends) during the treatment period. The exact time was recorded for each infusion and sample collection, and these times were used for the pharmacokinetic analysis.
|
Brain metastases patients with evidence of edema were placed on dexamethasone at the time of diagnosis without taper until whole brain radiation therapy was complete and then tapered thereafter. GBM patients were given dexamethasone depending on their clinical status. Patients with nausea or vomiting after the first administration of motexafin gadolinium were medicated with antiemetics (ondansetron, prochlorperazine, metoclopramide, famotidine, and/or granisetron) for subsequent administrations. Anticonvulsants (phenytoin and carbamazepine) were administered to some patients for the treatment or prevention of seizures. Other medications given to some patients included ranitidine, omeprazole, lorazepam, diphenhydramine, docusate, fluconazole, nystatin, oxycodone, acetaminophen, and ibuprofen.
Bioanalytical Methods
Whole blood was collected into collection tubes containing K3EDTA and the plasma isolated by centrifugation. Plasma was stored at –70°C until the time of analysis. At analysis, internal standard (a motexafin gadolinium structural analog synthesized and qualified at Pharmacyclics Inc, Sunnyvale, Calif) was added, and protein was precipitated using 4 volumes of acetonitrile. LC/MS/MS analysis was performed by injecting 5 µL supernatant onto a Zorbax Eclipse XDB C-18 column (150 x 3.0 mm, 3.5 µm) at a flow rate of 0.5 mL/min with isocratic elution using 60/40 v/v acetonitrile/100 mM ammonium acetate (adjusted to pH 4.3 using glacial acetic acid). Mass spectroscopy detection was done by monitoring 1089.6 
1089.6, 600 ms. The method had a linear range of 0.1 to 10.0 µg/mL.
During validation of the method, interday precision (percent coefficient of variation [%CV]) was 3.1%, 3.0%, and 5.3% at the low (0.3 µg/mL), mid (1.5 µg/mL), and high (7.5 µg/mL) drug concentration levels, respectively, with corresponding interday accuracy at each level of 108.7%, 107.5%, and 101.6% of nominal. The intraday precision (%CV) was 4.7%, and the intraday accuracy was 96.3% to 107.3% of nominal. MGd in unprocessed K3EDTA plasma was demonstrated to be stable (98%-102% of control) on the benchtop for at least 4 hours at room temperature, stable (97%-100% of control) after 3 freeze/thaw cycles, and stable (86%-94% of control) for 116 weeks when stored at –70°C. Processed samples were stable (97%-102% of control) for at least 153 hours in the autosampler and stable (97%-100% of control) for at least 149 hours in refrigerated (2-8°C) storage. The extraction recovery from K3EDTA plasma was 92% for MGd and 90% for the internal standard (PCI-0350). Ruggedness was demonstrated by validating the method on 2 LC/MS/MS platforms using 2 analysts and by showing that stable retention times could be achieved after injecting 128 samples as part of a single validation run.
A drug interference study was done by incubating plasma containing MGd at 3 concentrations spanning the linear range of the MGd assay for 2 hours in the presence of estimated Cmax levels of the following medications: dexamethasone, ranitidine, phenytoin, acetaminophen, ondansetron, prochlorperazine, metoclopramide, lorazepam, gadodiamide, diphenhydramine, granisetron, fluconazole, and gadolinium chloride. For each medication tested, the mean MGd concentration was within ±8% of the nominal concentration at each concentration level. Therefore, the LC/MS/MS assay for motexafin gadolinium was not significantly affected by these medications.
Data Analysis Methods
Data File Creation
A NONMEM data file was created from the existing clinical and bioanalytical databases for the purpose of this POP-PK analysis.7 Variables included dose, dosing time, sex, race, disease state, smoking status, alcohol use, and all of the variables listed in Table III. For patients with missing observations, a null value was assigned. For covariates measured intermittently throughout the study (eg, clinical laboratory parameters), the most recent prior measured value of the covariate was assigned to each plasma concentration record. For covariates measured the same day that plasma motexafin gadolinium concentration measurements were taken, the covariate value measured that day was used for all concentration records for that day. If no covariate measurement was taken prior to a plasma sample, then the next covariate measurement was assigned to that sample. Plasma samples with values measured below the lower limit of quantitation were excluded from the POP-PK analysis.
|
Concomitant medications were evaluated for an effect on motexafin gadolinium pharmacokinetics by grouping medications known to function as substrates/inhibitors or inducers of cytochrome P450 (CYP450) isozymes according to 6 categories: 1A2, 2C9, 2C19, 2D6, 2E1, and 3A4, 5, or 7. For medications that affected multiple CYP450 isozymes or functioned as both substrate/inhibitor and inducer, the medication was included in each category that was affected.
For each of the CYP450 isozyme categories identified above, an evaluation was made to determine if 1 or more concomitant mediations designated as an inducer for that category was given 3 to 7 days prior to the first motexafin gadolinium dose. If so, all of the concentration records for that patient were coded as positive for the presence of an inducer for that pathway. A similar evaluation was made for medications considered a substrate/inhibitor. However, only the concentration records less than 24 hours after administration of a substrate/inhibitor of a given isozyme category were coded as positive for that category.
Model Building
Concentration-time data were analyzed by nonlinear mixed-effects modeling using NONMEM, Version V, Level 1.0 (GloboMax LLC, Hanover, Md), to develop an appropriate POP-PK model using the first-order approximation method.7 NONMEM was run on a Hewlett-Packard UNIX production server equipped with a Fortran 77 compiler and running Apache Web Server Software, Version 1.3.6, and PERL, Version 5.6.0.
Preliminary NONMEM models examined were the 1-, 2-, and 3-compartment pharmacokinetic models after intravenous (IV) dose administration. Based on the NONMEM objective function value, the 3-compartment IV model (Advan 11, Trans 4) was used for further model building.
Interpatient variability was modeled assuming proportional error:
 | (1) |
is the estimate of the population mean, and 
j is the deviation from the population mean for the jth patient under the assumption that 
.8-10 Interindividual variability was included on clearance (CL) and volume of the central compartment (V1). Addition of interindividual variability on other model parameters was investigated but did not improve the model. Interoccasional variability was not explored.
Residual variability was modeled as a proportional error:
 | (2) |
is the residual proportional error term under the assumption that 
. Ten sample observation records from 9 patients were marked for exclusion as pharmacokinetic outliers; the absolute value of the NONMEM weighted residual was >4. This represented less than 0.7% of the sample observation data. After removal of outliers, combinations of proportional and/or additive error models were evaluated to characterize the distribution of residual variability. These alternative models degraded model performance, gave large standard errors for the error term, and/or resulted in high parameter correlation. Therefore, the proportional error models initially proposed were used for the final analysis.
Clearance and V1 demonstrated a statistically significant correlation (P < .0001). A term for pharmacokinetic parameter correlation was subsequently incorporated into the model, resulting in significant model improvement.
Evaluation of Covariates
Once the base model was built, individual patient pharmacokinetic parameters were calculated by the POSTHOC technique of NONMEM. Covariates tested for inclusion in the final model included disease type (ie, brain metastases or glioblastoma multiforme), concomitant medications, clinical laboratory values, and demographic characteristics. Additional covariates that were also evaluated in brain metastases patients from study A were primary tumor type (non-small-cell lung cancer, breast cancer, other/unknown primary tumor type) and number of prior cytotoxic/chemotherapy treatment regimens.
Covariate screening was performed using SAS (Version 6.12 for Windows 95).11 Continuous and binary covariates were screened for potential impact on POSTHOC CL and V1 estimates using stepwise linear regression. Categorical covariates were evaluated using general linear models. Scatter plots of plasma concentrations versus covariate overlaid with a nonparametric locally weighted scatter plot smoother (LOESS) were also used to help identify functional relationships. For categorical covariates, box and whisker plots of pharmacokinetic parameters for each of the groups were also used to identify differences between groups.
The final NONMEM model was built by adding potential covariates identified in the screening process to the base model expression incrementally. Each potential covariate was tested by NONMEM using 4 evaluation criteria:
After all significant covariates identified in the screening step were incorporated into the model, the remaining covariates (ie, those not identified as significant during screening) were evaluated for inclusion in the NONMEM model. Covariates that demonstrated significant population model improvement were incorporated into the updated base model, starting with the covariate that gave the greatest significant improvement. The model was then examined for parsimony by removing covariate terms (backward elimination), one at a time, to determine if the covariate resulted in significant model improvement (P < .005).
Following identification of the final model, the model was processed using the first-order conditional estimation method with interaction to generate individual patient posterior conditional POP-PK parameters.
Model Validation
An out-of-sample method was used to quantitate the predictive performance of the POP-PK model. Pharmacokinetic parameters for the omitted patient were then determined using the typical (population) parameter estimates obtained from modeling the remaining 242 patients. These parameter estimates were used to estimate motexafin gadolinium plasma concentrations for that patient.
The standardized mean prediction error (SMPE) statistic12 was computed by determining for each patient from the mean prediction error (MPE) and the standard deviation of the prediction errors (SDn(MPE)) as the following:
 | (3) |
SMPE is a measure of the number of standard deviations the MPE is from 0. For unbiased estimates, SMPE has an expected value of 0 and a standard deviation of 1. The mean SMPE and the 95% confidence interval for mean SMPE (for the out-of-sample patient observations) were computed. A bootstrap technique was used to develop the 95% confidence interval for the standard deviation of SMPE. The POP-PK model mean squared error (MSE) was computed to evaluate the precision of the model predictions.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXREFERENCES |
|---|
|
Evaluation of Covariates
The distribution of selected covariates is shown in Table III. Table V summarizes the covariate screening results obtained by stepwise linear regression and general linear models. For stepwise linear regression analysis, 9 covariates were deemed statistically significant (P < .05) for influencing untransformed CL, 3 covariates significant (P < .05) for untransformed V1, and 7 covariates significant (P < .05) for untransformed weighted residuals. Five of the 7 covariates identified as significant for influencing weighted residuals were also identified as potential covariates for either CL or V1.
|
Analysis of variance using general linear models revealed 3 covariates deemed statistically significant (
= 0.05) for influencing CL estimates, 2 covariates significant ( = 0.05) for influencing V1 estimates, and 5 covariates significant ( = 0.05) for influencing weighted residuals.
The covariates identified during the screening process were carried forward for further evaluation. Five of these covariates were found to significantly affect either CL or V1 in NONMEM: alkaline phosphatase; hemoglobin; sex; presence of CYP3A4, 5, or 7 inducer (phenytoin, n = 46 patients); and serum creatinine. In all cases, removal of the covariate resulted in significant model degradation. Therefore, no covariates were removed during the backward elimination step. The fully parameterized model was designated as the final POP-PK model.
NONMEM Population Parameter Estimates
All covariates with a P value .005 were retained in the model (see Table VI), and the final model resulted in good prediction of plasma concentrations (see Figures 1 and 2). Based on the POP-PK model parameters (and excluding outliers), the estimated value for CL was 1640 mL/h for a male patient and 1405 mL/h for a female patient with average levels of alkaline phosphatase (95.1 U/L), average levels of hemoglobin (13.3 g/dL), and average age (57 years). Estimated CL was increased by 30% for patients receiving medications classified as CYP450 3A4, 5, or 7 inducers. However, careful evaluation of the medications given in the trial suggests that this effect can almost entirely be attributed to phenytoin. Estimated CL decreases with increasing levels of alkaline phosphatase and increases with increasing levels of hemoglobin. Estimated CL decreases with age.
|
|
|
The estimated value for V1 was 7050 mL for a male patient and 5584 mL for a female patient with average levels of serum creatinine (0.75 mg/dL). Estimated V1 increased with increasing levels of serum creatinine. The POP-PK model was not improved with the addition of terms for either primary tumor type or number of cytotoxic/chemotherapy treatments. When patient covariates were included in the model, the interindividual variability was 24% for CL, 25% for V1, and 20% for CL:V1. The residual variability for the final model was 27%, suggesting that there was minimal model misspecification. The POSTHOC values obtained from the modeled patient data were consistent with the estimated model parameter estimates obtained in the FINAL POP-PK model (see Table VII).
|
Overall Model Bias and Predictability
The overall averages of the typical (population) parameter estimated values computed during the out-of-sample analysis process (using Method = FO) were 1439 mL/h for CL and 5867 mL for V1. These values compared favorably with those obtained from the FINAL POP-PK model and the mean POSTHOC patient parameter estimates (see Tables VI and VII). The strong correlation between actual concentrations and the out-of-sample predicted concentrations indicate that the model is well conditioned to predict the plasma concentrations for patients not included in the model parameter optimization process. The model was unbiased, with a slight tendency toward overestimating plasma concentrations.
When the model predictability was evaluated using standardized mean prediction errors, the observed results were consistent with an unbiased model. The average (95% confidence interval) standardized mean prediction error (SMPE) for the out-of-sample patient data was –0.103 (–0.227 to 0.021). The average (95% confidence interval) value of the standard deviation of the SMPE was 0.977 (0.864-1.110).
In the FINAL model, the residual modeling error was estimated at approximately 25%. The root mean square error (RMSE) for the out-of-sample prediction errors was 6.49 µg/mL. To evaluate the ability for the model to predict plasma concentrations across the wide range of observed plasma concentrations, the data were stratified according to observed plasma concentration values, and the relative RMSE (RMSE x 100/average concentration) was computed for each of these strata. For plasma concentration levels greater than 10 µg/mL, the relative RMSE was between 24% and 34% of the average concentration level for that stratum. For concentration levels below 10 µg/mL, the value of the RMSE dropped to less than 4 µg/mL.
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXREFERENCES |
|---|
There are several possible reasons why the CL estimate in the earlier study was lower than that predicted in the POP-PK analysis (ie, 1640 mL/h for a male patient and 1405 mL/h for a female patient). First, only patients with quantifiable samples out to 24 hours postdosing were included in the determination of CL for patients in the phase II study. Exclusion of patients with samples that were below the lower limit of quantitation would result in a downward bias in the estimated CL. Second, plasma concentrations of motexafin gadolinium in the earlier study were calculated from the amount of elemental gadolinium in plasma measured by ICP-AES. Therefore, any gadolinium-containing degradation products would have contributed to the total amount of motexafin gadolinium estimated in each plasma sample. In the current POP-PK study, plasma concentrations were measured using an LC/MS/MS assay that was specific for motexafin gadolinium. Third, only 10 patients were evaluated in the earlier study, and significant variability in the CL estimates was observed. Therefore, the earlier CL estimate may not be a good representation of the overall population CL.
The variable disease state was used to distinguish between patients with brain metastases and glioblastoma multiforme. Because brain metastases patients are more likely to have metastatic cancer distributed to other parts of the body compared to patients with glioblastoma multiforme, liver and kidney function could vary between these 2 patient populations, resulting in MGd pharmacokinetics that differ between the 2 disease types. However, disease state was not significant when tested as a covariate in the POP-PK model. This suggests that the plasma pharmacokinetics of motexafin gadolinium does not significantly differ between patients with brain metastases and glioblastoma multiforme and supports the pooling of data across studies containing patients with each of these disease types.
All covariates that were found to be significant in the final POP-PK analysis were also significant during covariate screening, except for age, which was significant in the final POP-PK analysis but not during covariate screening.
The 5th and 95th percentile value for each covariate evaluated in the POP-PK analysis was calculated based on the distribution of values for all individuals included in the final model (see Table III). Using the POP-PK model to predict values for CL or V1 based on covariate values that fall outside these percentile limits is not recommended due to the paucity of data.
For patients with GBM (studies B and C), the clinical impression was that fewer patients were excluded for liver enzyme abnormalities compared to patients with brain metastases (study A). Therefore, the distribution for some covariates (eg, clinical chemistry liver function parameters) may be more representative of the general GBM patient population than the general brain metastases population, and appropriate caution should be exercised when extrapolating conclusions from the POP-PK study to the general brain metastases patient population.
The plasma CL of motexafin gadolinium was 14% lower in women than in men. Examination of the variance-covariance matrix revealed that weight and sex were correlated (Pearson correlation coefficient = –0.36; P .0001). In addition, the median value for weight-normalized clearance was similar for men and women (21.9 and 20.7 mL/h/kg, respectively). Therefore, although sex resulted in the greatest decrease in the NONMEM objective function, sex could be functioning as a surrogate for weight. A 14% change in CL is not expected to have clinical significance.
Concomitant administration of phenytoin, a cytochrome P450 3A4, 5, or 7 inducer, increased motexafin gadolinium CL by about 30%. Induction of cytochrome P450 metabolizing enzymes would be expected to increase CL. However, administration of other potential CYP450 3A4, 5, or 7 inducing drugs (eg, carbamazepine, dexamethasone, beclomethasone, methylprednisolone, prednisolone, prednisone) did not result in a significant increase in CL. These results suggest that the other drugs classified as 3A4, 5, or 7 inducers may be less potent than phenytoin or that phenytoin may affect CL due to a mechanism other than cytochrome P450 isozyme induction (eg, induction of CYP450 reductase). Alternatively, the interaction could be due to a mechanism that is altogether unrelated to the cytochrome P450 enzyme system. In vitro studies suggest that motexafin gadolinium can be metabolized by human cytochrome P450 reductase (unpublished data).
Clearance decreased with increasing alkaline phosphatase concentration. At the 5th percentile alkaline phosphatase concentration (50 U/L), the predicted CL was 7% increased compared to that predicted at the population mean alkaline phosphatase concentration. Similarly, at the 95th percentile concentration (185 U/L), the predicted CL was decreased 13%. Although these changes are not expected to have clinical significance, it is possible that future studies could reveal a more significant effect in a patient population with alkaline phosphatase levels that exceed those seen in this study.
In liver disease, the plasma concentration of alkaline phosphatase rises due to biliary tract obstruction. Although liver toxicity was the dose-limiting toxicity identified in an earlier phase II study,6 thepresenceof tumor or bone disease could also lead to elevated alkaline phosphatase levels. Alkaline phosphatase levels are known to correlate with age and sex.13,14 However, these covariates were already incorporated into the POP-PK model. Therefore, the incremental effect attributed to alkaline phosphatase should not contain a contribution from these potentially confounding factors. Markers that are more specific for liver disease (alanine aminotransferase [ALT], aspartate aminotransferase [AST], and total bilirubin) were not covariates that significantly improved the POP-PK model. Alkaline phosphatase levels are known to increase with tumor burden; it is possible that the extent of tumor burden may correlate with motexafin gadolinium CL.
Clearance decreased with increasing age. The predicted CL increased by 10% at the 5th percentile age (35 years) and decreased by 8% at the 95th percentile age (76 years) relative to that predicted at the population mean age (57.1 years). These changes in CL are unlikely to have clinical implications. Many therapeutic agents exhibit decreased plasma CL with increasing age, which is generally attributed to decreasing renal function with increasing age.15 Not surprisingly, creatinine clearance and patient age were correlated in the POP-PK study (Pearson correlation coefficient = –0.57; P .0001). During covariate screening, creatinine clearance, but not age, was found to significantly influence motexafin gadolinium plasma clearance during stepwise linear regression analysis. Based on these observations, it is tempting to speculate that age may function as a surrogate for creatinine clearance in the POP-PK model. A relatively small effect of age on the plasma CL of motexafin gadolinium would be expected if a renal pathway were involved because only about 26% of administered motexafin gadolinium radioactivity was excreted in patient urine (unpublished data). However, it is also possible that other factors affecting plasma CL could correlate with patient age, such as hepatic function. A relationship between age and hepatic function is not as firmly established in the literature.
Although hemoglobin was determined to significantly reduce the NONMEM OFV (P = .0008), this covariate did not markedly reduce the level of interpatient variability. At the 5th percentile hemoglobin concentration (11.1 g/dL), the predicted CL decreased by 8% compared to that predicted for the population mean hemoglobin concentration (13.3 g/dL). At the 95th percentile concentration (15.5 g/dL), the predicted CL increased by 8% compared to that at 13.3 g/dL. These changes are not expected to have clinical significance.
We were not able to postulate a physiological rationale for an impact of hemoglobin on motexafin gadolinium clearance because motexafin gadolinium does not appear to bind significantly to erythrocytes. Due to the large number of covariates evaluated in this study, entrance of hemoglobin into the model may have been due to a Type I error.
Central compartment volume for women was about 21% lower than for men. Although weight and sex were correlated, the median value for weight-normalized V1 was about 18% lower for women than for men (95.9 and 78.6 mL/kg, respectively). Therefore, in the case of V1, sex was probably not functioning as a surrogate for weight, even though weight and sex were correlated.
A small effect of serum creatinine on V1 was observed in our study. Based on the final model, V1 is predicted to vary by ±8% of the population mean over the 5th to 95th percentile range of serum creatinine values, which is of doubtful clinical significance. An effect of serum creatinine on V1 was unexpected, and we were unable to identify a physiological rationale for this type of effect. A positive correlation between creatinine clearance and the volume of distribution of digoxin has been observed and attributed to decreased tissue distribution due to increased potassium levels caused by renal failure.16 However, this effect was opposite to that seen in our study. Serum creatinine was retained in the model with a P value of .001, which was the least significant for any covariate that entered the model. Furthermore, the 95% confidence interval for the serum creatinine theta estimate included zero. Therefore, entrance of serum creatinine into the model may represent a Type I error.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXREFERENCES |
|---|
13% from the population mean CL estimate. The other covariates examined did not significantly improve the model. The plasma pharmacokinetics of motexafin gadolinium did not significantly differ between patients with brain metastases and glioblastoma multiforme. |
APPENDIX |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXREFERENCES |
|---|
M. Croghan, Arizona Oncology Associates, Tucson, Arizona (IRB: PRN IRB); R. Pezner, City of Hope National Medical Center, Duarte, California; A. Rao, Kaiser Permanente Medical Center–Southern, Los Angeles, California; P. Eisenberg, Marin Oncology Associates, Greenbrae, California; Q. T. Le and S. Hancock, Stanford Medical Center, Stanford, California; J. Ford, UCLA Medical Center, Los Angeles, California; L. Gaspar, University of Colorado–Denver, Aurora, Colorado (IRB: Western IRB); M. Katin, S. R. Bonin, D. E. Dosoretz, and M. E. Keisch, Radiation Therapy Services, Inc, Fort Myers, Florida; I. Crocker, Emory University Hospital, Atlanta, Georgia; R. Timmerman, Indiana University Medical Center, Indianapolis, Indiana; M. Seiler and T. Cosgriff, Hematology and Oncology Services, New Orleans, Louisiana; T. Batchelor, Massachusetts General Hospital, Boston, Massachusetts (IRB: Dana Farber Cancer Institute); K. Levin and L. Gaspar, Wayne State University, Harper Hospital, Detroit, Michigan; A. Schorer, Minneapolis VA, Department of Veterans Affairs Medical Center, Minneapolis, Minnesota; R. Siemers, North Memorial Research Center, Robbinsdale, Minnesota; P. Kumar, Cancer Institute of New Jersey, New Brunswick, New Jersey (IRB: Johnson Medical School); J. Liebmann, New Mexico Oncology Hematology Consultants, Albuquerque, New Mexico; C. Collier, Amsterdam Community Cancer Program, Amsterdam, New York (IRB: PRN IRB); A. Frank, Riverview Cancer Care Medical Associates, Rexford, New York; R. Albright, University of Cincinnati Medical Center, Cincinnati, Ohio (IRB: St. Elizabeth Medical Center); P. Pickens, Abington Hematology Oncology Associates, Inc, Meadowbrook, Pennsylvania; H. Brereton, Radiation Medicine Associates of Scranton, Scranton, Pennsylvania (Scranton Temple Residency Program); M. Werner-Wasik Thomas Jefferson University, Philadelphia, Pennsylvania; S. McCachren and D. Arwood, Thompson Cancer Survival Center, Knoxville, Tennessee; A. Cmelak, Vanderbilt University Medical Center, Nashville, Tennessee; F. Mott, Scott & White Hospital, Temple, Texas; R. Komaki, The UT MD Anderson Cancer Center, Houston, Texas; M. Saunders, Tyler Cancer Center, Tyler, Texas; J. Rieke, Department of Radiation Oncology, University of Washington School of Medicine, Seattle, Washington; M. Mehta, University of Wisconsin, Madison, Wisconsin; C. Schultz, Medical Center of Wisconsin, Milwaukee, Wisconsin; E. Chang, University of Texas, Houston, Texas; L. Gaspar, University of Colorado–Aurora, Aurora, Colorado; M. Leibenhaut, Sutter Cancer Center, Sacramento, California; J. Suh, Cleveland Clinic, Cleveland, Ohio; R. Timmerman, Indiana Cancer Pavilion, Indianapolis, Indiana; J. Yamada, Memorial Sloan-Kettering Cancer Center, New York; W. Roa, Cross Cancer Center Institute, Edmonton, Alberta, Canada (Alberta Cancer Board); Y. Ung, Sunnybrook Regional Cancer Centre, Toronto, Ontario, Canada; A. Bezjak, Princess Margaret Hospital, Toronto, Ontario, Canada; S. Sagar, Hamilton Regional Cancer Center, Hamilton, Ontario, Canada (IRB: McMaster University); B. Fisher, London Regional Cancer Centre, London, Ontario, Canada; L. Souhami, Montreal General Hospital, Montreal, Canada; C. Carrie, Centre Leon Berard, Lyon, France (IRB: CCPPRB de Bicetre); J. M. Caudrelier, Department of Radiotherapy in Centre Oscar, Lille, France; J. J. Mazeron, Hopital de la Pitie–Salpetriere, Paris, France (IRB: CCPPRB de Bicetre); C. Haie-Meder, Institut Gustave Roussy, Villejuif, Cedex, France (IRB: Bicetre); F. Lagerward, Daniel den Hoed Cancer Center, Rotterdam, the Netherlands; J. Meerwaldt, Medisch Spectrum Twente, Enschede, the Netherlands; C. H. J Terhaard, Universitair Medisch Centrum Utrecht, Utrecht, the Netherlands; L. Stalpesrs, Academisch Centrum, Universiteit van Amsterdam, Amsterdam, the Netherlands; P. T. R Rodrigus, Dr. Bernard Verbeeten Instituut, Tilburg, the Netherlands; E. Levine, Christie Hospital NHS Trust, Manchester, United Kingdom; M. Brada, The Royal Marsden NHS Trust, Surrey, United Kingdom; T. Illidge, Wessex Cancer Centre, Hampshire, United Kingdom; I. Kunkler, Western General Hospital, Edinburgh, United Kingdom.
|
FOOTNOTES |
|---|
Submitted for publication May 19, 2004; Revised version accepted October 7, 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSAPPENDIXREFERENCES |
|---|
1. Mody TD, Fu L, Sessler JL. Texaphyrins: synthesis and development of a novel class of therapeutic agents. In: Karlin KD, ed. Progress in Inorganic Chemistry. New York: John Wiley; 2001: 551-598.
2. Mehta MP, Rodrigus P, Terhaard C, et al. Survival and neurologic outcomes in a randomized trial of motexafin gadolinium and whole-brain radiation therapy in brain metastases. J Clin Oncol. 2003;21: 2529-2536.
3. Mehta MP, Shapiro WR, Glantz MJ, et al. Lead-in phase to randomized trial of motexafin gadolinium and whole-brain radiation for patients with brain metastases: centralized assessment of magnetic resonance imaging, neurocognitive, and neurologic end points. J Clin Oncol. 2002;20: 3445-3453.
4. Magda D, Lepp C, Gerasimchuk N, et al. Redox cycling by motexafin gadolinium enhances cellular response to ionizing radiation by forming reactive oxygen species. Int J Radiat Oncol Biol Phys. 2001; 51: 1025-1036.[CrossRef][Medline] [Order article via Infotrieve]
5. Rosenthal DI, Nurenberg P, Becerra CR, et al. A phase I single-dose trial of gadolinium texaphyrin (Gd-Tex), a tumor selective radiation sensitizer detectable by magnetic resonance imaging. Clin Cancer Res. 1999;5: 739-745.
6. Carde P, Timmerman R, Mehta MP, et al. Multicenter phase Ib/II trial of the radiation enhancer motexafin gadolinium in patients with brain metastases. J Clin Oncol. 2001;19: 2074-2083.
7. Boeckmann AJ. NONMEM Users Guide: Part V. Introductory Guide. San Francisco: NONMEM Project Group, University of California at San Francisco; 1992.
8. Mizuta E, Tsubotani A. Preparation of mean drug concentration—time curves in plasma: a study on the frequency distribution of pharmacokinetic parameters. Chem Pharm Bull (Tokyo). 1985;33: 1620-1632.[Medline] [Order article via Infotrieve]
9. Lacey LF, Keene ON, Pritchard JF, Bye A. Common noncompartmental pharmacokinetic variables: are they normally or log-normally distributed? J Biopharm Stat. 1997;7: 171-178.[Medline] [Order article via Infotrieve]
10. Julious SA, Debarnot CA. Why are pharmacokinetic data summarized by arithmetic means? J Biopharm Stat. 2000;10: 55-71.[CrossRef][Medline] [Order article via Infotrieve]
11. SAS Institute. SAS Language and Procedures, Version 6. Cary, NC: SAS Institute; 1989.
12. Vozeh S, Maitre PO, Stanski DR. Evaluation of population (NONMEM) pharmacokinetic parameter estimates. J Pharmacokinet Biopharm. 1990;18: 161-173.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
13. Gordon T. Factors associated with serum alkaline phosphatase level. Arch Pathol Lab Med. 1993;117: 187-190.[Web of Science][Medline] [Order article via Infotrieve]
14. Wilding P, Rollason JG, Robinson D. Patterns of change for various biochemical constituents detected in well population screening. Clin Chim Acta. 1972;41: 375-387.[Medline] [Order article via Infotrieve]
15. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron. 1976;16: 31-41.[Web of Science][Medline] [Order article via Infotrieve]
16. Nagaraja NV, Park YJ, Jeon S, Sands CD, Derendorf H. Population pharmacokinetics of digoxin in Korean patients. Int J Clin Pharmacol Ther. 2000;38: 291-297.[Medline] [Order article via Infotrieve]
17. Sheiner L, Beal S. A note on confidence intervals with extended least squares parameter estimates. J Pharmacokinet Biopharm. 1987; 15: 93-98.[CrossRef][Medline]
[Order article via Infotrieve]
CiteULike 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  K. A. Bradley, I. F. Pollack, J. M. Reid, P. C. Adamson, M. M. Ames, G. Vezina, S. Blaney, P. Ivy, T. Zhou, M. Krailo, et al. Motexafin gadolinium and involved field radiation therapy for intrinsic pontine glioma of childhood: A Children's Oncology Group phase I study Neuro-oncol, January 1, 2008; 10(5): 752 - 758. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. De Stasio, D. Rajesh, J. M. Ford, M. J. Daniels, R. J. Erhardt, B. H. Frazer, T. Tyliszczak, M. K. Gilles, R. L. Conhaim, S. P. Howard, et al. Motexafin-Gadolinium Taken Up In vitro by at Least 90% of Glioblastoma Cell Nuclei Clin. Cancer Res., January 1, 2006; 12(1): 206 - 213. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
