5-fluorouracil: results from a multicentric randomized ...changes in the wbc count, neutrophil...

Vol. 4, 2039-2045, September 1998 Clinical Cancer Research 2039

Advances in Brief

Clinical Impact of Pharmacokinetically-guided Dose Adaptation of

5-Fluorouracil: Results from a Multicentric Randomized Trial in

Patients with Locally Advanced Head and Neck Carcinomas’

R#{233}gineFety,2 Fr#{233}d#{233}ricRolland,

Muriel Barberi-Heyob, Agnes Hardouin,

LoIc Campion, Thierry Conroy,

Jean-Louis Merlin, Alain Rivi#{232}re,

Genevieve Perrocheau, Marie-Christine Etienne,

and Gerard MilanoCentre Ren#{233}Gauducheau, 44805 Nantes [R. F., F. R., L. C., G. P.J;Centre Alexis Vautrin, 54511 Vandoeuvre-L#{232}s-Nancy [M. B-H.,T. C., J-L. M.]; Centre Fran#{231}oisBaclesse, 14021 Caen [A. H., A. R.];and Centre Antoine Lacassagne, 06050 Nice [M-C. E., G. M.], France

AbstractA significant link between 5-fluorouradil (SFU) plasma

concentration and its therapeutic activity has been demon-

strated in colon and head and neck cancer patients for SFU

used as a continuous infusion. Dose adjustment based on phar-

macokinetic follow-up has been proposed to decrease hemato-

logical and digestive toxidties, but the dinical impact of thisapproach has not yet been demonstrated. A randomized mad-

ticentric study was conducted to evaluate the dinical Interest of

SFU dose adaptation guided by pharmacokineties. One hun-

dred twenty-two head and neck cancer patients were randomly

assigned to receive induction chemotherapy with cisplatin (100

� thy 1) and SFU (96-h continuous infusion), either at

standard dose (St-arm; 4 g/m2) or at a dose adjusted according

to the SFU area under the curve (AUC0�; PK-arm). In total,106 patients were evaluable for tOXiCity and response. hi the

PK-arm (n = 49), SFU doses and area under the curve were

significanfly reduced during cycle 2 and cycle 3 (P < 0.001) as

compared with the St-arm (n = 57). Grade 3-4 neutropenia

and thrombopenia were significantly more frequent in theSt-arm as compared with the PK-arm (17.5% versus 7.6%,

respectively; P = 0.013). No grade 3-4 mucositis occurred in

the PK-arm, whereas 5.1% was observed in the St-arm (P <

0.01). The objective response rate was comparable in the two

treatment arms: 77.2% in the St-arm versus 81.7% in the

Received 3/13/98; revised 5/27/98; accepted 6/8/98.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 toindicate this fact.I Supported by Produits Roche, the French “F#{233}d#{233}rationNationale desCentres de Lutte Contre Ic Cancer,” and the “Ligue Nationale contre IcCancer, comit#{233}de la Meurthe et Moselle.” These data were presented inpart at the 1997 meeting of the American Society of Clinical Oncology.

2 To whom requests for reprints should be addressed, at Centre Ren#{233}Gauducheau, Boulevard Jacques Monod, 44805 Saint Herblain Cedex-France. Phone: 33-2-40-67-99-60. Fax: 33-2-40-67-97-05. Email:[email protected].

PK-arm. The present study is the first to demonstrate, in a

randomized design, the dinical interest of an individual SFU

dose adaptation based on pharmacokinetic survey, in terms of

therapeutic index improvement.

IntroductionDespite being one of the oldest anticancer drugs, 5FU� is

still increasingly used in cancer chemotherapy. SFU is consid-

ered not only as the standard drug for the treatment of advanced

colorectal cancer, but also as one of the major drugs for the

treatment of breast and head and neck carcinomas (1-3). Al-

though 5FU is a prodrug, significant relationships have been

evidenced between SRi plasma concentrations, drug pharma-

codynamics, side effects, and antitumor activity (4, 5). !n this

context, critical thresholds for 5FU plasma concentrations at risk

for toxicity have been demonstrated. Indeed, Au et a!. (6) noted

that, during a 5-day continuous infusion, steady-state SRi

plasma concentrations above 1 .5 p�M were significantly associ-

ated with leukopenia. Most data were obtained from studies

conducted in head and neck cancer patients treated by SRi-

based chemotherapy as induction treatment. Thyss et al. (7)

showed a significant relationship between elevated SRi AUC

(over 30,000 ng.h/ml) and the frequency of cycles with toxicity

(myelosuppression, mucositis, and diarrhea). More recently,

Vokes et a!. (8) confirmed these observations. The next logical

step was to use these pharmacokinetic data to propose individ-

ualized SRi dose adjustment. Santini et a!. (9) adopted such a

strategy in head and neck cancer patients. They proved that the

incidence of SRi-related side effects could be reduced by mdi-

vidually controlling SFU AUC at mid-cycle during the 5-day

continuous infusion, and, if required, by modifying the SRi

dose using a simple diagram. These data were further confirmed

by other authors (10). However, these last studies (9, 10) pre-

sented methodological limitations because they were based on

historical controls as comparators to SRi dose-adapted patients.

For this reason, it was decided to conduct a multicentric study in

which the clinical interest of SF0 dose adaptation guided by

pharmacokinetics could be evaluated on the basis of a random-

ized trial. Patients with locally advanced head and neck cancer

from three different institutions were enrolled in this study and

were randomly assigned to induction chemotherapy by cisplatin

and SRi at St-arm or at PK-arm.

Patients and Methods

This prospective study involved 122 patients (1 14 men and

8 women), with a median age of 55 years (range, 29-72). All

3 The abbreviations used are: SRi, 5-fluorouracil; St-arm, standard 5FUdose; PK-arm, adjusted SRi dose; AUC, area under the curve; CR,complete response; PR, partial response; ST. stabilization.

Research. on March 26, 2020. © 1998 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

http://clincancerres.aacrjournals.org/

St-arm

(n 61)

57

525

54(29-72)

1634

7

216

1410

60

83116

13112211

PK-arm(n 61)

49

48

55 (36-69)

1135

3

203

17

53

02

4070

169

1410

2040 Clinical Response after SFU Monitoring

Table 1 Pretreatment characteristics of patients randomly assigned toSt-arm or PK-arm

Characteristic

No. of evaluable patientsGender

MaleFemale

Median age, yr (range)Performance status (PS)

0

Primary siteOropharynxOral cavityHypopharynxLarynxDouble localizationUnknown

Stage of primary tumorT,T2T3T4Not determined

Node stageN0N1N2N3

patients had received no previous treatment and had histologi-

cally proven locally advanced squamous cell carcinoma of the

head and neck. Patients’ characteristics are listed in Table 1. The

eligibility criteria included age between 18 and 75 years; per-

formance status grade 0, 1, or 2; measurable disease, life cx-

pectancy of at least 3 months; adequate bone marrow function

(WBC >4.0 x l0�/l, neutrophil count >2.0 X iO�fl, and

platelet count >100 X l0�Il); adequate liver function (< twice

normal); and renal function (serum creatinine level <130 p�M).

Ethical committee approval was obtained and all patients gave

written informed consent.

Treatment Regimens. The treatment consisted of three

cycles of 5FIJ-cisplatin delivered every 2 weeks (11), followed

by response assessment and subsequent surgery and/or radiation

therapy. All patients were treated with curative intent. Patients

were randomly assigned to receive 5FU-cisplatin induction

chemotherapy with either SRi at St-arm or 5FU at PK-arm. In

both arms, and at day 1 of each cycle, the treatment consisted of

a 6-h prehydratation with 2 1 of 5% dextrose containing 6 g/l of

NaCl, 2 g/l of KC1, followed by cisplatin at a dose of 100 mg/m2

diluted in 500 ml of 10% mannitol and administered on a 2-h

infusion. Therapy with SRi started immediately after comple-

tion of cisplatin. SRi was infused i.v. during 96 h at a constant

rate with a volumethc pump.

In St-arm, SRi was delivered at a fixed dose of 4 g/m2/

cycle. Doses were modified according to pretreatment blood cell

counts and other toxicities. Patients with a WBC count �4.0 X

l0�Il, a neutrophil count �2.0 X i0911, or a platelet count�l00 x iO�/i received 100% of the dose. Patients with a WBC

count of 3.0-3.9 X i0911, a neutrophil count of 1.5-1.9 X iO�/l,

or a platelet count of 75-99 X i0�/l were given a total dose of

4 g/cycle. Patients with a WBC count �2.9 X l0�Il, a neutrophil

count �l.4 X !0�Il, or a platelet count �74 X l0�/l had

treatment course delayed for 1 week.

In PK-arm, patients received SRi at a dose defmed by the

AUC measured individually during treatment (12). During the

first cycle, the AUC was calculated from 0-48 h (AUCO_48h)

and during the 96-h infusion (AUCO�h) to identify patients

who accumulate SRi. For patients with an AUC0� lower than

3 X AUCO_�h, the SRi dose was adapted for cycles 2 and 3

according to diagram 1 (Fig. 1). For patients with an AUCO_�h

higher than 3 X AUCO_48h, the dose was adapted for cycles 2

and 3 according to diagram 2 (Fig. 1). During this first cycle,

SRi doses were not adapted, except when AUCO_�h was

greater than 20,000 ng.h/ml, because this value has been previ-

ously reported to be highly predictive for toxicity (12). In this

case, SRi infusion was stopped at mid-cycle. The SRi dose

administered at the beginning of cycle 2 was determined by

using diagrams 1 or 2 according to the AUCO_48h calculated at

cycle 1. The 5FU dose delivered at the beginning of cycle 3 was

the dose administered during cycle 2. During cycle 2 and cycle

3, AUCO_�h was controlled at mid-cycle and modified when

necessary. The initial dose was decreased from 15-50% when

AUCO_48b values ranged from 15,600-20,000 ng.h/ml for dia-

gram 1 and from 8,640-16,000 ng.h/ml for diagram 2. The dose

was increased (0-50%) when the AUCO.48h value was low

(below 10,400 ng.h/ml for diagram 1 and below 5,760 ng.h/ml

for diagram 2). An increase in SRi dose was possible only if the

changes in the WBC count, neutrophil count, or platelet count

were lower than 25% as compared with the previous cycle.

Without a dose reduction predicted by pharmacokinetics, and in

case of a WBC count at 3.0-3.9 X i0�/l, a neutrophil count at

1.5-1.9 X !0�/l, or a platelet count of 75-99 X i0�/l, the SRi

dose administered was a 4 g total dose/cycle. In case of a WBC

count �2.9 X iO�/l, a neutrophil count �1 .4 X iO�/l, or a

platelet count �74 X l0�Il, treatment was delayed for 1 week.

In the both arms, in case ofmucositis grade 1, the full dose

of SRi was reduced to a 4 g total dose/cycle, and in case of

grade higher than 1, treatment was delayed for 1 week. In case

of other toxicities and a serum creatiine level greater than 130

ELM, treatment was delayed for 1 week.Response and Toxicity Assessments. Assessments of

response and toxicity were performed every 2 weeks during

chemotherapy and 2 weeks after the completion of treatment. At

each visit, a physical examination was performed, with blood

cell counts and measurement of the serum creatiine level.

Response was evaluated by the product of the two perpendicular

lesion diameters, based on clinical, fibroscopic, ultrasound

and/dr computed tomography scan examination. CR was de-

fined as the disappearance of all lesions. PR was defined as a

tumor regression exceeding 50% without occurence of new

lesions. ST was defined as any tumor regression less than 50%.

Progression was defined as a tumor increase of more than 25%,

or as the occurence of any new lesion. Hematological and

nonhematological toxicities were evaluated 10 days and 14 days

after each cycle using the WHO criteria. Nonhematological

toxicities included mucositis and digestive tract toxicity (includ-

ing nausea, vomiting, and diarrhea).



150

50

0 10,4os 15,650 20,000 h/mi

AUCO-48h5,160 8,640

AUCO-48h

Clinical Cancer Research 2041

DIAGRAM I

% of initial 5FU dose

DIAGRAM 2

% of init�aI 5FU dose

1�,OOOng him

Fig. 1 Diagram for SRI dose delivery in patients having AUCO_%h lower than 3 X AUCo_�h (Diagram 1) or higher than 3 X AUCO_�h (Diagram2). The SRi doses were modified when AUCo_�h were above or below target AUCs at 13,000 ng.h/ml and 7,200 ng.h/ml for Diagram 1 and 2,

respectively. At the beginning of cycle 2, the SRi dose delivered was defmed according to AUCo_�h calculated at cycle 1. The initial dose wasdecreased by 15-50% when AUCO48h values ranged from 15,600-20,000 ng.h/ml for Diagram 1 and from 8,640-16,000 ng.h/ml for Diagram 2.The dose was increased (0-50%) when the AUCO_48h value was below 10,400 ng.h/ml for Diagram 1 and below 5,760 ng.h/ml for Diagram 2. Thedose delivered at the beginning of cycle 3 was the total dose administered during cycle 2. During cycle 2 and cycle 3, AUCO_48h was controlled atmid-cycle and the SRi dose was modified when necessary.

Pharmacokinetic Investigations. Plasma SRi levels

were measured in all patients included in the study. Two blood

samples/day were collected from days 1-4. Blood sampling on

day 1 was performed before 5FU administration (control prior

chemotherapy) and 2 h after SF0 infusion; on the following

days, blood sampling was done as follows: on days 2 and 3, at

8 a.m. and 5 p.m.; on day 4, at 8 a.m., and just before the end

of SRi infusion. Venous blood (5 ml) was collected into hep-

arinized tubes. The tubes were immediately centrifuged at 3000

rpm for 10 mm at 4#{176}C.Resulting plasma supernatants were

transferred to individual 3-mi polypropylene tubes and stored

frozen in the hospitalization ward. The frozen tubes were trans-

ferred to the laboratory once a day and stored at -20#{176}C.SRi

plasma concentrations were analyzed by high-performance liq-

aid chromatography (13) using UV detection (X = 262 nm). The

limit of sensitivity was 25 ng/ml. Intra- and interassay variations

were lower than 10%. The AUC was determined according to

the trapezoIdal rule. The AUC was calculated over the first part

of the cycle (AUCO48h) and during the whole pharmacokinetic

follow-up (AUCO_96h). For each patient, the pharmacokinetic

parameters representative of SFU exposure during the treatment

were the average AUCO_%h (sum of AUCO_%h for each cycle

divided by the number of cycles) and the AUCO_96h intensity

(sum of AUCO_�h for each cycle divided by the number of

weeks of treatment). Likewise, the average SFU dose/cycle

(sum of SRi dose for each cycle divided by the number of

cycles) and the 5FU dose intensity (sum of 5FU dose for each

cycle divided by the number of weeks of treatment) were

calculated for each patient.

Statistical Considerations and Analysis. The primary

end point of the study was the incidence of hematological

toxicity. The procedure tested a null hypothesis (HO) that the

occurrence of grade 3-4 hematological toxicity was similar in

both treatment groups versus the alternative hypothesis (Hi)

that occurrences were different between patients in the St-arm

and the PK-arm. A sample size of 61 patients/arm afforded

approximately 80% power to detect a 20% difference in grade

3-4 hematological toxicity between the two arms. Randomiza-

tion was stratified by center (3 centers). Categorical data were

analyzed using the x2 test. Continuous data were analyzed using

Student’s t or Mann-Whitney U two-sided tests.

The secondary end point was the equivalence of tumoral

response. The second procedure tested a null hypothesis (KO)

that the occurrence of objective response rate was different in

the two treatment arms (difference higher than 10%), versus the

alternative hypothesis (Ki) that the occurrence was identical in

the St-arm and the PK-arm. Objective response rates were

compared with the Dunnett and Gent one-sided test (14).

The statistical significance was considered at P < 0.05.

Results

Patient Characteristics. Among the 122 patients ran-

domly assigned to one of the two treatment arms, 16 patients

(13%) were subsequently found inevaluable for response and

toxicity (4 patients in the St-arm, and 12 patients in the PK-

arm). One patient had concomitant treatment with an alkylating

agent due to altered blood marrow function (polyglobulia), and

15 patients had major protocol violations: 3 patients had con-

comitant cisplatin dose reduction; 1 patient was adapted using

the diagram 1 instead of the diagram 2; 4 patients had an

incorrect dose reduction; S patients had an incorrect dose in-

crease; and 2 patients had both incorrect SRi dose increase and

dose reduction. Incorrect dose was considered as a percentage of

deviation from the recommended protocol dose higher than

20%. Of the remaining 106 evaluable patients, 57 were random-

ized to the St-arm, and 49 were randomized to the PK-arm. The

gender and age distribution between the two arms was not

significantly different. The distribution of patients within each



2042 Clinical Response after SRi Monitoring

Table 2 Doses and mean percentage of initial planned dose received/cycle for patients with St- arm or PK-arm

Cycle 1 Cycle 2 Cycle 3

n Dose (mg) ± SD (%) n Dose (mg) ± SD (%)n Dose (mg) ± SD (%)

CisplatinSt-arm 57 171 ± 17(100) 52 170± 18(100) 49 170±18(100)

49 174 ± 17 (99.8) 45 174 ± 16 (99.8) 41 175 ± 16(99.8)P NSb NS NS

SRiSt-arm 57 6798 ± 801 (99.1) 52 6692 ± 994 (98.1) 49 6230 ± 1242 (91.6)

PK-arm 49 6683 ± 1219 (96.0) 45 51 18 ± 670 (72.6) 41 4855 ± 1670 (68.9)P NS P<0.00l P<0.001

a The percentage of cisplatin dose was 99.8% for the three cycles because one obese patient received at each cycle 95% of the initial dose.b NS, not significant.

Table 3 5FU AUCO_%h (ng . Wmi) received/cycle

Cycle 1 Cycle 2 Cycle 3

St-arn� 31,220 ± 12,414 (n 53) 33,389 ± 1 1,476 (n 50) 31,081 ± 12,551 (n 42)PK-��” 29,525 ± 10,673 (n = 45) 24,909 ± 8,491 (n 42) 23,539 ± 7,492 (n 40)

P NS P<0.OOl P<0.0l

a Due to analytical problem, AUC0_�,.J, was not evaluable for all patients.b NS, not significant.

stratification variable (performance status, tumor site, tumor

stage, and node stage) was also comparable (Table 1).

Cisplatin and SFU Doses. The initial cisplatin dosereceived/cycle was not significanfly different between the

two treatment arms (Table 2).

Considering SRi, the percentage of the SRi initial planned

dose was sigmficanfly different in the two groups of treatment

during cycle 2 and cycle 3 (P < 0.001). In PK-arm, the changes

were more pronounced (72.6% for cycle 2 and 68.9% for cycle

3) as compared with the St-arm (98.1% for cycle 2 and 9 1.6%

for cycle 3). In the St-arm, changes corresponded to reductions

in SRi delivered dose according to pretreatment toxicity (3.9%

and 20.9% of the cycles had SRi dose reductions during cycle

2 and cycle 3, respectively). In PK-arm, changes were due either

to a decrease or to an increase in SRi doses according to the

pharmacokinetic protocol for dose adaptation, or to reductions

in SRi-delivered doses according to pretreatment toxicity. Dur-

ing cycle 2, SRi dose reductions occurred in 66.6% of the

cycles and dose increases occurred in 8.8% of the cycles. During

cycle 3, the changes occurred in 78.0% of the cycles (dose

reduction) and in 4.8% of the cycles (dose increase).

When dose modifications were performed, reductions

ranged from -40% to -50% in the St-arm and from -40 to

-48% in the PK-arm. The percentages of dose increases in the

PK-arm were low: 6% and 7% for cycle 2 and cycle 3, respec-

tively.

SRi AUC/Cycle. SRi AUCO_�h were available for pa-

tients in the St-arm and the PK-arm (Table 3). During cycle 1�

there was no significant difference between the two treatment

arms (AUCO_96h 31,220 ± 12,414 ng.h/ml and 29,525 ±

10,673 ng.h/ml for the St-arm and the PK-arm, respectively).

During cycle 2 and cycle 3, AUCO_%h were significantly lower

in the PK- arm as compared with the St-arm (P < 0.01). There

was less interpatient variability for AUC values in the PK-

guided arm in comparison with the ST-arm (coefficient of

variation: 31% versus 40%, P < 0.001, cycle 3).

Toxicity. Hematological and mucosal toxicities attribut-

able to SRi were significantly reduced in the PK-arm as com-

pared with the St-arm (Fig. 2). Grade 3-4 neutropenia and

thrombocytopenia were observed in 17.5% of the cycles in the

St-arm as compared with only 7.6% in the PK-arm (P = 0.013).

A complete absence of mucositis was observed in the PK-arm,

whereas 5.1% ofthe cycles in the St-arm experienced grade 3-4

toxicity (P < 0.01). There was no significant difference in

digestive tract toxicity between the two treatment arms.

Due to SRi-related toxicity, three patients (5.2%) in the

St-arm left the study before completing chemotherapy: two

patients after cycle 1 and one patient after cycle 2. In the

PK-arm, three patients (6,1%) left the study before completing

chemotherapy: one patient after cycle 1 and two patients after

cycle 2. As for cycle delay, the incidence was lower in the

PK-arm (14.9% of cycles, corresponding to 165 days) as com-

pared with the St-arm (21.2% of cycles, corresponding to 213

days), but the difference was not significant (P = 0.269).

Response. Patients (77.2%) in the St-arm (19 of 57 or

33.3% CR; 25 of 57 or 43.9% PR) and 81.7% of patients in the

PK-arm (14 of 49 or 28.6% CR; 26 of 49 or 53.0% PR)

responded to treatment. The objective response rates were sig-

nificantly equivalent (Dunnett and Gent, P = 0.03), with upper

90% confidence limit equal to 8.7% (Fig. 3).

Link between Clinical Events and Pharmacological

Data. Table 4 illustrates the link between hematological tox-

icities and mucositis with average AUC0�,/cycle and SFU

doses. The analysis was performed on the whole study group

(n 106 patients). All patients received one, two, or three

cycles. The mean value for the average AUCO�h was signifi-

cantly related to the severity of hematological toxicity (grade

3-4; P = 0.035). No significant relationship was found with



20� 17.5%

1�

10.

5’

(I)Ui-a

0U.0

5.1%

0

thrombocyto�-3.4

. St-arm

EIPK.aml

R

E1

8.1%

D�ttvs Toxicitygrad. 3-4

Fig. 3 Response rate 2 weeks after completionof 5FU-cisplatin chemotherapy for patients with

St-arm (0) and for patients with PK-arm (El).Test for equivalence significant: P = 0.03 (Dun-nett and Gent test).

77.2% 81.7%

53.0%

Coniplet. PMIaI

Rssponss RsponsObjectiveResponse(CR + PR)


Fig. 2 Grade 3-4 toxicity for patients with St-arm (EJ) and for patients with PK-arm (El). P0.013 for neutropenia and/or thrombocytopenia;P < 0.01 for mucositis; P not significant fordigestive toxicity (including nausea, vomiting,

and diarrhea).

100.

Cl) 70.I-.zUi

� 30�

20�

10.

0-

SRi dose. For mucositis, average AUCO_96h was higher in

patients with severe toxicities (grade 3-4), but the difference

was not significant as compared with patients with no or mod-

crate toxicity (grade 0-2).

Regarding objective response (Table 5), no significant as-

sociation was observed with available pharmacokinetic param-

eters (average, AUCO96h and AUCO_96h intensity) and SRi

doses (average dose/cycle and dose intensity).

DiscussionThe current role of chemotherapy in head and neck carci-

nomas is presently focused on the treatment of locally advanced

stages, for which continuous infusion SRi and cisplatin chem-

otherapy is considered as a reference protocol (15). This proto-

col has demonstrated a high response rate (80%), but severe

toxicities have been reported. Several investigators have shown

that interindividual variability in SRi clearance and AUC are

linked to the risk of severe toxicities (7, 16-18). These phar-

macokinetic parameters have also been reported to be correlated

with response and survival in patients receiving 5FU-cisplatin

induction chemotherapy (8, 19). In the present study, it was

Table 4 Analysis of clinical toxicity as a function of 5FU

pharmacological parameters (all patients)”

Average AUC�%h/

cycle (ng . h/ml)

Average dose/cycle

(mg)

Hematologic ToxicityGrade 0-2 27,622 ± 7,752 (65)” 6,104 ± 1,339 (75)Grade 3-4 3 1,45 1 ± 6,924 (26)”

P - 0.035

6,070 ± 980 (31)

NSC

Mucositis

Grade 0-2 28,866 ± 8,066 (86)” 6,098 � 1,248 (100)Grade 3-4 34,285 ± 6,814 (5)”

NS

6,225 ± 962 (6)

NS

a Values are given as mean ± SD (number of patients).b Average AUC#{216}_��/cyc1e in I 5 patients was not evaluable.

C NS, not significant.

possible to confirm the relationship between SFU pharmacoki-

netics and pharmacodynamics. Cycles with grade 3-4 hemato-

logical toxicity had significantly higher SRi systemic exposure

than cycles with grade 0-2 hematological toxicity. This obser-

vation is consistent with a study by Vokes et al. (8) who noted



2044 Clinical Response after SRi Monitoring

4 M. C. Etienne, E. Chatelut, M. Lavit, A. Pujol, P. Canal, and 0.

Milano, unpublished data.

Table 5 Analysis of clinical response as a function of SRi pharmacological parameters”

Average AUC�,�,/cycle (ng ‘ h/mi)

AUC�.%h intensity(ng ‘ h/mlJw)

Average dose/cycle (mg)

Dose intensity(mg/w)

Responders”NonrespondersP

29,065 ± 8,174 (75)C

28,180 ± 5,636 (l6)CNS”

15,766 ± 4,774 (75)C

16,401 ± 3,348 (l6)’�NS

6,133 ± 1,138 (87)5,887 ± 1,651 (19)

NS

3,378 ± 860 (87)3,425 ± 1,049 (19)

NSa Values are given as mean ± SD (number of patients).b Responders, patients with complete or partial responses; nonresponders, patients with stable or progressive disease.C Average AUC#{216}_%,,/cycle and AUC0_�F.J, intensity in 15 patients were not evaluable.d NS, not significant.

a significant inverse correlation between neutrophil nadir values

and SRi plasma concentration, and only a poor correlation

between SRi pharmacokinetics and mucositis. In the present

study, no link was found between SRI pharmacokinetics and

mucositis. As for tumoral response, no difference in average

AUCO_96h and AUCO_%h intensity was observed between pa-

tients who achieved a CR or PR and patients who had stable

disease or progression. This result is not consistent with previ-

ous studies (8, 19) that demonstrated that response and survival

were significantly related to high SRi plasma concentrations in

patients with head and neck cancer. This discrepancy can be

explained by the fact that in these previous studies, the patient-

to-patient variability may have been larger than in the present

one. Therefore, a closer relationship between tumor response

and SRi systemic exposure could be evidenced.

SRi therapeutic monitoring with SRi dose adjustment was

previously performed, and comparisons with historical groups

suggested that the SRi therapeutic index could be improved

with SRi dose adaptation based on pharmacokinetics (9). To

our knowledge, the present study is the first randomized trial to

investigate the clinical impact of pharmacokinetically-guided

dose adaptation. This work was centered on SRi clinical phar-

macology only, despite the potential role played by cisplatin in

the pharmacodynamics data analyzed herein. In fact, cisplatin

exhibits antitumor activity against head and neck tumors (17%

response rate; Ref. 20), and experimental data strongly suggest

that cisplatin enhances SRi cytotoxicity by indirecfly increasing

the intracellular levels of reduced folates (21, 22). In the present

study, the impact of individual SRi dose adaptation based on

pharmacokinetics on clinical outcome is striking. The present

randomized trial clearly demonstrates that patients with individ-

ually adapted SRi dose rate develop significanfly lesser toxic

manifestations than patients receiving a fixed SRi dose. In the

PK-arm, grade 3-4 neutropenia and thrombocytopenia were

reduced by more than 50%, and no grade 3-4 mucositis was

observed. These results reflect the dose reduction applied in the

PK-arm and the subsequent reduction in AUC values (Table 3).

SRi dose modifications were mainly dose reductions (66% and

78% of cycles 2 and cycles 3, respectively) and with few dose

increases (8.8% of cycles 2 and 4.8% of cycles 3), thus indicat-

ing that the target AUC in this protocol is very close to the

maximal tolerated dose.

The overall response rate remained within the range of

those usually observed in these tumors (iS). It should be noticed

that the SRi dose modifications performed did not alter the

response rate. In a retrospective study, Santim et a!. (1989)

reported similar results in monitored patients (9).

Previous work (9) shows that for the S-day continuous

infusion of SRi combined with cisplatin, the AUC predictive

for toxicity is 30 000 ng.h/ml. Our results (Table 4) for the 4-day

SRi infusion confirm that high-grade hematological toxicity is

associated to SRi AUC higher than 30,000 ng.h/ml (average

AUCO�h 31,451 ± 6,954 ng.h/ml). Taken together, these

data demonstrate that for 5- and/or 4-day continuous infusion of

SRi, the target SRi AUC for an optimal dose is 30,000 ng.b/ml.

An alternative approach based on pretreatment determination of

biological parameters instead of pharmacokinetics should be

investigated to anticipate individual SRi clearance and SRi

AUC. Such a strategy has already been applied successfully to

carboplatin with individual dosage calculation based on pre-

treatment glomerular clearance determined directly (23) or in-

directly (24). As for SRi clearance, correlations have been

demonstrated between this parameter and pretreatment hepatic

function test (serum alkaline phosphatase level), age, and the ac-

tivity of dihydropyrimidine dehydrogenase (25, 26, 27), the main

enzyme involved in the catabolism of SRi. However, these corre-

lations were too weak for a reliable prediction of SRi clearance.4

Interestingly, the development of new dihydropyrinildine dehydro-

genase inhibitors (such as ethinyluradil), when combined with SRi,

lead to the elimination of SRi mainly by kidney route and to a

decrease of the interindividual variability in SRi clearance. In this

case, it would be possible to predict SRi clearance based on

creatinine clearance and thus to evaluate the optimal individual

SRi dose before starting treatment (28).

For other drugs than anticancer agents like aminoglyco-

sides, the cost-saving benefit of therapeutic drug monitoring has

been demonstrated (29). In the present study, a pharmaco-

economic analysis was undertaken in a subgroup of 82 patients

(41 patients in each arm; data not shown). The medical cost

included three management groups. The first group included

hospitalization and patient care required to manage toxic events.

The second group included costs linked to SRi plasma concen-

tration monitoring. The third group included cost linked to

calculation of SRi dose adjustment. Overall, our analysis dem-

onstrated a reduction of 14.6% in the medical cost in the

PK-arm. More precisely, the respective costs (U. S. dollars) for

the ST-arm and the PK-arm were, $21,758 and $6,803 for

toxicity management; $0 and $13,798 for SRi monitoring; and

$16,835 and $12,325 for SRi dose adjustment. This information




indicates that economic approach should be included as an end

point in clinical pharmacokinetic studies.

In conclusion, this prospective study strongly suggests that

the exploitation of SRi pharmacokinetic-pharmacodynamic re-

lationships could improve cancer chemotherapy of locally ad-

vanced head and neck carcinomas. We hope that the present

study would show the interest, for other anticancer drugs, of

prospective studies conducted on the basis of randomized trials,

to establish the clinical value of individual drug dosage based on

real time pharmacokinetic follow-up.

References1. Sobrero, A. F., Aschele, C., and Bertino, J. R. Fluorouracil in

colorectal cancer-a tale of two drugs: implications for biochemicalmodulation. J. Clin. Oncol., 15: 368-381, 1997.

2. Harris, J. R., Morrow, M., and Bonnadonna, 0. Cancer of the breast.in: V. T. Dc Vita, S. Heilman, and S. A. Rosenberg (eds.), Cancer:

Principles and Practice of Oncology, Vol. 2, Ed. 4, pp. 1264-1332.Philadelphia: Lippincott, 1993.

3. Kish, J. A., Ersley, J. F., Jacobs, J., Weaver, A., Cummings, G., andAl Sarraf, M. A randomized trial of cisplatin + 5-fluorouracil infusionand cisplatin + 5-fluorouracil bolus for recurrent and advanced squa-mous cell carcinoma of the head and neck. Cancer (Phila.), 56: 2740-2744, 1985.

4. Trump, D., Egorin, M., Forest, A., Wilson, J., Remick, S., andTutsch, K. Pharmacokinetic and pharmacodynamic analysis of fluorou-racil during 72 hours continuous infusion with and without dipyri-damole. J. Clin. Oncol., 9: 2027-2035, 1991.5. Barberi-Heyob, M., Weber, B., Merlin, J. L., Dittrich, C., de Bruijn,E. A., Luporsi, E., and Guillemin, F. Evaluation ofplasma 5-fluorouracilnucleoside levels in patients with metastatic breast cancer: relationshipswith toxicities. Cancer Chemother. Pharmacol., 37: 110-116, 1995.

6. Au, J., Rustum, Y., Ledesma, E., Mittelman, A., and Creaven, P.Clinical pharmacological studies of concurrent infusion of 5 fluorouraciland thymidine in treatment of colorectal carcinomas. Cancer Res., 12:

2930-2937, 1982.

7. Thyss, A., Milano, G., Ren#{233}e,N., Vallicioni, J., and Schneider, M.Clinical pharmacokinetic study of5FU in continuous 5-day infusions forhead and neck cancer. Cancer Chemother. Pharmacol., 16: 64-66,1986.8. Vokes, E. E., Mick, R., Kies, M. S., Dolan, M. E., Malone, D.,Athanasiadis, I., Haraf, D. J., Kosloff, M., Weichselbaum, R. R., andRatain, M. J. Pharmacodynamics of fluorouracil-based induction chem-otherapy in advanced head and neck cancer. J. Clin. Oncol., 14: 1663-1671, 1996.

9. Santim, J., Milano, G., Thyss, A., Ren#{233}e,N., Viens, P., Schneider,M., and Demard, F. SRi therapeutic monitoring with dose adjustmentleads to an improved therapeutic index in head and neck cancer. Br. J.Cancer, 59: 287-290, 1989.

10. Withm, J., Lev#{233}que,D., Velten, M., and Klein, T. Surveillancepharmacocin#{233}tique avec adaptation de posologie du 5 fluorouracileadministr#{233}en perfusion continue. Bull. Cancer, 80: 439-445, 1993.11. Beauvillain, C., Mah#{233},M., Bourdin, S., Peuvrel, P., Bergerot, P.,Rivi#{232}re,A., Vignoud, J., Deraucourt, D., and Wesoluch, M. Final resultsof a randomized trial comparing chemotherapy plus radiotherapy withchemotherapy plus surgery plus radiotherapy in locally advanced re-sectable hypopharyngeal carcinomas. Laryngoscope, 107: 648-653,1997.

12. Fety, R., Rolland, F., Barben-Heyob, M., Merlin, J. L., Conroy, T.,Hardouin, A., Rivi#{232}re,A., and Milano, 0. Clinical randomized study ofSRi monitoring versus standard dose in patients with head and neckcancer: preliminary results. Anticancer Res., 14: 2347-2352, 1994.13. Christophidis, N., Mihaly, 0., Vadja, F., and Louis, W. Comparisonof liquid and gas-liquid chromatographic assays of 5 fluorouracil inplasma. Clin. Chem., 25: 83-86, 1979.

14. Dunnett, C. W., and Gent, M. Significance testing to establishequivalence between treatment, with special reference to data in form of2 X 2 tables. Biometrics, 33: 593-602, 1977.

15. Vokes, E. E., Weichselbaum, R. R., Lippman, S., and Hong, W.Head and neck cancer. N. Engl. J. Med., 328: 184-194, 1993.

16. Van Groeningen, C. J., Pinedo H. M., Heddes, J., Kok, R. M., deJong, A. P. J. M., Wand, E., Peters, 0. F., and Lankerma, J. Pharma-cokinetics of 5-fluorouracil assessed with a sensitive mass spectrometricmethod in patients on a dose escalation schedule. Cancer Res., 48:

6956-6961, 1988.

17. Yoshida, T., Araki, E., ligo, M., Fujii, T., Yoshino, M., Shimada,Y., Saito, D., Tajiri, H., Yamaguchi, H., Yoshida, S., Yoshino, M.,Onkura, H., Yoshimori, M., and Okazaki, N. Clinical significance ofmonitoring serum levels of 5-fluorouracil by continuous infusion inpatients with advanced colonic cancer. Cancer Chemother. Pharmacol.,26: 352-354, 1990.

18. Golberg, J. A., Kerr, D. J., Willmott, N., McKillop, J. H., andMcArdle, C. S. Pharmacokinetics and pharmacodynamics of locore-gional 5-fluorouracil. Br. J. Cancer, 57: 186-189, 1988.

19. Milano, 0., Etienne, M. C., Ren#{233}e,N., Thyss, A., Schneider, M.,Ramaioli, A., and Demard, F. Relationship between fluorouracil sys-temic exposure and tumor response and patient survival. J. Clin. Oncol.,12: 1291-1295, 1995.

20. Jacobs, C., Lyman, G., Velez-Garcia, E., Sridhar, K. S., Knight, W.,Hochster, H., Goodnough, L. T. L., Mortimer, J. E., Einhorn, L. H.,Schacter, L., Cherng, N., Dalton, T., Burroughs, J., and Rosencweig, M.A phase ifi randomized study comparing cisplatin and fluorouracil assingle agents and in combination for advanced squamous cell carcinomaof the head and neck. J. Clin. Oncol., 10: 257-263, 1992.

21. Etienne, M. C., Bernard, S., Fischel, J. L., Formento, P., Gioanni, J.,Santii, J., Demard, F., Schneider, M., and Milano, 0. Dose reductionwithout loss of efficacy of 5-fluorouracil and cisplatin combined withfolimc acid. in vitro study on human head and neck carcinoma cell lines.Br. J. Cancer, 63: 372-377, 1991.

22. Shirasaka, T., Shimamoto, Y., Ohshimo, H., Saito, H., and Fuku-shima, M. Metabolic basis on the synergistic antitumor activities of5-fluorouracil and cisplatin in rodent tumor models in vivo. CancerChemother. Pharmacol., 32: 167-172, 1993.

23. Calvert, A. H., Newell, D. R., Gumbrell, A., O’Reilly, S., Burnell,M., Boxall, F. E., Siddik, Z. H., Judson, I. R., Gore, M. E., andWiltshaw, E. Carboplatin dosage: prospective evaluation of a simpleformula based on renal function. J. Clin. Oncol., 7: 1748-1756, 1989.

24. Chatelut, E., Canal, P., Brunner, V., Chevreau, C., Pujol, A., Boneu,A., Roche, H., Houin, G., and Bugat, R. Prediction of carboplatinclearance from standard morphological and biological patient charac-teristics. J. Natl. Cancer Inst., 87: 573-580, 1995.

25. Harris, B. E., Carpenter, J. T., and Diasio, R. B. Severe 5-fluorou-racil toxicity secondary to dihydropyrimidine dehydrogenase defi-ciency. Cancer (Phila.), 68: 499-501, 1991.

26. fleming, R. A., Milano, G., Gaspard, M. H., Bargnoux, P. J., Thyss,A., Plagne, R., Ren#{233}e,N., Schneider, M., and Demard, F. Dihydropy-rimidine dehydrogenase activity in cancer patients. Eur. J. Cancer, 5:

740-744, 1993.27. Etienne, M. C., Lagrange, J. L., Dassonville, 0., Fleming, R. A.,Thyss, A., Ren#{233}e,N., Schneider, M., Demard, F., and Milano, G.Population study of dihydropyrimidine dehydrogenase in cancer pa-tients. J. Clin. Oncol., 12: 2248-2253, 1994.

28. Baker, S. D., Peang Khor, S., Adjei, A. A., Doucette, M., Spector,T., Donehower, R. C., Grochow, L. B., Sartorius, S. E., Noe, D. A.,Hohneker, J. A., and Rowinsky, E. K. Pharmacokinetic, oral bioavail-ability, and safety study of fluorouracil in patients treated with 776C85,an inactivator of dihydropyrimidine dehydrogenase. J. Clin. Oncol., 14:

3085-3096, 1996.29. Destache, C., Meyer, S., and Rowley, K. Does accepting pharma-cokinetics recommendations impact hospitalization ? A cost-benefitanalysis. The Drug Monit., 12: 427-433, 1990.



1998;4:2039-2045. Clin Cancer Res R Fety, F Rolland, M Barberi-Heyob, et al. patients with locally advanced head and neck carcinomas.of 5-fluorouracil: results from a multicentric randomized trial in Clinical impact of pharmacokinetically-guided dose adaptation

Updated version

http://clincancerres.aacrjournals.org/content/4/9/2039

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/4/9/2039To request permission to re-use all or part of this article, use this link



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

mailto:[email protected]



5-fluorouracil: results from a multicentric randomized ...changes in the wbc count, neutrophil...

Documents