Pharmacokinetic drug interactions with phenytoin (Part I).

Pharmacokinetic Drug Interactions

Clin. Pharmacokinet. 18 (I): 37-60, 1990 0312-5963/ 90/000 1-0037/$12.00/0 © ADIS Press Limited All rights reserved. CPKOO3244a

Pharmacokinetic Drug Interactions with Phenytoin (Part 1)1 Roger L. Nation, Allan M, Evans and Robert W. Milne School of Pharmacy, South Australian Institute of Technology, North Terrace, Adelaide, Australia

Contents

Summary ... .. ...................... , ....................................................................... .. ......................... ..... .. 38 I. Interactions Affecting the Pharmacokinetics of Phenytoin ................................................ 39 1.1 Pharmacokinetic Properties ............................... .. ............................. .... .......................... 39 1.2 Drugs Which May Affect the Gastrointestinal Absorption ..... .................................... 41 1.2.1 Nutritional Formulae .... ..... ..................... .......................... ....... ... ....... ............. ....... 41 1.2.2 Charcoal ......................................................... ........ ..... ............... .......... .. ..... .......... .. 41 1.2.3 Antacids .. ................ ......... .. ........... .... ..... ... ............. ... ............. ... ........... ............. ... .... 42 1.2.4 Sucralfate ........ ................................... ............ .......... ... .................. ..................... ...... 43 1.2.5 Theophylline ........... ... ... ... .......... ... ...... ........................ ............ ... ........... ........ ........ .. 43 1.2.6 Antineoplastic Agents ................. .... .............. ... .. ....... ....... ..... ... ........... ... ................. 43 1.2.7 Other Drugs .... ... ..................................................................................................... 43 1.3 Drugs Which May Alter the Plasma Protein Binding .......... ............. .......................... 43 1.4 Drugs Which May Increase the Metabolism ........ .... .... .......... .......... ............... ..... ........ 44 1.4.1 Folic Acid .......... .. ........... .... ... .... ..... .. ..... .. ....... .. .... ...... .. ................. ................. ......... 44 1.4.2 Alcohol .. ....................... ................................................................ ... ............ ... .. ... .... .45 1.4.3 Dexamethasone ....................................................... ... .. .. ... ... ....... .. ........ ........... .. ..... 46 1.4.4 Anticonvulsant Agents .............. ..................... .. ........ .. ... ......... ...... ..... ..... ....... ... ...... 46 1.4.5 Diazepam and Chlordiazepoxide ............................. ..... .... ... ....... .. ..... .......... ..... .... 47 1.4.6 Dichlorafphenazone .. ...... ...... ...... .. ....... .... ............................................................... 47 1.4.7 Rifampicin ... .... .. ....... .. .......... .. ....... ... .. .. ... .. ..... ... ...... .... ...... .. ....... ... ....... .. ..... ... ..... ... 47 1.4.8 Nitrofurantoin .... .......... ......... ....... ................... .. ............................ ........ ..... .. ....... .... 47 1.5 Drugs Which May Reduce the Metabolism ........................................................ ......... 47 1.5.1 Anticonvulsant Agents ................... ................ .. ....... ... ..... .................... .......... ......... 47 1.5.2 Oral Anticoagulants .. ..... ................................. ... .................. ......... ......... ..... ..... .... .. . 49 1.5.3 Antituberculosis Drugs ............................................................................... ............ 49 1.5.4 Amiodarone ............... ................. ............................................................................ 49 1.5.5 Allopurinol ...... ..... ..... ... ....... .... ....... ....... ....... ......... ........... ... ....... ......... ...... .... .... ...... 49 1.5.6 H2-Antagonists ..... .... .. ..... .. ............ ............ .. ........ .... .............. ... ............................. .. 50 1.5.7 Omeprazole ...... ........ ...... ................. ....... .... .... .... .... ... ..... ..... .. ... ..... .. ....... .. ....... ... ..... 51 1.5.8 Non-Steroidal Anti-Inflammatory Agents ........................................................ .... 51 1.5.9 Disulfiram .. .... ................. .......................... .... ............... .. .................. ........ .... ... ....... . 52 1.5. 10 Anti-Infective Agents ..... .. .................................... ........... ............ ......................... 52 1.6 Influence of Other Drugs on Plasma Phenytoin Concentration .. ........ .. .. ................... 53

I A complete reference list will appear in Part II of this article in the following issue of the JournaL

38

Clin. Pharmacokinel. 18 (I) 1990

2. Interactions Affecting the Pharmacokinetics of Other Drugs ............................................ 2.1 Gastrointestinal Absorption ........................................................................................... 2.2 Plasma Protein Binding .................................................................................................. 2.3 Drugs Whose Metabolism May Be Increased ............................................................... 2.3.1 Anticonvulsant Agents ........................................................................................... 2.3.2 Analgesics ................................................................................................................ 2.3.3 Theophylline ...........................................................................................................

Summary

53 53 54 55 56 58 60

Phenytoin, which is used primarily as an anticonvulsant agent, has a relatively low therapeutic index, and monitoring of plasma phenytoin concentration is often used to help guide therapy. It has properties which predispose it to an involvement in pharmacokinetic interactions, a large number of which have been reported. These properties include: low aqueous solubility and slow rate of gastrointestinal absorption; a relatively high degree of plasma protein binding; a clearance that is non-linear due to saturable oxidative biotransformation; and the ability to induce hepatic microsomal enzymes. Because of its narrow therapeutic range, drug interactions leading to alterations in plasma phenytoin concentration may be clinically important. Such interactions have often been reported initially as either cases of phenytoin intoxication or of decreased effectiveness. Drugs may modify the pharmacokinetics of phenytoin by altering its absorption, plasma protein binding, or hepatic biotransformation; alterations in the absorption and/ or biotransformation may lead to changes in both the unbound plasma phenytoin concentration and, as a result, the clinical effect. Preparations which may decrease the gastrointestinal absorption of phenytoin include nutritional formulae and charcoal. There are many reports of drugs which may increase (e.g. folic acid, dexamethasone and rifampicin) or decrease (e.g. valproic acid, sulthiame, isoniazid, cimetidine, phenylbutazone, chloramphenicol and some sulphonamides) the metabolism of phenytoin. It is important to bear in mind that, as a result of its non-linear clearance, changes in phenytoin absorption and/or biotransformation will lead to more than proportionate changes in plasma drug concentration. Drugs which may displace phenytoin from plasma albumin include valproic acid, salicylic acid, phenylbutazone and some sulphonamides. Although an alteration in the unbound fraction of phenytoin in plasma would not, in itself, be expected to alter the unbound plasma phenytoin concentration, the interpretation of total plasma concentrations for therapeutic drug monitoring may be confounded. Some drugs appear to alter phenytoin pharmacokinetics via dual mechanisms (e.g. valproic acid and phenylbutazone), while for other compounds the mechanism of interaction has not been fully elucidated. Phenytoin has been reported to alter the pharmacokinetics of a large number of drugs. The majority of these interactions arise because phenytoin is a potent inducer of cytochrome P450 microsomal enzymes, and therefore may increase the clearance of drugs which are extensively metabolised; drugs affected include carbamazepine, theophylline, methadone, prednisolone, dexamethasone, metyrapone and several cardiac antiarrhythmic agents. With all of these, the resultant decrease in plasma concentrations may be clinically important. Current evidence suggests that the various cytochrome P450 isozymes may be affected differentially and, therefore, that the metabolic clearance of some drugs may be more sensitive to the enzyme-inducing effects of phenytoin. In addition, phenytoin also appears to increase the clearance of some drugs which are eliminated predominantly by conjugation reactions [e.g. paracetamol (acetaminophen) and oxazepam]. Interestingly, phenytoin has been reported to decrease the metabolic clearance of a commonly coadministered anticonvulsant, phenobarbital. For a few compounds only [e.g. furosemide (frusemide) and thyroxine], it has been suggested that phenytoin decreases gastrointestinal absorption. Although the binding of drugs to plasma albumin is largely unaffected by this agent, plasma levels of aI-acid glycoprotein and sex hormone

Pharmacokinetic Interactions with Phenytoin

39

binding globulin may increase during phenytoin treatment, thereby altering the binding of some compounds.

Phenytoin (diphenylhydantoin; 5,5-diphenylimidazolidine-2,4-dione) is one of the most commonly used anticonvulsant agents and is also employed for the treatment of cardiac arrhythmias and for the control of certain pain conditions such as trigeminal neuralgia. As it is a drug of relatively low therapeutic index, plasma concentration monitoring of phenytoin is often used to guide therapy. The usually accepted therapeutic plasma concentration range for total (bound plus unbound) phenytoin is between 10 and 20 mg/L (Winter & Tozer 1986). While it is generally assumed that the unbound concentration in plasma is approximately equilibrated with that at the site of therapeutic action, earlier difficulties with the routine determination of the unbound plasma concentration of phenytoin meant that the therapeutic plasma concentration range for unbound phenytoin was not as well defined. However, it appears that this range is approximately I to 2 mg/L (Kilpatrick et al. 1984; Perucca 1984). As the concentration of phenytoin approaches and exceeds the upper end of the usual therapeutic range there is an increased probability that a patient will experience phenytoin toxicity, in particular eNS-related side effects (Winter & Tozer 1986). Many drug-drug interactions involving phenytoin have been reported. These include cases in which other drugs modify the pharmacokinetics of phenytoin as well as those where phenytoin, usually through its potent effect of inducing hepatic microsomal enzymes, alters the pharmacokinetics of other drugs. Both of these categories of drug interactions involving phenytoin are reviewed.

1. Interactions Affecting the Pharmacokinetics of Phenytoin It is important to briefly review the pharmacokinetics of phenytoin in order to fully appreciate the manner in which they may be influenced by other drugs.

1.1 Pharmacokinetic Properties Phenytoin, a weak acid with a pKa of 8.3, is practically insoluble in water. A parenteral preparation containing the sodium salt of phenytoin is available, although in most instances the drug is administered orally. While the rate of gastrointestinal absorption is slow, the extent of absorption from high quality products is almost complete (Winter & Tozer 1986). The apparent volume of distribution of phenytoin is approximately 0.6 to 0.7 L/kg of bodyweight. Plasma binding is almost exclusively to albumin; in individuals with normal plasma albumin concentration and in the absence of displacing agents, phenytoin is about 90% plasma bound (Winter & Tozer 1986). Phenytoin is cleared predominantly by metabolism, with less than 5% of an administered dose being excreted unchanged in the urine. The parahydroxylated metabolite, 5-(p-hydroxyphenyl)-5phenylhydantoin (P-HPPH), and its glucuronide conjugate, are excreted in urine and account for 60 to 90% of an oral dose of phenytoin. Lesser amounts appear as free and conjugated meta-hydroxylated and dihydrodiol metabolites (Chang & Glazko 1982). It would appear that the majority of the metabolites are formed via an epoxide intermediate (Winter & Tozer 1986). Importantly, the clearance of phenytoin is concentration-dependent within the usual therapeutic plasma concentration range, as a result of saturable biotransformation. The reported Michaelis-Menten constant (Km) and maximum rate of metabolism (V max) values vary considerably among patients (Winter & Tozer 1986). The percentage of a phenytoin dose excreted in urine as p-HPPH is essentially independent of dose (Kutt 1971) and would be expected to be relatively insensitive to changes in Km and Vmax. However, since only a small percentage of a dose is excreted unchanged

40

Clin. Pharmacokinet. 18 (/) 1990

in urine, the ratio of the urinary recovery ofphenytoin to that of the combined phenytoin and pHPPH will be influenced by the plasma concentrations of phenytoin, and also by alterations in Km and V max . The non-linear clearance of this agent assumes great importance in the design and interpretation of drug interaction studies. The effects of an enzyme inhibitor, for example, would be expected to be greater at higher plasma phenytoin concentrations. Hence, the failure to observe a significant effect of another drug on the pharmacokinetics of a single dose of phenytoin does not preclude the possibility of a clinically significant interaction when phenytoin is administered over the long term. Little is known about which cytochrome P450 isozyme(s) catalyses the biotransformation of phenytoin. While the drug itself is achiral, p-HPPH contains a chiral carbon and therefore 2 configurations are possible. The aromatic hydroxylation of phenytoin is product-stereoselective with the majority of p-HPPH recovered in urine being of the S-configuration. It is interesting that the major route of phenytoin metabolism, pro-S-phenytoin hydroxylation, is maintained in subjects with genetic deficiency for debrisoquine and mephenytoin hydroxylation (Fritz et al. 1987). Although it appears that the cytochrome P450 isozyme(s) responsible for the aromatic hydroxylation of S-mephenytoin is not involved in the formation of S-p-HPPH from phenytoin, the mephenytoin isozyme(s) may catalyse the formation of R-p-HPPH, which is a minor metabolite of the drug (Fritz et al. 1987). Applying physiological pharmacokinetic concepts (Wilkinson & Shand 1975), it is possible to isolate the mechanisms by which other drugs may influence the pharmacokinetics of phenytoin. For this agent, which is cleared predominantly by the liver, the area under the plasma concentration-time curve for total (AUC) and unbound (AUC u) species following the administration of a single dose is given by equations I and 2, respectively:

AUC=

fa" D

(Eq. l)

fa" D

(Eq. 2) CLint where fa is the fraction of the dose (D) absorbed, fu is the unbound fraction in plasma and CLint is the intrinsic clearance, a measure of enzymatic activity. With maintenance dosing at a set dosage interval (T), the average steady-state concentration of total (CSS) and unbound (OJ) phenytoin in plasma is given by equations 3 and 4: AUC u =

(Eq. 3)

CSS =

(Eq. 4) CLint Since the total urinary recovery of phenytoin and p-HPPH and its glucuronide conjugate may be used as an index of fa, and fu can be measured directly, it is possible to deduce whether the CLint of phenytoin has been altered by another drug (equations I and 3). Clearly, plasma concentrations of unbound phenytoin, generally assumed to be the pharmacologically important species, may be influenced by alterations in fa and/or CLint (equations 2 and 4). Changes in these concentrations may lead to alterations in pharmacological and/or toxicological response. While unbound phenytoin concentrations are not sensitive to changes in fu alone (equations 2 and 4), the importance of variations in the plasma binding of the drug should not be underestimated. For example, an increase in fu will lead to a decrease in the plasma concentration of total phenytoin while unbound concentration remains unchanged and, therefore, therapeutic monitoring of total drug concentrations may lead to inappropriate dosage adjustments. It is clear that other drugs may produce clinically important changes in the pharmacokinetics of phenytoin by altering its gastrointestinal absorption, its plasma protein binding and/or its hepatic biotransformation. Each of these mechanisms is discussed in turn.


1.2 Drugs Which May Affect the Gastrointestinal Absorption 1.2.1 Nutritional Formulae Decreased gastrointestinal absorption of phenytoin by concurrent nasogastric-tube feeding was first reported by Bauer (1982). In 10 neurosurgery patients receiving phenytoin suspension 300mg daily, the mean (± SD) serum phenytoin concentration decreased from 9.80 ± 3.27 mg/L to 2.72 ± 1.09 mg/L after 7 days of nasogastric-tube feeding with 'Isocal'. In another 10 patients receiving the same dose, the mean serum phenytoin concentration increased from 2.59 ± 0.96 mg/L to 10.22 ± 2.90 mg/L, 7 days after discontinuation of nasogastric-tube feeding. Bauer (1982) also found that the serum concentrations of phenytoin were reduced significantly in healthy volunteers given oral phenytoin while concurrently drinking 'Isocal'. Subsequent reports have confirmed that concurrent administration of phenytoin with nutritional formulae may significantly reduce phenytoin absorption (Cosh et al. 1987; Hatton 1984; Maynard et al. 1987; Pearce 1988; Worden et al. 1984). All of these reports have involved the nutritional formulae 'Isocal' or 'Osmolite'. Interestingly, Hooks et al. (1986) found that 'Osmolite' sorbed phenytoin in vitro. Both Krueger et al. (1987) and Nishimura et al. (1988) found the formula 'Ensure' to have no effect on phenytoin absorption in healthy volunteers. The mechanism of the interaction is not understood, although binding of phenytoin to nasogastric tube apparatus has been largely excluded if the tube is flushed after dosing (Cacek et al. 1986). Even when the phenytoin is administered 2 hours before or after the enteral feed, higher doses are required to obtain therapeutic concentrations (Maynard et al. 1987; Ozuna & Friel 1984). It is suggested that patients receiving oral forms of phenytoin be monitored carefully when feeding with a nutritional formula is commenced or ceased. Caution should also be exercised if changes are made in the type offormula administered. To minimise the clinical effect of the interaction, Pearce (1988) suggested that the phenytoin should be administered as a single daily dose of the suspension,

41

8

:::J 0;

.so c:

6

8 c:

~

c: ~ a.

§ ~

4

2

0~~~~======~-=-'-~~---4 0124

8

12

Time (h)

Fig. 1. Effect of SOg of activated charcoal on the serum concentrations of phenytoin SOOmg administered orally: 0 = without charcoal; • = plus charcoal SOg; 0 = plus charcoal SOg, after 1 hour (from Neuvonen et al. 1978, with permission).

diluted with water, during a period when the enteral feeding is routinely ceased, perhaps overnight. It should also be recognised that the effect of enteral feeds on phenytoin absorption may complicate the interpretation of drug-drug interaction reports involving this agent. 1.2.2 Charcoal Neuvonen et al. (1978) found that activated charcoal, given immediately following oral ingestion of a dose of phenytoin 500mg, almost completely prevented absorption of the drug. In addition, when a single dose of activated charcoal was administered 1 hour after a dose of phenytoin there was an estimated 80% reduction in absorption compared with when the phenytoin was administered alone (fig. 1). Recently, Mauro et al. (1987) examined the effect of gastrointestinal dialysis with activated charcoal, in combination with sorbitol, on the elimination of intravenously administered phenytoin in healthy subjects. Gastrointestinal dialysis reduced both the apparent half-life and AVe of phenytoin by 50%. In a previous case report of phenytoin and phenobarbital overdose (Mofenson et al. 1985), gastrointestinal dialysis with activated charcoal did not have any dramatic effect on serum phenytoin concentrations. However, Weichbrodt and Elliot

42

(1987) reported a favourable effect of repeated doses of activated charcoal and magnesium citrate in a patient who ingested at least 109 of phenytoin. The shortened recovery time in this patient, compared with conventionally treated overdose cases, prompted the authors to suggest that multiple-dose activated charcoal may be of value in treating cases of phenytoin overdose (Weichbrodt & Elliot 1987).

1.2.3 Antacids In 1975, Kutt cited an observation by Pippinger (unpublished data) that plasma phenytoin concentrations were low in 3 patients taking phenytoin and antacids (of unspecified composition) simultaneously. When the antacid preparations were given 2 to 3 hours after the ingestion of phenytoin, plasma drug concentrations rose 2- to 3-fold. In 1978, O'Brien et al. reported that the onset of poor seizure control in 2 epileptic patients receiving phenytoin coincided with concurrent antacid treatment for dyspepsia. Seizure control was improved in both patients when antacid treatment was withdrawn. As a consequence of these latter cases, this same group of workers investigated the influence of the antacids involved (aluminium hydroxide and magnesium hydroxide) on the disposition ofa single oral dose of phenytoin in 6 healthy males. Neither antacid altered the peak plasma concentration, the time of its occurrence, or the AVe of phenytoin. However, the authors conceded that because of the non-linear pharmacokinetics of the drug, steadystate plasma concentrations may be a more sensitive index of changes in the extent of gastrointestinal absorption (O'Brien et al. 1978). Subsequently, Kulshrestha et al. (1978) found that in 6 patients stabilised on oral phenytoin, concurrent dosing with therapeutic doses of 'Gelusil' (a proprietary antacid combining aluminium hydroxide and magnesium trisilicate) produced a slight, but statistically significant, fall in steady-state plasma phenytoin concentrations. Vpon cessation of the antacid, the values returned to the pre-treatment level. In another 6 patients, calcium carbonate was found to have no effect. In all 12 patients, there was no alteration in seizure frequency as a result of antacid treatment.

Clin. Pharmacokinet. 18 (1) 1990

ehapron et al. (1979) found calcium gluconate and aluminium hydroxide with magnesium hydroxide or magnesium trisilicate to have no effect on the AVe of phenytoin after a single oral dose. However, only 2 subjects were involved in this study, and for one it is apparent that the timing of blood sampling did not permit adequate definition of the plasma phenytoin concentration-time profiles. In 1981, Carter and others detailed the results of a study, reported earlier in summarised form (Garnett et al. 1979, 1980), in which 8 healthy subjects received a single oral dose of phenytoin alone (on 2 occasions) and in combination with 3 antacid preparations. The phenytoin AVe was decreased significantly by an aluminium hydroxide-magnesium hydroxide preparation, and by calcium carbonate. Although a mixture of aluminium hydroxide and magnesium trisilicate reduced the mean phenytoin AVe, the decrease was not significant. It should be emphasised that the dose volumes of the antacids employed in this study (between 40 and 90ml) were greater than those used by other investigators. The effect of 'Asilone', a product combining aluminium hydroxide, magnesium hydroxide and activated dimeticone, on the relative bioavailability of a single oral dose of phenytoin was investigated by McElnay et al. (1982). In 5 of the 6 healthy subjects studied, the AVe of phenytoin was decreased during concurrent antacid administration, and in 3 subjects this decrease was greater than 30%. However, in 1 volunteer, the AVe increased by 60% during the antacid treatment phase, and for this reason, the effect of the antacid failed to reach significance. In the same study, dimeticone alone was found to have no effect on phenytoin AVe. Hence, the few controlled studies investigating the influence of concurrent antacid administration on the disposition of phenytoin have produced conflicting results. Both the timing of antacid dosing and the volume of antacid used may play important roles in explaining these conflicts. Overall, it appears that the influence of antacids is variable, both between antacid preparations and between subjects. Future studies investigating the interaction will therefore need to consider carefully the


protocol adopted and subject numbers. It is generally recommended that, if antacids are to be used in patients receiving phenytoin. the administration of the 2 agents should be timed well apart. The clinical significance of the phenytoin-antacid interaction was recently reviewed by D' Arcy and McElnay (1987). 1.2.4 Sucra/fate In 1985, Smart and co-workers conducted a double-blind placebo-controlled study in 8 healthy volunteers to examine whether the disposition of a single oral dose of phenytoin 300mg was influenced by the concurrent administration of a dose ofsucralfate Ig; they found that the AVe ofphenytoin was significantly reduced (average 20%). In a crossover study involving 9 healthy volunteers, Hall et al. (1986) also found that when sucralfate was administered together with a single oral dose of phenytoin 500mg, the area under the serum phenytoin concentration-time profile (0 to 48h) was reduced, although by an average of only 9.5%. The authors stressed that for a drug such as phenytoin, a decrease in AVe can result from a change in either the rate or extent of absorption. The moderate reductions in serum phenytoin concentrations found in these 2 single-dose studies (Hall et al. 1986; Smart et al. 1985) may underestimate those which could occur with long term phenytoin dosing. Further studies are required to assess the effect of long term sucralfate administration on phenytoin concentrations under such conditions. 1.2.5 Theophylline In 2 patients stabilised on phenytoin, the commencement of theophylline therapy coincided with a decrease in phenytoin concentrations (Hendeles et al. 1979). In response to this observation, these workers examined the effect of theophylline on the pharmacokinetics of a single dose of phenytoin in 7 healthy adults. When phenytoin and theophylline were administered simultaneously, phenytoin AVe was reduced significantly by an average of 21 %. When the administration of the 2 drugs was separated by 2 hours, the mean phenytoin AVe decreased by 7%, although the change was not sig-

43

nificant. It was suggested by the authors that theophylline decreased the absorption of phenytoin when the 2 agents were administered at the same time (Hendeles et al. 1979). 1.2.6 Antineoplastic Agents There are a number of case reports in which the commencement of antineoplastic therapy in patients being treated with phenytoin has coincided with decreases in plasma phenytoin concentrations and/or a loss of seizure control (Bollini et al. 1983; Fincham & Schottelius 1979; Sylvester et al. 1984). In all of these cases it was suggested that antineoplastic therapy impaired the gastrointestinal absorption of phenytoin. More recently, Neef and de Voogd-van der Straaten (1988) reported the case of an epileptic patient, receiving phenytoin, carbamazepine and valproic acid, who experienced seizures and altered plasma anticonvulsant concentrations during antineoplastic therapy. The authors suggested that antineoplastic agents may also increase the metabolism of phenytoin, although the evidence presented is not convincing. 1.2.7 Other Drugs Hetzel et al. (1981) have suggested that cimetidine may increase the bioavailability of orally administered phenytoin, although it appears that additional factors may be involved (see section 1.5.6). Oral oxacillin may have been responsible for impaired phenytoin absorption in a patient receiving both drugs concurrently (Fincham et al. 1976). 1.3 Drugs Which May Alter the Plasma Protein Binding A large number ofxenobiotics have been shown to displace phenytoin from plasma protein binding sites, leading to an increased phenytoin unbound fraction. Displacement will have no effect on the steady-state unbound concentration of phenytoin in plasma and, therefore, clinical response is unlikely to be altered. However, as discussed in section 1.1, the total phenytoin concentration will be reduced. This altered relationship between total and unbound values may compromise therapeutic drug

44

Clin. Pharmacokinet. 18 (I) 1990

Table I. Drugs which may displace phenytoin from plasma protein binding sites Drug

Reference

Azapropazonea Diazoxide b Halofenate Heparin Ibuprofen Phenylbutazone a Salicylic acid

Geaney et al. (1983) Roe et al. (1975) Karch & Wardell (1977) Schulz et al. (1983) Bachmann et al. (1986) Lunde et al. (1970) Fraser et al. (1980) Leonard et al. (1981) Lunde et al. (1970) Paxton (1980) Hansen et al. (1979) Hansen et al. (1979) Lunde et al. (1970) Hansen et al. (1979) Lumholtz et al. (1975) Hansen et al. (1979) Hansen et al. (1979) Wesseling & Mols-Thurkow (1975) Bruni et al. (1980) Dahlqvist et al. (1979) Monks et al. (1978) Monks & Richens (1980) Patsalos & Lascelles (1977) Perucca et al. (1980) Sansom et al. (1980)

Sulfadiazinea Sulfadimethoxine Sulfafurazole Sulfamethizole a Sulfamethoxydiazine Sulfamethoxypyridazine Tolbutamide Valproic acid a

a b

May also decrease phenytoin metabolism (see section 1.5). May also increase phenytoin metabolism (see table II).

monitoring of total plasma phenytoin concentrations. Drugs for which there is evidence for displacement of phenytoin from plasma protein binding sites are listed in table I. It should be noted that a number of these drugs have been found to also alter the metabolism of phenytoin. The heparin-induced increase in phenytoin unbound fraction (table I) is mediated by liberation of non-esterified fatty acids (NEFAs). Caution may be needed when measuring the unbound fraction, and unbound concentration, of phenytoin in heparinised patients because of continued generation of NEFAs in vitro (Schulz et al. 1983).

1.4 Drugs Which May Increase the Metabolism 1.4.1 Folic Acid Many patients receiving phenytoin on a long term basis require folic acid supplementation (see section 2.3.13). The effect of this agent on phenytoin disposition has been carefully reviewed by Rivey et al. (1984) and MacCosbe and Toomey (1983). There have been numerous reports indicating that oral or intramuscular folic acid supplementation can decrease serum phenytoin concentrations (Baylis et al. 1971; Berg et al. 1983a,b, 1987; Furlanut et al. 1978; Glazko 1975; Jensen & Olesen 1970; Kutt et al. 1966; MacCosbe & Toomey 1983; Makki et al. 1980; Olesen & Jensen 1970) leading, in some of these cases, to a loss of seizure control (Baylis et al. 1971; Berg et al. 1983b; MacCosbe & Toomey 1983). In a 6-month study involving adult epileptic patients taking phenytoin with or without phenobarbital, Mattson et al. (1973) found that maintenance folic acid therapy significantly decreased the concentrations of phenytoin in CSF. Both total and unbound plasma phenytoin concentrations appear to be decreased by folic acid coadministration (Berg et al. 1983a; Furlanut et al. 1978). Therefore, the most likely mechanism of the interaction, as suggested earlier by Viukari (1968), is stimulation of phenytoin metabolism. Berg et al. (1983a) found that in 3 of 4 epileptic patients showing a reduced serum phenytoin concentration in response to folate supplementation, the urinary ratio of p-HPPH/phenytoin was increased. The magnitude of the increases closely reflected the changes in serum phenytoin concentration. In contrast, Furlanut et al. (1978) found no effect of folic acid on the urinary excretion of p-HPPH in healthy subjects receiving single oral doses of the parent compound. Olesen and Jensen (1970) had previously found that, following folic acid supplementation in epileptic patients, urinary p-HPPH excretion was decreased despite a reduction in serum phenytoin concentrations. However, attention should be drawn to the fact that since parahydroxylation is the major metabolic route for phenytoin, changes in serum phenytoin concentra-


tions may not transpose to large alterations in the urinary excretion of p- HPPH (see section 1.1). It should also be noted that an influence offolic acid on phenytoin bioavailability has not been totally excluded (Yuen 1984). In a recent study, Berg et al. (1987) examined retrospectively the effect of oral folic acid on serum phenytoin concentrations in 7 epileptic patients receiving long term oral therapy with phenytoin. Following the addition of folic acid 1mg daily, phenytoin concentrations were found to decrease by between 6 and 48%. The Km of phenytoin was significantly lower during folic acid therapy, while Vmax remained unchanged. The data supported the suggestion by Baylis et al. (1971) that administration offolic acid restored the impaired metabolism of phenytoin which resulted from the phenytoininduced depletion of body folate stores. If such a hypothesis is correct, caution is needed when interpreting data from studies investigating the interaction in healthy volunteers, whose folate levels are in the normal range. This may explain, in part, why Andreasen et al. (1971) found that in individuals without folate deficiency, the half-life of an intravenous dose of phenytoin was not influenced by long term folic acid therapy. Overall, the effect of folic acid on the disposition of phenytoin is variable. Certainly, in some individuals, folic acid supplementation can cause considerable decreases in serum phenytoin concentrations, leading to a loss of seizure control. For this reason, caution should be exercised when folic acid supplementation is commenced in patients receiving phenytoin, especially if body folate stores are greatly depleted.

1.4.2 Alcohol Very little is known of the effects of short or long term alcohol intake on the pharmacokinetics of phenytoin. The interaction was first investigated by Kater et al. (1969), who examined the disposition of phenytoin in 15 hospitalised alcoholic patients with a recent history of long term alcohol intake, and in 76 control subjects. Following the administration of oral phenytoin for 3 days, the mean (± SO) apparent elimination half-life of

45

phenytoin in plasma in the alcoholic group (16.3 ± 6.8 hours) was significantly less than in the control group (23.5 ± II hours). Unfortunately, the relative plasma phenytoin concentrations attained in the 2 groups were not reported, which makes interpretation of the data difficult. In addition, the alcohol intake of the control group was unknown. More recently, Sandor et al. (1981) conducted a longitudinal study in II hospitalised, male, chronic alcoholics. The patients were given oral phenytoin (except on 3 occasions when the dose was administered intravenously) on a continual basis for 20 days. During the first 6 days, patients were given oral alcohol, mixed with orange juice, every 2 hours while awake, to maintain a blood alcohol concentration of between 500 and 800 mg/L. The total and unbound clearance of phenytoin, determined from an intravenous dose, was determined on day 6 (alcohol ingestion phase), day 14 (early withdrawal phase), and day 20 (late withdrawal phase). Assessments of oral bioavailability were also made towards the end of each phase by comparing the phenytoin AUC over a dosing interval following oral dosing, to that after intravenous dosing. There was no change in the apparent bioavailability of phenytoin between the 3 phases. Compared with the alcohol ingestion phase, both the total and unbound clearance of phenytoin increased significantly during early and late alcohol withdrawal. However, these significant differences were only achieved when the data from 2 of the subjects, whose clearances were lower than in the other 9 subjects, were excluded from the statistical analysis. During alcohol ingestion, the mean clearance of phenytoin in the 9 subjects was similar to that reported in non-drinking healthy individuals. The authors postulated that alcohol induces the metabolism of phenytoin, but this is only unmasked when alcohol is withdrawn, suggesting therefore that it exhibits a dual effect on the metabolism of phenytoin. Hence, the effect of alcohol on phenytoin disposition is complex, and further controlled studies are needed to fully elucidate its mechanism and clinical significance. It should be noted that De Leacy et al. (1979) found that moderate alcohol intake did not influence the relationship between

46


phenytoin dose and steady-state concentration in 210 epileptic patients. 1.4.3 Dexamethasone In 1985, Wong et al. performed a retrospective study, and found that serum phenytoin concentrations were lowered when dexamethasone was administered concurrently. In 6 patients receiving an identical dose of phenytoin either with or without concurrent dexamethasone, the mean serum phenytoin concentration was, on average, halved during dexamethasone therapy. It was concluded that there was a significant interaction between the 2 agents, with possible mechanisms including enhancement of phenytoin clearance. Lawson et al. (1980) conducted a cross-sectional study in patients receiving phenytoin for prophylaxis of post-traumatic epilepsy. Patients with severe head injuries were also given dexamethasone to control cerebral oedema. Serum phenytoin concentrations 24 hours after a combined intravenous/intramuscular loading dose were higher in the patients receiving dexamethasone, and it was suggested that this drug decreased phenytoin hydroxylation. However, the effect of possible erratic absorption from the intramuscular sites on a single-point phenytoin determination in patients not matched for severity of trauma is not known. 1.4.4 Anticonvulsant Agents Phenobarbital There is some evidence to suggest that phenobarbital increases the clearance of phenytoin, but this finding is not universal. In the studies of Cucinell et al. (1965) and Kutt et al. (1969), a group of patients taking phenytoin was compared with another group given similar doses of phenytoin plus concomitant phenobarbital (45 to 120 mg/day). Both studies found a lower plasma concentration of phenytoin in the group receiving the combination, although in the study of Kutt et al. (1969) this difference was only minor and probably clinically insignificant. It is apparent that the effect of phenobarbital is variable. Kutt and colIeagues found that, either when phenobarbital was introduced or

when the dose was increased in patients stabilised on phenytoin, there was (after a period of 3 to 6 weeks) a decrease, no change or a slight increase in the plasma phenytoin concentration. Kristensen et al. (1969) determined the effect of pre-treatment with phenobarbital 1.5 to 2.5 mg/kg/ day for approximately 10 days on the apparent halflife of a single intravenous dose of phenytoin 100mg. There was a significant reduction for the 12 subjects studied. In 4 subjects the reduction was greater than 30%; while for the remainder there appeared to be only a minor change, suggesting a variable induction of phenytoin metabolism. A similar conclusion may be reached from the data of Morselli et al. (1971) obtained from a study in adults receiving phenytoin 2.4 to 6.8 mg/kg/day with or without concurrent phenobarbital 1.1 to 2.5 mg/kg/day. Addition of phenobarbital resulted in a significant decrease in plasma phenytoin concentrations, while its withdrawal from patients stabilised on both drugs led to an increase. These changes, however, were not universal, with phenobarbital causing no change in some patients. In those with a marked decrease in plasma phenytoin concentrations while taking phenobarbital, there was also a marked reduction in the urinary excretion ratio of phenytoin to phenytoin plus p-HPPH. Recently, Browne et al. (1988a) examined the effect of phenobarbital 90 to 120 mg/day on the clearance, half-life and volume of distribution of phenytoin, given intravenously as a stable isotope (150 mg), to 6 patients stabilised on oral phenytoin 250 to 450 mg/day. After 4 and 12 weeks of phenobarbital treatment, there was no significant effect on each of the measured parameters of phenytoin. In addition, there was no change in the average plasma concentration of phenytoin and the percentage of the dose excreted in urine as p-HPPH, the dihydrodiol metabolite or parent compound. In summary, studies of the effect of phenobarbital on the kinetics of phenytoin have produced conflicting results. The most compelling evidence suggests that, in most instances, phenobarbital will produce clinically insignificant changes in the plasma concentration of phenytoin.

47


Clonazepam In 1985, Saavedra et al. presented a case report showing that when clonazepam was added to phenytoin therapy, a reduction in the plasma concentration of phenytoin occurred coincident with an increase in seizure frequency. When the clonazepam was withdrawn, the plasma phenytoin concentration returned to previous levels. However, in 12 epileptic patients stabilised on phenytoin with or without other antiepileptic drugs, Johannessen et al. (1977) found that the addition of clonazepam produced no significant change in the plasma concentration of total phenytoin; similar findings were reported by Nanda et al. (1977). Huang et al. (1974) found an increase in plasma phenytoin concentrations in patients started on clonazepam therapy. However, these workers did acknowledge that the plasma concentrations of phenytoin, which were subtherapeutic before clonazepam was initiated, may have increased because of greater compliance during the study.

1.4.5 Diazepam and Chlordiazepoxide An early study by Vajda et al. (1971) reported an increase in plasma phenytoin concentrations in patients given either diazepam or chlordiazepoxide concurrently with phenytoin. In contrast, Richens and Houghton (1975), using a more specific analytical technique for the determination of phenytoin in plasma, found that diazepam significantly decreased the plasma concentration and half-life of phenytoin while chlordiazepoxide was without effect. However, in most cases, the decrease in the steady-state plasma phenytoin concentration caused by diazepam is unlikely to be clinically important. 1.4.6 Dichloralphenazone Riddell et al. (1980) administered a single intravenous dose of phenytoin to 5 healthy males before and after 13 days' therapy with dichloralphenazone. The mean apparent elimination half-life of phenytoin was reduced significantly from 20.8 ± 1.4 hours to 14.4 ± 2.5 hours and the mean clearance increased from 2.0 ± 0.7 L/h to 4.4 ± 1.2 L/h. Because phenazone (antipyrine) half-life was also decreased, the authors postulated that di-

chloral phenazone enhanced the metabolic clearance of phenytoin.

1.4.7 Rifampicin Rifampicin has been shown by Kay et al. (1985) to decrease the half-life of phenytoin in epileptic patients. The magnitude of the decrease was similar in patients taking rifampicin alone or in combination with isoniazid and ethambutol, and was unaffected by acetylator status. The metabolic induction was maintained during a 3-month period of treatment with all 3 antituberculous agents. 1.4.8 Nitrofurantoin In 1978, Heipertz and Pilz reported that when nitrofurantoin therapy was initiated in a patient with a seizure disorder which was controlled by phenytoin, serum phenytoin concentrations fell and there was a loss of seizure control. Upon withdrawal of nitrofurantoin, phenytoin concentrations returned to the previous level. The authors proposed that nitrofurantoin either inhibited the absorption, or increased the rate of metabolism, of phenytoin. 1.5 Drugs Which May Reduce the Metabolism

1.5.1 Anticonvulsant Agents Valproic Acid A number of early reports indicated that the in'troduction ofvalproic acid to patients stabilised on phenytoin resulted in a decrease in the plasma concentration of total phenytoin (Sansom et al. 1980; Wilder et al. 1978). This finding is in keeping with the decrease in the plasma protein binding of phenytoin induced by valproic acid (Bruni et al. 1980; Dahlqvist et al. 1979; Monks & Richens 1980; Monks et al. 1978; Patsalos & Lascelles 1977; Perucca et al. 1980; Sansom et al. 1980). However, when valproic acid was started in patients stabilised on phenytoin, the relative magnitude of the observed increase in the total phenytoin clearance was less than the increase in the plasma unbound fraction, suggestive of a concomitant reduction in

48

intrinsic clearance (Bruni et al. 1980; Perucca et al. 1980). That intrinsic clearance decreased was also supported by an observed increase in the urinary phenytoin: p-HPPH ratio (Perucca et al. 1980). When valproic acid was started in patients on a maintenance phenytoin dose regimen, the plasma concentration of total phenytoin decreased in the short term while the concentration of unbound drug in plasma remained essentially unchanged, but over a longer period the unbound concentration increased and the total plasma concentration returned to initial levels or higher (Bruni et al. 1980; Perucca 1984). The proposed mechanism of interaction, that an increase in the unbound fraction of phenytoin is accompanied by a reduction in intrinsic metabolic clearance, is supported by the study of Koch et al. (1981) performed in the Rhesus monkey. There are considerable clinical implications for the dual effect of valproic acid on phenytoin pharmacokinetics. After the addition of valproic acid to the drug regimen of a patient previously stabilised on phenytoin, the appearance of signs of phenytoin toxicity may not be apparent for a number of weeks until steady-state plasma concentrations of unbound phenytoin are re-established at a higher level. Due to the altered plasma binding, therapeutic drug monitoring based on total concentrations of phenytoin may be misleading. Carbamazepine Examination of the effect of carbamazepine on the pharmacokinetics of phenytoin has produced conflicting findings. Hansen et al. (1971 b) detected a reduction in half-life and, in some patients, a decrease in the steady-state plasma concentration of phenytoin when carbamazepine therapy was initiated. In contrast, Zielinski and Haidukewych (1987) and Zielinski et al. (1985) found that carbamazepine caused an increase in the steady-state total plasma concentration of phenytoin in some patients, and no change in the remainder. Gratz et al. (1982), in a study of 3 patients who had achieved steady-state plasma phenytoin concentrations, detected an increase in these after dosing with carbamazepine. Browne et al. (1988b), in another study

Clin. Pharmacokinc/. 18 (l) 1990

of 6 patients stabilised on a constant dose of phenytoin, gave an intravenous dose of stable isotope-labelled phenytoin before and after 12 weeks' treatment with carbamazepine 600 to 800mg daily. There was a decrease in the clearance of phenytoin to p-HPPH and the dihydrodiol metabolite, and an increased phenytoin half-life. Moreover, the average plasma phenytoin concentration increased after carbamazepine treatment commenced. While examination of the effect of carbamazepine on the pharmacokinetics of phenytoin has produced conflicting findings, most of the evidence suggests that coadministration of carbamazepine will lead to an increase in the plasma phenytoin concentration. The effect of phenytoin on the pharmacokinetics of carbamazepine is discussed in section 2.3.1. Sulthiame Pre-treatment with sulthiame has been shown to increase the half-life of phenytoin given as a single intravenous dose (Hansen et al. 1968). In both the control and treatment phases of this study, similar initial plasma phenytoin concentrations were achieved, an important consideration in view of the non-linear pharmacokinetics of the drug under review. Hansen et al. (1968) examined the effect of treatment with sulthiame 200 to 800 mg/day for at least 20 days on the plasma concentration of phenytoin in 4 patients who had received phenytoin 3.6 to 6.4 mg/kg/day for at least 12 months. In all patients, phenytoin concentrations increased after initiation of sulthiame. Similarly, Olesen and Jensen (1969) performed a study in 7 patients who had been stabilised on phenytoin (and phenobarbital) over a period of several years. The addition of sulthiame triggered a rapid rise in plasma phenytoin concentrations in all patients, with the higher values being maintained until sulthiame was withdrawn. In a later study, Houghton and Richens (l974a) found thatsuIthiame 400 to 600 mg/ day significantly increased the steady-state plasma concentration of phenytoin. In a cross-sectional study, Houghton and Richens (I 974b) compared 2 groups of patients receiving similar doses ofphenytoin in addition to other antiepileptic drugs. The


incidence of symptoms of toxicity attributable to phenytoin, and the mean plasma phenytoin concentration, were both significantly greater for the group receiving sulthiame. In summary, it would appear that sulthiame reduces the metabolic clearance of phenytoin, giving rise to increased plasma concentrations and a greater likelihood of phenytoin toxicity. While sulthiame is not commonly used for the treatment of epilepsy, caution should be exercised if it is used in patients receiving phenytoin. Progabide While conducting studies to determine the efficacy of progabide as an anti epileptic drug, a number of groups found that, in some patients previously stabilised on phenytoin, the plasma concentration of that agent increased during the trial period with progabide (Crawford & Chadwick 1986; Dam et al. 1983; Schmidt & Utech 1986; Weber et al. 1985). This finding, however, has not been universal (Loiseau et al. 1983). In healthy volunteers given a single dose of phenytoin, Bianchetti et al. (1987) found that pre-treatment with progabide for r5 days did not significantly affect the AUC of phenytoin. However, the small but significant decrease in the weight-adjusted apparent oral clearance, and increase in the half-life and percentage of the phenytoin dose excreted unchanged in urine, suggested a modest reduction in the metabolic clearance of phenytoin by progabide.

1.5.2 Oral Anticoagulants Dicoumarol increased the steady-state serum concentration and the elimination half-life of phenytoin in healthy volunteers (Hansen et al. 1966) and also raised serum phenytoin concentration in patients previously stabilised on phenytoin (Skovsted et al. 1976). In addition, these latter authors found that dicoumarol caused a marked increase in the half-life of phenytoin given as a single intravenous dose to volunteers. While phenprocoumon caused a moderate increase in both the half-life and steady-state serum concentrations of phenytoin, warfarin and phenindione were without effect (Skovsted et al. 1976).

49

1.5.3 Antituberculous Drugs Murray (1962) and Miller et al. (1979) reported an increased incidence of phenytoin toxicity when isoniazid was given concurrently. A similar finding was reported by Kutt et al. (1970) in patients receiving phenytoin and isoniazid, together with either cycloserine or aminosalicylic acid (PAS). Slow acetylators of isoniazid appeared more likely to experience toxicity coincident with an elevated plasma concentration of phenytoin (Brennan et al. 1970; Kutt et al. 1970). The influence of PAS and cycloserine on the pharmacokinetics of phenytoin is unclear: in 2 patients receiving long term treatment with phenytoin it was found that isoniazid and PAS alone had no effect on plasma phenytoin concentrations, but a marked increase was observed when both antituberculous drugs were coadministered (Kutt et al. 1964, 1970). Thus, coadministration of PAS with isoniazid may further increase the likelihood of phenytoin toxicity. Interestingly, studies in vitro suggest that PAS may potentiate the inhibitory effect of isoniazid on phenytoin metabolism (Kutt et al. 1970). 1.5.4 Amiodarone A number of case reports have appeared in the literature describing an increase in the incidence of phenytoin side effects, and in plasma phenytoin concentration, following the introduction of amiodarone to patients stabilised on phenytoin (Gore et al. 1984; McGovern et al. 1984; Shackleford & Watson 1987). For the 1 patient studied by Gore et al. (1984), the increase in both total and unbound plasma phenytoin concentrations suggested a decreased intrinsic clearance. 1.5.5 Allopurinol Yokochi et al. (1982) reported a case of toxicity attributed to phenytoin following the introduction of allopurinol therapy to a child with Lesch-Nyhan syndrome who was already receiving phenytoin. While the doses of phenytoin and other anticonvulsants remained constant, there was a significant increase in the serum concentration of phenytoin. The relevance of the Lesch-Nyhan syndrome to the suspected interaction remains unknown.


50

1.5.6 H2-Antagonists Cimetidine A number of reports have indicated that cimetidine increases plasma phenytoin concentrations when added to the dosage regimen of patients previously stabilised on oral phenytoin (Algozzine et al. 1981; Hetzel et al. 1981; Levine et al. 1985; Neuvonen et al. 1981; Phillips & Hansky 1984; Salem et al. 1983; Watts et al. 1983). That this interaction is clinically important is highlighted by reports of cimetidine-precipitated phenytoin intoxication (Hetzel et al. 1981; Phillips & Hansky 1984; Salem et al. 1983; Watts et al. 1983). Therefore, concomitant administration of cimetidine and phenytoin is potentially hazardous, especially when cimetidine treatment is started or stopped in patients receiving phenytoin. Although the mechanism of the interaction is not fully understood, studies investigating the effect of oral cimetidine on the disposition of a single intravenous dose of phenytoin (Bartle et al. 1983; Frigo et al. 1983) suggest that inhibition of phenytoin clearance by cimetidine is an important factor. Bartle et al. (1983) found that, in 8 healthy male volunteers, long term administration of a range of cimetidine doses (400, 1200 and 2400mg daily) reduced the clearance of a single intravenous dose of phenytoin 250mg by 10, 11 and 21 %, respectively. Frigo et al. (1983) have reported that long term administration of cimetidine 1200mg daily produced a small but statistically significant decrease (mean 12%) in the clearance of phenytoin, administered as a single intravenous dose of 250mg to healthy volunteers. In this same study, cimetidine was found to have no effect on the volume of distribution of phenytoin. These small effects of cimetidine on phenytoin clearance are in contrast to the significant increases in steady-state plasma phenytoin concentrations caused by cimetidine in patients receiving oral phenytoin on a long term basis. The most likely cause of this difference is that the plasma phenytoin concentrations attained during maintenance oral dosing are generally greater than those achieved following a single oral dose. As a result of the concentration-dependent meta-

bolic clearance of phenytoin, the effects of a metabolic inhibitor, such as cimetidine, would be expected to be greater at higher plasma phenytoin concentrations. Alternatively, or additionally, cimetidine may alter the bioavailability of phenytoin. It is interesting to note that Hetzel et al. (1981) found that when cimetidine 1200mg daily was added to the dosage regimen of 4 epileptic patients stabilised on oral phenytoin 260 to 400mg daily, the increase in plasma phenytoin concentration (13 to 33%) was accompanied by an increase in the 24hour urinary recovery of p-HPPH and phenytoin. The magnitude of the mean increases (122.5mg to 201.5mg and 6.8mg to 12.0mg for p-HPPH and phenytoin, respectively) led the authors to postulate that either cimetidine affects an alternative pathway of phenytoin metabolism, or it increases phenytoin bioavailability. The 3 studies which have examined the effect of cimetidine on the pharmacokinetics of a single oral dose of phenytoin (Hsieh et al. 1986; Iteogu et al. 1983; Sambol et al. 1989) have found increases in AVe of variable magnitude. Iteogu et al. (1983) reported that in 5 healthy subjects, AVe increased by between 41 and 104%, during long term cimetidine treatment. In contrast, Hsieh et al. (1986) reported that cimetidine had only marginal effects on phenytoin AVe. Very recently, Sambol et al. (1989) found that long term dosing with cimetidine significantly reduced the apparent oral clearance of a single dose of phenytoin by an average of 16%. Ranitidine and Famotidine Ranitidine was found to have no effect on steady-state plasma phenytoin concentrations when administered for 2 weeks to 4 epileptic patients receiving long term oral phenytoin (Watts et al. 1983). The lack of effect prompted the authors to suggest that ranitidine may be a more suitable H2-antagonist than cimetidine for use in epileptic patients receiving phenytoin. However, Bramhall and Levine (1988) have reported a case study in which plasma phenytoin concentration increased by approximately 50% after commencement of ranitidine therapy. Following withdrawal of the H2-ant-


agonist, the phenytoin concentration returned to the original value. Recently, Sambol et al. (1989) found famotidine to have no effect on the pharmacokinetics of a single oral dose of phenytoin in healthy subjects. 1.5.7 Omeprazole Gugler and Jensen (1985) investigated the effect of a 7-day course of oral omeprazole 40mg daily on the disposition of a single intravenous dose of phenytoin 250mg. Omeprazole caused a statistically significant decrease in the apparent phenytoin clearance (mean ± SEM) from 25.1 ± 2.0 to 21.4 ± 1.8 ml/h/kg. It increased the half-life of phenytoin, but had no effect on its plasma protein binding and volume of distribution; nor did it influence the urinary excretion of p-HPPH. The authors concluded that omeprazole inhibited the oxidative metabolism of phenytoin. In a double-blind crossover study, Prichard et al. (1987) investigated the effect of omeprazole 40mg daily on the disposition of a single oral dose of phenytoin in 10 healthy males. Omeprazole produced a significant increase in the AUC(O-72h) of phenytoin from (mean ± SEM) 121.6 ± 14.0 to 151.4 ± 13.6 mg/L·h. The effect of omeprazole on phenytoin administered as maintenance therapy has yet to be determined. Certainly, at higher plasma phenytoin concentrations, the influence of omeprazole may be intensified and hence caution should be excercised if the 2 agents are to be coadministered. 1.5.8 Non-Steroidal Anti-Inflammatory Agents Azapropazone (Apazone) Two case reports of azapropazone-precipitated phenytoin toxicity (Geaney et al. 1982; Roberts et al. 1981) prompted Geaney et al. (1983) to investigate the interaction between the 2 drugs. Five healthy volunteers were stabilised on long term oral phenytoin. When azapropazone 600mg twice daily was administered, the plasma concentrations of phenytoin increased, necessitating a reduction in phenytoin dosage to maintain the previous levels. In all volunteers there was an increase in Km while Vmax remained unaltered. The fall in the apparent

51

phenytoin clearance coincided with a decrease in the plasma and urinary p-HPPH/phenytoin ratio. The authors concluded that azapropazone inhibits the para-hydroxylation of phenytoin in a competitive manner. The decreased intrinsic clearance could lead to an increase in the unbound concentration of phenytoin in plasma, predisposing to toxicity. Azapropazone was also found to cause a small degree of displacement of phenytoin from plasma binding sites (Geaney et al. 1983), a factor which should be considered when interpreting total plasma phenytoin concentrations for therapeutic drug monitoring. In view of these findings, it was advised that azapropazone is better avoided in patients receiving phenytoin (Geaney et al. 1983). Phenylbutazone Andreasen et al. (1973) found that 5 days' pretreatment with phenylbutazone increased the halflife of phenytoin, administered as a single intravenous dose, in all of 28 healthy volunteers. In 6 epileptic patients taking phenytoin, either alone or in combination with carbamazepine, administration of phenylbutazone for 2 weeks was found to cause an increase in total and unbound serum phenytoin concentrations (Neuvonen et al. 1979). One patient experienced phenytoin toxicity which was reversed by phenylbutazone withdrawal. It is of interest that in the first days of phenylbutazone coadministration there was a significant lowering of total phenytoin concentration. It was concluded that because phenylbutazone inhibited both the metabolism and plasma binding of phenytoin, the increase in concentration of total phenytoin would underestimate the rise in that of unbound drug. Ibuprofen In 1982, Sandyk described a case of phenytoin toxicity which may have been due to the coadministration of ibuprofen. Subsequently, Bachmann et al. (1986) examined the effect oflong term dosing with ibuprofen on the pharmacokinetics of phenytoin, administered as a single oral dose, in 10 healthy volunteers. Although ibuprofen decreased the plasma protein binding of phenytoin slightly, it was found to have no effect on the ap-

52

C1in. Pharmacokine/. 18 (/J 1990

parent oral clearance (for total and unbound drug) or the apparent volume of distribution of the anticonvulsant. 1.5.9 Disulfiram The first study to examine the effect of disulfiram on phenytoin disposition was conducted by Olesen (1966). When 4 male patients, stabilised on oral phenytoin, were given daily doses of disulfiram, the serum phenytoin concentrations increased by between 100 and 500%, and 2 patients experienced symptoms of phenytoin toxicity. Upon withdrawal of disulfiram, the concentrations decreased. In 1967, Olesen found that disulfiramprecipitated increases in serum phenytoin concentrations coincided with decreases in the rate of urinary excretion of p-HPPH, and concluded that disulfiram inhibited the metabolic transformation of phenytoin to p-HPPH. Such a mechanism is supported by the results of a study by Svendsen et al. (1976) examining the effect of disulfiram on the pharmacokinetics of phenytoin administered as a single intravenous dose to 5 volunteers. While the apparent clearance of phenytoin was markedly reduced by disulfiram, the volume of distribution was unaltered. The significance of the interaction is highlighted by literature reports of disulfiram-induced phenytoin toxicity (Brown et al. 1983; Kiorboe 1966; Taylor et al. 1981). Clearly, when disulfiram is added to, or removed from, the dosage regimen of patients receiving phenytoin, dosage adjustment may be required. 1.5.10 Anti-Infective Agents

Metronidazole Inconclusive evidence of an effect of metronidazole on the disposition of phenytoin is provided by an anecdotal report (Picard 1983) noting that in patients receiving maintenance phenytoin, toxic plasma concentrations of that drug were found after metronidazole was added to the daily drug regimen. In 1985, Jensen and Gugler found that there was no change in phenytoin pharmacokinetics in

healthy subjects who had ingested a single oral dose of phenytoin before and after 6 days' therapy with metronidazole. Recently, Blyden et al. (1988) administered a single intravenous dose of phenytoin alone and during metronidazole treatment to 7 healthy volunteers, and found that metronidazole decreased the clearance of phenytoin by an average of 15%. An increase in the elimination half-life and volume of distribution of phenytoin was also detected, although plasma protein binding of the drug was not altered. Sulphonamides In 1975, Lumholtz and co-workers found that phenytoin half-life was increased by concurrent sulfamethizole administration. Subsequently, Hansen et al. (1979) studied the effect of a range of sulphonamides on phenytoin disposition. Sulfaphenazole, sulfadiazine, sulfamethizole, sulfamethoxazole and cotrimoxazole (sulfamethoxazole plus trimethoprim) all increased the half-life of an intravenous dose of radiolabelled phenytoin, prompting the authors to suggest that these sulphonamides inhibit the hepatic metabolism of the latter. Sulfamethoxypyridazine, sulfadimethoxine and sulfamethoxydiazine had no effect on phenytoin half-life. Additionally, in epileptic patients receiving long term oral phenytoin, coadministration of sulfaphenazole, sulfadiazine, sulfamethizole or cotrimoxazole increased serum phenytoin concentrations, although relatively small subject numbers were investigated. Case reports of phenytoin intoxication induced by cotrimoxazole (Gillman & Sandyk 1985; Wilcox 1981) emphasise the caution required when this antibacterial combination is administered to patients receiving phenytoin. It should also be noted that trimethoprim alone has been found to increase the elimination half-life of an intravenously administered dose of phenytoin (Hansen et al. 1979). Chloramphenicol The only study investigating the influence of chloramphenicol on phenytoin pharmacokinetics in humans was conducted by Christensen and Skovsted (1969). They found that this agent, when


added to the daily drug regimen of 2 patients stabilised on phenytoin, caused an increase in serum phenytoin concentrations. In 3 other patients, the elimination half-life of a radiolabelled intravenous dose of phenytoin was increased considerably. These workers postulated that chloramphenicol inhibited the metabolic transformation of phenytoin, and stressed that the antibiotic should be used with caution in patients receiving phenytoin. Subsequent to this study, a number of case reports have appeared in the literature indicating that administration of chloramphenicol to patients receiving phenytoin can result in phenytoin toxicity (Ballek et al. 1973; Cosh et al. 1987; Koup et al. 1978; Rose et al. 1977). 1.6 Influence of Other Drugs on Plasma Phenytoin Concentration There are a number of other drugs which have been reported to increase or decrease plasma phenytoin concentration (table II). Many of these reports are anecdotal, and the possible influence of other factors, such as compliance and dietary changes, cannot be excluded. Drugs for which there is evidence of no effect on phenytoin disposition are listed in table III. It should be noted that in a number of these studies, phenytoin was administered as a single dose. Failure to detect an interaction under such conditions does not preclude the possibility of an effect when phenytoin is administered in the long term, as discussed in section 1.1.

2. Interactions Affecting the Pharmacokinetics of Other Drugs 2.1 Gastrointestinal Absorption When compared with drug-free healthy subjects, epileptic patients receiving phenytoin and phenobarbital have been reported to exhibit a significantly smaller diuretic response to furosemide (frusemide) administered either orally or intravenously (Ahmad 1974). It was suggested that the sensitivity of the renal tubule to the diuretic action of furosemide was reduced by anticonvulsant therapy.

53

Table II. Drugs for which there is some evidence for an effect on plasma phenytoin concentration Drug

Reference

Increased plasma phenytoin concentration Dextropropoxyphene Hansen et al. (1980) (propoxyphene) Felbamate Fuerst et al. (1988) Graves et al. (1989) Wilenksy et al. (1985) Fluconazole Mitchell & Holland (1989) Imipramine Perucca & Richens (1977) Methsuximide Rambeck (1979) Methylphenidate Garrettson et al. (1969) Miconazole/flucytosine Rolan et al. (1983) Nafimidone Treiman & Ben-Menachem (1987) Nifedipine Ahmad (1984) Oral contraceptives DeLeacy et al. (1979) Phenyramidol Solomon & Schrogie (1967) Pindolol Greendyke & Gulya (1988) Ticrynafen Ahmad (1981) Trazadone Dorn (1986) Decreased plasma phenytoin Diazoxide Influenza vaccination Oxacillin

concentration Roe et al. (1975) Sawchuk et al. (1979) Fincham et al. (1976)

Furthermore, after oral administration of furosemide the peak diuretic response was considerably delayed in the epileptic patients, and this was attributed to delayed oral absorption of furosemide, perhaps the result of a phenytoin-induced decrease in the spontaneous activity of gastrointestinal smooth muscle (Ahmad 1974). Subsequently, Fine et al. (1977) reported that phenytoin caused malabsorption of furosemide from the gastrointestinal tract. They performed an intravenous and oral crossover pharmacokinetic study in 5 healthy volunteers, both before and after a 10-day course of phenytoin 300mg daily. The serum clearance of furosemide, determined after intravenous dosing, was not significantly different between control and phenytoin-treatment phases, a finding in keeping with the relatively minor role of hepatic biotransformation in the clearance of furosemide (Fine et al. 1977). However, the mean (± SEM) absolute oral bioavailability of furosemide was 80.9 ± 6.5

54

Clin. Pharmacokinel. 18 (1) 1990

Table III. Drugs for which there is evidence of no effect on the disposition of phenytoin Drug

Reference

Olesen (1967) Richens & Houghton (1975) Callaghan et al. (1983) Callaghan et al. (1983) Rameis (1985) Bachmann et al. (1984); Milne et al. (1988) Ethosuximide Richens & Houghton (1975) Famotidine Sambol et al. (1989) Kapetanovic et al. (1988) Flunarizine Lithium carbonate Perrild et al. (1983) Paracetamol (acetaminophen) Neuvonen et al. (1979) Phenindione Skovsted et al. (1976) Ranitidine Watts et al. (1983) Gotz et al. (1984); Thioridazine Sands et al. (1987) Tolfenamic acid Neuvonen et al. (1979) Warfarin Skovsted et al. (1976)

Calcium carbimide Chlordiazepoxide Cholestyramine Colestipol Digoxin Erythromycin

and 39.0 ± 5.4% in the control and phenytoin phases, respectively, and there was a corresponding and significant reduction in the maximum serum concentration of furosemide during phenytoin treatment. Unfortunately, there was no mention of the time of occurrence of the peak serum furosemide concentration after oral dosing, or of the diuretic response to oral or intravenous furosemide in the control and phenytoin phases (Fine et al. 1977). As a result, it is not possible to fully compare the study of Fine and co-workers with that of Ahmad (1974). In contrast with the findings on furosemide, Keller et al. (1981) have reported that pre-treatment with phenytoin did not alter the disposition of hydrochlorothiazide administered orally. There is some evidence that phenytoin treatment may decrease the gastrointestinal absorption of thyroxine and folic acid (see section 2.3.13). In addition, Rowland and Gupta (1987) suggested that treatment with phenytoin leads to decreased gastrointestinal absorption of cyclosporin. However, as discussed in section 2.3.8, other mechanisms may be involved in the phenytoin-cyclosporin interaction. Coadministration of phenytoin has been re-

ported to increase both the rate and extent of absorption of isoxicam, administered orally as a single dose (Caille et al. 1987). The conclusion concerning the increased extent of absorption was based on a small (19%), but significant, increase in the AUC of isoxicam during the phenytoin phase. However, in the absence of data from intravenous administration of isoxicam, other mechanisms, such as an effect of phenytoin on isoxicam clearance, cannot be fully discounted. 2.2 Plasma Protein Binding At therapeutic plasma concentrations, the molar concentration of phenytoin is very low in comparison with that of its major binding protein, albumin. It is not surprising, therefore, that phenytoin, at clinically relevant concentrations, causes a negligible increase in the unbound fraction of valproic acid (Cramer & Mattson 1979; Monks et al. 1978), clonazepam (Khoo et al. 1980), dicoumarol (Hansen et al. 1971a) and tolbutamide (Fernandez et al. 1985). Treatment with phenytoin causes an increase in the plasma concentration of a I-acid glycoprotein (Routledge et al. 1981; Scott et al. 1983; Tiula & Neuvonen 1982) and an associated decrease in the unbound fraction of lidocaine (lignocaine) [Routledge et al. 1981]. This would need to be considered when interpreting total plasma lidocaine concentrations in the clinical setting (see section 2.3.5). Borga et al. (1969) found that phenytoin increased the plasma unbound fraction of a number of tricyclic antidepressant agents in vitro. However, this may not reflect the in vivo situation in a patient on long term phenytoin therapy, because of the phenytoin-induced increase in the plasma concentration of ai-acid glycoprotein which is an important binding protein for basic drugs, including tricyclic antidepressants (Piafsky 1980). The plasma concentration of sex hormone binding globulin is increased in patients receiving phenytoin (Odlind & Olsson 1986; Victor et al. 1977) and this may alter the plasma binding of oral contraceptive steroids (see section 2.3.6). Although early studies suggested that phenytoin displaced


Table IV. Summary of evidence for microsomal enzyme induction by phenytoin Indicator of enzyme induction by phenytoin

Reference

Increased urinary excretion of 6!l-hydroxycortisol

Choi et al. (1971) Crowley et al. (1988) Frey & Frey (1983) Haque et al. (1972) Roots et al. (1979) Werk et al. (1964, 1971)

Increased urinary excretion of O-glucaric acid

Cunningham & Evans (1974) Perucca et al. (1979, 1984)

Increased concentration of hepatic cytochrome P450

Pirttiaho et al. (1978, 1982) Sotaniemi et al. (1978)

Increased clearance of phenazone (antipyrine)

Evans et al. (1985) Freeman et al. (1984) Marquis et al. (1982) Perucca et al. (1979, 1984) Petruch et al. (1974) Pirttiaho et al. (1978, 1982) Prescott et al. (1981) Shaw et al. (1985) Sotaniemi et al. (1978)

thyroxine from thyroxine-binding globulin (Oppenheimer & Tavernetti 1962; Wolff et a1. 1961), later investigations found that the serum unbound fraction of thyroxine was not altered by phenytoin (Faber et a1. 1985; Larsen et a1. 1970). 2.3 Drugs Whose Metabolism May Be Increased Phenytoin is a potent inducer of hepatic microsomal enzymes, as indicated by the ample evidence both from studies employing non-invasive indices (such as measurement of the urinary excretion of the endogenous compounds 6{j-hydroxycortisol and D-glucaric acid), and from those using invasive techniques (such as determination of the concentration of cytochrome P450 in liver biopsy specimens and investigation of the clearance ofthe marker drug phenazone) [table IV]. The degree of enzyme induction, as measured by changes in phenazone clearance or urinary excretion of D-glucaric acid, is correlated positively with phenytoin dose or plasma phenytoin concentration (Perucca

55

et a1. 1984; Shaw et a1. 1985). As gauged from the effect on theophylline clearance, phenytoin induction of microsomal enzymes is of similar magnitude in young and elderly adults, and additive to the enzyme induction associated with cigarette smoking (Crowley et a1. 1988). Enzyme induction is apparent in neonates after fetal exposure to phenytoin (Rating et a1. 1983). It is not only the synthesis of cytochrome P450 enzymes that may be induced by phenytoin, but also VDP-glucuronosyltransferase (see sections 2.3.2 and 2.3.11), and plasma proteins such as a I-acid glycoprotein and sex hormone binding globulin (see section 2.2). In association with enzyme induction, treatment with phenytoin appears to cause an increase in liver size and absolute liver blood flow, although liver blood flow per unit weight ofliver seems to be unaffected (Pirttiaho et a1. 1982). There is some evidence to suggest that phenytoin has a differential enzyme-inducing effect on the multiple forms of cytochrome P450. Shaw et a1. (1985) have demonstrated that the phenytoininduced increase in the total clearance of phenazone resulted from increased formation clearances of 4-hydroxyphenazone, mainly, and 3-hydroxymethylphenazone, with minimal effect on the formation clearance of norphenazone. It was concluded that phenytoin treatment led to the induction of the cytochromes P450 controlling the 4-hydroxylation and 3-methylhydroxylation of phenazone, while the form of cytochrome P450 involved in the N-demethylation pathway was unaffected (Shaw et a1. 1985). Differential induction of the various forms of cytochrome P450 is in keeping with the effect of phenytoin on the pharmacokinetics of the individual enantiomers of misonidazole (see section 2.3.9). There are many drugs for which the AVC after a single dose, or the steady-state plasma concentration after long term dosing, is decreased by phenytoin therapy; these interactions are discussed in this section. In the majority of cases, the most likely mechanism of interaction is enzyme induction by phenytoin. However, with some studies, the possibility cannot be excluded that phenytoin may have altered the absorption and/or plasma

56

protein binding of the drug concerned, although these mechanisms appear to affect a few drugs only (see sections 2.1 and 2.2). Furthermore, some of the drugs included in this section contain a chiral centre (e.g. misonidazole, methadone, mexiletine, disopyramide). Unfortunately, with the one exception ofmisonidazole (see section 2.3.9), no attempt has been made by investigators to elucidate the effect of phenytoin on the pharmacokinetics of the individual enantiomers. The potential complications of employing non-stereoselective assays in pharmacokinetic studies have been discussed recently by Evans et al. (1988). 2.3.1 Anticonvulsant Agents Phenytoin interacts pharmacokinetically with a number of anticonvulsants, and in many cases the interaction is bidirectional. The effect of other anticonvulsants on the pharmacokinetics ofphenytoin is discussed in sections 1.4.4 and 1.5.1.

Carbamazepine Phenytoin appears to increase the clearance of carbamazepine, a drug which is cleared primarily by biotransformation. The plasma concentration to dose ratio of carbamazepine was lower in patients also receiving phenytoin than in patients on carbamazepine alone, a finding which has been reported both from studies within subjects (Cereghino et al. 1975) and from cross-sectional studies (Battino et al. 1980; Christiansen & Dam 1973; Johannessen & Strandjord 1975; Perucca & Richens 1980). In addition, when the phenytoin dose was reduced in patients on combination therapy, the plasma carbamazepine concentration increased (Zielinski & Haidukewych 1987). The ratio of the plasma concentration of carbamazepine to that of its epoxide metabolite appeared to decrease as plasma phenytoin concentration increased in patients treated with both anticonvulsant agents (Dam et al. 1975). These findings were attributed to an increase in the clearance of carbamazepine by phenytoin. It seems, however, that concomitant phenytoin therapy may not lead to a maximal induction of carbamazepine clearance, since the addition of phenobarbital or primidone to the


therapeutic regimen of patients stabilised on carbamazepine and phenytoin resulted in a further decrease in plasma carbamazepine concentration (Cereghino et al. 1975; Christiansen & Dam 1973). However, the magnitude of the additive inducing effect of the barbiturates may be quite small and possibly clinically insignificant (Johannessen & Strandjord 1975; Tomson et al. 1987). Carbamazepine has antidiuretic properties which may lead to hyponatraemia and low plasma osmolality (Perucca et al. 1978a; Perucca & Richens 1980; Sordillo et al. 1978). The magnitude of the antidiuretic effect is positively correlated with plasma carbamazepine concentration (Perucca et al. 1978a; Perucca & Richens 1980). It has been observed that there is a higher incidence of hyponatraemia in patients on carbamazepine monotherapy compared with those on combination therapy with carbamazepine and phenytoin (Perucca et al. 1978a); and that phenytoin may reverse water intoxication induced by carbamazepine (Sordillo et al. 1978). Potentially, these observations may be the result of a pharmacodynamic interaction between phenytoin and carbamazepine, because the former appears to inhibit the release of antidiuretic hormone from the posterior pituitary gland, at least in certain patients (Fichman et al. 1970; Tanay et al. 1979). However, an elegant study by Perucca and Richens (1980) strongly suggests that this reversal of carbamazepine-induced water intoxication is more likely to be the result of a pharmacokinetic interaction. The diuretic response to a water load was significantly greater in patients receiving long term combination therapy with carbamazepine and phenytoin compared with carbamazepine dose-matched control subjects receiving that drug alone. It should be noted that the patients on combination therapy had significantly lower serum carbamazepine concentrations. That a pharmacodynamic interaction was not contributing greatly to the different diuretic response between the above 2 treatment groups was suggested by the finding in 2 other patient groups, which were matched for serum carbamazepine concentration, that there was no difference in diuretic response between patients on combined therapy and those


on carbamazepine alone. Moreover, in a withinsubject study in patients stabilised on carbamazepine, the short term administration of phenytoin had no significant influence on carbamazepineinduced antidiuresis. Thus, the reversal by long term phenytoin therapy of carbamazepine-induced antidiuresis is probably the result of a substantial reduction of serum carbamazepine concentration (Perucca & Richens 1980). Patients stabilised on combination therapy with carbamazepine and phenytoin should be observed carefully for the possible emergence of water intoxication or other adverse effects of carbamazepine when phenytoin is discontinued or the dosage is reduced (Perucca & Richens 1980; Zielinski & Haidukewych 1987). Primidone The anticonvulsant agent primidone is biotransformed to 2 pharmacologically active metabolites, phenylethylmalonamide (PEMA) and phenobarbital. It has been reported that the commencement of phenytoin therapy in patients stabilised on primidone coincided with a large increase in plasma phenobarbital concentration (Porro et al. 1982; Schmidt 1975) and, in patients already receiving both primidone and phenytoin, increasing the phenytoin dose resulted in a further increase in plasma phenobarbital concentration (Garrettson & Gomez 1977; Lambie et al. 1976; Reynolds et al. 1975; Schmidt 1975). In a number of cases, the increased concentration of phenobarbital was implicated as a cause of CNS toxicity. The plasma concentration of primidone appeared to be unaffected or slightly decreased by concomitant administration of phenytoin (Fincham et al. 1974; Porro et al. 1982; Reynolds et al. 1975; Schmidt 1975; Windorfer & Sauer 1977). As a result, when compared with patients receiving primidone alone, patients on combination therapy with primidone and phenytoin had considerably higher plasma concentration ratios of phenobarbital: primidone (Callaghan et al. 1977; Fincham et al. 1974; Lambie & Johnson 1981; Porro et al. 1982; Reynolds et al. 1975; Schmidt 1975) and, in ad-

57

dition, higher ratios ofPEMA : primidone (Lambie & Johnson 1981).

Various mechanisms have been proposed for the elevated plasma concentration ratio of phenobarital : primidone in patients receiving phenytoin, including the suggestion that phenytoin may increase the clearance of primidone to phenobarbital (Callaghan et al. 1977; Fincham et al. 1974), or inhibit the clearance of phenobarbital derived from primidone (Schmidt 1975) or both of the above (Lambie & Johnson 1981; Porro et al. 1982; Reynolds et al. 1975). There is some evidence to support the third proposition. That the clearance of primidone to phenobarbital is increased by phenytoin is suggested from a study by Cloyd et al. (1981), in which it was found that the apparent oral clearance of primidone was significantly higher while patients were also receiving phenytoin. In addition, in an intensive study of the plasma and urinary concentrations of primidone, phenobarbital and PEMA in an epileptic patient, the plasma concentration and urinary excretion of primidone decreased while the urinary output of PEMA and phenobarbital significantly increased after the introduction of phenytoin (Porro et al. 1982). These results were interpreted as being consistent with a phenytoininduced increase in the metabolic clearance of primidone. It is apparent, however, that phenytoin also inhibits the clearance of phenobarbital derived from primidone, and this is discussed in greater detail in section 2.4.2. Clonazepam Pre-treatment with phenytoin caused an approximate 50% increase in the apparent oral clearance of clonazepam following a single dose (Khoo et al. 1980). This was consistent with the observed decrease in plasma clonazepam concentrations after the introduction of phenytoin therapy to patients previously stabilised on clonazepam (Sjo et al. 1975). Interestingly, the plasma concentrations of 7-amino-clonazepam, a major metabolite, were also decreased; this was taken to indicate that if enzyme induction by phenytoin was the mechanism of the interaction, some other metabolic pathway was probably involved (Sjo et al. 1975). Alternatively,


58

phenytoin may increase the clearance of both clonazepam and 7-amino-clonazepam or decrease the absorption of the parent drug. Phenytoin treatment did not cause a significant change in the unbound fraction of clonazepam in plasma (Khoo et al. 1980). The clinical relevance of this interaction is difficult to predict, given the relatively poor relationship between the clinical effect and plasma concentration of clonazepam. Valproic Acid In patients receiving concomitant phenytoin, the plasma concentration to dose ratio ofvalproic acid was considerably lower than in patients on valproic acid alone (May & Rambeck 1985; Mihaly et al. 1979; Sackellares et al. 1981). Because phenytoin does not affect the gastrointestinal absorption of valproic acid (Perucca et al. 1978b), the most likely explanation for the lower plasma valproic acid concentrations is that phenytoin increases the hepatic intrinsic clearance and/or plasma unbound fraction of valproic acid. There are conflicting findings on the effect of phenytoin on the plasma protein binding of valproic acid at clinically relevant concentrations (Cramer & Mattson 1979; Monks et al. 1978). As with clonazepam, the clinical significance of the interaction is uncertain.

2.3.2 Analgesics Paracetamol (Acetaminophen) and Acetanilide Pretreatment with phenytoin has been reported to decrease the steady-state plasma concentrations of acetanilide administered orally, and its metabolite paracetamol (Cunningham & Evans 1981). This finding is consistent with phenytoin having induced the enzymes involved in both the oxidative metabolism of acetanilide to paracetamol and the conjugative metabolism (glucuronidation and/ or sulphation) of the derived paracetamol, although the possibility of an effect of phenytoin on the extent of acetanilide absorption should not be overlooked. Studies on the disposition of paracetamol in epileptic patients receiving enzyme-inducing anticonvulsants, including phenytoin, indicated that these patients had a higher clearance of

paracetamol than control subjects (Perucca & Richens 1979a; Prescott et al. 1981) and that this was the result of enhanced glucuronide conjugation of paracetamol (Prescott et al. 1981). Unfortunately, the reports of these studies (Perucca & Richens 1979a; Prescott et al. 1981) do not include the individual data for the patients receiving phenytoin as the sole anticonvulsant agent. Methadone Finelli (1976) reported that opioid withdrawal symptoms developed 4 days after the start of phenytoin therapy in a methadone-maintained former heroin addict. Subsequently, Tong et al. (1981) reported a similar finding in 5 methadone-maintained volunteers in whom plasma methadone concentrations were determined before and after the introduction of phenytoin. The emergence of opioid withdrawal symptoms was associated with a significantly decreased trough plasma concentration and AUC of methadone. Methadone is eliminated primarily by biotransformation and Tong et al. (1981) also measured the recovery of the major urinary excretion product, the pyrrolidine metabolite. During phenytoin administration, the ratio of 24-hour urinary methadone to pyrrolidine metabolite excretion decreased significantly; further, there was a significant increase in the ratio ofpyrrolidine excretion to methadone AUe. While decreased gastrointestinal absorption of methadone could not be excluded as a contributing factor, it was concluded that this clinically significant interaction was probably due to an increase in the biotransformation rate of methadone. Adjustments to methadone dosage regimens should be anticipated when phenytoin is initiated or discontinued in methadone-maintained patients (Tong et al. 1981), and similarly in patients receiving methadone for the control of pain. Pethidine (Meperidine) Pethidine is biotransformed by hepatic carboxylesterase to inactive pethidinic acid and by the microsomal hydroxylation enzyme system to norpethidine, a metabolite with CNS-excitatory activ-


59

Non-smokers

Halt-hfe Clearance

Volume of dIstribution

Fig. 2. Effect of phenytoin pre-treatment on the proportionate change in half-life. clearance and volume of distribution of theophylline in old (_) and young (0) smokers and non-smokers (after Crowley et al. 1988).

ity but weak analgesic activity. The effect of coadministration of phenytoin on the disposition of an oral and intravenous dose of pethidine has been examined in 4 healthy subjects (Pond & Kretschzmar 1981). Phenytoin dosing led to a significant increase in both the systemic blood clearance of pethidine (about 20% increase) and the area under the blood norpethidine concentration-time curve (about 50% increase) after the intravenous administration of pethidine. The renal clearance

and unbound fraction of pethidine in plasma did not change; but the oral bioavailability fell from a mean of 61% (control) to a mean of 43% during phenytoin dosing. It was concluded that the likely mechanism of the interaction was enhancement by phenytoin of pethidine clearance, including that by N-demethylation. The clinical significance of the interaction is uncertain but patients on long term phenytoin therapy may require a higher rate of pethidine dosing (Pond & Kretschzmar 1981).

60

2.3.3 Theophylline Phenytoin has been reported to increase the clearance and decrease the half-life of the bronchodilator theophylline, both in patients (Marquis et al. 1982; Reed & Schwartz 1983; Sklar & Wagner 1985) and in healthy volunteers (Crowley et al. 1987, 1988; Marquis et al. 1982; Miller et al. 1984). This effect, which may be clinically important, has been attributed to enzyme induction. Theophylline is cleared mainly by oxidative biotransformation and its clearance is increased by a number of agents, including the hydrocarbons in cigarette smoke (Jusko et al. 1978). Interestingly, Reed & Schwartz (1983) reported that the commencement ofphenytoin therapy in a patient who was a heavy smoker resulted in a further increase in the clearance of theophylline, suggesting that phenytoin and cigarette smoking may have an additive inducing effect on theophylline biotransformation. That such an additive effect occurs has been confirmed by studies conducted in young adults (Crowley et al. 1987, 1988) and in elderly subjects (Crowley et al. 1988). Pre-treatment with phenytoin for 2 weeks increased the clearance of intravenously administered theophylline to an approximately equal degree in both smokers and non-smokers, and, importantly, this was the case in both age groups (fig. 2) [Crowley et al. 1988]. Whereas smoking appeared to increase the partial clearances of theophylline to 3-methylxanthine (3-MX) and to 1methy1uric acid (1-MU) to a greater extent than that to 1,3-dimethyluric acid (1,3-DMU), pheny-


toin increased the partial clearances for all 3 metabolites to an approximately equal degree in both old and young subjects (Crowley et al. 1988). This is of interest in view of the suggestion that the 8oxidation oftheophylline to form 1,3-DMU is carried out by a cytochrome P450 isozyme(s) different from the form(s) involved in the demethylation reactions which produce 3-MX and 1-MU (Birkett et al. 1988; Crowley et al. 1988). The studies of Crowley et al. (1987, 1988) are important for a number of reasons. First, they demonstrate that phenytoin increases the clearance of theophylline, a drug of low therapeutic index which requires careful selection of dose. Secondly, they show that phenytoin and cigarette smoking have an additive inducing effect on the biotransformation of theophylline. Thus, it cannot be assumed that cigarette smoking alone leads to a maximal induction of theophylline clearance, and recognition of this may be important clinically. Finally, it has been demonstrated that the ability of phenytoin to induce the metabolism of theophylline is maintained in old age, an interesting finding in view of conflicting results with other inducers and substrates in the elderly (Crowley et al. 1988).

(Part II of this article, including references, will appear in the following issue of Clinical Pharmacokinetics). Authors' address: Dr Roger L. Nation. School of Pharmacy, South Australian Institute of Technology, North Terrace, Adelaide 5000, Australia.

Pharmacokinetic drug interactions with phenytoin (Part II).

Pharmacokinetic interactions with calcium channel antagonists (Part I).

An updated comparison of drug dosing methods. Part I: Phenytoin.

Pharmacokinetic interactions between theophylline and other medication (Part I).

Pharmacokinetic drug interactions with rifampicin.

Pharmacokinetic drug interactions with oral contraceptives.

Pharmacokinetic drug interactions of macrolides.

Pharmacokinetic mechanisms of drug interactions.

Pharmacokinetic interactions with calcium channel antagonists (Part II).

Pharmacokinetic drug interactions in anaesthetic practice.

Pharmacokinetic drug interactions of afatinib with rifampicin and ritonavir.

Pharmacokinetic interactions between theophylline and other medication (Part II).

pharmacodynamic review (Part I).

Metabolism-related pharmacokinetic drug-drug interactions with tyrosine kinase inhibitors: current understanding, challenges and recommendations.

Therapeutic and pharmacokinetic effects of increasing phenytoin in chronic epileptics on multiple drug therapy.

Pharmacokinetic drug interactions in liver disease: An update.

Pharmacokinetic and pharmacodynamic drug interactions between antiretrovirals and oral contraceptives.

[Pharmacokinetic interactions].

Cytochrome P450 enzyme mediated herbal drug interactions (Part 2).

Cytochrome P450 enzyme mediated herbal drug interactions (Part 1).

Clinically significant pharmacokinetic drug interactions of antiepileptic drugs with new antidepressants and new antipsychotics.

Pharmacokinetic drug interactions with clopidogrel: updated review and risk management in combination therapy.

Clinical pharmacokinetic drug interactions associated with artemisinin derivatives and HIV-antivirals.

Venetoclax (ABT-199) Might Act as a Perpetrator in Pharmacokinetic Drug-Drug Interactions.