molecules that could regulate tubulinassembly and function in cells. It isimportant to elucidate the conformationand structure of these binding sites tocomprehend the cellular control mecha-nisms for microtubules. In addition, suchinformation may prove essential forunderstanding the chemotherapeuticproperties of taxol.

References

(/) SCHIFF PB, HORWITZ SB: Taxol stabilizesmicrotubules in mouse fibroblast cells. ProcNatl Acad Sci USA 77:1561-1565, 1980

(2) MCGUIRE WP, ROWINSKY EK, ROSENSHEINNB, ET AL: Taxol: A unique antineoplasticagent with significant activity in advancedovarian epithelial neoplasms. Ann Intern Med111:273-279, 1989

(3) HOLMES FA, WALTERS RS, THERIAULT RL, ETAL: Phase II study of taxol, an active drug inthe treatment of metastatic breast cancer. JNatl Cancer Inst 83:1797-1805, 1991

(4) EINZIG Al, HOCHSTER H, WERNIK PH, ET AL:A phase II study of taxol in patients withmalignant melanoma. Invest New Drugs9:59-64, 1991

(5) ROWINSKY EK, BURKE PG, KARP JE, ET AL:Phase I and pharmacodynamic study of taxolin refractory acute leukemias. Cancer Res49:4640-4647, 1989

(6) SCHIFF PB, HORWITZ SB: Taxol assemblestubulin in the absence of exogenousguanosine 5'-triphosphate or microtuble-associated proteins. Biochemistry 20:3247-3252, 1981

(7) KUMAR N: Taxol-induced polymerization ofpurified tubulin. Mechanism of action. J BiolChem 256:10435-10441, 1981

(8) SCHIFF PB, FANT J, HORWITZ SB: Promotionof microtubule assembly in vitro by taxol.Nature 277:665-667, 1979

(9) PARNESS J, HORWITZ SB: Taxol binds to poly-merized tubulin in vitro. J Cell Biol 91:479-487, 1981

(10) MERRILL BM, WILLIAMS KR, CHASE JW, ETAL: Photochemical cross-linking of theEscherichia coli single-stranded DNA-bindingprotein to oligodeoxynucleotides. Identifica-tion of phenylalanine 60 as the site of cross-linking. J Biol Chem 259:10850-10856, 1984

(11) NATH JP, EAGLE GR, HIMES RH: Direct pho-toaffinity labeling of tubulin with guanosine5'-triphosphate. Biochemistry 24:1555-1560,1985

(12) WOLFF J, KNIPLING L, CAHNMANN HJ, ET AL:Direct photoaffinity labeling of tubulin withcolchicine. Proc Natl Acad Sci USA 88:2820-2824, 1991

(13) SHELANSKI ML, GASKIN F, CANTOR CR:Microtubule assembly in the absence ofadded nucleotides. Proc Natl Acad Sci USA70:765-768, 1973

(14) RINGEL I, HORWITZ SB: Taxol is converted to7-epitaxol, a biologically active isomer, incell culture medium. J Pharmacol Exp Ther242:692-698, 1987

(75) GASKIN F, CANTOR CR, SHELANSKI ML: Tur-bidimetric studies of the in vitro assembly anddisassembly of porcine neurotubules. J MolBiol 89:737-755, 1974

(16) LAEMMLI UK: Cleavage of structural proteinsduring the assembly of the head of bac-teriophage T4. Nature 227:680-685, 1970

(17) WANI MC, TAYLOR HL, WALL ME, ET AL:Plant anlitumor agents. VI. The isolation andstructure of taxol, a novel antileukemic andantitumor agent from Taxus brevifolia. J AmChem Soc 93:2325-2327, 1971

(18) ZAREMBA TG, LEBON TR, MILLAR DB, ET AL:Effects of ultraviolet light on the in vitroassembly of microtubules. Biochemistry23:1073-1080, 1984

(19) MANDELKOW EM, HERRMANN M, RUHL U:Tubulin domains probed by limited pro-teolysis and subunit-specific antibodies. JMol Biol 185:311-327, 1985

Pharmacokinetics ofPamidronate in Patients WithBone Metastases

S. Leyvraz,* U. Hess, G. Flesch,J. Bauer, S. Hauffe, J. M. Ford,P. Burckhardt

Background: Pamidronate is a second-generation bisphosphonate used in thetreatment of tumor-induced hyper-calcemia and in the management ofbone metastases from breast cancer,myeloma, or prostate cancer. Thepharmacokinetics of pamidronate isunknown in cancer patients. Purpose:To determine the influence of the rateof administration and of bone metabo-lism, we studied the pharmacokineticsof pamidronate at three different infu-sion rates in 37 patients with bonemetastases. Methods: Three groups of11-14 patients were given 60 mg pam-idronate as an intravenous infusionover a period of 1, 4, or 24 hours.Urine samples were collected in thethree groups of patients. Plasma sam-ples were obtained only in the 1-hourinfusion group. The assay of pamidro-nate in plasma and urine was per-formed by high-performance liquidchromatography with fluorescencedetection after the derivatization ofpamidronate with fluorescamine.Results: The body retention (BR) at0-24 hours of pamidronate represented60%-70% of the administered doseand was not significantly modified bythe infusion rate. In particular, theBR at 0-24 hours was not reduced atthe fastest infusion rate. Amongpatients, a threefold variability in BRat 0-24 hours occurred, which wasrelated directly to the number of bone

metastases and, to some extent, tocreatinine clearance. At 60 mg/hour,the plasma kinetics followed a multiex-ponential course characterized by ashort distribution phase. The mean (±SD) half-life of the distribution phasewas 0.8 hour (±0.3), the mean (±SD)of the area under the curve for drugconcentration in plasma x time at0-24 hours was 22.0 ± 8.8 (xmol/L xhours, and the mean (±SD) of themaximum plasma concentration was9.7 (xmol/L (±3.2) . Pharmacokinetievariables remained unchanged afterrepeated infusions applied to fourpatients. Clinically, the three infusionrates were equally well tolerated with-out significant toxicity. Conclusions:The 1-hour infusion rate could be pro-posed as kinetically appropriate forthe administration of pamidronate topatients with metastatic bone diseases.[J Natl Cancer Inst 84:788-792, 1992]

Bisphosphonates are structural ana-logues of pyrophosphate, a natural reg-ulator of bone mineral precipitation anddissolution. Pamidronate is a second-generation bisphosphonate that stronglyinhibits bone resorption without interfer-ing with bone mineralization (7,2). Itsactivity, measured by inhibition of boneresorption, is more potent when com-pared with the activity of etidronate andclodronate, probably because of a directaction on osteoclast precursors (1-4).

Pamidronate has been used widely inbenign clinical conditions such asPaget's disease (5) and osteoporosis (6).In malignancy, it has been administeredto treat tumor-induced hypercalcemia(7-9) and to reduce morbidity caused bybone metastases (10,11). The intra-venous route of administration is gener-ally preferred for the treatment ofmalignant hypercalcemia or bone meta-

Received September 11, 1991; revised January21, 1992; accepted January 30, 1992.

S. Leyvraz, J. Bauer (Centre Pluridisciplinaired'Oncologie), P. Burckhardt (Department of Inter-nal Medicine), University Hospital Lausanne,Switzerland.

U. Hess, Kantonsspital, St. Gallen, Switzerland.G. Flesch, S. Hauffe, J. M. Ford, Ciba Geigy

Limited, Basel, Switzerland.*Correspondence to: Serge Leyvraz, M.D.,

Centre Pluridisciplinaire d'Oncologie, UniversityHospital—Niveau 10, 1011 Lausanne, Switzerland.

788 Journal of the National Cancer Institute

stases, because of poor absorption by thegut (12) and gastric irritation caused byoral administration. The optimal sched-ule and infusion rate have not yet beendefined, however, and might be clarifiedby pharmacokinetic information.

The pharmacokinetic properties ofbisphosphonates have been determined inlaboratory animals {13,14) in which halfof the dose was accumulated in the skel-eton and half excreted unchanged in theurine. Accumulation of bisphosphonatesin the reticuloendothelial system of ani-mals varies between species (15) andbetween individual bisphosphonates(16), suggesting different distributionkinetics—factors that mandate cautiousinterpretation when extrapolating datafrom animals to humans.

Pamidronate pharmacokinetics werestudied in rats following oral and intra-venous administration of l4C-labeledcompound (17). Bone accumulation wasconstant at 25%-35% of dose. The up-take of labeled pamidronate by thereticuloendothelial system was, however,dose dependent, with less than 3% in theliver at 0.01 mg/kg and 30% in the liverat 10 mg/kg. Because the liver is animportant storage compartment with ashort half-life (ti/2) of retention, furtherincrease of bone accumulation occurredwith time. Release of pamidronate fromthe skeleton was extremely slow with anestimated tm of at least 300 days.

In humans, pharmacokinetic data onbisphosphonates are limited. In healthysubjects, means of 73% and 81% of theadministered dose of clodronate wererecovered unchanged in the urine within24-48 hours, with no difference betweenthe three doses tested (18,19). In sixpatients with metastatic breast cancer,total urinary excretion was similar at75% (20). However, in patients withPaget's disease, mean urine excretionwas lower at 58%. Because clodronatewas administered as a slow infusion overa 5-day period, it was concluded thatslow infusion increases accumulation inthe body and decreases urine excretion(21). It is not known by how much thepharmacokinetics of bisphosphonatescould be affected by the rate of admin-istration or by bone metabolism.

In initial clinical studies with pami-dronate, multidose regimens have beenused for a period of 6-9 days, but,recently, single-day infusions have beenshown to be equally effective and arecurrently recommended (7). We report

the pharmacokinetic study of pami-dronate given to cancer patients withbone metastases. Pamidronate wasadministered as a single infusion, at con-stant dose, but at three different infusionrates, to test the influence of infusionrate on pharmacokinetic variables.

Subjects and Methods

Study Subjects

The study was conducted at the CentrePluridisciplinaire d'Oncologie, Univer-sity Hospital, Lausanne, and at theKantonsspital, St. Gallen, Switzerland.It was approved by the ethical committeeof both institutions. Informed consentwas obtained from every patient. Entrycriteria included an age limit greater thanor equal to 18 years, a histological diag-nosis of cancer with radiological evi-dence of bone metastases, serum cre-atinine level less than 150 (xmol/L orbilirubin level less than 25 (Jimol/L, andnormal levels of liver enzymes. Exceptfor one patient with a serum calciumlevel at 2.91 mmol/L, the serum calciumlevel was not above 2.75 mmol/L. Thepatients were not treated previously withbisphosphonates. No new medicationsand no chemotherapy were started for atleast 72 hours prior to entry or at anytime during the study. Patients on hor-mone therapy were eligible, providedthey did not have a change in treatmentschedule for at least 2 weeks. A total of37 patients were enrolled in the study;their characteristics are summarized inTable 1.

Treatments

The study was performed in twophases over a 15-month period. Duringthe first phase, 23 patients were ran-domly assigned to receive 60 mg pam-idronate, diluted in 500 mL of 0.9%saline, as an intravenous infusion over aperiod of 4 or 24 hours. To ensure ade-quate urine flow, a total of 2 L of intra-venous fluid was given during the day ofpamidronate administration and the fol-lowing day. Blood samples were notdrawn during this phase of the study.When analysis disclosed that the amountof drug excreted in urine was similar atboth infusion rates, we started the secondphase of the study.

During the second phase, an additionalgroup of 14 patients was given 60 mgpamidronate, diluted in 250 mL of 0.9%saline, over a period of an hour withouthydration. Blood and urine samples werecollected in this group of patients.

Among this group of patients, fourwere selected in whom the 1-hour infu-sion was repeated at 4- to 5-week inter-vals at a maximum of four infusions.

Urine Collection and Blood Sampling

During the first part of the study, urinecollections were performed at 4-hourintervals for 12 hours (three collections)and then at 12-hour intervals for a further36 hours (three collections). During thesecond part of the study, collection ofblank samples and two collections ofurine were performed from 0 to 4 hoursand from 4 to 24 hours.

Table 1. Patient characteristics

No. of patientsMale/femaleMean age, y (range)Mean height, cm (range)Mean weight, kg (range)Median creatinine clearance, mL/min (range)No. of breast carcinomasNo. of prostatic carcinomasNo. of bronchial carcinomasNo. of hypernephromasNo. of bone metastases*

15

Pamidronate infusion60mg/4 h

122/10

67 (56-80)162 (135-175)

62 (34-85)65 (37-98)

1011

—

372

60 mg/24 h

110/11

63 (28-78)157(148-165)

60(52-71)66 (48-102)

11———

272

rates60 mg/ 1 h

143/11

62 (42-73)162(139-178)

65 (52-75)68 (35-110)

112

—1

194

•Evaluated on standard x rays.

Vol. 84, No. 10, May 20, 1992 REPORTS 789

Blood samples were obtained onlyduring the second part of the study,before and then every 15 minutes for 2hours after the start of the infusion andthen at 2.5, 3, 5, and 24 hours.

Assay of Pamidronate in Urine andPlasma

The assay of pamidronate in plasmaand urine was performed by high-performance liquid chromatography withfluorescence detection, after derivatiza-tion of pamidronate with fluorescamine(22,23). The limits of quantitation were1.4 (jimol/L in plasma and 1.8 jjimol/L inurine.

Clinical Assessment

In the first part of the study, serumcalcium, phosphate, creatinine, andcreatinine clearance levels were assessedbefore treatment.

Because of concerns about the pos-sibility of increased toxicity of the morerapid infusion rate used during the sec-ond part of the study, the followingparameters were measured in serumbefore infusion and at 24 and 48 hoursafter infusion: urea, creatinine, calcium,phosphate, albumin, bilirubin, alkalinephosphatase, and transaminases. In addi-tion, the following parameters weremeasured in urine: calcium, phosphate,creatinine, sodium, protein, hydroxy-proline, and lysozyme. Routine urinal-yses were done for the presence of bloodand protein.

In all patients, the number of bonemetastases was evaluated on standardx rays and reported in three groups—those with fewer than five metastases,those with between five and 15 meta-

stases, and those with more than 15 met-astases. In seven patients, a moreaccurate determination of the number ofbone metastases was obtained by bonescan. Clinical examinations for toxicitywere performed daily during the infusionperiod and then 2 to 4 weeks later.

Pharmacokinetic Calculations

Noncompartmental techniques wereused in the pharmacokinetic analysis(24). The total urinary excretion (TUE)at 0-24 hours was the total amount ofpamidronate excreted in urine over 24hours. The body retention (BR) at 0-24hours was estimated as follows: BR (%of dose) = dose (100%) - TUE (% ofdose); fecal excretion was, in all proba-bility, negligible as known from animalexperiments (17).

The area under the curve (AUC) fordrug concentration in plasma X time at0-24 hours, expressed as (u,mol/L) xhours, was determined by the linear tra-pezoidal rule from 0 to 24 hours. Cmax((i-mol/L) was the maximal plasma con-centration, and Tmax (hour) was the timeat which Cmax was reached. The apparentplasma distribution half-life (tl/2 inhours) of pamidronate was calculatedfrom the slope of the linear least-squaresregression line through the concentra-tion-time points in the time interval from1 to about 2 hours. The total plasmaclearance (CL, [L/hour]) was calculatedas the dose/AUC (0 - ), taking AUC(0-24 hours) as AUC (0 - oo) because all24-hour plasma concentrations werebelow the limit of detection and taken aszero. The renal clearance (CLr [L/hour])was calculated as the TUE (0-24 hour)/AUC (0-24 hour). The nonrenal clear-ance (CLnr [L/hour]) was CL, - CLr.

No statistical comparison of the treat-ments was performed. This study beingthe first pharmacokinetic trial with pami-dronate disodium, no information on theintersubject variability was available.The results of this trial showed a highinterpatient variability in the urinaryexcretion, and, therefore, the number ofpatients included in each group of thestudy was too low to perform meaningfulstatistical tests with sufficient power.

Results

Pharmacokinetics in Urine

The TUE over 24 hours of eachpatient and the mean (±SD) values areshown graphically in Fig. 1, according tothe three different infusion rates. Themean TUE (0-24 hours) (±SD) was notsignificantly different and was measuredat 31% ± 15.2%, 34.9% ± 13.9%, and41.0% ± 15.4% of the 60 mg of pami-dronate administered over periods of 1hour, 4 hours, and 24 hours, respec-tively. Considerable between-patientvariability was noted.

To explain the wide disparity of TUE(0-24 hours) among patients, TUE wascorrelated with creatinine clearance andwith tumor bone involvement. A weakrelationship was found between cre-atinine clearance and TUE (0-24 hours)by linear least-squares regression anal-ysis (r = 0.42), but a stronger associa-tion was demonstrated between thenumber of bone metastases and BR (0-24hours). The mean (±SD) of the BR(0-24 hours) values in patients withfewer than five bone metastases was50.6% ± 11.8% and increased withincreasing number of bone metastases to76.4% ± 12.0% in patients with more

20 30

Time [h]

40

70

60

50

|60 mg in 4 h

20 30

Time [h]

20 30

Time [h]

Fig. 1. Total unnary excretion [TUE (0-24 hours)] of each patient after an infusion of 60 mg of pamidronate disodium given over 24-, 4-, and 1 -hour periods at aconstant infusion rate. Mean values are presented as dash lines.

790 Journal of the National Cancer Institute

than 15 metastases (Fig. 2). These find-ings were confirmed in the subgroup ofseven patients, where an accurate countof bone involvement could be made bybone scintigraphy and where the coeffi-cient of correlation was 0.82 (linearleast-squares regression). No correlationwas found between serum and urineparameters such as calcium, phosphate,alkaline phosphatase, hydroxyproline,and TUE (0-24 hours).

Pharmacokinetics in Plasma

After the 1-hour infusion of 60 mg ofpamidronate disodium, the mean (±SD)of the AUC for drug concentration inplasma x time at 0-24 hours was 22.0± 8 . 8 (itnol/L X hours; the mean Cmax(±SD) was 9.7 (xmol/L ± 3.2 in 14patients. The median value of Tmax was0.8 hour. The mean (±SD) plasma con-centration times profile is shown in Fig.3. The decline of pamidronate concentra-tions in plasma followed a multiexponen-tial course. The distribution of pami-dronate was rapid with a mean (±SD)apparent tV2 of 0.8 hour (±0.3). Plasmaconcentrations of pamidronate were nearthe limit of quantitation in most patientsat 3 hours and below the limit at 5 and24 hours. Therefore, no terminal tmcould be estimated for the slow releaseof pamidronate from bone.

Two hours after the start of the infu-sion, a slight increase in plasma con-centration occurred in the majority ofpatients, possibly corresponding to arelease of pamidronate from an unknowncompartment. The mean (±SD) CL, was10.7 L/hour (±3 .7 ) , and the mean

100-

8 0

6 0 -

4 0 -

2 0 -

o-

BR% of

N-

r11•

BR (0-24 hours) was 50% in patientswith few bone metastases and 75% inthose with numerous bone metastases.The relationship was stronger in a subsetof patients in whom a more accuratedetermination of the number of metasta-tic sites was performed by countinglesions on bone scan.

The influence of renal function on uri-nary excretion of pamidronate is un-known. In the population studied whichhad either a normal or a slightly reducedcreatinine clearance, the renal functionhad a weak effect on TUE (0-24 hours).However, in a study in which patientswith osteodystrophy had a creatinineclearance of only 5-35 mL/minute, TUE(0-24 hours) of Tc-99m-etidronate was11.4% of the administered dose (25).The disparity in skeletal invasion and inrenal function could explain the almostthreefold variability in BR (0-24 hours)among patients in our study.

The plasma pharmacokinetics of pami-dronate at 60 mg/hour followed a multi-exponential course and was characterizedby a short distribution phase in plasma,followed by a rapid elimination of un-changed drug in urine. The distributionof pamidronate was rapid with a tm of0.8 hour. Clodronate was also removedpromptly from the blood, with a tmbetween 1.8 hours and 2.3 hours inhealthy subjects (16) and in patients withbone disease (18,20). The total clearanceof pamidronate was 10.7 L/hour with aCLnr clearance of 7.5 L/hour. For clodro-nate, values varied between studies, witha total clearance of 6.4 and 11.5 L/hour,possibly due to different methods ofmeasurement, a factor which preventsany direct comparison with our data.Plasma concentration increased slightly 2hours after the start of the infusion. Pam-idronate may have been released fromthe reticuloendothelial system, where itcould have accumulated temporarily. Asimilar observation has been made inanimals but not previously in humans. Inmice, pamidronate accumulated in spleenand liver in contrast to clodronate andetidronate.

When pamidronate was infused at adose of 60 mg over a 1-hour period at 4-to 5-week intervals, its pharmacokineticvalues remained unchanged. With thisschedule and dose, the bone compart-ment was not saturated. Similar conclu-sions were reached after five dailyinfusions of clodronate, where no signifi-

cant variations were detectable betweendaily infusions (20).

Clinically, the tolerance of patients forall three infusion rates was similar, withonly minimal toxicity. In particular, norenal toxicity was observed after the1-hour infusion rate. Previous studieshave shown a transient increase inplasma creatinine (26) and rare cases ofrenal failure (27) after etidronate andclodronate administration. The plasmacreatinine did not rise within 48 hoursafter the infusion of pamidronate. Pro-teinuria remained within normal range.The minimal elevation of aspartateaminotransferase seen in one patient hadno clinical significance and could berelated to a possible uptake of pami-dronate into the reticuloendothelial sys-tem. A larger number of patients,however, should be studied for toxicityat the fastest rate of infusion beforesafety can be ensured.

On the basis of the pharmacokineticdata combined with the clinical findings,we conclude that BR (0-24 hours) ofpamidronate was dependent on theamount of bone metastases and was notinfluenced by the rate of infusion. Thus,the 1-hour infusion rate can be proposedas a kinetically relevant treatment sched-ule for patients with malignant bonediseases.

References

(/) FLEISCH H: Bisphosphonates—history andexperimental basis. Bone 8:23-28, 1987

(2) SHINODA H, ADAMEK G, FELIX R, ET AL:Structure-activity relationships of variousbisphosphonates. Calcif Tissue Int 35:87-99,1983

(3) BOONEKAMP PM, VAN DER WEE P A L S L, VANWIJK-VAN LENNEP MML, ET AL: Two modesof action of bisphosphonates on osteolyticresorption of mineralized bone matrix. BoneMin 1:27-39, 1986

(4) FAST DK, FELIX R, DOWSE C, ET AL: Theeffects of diphosphonates on the growth andglycolysis of connective-tissue cells in cul-ture. Biochem J 172:97-107, 1978

(5) THIEBAUD D, JAEGER P, GOBELET C, ET AL: Asingle infusion of bisphosphonate AHPrBP(APD) as treatment of Paget's disease ofbone. Am J Med 85:207-212, 1988

(