Increasing numbers of people are seeking symptomatic relief of psychiatric disorders by using dietary supplements. Since this is generally without medical supervision (Fugh-Berman and Cott, 1999; Wong et al., 1998), it is essential that clinicians avail themselves of the extensive literature available on natural products so that potential problems can be avoided. Herb—drug interactions can be of two primary types: pharmacokinetic and pharmacodynamic. Dynamic interactions are those having to do with the mechanism of action, e.g. where the drug's pharmacologic actions may be in opposition to or in addition to one another. Pharmacokinetic interactions are the result of alterations in the absorption, distribution, metabolism or excretion of medications when given together with specific drugs, foods or supplements. Interactions between botanical products and prescribed medications could increase or decrease the action of the drug, though the majority of interactions are likely to go unnoticed due to their having only minor effects on drug plasma levels. For a few, this is not the case, and illustrations will be given to provide a basis by which many of the most serious interactions can be prevented. Since kinetic interactions are of much greater relative importance, only they will be considered here. While some pharmacokinetic information on herbal medicines is available (De Smet and Brouwers, 1997), much more is needed in order to put in proper perspective the vagaries of clinical anecdotes and in vitro experiments. Therefore, this chapter will not be a laundry list of putative interactions, but will offer an explanation of the primary types of kinetic interactions as well as a critical summary of reports concerning specific, illustrative herb—drug interactions, especially St John's wort (Hypericumperforatum; SJW).
Biotransformation of foreign exogenous compounds, such as drugs, takes place by various enzymatic reactions traditionally classified as phase I or phase II. Both phase I and phase II enzymes result in increased polarity of lipophilic compounds, thus aiding in their eventual elimination. The cytochrome P450 (CYP450) system is a family of enzymes, particularly concentrated in the liver and intestinal mucosa but also found in the kidneys, skin, lungs and other tissues. These enzymes are important for phase I drug metabolism. They normally attach small polar groups such as hydroxy or car-boxyl groups to their substrates to produce a more water-soluble compound. Phase II metabolism involves conjugation, primarily as glucuronides and sulphates but also with glycine, glutathione, acetylation and methylation. While less is known about these reactions, they can also be induced and inhibited, resulting in drug—drug or supplement—drug interactions (Liston et al., 2001). The phase I cytochrome P450 enzymes catalyse predominantly oxidative reactions. These P450 enzymes may be thought of as a 'garbage disposal' that helps rid the body of various poisons and toxins before they harm us (Vogel, 2001a). Thus, foods or foreign substances that induce the enzymes may have served a beneficial role in our evolution. While twelve gene families have been identified, three categories of these enzymes are of the greatest significant in humans: the CYP2C, 2D6 and 3A4 (Hardman et al., 1996; Caraco, 1998). CYP2C (particularly 2C9 and 2C19) is responsible for the metabolism of many anticonvulsants, proton pump inhibitors, antidepressants and NSAIDS (non-steroidal antiinflammatory drugs). CYP2D6 is found in the liver, intestine, kidneys and brain where it mediates oxidative metabolism of many antidepressants, beta blockers, antipsychotics and other medications. This enzyme shows genetic polymorphism, and 'poor metabolizers' make up approximately 7 per cent of the caucasian population. While this enzyme is not itself induced, it has been suggested that persons deficient in CYP2D6 may be more susceptible to certain drug interactions after induction of other enzymes (Spina et al., 2001). CYP3A4 is the most abundant hepatic enzyme and accounts for the oxidation of over half of all medications subject to oxidative metabolism. Due to the liberal presence of the enzyme in the enterocytes of the small intestine, orally administered substrates of CYP3A4 undergo a significant extrahepatic metabolism prior to absorption. While the activity of CYP3A4 is monomorphically distributed, there is significant interindividual variation. The benzodiazepine, alprazo-lam, is a rather pure substrate for CYP3A4 and is even used as a marker for the enzyme.
Many foods and drugs induce or inhibit (or both) the activity of CYP450 enzymes. For example, multiple oral dosing (twice daily for 10 days) of grapefruit juice to rats has been reported to inhibit the intestinal metabolism of nifedipine while simultaneously inducing liver microsomal metabolism (Mohri et al., 2000). Induction is a slow process, since it depends on the rate of synthesis of new enzymes. It is usually noticeable after a few days, and may be maximal after two weeks. Inhibition is more rapid, and can become maximal within the first 24 hours of exposure to the inhibitor — but likewise may reverse more rapidly.
Herbal products usually contain numerous pharmacologically active constituents including essential oils, tannins, coumarins, anthraquinones, saponins, glycosides, anthocyanins, alkaloids and flavonoids, all of which may potentially participate in herb—drug interactions. In vitro studies have shown the ability of plant saponins to inhibit the CYP450 enzymes (Kim et al., 1997). Some coumarins may also inhibit specific CYP isoenzymes (Tirillini, 2000; Ohnishi et al., 2000). While a large amount of in vitro data are available regarding the ability of plant flavonoids to inhibit CYP isoenzymes (Obermeier et al., 1995; Henderson et al., 2000), the effects vary with the tissue being studied (Ueng et al., 2000). They may also be additive. A synergy between the coumarins and the flavonoids may be important in regard to grapefruit inhibition (Tirillini, 2000). For example, the naringenin bioflavonoids and the coumarins of grapefruit juice are reported to inhibit intestinal CYP3A4 and may cause clinically significant drug interactions with felodipine, cyclosporin, terfinadine and diazepam (Lown et al., 1997a; Fuhr, 1998; Özdemir et al., 1998). The flavonoid quercetin is a constituent of many herbal products, including SJW. It has been shown to inhibit CYP3A4 in vitro (Li et al., 1994). Cruciferous vegetables such as Brussels sprouts and broccoli induce CYP1A2 (Kall et al., 1996; Fontana et al., 1999). This enzyme metabolizes many carcinogens, including tobacco-related compounds and char-grilled meat. Red wine has also been reported to inhibit CYP3A4 (Offman et al., 2001). A high-protein diet may increase liver metabolism while a low-protein diet may reduce it (Walter-Sack and Klotz, 1996; Stowe et al., 2000).
Although in vitro screening is a common and non-invasive means of screening for potential drug interactions mediated by the cytochromes, it has considerable limitations that may prevent generalization to clinical situations. For instance, in vitro drug and enzyme concentrations must approximate those attained in vivo, since enzyme specificity may be lost at elevated concentrations. False positives may be generated when crude extracts are incubated directly with hepatocytes (often at thousands of times the physiological plasma level). The incubates often contain constituents which would never be absorbed if orally ingested. Additional contributing factors that are difficult to simulate are genetic and environmental influences on enzyme expression, the extent of protein binding, hepatic blood flow and extra-hepatic elimination. Finally, the phase 2 enzymes are also subject to polymorphism, can be induced and inhibited, and are subject to rate limiting kinetics due to the availability of cofactors, and the overall redox status of the organism.
Thus, whole-animal or human clinical studies are by far the most useful measures of metabolic alterations since they incorporate the variables mentioned above and take into account the effects of stomach acids, digestive enzymes, transport systems, absorption, and so on.
St John's wort
In spite of recent reports concerning interactions of SJW with prescription medications, its safety record is excellent (Schulz, 2001) and it is still considered a first-line treatment for mild to moderate depression in Europe (Di Carlo et al., 2001).
Although SJW shows monoamine oxidase (MAO) inhibition in vitro, this effect has not been displayed in vivo, nor have there been any reported cases of MAO inhibitor-associated hypertension in humans using SJW (Cott, 1997; Upton et al., 1997). Although SJW has been reported to non-selectively inhibit uptake of serotonin, nor-epinephrine and dopamine (and other amines as well) into synaptosomes in vitro (Müller et al., 1997) and in vivo (Neary et al., 2001), the concentrations required are unrealistically high (approximately 1,000 times less potent than synthetic uptake inhibitors). In addition, the side-effects of SJW are not at all similar to those of serotonin (or other amine) uptake inhibitors (Woelk et al., 1994; Ernst et al., 1998). The phloroglucinol derivative hyperforin (up to 5 per cent of total extract) is believed to be the primary uptake inhibitor (Chatterjee et al., 1998; Müller et al., 1998). However, more recent studies suggest that neither SJW nor hyperforin are true uptake inhibitors since they do not bind to the uptake site like synthetic uptake inhibitors (Singer et al., 1999; Jensen et al., 2001). Rather, they appear to release monoamines from synapto-somes, yielding the same net effect in the in vitro assay (Gobbi et al., 1999). This 'pseudo' non-selective uptake inhibition has been suggested to be a release of transmitter related to increasing intracellular sodium concentration (Singer et al., 1999), calcium mobilization (Koch and Chatterjee, 2001), and ion channel modulation (Krishtal et al., 2001). Finally, the whole notion of the relevance of hyperforin and uptake inhibition is brought into question when one considers that the relatively hyperforin-free (<0.2 per cent) formulation, Ze 117 (Wurglics et al., 2001), shows clinical antidepressant efficacy when compared with placebo (Schrader et al., 1998) and equivalence to 20 mg/day fluoxetine (Schrader, 2000) and 150mg/day imipramine (Woelk, 2000) in major depression. SJW treatment also fails to inhibit uptake in human depressed patients, unlike tricyclic and SSRI (selective serotonin re-uptake inhibitor) antidepressants (Uebelhack and Franke, 2000, 2001). Thus, the true mechanism(s) of antidepressant action for SJW is yet to be determined (Gobbi et al., 2001).
Remaining putative antidepressant mechanisms include direct neurotransmitter release (Chatterjee et al., 2001), alteration of neuronal membrane fluidity (Eckert and Müller, 2001), adenosine receptor antagonism (Müller et al., 2000) and inhibition of free radical production (Hunt et al., 2001). Surprisingly, rat atrial tissue preparations showed SJW extracts to have serotonin antagonist activity and negative chronotropic and inotropic actions (Straumann et al., 2001).
Since the recent publication of data showing that SJW reduced plasma levels of indinavir (Piscitelli et al., 2000) and cyclosporin (Ruschitzka et al., 2000), there has been increasing interest in determining the extent of the interaction problem with this herb. Commercially available SJW extracts were examined for the potential to inhibit the human CYP enzymes, 1A2, 2C9, 2C19, 2D6 and 3A4 (Obach, 2000). Crude SJW methanolic extracts showed inhibition of all these enzymes at very high concentrations — IC50s ranged from 10 to 1,000 fxg/mL. The flavonoid compound I3,II8-biapigenin inhibited 3A4, 2C9 and 1A2 activities with IC50 values of 0.08, 4.0 and 3.7 (xM, respectively. Hyperforin inhibited 2D6, 2D9 and 3A4 with IC50s of 1.6, 4.4 and 2.3 fxM, respectively. The significance of these data is uncertain because the concentrations were higher than those attained clinically, e.g. hyperforin maximum plasma level is reported to be 280 nM (150 ng/mL) (Biber et al., 1998). In addition, the activities of isolated chemical constituents may not be relevant to whole or crude plant extracts. However, within physiologically relevant concentrations, the SJW constituent hyperforin induces CYP3A4 in hepatocyte cells via the pregnane X nuclear receptor (Ki = 27 nM) (Moore et al., 2000) and the steroid X receptor (Wentworth et al., 2000). Thus, the preponderance of data suggest that hyperforin may be the constituent responsible for enzyme and transport protein induction.
In contrast, rats treated with oral doses of 300 mg/kg SJW extract for 10 days showed no changes in CYP450 liver enzyme activity (Nöldner and Chatterjee, 2001). Rats treated orally with SJW did show reduced plasma levels of warfarin, however (Nöldner and Chatterjee, 2001). Together, these data suggest (in rats at least) that the metabolic induction by SJW takes place in the intestine, rather than in the liver.
Direct (in vivo) evidence of SJW interaction with CYP450 is more useful for predicting clinical interactions. One study to evaluate effects on CYP3A4 was conducted on 13 healthy volunteers given 300 mg standardized extract SJW t.i.d. for 14 days. Urinary excretion ratios (over 24 hours) of 6-beta-hydroxycortisol/cortisol were used as an index of 3A4 activity both before and after 14 days of SJW treatment. A significant increase (from a baseline ratio of 7.1 to an endpoint of 13) was seen and the authors concluded that SJW is an inducer of CYP3A4 (Roby et al., 2000).
In another study, the effects of SJW on the activity of CYP2D6 and 3A4 were assessed in 7 normal volunteers (Markowitz et al., 2000). Probe substrates included dextromethorphan (for 2D6 activity) and alprazolam (for 3A4 activity). They were administered orally with and without the coadministration of a standardized 300-mg extract of SJW 3 t.i.d. for 3 days. Urinary concentrations of dextromethorphan and dextrorphan were quantified. Plasma samples were collected (0—60 h) for alprazolam pharmacokinetic analysis sufficient to estimate tmax, Cmax, t1/2, and AUC (area under the concentration—time curve). No statistically significant differences were found in any estimated pharmacokinetic parameters, suggesting that short-term treatment with SJW is unlikely to inhibit CYP2D6 or CYP3A4 activity. The dosing regimen, however, was too short to draw conclusions regarding induction.
Similar probe methodology was used to examine 3A4 and 2D6 in 16 healthy volunteers divided into extensive and poor metabolizers. SJW extract (300 mg t.i.d.) was administered for eight days. There was a tendency for induction of 3A4, but there was no effect on 2D6. No significant (inhibitory) effect on either enzyme was seen after an acute dose of SJW (Ereshefsky et al., 1999). The same group of investigators also reported on the effect of acute and eight days' SJW treatment in 16 subjects on CYP1A2 and the phase 2 enzyme N-acetyltransferase (NAT2) using a caffeine probe methodology. The results showed no significant interactions with CYP1A2 or NAT2 metabolic pathways (Gewertz et al., 1999).
In a recent study, 12 healthy subjects (5 female, 7 male) received SJW extract for 2 weeks (Wang et al., 2001). Probe drugs were given to estimate the acute and chronic enzyme activity of CYP2C9, CYP1A2, CYP2D6 and CYP3A (oral midazolam for intestinal wall and hepatic enzyme, and intravenous midazolam for hepatic enzyme) before and after 2 weeks of SJW administration (300 mg 3 t.i.d.). Short-term administration of SJW had no effect on CYP activities. Long-term SJW administration resulted in a significant and selective induction of CYP3A activity in the intestinal wall. The activities of CYP2C9, CYP1A2 and CYP2D6 were unaltered (Wang et al., 2001). These data are consistent with other human findings and indicate that CYP3A4 is the only P450 enzyme affected by SJW. Of interest here is a report that the low-hyperforin formulation Ze 117 does not interact with the 3A4 substrate oral contraceptives (Brattstrom, in press).
P-glycoprotein (Pgp) is an ATP-dependent pump that moves substrates out of cells. It is an inducible membrane transport protein that was initially discovered by cancer researchers studying multidrug resistance to certain cytotoxic anticancer drugs (Johnstone et al. , 2000). This resistence was found to result in cross-tolerance or cross-resistance to structurally unrelated compounds due to an overexpression of a family of transporter proteins (Pgp) under the control of the multidrug-resistance (MDR-1) gene (Yu, 1999). Pgp is found in normal human renal, intestinal and biliary epithelia, adrenals, testis and pregnant uterus where it is a barrier to xenobiotic accumulation and a determinant of oral bioavailability of many drugs (Tanigawara, 2000). It is also found in both the choroid plexus and the cerebral endothelium where it contributes to the blood—brain barrier and limits entry of drugs into the brain (Sugiyama et al., 1999). Pgp is expressed in normal human T lymphocytes where it appears to participate in the transport of cytokines (IL-2, IL-4 and IFN-gamma) (Drach et al., 1996).
Pgp can also be affected by a range of naturally occurring compounds. Some of these, like grapefruit juice, also modulate CYP450 (Tirillini, 2000), although this may be a random rather than an intrinsic linkage between the two systems (Kim et al.,
1999). Also, while grapefruit juice has been reported to enhance Pgp transport (Soldner et al., 1999) the effects appear to be rather weak (Becquemont et al., 2001). With drugs that are substrates of both Pgp and 3A4 (such as indinavir and cyclosporin), pre-systemic metabolism would take place in a synergistic fashion (Hochman et al., 2000). This could result in large decreases in plasma levels by agents that induce expression of both proteins. Reactive oxygen species (ROS) downregulate the expression of Pgp (Wartenberg et al., 2000). Since many medicinal plant constituents are antioxidants, this mechanism could play a significant role in the proposed interactions. Several naturally occurring flavonoids (many of which are antioxidants) bind the protein with high affinity (Maitrejean et al., 2000). Rosemary (Rosmarinus officinalis) extracts appear to inhibit transport into cells expressing Pgp by prevention of binding of the substrate to the Pgp protein (Plouzek et al., 1999). The antioxidants in rosemary are polyphenols, rather than flavonoids (Offord et al., 1997). Methoxyflavones from orange juice are reported to inhibit Pgp-mediated transport of vinblastine into Caco-2 cells (Takanaga et al., 2000) while the antioxidant flavones, quercitin and kaempferol, induced expression of UDP-glucuronosyltransferases and Pgp protein in Caco-2 cell monolayers (Bock et al., 2000).
SJW has recently been reported to induce Pgp as well as CYP3A4. The administration of SJW extract to rats or to humans for 14 days resulted in a 3.8-fold or 1.4-fold increase, respectively, of intestinal Pgp expression (Dürr et al., 2000). However, the low-hyperforin formulation, Ze 117, lacks interaction potential with the Pgp substrate, digoxin (Brattström, in press). On the other hand, inhibition of Pgp can greatly increase transfer of certain drugs into tissues where they normally do not go, such as the HIV-1 protease inhibitor, nelfinavir, into testes and brain (Choo et al., 2000). Thus, the therapeutic efficacy of many drugs might be increased by enhancing their tissue perfusion. Several categories of drugs including antihistamines and diuretics have been reported to result in significant inhibition of Pgp at therapeutically relevant concentrations (Ibrahim et al., 2001).
The acute rejection of cardiac grafts in two male patients in their early sixties was recently reported (Ruschitzka et al., 2000). In both cases, immunosuppression was maintained with a standard triple therapy of azothiaprine, cyclosporin and corticos-teroids. Both patients were hospitalized because of early signs of rejection three weeks after beginning standardized SJW at 300 mg three times per day. In both cases, cessation of SJW led to an increase in cyclosporin levels; both patients were eventually stabilized and they recovered (Ruschitzka et al., 2000). In another report (Barone et al.,
2000), a 29-year-old woman who received a cadaveric kidney and pancreas transplant, with stable organ function and stable cyclosporin concentrations began self-medicating with SJW. After taking SJW supplements for four to eight weeks, her cyclosporin concentrations became subtherapeutic; this was associated with organ rejection. Four weeks after stopping SJW, her cyclosporin concentrations again became therapeutic. Two other kidney transplant recipients developed marked reduction of cyclosporin therapeutic activity after self-initiation of SJW but they did not reject the organ (Mai et al., 2000; Moschella and Jaber, 2001). Thirty patients at one institution in Germany were reported to have precipitous drops in cyclosporin plasma levels after starting SJW (Breidenbach et al., 2000). Two additional patients were just reported to have lowered plasma concentrations of cyclosporin due to SJW (Turton-Weeks et al., 2001).
Cyclosporin is known to be a substrate of 3A4, but 3A4 induction by SJW cannot explain the magnitude of the cyclosporin interaction. Much of the oral bioavailability variation in cyclosporin was previously ascribed to 3A4 variability. However, this variability is now known to be due to Pgp variably reducing the rate of intestinal absorption (Lown et al., 1997b). Thus, SJW extracts may have reduced oral bioavailability of cyclosporin by inducing Pgp as well as 3A4 (Dürr et al., 2000). In any event, since the potential SJW interaction with cyclosporin is marked, coadministration of the two agents should be avoided.
Since digoxin is a known substrate of Pgp transport, but is not metabolized by P450 enzymes, a recent clinical study on the interaction of digoxin and SJW suggests that Pgp modulation may be induced. Healthy volunteers were brought to steady state after five days' treatment with digoxin (Johne et al., 1999). The subjects continued to receive digoxin (0.25 mg/day) either with placebo (n = 12) or with 900 mg/day SJW (LI160; n = 13) for another 10 days. No statistically significant changes were observed after the first dose of SJW extract. However, 10 days of treatment with the extract resulted in a 25 per cent decrease of digoxin AUC (P = 0.0035) and a reduction in trough concentrations and Cmax of 33 per cent (P = 0.0023) and 26 per cent (P = 0.0095), respectively. SJW has recently been reported to induce Pgp. The administration of SJW extract to 8 healthy males over 14 days resulted in an 18 per cent decrease of digoxin concentration after a single dose of 0.5 mg (Dürr et al., 2000).
HIV patients are very likely to be taking a number of medications concurrently. They are also likely to be taking SJW for various reasons. Several reports suggest extreme caution should be used with respect to the potential for drug interactions in this sensitive group. Piscitelli and colleagues (2000) carried out a clinical study on the effects of SJW on plasma levels of the HIV protease inhibitor, indinavir, in healthy, non-HIV subjects. A baseline steady state with indinavir (3 X 800 mg) over 24 hours was established and, after a fourth dose on the next day, kinetic parameters were established. The same dosing regime was repeated after fourteen days of standardized SJW extract consumption at 3 X 300 mg/day. There was a large (57 per cent) reduction in the indinavir AUC after the SJW therapy. While the exact mechanism of this interaction is unclear, indinavir is a substrate of CYP3A4. However, as with cyclosporin, indinavir is also a substrate of Pgp (Choo et al., 2000).
A crossover study examined the effect of SJW (LI160) extract on a single dose of phen-procoumon (an anticoagulant closely related to warfarin) in ten healthy males aged 18 to 50 (Maurer et al., 1999). Subjects received SJW (300 mg t.i.d.) or placebo; on day 11 each received a single dose of phenprocoumon (12 mg). SJW resulted in a significant decrease (~17 per cent; P = 0.007) in the AUC of free phenprocoumon compared with placebo.
A letter to The Lancet from the Swedish Medical Products Agency reported seven cases where patients stabilized on warfarin had experienced reduced bleeding times during SJW consumption. No thromboembolic complications were noted, and either the SJW was discontinued or the warfarin dose was adjusted. The authors suggested the cause to be an interaction between SJW and CYP2C9 (the primary liver enzyme associated with warfarin metabolism) although there was no direct evidence for this (Yue et al., 2000). Another possible explanation for the interaction is reduced intestinal absorption due to induction of Pgp. In support of this possibility, rats treated orally with SJW did not show changes in liver enzyme activity but did show reduced plasma levels of orally administered warfarin (Noldner and Chatterjee, 2001).
The same Lancet letter cited above (Yue et al., 2000) also mentions reports of intermenstrual (n = 8) or changed (n = 1) menstrual bleeding in women (aged 23—31 years) who had been taking long-term oral contraceptives and had recently started taking SJW (within the previous week in five of the cases). No details are given but the authors suggest that induction of 3A4 by SJW is responsible, since steroids are known substrates of CYP3A4. These reports resulted in the Swedish MPA contacting marketers of SJW and requesting that a warning be added to the labelling and that studies on the extent and implications of these interactions be carried out. Isolated reports of irregular bleeding are still occasionally being reported (Ratz et al., 2001) but the significance is undetermined. Exaggerated symptoms of low-dose oral contraceptives have been reported during concomitant administration with the antidepres-sant nefazodone, a CYP3A4 inhibitor (Adson and Kotlyar, 2001), so this effect is theoretically possible. To date, there have been no reports of decreased plasma levels of steroid hormones or of unwanted pregnancies associated with SJW.
Increased bioavailability of theophylline in human subjects (increased Cmax, AUC and elimination half-life) has been reported when it is combined with certain food substances including piperine from black pepper (Bano et al., 1991) and a high-carbohydrate, low-protein diet (Walter-Sack and Klotz, 1996). Theophylline has been reported to be metabolized (by demethylation) to a significant degree by CYP1A2 in human liver microsomes (Ha et al., 1995; Lee et al., 1998). As already noted, 1A2 enzymes are induced by tobacco, char-broiled meat, cruciferous vegetables, and a highprotein diet. There are no in vivo data that show an interaction between SJW and 1A2. On the contrary, eight days' treatment with SJW in 16 subjects showed no effects on 1A2 (Gewertz et al., 1999). While an interaction between theophylline and SJW has been cited dozens of times in the literature, the published report referred to is a letter to the editor regarding a single case of a 42-year-old woman (Nebel et al., 1999). She smoked half a pack of cigarettes daily and took eleven other prescription medications, most of which affect CYP enzymes, in addition to taking SJW for the previous two months. On cessation of SJW, her plasma theophylline levels rose within seven days (Nebel et al., 1999). This same paper also discussed unpublished in vitro data suggesting induction of 1A2 with pure hypericin at concentrations several hundred times greater than those found in plasma. This report is difficult to evaluate and does not constitute evidence for an SJW—theophylline interaction. CYP2E1 may also be involved in theophylline metabolism (Kharasch et al., 1993). Alcohol is known to induce this enzyme, but no mention was made in the report of alcohol consumption or of the other dietary factors influencing 1A2. Smoking is a known inducer of CYP1A2 and has been reported to reduce plasma levels of the CYP1A2 substrate, clozapine, by 20—40 per cent, while cessation of smoking elevated clozapine levels by 72 per cent (Meyer, 2001). Until additional in vivo data are available for SJW and 1A2, little can be said about interaction potential.
Eight healthy volunteers received 100 mg of carbamazepine twice daily for 3 days, 200 mg twice daily for 3 days, and then 400 mg once daily for 14 days (Burstein et al., 2000). On the last day, blood samples were taken. The subjects then took 300 mg SJW (0.3 per cent hypericin) 3 times daily with meals and with carbamazepine (400 mg) for an additional 14 days. On day 35, plasma samples were analysed for carbamazepine and its metabolite carbamazepine-10, 11-epoxide. There were no significant differences before and after the administration of SJW in carbamazepine concentrations at peak, trough or AUC. This suggests that SJW is either not a particularly powerful CYP3A4 inducer or that it cannot induce carbamazepine metabolism beyond the extent to which it induces itself.
The concern about potential interactions between SJW and other antidepressants probably stems from reports about its ability to inhibit MAO and serotonin uptake. The theory goes that combination of an antidepressant and a MAO inhibitor could result in hypertensive crises or combination with an uptake inhibitor could result in serotonin syndrome. There have been no reports suggesting that MAO inhibitor side-effects have ever occurred with SJW and this is in line with current evidence suggesting that MAO inhibition may be an in vitro artefact (Cott, 1997). There are a few case reports of 'serotonin syndrome' in the United States, but, interestingly, none in Europe, where SJW has been used extensively for many years (Schulz, 2001). One report concerned four cases of elderly patients described as having 'mild serotonin syndrome' but were consistent with exaggerated side-effects of sertraline, namely, nausea, vomiting and restlessness (Lantz et al., 1999). All patients were stable on sertraline and experienced these effects within 3 to 4 days of adding SJW. There are many conflicting literature references to drug metabolism, and sertraline is certainly an example. While most references do not list sertraline as a substrate of CYP3A4 (Xu et al., 1999), there is a case report of a 12-year-old boy on sertraline who experienced a serotonin syndrome when erythromycin, a known CYP3A4 inhibitor, was added (Lee and Lee, 1999). There is evidence, presented earlier, that acute doses of SJW may have a mild inhibiting action on CYP3A4. Since these patients were all stable on sertraline at the time they initiated SJW, this response can be explained by an increase in sertraline plasma levels — a pharmacokinetic effect, rather than a pharmacodynamic effect. A fifth elderly patient in the Lantz et al. (1999) report was stable on nefazodone when she added SJW. A similar exaggerated serotonergic response resulted which is consistent with increased blood levels of nefazodone due to acute inhibition of CYP3A4 (Feucht and Weissman, 2000). The opposite effect could be predicted if the SJW had been started first, followed by the antidepressant. This is in fact the result of a clinical trial of amitriptyline and SJW (Roots et al., 2000). In this study, 12 depressed patients received 900 mg SJW extract along with 75 mg twice daily of amitriptyline for 14 days. Reductions in AUC of 21.7 per cent were seen for amitriptyline and 40.6 per cent for nortriptyline. Levels of amitriptyline and its metabolite continuously decreased over the 14-day period, consistent with enzyme induction. Amitriptyline is another drug where considerable contradiction exists in the literature regarding its metabolism. David Flockhart's website lists amitriptyline as a substrate for CYP1A2, 2C19, 2C9 and 2D6, while Feucht and Weissman (2000) also list it as a substrate for CYP3A4 and glucuronyl transferase.
Was this article helpful?
Are You Depressed? Heard the horror stories about anti-depressants and how they can just make things worse? Are you sick of being over medicated, glazed over and too fat from taking too many happy pills? Do you hate the dry mouth, the mania and mood swings and sleep disturbances that can come with taking a prescribed mood elevator?