CLINICAL PHARMACOKINETIC CONCEPT
`
`Clin. Pharmacokinet. 25 (4): 300-328, 1993
`0312-5963/ 93/ 0004-0300/ $14.50/0
`© Adis International Limited. All rights reserved.
`CPKl357
`
`Individual Variation in First-Pass Metabolism
`
`Yun K. Tam
`Faculty of Pharmacy and Pharmaceutical Sciences, University of Alberta, Edmonton, Alberta, Canada
`
`Contents
`300
`301
`301
`307
`308
`31/
`311
`312
`314
`315
`317
`318
`318
`
`Summary
`
`Summary
`I. Physiological Factors
`1.1 Gastrointestinal Effects
`1.2 Liver
`1.3 Lung
`2. Factors Contributing to Interindividual Variation
`2.1 Age
`2.2 Diseases
`2.3 Enzyme Induction and Inhibition
`2.4 Food Effects
`2.5 Genetic Polymorphism and Ethnic Differences
`2.6 Gender
`3. Conclusions
`
`Individual variation in pharmacokinetics has long been recognised. This variability is ex(cid:173)
`tremely pronounced in drugs that undergo extensive first-pass metabolism. Drug concentrations
`obtained from individuals given the same dose could range several-fold, even in young healthy
`volunteers.
`In addition to the liver, which is the major organ for drug and xenobiotic metabolism, the
`gut and the lung can contribute significantly to variability in first-pass metabolism. Unfortunately,
`the contributions of the latter 2 organs are difficult to quantify because conventional in vivo
`methods for quantifying first-pass metabolism are not sufficiently specific. Drugs that are mainly
`eliminated by phase II metabolism (e.g. estrogens and progestogens, morphine, etc.) undergo
`significant first-pass gut metabolism. This is because the gut is rich in conjugating enzymes. The
`role of the lung in first-pass metabolism is not clear, although it is quite avid in binding basic
`drugs such as lidocaine (lignocaine), propranolol, etc.
`Factors such as age, gender, disease states, enzyme induction and inhibition, genetic poly(cid:173)
`morphism and food effects have been implicated in causing variability in pharmacokinetics of
`drugs that undergo extensive first-pass metabolism. Of various factors considered, age and gender
`make the least evident contributions, whereas genetic polymorphism, enzymatic changes due to
`induction or inhibition, and the effects of food are major contributors to the variability in first(cid:173)
`pass metabolism. These factors can easily cause several-fold variations. Polymorphic disposition
`of imipramine and propafenone, an increase in verapamil first-pass metabolism by rifampicin
`(rifampin), and the effects of food on propranolol, metoprolol and propafenone, are typical ex(cid:173)
`amples.
`
`Page 1
`
`RB Ex. 2023
`BDSI v. RB PHARMACEUTICALS LTD
`IPR2014-00325
`
`

`
`First-Pass Metabolism
`
`301
`
`Unfortunately, the contributions of these factors towards variability are unpredictable and
`tend to be drug-dependent. A change in steady-state clearance of a drug can sometimes be ex(cid:173)
`acerbated when first-pass metabolism and systemic clearance of a drug are simultaneously altered.
`Therefore, an understanding of the source of variability is the key to the optimisation of therapy.
`
`First-pass metabolism is a phenomenon char(cid:173)
`acterised by significant presystemic removal of
`drugs after oral administration. Drugs that undergo
`extensive first-pass metabolism often have low and
`variable bioavailability. It is common to see a 3-
`to 5-fold difference in the extent of absorption and
`the same or a greater divergence in the resultant
`steady-state plasma concentrations among individ(cid:173)
`uals.
`Interindividual variation in pharmacokinetics
`and its potential effect on drug therapy have long
`been recognised (Alvan 1978; Breimer 1983; Row(cid:173)
`land 1985; Vesell & Penno 1983). However, spe(cid:173)
`cific discussion of variability in first-pass metab(cid:173)
`olism is scanty (Pond & Tozer 1984; Rowland
`1985).
`This review examines factors that could be re(cid:173)
`sponsible for interindividual variation. Due to the
`vastness of the subject, it would be impossible to
`cover all drugs affected. Examples highlight the
`contribution of specific factors to interindividual
`variability. Literature data have been compiled to
`document the variability in pharmacokinetics of
`drugs that undergo extensive first-pass gut and liver
`metabolism (tables I, II). This review consists of 2
`sections. Section I deals with physiological factors
`that determine first-pass metabolism. Section 2
`presents specific factors that may contribute to
`interindividual variability.
`
`1. Physiological Factors
`
`During oral absorption, a drug has to traverse
`the gastrointestinal tract, the liver, the heart and
`the lungs and back to the heart sequentially before
`gaining access to the body's general circulation. The
`gastrointestinal tract, liver and lungs are capable of
`removing drugs. First-pass metabolism becomes
`important when a significant percentage of a dose
`is eliminated by I or more of these organs during
`
`the first passage of the drug through these organs.
`With this unique anatomical arrangement, the oral
`bioavailability (F) of a drug is dependent on the
`extent of first-pass removal by each organ. Quan(cid:173)
`titatively, F is expressed as:
`
`(Eq. I)
`
`where Fg, Fl and Fp are the fractions of drug that
`survive passage through the gastrointestinal tract,
`the liver and the lung, respectively, before entering
`the systemic circulation.
`
`1.1 Gastrointestinal Effects
`
`In vitro and in vivo studies have shown that
`many drugs and xenobiotics are metabolised in the
`gastrointestinal tract (see the review by lIett et al.
`1990). Metabolism occurs either in the gut lumen
`or the gut wall, and huge interindividual and in(cid:173)
`terspecies differences in these processes have been
`recorded.
`
`1.1.1 Gut Lumen
`Metabolic enzymes in the gut lumen originate
`from exocrine glands, cells that are shed from the
`mucosal lining and gut flora (Renwick 1982). In(cid:173)
`testinal micro-organisms, consisting of a complex
`mixture of aerobic and anaerobic organisms, are
`sensitive to the acidic environments of the upper
`gastrointestinal tract, and, therefore, are mostly lo(cid:173)
`cated in the colon and the lower portion of the
`small intestine.
`Enzymes produced by the host are more active
`in the lumen of the upper intestine and are inac(cid:173)
`tivated by the micro-organisms in the hind gut.
`There are few examples in the literature to show
`that these enzymes are important in the presys(cid:173)
`temic removal of drugs, partly because most of
`these enzymes are present in the gut wall. The hy-
`
`Page 2
`
`

`
`302
`
`C/in. Pharmacokinet. 25 (4) 1993
`
`Table I. Drugs or xenobiotics that are metabolised in the gut
`
`Drug or xenobiotic
`
`Mode of metabolism
`
`Reference
`
`lumen
`
`wall
`
`Alacepril
`Ethanol
`Ampicillin esters
`pivampicillin
`talampicillin
`Aspirin
`Buprenorphine
`Chlorpromazine
`Digoxin
`Fenoterol
`Flurazepam
`Hexamethylmelamine
`Isoprenaline (isoproterenol)
`Isosorbide dinitrate
`Levodopa
`
`Lorazepam
`Metronidazole
`
`Morphine
`
`Naloxone
`Estrogens
`estradiol
`ethinylestradiol
`
`Paracetamol (acetaminophen)
`Phenacetin
`Phenylethylamine
`Progestogens
`norethisterone (norethindrone)
`levonorgestrel
`Retinoic acid
`
`Sulfasalazine
`
`Sulfinpyrazone
`Sulphonamides
`Tyramine
`
`Matsumoto et al. (1986)
`Lamboeuf et al. (1981, 1983)
`
`von Daehne et al. (1970)
`Jeffery et al. (1978)
`Harris & Riegelman (1969); Iwamoto et al. (1981)
`Brewster et al. (1981); Mistry & Houston (1987)
`Curry et al. (1971)
`Dobkin et al. (1983); Robertson et al. (1986)
`Koster et al. (1985)
`Mahon et al. (1977)
`Klippert et al. (1983)
`George (1981); lIett & Davis (1982)
`Posadas et al. (1988)
`Iwamoto et al. (1987a); Sasahara et al. (1981);
`Granerus et al. (1973)
`Gerkens et al. (1981); Kraus (1978)
`Goldman et al. (1986); Koch & Goldman (1979);
`Koch et al. (1979)
`Brunk & Delle (1974); Dahlstrom & Paalzow (1978);
`Iwamoto & Klassen (1977); Mistry & Houston
`(1987); Siiwe et al. (1985)
`Mistry & Houston (1987)
`
`Diczfaluzy et al. (1961, 1962)
`Back et al. (1981a, 1982); Rogers (1987a); Stead et
`al. (1987); Schwenk et al. (1982)
`lIett & Davies (1982)
`Borm et al. (1983)
`Garcha et al. (1985); Worland & lIett (1983)
`
`Back et al. (1978)
`Stead et al. (1987)
`Cullum & Zile (1985); Ganguly (1969); Zile et al.
`(1982)
`Peppercorn & Goldman (1972); Schroder &
`Gustafsson (1973)
`Renwick et al. (1982); Strong et al. (1984)
`Krebs et al. (1947)
`Garcha et al. (1985); lIett & Davis (1982); Yasuhara
`et al. (1986)
`
`II
`I
`
`II
`
`II
`I
`
`II
`
`II
`
`II
`
`II
`II
`
`II
`I
`I
`
`II
`
`1&11
`
`II
`
`Abbreviations: I, II = Phase I and Phase II metabolism, respectively.
`
`drolysis of pivampicillin (Shindo et at. 1978) is an
`example where luminal esterases may have been
`involved.
`Intestinal flora are capable of performing a
`number of metabolic reactions, most of which are
`degradative, reductive and hydrolytic (Back &
`Rogers 1987; lIett et at. 1990; Renwick 1982). {3-
`
`Glucuronidase, {3-glycosidase and sulphatase in in(cid:173)
`testinal micro-organisms are responsible for hydro(cid:173)
`lysing biliary excreted estrogen conjugates (Hawk(cid:173)
`sworth et at. 1971). These processes are important
`for the enterohepatic recirculation of these ster(cid:173)
`oids. Failure of oral contraceptives containing nor(cid:173)
`ethisterone (norethindrone), ethinylestradiol and
`
`Page 3
`
`

`
`First-Pass Metabolism
`
`303
`
`mestranol has been associated with concurrent
`administration of antibiotics (Back et al. 1978;
`E1chdere et al. 1987); these agents have been shown
`to significantly lower the quantities of micro(cid:173)
`organisms in the lower gastrointestinal tract (AI(cid:173)
`dercreutiz et al. 1984; Gorbach 1984). This change
`is related to the decrease in estrogen reabsorption.
`The huge variability of digoxin bioavailability
`was found to be related to the dissolution rate of
`the oral formulation (Lindenbaum et al. 1973,
`198 I b). When the underlying mechanisms of this
`variability were investigated, Dobkin et al. (1983)
`showed that digoxin was inactivated via enzymatic
`reduction by gut flora and that this process was
`reversed with antibiotic therapy (Lindenbaum et
`al. 198Ia). Vrinary recovery of the reduced prod(cid:173)
`ucts, dihydro-digoxin and dihydro-digoxigenin,
`varied more than 12-fold (Lindenbaum et al.
`198 I b). The yields of these reduced products were
`higher when the dissolution rate of a formulation
`was slower. This observation was attributed to a
`higher percentage of the dose being delivered to the
`terminal portion of the gut where intestinal flora
`resided. Independent of the dosage form used, ap(cid:173)
`proximately 10% of patients excreted massive
`amounts of reduced digoxin products (>40% of
`combined urinary digoxin and its reduced metab(cid:173)
`olites). The reason for this observation was un(cid:173)
`clear. Alam et al. (1988) have shown, however, that
`there were interethnic variations in the occurrence
`of bacteria that reduced digoxin and that levels of
`micro-organisms were proportional to the extent of
`digoxin reduction in vivo.
`
`1.l.2 Gut Wall
`Metabolism of drugs and xenobiotics in the gut
`wall has been reviewed extensively (Back & Rogers
`1987; Caldwell & Marsh 1982; Bett & Davies 1982;
`Bett et al. 1990). Enzyme activities, in general, are
`higher in the mucosal epithelial cells of the duo(cid:173)
`denum and jejunum, and these activities decrease
`distally. Numerous metabolic reactions occur in the
`gut wall. These include phase I reactions such as
`oxidation, reduction and hydrolysis, and many
`phase II conjugation reactions including glucuron-
`
`idation, sulphation, N-acetylation, O-methylation
`and glutathione and glycine conjugation.
`
`Phase I Enzymes
`Enzyme activity in the small intestine is lower
`than in the liver (Back & Rogers 1987; Caldwell &
`Marsh 1982; Pacifici et al. 1988a,b). In humans,
`the liver to intestine cytochrome P450 (CYP450)
`ratio has been reported as ",,20, suggesting that the
`contribution of intestinal phase I biotransforma(cid:173)
`tion to the overall metabolism of a drug is unlikely
`to be important (Back & Rogers 1987). Neverthe(cid:173)
`less, there are examples of gut first-pass metabol(cid:173)
`ism that involve phase I reactions (table I). Areas
`under the plasma concentration curves (AVC) of
`levodopa were 3 times higher in elderly patients
`with Parkinson's disease than in young healthy
`volunteers (Evans et al. 1980). This suggests that
`there is an age-related reduction in gut wall decar(cid:173)
`boxylation. The bioavailability of hexamethylam(cid:173)
`ine in rats was low (8 to 27%), implying that N(cid:173)
`dealkylation, the major metabolic pathway, was
`highly variable in the gut wall (Klippert et al. 1983).
`Gut mucosal enzymes could be induced by
`agents such as phenobarbital, food substances such
`as cruciferous vegetables (brussels sprouts, cab(cid:173)
`bage, broccoli, cauliflower and spinach) [Pantuck
`et al. 1979], charcoal-broiled food (Conney et al.
`1977) and cigarette smoking (Pantuck et al. 1974).
`Starvation (Stohs et al. 1976), dietary deficiency of
`iron or cysteine/methionine (Edes et al. 1979) have
`been known to decrease intestinal aryl hydrocar(cid:173)
`bon hydroxylase activity. The importance of these
`factors in contributing to interindividual variation
`in bioavailability is not known. This is an area that
`requires more study.
`
`Phase II Enzymes
`While the concentration of phase I enzymes is
`relatively low in the gut, conjugative phase II en(cid:173)
`zyme activity is comparable to that in the liver
`(Back & Rogers 1987; Caldwell & Marsh 1982;
`Cappiello et al. 1989, 1991; Pacifici et al. 1988a;
`Romiti et al. 1992). The most important phase II
`reactions are glucuronidation and sulphation. A
`number of examples involving first-pass gut con-
`
`Page 4
`
`

`
`304
`
`Clin. Pharmacokinet. 25 (4) 1993
`
`Table II. Bioavailability of some highly extracted drugs in healthy volunteers and their source of variability (updated from Pond &
`Tozer 1984)
`
`Variability Comments
`(fold)
`
`Reference
`
`Drug
`
`Analgesics
`Codeine
`Morphine
`Pentazocine
`
`Pethidine
`(meperidine)
`
`F
`(%)
`
`42-71
`15-64
`11-32
`
`47-73
`
`Dextropropoxyphene 29-70
`
`Cardiovascular
`Alprenolol
`
`1-15
`
`Dilevalol
`Diltiazem
`Diprafenone
`
`Encainide
`
`12
`42 ± 18
`11 ± 1 (50mg)
`33 ± 26 (150mg)
`7-82
`
`Etilefrine
`
`17-35
`
`Felodipine
`Hydralazine
`
`4.4-36
`16.2 ± 6.3 (FA)
`35.4 ± 7.8 (SA)
`
`Indoramin
`2.1-77.2
`Isosorbide dinitrate 26
`Labetalol
`33 ± 3
`
`Lidocaine
`(lignocaine)
`
`38.7 ± 4.9
`
`Lorcainide
`
`0-89
`
`Metoprolol
`
`Nifedipine
`
`Nitrendipine
`Oxprenolol
`
`50
`
`56
`
`5-30
`40
`
`1.7
`4.3
`2.9
`
`1.6
`
`2.4
`
`15
`
`10
`3.1
`
`12
`
`2.1
`
`8.2
`5.5
`
`2.8
`
`7
`
`3.7
`
`3.9
`
`Findlay et al. (1977)
`Brunk & Delle (1974); Sliwe et al. (1981)
`Ehrnebo et al. (1977); Neal et al. (1979); Pond
`et al. (1980)
`Edwards et al. (1982b); Mather & Tucker
`(1976); Neal et al. (1979); Szeto et al. (1977);
`Verbeeck et al. (1981)
`Giacomini et al. 1980); Gibson et al. (1977,
`1980); Gram et al. (1979); Inturrisi et al.
`(1982); Pond et al. (1980, 1981)
`
`Ablad et al. (1974); Alvan et al. (1977b,c);
`Collste et al. (1979)
`
`Frishman et al. (1991); Tenero et al. (1989)
`Herman et al. (1983)
`Trenk et al. (1989)
`
`Gomoll et al. (1981); Roden et al. (1982);
`Wang et al. (1982); Winkle et al. (1981);
`Woosley et al. (1981)
`Hengstmann et al. (1982)
`
`Bailey et al. (1991); Blychert et al. (1991)
`Ludden et al. (1982); Reece et al. (1980);
`Semple et al. (1990); Shepherd et al. (1982,
`1984); Talseth (1976a,b)
`Pierce et al. (1987)
`Schaumann (1989)
`Homeida et al. (1978); Mantyla et al. (1980)
`
`i by cirrhosis
`
`i by cirrhosis; ~ by
`phenytoin
`
`i by cirrhosis, repeated
`doses, renal failure
`
`i by increasing single
`dose; ~ by pentobarbital
`(phenobarbitone)
`
`Polymorphic metabolism
`(DB type)
`
`i by oral
`dehydroergotamine
`i by grapefruit juice
`i by increasing single
`dose; ~food
`
`i by cirrhosis, food; ~ by
`pentobarbital
`i by cirrhosis; ~ by
`cigarette smOking,
`anticonvulsant drugs
`
`Bennett et al. (1982); Boyes et al. (1971);
`Colli et al. (1988); Drayer (1976); Huet &
`LeLorier (1980); Huet et al. (1978); Perucca &
`Richens (1979); Tschanz et al. (1977);
`Villeneuve et al. (1987)
`i by increasing single dose Jlinchen et al. (1979); Kates et al. (1983);
`and repeated
`Meinertz et al. (1979)
`administration
`Polymorphic metabolism
`(DB type); i by food,
`cirrhosis, repeated
`administration
`i by grapefruit juice
`
`Frishman et al. (1991); Haglund et al. (1979);
`Lennard et al. (1982); Melander et al. (1977b);
`Regardh & Johnsson (1980)
`
`Bailey et al. (1991); Kleinbloesem et al.
`(1984); McAllister et al. (1982)
`i by grapefruit juice
`Mikus et al. (1987); Soons et al. (1991)
`i by inflammatory diseases Frishman (1979); Frishman et al. (1991);
`Mason & Winer (1976)
`
`Page 5
`
`

`
`First-Pass Metabolism
`
`305
`
`Table II. Contd
`
`Drug
`
`Propranolol
`
`F
`(%)
`
`0-28
`
`Variability Comments
`(fold)
`
`Reference
`
`t by food, cirrhosis,
`Cleveland & Shand (1972); Evans & Shand
`(1973); Frishman et al. (1991); Homeida et al.
`inflammatory diseases,
`(1987); Kornhauser et al. (1978); Kowey et al.
`chlorpromazine,
`(1989); McLean et al. (1980); Melander et al.
`hydralazine, propafenone,
`(1977b); Perucca et al. (1984); Pessayre et al.
`repeated administration;
`t in males; ~ by cigarette
`1978); Schneck & Vary (1983); Schneider &
`smoking; sustained release Bishop (1982); Silber et al. (1982); Shand &
`formulation
`Rangno (1972); Shand et al. (1970); Vestal et
`al. (1979b); Walker et al. (1986); Walle et al.
`(1981, 1989); Watson et al. (1987); Wood et
`al. (1978)
`Axelson et al. (1987); Ho"mann et al. (1983);
`Lee et al. (1987); Siddoway et al. (1987)
`
`Polymorphic metabolism
`(DB type); t by cirrhosis,
`food
`
`Propafenone
`
`5-50
`
`10
`
`Quinidine
`
`>70
`
`Verapamil
`
`12-33
`
`2.8
`
`t with age, by cimetidine;
`~ by rifampicin
`
`Greenblatt et al. (1977); Guentert et al. (1979);
`Holford et al. (1981); Ochs et al. (1978); Ueda
`et al. (1976); Yu et al. (1982)
`Barbarash et al. (1987); Eichelbaum et al.
`(1980, 1981); Freedman et al. (1981); Kates et
`al. (1981); McA"ister & Kirsten (1982); Rahn
`et al. (1985); Reiter et al. (1982); Shand et al.
`(1981); Smith et al. (1984); Storstein et al.
`(1984); Wagner et al. (1982)
`
`Shi et al. (1987)
`
`Fotherby (1983); Orme et al. (1983)
`
`Me"strom et al. (1982); Schulz et al. (1983)
`Bondesson & Linstrom (1988)
`Br0sen & Gram (1988); Sallee & Pollock
`(1990)
`
`Br0sen & Gram (1988); Gram & Christiansen
`(1975); Potter et al. (1982); Sallee & Pollock
`(1990)
`Cooper & Ke"y (1979); Simpson et al. (1978)
`Saleh et al. (1990)
`Dahl (1976)
`Guentert et al. (1990)
`Alvan et al. (1977a); Gram & Over0 (1975)
`Love et al. (1981)
`
`Polymorphic metabolism
`(DB type)
`
`Polymorphic metabolism
`(DB type)
`
`t by cirrhosis; with age
`
`Moore et al. (1975); Nation et al. (1977b);
`Pentikainen et al. (1978)
`Ritschel et al. (1977)
`t by increasing single dose Christophidis et al. (1978)
`Regardh et al. (1990)
`Zimm et al. (1983)
`Bateman et al. (1980)
`Hartvig et al. (1990)
`Krause et al. (1991)
`
`t by a"opurinol
`
`64
`
`Oral contraceptives
`Norethisterone
`(norethindrone)
`40
`Ethinylestradiol
`Psychotherapeutic agents
`47.7 ± 11
`Amitriptyline
`12-81
`Clozapine
`Desipramine
`56 ± 4 (EM)
`73 ± 12 (SM)
`86 ± 13 (PM)
`39 ± 7 (EM)
`42 ± 19 (SM)
`71 ± 8 (PM)
`
`Imipramine
`
`Loxapine
`Medifoxamine
`Methotrimeprazine
`Moclobemide
`Nortriptyline
`Zimelidine
`Miscellaneous
`Clomethiazole
`
`Coumarin
`Fluorouracil
`Omeprazole
`Mercaptopurine
`Metoclopramide
`Tacrine
`Terguride
`
`21
`33-74
`66 ± 18
`46-59
`34-71
`
`12 ± 7
`
`3 ± 3
`28
`25-117
`16 ± 11
`32-100
`17.4
`18.6
`
`3-5
`
`3.5
`
`1.9
`6.8
`
`2.7
`
`3.3
`2.2
`
`1.3
`2.1
`
`4.7
`
`3.1
`6
`6.6
`
`Abbreviations and symbols: t = increased; ~ = decreased; DB = debrisoquine; FA = fast acetylator; SA
`EM = rapid extensive metaboliser; SM = slow extensive metaboliser; PM = poor metaboliser.
`
`slow acetylator;
`
`Page 6
`
`

`
`306
`
`Clin. Pharmacokinet. 25 (4) 1993
`
`jugation are cited in the literature (table I). Foth(cid:173)
`erby (1983) recorded huge variability in AUC val(cid:173)
`ues after oral administration of ethinylestradiol (10-
`fold) and norethisterone (5-fold), which are known
`to be conjugated in the gut wall. Similarly, Shi et
`al. (1987) reported a 3- to 5-fold variation in nor(cid:173)
`ethisterone availability. This huge variation was
`postulated by the investigators to be linked to gen(cid:173)
`etic differences. Gourlay et al. (1986) showed that
`morphine absorption has a huge interindividual
`variation (4.3-fold), possibly due to variations in
`conjugating enzyme activity. The low bioavaila(cid:173)
`bility of this opioid analgesic (26%) combined with
`large interindividual difference could lead to sub(cid:173)
`therapeutic concentrations in patients.
`Gut wall metabolism can be modified by many
`factors. The best known examples are competitive
`interaction between ethinylestradiol and either
`paracetamol (acetaminophen) or ascorbic acid in
`humans (Back et al. 1981b; Rogers et al. 1987a,b).
`The AUC of ethinylestradiol increased 22 and 48%
`with the coadministration of the respective drugs.
`Paracetamol and ascorbic acid compete with eth(cid:173)
`inylestradiol for the limited sulphate pool, result(cid:173)
`ing in an undesirable increase in systemic estrogen
`exposure. In these studies, the extent of interaction
`in the gut wall may have been overestimated be(cid:173)
`cause the liver contribution was not taken into ac(cid:173)
`count. Although there are other similar examples
`cited in the literature, such as interactions between
`morphine and fenoterol (Koster et al. 1985), mor(cid:173)
`phine and orciprenaline (meta proterenol) [Koster
`et al. 1985], tyramine and monoamine oxidase
`(MAO) inhibitors (Yasuhara et al. 1986), etc., again
`the significance in the alteration of first pass through
`the gut is not known, because the liver contribu(cid:173)
`tion was not quantified.
`
`1.1.3 Estimation of Gut Absorption
`Despite the difficulty involved in the quantifi(cid:173)
`cation of gut first-pass metabolism, there are meth(cid:173)
`ods available for estimating drug absorption from
`the gut (Harris & Riegelman 1969; Weiss 1990).
`When equal doses of a drug are administered via
`intraperitoneal and oral routes, the ratio of oral
`(AUCpo) to intraperitoneal (AUCip) AUC values
`
`provides an estimate of gut bioavailability (Harris
`& Riegelman 1969; equation 2). The validity of this
`method is based on the assumption of linear phar(cid:173)
`macokinetics:
`
`Fg = AUCpo/AUCip
`
`(Eq. 2)
`
`This approach has been used to estimate the gut
`availability of phenol (Cassidy & Houston 1980,
`1984), naphthol (Mistry & Houston 1985), mor(cid:173)
`phine, naloxone and buprenorphine (Mistry &
`Houston 1987). This method, however, did not
`permit differentiation of incomplete absorption and
`first-pass gut effects. To distinguish between these,
`appropriate metabolite data are required.
`For ethical reasons, it would be difficult to ad(cid:173)
`minister a drug intraperitoneally to humans for the
`estimation of gut absorption (Harris & Riegelman
`1969). Weiss (1990) proposed a model of hepatic
`first-pass metabolism. The basic assumption of this
`method was that the disposition of the drug and
`metabolites
`followed
`linear pharmacokinetics.
`Administration of the parent drug via the oral and
`intravenous routes were required. In addition to
`parent drug concentrations, plasma or urine data
`of a primary metabolite were necessary. The equa(cid:173)
`tion developed by Weiss (1990) is presented in
`equation 3:
`
`(Eq. 3)
`
`where Fm is the fraction of drug eliminated he(cid:173)
`patically after intravenous administration, Rmi is
`the dose-normalised AUC ratio of any primary
`metabolite following oral and intravenous admin(cid:173)
`istration of the parent drug. This method has been
`used by Weiss (1990) to estimate the gut absorp(cid:173)
`tion of lidocaine (lignocaine), triamterene, ketan(cid:173)
`serin, metronidazole, caffeine, phenazone (antipyr(cid:173)
`ine) and nomifensine. This method, unfortunately,
`cannot be used to quantify gut absorption if the
`drug undergoes significant first-pass gut metabol(cid:173)
`ism. The reason is that Fg can be greater than I,
`which is theoretically unsound, if Rmi is larger than
`1. This could happen when extensive first-pass gut
`metabolism occurs.
`Although it is difficult to quantify first-pass gut
`
`Page 7
`
`

`
`First-Pass Metabolism
`
`307
`
`metabolism, evidence of its occurrence in animals
`and humans is obtained from in vitro intestinal
`preparations and in vivo catheterisation of the por(cid:173)
`tal vein. The latter procedure has been performed
`in only a limited number of patients because of
`ethical reasons. During the absorption phase, a
`higher concentration of metabolites in the portal
`vein compared with a peripheral vein is indicative
`of presystemic gut metabolism. This approach was
`used by Mahon et al. (1977) to study the metab(cid:173)
`olism of flurazepam by the small intestine in
`patients. High concentrations of the mono- and
`dide-ethyl metabolites were detected in the portal
`vein, with lower concentrations in the hepatic vein
`and peripheral blood.
`
`1.2 Liver
`
`The liver is the major organ of biotransforma(cid:173)
`tion of drugs. Enzyme activity in this organ ranges
`from I to 50 times that observed in the gut wall
`(Back & Rogers 1987; Caldwell & Marsh 1982;
`Cappiello et al. 1989 1991; Pacifici et al. 1988a,b;
`Romiti et al. 1992). The liver receives approxi(cid:173)
`mately 25% of cardiac output. With its unique ana(cid:173)
`tomical position, this organ is perfused by part of
`the splanchnic IPortal circulation and hepatic ar(cid:173)
`teries, supplying roughly 75 and 25% of the total
`organ flow, respectively. The rate of drug elimin(cid:173)
`ation by the liver is dependent on the hepatic blood
`flow rate (Q), its intrinsic ability to metabolise, bind
`and excrete (through the biliary tract) drugs, (CLint),
`and the free fraction of drug in blood or plasma
`(fu).
`
`1.2.1 Theoretical Assessment of
`Hepatic Drug Removal
`There are a number of physiological models
`proposed to describe drug elimination by the liver.
`The characteristics of these models and their ap(cid:173)
`plications have been critically reviewed by Morgan
`and Smallwood (1990) and, therefore, are not dis(cid:173)
`cussed in this review.
`The 2 most studied models are the venous equi(cid:173)
`libration or well stirred model (Gillette 1971; Nies
`et al. 1976; Pang & Rowland 1977; Rowland et al.
`
`1973), and the undistributed sinusoidal or parallel(cid:173)
`tube model (Bass et al. 1976; Keiding 1976; Keid(cid:173)
`ing & Andreasen 1979; Winkler et al. 1973, 1974,
`1979). The basic assumptions of the venous equi(cid:173)
`libration model are instantaneous drug distribu(cid:173)
`tion and that the effiuent free drug concentration
`is in equilibrium with that in the liver. For the
`undistributed sinusoidal model, the liver is as(cid:173)
`sumed to be made up of a number of identical
`tubes, with enzymes distributed evenly around
`them, arranged in parallel to each other. At any
`point along the tube, unbound drug species are
`equilibrated between sinusoids and hepatocytes.
`Equations for the 2 models were derived with the
`assumption that linear pharmacokinetics hold. The
`2 models have been shown by Gray and Tam
`(1987), and Roberts and Rowland (1986a,b) to be
`the extreme cases of the series-compartment and
`dispersion model, respectively.
`Using Fick's principle, Pang and Rowland (1977)
`have shown that hepatic drug clearance is related
`to Q and E (equation 4):
`
`(Eq. 4)
`
`where E is the ratio of inlet (Cin) minus outlet con(cid:173)
`centrations (Cout) to Cin. E has a value ranging from
`o to 1. Drugs are said to have a low, medium or
`high extraction ratio when E is less than 0.3, be(cid:173)
`tween 0.3 and 0.7 or higher than 0.7, respectively.
`E is dependent on the 3 physiological parameters:
`Q, CLint and fu. The relative contributions of these
`parameters to E and subsequently to hepatic clear(cid:173)
`ance (CLH) are model dependent (equations 5a,b):
`
`CLH = Q[(fuCLin!)/(Q + fuCLin!)]
`(venous equilibration model)
`CLH = Q[l - e-(fuCLint/Q)]
`(undistributed sinusoidal model)
`
`(Eq. 5a)
`
`(Eq. 5b)
`
`By separately substituting the model dependent
`values of E from equations 5a and b into equation
`6a, liver bioavailability, Fl, can be estimated using
`equations 6b and c, respectively:
`F( = I - E
`(basic equation)
`
`(Eq. 6a)
`
`Page 8
`
`

`
`308
`
`Clin. Pharmacokinet. 25 (4) 1993
`
`Fl = [1/(1 + fuCLinJCm
`(venous equilibration model)
`Fl = e-(fuCLintlQ)
`(undistributed sinusoidal model)
`
`(Eq. 6b)
`
`(Eq. 6c)
`
`The overall changes in drug absorption and elim(cid:173)
`ination are reflected by the average steady-state drug
`concentration during multiple oral administration,
`(equation 7a):
`
`C;o = FD/CLST
`(basic equation)
`
`(Eq. 7a)
`
`where CLs is systemic clearance, 0 is the dose and
`T the dosage interval. For the sake of simplicity, it
`is assumed that the liver is the only organ by which
`the drug is eliminated. Therefore, CLs is equal to
`CLH. When the appropriate values of CLH from
`equations 6b and 6c are used to substitute CLs in
`equation 7a, equations 7b and 7c result:
`
`C~ = (D/T)/fuCLint
`(venous equilibration model)
`
`(Eq. 7b)
`
`C;o = [(D/T)(e-(fuCLintlQ»)]fQ[I - e-(fuCLint/Q)]
`(undistributed sinusoidal model)
`(Eq. 7c)
`
`Since the undistributed sinusoidal and venous
`equilibration models provide the limits of predic(cid:173)
`tion when physiological processes are perturbed, it
`is important to appreciate the qualitative and
`quantitative differences between model predictions
`so that this knowledge can be used to rationalise
`the effects of potential perturbations on first-pass
`metabolism and overall pharmacokinetics. A sim(cid:173)
`ulation has, therefore, been performed to evaluate
`the effects of enzymatic, protein binding and blood
`flow changes on bioavailability (fig. I) and average
`steady-state concentrations (fig. 2) of drugs with
`extraction ratios (E) ranging from 0.1 to 0.9.
`According to the results of this simulation, there
`are 2 observations to be made for drugs with low
`E values:
`(a) model difference in predictions of F, CLH
`or C~ is minimal (figs I and 2).
`(b) perturbations of any single physiological
`parameters have very little effect on F (fig. I).
`
`Regardless of the model used, perturbation of
`any single physiological parameter produces qual(cid:173)
`itatively similar changes in F (fig. I). Furthermore,
`F is changed to the same extent if alterations in
`CLint or Fu are of the same magnitude (fig. I). For
`example, if enzyme function is induced and CLint
`is increased 2 times, the reduction in F is the same
`as increasing the free fraction of the drug in blood
`twice. Both models predict an increase in F with
`an increase in Q (fig. Ic). The difference betseen
`the 2 models is their quantitative predictions of F;
`the divergence increases as E values increase. This
`implies that drugs that undergo extensive metab(cid:173)
`olism are more sensitive to physiological pertur(cid:173)
`bations if their pharmacokinetics are better de(cid:173)
`scribed by the undistributed sinusoidal model.
`Although the models predict similar trends in
`the changes of average steady-state oral drug con(cid:173)
`centrations when enzyme function and protein
`binding are altered (fig. 2), the magnitude of change
`predicted by the venous equilibration model is in(cid:173)
`dependent of the extraction ratio of a drug. This
`uniqueness can be appreciated if we examine equa(cid:173)
`tion 7b, which shows that C~~ is inversely related
`to Fu and CLint. In the case of the undistributed
`sinusoidal model, Fu and CLint are involved with
`Q in an exponential function, the significance of
`this function in the contribution to C~~ is depend(cid:173)
`ent on the ratio of fu and CLint to Q. The higher
`the ratio, implying a higher E value and more ex(cid:173)
`tensive first-pass metabolism, the more sensitive is
`C~~ to fu or CLint perturbations.
`Figure 2 shows that the venous equilibration
`model predicts no effect of Q on C~~ whereas the
`undistributed sinusoidal model predicts that C~~ is
`sensitive to Q changes; the sensitivity increases as
`E increases. The model difference can be appreci(cid:173)
`ated by comparing equations 7b and 7c. The for(cid:173)
`mer involves no Q in the equation, whereas Q is
`involved in an exponential function in the latters.
`
`1.3 Lung
`
`After oral administration, the lung is the last
`barrier through which a drug has to survive before
`entering the systemic circulation. This organ has
`
`Page 9
`
`

`
`First-Pass Metabolism
`
`309
`
`Venous equilibration model
`
`Induction
`Inhibition
`2 .0-r-....,-~-