Pharmacology of 21-Day Oral Etoposide Given in Combination ...AUC3 of etoposide and hematological toxicities, particularly leukopenia and neutropenia (2-7). One study also examined

Vol. 3, 719-725, May 1997 Clinical Cancer Research 719

Pharmacology of 21-Day Oral Etoposide Given in Combination with

i.v. Cisplatin in Patients with Extensive-Stage Small Cell Lung

Cancer: A Cancer and Leukemia Group B Study

(CALGB 9062)’

Antonius A. Miller,2 Gary L. Rosner,

Mark J. Ratain, Donna R. Hoffis, Mark R. Green,

and Richard L. SchilskyUniversity of Tennessee, Memphis, Tennessee 38163 [A. A. M.];Duke University, Durham, North Carolina 27710 [0. L. R., D. R. H.];University of Chicago, Chicago, illinois 60604 [M. J. R., R. L. S.];and University of California, San Diego, California 92103 [M. R. 0.]

ABSTRACTThis was a pharmacological companion study to a ran-

domized Phase ifi trial comparing 21-day oral versus 3-day

i.v. etoposide in combination with i.v. cisplatin in patients

with extensive-stage small cell lung cancer. Etoposide

plasma concentrations were measured in patients random-

ized to the 21-day schedule and correlated with toxicity and

tumor response.

Patients were treated with etoposide (50 mg/m2/day)

orally for 21 days and cisplatin (33 mglm2/day) i.v. for 3

consecutive days every 28 days for 6 courses. Plasma sam-

plea before the daily etoposide dose (trough concentrations)

and complete blood counts were obtained weekly during

treatment. The average of three etoposide concentrations

(Er) per course was calculated.

Of 158 patients registered to this schedule of the study,

150 were eligible. In 106 patients, etoposide samples wereobtained at least in the first course in which the mean E�

was 039 �ml (SD 0.29). In 102 patients (missing albu-

mm values in 4 of 106 patients), the concentration of etopo-

side not bound to protein (E�) was estimated based on the

following equation: percentage unbound = (1.4 x total bil-

irubin) - (6.8 x albumin) + 34.4. Regression analysis re-

vealed that increasing age was correlated with higher E�

(r 0.27; two-tailed P< 0.01) and E� (r 0.31; two-tailed

P < 0.01). Higher E� and Ef� values were associated with

lower WBC counts and absolute neutrophil counts after the

Received 8/29/96; revised 1/24/97; accepted 1/24/97.The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to

indicate this fact.1 Supported by National Cancer Institute (Bethesda, MD) GrantsCA47555, CA33601, CA41287, CA11789, and CA44691 and by theBristol-Myers Oncology Division. Conducted by the Cancer and Leu-kemia Group B (R. L. S., chairman; central office, Chicago, IL).2 To whom requests for reprints should be addressed, at University ofTennessee, 3 North Dunlap, Memphis, TN 38163. Phone: (901) 448-5817; Fax: (901) 448-5033; E-mail: [email protected].

first treatment course in 83 patients with nadir counts. Usingmultiple linear regression, a pharmacodynamic model was

developed that included E� or E�, age, and alkaline phos-

phatase. An interaction with bone marrow results at Wag-

nosis was found, indicating a sharper decline in nadir counts

with increasing E� or E� when the marrow was involvedwith small cell lung cancer. This model explained 29% of the

variation for WBC nadirs (P < 0.001) and 31 % of thevariation for absolute neutrophil count nadirs (P < 0.001).

Neither E� nor E� showed a significant correlation with

tumor response.

A pharmacokinetic relationship between E� or E�

and age was found. A pharmacodynamic model could bedeveloped for toxicity but not for tumor response.

INTRODUCTION

Clinical oncologists are well aware of the narrow therapeu-

tic index of most antineoplastic drugs, yet our understanding of

the pharmacodynamics of cancer therapy is limited (1). The

desired effect is a reduction in tumor volume, which is generally

optimized by maximizing the dose. Because of the narrow

therapeutic index, most dosing strategies focus on the dose-

related adverse effects of treatment, especially myelosuppres-

sion. Pharmacodynamic studies of antineoplastic drugs are corn-

plicated by the temporal discrepancy between measured drug

concentrations and clinical outcome. Furthermore, patients gen-

erally cannot be observed daily, leading to the maximum ob-

served effect being less than the true maximum effect. Despite

these difficulties, insight into pharmacodynamic relationships

may improve our therapeutic strategies.

Studies of the pharmacokinetics (study of drug concentra-

tion versus time) and pharmacodynamics (study of drug dose or

concentration versus clinical effects) of etoposide have recently

shown promising results. After i.v. administration, relationships

have been noted between steady-state concentrations or the

AUC3 of etoposide and hematological toxicities, particularly

leukopenia and neutropenia (2-7). One study also examined the

duration of exposure to various concentrations of etoposide;

antitumor activity was associated with the maintenance of lower

levels than those found to be associated with hematological

toxicity (7). This led to the hypothesis that the schedule of

etoposide administration may affect efficacy and toxicity and

3 The abbreviations listed are: AUC, area under the curve; CALGB,Cancer and Leukemia Group B; E#{128}�,average of three etoposide concen-trations per course; � the concentration of etoposide not bound toprotein; ANC, absolute neutrophil count; ANC�, ANC (at nadir); WBC,,WBC count (at nadir); Em,,�, sigmoid maximum response model.

Research. on January 19, 2021. © 1997 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

http://clincancerres.aacrjournals.org/

720 Pharmacology of 21-Day Oral Etoposide

that prolonged exposure to low concentrations may improve the

therapeutic index for this phase-specific cytotoxic drug. Because

etoposide is also available in an oral formulation, prolonged

administration (for instance, for 2 1 days) is feasible without the

need for infusion pumps (8). Disadvantages of this approach are

the variable bioavailability of oral etoposide (9-12) and the

availability of only one capsule size (50 mg) in the United

States. Some dose approximations on a schedule of 50 mg/m2/

day for 21 days are therefore necessary. Phase II studies in small

cell lung cancer suggested that the 21-day oral schedule of

etoposide might be more efficacious (13-15). Pharmacody-

namic relationships between etoposide levels and hematological

toxicity (leukopenia and neutropenia) have also been demon-

strated for the 21-day schedule (16-19). Statistically significant

correlations between weekly measurements of E� before a once-

daily dose (trough concentrations) and the nadir leukocyte or

neutrophil counts have been found and expressed in pharmaco-

dynamic models (16-18).

In 1990, the CALOB designed a study that asked the

following questions: (a) Does etoposide have schedule-depen-

dent effects in small cell lung cancer? and (b) Do measurements

of E� during prolonged exposure yield clinically important

correlations with hematological toxicity and tumor response?

This resulted in a randomized Phase III trial testing the schedule

dependency of 2 1-day oral versus 3-day i.v. etoposide in com-

bination with iv. cisplatin in extensive-stage small cell lung

cancer. All patients randomized to the 21-day schedule had

plasma samples taken to measure E�. Pharmacological ap-

proaches, if proven superior, could replace our current practice

of dose normalized for body surface area. Dose normalization

based on body weight or predicted surface area fails to achieve

consistent drug exposure (20, 21). All patients received combi-

nation chemotherapy with etoposide and cisplatin. Although the

cisplatin may have confounded the results regarding clinical and

pharmacological questions about etoposide, a trial of single-

agent etoposide in untreated extensive-stage small cell lung

cancer would not have received wide acceptance in the cooper-

ative group at the time and was subsequently found to be

inferior to combination chemotherapy (22, 23).

The two schedules of etoposide did not result in differences

in treatment outcome with respect to tumor response and sur-

vival (24). A significantly greater rate of severe or life-threat-

ening hematological toxicity was noted on the 21-day oral

etoposide schedule; 83% of patients on this schedule experi-

enced grade 3 or 4 leukopenia during treatment compared to

62% of patients on the 3-day i.v. etoposide schedule (24). This

report provides the results of the pharmacological component of

the study.

PATIENTS AND METHODSEligibility criteria and treatment schedules were published

previously (24). When patients were randomized to the 21-day

schedule, they were also registered to the pharmacological com-

panion trial. Treatment on this schedule consisted of etoposide

(50 mg/m2/day) orally for 2 1 days and cisplatin (33 mg/m2/day)

on days 1, 2, and 3 by i.v. infusion over 1 h in 250 ml of normal

saline or 3% sodium chloride solution (after hydration with at

least I liter of normal saline over 1-2 h) every 28 days for 6

courses. The commercially available, soft gelatin capsules of

etoposide (VePesid) were used for all patients as a once-daily

dose. Bristol-Myers Oncology Division (Princeton, NJ) pro-

vided the drug free of charge for 15 patients who could not

otherwise afford it (IND 38,375). Because etoposide is available

in the United States only in 50-mg capsules, it was necessary to

make some approximations in the calculated daily dose. For

example, if a patient was calculated to receive 85 mg/day, the

amount of etoposide given was 100, 100, and 50 mg on 3

consecutive days, respectively, and the schedule was repeated

for 21 days (i.e., average daily dose = 83.3 mg). Patients

received a calendar with the number of capsules to be taken

written in for each treatment day to insure accuracy and corn-

pliance. No radiation therapy except for brain metastases was

administered while patients received chemotherapy.

Weekly samples were obtained for complete blood counts

and determination of etoposide plasma concentrations. Proce-

dures developed in a previous study at one institution (16) were

followed in this cooperative group study. The time period be-

tween the previous day’s drug administration and blood sam-

pling was 24 h. Blood was collected before the daily dose of

etoposide in heparinized green-top tubes and immediately cen-

trifuged at 1000 x g (2000-3000 rpm in a tabletop centrifuge)

for 10 mm. Plasma was transferred into a polypropylene tube,

stored at -20#{176}C, batched, and sent on dry ice by overnight

express mail to the laboratory at the University of Tennessee.

Plasma was analyzed for E� by high-performance liquid chro-

matography according to a previously published method (16).

The lower limit of detection was 0.05 p.g/ml. The day-to-day

coefficients of variation for etoposide measurements were less

than 10%, as in the previous study (16). The measured values

for the E� in microgram/milliliter plasma were entered into the

central CALOB database that also contained all the clinical

information for this trial. E� was estimated based on an

equation developed by Stewart et a!. (25): percentage un-

bound = (1.4 X total bilirubin) - (6.8 X albumin) + 34.4.

Statistics. Total E� and Ef�� for a given course were

computed by averaging the available data for each patient. We

also generated an adjusted trough concentration to account for

the fact that patients varied the number of 50-mg tablets each

day to approximate a dose of 50 mg/m2. This adjusted concen-

tration statistic, produced by dividing each measured concentra-

tion by the oral dose of etoposide taken by the patient the

previous day, estimated the trough concentration of etoposide

per milligram of oral etoposide. The adjusted average daily

trough concentration of oral etoposide, standardized for body

surface area, was estimated by averaging these adjusted concen-

trations computed for a patient in a given cycle and multiplying

the result by 50, giving an estimated trough concentration during

a treatment cycle for each patient, based on the prescribed daily

dose of 50 mg. Thus, the adjusted average daily trough concen-

tration of etoposide is an estimate of total E� expected for the

patient for every 50 mg of oral etoposide. Furthermore, multi-

plying these latter estimates of a patient’s daily average trough

concentration by the patient’s body surface area produced an

estimate of the average daily trough concentration expected for

that patient if receiving a constant daily dose of 50 mg/m2.

Although daily concentrations correlated with the previous

day’s dose, the average concentration of the cycle exhibited



Clinical Cancer Research 721

much lower correlation. In addition, pharmacodynamic analyses

using adjusted trough concentrations did not differ appreciably

from analyses based on the unadjusted average concentrations,

so we only report results using the average of measured con-

centrations.

The coefficient of variation, expressed as a percentage, was

defined as 100 times the SD divided by the mean. Interpatient

variability for each treatment course summarizes the variability

in patient-specific average concentrations within the course and

was estimated by the coefficient of variation for the patient-

specific average concentrations within the course. Intrapatient

variability across courses, calculated using data from patients

with concentrations from at least three courses, was estimated

by computing for each patient the coefficient of variation of

his/her course-specific average concentrations.

Means are presented with SDs or SEs as indicated. All

statistical significance tests are two-sided, and significance is

assessed at the 5% level. Some analyses used logarithms of data,

such as concentrations or nadir blood counts. Measured concen-

trations equal to 0 p.g/ml were changed to 0.03 before taking

logs, because the lowest detectable value was 0.05 p.g/ml. Nadir

neutrophil counts equal to zero were set to 0.05 X l03/p.1 if the

logarithm was used. Linear regression was used to model rela-

tionships. The E� model was fit using nonlinear regression

with the logarithm of the parameters to ensure positive values.

Nonparametric group comparisons were tested for statistical

significance using the Wilcoxon rank-sum test if examining two

groups and the Kruskal-Wallis test if examining three or more

groups. Correlations across cycles were estimated nonparametri-

cally, based on the ranks of the data (Spearman’s rank correla-

tion).

Pharmacodynamic analyses of nadir counts excluded pa-

tients without monitored blood counts recorded after day 14 of

cycle 1 because the nadir occurred around day 21 . Furthermore,

analyses of concentrations within a cycle or over multiple cycles

of chemotherapy accounted for within-patient correlation using

a linear mixed-effects model (26) fit with the function Ime

written in the S programming language (27). All analyses were

carried out using the S-Plus statistical package (28).

RESULTS

Of 158 patients registered to the oral etoposide schedule,

150 were eligible, and 106 of these had samples for E� available

at least for the first course (Table 1). The mean E� for 106

patients in the first course was 0.39 p.g/ml (SD = 0.29; range,

0.03-1.54 p.g/ml). The mean � for 102 patients for whom

albumin concentrations were recorded was 0.044 p.g/ml (SD =

0.039; range, 0.002-0.220 p.g/ml) in the first course (Table 2).

Interpatient variability in course 1 was 75% for E� and 90% for

� Because dose adjustments could have introduced more

interpatient variability, the study prohibited dose adjustments

for etoposide; only cisplatin doses were adjusted based on

toxicity, including hematological toxicity. Nevertheless, courses

2-6 exhibited higher interpatient variability, ranging from 86-

99% for E� and from 106-125% for � with the greatest

variability occurring in the second treatment course. Overall, the

median interpatient coefficient of variability was 93% for E�

and 109% for Ef�e for courses 1-6. The mean intrapatient

Table 1 Characteristics of 106 patients with E� at least in the firstcourse

Characteristic No.

GenderMale 81Female 25

Age (yrs)<50 650-59 2960-69 4270+ 29

Performance status0 231 442 39

RaceWhite 95Black 10

Other 1Bone marrow involvement

Positive 35Negative 69Unknown 2

Laboratory values recordedBilirubin 106Alkaline phosphatase 106Creatimne 106Albumin 102Neutrophil nadir counts 83

Tumor responseComplete 17Partial, regression 50Stable disease 15Progression 7Unevaluable 17

variability in the first course, expressed as the coefficient of

variation, was 52% (median, 39%; range, 6-173%) for both E�

and Efree. The mean intrapatient variability across courses was

50% (median, 39%; range, 12-198%), examining patients with

at least 3 treatment courses (72 patients for E� and 69 patients

for Ef�e). Correlations for E� and Ef�e averaged 0.60 and 0.64,

respectively, between consecutive courses.

Age was significantly correlated with E� (r = 0.27; P =

0.006) and Efr� (r = 0.31; P = 0.002) in the first course (Fig.

1). An analysis using only the first concentration among patients

receiving 100 mg of oral etoposide on the previous day (73

patients) confirmed this result, indicating that this correlation

was independent of the dose administered on the day before the

E� was determined. Age did not correlate significantly with

measures of renal (i.e. , serum creatinine) or hepatic function

(i.e., bilirubin, aspartate aminotransferase, and alkaline phos-

phatase). E� was not significantly correlated with other meas-

ured patient characteristics (height, weight, body surface area, or

body mass index) or with clinical laboratory values (bilirubin,

aspartate aminotransferase, alkaline phosphatase, or serum crc-

atinine). Liver metastases were detected in 44 of 106 patients

and were correlated with higher alkaline phosphatase values

(P < 0.001). Bone metastases were diagnosed in 51 patients but

did not correlate with alkaline phosphatase values (P 0.46).

E� did not differ significantly by sites of metastatic disease.



In

Fig. 1 Relationship between age (years) and E� or E� (microgram/milliliter plasma).


Table 2 Median, mm imum, and max imum E� by course

Course

Total etopos ide E� Efr,,,

n Median Mm” Max n Median Mm Max

1234S6

1068573634324

0.310.340.340.350.370.32

000000

1.542.702.042.262.761.80

1028170614324

0.0310.0310.0310.0340.0350.027

000000

0.2200.3720.2910.3230.3940.210

a Mm, minim um; Max, maximum.

The median (range) E� values for performance status 0, 1,

and 2 were 0.23 (0-1.08), 0.30 (0-1.26), and 0.39 (0-1.54)

p.g/ml, respectively. For Ef�, the corresponding values were

0.027 (0-0. 13), 0.023 (0-0. 12), and 0.045 (0-0.22) p.g/ml.

These differences by performance status were statistically sig-

nificant for E� (P = 0.023) and Ef� (P = 0.009) when

analyzed without adjustment for administered dose. An analysis

comparing the first concentration across levels of performancestatus among patients who received 100 mg of etoposide the

previous day (73 patients) showed that only E� was signifi-

candy higher among patients with a performance status of 2

versus those with a lower performance status. However, this

finding may be attributable to a confounding relationship be-

tween performance status and albumin. The albumin values

were significantly (P = 0.017) lower in patients with a perform-

ance status of 2. No significant relationship existed between

albumin values and age.

The effects of the previous day’s dose and course among

patients receiving up to six courses (Table 2) of oral etoposide

were explored to answer the question of whether concentrations

tended to change systematically by week within a given course,

by course, or by dose. The results showed only that the concen-

trations increased slightly but significantly (P < 0.01) with

increasing course.

Higher E� and Ef� values were associated with lower

WBCn and ANC� values after the first treatment course in 83

patients in whom weekly counts allowed an estimate of the

nadir. A pharmacodynamic model, using multiple linear regres-

sion, showed that age, drug concentration (E� or E�), and

alkaline phosphatase were significant correlates with observed

WBCn and ANC�. An example of the predictor equations fol-

lows, and details of the model are given in Table 3:

log10(WBC,) = -0.030 - 0.007 X age - 0.0330 X E�

E� was compared by race and gender. Median (range) E�

during the first course among white patients was 0.30 (0-1.54)

�i.g/rnl and was 0.42 (0.16-1.26) among African-American pa-

tients (P = 0.17). For Ef�, the corresponding values were 0.03

(0-0.22) p.g/ml among whites and 0.05 (0.022-0.13) p.g/ml

among African-American patients (P = 0.06). Median (range)

E� during the first course among males was 0.30 (0-1.54)

ii.g/ml and was 0.35 (0. 12-1 .26) among females (P = 0.33). ForE�, the corresponding values were 0.03 (0-0.22) p.g/ml

among males and 0.03 (0.01-0.13) p.g/ml among females (P =

0.72).

+ 0.379 X log10(alkaline phosphatase)

Adjusted average daily trough concentrations were significantly

associated with WBCn and ANC� (multiple R2 = 0.26 for both;

P < 0.01 and P = 0.02, respectively). Using only a single

etoposide concentration obtained 1 week from the start of treat-

ment gave similar results (multiple R2 between 0.23 and 0.27),

although the sample size was smaller. No significant association

between the pretreatment counts and nadir counts for WBC

count and ANC was found. The correlations between the loga-

rithms of pretreatment and nadir counts were 0.12 ( P = 0.289)

and 0. 12 (P = 0.302) for WBC count and ANC, respectively.




Table 3 Pharmacodyn amic models

Variable Estimate SE PRegression

R2

Model for log10 (WBC�)InterceptAgeBclog,0 (alk. phos.Y’

Model for log10 (WBC�)InterceptAgeE�log,0 (alk. phos.)

Model for log10 (ANC�)Intercept

AgeE�

log10 (alk. phos.)Model for log10 (ANC�)

InterceptAgeE�log10 (alk. phos.)

-0.030-0.007-0.330

0.379

-0.127-0.007

-0.3320.409

-0.153

-0.017-0.363

0.530

-0.312-0.017-2.973

0.583

0.3660.0040.1050.145

0.3800.004

0.8090.151

0.583

0.0060.159

0.222

0.5930.0061.2070.225

0.9360.0440.0020.011

0.7390.0580.0050.008

0.7940.0030.026

0.020

0.6000.0050.0160.012

0.25

0.23

0.25

0.27

a a&. phos., alkaline phosphatase.

Significantly different relationships between E� and E�

and nadir counts existed in the data set, depending on whether

the patient did or did not have bone marrow involvement. An

interaction term in the phannacodynamic model indicated that

patients with bone marrow involvement from small cell lung

cancer exhibited steeper declines in WBC count (P = 0.032)

and ANC (P = 0.021) as E� increased. The model using E� is

shown in Table 4 and is graphically illustrated in Fig. 2.

Previous efforts to describe the pharmacodynamics of i.v.

etoposide included sigmoid E,,,,� models (3, 6, 29). A sigrnoid

E,,,.,� model fit to our data of trough concentrations after oral

etoposide and surviving fractions of WBC count or ANC (i.e.,

nadir divided by pretreatment counts) resulted in curves that

seemed linear over the range of average concentrations studied.

The R2 (1 - residual variation/total variation) values corre-

sponding to the fitted E,,.� models ranged between 0.05 and

0.07. The linear regression of the log of the nadir count on E�

alone had an R2 of 0.14 (WBC count) and 0.10 (ANC). There-

fore, we saw no advantage in using an E,,,� model over simple

linear regression. Other previously published pharrnacodynamic

models for 21-day oral etoposide (16-18) were also evaluated

but resulted in regression R2 � 0.07; these models were there-

fore inferior to the models in Tables 3 and 4.

E� and E� were correlated with the various categories of

tumor response, but no significant relationships were found.

Exposure to etoposide above threshold concentrations of 0.1,

0.3, and 0.5 p.g/ml was also evaluated in relation to tumor

response, but without statistically significant results.

DISCUSSIONThis study was conducted in the cooperative group setting

in which 24 member institutions in CALGB submitted plasma

samples to a central laboratory for etoposide measurement.

Although complete data were not available on all 150 eligible

patients, a substantial number of patients had at least 3 samples

Table 4 Pharmacodynamic models with interaction term foruninvolved versus involved bone marrow

Variable Estimate SE PRegression

R2

Model for log10 (WBC�) 0.29

Intercept -0.076 0.387 0.845Age -0.008 0.004 0.032Etree -0.617 1.092 0.574log10 (alk. phos.)aMarrow involved”

0.3820.132

0.1580.098

0.0180.182

Interaction term -3.487 1.534 0.026Model for log,0 (ANC�) 0.31

Intercept -0.068 0.600 0.910Age -0.019 0.006 0.002EfrCI, -0.764 1.624 0.640

log,0 (alk. phos.)Marrow involved”

0.4790.224

0.2350.146

0.0460.130

Interaction term -4.630 2.279 0.046

a alk. phos., alkaline phosphatase.b 1 = yes, 0 = no.

for E� taken in the first course (n = 106), and the corresponding

nadir count was recorded (n = 83). Limitations to the study are

that: (a) etoposide doses varied in patients alternating 50 and

100 mg/day; (b) only trough concentrations of etoposide were

measured; (c) cisplatin may have confounded the pharmacolog-

ical results; and (d) blood counts and E� were only obtained

once a week. Nevertheless, we find the following results re-

markable: a pharmacokinetic relationship between E� or Ef�

and age was found, and a pharmacodynamic model could be

developed for toxicity but not for tumor response.

A consistent finding in pharmacological reports on etopo-

side administered i.v. (2-7) or p.o. (16-19) is the moderate

intrapatient and pronounced interpatient variability in concen-

trations. On the other hand, results from this and other trials

have demonstrated large interpatient variability in clinical re-

sults (24). Therefore, it seemed logical to investigate what

impact the pharmacological variability has on clinical outcomes.

The desired result is turnbr response. Clark et aL were able to

relate the AUC for etoposide to tumor response; however, their

regimen consisted of 5 or 8 days of daily i.v. etoposide (7). The

duration of exposure to concentrations between 0.5 and 2.0

1j.g/ml was correlated with tumor response, whereas the duration

of concentrations above 3.0 p.g/ml was correlated with neutro-

penia. In our trial, trough concentrations between 0.03 and 1.54

1j.g/ml were maintained for 21 days, and a significant correlation

with tumor response could not be demonstrated.

The dose-limiting toxicity for etoposide is leukopenia and

neutropenia. In this CALGB study, 83% of patients treated with

21-day oral etoposide experienced treatment-related leukopenia

and neutropenia of grade 3 or 4 severity (24). Various previous

reports have demonstrated correlations between leukopenia or

neutropenia and concentrations at steady state (2-5) or the AUC

(6, 7) of infusional etoposide. Studies of prolonged oral etopo-

side relied more on trough levels (16-18) or a calculated mean

concentration (19). Indeed, in one study of 21-day oral etopo-

side, the trough concentrations did show a pharmacodynamic

relationship with neutropenia, but the AUC did not (18). In a

preliminary single-institution study, the following pharmacody-

namic model was developed in 27 evaluable patients:



1.0� 0.8

� 0.60.5

0.4

0.3

0.2

0.1

0.0 0.05 0.10 0.15 0.20

Average Free Etoposide


6.0

5.0

4.0

3.0

2.0

Fig. 2 Relationship betweenWBC� and E� depending onbone marrow involvement.

ANC� = 0.32(1 + ANC� X e247 X Ec)

in which ANC� is the pretreatment neutrophil count (16). This

model was subsequently validated in an additional 21 evaluable

patients (17). Zucchetti et a!. (18) independently validated this

model in 39 patients and compared it to their following modi-

fication:

ANC� = 0.44(1 + ANC� x e_2M X EC)

This larger cooperative group study arrived at different

models based on multiple regression of the data accumulated in

the central data base (Tables 3 and 4). Age, which had pharma-

cokinetic importance, was also incorporated in the pharmaco-

dynamic model. The relationship between E� and race reached

borderline significance (P = 0.06), although only 10 African-

American patients were enrolled in the study (Table 1). This

observation warrants attention in future studies. Neither E� nor

Efrec differed significantly by gender or sites of metastatic

disease. Performance status was related to E�, but this relation-

ship was not independent of dose. E� was related to perform-

ance status independent of dose, but this relationship was prob-

ably confounded by albumin concentrations that determined

how much etoposide was bound. This may explain why per-

formance status did not contribute to our pharmacodynamic

model, which is different from a model for i.v. etoposide in

which performance status did play a role (2). Pretreatment

counts also did not explain a significant amount of the residual

variation, contrary to previous models (16-18).

Using the estimated E� in the model generally resulted in

an improvement over using the measured E#{128},.E� was calcu-

lated based on a formula validated by Stewart et a!. that includes

bilirubin and albumin (25). Joel et al. have confirmed that E�

is correlated with measures of liver function, including bilirubin,

albumin, and alkaline phosphatase (30). They also found that

etoposide clearance was significantly correlated with alkaline

phosphatase and bilirubin. In our study, E� was not related to

sites of metastatic disease. However, liver metastasis was asso-

ciated with significantly higher alkaline phosphatase levels. This

may explain why alkaline phosphatase contributed to our phar-

rnacodynamic model.

The fact that significantly different models existed in the

data base depending upon whether the patient was diagnosed

with or without bone marrow involvement from small cell lung

cancer is a novel observation. Although all patients had exten-

sive-stage disease, bone marrow involvement may be an mdi-

cator of total metastatic tumor burden. Alternatively, the steeper

decline of WBC count and ANC with increasing E� in patients

with marrow involvement may indeed reflect a limited capacity

of the residual normal bone marrow to regenerate after chemo-

therapy. Although all these pharmacodynamic models described

statistically significant relationships, the multiple R2 values and

visual inspection of Figs. 2 and 3 indicate that application to an

individual patient is severely limited. Even the model with and

without bone marrow involvement explained only 29 and 31%

of the variation in WBC� and ANC�, respectively. Although the

mathematical models are statistically significant, plasma level

monitoring of etoposide is unlikely to be clinically beneficial.

The National Cancer Institute of Canada recently con-

ducted a retrospective analysis of data on 608 patients with

limited-stage small cell lung cancer to evaluate the prognostic

importance of age (31). Patients were treated with cyclophos-

phamide, doxorubicin and vincristine, and etoposide plus cis-

platin in an immediate or delayed alternating fashion. Response

rates, survival, and hematological toxicity were not influenced

by age. In our previous report ofthe clinical results of this study,

logistic regression was used to examine the joint effect of age,




gender, and performance status and their interactions with treat-

ment on grade 4 toxicity (24). Older patients and patients with

poorer performance status had a significantly greater chance of

experiencing grade 4 toxicity, whereas gender had no significant

impact. We now report a pharmacokinetic relationship between

E� and age that has not been observed before and warrants

further study. In the future, it may be possible to make dosing

recommendations based on age. CALGB is currently conducting

a Phase H study of 21-day oral etoposide in relapsed non-

Hodgkin’s lymphoma with an accrual goal of 82 eligible pa-

tients. This study has the advantage of using oral etoposide as

the single agent. This should allow us to further evaluate the

pharmacodynamic models without the confounding influence of

cisplatin on toxicity and tumor response.

REFERENCES

1. Ratain, M. J., Schilsky, R. L., Conley, B. A., and Egorin, M. J. Phar-macodynamics in cancer therapy. J. Clin. Oncol., 8: 1739-1753, 1990.

2. Ratain, M. J., Mick, R., Schilsky, R. L., Vogelzang, N. J., andBerezin, F. Pharmacologically based dosing of etoposide: a means ofsafely increasing dose intensity. J. Clin. Oncol., 9: 1480-1486, 1991.

3. Karlsson, M. 0., Port, R. E., Ratain, M. J., and Schemer, L. B. Apopulation model for the leukopenic effect of etoposide. Clin. Pharma-col. & Ther., 57: 325-334, 1995.

4. Kunitoh, H., and Watanabe, K. Phase I/H and pharmacologic study oflong-term continuous infusion etoposide combined with cisplatin inpatients with advanced non-small cell lung cancer. J. Clin. Oncol., 12:83-89, 1994.

5. Minami, H., Shimokata, K., Sake, H., Saito, H., Ando, Y., Senda K.,Nomura, F., and Sakai, S. Phase I clinical and pharmacokinetic study ofa 14-day infusion of etoposide in patients with lung cancer. J. Clin.Oncol., 11: 1602-1608, 1993.

6. Stewart, C. F., Arbuck, S. G., Fleming, F. A., and Evans, W. E.Relation of systemic exposure to unbound etoposide and hematologictoxicity. Clin. Pharmacol. & Ther., 50: 385-393, 1991.

7. Clark, P. I., Slevin, M. L., Joel, S. P., Osborne, R. J., Talbot, D. I.,Johnson, P. W. M., Reznek, R., Masud, R., Gregory, W., and Wrigley,P. F. M. A randomized trial of two etoposide schedules in small celllung cancer: the influence of pharmacokinetics on efficacy and toxicity.J. Clin. Oncol., 12: 1427-1435, 1994.

8. Hainsworth, J. D., Johnson, D. H., Frazier, S. R., and Greco, A.Chronic daily administration of oral etoposide: a Phase I trial. J. Clin.Oncol., 7: 396-401, 1989.

9. D’Incalci, M., Farina, P., Sessa, C., Mangioni, C., Conter, V.,Masera, G., Rocchetti, M., Pisoni, M. B., Piazza, E., Beer, M., andCavalli, F. Pharmacokinetics of VP-l6-2l3 given by different admin-istration methods. Cancer Chemother. Pharmacol., 7: 141-145, 1982.

10. Stewart, D. J., Nundy, D., Maroun, J. A., Tetreault, L., and Prior,J. Bioavailability, pharmacokinetics, and clinical effects of an oralpreparation of etoposide. Cancer Treat. Rep., 69: 269-273, 1985.

11. Smyth, R. D., Pfeffer, M., Scalzo, A., and Comis, R. L. Bioavail-ability and pharmacokinctics of etoposide (VP-16). Semin. Oncol., 12(Suppi. 2): 48-51, 1985.

12. Hande, K. R., Krozely, M. G., Greco, F. A., Hainsworth, J. D., andJohnson, D. H. Bioavailability of low-dose oral etoposide. J. Clin.Oncol., 11: 374-377, 1993.

13. Johnson, D. H., Greco, F. A., Strupp, J., Hande, K. R., andHainsworth J. D. Prolonged administration of oral etoposide in patientswith relapsed or refractory small cell lung cancer: a Phase II trial.J. Clin. Oncol., 8: 1613-1617, 1990.

14. Einhorn, L. H., Pennington, K., and McClean, J. Phase II trial ofdaily oral VP-16 in refractory small cell lung cancer: a Hoosier Oncol-ogy Group study. Semin. Oncol., 17 (Suppl. 2): 32-35, 1990.

15. Murphy, P. B., Hainsworth, J. D., Greco, F. A., Hande, K. R.,DeVore, R. F., and Johnson, D. H. A Phase H trial of cisplatin andprolonged administration of oral etoposide in extensive-stage small celllung cancer. Cancer (Phila.), 69: 370-375, 1992.

16. Miller, A. A., lolley, E. A., Niell, H. B., Griffin, J. P., and Mauer,A. M. Pharmacodynamics of prolonged oral etoposide in patients withadvanced non-small cell lung cancer. J. Clin. Oncol., 11: 1179-1188,1993.

17. Miller, A. A., and Tolley, E. A. Predictive performance of a pharma-codynamic model for oral etoposide. Cancer Res., 54: 2080-2083, 1994.

18. Zucchetti, M., Pagani, 0., Tom, V., Sessa, C., D’Incalci, M., DcFusco, M., de Jong, J., Gentili, D., Martinelli, G., Tinazzi, A., andCavalli, F. Clinical pharmacology of chronic oral etoposide in patientswith small cell and non-small cell lung cancer. Clin. Cancer Res., 1:1517-1524, 1995.

19. Minami, H., Ando, Y., Sakai, S., and Shimokata, K. Clinical andpharmacologic analysis of hyperfractionated daily oral etoposide.J. Clin. Oncol., 13: 191-199, 1995.

20. Reilly, J. J., and Workman, P. Normalisation of anticancer drugdosage using body weight and surface area: is it worthwhile? Cancer

Chemother. Pharmacol., 32: 411-418, 1993.

21. Miller, A. A., Mauer, A. M., and Iolley, E. A. Dosing of oraletoposide normalized for body surface area. Cancer Invest., 13: 487-491, 1995.

22. Clark, P. I. Oral etoposide alone is inadequate palliative chemo-therapy for small cell lung cancer: a randomized trial. Proc. Am. Soc.Clin. Oncol., 15: 377, 1996.

23. Harper, P., Underhill, C., Ruiz de Elvira, M. C., Rudd, R., Souhami,R., Spiro, S., Trask, C., Partridge, M., Iobias, J., Ledermann, J., Eraut,D., and James, L. on behalf of the London Lung Cancer Group. Arandomized study of oral etoposide versus combination chemotherapy inpoor-prognosis small cell lung cancer. Proc. Am. Soc. Clin. Oncol., 15:27, 1996.

24. Miller, A. A., Herndon, J. E., Hollis, D. R., Ellerton, J., Langleben,A., Richards, F., and Green, M. R. Schedule dependency of 21-day oralversus 3-day intravenous etoposide in combination with intravenouscisplatin in extensive-stage small cell lung cancer: a randomized Phaseifi study of the Cancer and Leukemia Group B. J. Clin. Oncol., 13:1871-1879, 1995.

25. Stewart, C. F., Fleming, R. A., Arbuck, S. G., and Evans, W. E.Prospective evaluation of a model for predicting etoposide plasmaprotein binding in cancer patients. Cancer Res., 50: 6854-6856, 1990.

26. Crowder, M. J., and Hand, D. J. Analysis of Repeated Measures.London: Chapman and Hall, 1990.

27. Becker, R. A., Chambers, J. M., and Wilks, A. R. The New SLanguage: A Programming Environment for Data Analysis and Graph-

ics. Paciflcs Grove, CA: Wadsworth & Brooks/Cole Advanced Booksand Software, 1988.

28. Statistical Sciences. S-Plus User’s Manual, Version 3.2. Seattle,WA: StatSci, a Division of MathSoft Inc., 1993.

29. Mick, R., and Ratain, M. J. Modeling interpatient pharmacody-namic variability of etoposide. J. NatI. Cancer Inst., 83: 1560-1564,1991.

30. Joel, S. P., Shah, R., Clark, P. I., and Slevin, M. L. Predictingetoposide toxicity: relationship to organ function and protein binding.J. Clin. Oncol., 14: 257-267, 1996.

31. Siu, L. L, Shepherd, F. A., Murray, N., Feld, R., Pater, J., and Zee,B. Influence of age on the treatment of limited-stage small cell lungcancer. J. Clin. Oncol., 14: 821-828, 1996.



1997;3:719-725. Clin Cancer Res A A Miller, G L Rosner, M J Ratain, et al. cancer: a cancer and leukemia group B study (CALGB 9062).

lungwith i.v. cisplatin in patients with extensive-stage small cell Pharmacology of 21-day oral etoposide given in combination

Updated version

http://clincancerres.aacrjournals.org/content/3/5/719

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/3/5/719To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



Documents

Pharmacology of 21-Day Oral Etoposide Given in Combination ...AUC3 of etoposide and hematological toxicities, particularly leukopenia and neutropenia (2-7). One study also examined