| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Clinical Trials |
Medical Oncology, Sydney Cancer Centre, New South Wales 2050 [L. P. R., M. B., P. J. B., M. J. M., J. F. B., S. J. C.], and Department of Pharmacology, University of Sydney, New South Wales 2006 [L. P. R., K. S., J. P. S., J. M. H.], Australia
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Sixteen adult patients with incurable cancer were studied. The EBT was performed on day 1 and breath sampled after the i.v. injection of 4 µCi of 14C-erythromycin. The breath 14CO2 flux (CERt) was estimated at 11 time points over 2 h. On day 2, the EBT was repeated midway through a 10-min infusion of 100 mg of erythromycin lactobionate, and the plasma pharmacokinetics of erythromycin were determined.
The infusion of 100 mg of erythromycin did not modify the EBT results significantly. The values of the conventional EBT parameter CER20 min obtained on day 1 were comparable for most subjects (0.03–0.06% dose/min), with the exception of an individual receiving the known CYP3A4 inducers dexamethasone and phenytoin who returned a value of 0.14% dose/min. There was no significant correlation between any of the conventional EBT parameters and erythromycin clearance. However, two parameters reflecting early emergence of breath radioactivity (1/TMAX and CER3 min/CERMAX) correlated significantly with erythromycin clearance (P = 0.005 and 0.006, respectively).
Novel parameters derived from the EBT are significantly correlated with the clearance of erythromycin even in the presence of confounding factors, such as metastatic liver disease, altered protein binding, and comedication. These parameters may enable dose optimization of cytotoxics metabolized by CYP3A4.
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The majority of drugs are cleared by hepatic metabolism, particularly
by the cytochrome P450 mixed function oxidases. The principal hepatic
isoform of these oxidases,
CYP3A4,3
is thought
to be a significant contributor to the metabolism of 
60% of all
drugs (3)
. Of particular interest in oncology is the role
of CYP3A4 in the metabolism of cytotoxic drugs, such as paclitaxel,
docetaxel, vinorelbine, vincristine, irinotecan, topotecan, ifosfamide,
cyclophosphamide, and tamoxifen (4)
. Therefore, a
clinically applicable technique capable of predicting CYP3A4-mediated
clearance would prove useful in the rational dose selection of many
cytotoxic drugs. Presently, doses are mostly selected according to the
patient’s BYSA, although there is little scientific rationale
for this practice (5)
.
Several in vivo probes of CYP3A4 have been proposed and their relative merits discussed (3 , 6, 7, 8, 9, 10, 11, 12) . Arguably, the most widely accepted is the EBT (13) . Erythromycin is N-demethylated by CYP3A4, and the cleaved methyl group is released as formaldehyde and, eventually, as CO2. The inclusion of a 14C radiolabel on the N-methyl moieties results in the production of 14CO2, which is then exhaled. After the administration of 3–4 µCi of 14C-erythromycin i.v., the rate of release of 14CO2 in expired breath is thought to reflect hepatic CYP3A4 activity (13) . Significant correlations have been observed between the 14CO2 flux at a particular time, or its time-integral, with the clearance of drugs with significant CYP3A4 metabolism, such as cyclosporine (14, 15, 16) , midazolam (8) , and others (3) . Also, Baker et al. (17) established a correlation between the EBT and docetaxel clearance normalized for BYSA. However, other studies attempting to correlate drug clearance with the EBT have provided negative findings (9 , 10 , 12) , and erythromycin itself has, surprisingly, not been studied. Therefore, whether the EBT can be used as a reliable predictor of the clearance of drugs metabolized by CYP3A4 requires further clarification (18) .
In animal studies, it has been shown that the EBT most accurately
reflects the induction of CYP3A4 by dexamethasone when additional,
unlabeled, erythromycin is coadministered with the
14C-erythromycin (13)
. This could
occur if CYP3A-mediated drug metabolism of tracer erythromycin (a
tracer dose of 4 µCi equates to 0.05 mg of erythromycin) is not
identical to that observed with pharmacologically relevant doses (
100
mg). The addition of unlabeled erythromycin could, therefore, improve
the accuracy of the EBT.
For drugs metabolized principally by the liver and with moderate extraction ratios (such as erythromycin), the total body clearance (CL) is approximated by the product of the fraction unbound in plasma (fUB) and the intrinsic clearance (CLINT). The latter represents the drug metabolism activity of the liver expressed as a clearance. From the presence of fUB in the product, it is evident that protein binding could be a major confounder of the relationship between drug clearance and CYP3A4-mediated intrinsic clearance. This concern is all of the more relevant given that plasma AAG, which is the major binding protein for many drugs, including erythromycin (19) , is increased significantly in patients with advanced malignancies (20 , 21) . The other possible confounder is the presence of pathways of elimination, such as renal clearance, biliary excretion, and non CYP3A4-mediated metabolism (12) . Hence, even if the EBT is a good predictor of CYP3A4 activity, the above listed factors could nullify its usefulness with regards to predicting total body clearance of drugs. This may be further exacerbated by the presence of comorbidity and comedication, which can affect drug metabolism and the EBT (13) .
The objectives of this study were to investigate the relationship between the EBT and erythromycin clearance in cancer patients, and to determine the effects of erythromycin dose and protein binding on this relationship.
|
PATIENTS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The study was carried out as a two-period nonrandomized design. On D1, a conventional EBT was performed. Fresh plasma was collected for the analysis of erythromycin protein binding and measurement of AAG. On the following day, the EBT was performed midway through a 10-min infusion of 100 mg of erythromycin lactobionate (Erythrocin-I.V., Abbott Australasia Pty Ltd., Kurnell, Australia) reconstituted in 100 ml of 0.9% saline. To remove any confounding influence of mode of administration, the D1 EBT was carried out midway through a 10-min infusion of drug-free saline. Subjects were not fasted, and they were instructed to have similar breakfasts on the 2 days. Physical activity, position, and time of day may all affect the endogenous production of CO2; hence, all EBTs were initiated between 9 and 11 a.m. Also, patients were seated in a semirecumbent position for at least 15 min before the injection of tracer erythromycin and throughout the breath collection period.
EBT.
The 4-µCi dose of 14C erythromycin
(N-methyl-14C, 55 mCi/mmol; NEN Life
Science Products Inc, Boston, MA) was diluted into 5 ml of dextrose
saline and loaded into a 5-ml syringe. A low-volume 0.22-µm sterile
filter (Millex-GV, Millipore, Bedford, MA) was attached, and the air
was chased through. After obtaining the total weight of the syringe,
the dose was administered as a bolus through the side port of the
infusion set. Breath collection was carried out by blowing through a
straw into 2 ml of capture solution in glass vials. The capture
solution consisted of benzethonium hydroxide (Sigma Chemical Co, St.
Louis, MO) in 50:50 methanol/ethanol v/v to which a trace of
phenolphthalein was added. A change in color from pink to clear
indicated the capture of 1 mmol CO2. Breath was
collected pre-dose and then at 3, 8, 20, 30, 40, 55, 65, 75, 95, 115,
and 145 min after the injection of radiotracer. Vials were capped and
stored on ice until 10 ml of scintillant cocktail (Ultima Gold,
Packard, Sydney, New South Wales, Australia) could be added. The
samples were vortex-mixed and left standing at room temperature in the
dark before liquid scintillation counting (Tricarb 1500, Packard). The
flux of exhaled 14CO2 at
each time point (CERt), expressed as a
percentage of dose/min, was calculated from:
 |
weight was the mass (g) of solution injected. The
product of BYSA·5 allows the normalization of total
CO2 output of individuals on the basis of 5 mmol
exhaled/min/m2
in resting individuals
(13)
. The maximum flux,
CERMAX, and the time at which it occurred,
TMAX, were determined directly from
observed values. The areas under the flux curve for the first hour
(AUCER0
1 h)
and extrapolated to infinity
(AUCER0
) were
estimated by standard trapezoidal methods. In the case of
AUCER0,
extrapolation was carried out after identifying the terminal log-linear
decrease in flux (KCER).
Because the values of TMAX are limited
to actual sampling times, this causes a lack of continuity in this
variable. Therefore, a model-dependent
TMAX value
(TMAX(model)) was also obtained by
first fitting the breath-flux data to the equation:
 |
 |
Erythromycin Pharmacokinetics.
On D2, blood samples (5 ml) were collected from a contralateral cannula
into heparinized containers before, midway, and at the end of the
erythromycin infusion and then 5, 10, 20, 40, 50, 60, 80, 100, 120,
160, 200, 240, 300, and 360 min after. The samples were immediately
centrifuged at 4°C for 10 min, and plasma was transferred into
polypropylene vials, which were stored at -70°C.
Samples were analyzed using a Shimadzu QP8000 (LC/mass spectrometry) Spectrometer (Shimadzu Scientific Instruments, Rydalmere, New South Wales, Australia) connected to a Shimadzu LC system via the electrospray interface. Firstly, samples, standards (0.01, 0.1, 1, 5, and 20 µg/ml) and quality controls (0.02 and 10 µg/ml) were extracted using Nexus solid phase extraction columns (Varian Australasia Pty Ltd., Mulgrave, Victoria, Australia) after spiking with 1 µg of the internal standard, oleandomycin. Samples were loaded onto the unconditioned columns, which were rinsed sequentially with 1 ml of water and 1 ml of 20% aqueous methanol and eluted with 1 ml of methanol into 1.5-ml Eppendorf tubes. The eluant was dried under vacuum in a centrifuge (SpeedVac SC200, Savant Instruments, Sydney, New South Wales, Australia) after which it was reconstituted in 100 µl of the LC mobile phase [20 mM ammonium acetate (pH 5.5) mixed 1:1 with acetonitrile, v/v]. The samples were vortex-mixed and centrifuged, and 0.2–10 µl were injected onto the column (Symmetry C-8, 2 x 150 mm, Waters Australasia Pty Ltd., Rydalmere, Australia) using a cooled (15°C) autoinjector. The entire column outflow (0.2 ml/min) was nebulized at 230°C with a N2 flow of 4 liters/min. The probe, curved desoluation line, and deflector plate voltages were set at 3.5 keV, -30 V, and 55 V, respectively in positive ion mode. Erythromycin and oleandomycin were monitored as their corresponding [M+H] ions (m/z of 734.7 and 688.6, respectively). Using this method, the mean recoveries of erythromycin and oleandomycin were 64.1 and 81.6%, respectively. Inter- and intraday variability and total accuracy averaged 7.3, 7.7, and 99.0%, respectively, over the range of erythromycin concentrations studied (0.01–20 µg/ml).
The area under the concentration curve
(AUC0) and the
area under the first moment of the concentration curve
(AUMC0) were
estimated using the trapezoidal method and extrapolated to infinity.
Standard pharmacokinetic parameters, such as total body clearance
(CL), volume of distribution at steady-state
(VdSS), and mean residence time
(MRT) were calculated from:
 |
 |
 |
is the duration of infusion. The terminal elimination half-life (t1/2(z)) was estimated from the log-linear regression of the last five to six time points of the concentration versus time curve.
Protein Binding of Erythromycin.
Erythromycin binding was investigated in freshly collected plasma at
two concentrations, 0.4 and 2.1 µg/ml. Briefly, 9 nCi of
14C-erythromycin were added alone or with
erythromycin as the free base (ICN Biomedicals Inc., Aurora, OH) to 0.6
ml plasma and incubated at 37°C for 5 min. The incubate was
transferred to ultrafiltration devices (10,000 mw cutoff,
Centrifree, Amicon Corp., Danvers, MA) and centrifuged in a fixed rotor
centrifuge for 7 min. Aliquots (100 µl) of the ultrafiltrate and
reservoir were counted by liquid scintillation, and the fraction
unbound in plasma (fUB) was calculated
as the ratio of the respective dpm counts.
The plasma concentration of AAG was determined using a commercial immunodiffusion kit (NOR Partigen, Dade Behring Inc., Newark, DE).
Statistics.
The differences in EBT parameters on days 1 and 2 were analyzed by
Student’s paired t-tests, except for
TMAX data for which a Wilcoxon Signed
Ranks Test was used. A Student’s paired t test was also
used to evaluate the effect of erythromycin concentration on protein
binding. Correlations between CL and EBT parameters
(CERMAX,
CER20 min,
AUCER01 h,
AUCER0, and
TMAX) from each day were sought using
Pearson’s correlation. The Ps so obtained were corrected
for the multiple comparisons using the Dunn-Sidak test. Ps
obtained from other correlations were uncorrected except where
specified otherwise.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The median age was
57 years, and equal numbers of males and females were recruited. The
values of BYSA ranged from 1.48 to 2.40 m2
. Two
female subjects did not complete D2 of the study (lack of successful
venous access and personal reasons). Because of a technical difficulty,
the exact dose of 14C-erythromycin could not be
determined in one additional subject on D2. However, the D1 EBT results
from these individuals are included in the pertinent analyses.
|
. In general, the values of
CERMAX and
CER20 min were close to their respective
medians (±50% of median), with the exception of a subject being
treated with both phenytoin and dexamethasone who had the highest
values. Statistical analysis did not reveal any significant differences
between the 2 days for any of the tabulated parameters. There was a
trend toward a reduced TMAX on D2, but
this did not reach statistical significance (P =
0.069). Because of the presence of confounding factors (liver disease,
comedication, and so forth), the EBT results were not analyzed
for gender effects.
|
5-fold between patients and
ranged from 10.48 to 47.01 liters/h. The relevant pharmacokinetic
parameters are listed in Table 3
. There
was no significant correlation between clearance and either BYSA or age
(P > 0.7). D1 EBT and erythromycin plasma
concentration profiles of representative subjects with high and low
CL values are shown in Fig. 1
.
|
|
. Erythromycin clearance decreased with
increased protein binding, but this was not statistically significant
(r2
= 0.20; P = 0.11).
The volume of distribution at steady-state
(VdSS) was strongly correlated with
clearance (r2
= 0.74;
P = 8.7 x
10-5).
|
1 h, and
AUCER0) on
either of the 2 days and the measured erythromycin clearance
(r2
< 0.22; Dunn-Sidak
P > 0.5). Subjects with high clearance values tended
to have early emergence of breath radioactivity on D1, and
consequently, low TMAX values
(e.g., see Fig. 1
). This did not apply in the case of the D2
breath-test results, and a possible reason for this is discussed later.
CL was negatively and significantly correlated with
TMAX
(r2
= 0.49; Dunn-Sidak
P = 0.025). However, the corresponding line of
regression indicated that CL would take on negative values
for TMAX > 30 min. Because
CL is expected to range from 0 to infinity, the correlation
between CL and the D1 1/TMAX was
investigated and found to be significant also
(r2
= 0.49; P = 0.005). Using the
inverse of the TMAX value estimated
from the fitting of the breath-flux data to equations 2 and 3
(TMAX(model)) yielded a stronger
correlation (r2
= 0.64;
P = 0.001). Another indicator of early breath-emergence
was
CER3 min/CERMAX,
which was also significantly correlated with CL
(r2
= 0.47; P =
0.006).
Plotting the CL data against
1/TMAX(model) (Fig. 3)
revealed that most were well described
by linear regression, with the exception of a single outlier
(Studentized Residual = 4.3). Removal of this outlier yielded the
linear regression (r2
= 0.85;
P = 7.5 x
10-6) described by:
 |
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The conventionally derived EBT parameters
(CER20 min,
CERMAX,
AUCER01 h, and
AUCER0) were
not affected by the presence of additional erythromycin in the D2 EBT
and were not predictive of erythromycin clearance (CL) on
either day. Instead, we found that indicators of early emergence of
breath radioactivity (1/TMAX,
CER3 min/CERMAX)
from the D1 EBT were significantly correlated. Presumably, early
emergence reflects rapid hepatic uptake and metabolism of erythromycin,
which are requisite for high total clearance. In other words, the
1/TMAX may reflect early drug distribution
and elimination rather than just CYP3A4 metabolism.
Interpolating the breath flux data to yield a model-dependent 1/TMAX(model) value resulted in stronger correlations with CL than with 1/TMAX itself. This is likely to be attributable to the continuity of the latter, whereas TMAX values are limited to actual breath sampling times. Because the peak occurred in all cases before 30 min, a modified EBT using 1/TMAX will remain a practical procedure.
The correlations between both 1/TMAX and 1/TMAX(model) and CL identified one subject as an outlier. It is unclear whether this was a random finding or a particular feature of this subject, who is a frequent user of both alcohol and cannabis. Because the production of carbon dioxide from the oxidation of erythromycin is folate-dependent, a reduction in the folate-dependent intermediate step could potentially explain the apparent delay (i.e., low 1/TMAX) in the EBT test results of the outlier. However, the vitamin B12, serum folate, and erythrocyte folate were all within the normal range for this individual, although the serum homocysteine was slightly elevated (25 µmol/liter; normal range, 7–23 µmol/liter), indicating a mild decrease in functional folate status.
The relationship between 1/TMAX and CL was not significant using the D2 EBT results. Indeed, there was a shift in the TMAX on D2 that varied from subject to subject. Because the EBT was administered midway through the erythromycin infusion, this is likely to represent a protein-binding disequilibrium phenomenon. In any case, in the absence of any effect of the administration of 100 mg of erythromycin on the EBT, it is possible to use the conventional tracer-only EBT without coadministered erythromycin in future studies.
Although the conventional EBT was not predictive of erythromycin CL, there was some evidence of relationship to CYP3A4 activity. A subject receiving the known inducers phenytoin and dexamethasone had a CER20 min value some 3.5 times the median. Also, a subject with extensive metastatic involvement of the liver (as reflected by significantly abnormal liver function tests) had one of the lowest CER20 min values (0.028%/min) and a low erythromycin CL (10.5 liters/h). However, the "induced" subject had a relatively modest erythromycin CL (32.5 liters/h), and the subject with the highest CL value (47.0 liters/h) actually had a low CER20 min value (0.035%/min). It may be that the EBT does correlate with CYP3A4 activity, but because of protein-binding effects and/or the presence of other pathways of elimination, it does not correlate with drug clearance.
Other results show both the fUB of erythromycin and the plasma concentrations of AAG to be highly variable between subjects but interrelated. The CL was correlated to its volume of distribution (VdSS), which indicates a strong influence of protein binding on the metabolism and disposition of erythromycin. This is in agreement to studies performed in cirrhotic patients, in whom erythromycin pharmacokinetics are modified partly through the altered protein binding of erythromycin (24) . Our results indicate that the conventional parameters of the EBT (based on the 14CO2 flux) are likely to be confounded by protein binding, the variability of which may obscure that due to intersubject differences in the activity of CYP3A4. This may be of relevance to EBT studies in other patient populations.
Many anticancer drugs are metabolized by CYP3A4, but there are no readily available techniques for pharmacologically based dosing of individual patients. In cancer patients, our study shows that the EBT is predictive of the clearance of the prototypic CYP3A4 substrate, erythromycin, when using parameters that reflect the kinetics of breath radioactivity. We now intend to prospectively evaluate equations incorporating 1/TMAX for the clearance of anticancer drugs.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported by grants from the New South Wales
Cancer Council and the University of Sydney. This project was an
initiative of the Pharmacology Interest Group of the Sydney Cancer
Centre. 
2 To whom requests for reprints should be
addressed, at Medical Oncology, Sydney Cancer Centre, Missenden Road,
Camperdown, New South Wales 2050, Australia. Phone: 61-2-9515-7376;
Fax: 61-2-9519-1546; E-mail: lrivory{at}canc.rpa.cs.nsw.gov.au
3 The abbreviations used are: CYP3A4,
cytochrome P450 3A4; BYSA, body surface area; EBT, erythromycin breath
test; AAG, 
1-acid glycoprotein; D1, day 1; D2, day 2;
LC, liquid chromatography.
Received 2/25/00; revised 4/28/00; accepted 5/ 4/00.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1-acid glycoprotein and propranolol binding. Clin. Pharmacol. Ther., 32: 659-663, 1982.[Medline]
1-acid glycoprotein and albumin. Clin. Pharmacol. Ther., 32: 295-302, 1982.[Medline]
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. 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]
|
|
|
|
 |
|
 A. J. McLean and D. G. Le Couteur Aging Biology and Geriatric Clinical Pharmacology Pharmacol. Rev., June 1, 2004; 56(2): 163 - 184. [Abstract] [Full Text] [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]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
) and high (•)
erythromycin clearance values (10.6 and 47.0 liters/h, respectively).
···· and 
