Portal Levels and Hepatic Clearance of 5-Fluorouracil ... · of 5-FUra metabolism since, in some studies, plasma clearance of 5-FUra exceeds the hepatic blood flow (13, 20, 25, 32)

[CANCER RESEARCH 41, 1916-1922. May 1981]

Portal Levels and Hepatic Clearance of 5-Fluorouracil after

Intraperitoneal Administration in Humans

James L. Speyer,1 Paul H. Sugarbaker, Jerry M. Collins, Robert L. Dedrick, Raymond W. Klecker, Jr., and

Charles E. Myers

Biochemical Pharmacology Section, Clinical Pharmacology Branch [J. L. S., R. W. K.. C. E. M.\ and Surgery Branch Â¡P.H. S.], Clinical Oncology Program. Divisionof Cancer Treatment, National Cancer Institute, and Division of Research Services [J. M. C., R. L. D.j. NIH, Bethesda. Maryland 20205

ABSTRACT

Intrahepatic tumor is a major problem in clinical oncology.While direct intravascular infusions provide high local drugconcentrations and variable rates of tumor response, they arelimited by technical considerations and complications. In thisstudy, we have tested whether high portal venous and hepaticarterial concentrations of 5-fluorouracil (5-FUra) can be

achieved by administering drug via peritoneal dialysis. Fourpatients with metastatic colon carcinoma had a Tenckhoffcatheter surgically implanted. During dialysis therapy with 4mW 5-FUra, simultaneous samples of peritoneal fluid and of

portal venous, hepatic venous, and peripheral venous, andarterial blood were obtained, and 5-FUra concentrations were

determined.Mean peak portal vein drug concentrations were 60 Â¡Â¿Mand

exceeded the measured concentrations in the other vessels.Total drug exposures as measured by concentration x time(mW x min) during Exchange 1 were: portal, 3.8 Â± 0.65;hepatic vein, 0.97 Â±0.44; peripheral vein, 0.90 Â±0.32; andarterial, 1.1 Â±0.26. During Exchange 7, total drug exposureswere: portal, 6.3 Â± 1.4; hepatic vein, 2.5 Â± 1.3; peripheralvein, 2.3 Â±1.1; and arterial, 2.7 Â± .85. The fraction of i.p.drug that exited the peritoneal cavity through the portal venoussystem ranged from 0.29 to 1.0. This variation resulted in partfrom uncertainty in estimating portal blood flow and gastrointestinal drug elimination. Calculated hepatic extraction was67% (range, 0.23 to 0.89). Extrahepatic metabolism was demonstrated. Measured 5-FUra concentrations compared favorably to values predicted by a pharmacokinetic model for 5-

FUra.Dialysis therapy (i.p.) with 5-FUra provides a means of

achieving high drug concentrations for treating both i.p. andintrahepatic tumor. Further clinical testing of this route ofadministration is warranted.

INTRODUCTION

While primary hepatic cancers are not common, tumorspread to the liver is frequent and constitutes a major problemin clinical oncology. Intrahepatic tumor is often the primarycause of symptoms and the limiting factor in survival (5, 22,38, 43). In autopsy series of patients with gastrointestinalcancer, approximately 50% of patients dying of their cancershave been found to have tumor metastatic to the liver (5, 8).Colorectal cancer recurs most frequently in local sites andsecondarily in the liver (7, 14, 18) as a result of metastatic

spread through the portal venous drainage of the primary site.Other tumors, such as breast cancer and small-cell lung can

cer, more probably spread to the liver via the systemic (hepaticarterial) circulation. A variety of approaches have been takento treat intrahepatic tumor. These include surgery (15), radiation therapy (31, 41), systemic chemotherapy (26), and locallydirected chemotherapy administered via hepatic arterial (2,13,27, 29, 39) or portal venous (1, 3, 6, 36, 40) infusion. Althoughobjective responses have been seen in response to both hepatic arterial and portal venous infusions, there are serioustechnical problems with and complications from prolongedinfusions via intravascular catheters (4, 16).

We have been conducting trials with i.p. instillation of 5-FUra2 for treatment of localized intraabdominal cancers, pri

marily ovarian and colon carcinoma (23, 32, 33). In thesestudies, drug was delivered via peritoneal dialysis throughsemipermanent indwelling catheters which provide easy andrepeated access for the delivery of prolonged instillations ofdrug in large volumes of fluid. This form of therapy results inthe delivery of high concentrations of drug to the i.p. spacewhile systemic exposure to drug is minimized. For 5-FUra,there is a greater than 2-log difference between peritoneal fluid

and plasma levels. On the basis of these findings, we considered the possibility that i.p. therapy may also be a simpler routethan arterial or portal vein infusion for prolonged delivery ofdrug directly to the liver.

Although many substances placed into the peritoneal spaceare removed via the portal circulation, the routes of absorptionof i.p. antineoplastic agents have not been defined in humans.Furthermore, it is unclear that the liver constitutes the sole siteof 5-FUra metabolism since, in some studies, plasma clearanceof 5-FUra exceeds the hepatic blood flow (13, 20, 25, 32).

In the following studies, we simultaneously measured peripheral, portal, and hepatic venous as well as systemic arterial 5-FUra concentrations during i.p. 5-FUra dialysis therapy. These

data were necessary to assess the following, (a) Is Â¡.p.administration of 5-FUra a feasible alternative route for treating

intrahepatic mÃ©tastasesor for administering adjuvant therapyto Gl cancers; i.e., what concentrations of drug are achievedin the portal venous and hepatic arterial circulation? (b) Howmuch of the i.p. drug is absorbed through the portal circulationduring 5-FUra dialysis? (c) What is the quantitative role of theliver in the metabolism of 5-FUra when administered by this

route? Furthermore, these data provide a unique opportunityto test our pharmacokinetic model published previously (10).

1To whom requests for reprints should be addressed, at National Cancer

Institute, NIH, Building 10, Room 6N102, Bethesda, Md. 20205.Received May 12, 1980; accepted February 3, 1981

2 The abbreviations used are: 5-FUra, 5-fluorouracil; Gl, gastrointestinal; 5-FdUrd, 5-fluoro-2'-deoxyuridine; PA, permeability area product; V,P, volume of

peritoneal fluid; AUC, area under the curve of concentration versus time; PeA,peripheral artery; fpo, fraction of drug absorbed through the portal system; PoV,portal vein; HV, hepatic vein; EH, hepatic extraction; Q, blood flow; PeV, peripheral vein.

1916 CANCER RESEARCH VOL. 41

Research. on November 26, 2020. © 1981 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

Portal Uptake of Lp. 5-FUra

METHODS

Patients. Four patients with a confirmed histolÃ³gica!diagnosis of adenocarcinoma of the colon were treated (Table 1).One patient (Patient 3) had received prior chemotherapy withi.V. 5-FUra.

Surgery and Catheter Placement. Informed consent wasobtained from all patients prior to surgery. A No. 5 Frenchpolyethylene catheter was inserted into the right antecubitalvein and advanced over a guidewire into a right main hepaticvein. During the procedure, patients were monitored by electrocardiogram for arrhythmias. The position of the catheter wasconfirmed fluoroscopically by hand injection of Conray 60(Mallinckrodt, Inc., St. Louis, Mo.) contrast material and documented by a spot film. All 4 patients required an operativeprocedure for placement of a Tenckhoff catheter. Patient 2also underwent a left colectomy with end-to-end anastomosis.

At the time of laparoscopy and Tenckhoff catheter placement,a 16-gauge Intracath (Deseret Co., Sandy, Utah) radiopaque

catheter was placed into the left branch of the portal vein byrecanalization of the obliterated umbilical vein. The ligamentumteres was located approximately 2 cm from the umbilicusthrough the abdominal incision. Intraportal position of the catheter was confirmed by hand injection of Conray 60 and documented by a spot film. A Tenckhoff semipermanent dialysiscatheter with a single Dacron cuff was placed through theabdominal wall via a s.c. tunnel. A peripheral vein catheter inthe left antecubital fossa and a radial artery catheter were alsoestablished. The patency of the lines was maintained with asolution of 1000 units of sodium heparin in 1 liter of 0.9% NaCIsolution. After dialysis treatment and plasma sampling, thepositions of the portal and hepatic vein catheters were reconfirmed radiographically. All catheters, except the Tenckhoffcatheter, were removed after the initial course of dialysis treatment (within 48 hr after surgery). The Tenckhoff catheterremained in place for subsequent courses of treatment, andthe patients were instructed in the daily dressing and care oftheir own catheters (32).

5-FUra Dialysis Therapy. All patients were treated with 8consecutive exchanges of dialysate containing 5-FUra, starting

as soon as their condition following general anesthesia wasstable. Each exchange consisted of a 4-hr dwell time and then

drainage. The dialysate solution consisted of 2 liters of 1.5%Inpersol (Abbot Laboratories, North Chicago, III.) containing 8mEq of KCI, 50 mEq of NaHCO3, and 1040 mg (4 mvi) of 5-

FUra (Roche Laboratories, Nutley, N. J.). Patients were treated

every 2 weeks unless toxic symptoms or disease progressioncaused an alteration of schedule or a change in therapy.

Drug Level Measurement. During the first and seventh exchanges of the initial course of 5-FUra dialysis, timed samples

of arterial, peripheral venous, hepatic venous, and portal venous blood, as well as of peritoneal fluid, were obtained. 5-

FUra and FdUrd concentrations were measured in plasma andperitoneal fluid as reported elsewhere (32). The assay consistsof a 2-stage sample clean-up procedure with anion-exchange

chromatography and an organic extraction with ethyl acetate.Samples were then measured in a reverse-phase high-pressureliquid chromatography system. An internal standard of [3H]-5-

FUra was used to normalize differences in extraction, and thepeak height:dpm ratio was used to calculate the concentrationof the drug. A set of known standards was run with each setof patient samples. Replicate samples are reproducible towithin 5%.

CALCULATIONS

For purposes of calculation, it is assumed that (a) there isequilibration of drug between blood cells and plasma and (fa)that the concentration of drug in the peripheral artery is equalto that in the hepatic artery since significant metabolism of 5-

FUra by blood has not been reported.Peritoneal clearance, expressed as the product of permea

bility and area (PA), was obtained from the slope of the logarithm of peritoneal fluid concentration (C) versus time (f):

In C,p (f) = In C,P(0) - PAW(A)

where C/P is the concentration of drug in the peritoneal fluidand V/p is the volume of the fluid (2 liters). Apparent clearanceis calculated as (dose absorbed)/AUCPexi. This definition differsfrom the usual one for total body clearance because it incorporates any "first-pass" elimination and expresses an effective

value which would be rigorously obtained only for a linearsystem. AUCPe/, is an expression of C x t for the peripheralartery. For each of the sampled fluids, AUC was determined bythe trapezoidal rule from zero to the last plasma point and thenby first-order extrapolation to infinite time. A 10-min half-timewas used for this extrapolation, since we expected 5-FUra

disappearance similar to that observed following an i.v. bolus,once peritoneal drug input ceased. This extrapolation usuallyrepresented 1 to 2% of the total AUC and was always less than5%. No correction was made for residual drug present before

Table 1

Disease status and surgery performed on study patients

PatientAge1

572

563

464

21SexMFMMPrior

treatmentNoneNonei.v.

5-FUraNonePrior

surgerySigmoid

colonresectionSleeve

resection of sigmoid colon with sigmoidcolostomySigmoid

colonresectionAbdominal

perineal resection, endsigmoidcolostomyDisease

at time ofsurgeryRetroperitoneal

andhepaticmÃ©tasta

sesPelvicmÃ©tastasesHepatic

mÃ©tastasesHepatic

mÃ©tastasesSurgery

performedLaparoscopy,

liver biopsy, umbilical veincatheterization,Tenckhoff

catheterinsertionLaparoscopy,liver biopsy, um

bilical veincatheterization.Tenckhoffcatheterinsertion.left

hemicolectomyLaparoscopy,liver biopsy, um

bilical veincathetorization.TenckhoffcatheterinsertionLaparoscopy,

liver biopsy, umbilical veincatheterization,Tenckhoff

catheter insertion

May 1981 1917



J. L. Speyer et al.

Exchange 7 since, in all vessels measured, 5-FUra concentrations were 1 log greater at the end of the dwell phase thanwere those observed immediately before the next instillation.

As seen in the flow diagram (Chart 1), PoV concentration isdependent upon the fraction of the absorbed dose which entersthe body through the portal system (fPo):

c -Â»-'POVfpÂ°PA

(B)

Integration of Equation B with respect to time and rearrangement gives:

(AUCPoV - AUCG)) QPO,

Dose absorbed(C)

This calculation of fPo is not model dependent but is definedonly by portal anatomy.

As seen in the flow diagram (Chart 1), drug enters the liverfrom both the hepatic artery (HA) and PoV and leaves via theHV only. EHis calculated as:

(Drug in) - (drug out) QPoV CPoV + QHA CHA - QHV CH

(Drug in)(D)

Since 5-FUra elimination processes are saturable, EH depends upon hepatic 5-FUra concentration, which we assume

to be approximated by CHV.EH is time dependent in our protocol, since CHVvaries with time. Two measures of EH can beevaluated: instantaneous EH (Equation D); and overall liverextraction, obtained by replacing the concentration terms withcorresponding AUC terms:

QPoVAUCPoV + QHA AUCHA - QHVAUCHV

QPoVAUCPOV + OHA AUCHA(E)

Overall extraction is a measure of the fraction of the total 5-

FUra dose delivered to the liver, including recirculated drug,which is eliminated by the liver. As long as CHV<K Km, instantaneous extraction remains constant and is identical to theoverall extraction. This condition is not strictly met by any dataset in this study. Therefore, the overall extraction values reported here must be interpreted as nominal values for thewhole exchange period during which instantaneous extractiondynamically varies.

RESULTS

Chart 2 demonstrates the measured concentrations of 5-

FUra in Patient 1 during Exchanges 1 and 7. The drug concentrations in the peritoneal fluid and peripheral vein are similarto those reported previously (32). Over the 4-hr dwell time,

concentration x time was 318 times greater in the fluid than inthe peripheral venous plasma. Portal venous levels exceedperipheral venous, hepatic venous, and arterial levels at alltime points and reached a maximum of 22 to 120 JUMwithin 30min. There is a steeper slope to the decline of peritoneal fluidlevels and plasma levels in Exchange 7 than in Exchange 1.AUC data are presented for all 4 patients in Table 2 along withcalculated apparent body clearance and PA.

The AUCpov is greater than the AUC of any of the other 3vessels measured. At Exchange 1, it is 4.2 times AUCPey, 3.6times AUCPe/,, and 4.0 times AUCHv- In Exchange 7, the ratiois lower, although the absolute differences measured are even

greater: AUCPoVis 2.8 times AUCPei,, 2.5 times AUCPeâ€ž,and2.6 times AUCW.

There is a sizable difference between portal and hepatic veinlevels. While portal and hepatic venous and arterial concentrations were measured, the mass of drug entering the liver is alsodependent upon the flow in each of these vessels which wasestimated. Depending on the assumed ratio of portal venousto hepatic arterial blood flow, mean EH (Table 3) ranged from0.71 to 0.74 in Exchange 1 and 0.61 to 0.65 in Exchange 7.

In 3 of 4 patients, drug was absorbed from the peritonealcavity more rapidly in the seventh than in the first exchange.This is illustrated by the steeper slope of fluid disappearancein Chart 2 and resulted in increased values for PA. These

- PeV PcA

Chart 1. Flow diagram for pharmacokinetic modeling of 5-FUra (5FU) andcalculation of EHand fPo.C, concentration.

10.000

1000 -

100

10

0.160 120 180 240 60 120 180 240

TIME (Minutes)

Chart 2. Experimental data for Patient 1, Exchanges 1 (left) and 7 (right). â€¢Cip; A, Cpov!^, CpeAÂ¡O, Ce,v; A, CHV.Curves do not represent model simulations.




Portal Uptake of i.p, 5-FUra

Table 2Area under concentration-time curve (AUC) and other pharmacokinetic parameters for exchanges 1 and 7

AUC units are mM X min.

Patient1Exchange1Exchange

7Patient

2Exchange1Exchange

7Patient

3Exchange1Exchange

7Patient

4Exchange1Exchange

7AUCâ€ž190130340280260370530320AUCp.,0.661.31.85.10.741.71.12.5AUCp.v0.520.871.85.70.401.00.861.6AUCpov2.83.55.710.03.56.73.25.1AUCHV0.380.492.26.50.311.40.971.8CLa

(liters/

min)116.04.21.5104.35.93.1PA(ml/

min)2329233323201426

Exchange! 330 Â±73 1.1 Â±0.26 0.90 Â±0.32

Exchange 7b 275 Â±52 2.7 Â±0.85 2.3 Â±1.1

3.8 Â±0.65 0.97 Â±0.44

6.3 Â±1.4 2.5 Â±1.3

7.8 Â±1.6 21 Â±2.3

3.7 Â±0.95 27 Â±2.7a CL, apparent body clearance.b Mean Â±S.E.

Table 3Dependence of EHaon ratio of portal venous to hepatic arterial blood flow after

i.p. 5-FUra

u, . OÃ.,-,at following ratios

2:1 3:1 4:1

Patient 1Exchange 1Exchange7Patient

2Exchange 1Exchange7Patient

3Exchange 1Exchange7Patient

4Exchange 1Exchange7Exchange

1CExchange

7Ca

F drug in0.82

0.820.50

0.230.88

0.720.62

0.670.71

Â±0.0880.61

Â±0.13â€”drug out0.83

0.830.53

0.270.89

0.740.64

0.690.72

Â±0.0830.63

Â±0.120.84

0.840.55

0.290.89

0.750.66

0.700.74

Â±0.0790.65

Â±0.12

drug inb Assumed ratios of portal vein:hepatic arterial blood flow.c Mean Â±S.E.

values for PA are higher than predicted from molecular weightand exceed the values reported previously in patients whowere not immediately postoperative. In all patients, peakplasma levels were higher during the later exchange. A furtherdifference is the decreased apparent clearance noted in thelater exchange for all patients.

A final observation is the small but consistent differencebetween peripheral arterial and peripheral venous drug concentration that was seen in 3 of 4 patients. It can be expressedas a ratio of AUCPel,/AUCpe/, and had a mean value of 0.78.

DISCUSSION

Therapy (i.p.) with 5-FUra results in portal vein drug levels

that may make this route a feasible one for perfusing the liverwith cytotoxic concentrations of drug. Peak portal venous

levels of greater than 10 6 M are 4 to 10 times peripheral

venous levels. These levels exceed those which can be safelymaintained by peripheral infusion (20). The dose of drug delivered to the liver by i.p. 5-FUra treatment and the drug concen

trations achieved compare favorably with levels achievable byhepatic arterial infusions. If a 70-kg human with a hepatic

artery flow of 300 ml/min is given an arterial infusion of 10mg/kg/24 hr (a typical dose), the mean arterial concentrationwould be 12 /Â¿Mwhile the portal vein concentration would beless than 1 /Â¿M(assuming total body clearance of 5 liters/min,regardless of first-pass effects). In this study, the mean 5-FUra

portal vein concentration (&UCPoV/4 hr) achieved by the i.p.route was 21 Â¡IM,and the mean arterial concentration was 8/Â¿M.Total drug delivered i.p. was 90% of 1040 mg/4 hr orabout 75 mg/kg/24 hr, of which 29 to 100% was delivereddirectly to the liver (see below).

In addition to the delivery of comparable amounts of drug,the i.p. route offers some other attractive advantages overarterial infusions, (a) Drug instillation is simple without the needfor an indwelling arterial catheter and infusion pump with attendant risks of thrombosis, infection, and catheter slippage(4, 16). (b) Therapy (i.p.) more likely provides uniform distribution to the hepatic parenchyma since drug enters the portalsystem distally; direct intravascular infusions are often associated with streaming and nonuniform distribution of drug (17);moreover, arterial infusions are usually to either the right or leftlobe only, since it is quite difficult to keep a catheter placed inthe common hepatic artery, (c) In an adjuvant setting, the i.p.route provides the highest plasma levels in the portal system,which is the putative route by which intraabdominal tumorseeds the liver, (d) Many patients with intrahepatic disease mayalso have intraabdominal disease, which might be more likelyto benefit from i.p. than from local hepatic arterial infusions. Itis not our intent to state that portal infusion via i.p. therapy issuperior to hepatic arterial infusions. There are certainly existent data which support therapy of intrahepatic tumor by bothroutes (1 -3, 6, 13, 27, 29, 34, 39, 40). The relative advantages

of these 2 routes of administration or their combination shouldbe the subject of further studies which address both pharmacokinetic factors and tumor blood supply.

May 1981 1919



J. L. Speyer et al.

An increase in PA and plasma levels and a decrease inapparent body clearance between Exchanges 1 and 7 werenoted in 3 of 4 patients. It is possible that prolonged exposureof the peritoneal membrane to drugs might alter the normalcharacteristics of the membrane and permit increased drugegress from the abdomen, as evidenced by the increase in PA.The decreased clearance is consistent with a recent report byKirkwood ef al. (24) indicating increased plasma half-life of 5-

FUra during the course of 5 daily infusions. No concentrationchange between exchanges was noted in our Phase I study(32), but only 4-hr levels were generally determined, and those

patients were not immediately postoperative. Even in the present study, 4-hr levels are not statistically different, but a full

AUC analysis demonstrates the difference.Two of 4 patients had a decrease in WBC during therapy. In

Patient 2, WBC fell to 500 and was accompanied by Esche-

richia coli sepsis, without bacterial peritonitis or seeding of thecatheter, which developed 11 days posttreatment. In Patient 4,WBC fell to 2,700. All platelet counts remained above 150,000,and there were no changes in hemoglobin. Patient 2 also hadthe highest plasma levels of drug. She was the only patient whohad extensive surgical violation of peritoneal integrity in thecourse of a left colectomy. This marrow suppression was incontrast to our earlier finding of the safety of this dose (32).The earlier sample may not have been large enough to demonstrate this effect. Alternatively, the immediate postoperativestate may have resulted in higher plasma levels through fasterabsorption from the abdominal cavity in the absence of anintact peritoneal lining or decreased metabolism and clearancesecondary to anesthesia or other perioperative events. Whileit is attractive to speculate about these possibilities, there arenot enough patients here to properly address this question.

The fraction of drug removed by the portal system (Table 4)was calculated using the term AUCQ/, where AUCG/ representsthat amount of recirculating drug which is not cleared by thetissues of the Gl tract before the blood enters that portalsystem. Dose and AUCPoâ€žare determined experimentally. Metabolism by the Gl tract can range from 0 (AUCG) = AUCPe/,), inwhich case the AUCPoV is a function of both i.p. drug and

Table 4Fraction of i.p. 5-FUra absorbed through portal system (fpja

Qpov (1liter/min)AUCa

=AUCp./Patient

1Exchange1Exchange

7Patient2Exchange1Exchange



70.290.290.520.730.360.680.300.31AUCG,

=00.370.450.68>10.460.910.460.59QPovb

(1.5liter/min)AUCOI

=AUCPeAc0.430.430.79>10.541.00.440.47AUCa,=00.560.67>1>10.70>10.680.89'

f is calculated using:

Dose

6 PoV flow obtained from published values (9, 11, 12, 21, 28, 30, 35-37,

42)Â¿c Gl elimination of recirculating arterial drug is assumed and is included in

calculations over a range of 0 to 100%.

recirculating drug, to 100% removal of drug from recirculatingblood (AUCG, = 0), in which case AUCPoVis exclusively fromi.p. drug. We are unable to further define this term, although Glmucosa is a probable site for extrahepatic fluoropyrimidinemetabolism. In dogs, Gustavsson era/. (19) have reported that14.6% of a peripheral infused dose of 5-FUra is removed in

the prehepatic splanchnic bed.Values for portal flow were obtained from the literature

because of the technical difficulty of obtaining this informationin humans. The range of portal flow used here is 1000 to 1500ml/min. Portal flow has been measured with electromagneticflow probes placed around (28) or in (35) the vessel, or byindicator [Lipiodol (30), 51Cr-RBC (21 ), 125l-Risa (9), 133Xe(36),

indocyanine green (12)] washout studies, all of which arecumbersome. Portal flow is not directly or uniformly related tototal hepatic blood flow (37) which is more easily measured.Furthermore, these studies indicate variable flow in the samepatient at different times and under different conditions (11,42). Therefore, to determine flow directly, it would have beennecessary to perform flow studies at the same time as treatmentand sampling from 5 separate catheters.

The demonstration of substantial nonportal drainage fromthe peritoneal cavity is an important finding. Qualitatively, theexistence of nonportal drainage is demonstrated by the lowerhepatic vein levels (compared to arterial levels) seen in 3 of 4patients. If all drainage were via the portal route, hepatic veinconcentrations (AUC) would equal the arterial concentration (Â¡fthe liver were the sole eliminating organ) or exceed arterialconcentration (when extrahepatic elimination was present).Due to uncertainties in the fraction of Gl tract metabolism ofrecirculating drug and of portal vein flow, our quantitativeestimates for the fraction of drug which is absorbed throughthe portal system range from 29 to 100% (Table 4). Otherpossible routes for drug to disappear from the abdomen includedirect binding to tissue, i.p. metabolism, egress via lymphaticsor nonportal (i.e., inferior mesenterio to inferior vena cava)venous routes, or retroperitoneal or muscle wall absorption.Direct entrance of drug to peritoneal tumor is the aim of i.p.therapy. It is possible that some drug penetrates peritonealtissues and does not physically leave the abdomen and is nolonger detectable as free drug. There is, however, no evidenceof drug metabolism in peritoneal fluid. No FdUrd was detectedin the fluid or plasma of 2 of these patients or in 2 patientstreated similarly as part of another trial. From these data, weare unable to affix a value to lymphatic or nonportal venousdrainage, other than it cannot exceed (1 â€”fPo).

Calculation of hepatic extraction also requires values forportal and hepatic arterial blood flow. As Table 3 indicates,however, the actual values for extraction are relatively insensitive to the ratio of portal to hepatic arterial flow over the rangewhich has generally been reported for flow measurements inhumans. Ensminger ef al. (13) reported hepatic extractions for5-FUra of 0.22 to 0.45 in similar studies using high-dose

hepatic arterial infusions, during which HV concentrations of24 to 94 ftM were reported. It is possible that the lower HVconcentrations achieved here indicate less saturation and,therefore, greater extraction.

The data obtained in this study provide additional documentation of the existence of extrahepatic clearance. The lowerAUCPevcompared to AUCPeX,(p < 0.05) suggests some elimination of 5-FUra even in peripheral tissues such as muscle.




Portal Uptake of i.p. 5-FUra

However, the total amount of clearance is not large: 24%extraction of a flow rate of 1 to 2 liters/min relative to totalbody clearance of about 5 liters/min.

Chart 3 presents a comparison of 5-FUra concentrations

calculated from our published model (10) and Equation Bversus measured concentrations in the peritoneal fluid, PoV,HV, and peripheral arterial plasma of Patient 3. The trendsobserved are typical of all 4 patients. The concentrations in all4 fluids were reasonably estimated by the model predictionsfor Exchange 7. The experimental data for hepatic vein andarterial plasma in Exchange 1, however, peak later and declinemore slowly than the simulations predict. The 2 major possibilities for this discrepancy are (a) fundamental failure of themodel structure or (fa) changes in patient parameters betweenexchanges.

In terms of model structure, one explanation for the lowerpeak concentrations and slower decline in Exchange 1 wouldbe the "storage" of 5-FUra by reversible formation of nucleo-

tides and other anabolic products. These anabolic productscould function as a large compartment which introduces a lagtime while it is being "filled up." The elimination kinetics could

be dominated by the rate of release from such a depot. Additionof such a compartment to our model did provide excellentagreement with the Exchange 1 data but forced a loss ofagreement for Exchange 7. Thus, the original model structurehas been retained.

It is quite plausible that changes in hepatic blood flow ratesand enzyme activity related to perioperative events stronglyaffect the first exchange, since it was carried out within 3 to 4hr of general anesthesia and surgery for catheter placement.Exchange 7, an additional 24 hr later, may have been carriedout under nearly "normal" values for patient parameters.

The ability of our model to predict changes in hepatic extraction as CHVÂ¡svaried as demonstrated in Chart 4 for Patient 2,Exchange 7. This patient and exchange were selected sincethe concentration in the HV was the highest measured in this

10,000

1000

I

C.V.

120 180 240 300 0 60

TIME (Minutes)

120 180 240 300

.563

.606

.449

392

335

Â¿78

Â¿21

.164

.107

.06

Chart 3. Model predictions and experimental data for Patient 3, Exchangesi (letti and 7 (right).

Chart 4. Instantaneous EH versus hepatic vein concentration (C,lv) of 5-FUrafor Patient 2, Exchange 7. Continuous curve represents model predictions, andindividual points are calculated from the data (Equation F). Half-saturating concentration of 15 JIM, maximum liver elimination capacity of 18 fimol/min/liter,and portal vein:hepatic artery blood flow ratio of 2:1 are assumed.

study and was greater than our assumed KM (half-saturating

concentration) of 15 JIM for the first half of the exchange andbelow the Km for the second half. No other data set exhibitedsufficient variation in CHVto challenge the model. The experimentally observed "instantaneous" extraction compares fa

vorably to the model calculations.Although it is difficult to fully validate any human pharmaco-

kinetic model due to limitations on sampling sites and otherethical considerations, the extensive data collected in thisstudy provide a unique opportunity for model testing. We areencouraged by the agreement shown in the Exchange 7 simulations and the instantaneous extraction calculations.

ACKNOWLEDGMENTS

We wish to thank Fred Gianola for assistance in sample collection and patientfollowup, Kathy Moore for manuscript preparation, and the professional staff ofthe Surgery Branch for their attention to the management of these patients.

REFERENCES

1. Almersjo, O., Gustavssan, B., and Hafstrom, L. Results of regional portalinfusion of 5-fluorouracil in patients. Ann. Chir. Gynaecol., 65. 27-32,1976.

2. Ansfield, F. J.. Ramiriz, G., Davis, H. L., Wirtanen, G. W., Johnson, R. O.,Bryan, G. T., Manale, F. B.. Borden. E. C.. Davis, T. E., and Esmaili. M. E.Further clinical studies with intrahepatic arterial infusions with 5-fluorouracil.Cancer (Phila.), 36. 2413-2417, 1975.

3. Ariel, I. M. Hepatic mÃ©tastasesfrom rectal and colon cancers. N. Y. State J.Med., 72. 26-29, 1972.

4. Band, J. D.. and Maki, D. G. Infection caused by arterial catheters used forhemodynamic monitoring. Am. J. Med., 67: 735-741, 1979.

5. Bengmark, S., and Hafstrom, L. The natural history of primary and secondarymalignant tumors of the liver. The prognosis for patients with hepatic diseasefrom colonie and rectal carcinoma by laparotomy. Cancer (Phila.), 23: 198-202, 1969.

6. Bevan, P. G. Cytotoxic perfusion of the liver via the umbilical vein formÃ©tastasesin carcinoma of the colon. Br. J. Surg.. 60. 369-375, 1973.

7. Cass, A. W., Million, R. R., and Pfaff, W. W. Patterns of recurrence followingsurgery alone for adenocarcinoma of the colon and rectum. Cancer (Phila.),37. 2861-2865, 1976.

8. Cedermark, B. J., Schulte, S. S., Bakshi, S., Parthasarathy, K. L.. Mittleman,A., and Evans, J. T. The value of liver scan in the follow-up study of patientswith adenocarcinoma of the colon and rectum. Surg. Gynecol. Obstet., 144:745-748, 1977.

9. Chiandussi, L., Goeco, F., Sardi. G., Vaccarino, A., Ferraris, C. M., andCurti, B. Estimation of hepatic arterial and portal venous blood flow by directcatheterization of the vena porta through the umbilical cord in man; prelim-

May 1981 1921



J. L. Speyer et al.

inary results. Acta Hepatosplen., 15: 166-171, 1968.10. Collins, J. M., Dedrick, R. L., King, F. G., Speyer, J. L., and Myers, C. E.

Nonlinear pharmacokinetic models for 5-fluorouracil in man: intravenousand intraperltoneal routes. Clin. Pharmacol. Ther., 28: 235-246, 1980.

11. Culbertson. J. W.. Wilkins, R. W., Ingelfinger, F. J., and Bradley, S. E. Theeffect of the upright posture upon hepatic blood flow in normotensive andhypertensive subjects. J. Clin. Invest., 30. 305-311, 1951.

12. Dencker, H., Gothlin, J., Olin, T., and Tibblin. S. Portal circulation in humansstudied by dye-dilution technique. Eur. Surg. Res., 4: 81-89, 1972.

13. Ensminger, W. D.. Rosowsky, A., Raso, V.. Levin. D. C., Glode. M.. Come,S., Steele, G., and Frei. E., III. A clinical-pharmacological evaluation ofhepatic arterial infusions of 5-fluoro-2'-deoxyuridine and 5-fluorouracil. Can

cer Res.. 38. 3784-3792, 1978.14. Floyd, E. C., Corley, R. G., and Conn, I. Local recurrence of carcinoma of

the colon and rectum. Am. J. Surg., 709: 153-159, 1965.15. Foster, J. H. Survival after liver resection for cancer. Cancer (Phila.), 26:

493-502, 1970.16. Goldman. M. L., Bilbao, M. K.. Rosch, J., and Dotter, C. T. Complication of

indwelling chemotherapy catheter. Cancer (Phila.), 36: 1983-1990, 1975.17. Grady, E., Nolan, F., Crumbley, A.. LaRose, J., and Cheek, W. Internal

hepatic radiotherapy II. Am. J. Roentgenol. Radium Ther. NucÃ.Med., 124:596-599. 1975.

18. Gunderson, L. L., and Sosin, H. Areas of failure found at reoperation (secondor symptomatic look) following "curative" surgery for adenocarcinoma of

the rectum. Cancer (Phila.). 34: 1278-1292. 1974.19. Gustavsson. B. G.. Brandberg, A., Regardh, C. G., and Almersjo, O. E.

Regional and systemic serum concentrations of 5-fluorouracil after portaland intravenous infusion: an experimental study in dogs. J. Pharmacokinet.Biopharm., 7:665-673, 1979.

20. Hillcoat. B. L.. McCulloch, P. B.. Figueredo, A. T., Ehsan. M. H., andRosenfeld, J. M. Clinical response and plasma levels of 5-fluorouracil inpatients with colonie cancer treated by drug infusion. Br. J. Cancer, 38:719-724, 1978.

21. Huet, P. M.. LaVoie, P., Legare, A., and Viallet, A. Combined hepatic vein,umbilico portal vein, and superior mesenteric artery catheterization in portalhypertension: estimation of the portal fraction of total hepatic flow in cirrhoticpatients. Yale J. Biol. Med., 48: 55-66, 1975.

22. Jaffee, B. M., Donegan, W. L.. Watson. F.. and Spratt, J. S. Factorsinfluencing survival in patients with untreated hepatic mÃ©tastases. Surg.Gynecol. Obstet.. 127.: 1-11. 1968.

23. Jones, R.. Myers, C.. Guarino, A., Dedrick, R., Hubbard, S.. and DeVita, V.T. High volume Â¡ntraperitoneal chemotherapy ("belly bath") for ovarian

cancer. Cancer Chemother. Pharmacol., 7: 161-168. 1978.24. Kirkwood. J. M., Ensminger, W.. Rosowsky, A., Papathanasapolous, and

Frei, E., III. Comparison of pharmacokinetics in 5-fluorouracil and 5-fluorouracil with concurrent thymidine infusions in a phase I trial. Cancer Res.,40: 107-113, 1980.

25. MacMillan. W. E., Wolberg, W. H., and Welling. P. G. Pharmacokinetics offluorouracil in humans. Cancer Res., 38: 3479-3482. 1978.

26. Moertel. C. G.. Schutt. A. J.. and Hahn, R. G. Therapy of advanced colorectalcancer with a combination of 5-fluorouracil methyl 1-3-cis(2 chloroethyl)-!-

nitrosourea and vincristine. J. Nati. Cancer Ins., 54: 69-71, 1975.27. Oberfield, R. A., McCaffrey, J. A., Polio, J., Clouse, M. E., and Hamilton, T.

Prolonged and continuous percutaneous intra-arterial hepatic chemotherapyin advanced metastatic liver adenocarcinoma from colorectal primary. Cancer (Phila.), 44: 414-423, 1979.

28. Ohnhaus. E. E. Methods of assessment of the effects of drugs in liver bloodflow in man. Br. J. Clin. Pharmacol., 7: 223-229, 1979.

29. Petrkek, J. A., and Minton, J. P. Treatment of hepatic mÃ©tastases bypercutaneous hepatic arterial infusion. Cancer (Phila.), 43: 2182-2188.

1979.30. Reichte, F. A., Sovak. M.. SÃ¤ulen, R. L., and Rosemond. G. P. Portal vein

blood flow determination in the unanesthetized human by umbilico portalcannulation. J. Surg. Res., 12: 146-150, 1972.

31. Sherman, D. M., Weichselbaum, R., Order, S. E., Cloud, L., Trey, C., andPiro, A. J. Palliation of hepatic mÃ©tastases.Cancer (Phila.), 41: 2013-2017,

1978.32. Speyer, J. L.. Collins, J. M.. Dedrick, R. L., Brennan, M. F.. Londer. H..

DeVita, V. T., and Myers, C. E. Phase I and pharmacological studies of 5-fluorouracil administered intraperitoneally. Cancer Res., 40: 567-572,1980.

33. Speyer, J. L., and Myers, C. E. The use of peritoneal dialysis for delivery ofchemotherapy to intraperitoneal malignancies. Recent Results Cancer Res.,74: 274-279, 1980.

34. Storer, E. H., and Akin, T. J. Chemotherapy of hepatic neoplasms. Am. J.Surg., 111: 56-58, 1966.

35. Strandell, T., Delin, A., Erwald. R., Kulling, K. G., Lundberg, P.. Marions, D.,and Weichel, K. L. Portal blood flow measurement by a catheter tip electromagnetic velocity probe in awake man. ActaChir. Scand., 139: 7-10, 1973.

36. Strandell, T., Erwald, R., Kulling. K. G., Lundberg, P., Marions, O., andWeichel, K. L. Simultaneous determination of portal vein and hepatic arteryblood flow by indicator dilution technique in awake man. Acta Med. Scand.,797: 139-140, 1972.

37. Strandell, T.. Erwald, R., Kulling, K. G., Lundberg, P., Marions, O., andWeichel, K. L. Measurement of dual hepatic blood flow in awake patients. J.Appi. Physiol., 35: 755-761, 1973.

38. Taylor, F. W. Cancer of the colon and rectum. A study of routes of mÃ©tastasesand death. Surgery (St. Louis). 52: 305-306, 1962.

39. Taylor, I. Cytotoxic perfusion for colorectal liver mÃ©tastases. Br. J. Surg.,65. 109-114, 1978.

40. Taylor, I., Brooman, P., and Rawling. J. T. Adjuvant liver perfusion incolorectal cancer: initial results of a clinical trial. Br. Med. J., 2:1320-1322,

1977.41. Turek-Masschesder, M., and Kazem, I. Palliative irradiation for liver mÃ©tas

tases. JAMA (J. Am. Med. Assoc.), 232: 625-628, 1975.42. Wade. O. L., Combes, B., Childs, A. W.. Wheeler, N. O., Cournand, A., and

Bradley, S. E. The effect of exercise on the splanchnic blood flow andsplanchnic blood volume in normal man. Clin. Sci. (Oxf.), 75: 457-463,

1956.43. Wood, C. B., Gillis, C. R., and Blumgart, L. H. A retrospective study of the

natural history of patients with liver mÃ©tastasesfrom colorectal cancer. Clin.Oncol., 2: 285-288, 1976.




1981;41:1916-1922. Cancer Res James L. Speyer, Paul H. Sugarbaker, Jerry M. Collins, et al. Intraperitoneal Administration in HumansPortal Levels and Hepatic Clearance of 5-Fluorouracil after

Updated version

http://cancerres.aacrjournals.org/content/41/5/1916

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://cancerres.aacrjournals.org/content/41/5/1916To request permission to re-use all or part of this article, use this link



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

mailto:[email protected]



Documents

Portal Levels and Hepatic Clearance of 5-Fluorouracil ... · of 5-FUra metabolism since, in some studies, plasma clearance of 5-FUra exceeds the hepatic blood flow (13, 20, 25, 32)