From a clinical standpoint, important features of enzyme induction can be summarized:
• The process is relatively slow, i.e. usually days or even weeks.
• The potential changes in drug concentrations can be great enough to cause treatment failure.
• The induction process is usually, but not always, reversible over a similar time frame to its appearance, although reversal can be slower.
• Where a patient is stabilized on a high ‘induced’ drug dosage, if there is a treatment break of up to several days, drug accumulation and toxicity will occur.
The timescale of the induction process does largely depend on the potency of the inducer. Pentobarbitone causes a marked decrease in nortriptyline blood concentrations within only two days, doubling its clearance. Conversely, anecdotal evidence regarding teenagers suggests that induction of ethanol metabolism is a much more protracted process. Generally, the decline in drug levels caused by induction will lead to a commensurate loss of drug efficacy. Some of the most clinically relevant drug interactions caused by enzyme induction are described below.
Anti-epileptic agents
Drug combinations
In approximately oneIIhird of cases of epilepsy, control of the condition can only be achieved with a combination of anti epileptic drugs (AEDs) which are also known as anticonvulsants. This combination therapy can lead to potential problems with the induction effects of carbamazepine (CYPs 2C9, 2C19, 3A4; Histories 1 and 2), phenytoin (CYPs 1A2, 3A4) and phenobarbitone (CYPs 1A2, 2C8, 3A4). In a combination of inducing anticonvulsants, co- administered compounds metabolized by these CYP enzymes will have their plasma concentrations significantly reduced. A good example of this is valproic acid, where plasma levels can be reduced by 80 per cent in the presence of phenobarbitone and by half with phenytoin and nearly 70 per cent with carbamazepine co-administration. Felbamate clearance is modestly accelerated by carbamazepine whilst among the second generation of AEDs, drugs such as topiramate and tiagabine are also cleared more rapidly in the presence of the AED inducers.
It is important to realize, however, that there are many alternative anticonvulsant drugs which are not inducers, such as valproate (an inhibitor of CYPs) lamotrigine, pregabalin, zonisamide, levetiracetam and topiramate. Gabapentin is also in this category and it is not actually metabolized at all and is cleared entirely renally as a parent drug. At the moment it is unlikely that these non-inducing AEDs will replace the inducing AEDs, which remain the main drugs of choice in epilepsy, despite their effects on other drugs. The newer agents are used both as first4ine therapy and as substitutes for the inducing AEDs when the induction process presents an insurmountable clinical problem.
Drug withdrawal
It is generally stated that no AED, whether inducer or not, should be abruptly withdrawn. In regimens containing inducing AEDs, any changes or drug withdrawals can lead to acute problems related to the rapid reversal of the induction process (History 3). The remaining drug plasma levels might rise over the following few days as the inductive effect recedes and clinical signs of an intensification of the pharmacological effect will gradually become apparent. It is most desirable to anticipate this effect by tapering the dosage of the other drugs over days or weeks as appropriate. However, this is not such an easy process, as there is relatively little literature on how long it takes for the effects of standard inducers to fully wear off. There is some evidence that in some drugs it can take longer to disappear than the original onset time. It is better to taper the dose, or the patient might be subject to increased side effects, which they may or may not complain about. Overall, it is important that the drug levels remain within the therapeutic window and toxicity is avoided.
Epilepsy and brain tumours
The majority of brain tumours originate from malignant glial cells, known as gliomas. Although brain tumours account for less than 2 per cent of adult cancers, they are ten-fold more common in children. Between 50 and 70 per cent of patients with gliomas of various types suffer from epilepsy and studies have shown that use of the inducing AEDs present several clinical problems in this context. Antineoplastic drugs are a vital part of the therapy of gliomas and many of these drugs are cleared by the main inducible CYPs, so the inducing AEDs can accelerate the clearance of the anticancer drugs, reducing their effectiveness and in some cases change or accentuate their toxicity. The taxanes (paclitaxel and docetaxel) illustrate this point, as their clearance (CYP2C8) is increased by the inducing AEDs by about 1.5 – 2 fold. What is interesting, is that the usual adverse reactions caused by taxanes when given alone (gut toxicity and myelosuppression) change to peripheral neuropathy in the presence of inducing AEDs. This illustrates a re-routing of taxane metabolism and toxicity by induction. Some authors have recommended that either inducing AEDs should not be used with taxanes at all, or the anticancer agent’s dosage should be increased by around half to ensure adequate therapeutic effect. During a weekly treatment cycle, docitaxel is generally administered after dexamethasone premedication to minimize fluid retention. Interestingly, as if the toxicity of the taxane was not bad enough, dexamethasone can have an amphetamine–ike ‘speeding’ effect which may last several days. This is not uncommon, but it is not considered a major side effect. Although dex-amethasone is also a CYP inducer, the benefit in the reduction of side-effects favours the use of the steroid.
Vinca alkaloid (vinblastine, vincristine, vindesine) clearance is increased by more than 60 per cent in the presence of inducing AEDs. Phenytoin increases the clearance of the camptothecin derivative irinotecan, which is a CYP3A4 substrate, by more than 50 per cent. Teniposide and etoposide are also CYP3A4 substrates; phenytoin and phenobarbi-tone can treble their clearances, requiring dosage increase to maintain antitumour effect. A clinical study has shown that teniposide ’ s efficacy is actually poorer with inducing AEDs than without.
Other drug combinations
Anticonvulsants are co-administered with other CNS modulating drugs, such as antipsychotics, tricyclic antidepressants (TCAs), benzodiazepines and newer agents such as SSRIs. With respect to enzyme induction, anticonvulsants can greatly accelerate the clearance of antipsychotics like haloperidol and benzodiazepines such as midazolam, although temazepam clearance is not dependent on 3A4 and is not affected by inducers of this isoform. CYP3A4 inducers can also accelerate the clearance of some TCAs.
OTC (over the counter) herbal preparations
It has been estimated that nearly half of the US population has used herbal or complementary medicine in their lifetime and this rises to nearly 70 per cent of those suffering from life-threatening conditions such as HIV or cancer. Unfortunately, for many reasons, patients do not necessarily inform their medical practitioners of their use of these products. Although the clinical potency of many herbal remedies is at best arguable, the most popular, St John’s Wort (hypericum perforatum), has been extensively clinically investigated and found to be effective in mild to moderate depression and anxiety. It contains the active components, hypericin, pseudohypericin and hyperforin, as well as several flavonoids, xanthones and many other chemicals. Of course, patients are usually unaware that hyperforin, (unlike the other constituents) is a powerful activator of human PXR, leading to induction of CYPs 3A4 and 2C9. Hyperforin itself undergoes extensive metabolism by CYP3A4 to hydroxylated products. It is safe to assume that St John’s Wort will accelerate the clearance of any and all CYP3A4 and CYP2C9 substrates, which, as you know, involves the majority of prescription drugs. A 14-day course of St John ’ s Wort and dosage of a total of 900mg daily is capable of doubling alprazolam clearance. The herb has been shown to accelerate the metabolism of cyclosporine in renal transplant patients (at only 600mg daily), potentially leaving them at risk of tissue rejection unless the immunosuppressant’s dosage was increased by 60 per cent. This is also the case with the newer anti-rejection drugs tacrolimus and everolimus. Moreover, the inducing effect of the herb complicates the pharmacology of cyclosporine, as the increased production of metabolites is associated with greater nephrotoxicity, although the metabolites also exert an immunosuppressive effect. St Johns Wort also increases the clearances of amitriptyline and indinavir, although it does not increase the clearance of carbamazepine as this drug already maximizes its own induction. Confusingly, the herb has some inhibitory effects on CYPs in vitro, but this has not been so apparent in vivo. St John ’s Wort is also a potent inducer of the transporter P-glycoprotein (section 4.4.7) which is probably how it reduces digoxin levels by accelerating the rate it is pumped out of intestinal cells back into the gut.
Unsurprisingly, many patients who are undergoing a variety of conventional therapies are likely to become depressed and turn to herbal remedies such as St John’s Wort for relief. A high proportion of cancer patients are known to use such remedies, as the processes of chemotherapy, radiotherapy and recovery from surgery can range from merely gruelling to unendurable.
There are several key clinical problems to summarize which may be encountered by the use of St John’s Wort in particular:
• Many patients do not consider herbal remedies as ‘real drugs’ so they do not inform their medical practitioner they are taking or stopping them (History 4). In addition, patients often fully expect these remedies to be wholly beneficial without any side – effects.
• The onset of the inductive effect of St John’s Wort is variable, with some studies suggesting it must be taken for weeks to see a full effect, whilst in contrast, its impact was demonstrated within three days in the case of cyclosporine co-administration in the renal transplant patients. This may be related to the hyperforin content of the extracts.
• The patient may abruptly terminate their self-medication if they feel better or encounter side-effects (History 4).
• Hyperforin, pseudohypericin and hypericin have long half lives (16-42 hours) so the timeframe of the diminution of the inductive effect once self-medication ceases may be unpredictable. Some studies have shown it can take up to three weeks for the effects to subside.
• Quality, purity and content of the active ingredients in herbal preparations can vary widely according to methods of preparation.
This final point is crucial: one study found a 62-fold difference between the hyperforin contents of various commercially available preparations. Clearly, some of these preparations contain so little hyperforin that they are unlikely to even show any pharmacological effect, besides any enzyme inductive actions.
Another problem with herbal preparations is that although many are available, few have undergone even preliminary evaluation for their effects on the clearance of drugs. Green tea (contains epigallocatechin gallate) and sleep promoting valerian extract have both so far shown no inductive effects on drug clearance, although garlic significantly reduced the bioavailability of the anti-HIV drug saquinavir, possibly through its effects on transporters. The effects of Echinacea are more complex. This herb has been used for centuries to combat the symptoms of colds and ‘flu, although its efficacy in the prevention of common colds and its actions on the immune system are unproven. It may induce CYP1A2 and CYP2C9, although it has a mixed and unconfirmed effect on CYP3A4. As with many other remedies, such as ginger (a possible inducer -n vitro) ginseng and kava-kava, the enzyme–nducing properties of most herbal preparations have not been substantiated or systematically investigated. A study has suggested that Gingko biloba extract maybe an inducer of CYP2C19, as it accelerated omeprazole clearance. One point important to emphasize, is that assuming various herbal remedies do contain active and potent substituents, there is virtually nothing known clinically about what effects mixing herbal remedies might have, in terms of pharmacology and toxicity. This area is unfortunately left for patients to discover for themselves. It has been reported that a combination of St John’s Wort and kava-kava caused acute hepatitis in a 48-year-old female and although this possible interaction should be investigated further, it would seem to be reasonable to recommend to patients not to take these remedies simultaneously.
It is important that patients are asked if they have taken, or would consider taking, herbal remedies during a drug treatment regimen. For instance, this may be a problem where a course of conventional antidepressants is embarked upon and the patient’s symptoms do not improve quickly enough. Consequently they may understandably resort to assistance from an herb extract such as St John’s Wort. Even if the patient is aware of the potential impact St John’s Wort might have on their drug therapy, they should still feel confident in discussing this with their healthcare practitioner. From the practitioner’s point of view, since this herb does have proven efficacy, it may be worthwhile considering accommodating its inducing effects in a drug regimen, if it is judged to be in the interests of the patient. In some cases, such as with statin therapy, a CYP3A4 substrate such as simvastatin could be substituted with pravastatin (cleared virtually unchanged), or rosuvastatin (<10 per cent cleared by CYP2C9). In other cases, the induced drug dosage could be increased to compensate for the inductive effect of St John’s Wort.
Anticoagulant drugs
Atrial fibrillation (AF) promotes the formation of microemboli which increases a patient’s risk of a stroke by five fold. As well as AF, those at risk of other blood clotting related conditions such as deep vein thrombosis (DVT) and valve replacement recipients can be protected with anticoagulants such as warfarin. Interestingly, in the UK it has been suggested that more than 30 per cent of patients who could benefit from warfarin do not receive the drug for various reasons. In practice, warfarin is monitored pharmacodynami-cally, that is to say its effect is monitored rather than its concentration. The measurement ‘INR’ is used which is the ‘international normalized ratio’, which effectively assumes that normal blood clotting time is taken as around 1, so for warfarin to anticoagulate enough to successfully treat AF, an INR of 2.5 is required. So through frequent clinical monitoring, the dose is adjusted to achieve an INR value of approximately 2-3. At the other end of the scale, those with mechanical mitral valve replacement require maintenance of an INR of 3-3.5. Warfarin is given as a mixture of two isomers, S and R. The S isomer is up to five-fold more potent an anticoagulant than the R isomer and the S is cleared by CYP2C9, whilst the R is metabolized by CYP1A2 and CYP2C19. CYP3A4 has a minor role for each isomer. Inducers of these enzymes will make a substantial reduction in the plasma levels of warfarin and after a lag period, its anticoagulant effects will recede (History 5). There is no substitute for checking INR to ensure that they remain within the therapeutic window. If an enzyme-Inducing drug is withdrawn, there is the danger of accumulation of the anticoagulants, which will lead to haemorrhaging. It appears that most clinical problems with warfarin seem to occur as a result of the effect of inhibitors rather than inducers and some groups of patients on warfarin are more at risk than others from drug interactions, particularly those receiving cancer chemotherapy. When warfarin patients undergo minor surgical procedures, most authorities recommend they should continue to take the drug, as withdrawal’s problems outweigh its benefits. Prior to major surgical procedures, 4-5 warfarin doses might be withheld, letting the INR fall to around 1.5. If an inducer is part of the regimen and this drug is stopped also, it is worth remembering that its inductive effect will wear off in a few days and dosage adjustments may have to be made when the full regimen is resumed. It is interesting to note that those with mechanical valves must be anticoagulated constantly and if warfarin is stopped for a major procedure, another agent such as heparin is used during hospitalization to ensure they do not develop life- threatening emboli.
Oral contraceptives/steroids
The CYP3A4 inducers including St John’s Wort can accelerate the clearance of ethiny-loestradiol; this is a particular concern with low- dose oral contraceptive preparations. Increasing the contraceptive dose, or a recommendation to use other methods of contraception, may negate this effect. Corticosteroids are potent inducers of CYP3A4 and CYP2B6 and they can cause the clearance of inducing AEDs such as phenytoin to be accelerated by a factor of nearly three. It is recommended that when inducing AEDs are used during steroid therapy they are closely monitored to ensure that their levels do not fall out of the therapeutic window. This effect also works the other way around, as the AEDs can cause the acceleration of exogenous and endogenous steroid clearance. In general, inducers of CYP3A4 will accelerate the clearance of corticosteroids.
Antiviral/antibiotic drugs
Of the newer anti-HIV antiviral compounds, ritonavir, nevirapine, indinavir and saquinavir are all metabolized by CYP3A4, so it is possible that inducers may affect their clearances in vivo. However, this situation is complicated by the fact that for example ritonavir is a potent inhibitor of CYP3A4 and induces its own metabolism. This induction effect means that at least 14 days’ therapy is required before plasma levels stabilize. Potent inducers such as rifampicin do exert some effect on ritonavir plasma levels, but only to a relatively modest (~35 per cent) degree. It is believed that the observed acceleration of the clearance of drugs co-administered with protease inhibitors is most likely to be due to the induction of CYP2B6, 2C8/9 and to a lesser extent CYP2C19, as these CYPs are not inhibited by the proteases.
Of course any changes in the plasma levels of an antibiotic or antiviral agent can lead to subcurative drug concentrations and a possible selection of resistant variants of the infectious agent, so plasma levels should be closely monitored to ensure minimum inhibitory concentrations (MICs) are exceeded while toxicity is minimized. St John’s Wort is known to cause indinavir levels to fall below the MIC. Certainly abruptly stopping and restarting inducing antibiotics such as rifampicin (Histories 6 and 7) will lead not only to resistance, but also to severe disruption of the clearances and efficacies of co -prescribed agents. Co-administered drug levels will climb above the therapeutic window until the inducing effect is re-established. Patient drug tolerance may be severely impaired during this period.
Anti-cancer drugs
As mentioned above, many antineoplastic agents are cleared or activated by CYP metabolism and changes in their plasma levels can have serious repercussions in terms of toxicity and therapeutic effects. The acceleration of taxane metabolism (CYP2C8) leads to their clearance of inactive metabolites and a loss of efficacy. Other antinoplastics, such as thiotepa are cleared by CYP2B6 and CYP3A4, but both parent and metabolite (tepa) are equally therapeutically effective, so dose modification in the presence of inducers is easier to manage clinically. The effects of induction on anti cancer ‘pro-drugs – can be more complex. Cyclophosphamide and its sister compound ifosfamide are probably the most widely used of this type. These agents are part of several cancer therapeutic regimens, for either primary or metastatic disease. Cyclophosphamide is activated to an active 4-hydroxy metabolite mainly by CYP2B6, CYP2C19, CYP3A4 and to a lesser extent, CYP2C9. The 4-hydroxy cyclophosphamide enters cells and then rearranges itself into a highly reactive phosphoramide mustard derivative, which is similar to the blister forming antipersonnel mustard agents used in the Great War (1914-1918). This metabolite indiscriminately kills any growing cells (malignant or non- maligant) by alkylating their DNA, leading to a therapeutic effect, but an appalling list of toxic effects, including hair loss, gut damage, nausea and vomiting, as well as cystitis, nephro/neurotoxicity and immunosuppression. Ifosfamide tends to be more neurotoxic than cyclophosphamide, mainly because it is much more subject to side-chain dechloroethylation than the cyclo derivative.
Cyclophosphamide and ifosfamide can induce their own metabolism within a treatment cycle of a few days and AED inducers such as carbamazepine accelerate this process markedly. With cyclophosphamide this leads to an increase in production of the 4-hydroxy metabolite by more than 70 per cent, leading to intolerable toxicity. As underlined in the previous section on epilepsy and brain tumours, it is recommended that non–nducing AEDs are used in patients taking cyclophosphamide and ifosfamide.
The role of CYPs in the metabolism and efficacy of the topoisomerase inhibitor irenote-can makes this agent vulnerable to the effects of enzyme induction. Irinotecan is used mainly in colon cancer and it is a prodrug, which is then hydrolyzed by carboxyesterases to its active metabolite SN-38, which is normally cleared by glucuronidation to the inactive SN-38G. Irinotecan is cleared by CYP3A4 to several metabolites and St John’s Wort accelerates this process thus restricting availability of the parent drug for SN-38 production, which in turn causes SN-38 levels to fall by more than 40 per cent. Although this alleviates the side effects, sadly it also means that the drug loses its therapeutic effect and the herb should not be taken with irinotecan. It is important to note that any CYP3A4 inducer would cause this effect and some, such as phenobarbitone, also induce glucuronidation, so the parent and the active metabolite’s clearances will both be accelerated, seriously eroding efficacy.
