| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Clinical Trials |
Division of Hematology/Oncology [J. H., A. S., L. H. B.], General Clinical Research Center [P. B. W.], and Biostatistics Care [M. S.], University of Michigan Comprehensive Cancer Center, Ann Arbor, Michigan 48109, and Rhône-Poulenc Rorer, Antony Cedex 92165, France [R. B.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-1-acidic
glycoprotein. The natural log of ERMBT accounted for 67% of the
interpatient variation in CL. Multivariate analysis showed that the
natural log of ERMBT and albumin together accounted for 72% of the
interpatient variation in CL. The greatest toxicity was seen in
patients with the lowest ERMBT. Hepatic CYP3A4 activity is the
strongest predictor of docetaxel CL and accounts for the majority of
interpatient differences in CL. Patients with low CYP3A4 activity are
at risk for having decreased CL and may thus experience increased
toxicity from docetaxel. Those with high activity may be receiving a
suboptimal dose. By measuring CYP3A4 activity, the ERMBT may be
clinically useful in tailoring doses of CYP3A4 substrates, such as
docetaxel, in certain individuals. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The cytochrome P450 system is a family of heme-containing enzymes responsible for the metabolism of nonendogenous substances, such as drug molecules and toxins (2) . Specific isoenzymes are designated by the prefix "CYP," and to date, over 30 isoenzymes have been purified, cloned, sequenced, and characterized in humans. CYP3A4 is found in the liver and small bowel. The CYP3A4 enzyme displays large interpatient differences in both content and catalytic activity in humans (3) . These differences exist even in the absence of medications known to induce or inhibit the enzyme and are thus likely to reflect genetic variability (3) .
The ERMBT3 specifically measures the in vivo activity of hepatic CYP3A4 (3 , 4) and is therefore ideally suited for drugs given i.v., such as docetaxel. This test is based on the fact that CYP3A4 is the major enzyme responsible for erythromycin metabolism, liberating a carbon atom that is exhaled as carbon dioxide. After injecting a trace dose of radiolabeled erythromycin, the rate of radiolabeled carbon dioxide exhaled is measured. This test has been validated as a specific measure of CYP3A4 activity (3) .
In vitro studies suggest that CYP3A4 is the major enzyme involved in docetaxel metabolism (5) . Docetaxel is excreted in the feces with less than 8% excreted unchanged (6) . It is highly bound to plasma proteins including AAG. Pharmacokinetic data show significant variation in CL between patients (7) . Decrease in total body CL is associated with increased frequency and severity of side effects (8) . Mulitvariate analysis demonstrates that interpatient variability of CL is predicted by body surface area, AAG plasma level, and elevated hepatic enzymes; however, these parameters do not account for total variability (7) . Because docetaxel binds chiefly to AAG, high levels of this protein presumably limit the free fraction of docetaxel available for CL in the liver. Due to lower CL of docetaxel in patients with elevated transaminases, it is currently recommended in the United States package insert that patients with an ALT > 1.5x the ULN concomitant with alkaline phosphatase > 2.5x ULN or bilirubin > ULN do not receive this drug. The European Summary of Product Characteristics alternatively recommends a dosage reduction of 25% in patients with concomitant elevations of ALT and alkaline phosphatase.
Previous studies have shown modest activity for docetaxel in patients
with advanced sarcomas (response rates of 17% and 18%), thus we
decided to try this agent in our sarcoma population (9
, 10)
. We hypothesized that variability in CYP3A4 activity
accounts for interpatient differences in docetaxel CL and toxicity.
Thus, the ERMBT was used to measure CYP3A4 activity in sarcoma patients
receiving docetaxel (Fig. 1)
.
|
|
PATIENTS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Study Schema and Procedures.
Three ERMBTs were administered to each patient at the GCRC. The
test was done on three occasions in case either the dexamethasone
premedication or the chemotherapy introduced variability. The first
study served as a baseline and was done before admission for docetaxel
(range, 1–19 days before admission). The second study was done 24 h after dexamethasone premedication was initiated and 1 h before
the infusion of docetaxel. The third study was done 3 h after the
infusion of docetaxel had been completed. A total of 3 µCi (<0.1
mmol) of [14C-N-methyl]erythromycin
obtained from Metabolic Solutions, Inc. (Nashua, NH) was given i.v.
Exhaled carbon dioxide was trapped in a solution of hyamine hydroxide,
ethanol, and a blue indicator before and 20 min after the injection.
The percentage of 14C exhaled/h =
49.496 x (percentage of 14C exhaled/min at
20 min - baseline counts) (11)
.
Docetaxel was administered at 100 mg/m2 i.v. over 1 h. Each patient received the following standard premedications: (a) dexamethasone (8 mg, p.o.) every 12 h starting the day before infusion and continuing for 5 days; and (b) granisetron (2 mg, p.o.) 30 min before the infusion. Blood was drawn for pharmacokinetic analysis over 24 h with specimens at optimal sampling times: just before the end of infusion and at 0.25, 0.75, 3.00, 6.50, and 24.00 h after the end of infusion (12) . Case report forms documented actual sampling times as well as the actual beginning and end times of the infusion. Docetaxel was assayed from frozen plasma samples using reverse-phase high-performance liquid chromatography (13) . All samples were assayed at the same time. A Bayesian criterion was used to calculate the docetaxel plasma concentration area under the curve based on measured drug levels and population pharmacokinetic parameters using NONMEM software (12 , 14) . The area under the curve was computed as dose/CL. CL was directly estimated by fitting the model (12 , 14) . Patients had baseline liver chemistries and serum AAG levels drawn. One patient did not have a baseline serum ALT recorded, and the two patients had insufficient sample to determine the AAG level. Baseline and weekly hematological counts were checked. Granulocyte-colony stimulating factor was not given during this first cycle.
Subsequent cycles of chemotherapy were given every 3 weeks if the patient was judged by us to be deriving clinical benefit. The pharmacokinetic data were obtained only during the initial cycle. Treatment was stopped if the disease progressed, if intolerable toxicities developed, or at the patient’s request. Radiological evaluation of patients was done at the discretion of the treating physician to document clinical response or progression.
Statistical Analysis.
Although the distribution of the ERMBT in a large sample of
normal controls is positively skewed, the distribution in this small
sample is approximately normal. The paired t test was used
to compare the breath test values obtained at the three time points.
Pearson’s correlation coefficient (r) further describes the
strength of the linear relationship and is presented along with the
P. Nonparametric analogues to these test procedures were
conducted for confirmation but are not reported.
CL, as determined by pharmacokinetic modeling, is approximately normal. The association between CL and patient characteristics was evaluated using simple and multiple least-squares regression models. Variables considered as potential predictors of CL in this model included ERMBT, AAG, albumin, ALT, and alkaline phosphatase. A natural log transformation was applied if indicated by univariate residual plots. For instance, the relationship between the ERMBT and CL was not linear over the entire range of values, particularly at the lowest ERMBT values, and the low ERMBT values had a better fit when ln ERMBT was used. The best model for predicting CL was determined from the set of predictors using a step-wise procedure based on the largest F statistic for each variable’s contribution to the model. The first factor included using the step-wise procedure is the best single predictor of CL. Two patients missing certain lab test results were excluded from analyses that involved these lab tests. All two-way interactions were examined for the final best model. All analyses were conducted using the SAS software package.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The ERMBT varied >20-fold, largely
because a few patients had very low breath test results. The CL of
docetaxel varied >5-fold among these patients. These ERMBT data
exhibit similar variability to that seen in 107 normal controls studied
previously (ERMBT mean = 2.56; SD, 1.06; range, 0.77–6.15). There
was excellent correlation between ERMBT results on all three occasions
(r = 0.81–0.92, P < 0.0001). Thus,
there is no evidence for a significant effect of dexamethasone
premedication on the ERMBT. There is a tendency for the test value to
fall modestly after docetaxel administration (ERMBT #3
versus ERMBT #2, P = 0.0047), likely
reflecting competition between erythromycin and docetaxel for
metabolism by CYP3A4. The ERMBT #2 was used in subsequent analyses.
|
demonstrates the
relationship between CL and the phenotypic measurement of CYP3A4. As
expected, those patients with reduced CL had the lowest ERMBT results.
As mentioned in the statistical section, the low ERMBT values had a
better fit when ln ERMBT was used, with
r2 = 0.67, P = 0.0001.
Even if the ERMBT was not log-transformed, the relationship between CL
and the ERMBT remained strong (r2 =
0.49, P = 0.001). Albumin was also a good predictor of
CL (r2 = 0.46, P = 0.001). Alkaline phosphatase was not as strong a predictor of CL. One
patient with osteosarcoma had an exceedingly high alkaline phosphatase
(5769) from bone rather than liver origin and was therefore included in
the study. When this outlier was removed from the analysis, alkaline
phosphatase was a weak predictor of CL (r2 =
0.36). The relationships between CL and ALT or AAG were also weak
(r2 = 0.23 and 0.12, respectively),
but in the anticipated inverted direction. Thus, the ERMBT proved to be
the best single predictor of docetaxel CL.
|
As described in the statistical section, simple and multiple regression analyses were conducted to model the best predictors of CL. The best two-factor model for predicting CL was determined to be ln ERMBT and albumin (r2 = 0.72). No other factors added significantly to this model, and no statistically significant codependence between variables was detected. In the two-factor model containing untransformed ERMBT and AAG, both variables were independently predictive of CL (r2 = 0.61), with higher AAG predicting lower CL.
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The most significant independent variable examined to predict docetaxel CL in this study was hepatic CYP3A4 activity as measured by ERMBT. Previously, the best predictors of docetaxel pharmacokinetics that had been identified were abnormal liver chemistries and serum AAG level. Albumin was also a significant predictor of CL, but its effect was of low magnitude compared with that of AAG (9) . In this study, AAG and the ERMBT results do not appear to be correlated, and the ERMBT is a superior predictor of CL. In fact, the ERMBT and albumin are the strongest predictors of CL, and together they account for 72% of the interpatient variability in CL. Why is the AAG a poorer predictor in this study? The pharmacokinetic data and AAG levels in this study were determined in the same manner and by the same investigator as these prior data, so quality control should not be a factor. The small sample size of the present study may account for this difference. Bruno’s study (8) included 547 patients and used a statistical approach with Bayesian estimates directed toward subpopulations with extreme covariate values (presumably clinically at a higher risk). This requires a large sample size. This approach is not applicable to our study of 21 patients. The variability of AAG levels may also be important. Docetaxel is strongly protein-bound to AAG, thus high AAG levels should decrease docetaxel CL. In the previous work (7) , this was evident for high AAG levels; however, the model did not improve the prediction of CL for patients with low AAG levels (<0.88 g/liter). Perhaps our study did not include as many patients with "high" AAG levels, and thus AAG was not found to be an important predictor of docetaxel CL. The AAG reflects the free fraction of drug available for CL in the liver, whereas CYP3A4 activity reflects the individual’s inherent ability to metabolize the drug. Although this is a small study, the fact that CYP3A4 activity was a better predictor is an important new finding and will hopefully improve our understanding of the pharmacokinetics of docetaxel. Larger studies are under way and will help clarify the importance of AAG and CYP3A4 activity in relationship to CL.
Although toxicities (namely, grade 3 or 4 neutropenia) were seen in virtually all patients at this dose level, the observation that an ERMBT result of <1 in this study was associated with the severest toxicities is interesting. Although marked interpatient heterogeneity in the expression of CYP3A4 genes is known to exist, the clinical implications of having a "high" or "low" CYP3A4 activity were not necessarily known. In this study "low" CYP3A4 activity (ERMBT < 1) appears to be associated with excess toxicity. It implies that consideration should be given to modifying the dose for patients with such a low ERMBT. Similarly, one would wonder whether patients with "high" CYP3A4 activity are at risk of getting subtherapeutic doses of CYP3A4 substrates such as docetaxel. This hypothesis should be studied further.
In this study, single agent docetaxel was given, but docetaxel is now commonly included in combination regimens. Because the additional drugs are often myelotoxic but not CYP3A4-mediated, the ERMBT may be helpful in sorting out the specific effect docetaxel makes with regard to the toxicity of the regimen.
Logistically, the ERMBT is easy to administer and poses a minimal risk to patients. It proved to be highly reproducible. It was not significantly affected by dexamethasone premedication, and this observation is consistent with prior data (15) that showed no substantial effect of dexamethasone premedication on docetaxel CL. The test can therefore be done at a convenient time, days or weeks ahead of the planned administration of docetaxel.
In conclusion, CYP3A4 activity was found to be the most significant independent variable for predicting the CL of docetaxel. The ERMBT, perhaps together with serum albumin or AAG, may be clinically useful in individualizing the dose of docetaxel and possibly other drugs that are CYP3A4 substrates, including several new antineoplastic agents. It may also be helpful in individualizing the dose of docetaxel in patients with markedly elevated hepatic enzymes who are currently excluded from treatment with docetaxel.
We plan to initiate a study using tailored dosing of docetaxel based on CYP3A4 activity. We postulate that by tailoring the dose, variability in patient drug exposure could be diminished. Hopefully, modern pharmacogenetics can be put into practical use in the clinic and allow us to treat patients with better precision and less toxicity.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported by General Clinical Research Center
Grant M01 RR00042. 
2 To whom requests for reprints should be
addressed, at University of Michigan Comprehensive Cancer Center, 1500
East Medical Center Drive, 7216 CCGC, Ann Arbor, MI 48109-0948.
3 The abbreviations used are: ERMBT, erythromycin
breath test; AAG, -1-acidic glycoprotein; ULN, upper limit of
normal; ALT, alanine aminotransferase; ln ERMBT, natural log of ERMBT;
GCRC, General Clinical Research Center; CL, clearance.
Received 8/24/99; revised 12/ 1/99; accepted 12/ 6/99.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  L.-S. Tham, N. H.G. Holford, S.-Y. Hor, T. Tan, L. Wang, R.-C. Lim, H.-S. Lee, S.-C. Lee, and B.-C. Goh Lack of Association of Single-Nucleotide Polymorphisms in Pregnane X Receptor, Hepatic Nuclear Factor 4{alpha}, and Constitutive Androstane Receptor with Docetaxel Pharmacokinetics Clin. Cancer Res., December 1, 2007; 13(23): 7126 - 7132. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. Lichtman, H. Wildiers, E. Chatelut, C. Steer, D. Budman, V. A. Morrison, B. Tranchand, I. Shapira, and M. Aapro International Society of Geriatric Oncology Chemotherapy Taskforce: Evaluation of Chemotherapy in Older Patients--An Analysis of the Medical Literature J. Clin. Oncol., May 10, 2007; 25(14): 1832 - 1843. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J Alexandre, E Rey, V Girre, S Grabar, A Tran, V Montheil, F Rabillon, V Dieras, V Jullien, P Herait, et al. Relationship between cytochrome 3A activity, inflammatory status and the risk of docetaxel-induced febrile neutropenia: a prospective study Ann. Onc., January 1, 2007; 18(1): 168 - 172. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. M. Bosch, A. D.R. Huitema, V. D. Doodeman, R. Jansen, E. Witteveen, W. M. Smit, R. L. Jansen, C. M. van Herpen, M. Soesan, J. H. Beijnen, et al. Pharmacogenetic Screening of CYP3A and ABCB1 in Relation to Population Pharmacokinetics of Docetaxel. Clin. Cancer Res., October 1, 2006; 12(19): 5786 - 5793. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Michael, M. Thompson, R. J. Hicks, P. L. Mitchell, A. Ellis, A. D. Milner, J. Di Iulio, A. M. Scott, V. Gurtler, J. M. Hoskins, et al. Relationship of Hepatic Functional Imaging to Irinotecan Pharmacokinetics and Genetic Parameters of Drug Elimination J. Clin. Oncol., September 10, 2006; 24(26): 4228 - 4235. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. E. Broker, F. Y.F.L. de Vos, C. J. van Groeningen, B. C. Kuenen, H. E. Gall, M. H. Woo, M. Voi, J. A. Gietema, E. G.E. de Vries, and G. Giaccone Phase I Trial with BMS-275183, a Novel Oral Taxane with Promising Antitumor Activity. Clin. Cancer Res., March 15, 2006; 12(6): 1760 - 1767. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. J. Komoroski, R. A. Parise, M. J. Egorin, S. C. Strom, and R. Venkataramanan Effect of the St. John's Wort Constituent Hyperforin on Docetaxel Metabolism by Human Hepatocyte Cultures Clin. Cancer Res., October 1, 2005; 11(19): 6972 - 6979. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Yamamoto, T. Tamura, H. Murakami, T. Shimoyama, H. Nokihara, Y. Ueda, I. Sekine, H. Kunitoh, Y. Ohe, T. Kodama, et al. Randomized Pharmacokinetic and Pharmacodynamic Study of Docetaxel: Dosing Based on Body-Surface Area Compared With Individualized Dosing Based on Cytochrome P450 Activity Estimated Using a Urinary Metabolite of Exogenous Cortisol J. Clin. Oncol., February 20, 2005; 23(6): 1061 - 1069. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. C. Dees and P. B. Watkins Role of Cytochrome P450 Phenotyping in Cancer Treatment J. Clin. Oncol., February 20, 2005; 23(6): 1053 - 1055. [Full Text] [PDF]
|
|
|
|
|
|
S. D. Baker, R. H. N. van Schaik, L. P. Rivory, A. J. ten Tije, K. Dinh, W. J. Graveland, P. W. Schenk, K. A. Charles, S. J. Clarke, M. A. Carducci, et al. Factors Affecting Cytochrome P-450 3A Activity in Cancer Patients Clin. Cancer Res., December 15, 2004; 10(24): 8341 - 8350. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. H. J. Mathijssen, F. A. de Jong, R. H. N. van Schaik, E. R. Lepper, L. E. Friberg, T. Rietveld, P. de Bruijn, W. J. Graveland, W. D. Figg, J. Verweij, et al. Prediction of Irinotecan Pharmacokinetics by Use of Cytochrome P450 3A4 Phenotyping Probes J Natl Cancer Inst, November 3, 2004; 96(21): 1585 - 1592. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B.-C. Goh, S.-C. Lee, L.-Z. Wang, L. Fan, J.-Y. Guo, J. Lamba, E. Schuetz, R. Lim, H.-L. Lim, A.-B. Ong, et al. Explaining Interindividual Variability of Docetaxel Pharmacokinetics and Pharmacodynamics in Asians Through Phenotyping and Genotyping Strategies J. Clin. Oncol., September 1, 2002; 20(17): 3683 - 3690. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Arora, M. L. Cate, C. Ghosh, and P. L. Iversen Phosphorodiamidate Morpholino Antisense Oligomers Inhibit Expression of Human Cytochrome P450 3A4 and Alter Selected Drug Metabolism Drug Metab. Dispos., July 1, 2002; 30(7): 757 - 762. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. I. Block and C. Gyllenhaal Clinical Corner: Herb-Drug Interactions in Cancer Chemotherapy: Theoretical Concerns Regarding Drug Metabolizing Enzymes Integr Cancer Ther, March 1, 2002; 1(1): 83 - 89. [Abstract] [PDF]
|
|
|
|
|
|
C. F. Stewart and E. G. Schuetz Need and Potential for Predictive Tests of Hepatic Metabolism of Anticancer Drugs Clin. Cancer Res., September 1, 2000; 6(9): 3391 - 3392. [Abstract] [Full Text]
|
|
|
|
|
|
J. M. Collins Cytochrome P-450 and Other Determinants of Pharmacokinetics, Toxicity, and Efficacy in Humans Clin. Cancer Res., April 1, 2000; 6(4): 1203 - 1204. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
