JLE

Epileptic Disorders

MENU

A review of pharmacokinetic drug interactions between antimicrobial and antiseizure medications in children Volume 23, issue 2, April 2021

Epilepsy is one of the most common neurological disorders affecting children [1] and requires long-term treatment to prevent or limit seizure recurrence. Traditional antiseizure medications (ASMs) can exhibit several drug-drug interactions (DDIs) [2], and new compounds with high interaction potential such as stiripentol and cannabidiol have recently entered the market with indications for specific paediatric epileptic syndromes.

In the general population, infectious diseases have constituted the most serious health issue in the world up to the mid 20th century when, in developed countries, chronic degenerative diseases began to dominate the scenario [3]. Remarkably, infectious diseases still represent a major problem in children, and the COVID-19 emerging pandemia has renewed interest in the topic [4]. It has been estimated that a median of 14 infectious episodes occurs at age 0-3 years [5], and antimicrobial drugs constitute 60% of all medical prescriptions in children [6].

So far, several studies have assessed the DDIs occurring between ASMs and drugs used for treatment of diseases, most often comorbid with epilepsy in adults and elderly patients [7, 8]. Conversely, no updated systematic evidence exists for DDIs in paediatric age [9] despite the burden of the condition. According to one study performed almost 15 years ago, in primary care, 3.0% of children on chronic antiepileptic therapy were coprescribed therapeutic agents, which could give rise to clinically serious DDIs [10].

A few years ago, a review was performed on DDIs between ASMs and between ASMs and other drugs [2]. The aim of this review was to focus on a more specific aspect, that is the pharmacokinetic DDIs occurring between ASMs and drugs used for the treatment of infectious diseases in children. This focus relies on three specific aspects. First, while adult neurologists are generally aware of DDIs between ASMs and other drugs because of the high frequency of comorbidities in adult and elderly patients with epilepsy, paediatric neurologists are less aware of this issue. Second, although the pharmacokinetics in children is more similar to that in adults than neonates or infants, several differences exist. For several agents, children show higher drug clearance than adults and in general have higher pharmacokinetic variability that may be enhanced by enzyme induction or inhibition [11]. This also applies specifically to ASMs [12]. Third, antimicrobial drugs are the most frequently prescribed drugs in children [6] and are the source of a myriad of dangerous DDIs.

Identification of DDIs is an important component of preclinical development of any drug and is accomplished by in vitro and in vivo studies. This information, combined with that obtained from studies on healthy volunteers and from patients during clinical development, is stored in large data bases and used by drug compendia to allow prediction of DDIs that, in the majority of cases, are not verified through specific studies in humans.

Accordingly, in this review, as a first step, all potential DDIs were identified for all drugs, and as a second step, clinical confirmation was searched and concordance between these findings evaluated.

Search methods for systematic identification of drug interactions

We systematically searched all DDIs between ASMs and drugs pertaining to the class “anti-infectives for systemic use” (Class J) of the Anatomical Therapeutic Chemical (ATC) classification system [13].

ASMs searched were: brivaracetam (BRV), cannabidiol (CBD), carbamazepine (CBZ), clobazam (CLB), clonazepam (CNP), eslicarbazepine (ESL), ethosuximide (ETS), felbamate (FBM), gabapentin (GBP), lacosamide (LCM), lamotrigine (LTG), levetiracetam (LEV), midazolam (MDZ), oxcarbazepine (OXC), perampanel (PER), phenytoin (PHT), phenobarbital (PB), pregabalin (PGB), rufinamide (RFN), stiripentol (STP), topiramate (TPM), valproic acid (VPA), vigabatrin (GVG), and zonisamide (ZNS).

We searched all anti-infectives authorized for paediatric use (age< 16 years) for possible DDIs with ASMs in two leading publicly accessible drug compendia [14]. In the case of antiviral drugs, Liverpool Interaction checker and the list of interactions of experimental drug therapies for COVID-19 were also consulted [15]. Drugs with indication for rare disorders (for example, leprosy) were excluded. When a potential interaction between any anti-infective and ASM emerged, the literature was searched for available clinical evidence through MEDLINE (accessed by PubMed: name of the anti-infective drug AND antiepileptic drugs AND drug-interaction) and the Summary of Product Characteristics (SPC) for each drug were consulted.

Mechanisms of pharmacokinetic drug interactions of ASMs

Pharmacokinetic interactions occurring at the metabolic level

The most frequent and clinically significant DDIs of ASMs occur at the metabolic level and are consequent to their oxidative metabolism and/or glucuronidation and to their effect on the enzymatic systems involved in metabolism of drugs, mainly the cytochromes P450 (CYPs) and the uridine glucuronyl transferases (UGTs) [16, 17].

All CYPs function as mono-oxygenases. They are classified into several families and subfamilies, have a characteristic substrate specificity and their activity is genetically determined [16, 17]. The UGTs catalyse drug glucuronidation and are less substrate specific than CYPs. They include two enzymatic families (UGT1 and UGT2), each with eight isoenzymes [16].

The activity of these enzymes may be induced or inhibited by other drugs. Enzyme induction consists of enhanced metabolic clearance of the drug which is the substrate of the induced enzyme, consequently leading to a decrease in its serum concentration. This process is characterized by increased synthesis of the enzymes involved in drug metabolism and requires time before the new steady state concentration of the victim drug is reached. Enzyme inhibition results in decreased metabolism and increased concentrations of the affected drug and takes place rapidly after the administration of the inhibiting agent [2, 17].

Although all pharmacokinetic DDIs can be predicted by the knowledge of the effects that perpetrators have on all CYP and UGT isoenzymes that metabolize the affected drug, the magnitude of the interaction is subject to high variability. The degree of interaction may be modulated by drug dose, genetic factors, different strength of the effect of perpetrators and the extent of the metabolic transformation of the enzymatic pathway of the affected drug that has been induced or inhibited as well as several other factors [18, 19].

Several traditional ASMs (PB, PHT, CBZ) are broad-spectrum strong enzyme inducers as they can induce the activity of many CYP450 enzymes and/or UGT isoenzymes; several second-generation ASMs are weak enzyme inducers (ESL, OXC, FBM, RFN, TPM at doses higher than 200 mg/day and PER at doses higher than 8 mg/day). VPA, FBM, STP, CBD, and BRV have enzyme inhibiting properties although some of them can also exert inducing effects on the same or other enzymes (OXC, STP, FBM, CBD) [20].

Pharmacokinetic interactions affecting absorption and distribution

Although DDIs concerning drug metabolism are the most important, DDIs can also occur at the level of gastrointestinal absorption, renal excretion, or distribution of the drug [2, 20].

Drug absorption, disposition and elimination are influenced by some transmembrane polypeptides. These proteins are divided into an ATP-binding cassette (ABC) family and a solute carrier (SLC) family [19]. The ABC family include P-glycoprotein 1 (permeability glycoprotein, Pgp), also known as multidrug resistance protein (MDR1). This is a protein of the cell membrane that pumps many foreign substances out of cells and is extensively expressed in the intestinal epithelium, in excretory cells in the liver and kidney, and in endothelial cells related to the blood-brain barrier, thus regulating absorption, excretion and distribution of a wide range of compounds [21]. Breast cancer resistance protein (BCRP; ABCG2) limits intestinal absorption of low-permeability substrate drugs and mediates biliary excretion of drugs and metabolites [22]. SLC family transporters include several transporting polypeptides [23]. The activity of all these peptides can also be induced or inhibited [24]. Interestingly, while traditional ASMs (CBZ, PHT, PB) [19] are inducers, some of the newer ASMs (CBD [25], STP [26] and BRV [27]) are inhibitors of some of these protein transporters.

Mechanisms regulating pharmacokinetic interactions

Although the consequences of all these interactions are complex and cannot be easily predicted, it has recently been shown that there is some coordination between all these mechanisms. Pregnane X receptor (PXR) is a nuclear receptor whose primary function is to sense the presence of foreign and possibly toxic substances. This receptor up-regulates the expression of cytochrome P450 genes (mainly CYP3A4) and that of conjugating enzymes such as glutathione S-transferase as well as the expression of efflux proteins such as some OATP transporters and P-gp. Many P-gp substrates overlap with CYP3A4 substrates, and several drugs that are CYP3A4 substrates are also P-gp substrates [19].

Specific aspects of interactions in children

For some drugs, the level of interaction in paediatric patients might be different leading to a higher or lower DDI potential compared to adults [11, 28].

Studies conducted on the expression of drug-metabolizing enzymes during ontogeny show that in liver samples of subjects from 1 to 10 years, CYP2C9 and CYP2C19 expression is 40-50% of that reported in adults, while the expression of CYP3A4 increases over these years and adult levels are generally achieved after the third year of age [29].

In a systematic literature review conducted for all drugs, only 31 paediatric studies on DDIs were identified corresponding to only 24 cases, and comparisons were possible between studies conducted in adults and paediatric patients (also neonates and infants). The magnitude of the interaction, as measured by the area under the curve and the clearance of the affected drug in the presence and absence of the perpetrator, was higher than, similar to, or lower than the corresponding ratio in adults in 10, 15, and 8 cases respectively [11].

The main enzymes involved in the metabolism of ASMs authorized in children and their inducing or inhibiting properties on enzymatic systems and membrane transporter proteins are reported in table 1 table 1.

Identification of drug interactions

Among more than 400 drugs included in Class J of the ATC system, 147 were found in the compendia [14]. Of these, 47 have potential DDIs with at least one ASM and are authorized for paediatric patients (18 antibacterials, six anti-mychotics and 23 antivirals including drugs used for COVID-19). Analysis of the literature allowed the identification of 111 records on DDIs between the 47 identified antimicrobials and ASMs. All potential DDIs and a summary of clinical data concerning each interaction are reported in tables 2, 3, 4 tables 2, 3, 4. Potential DDIs are ranked in parenthesis as “serious” (may result in potentially serious clinical consequences, and the combination should be avoided), “monitor” (not possible to avoid the combination, and clinical and laboratory monitoring is required which may lead to dosage adjustment), “minor” (limited clinical value, and dosage adjustment is usually not necessary). A list and a more detailed description of the clinical studies and case reports on the identified DDIs are presented in supplementary tables 1, 2, 3of the supplementary material. The most relevant DDIs are summarized in the following paragraphs.

Drug interactions between ASMs and antibacterials/antimycotics

Although antibacterials and antimycotics are usually taken for a limited amount of time and are rarely administered as a chronic therapy, there are examples of DDIs that can have very significant clinical consequences.

Interactions characterized by increased levels and/or effects of ASMs

Several antimicrobics are strong CYP inhibitors and may increase blood levels of ASMs with consequent risk of CNS toxicity [18, 20]. Among drugs used in children, ciprofloxacin, chlarythromicin, erythromycin, isoniazid, metronidazole, and fluconazole can increase levels of CBZ mainly through CYP3A4 inhibition; chloramphenicol, isoniazid, sulfamethoxazole and trimethoprim and the antimycotics fluconazole, voriconazole and isoniazid (in slow acetylators) can increase PHT concentrations with possible toxicity through CYP2C19 and/or CYP2C9/10 inhibition (supplementary tables 1, 2).

Given the lack of clinical studies, inhibition of metabolism of other ASMs by antibacterials can only be predicted (tables 2, 3).

Interactions characterized by reduced levels and/or effects of ASMs

The combination of antibacterials and ASMs may be less commonly associated with reduced ASM levels and possible loss of antiseizure activity. This effect generally derives from enzyme induction. Ciprofluoxacin and levofloxacin have been found to lower PHT levels with unknown mechanisms [2]. Rifampicine, a strong enzyme inducer of several cytochrome and glucuronidation enzymes and transport proteins, can be predicted to induce metabolism of several traditional and new ASMs. Amikacine and rifapentine have some inducing properties.

The fall of VPA levels with possible loss of seizure control induced by carbapenem is a case of antibacterial-induced reduction in the concentration of an ASM which is not due to enzyme induction [30].

Interestingly, this DDI was not predicted by in vitro studies and the first observation was in a child [31]. Moreover, this was also the only interaction that was assessed based on a retrospective study in a population of paediatric patients [32]. The suggested mechanism is the inhibition of acyl peptide hydrolase, a hepatic enzyme that catalyses the hydrolysis of VPA-glucuronide, leading to increased elimination of VPA [33, 34]. Ertapenem and meropenem seem to have greater effects on VPA levels than imipenem/cilastatin [35].

Interactions characterized by reduced levels/efficacy of antibacterials/antimycotics

There are also examples of ASMs that alter the metabolism of antibacterial and antimycotic drugs. These DDIs usually involve induction of oxidative metabolism with a possible loss of antimicrobic activity. Lower levels of doxycycline levels have been found in patients treated with PB and/or PHT. Metabolism of metronidazole is induced by PB. In one study in healthy subjects, PHT caused a 70% reduction in voriconazole plasma levels, which led to a request to double the voriconazole dose in order to maintain effective antimycotic concentrations [16, 20].

Interactions characterized by increased levels/efficacy of antibacterials/antimycotics

More recently, new ASMs with strong inhibiting properties, namely CBD and STP, have been approved for use for specific paediatric syndromes. These drugs may potentially inhibit the metabolism of several antibacterials and antimycotics.

Concordance between prediction and clinical observations

In the vast majority of cases, concordance has been found between predicted and clinically observed DDIs. However, this is not always the case. One study failed to confirm the predicted DDI between metronidazole and MDZ [36] and there are also clinical studies or case reports signalling a DDI not predicted by drug compendia. PHT was found to increase chloramphenicol levels [37]. Primidone [38] and ETS levels [39] were increased by chloramphenicol and isoniazid, respectively, and rifampicine decreased the active OXC metabolite [40]. Importantly, cases of discordant findings also exist. OXC levels are predicted to be decreased by clarithromycin. However, a 10-year-old boy treated with OXC developed neurological signs clearly related to OXC toxicity a few days after the combination of this agent with clarithromycin, in the absence of any change in OXC plasma levels. Clarithromycin-induced inhibition of the efflux proteins of the blood-brain barrier, which are over-expressed in drug-resistant patients, may increase brain OXC concentrations with consequent toxic symptoms [41] (see supplementary tables 1, 2). All the potential DDIs between ASMs and antibacterials and antimycotic drugs and a summary of the clinical evidence for each identified DDI are presented in tables 2, 3.

Drug interactions between ASMs and antivirals

In this section, antiviral drugs are classified according to their use against specific viral diseases, with special attention to chronic conditions such as HIV/AIDS, hepatitis C, hepatitis B and COVID-19.

Interactions between ASMs and antiretroviral therapies used as treatment for HIV/AIDS

Children may be infected in 0.7% cases through mother-to-child transmission [42]. This disease can be especially harmful to infants and children, with one study showing that 52% of untreated children born with HIV in Africa died by the age of two [43]. Notably, the WHO recommends treating all children less than five years old [43].

A guideline for patients who need to be treated with a combination of antiviral agents for HIV/AIDS and ASMs was delivered in 2012 [44]. Today, however, we can foresee more DDIs, mainly because of the introduction of new ASMs with strong inhibitory effects.

Several antivirals are comprised of a combination of two or more agents, one of which is a mechanistic inhibitor of CYP3A4 such as ritonavir or cobicistat. These agents are intended to enhance the effect of the combined protease inhibitor by inhibiting its metabolism [45, 46], but they have also some inducing properties on other enzymes. Hence, it is not surprising that these agents may have DDIs with a great variety of drugs, including ASMs undergoing oxidative metabolism.

Several clinical studies have documented two-way inducing effects between antivirals and ASMs. This is the case for DDIs between lopinavir/ritonavir and PHT, with increased clearance of both the antiviral, possibly via CYP3A4 induction, and PHT, possibly via CYP2C9 induction [47]. Coadministration of CBZ with efavirenz significantly reduces exposure to both drugs [2].

Documented examples of inducing effects of ASMs on antivirals include the effect of CBZ and OXC on the metabolism of atazanavir [48]; contrasting results are reported on the effect of OXC on dolutegravir [48, 49]. An example of an inducing antiviral is zidovuline, which has been shown to decrease PHT levels [18].

VPA has inhibiting properties and, as predicted, inhibits zidovuline metabolism and can double the level of the antiviral agent [2, 18]. However, in some patients, the combination of VPA and dolutegravir was associated with lower antiviral concentrations [50]. This unexpected finding may be due to reduced dolutegravir absorption by the excipients contained in some VPA gastro-resistant oral formulations.

Concerning the influence of antivirals on VPA metabolism, efavirenz or lopinavir/ritonavir have been shown to reduce VPA levels in one patient with bipolar disorder, possibly through the induction of VPA glucuronidation. However, in a pharmacokinetic study, the administration of a dose of VPA of 500 mg/day in HIV-1 infected patients receiving efavirenz or lopinavir/ritonavir and in patients who had discontinued the antiretroviral therapy did not result in significant effects on efavirenz and lopinavir concentrations; further, no concerns arose that coadministration with these antivirals agents significantly influences trough concentrations of VPA [51].

The inducing effects of ritonavir on UGT enzymes may be responsible for the decreased levels of LTG observed in a clinical study conducted in healthy volunteers [52]. In this study, lopinavir/ritonavir led to a clear reduction in LTG levels, atazanavir/ritonavir had a moderate effect in reducing LTG exposure, and atazanavir alone did not significantly influence LTG concentrations [53].

An interesting example of the complexity of pharmacokinetic interactions is that between etravirine and CLB. Etravirine may reduce CLB levels through CYP3A4 induction. However, this antiviral is also a weak CYP2C19 inhibitor and may increase the concentration of desmethylclobazam, the active metabolite of CLB. Indeed, a case of etravirine-induced CLB toxicity has been described [54].

Clinical studies and case reports generally confirm predicted DDIs; some exceptions, however, have been documented. Notably, OXC was not found to reduce dolutegravir or efavirenz levels [50, 55].

Interactions between ASMs and other antiviral agents

Even in the case of hepatitis viruses, there is a risk of mother-to-child transmission. This risk is higher for hepatitis B than for hepatitis virus C [56]. Both hepatitis B and C place a child at high risk of subsequent chronic hepatitis. For hepatitis C, when treatment cannot be deferred, different first-line treatment options exist in the paediatric population [57]; as these agents are often substrates and inhibitors of P-gp, several potential DDIs with ASMs may be predicted [58]. Fewer drugs are available in children with hepatitis B and infrequently they cause DDIs [59]. Agents used to treat influenza do not show DDIs.

DDIs between ASMs and drugs used to treat COVID-19 are interesting. In the pandemia caused by Sars-CoV-2 infection, children usually present with mild clinical symptoms. Some patients, however, may progress rapidly and develop respiratory failure [4]. No drugs are currently approved and several agents already available on the market for other indications have been repurposed [60]. This is the case for atazanavir/ritonavir, darunavir/cobicistat and lopinavir/ritonavir (see section on HIV/AIDs drugs). Other drugs repurposed from other diseases (chloroquine, hydroxychloroquine, nitazoxanide) or not yet on the market (remdesvir) are reported in table 4. Some non-antiviral agents used for COVID-19, such as chloroquine and hydroxychloroquine, also exhibit important potential DDIs as their metabolism is induced by ASM enzyme-inducers.

Potential DDIs between antiviral drugs and ASMs in paediatric patients and a summary of the clinical evidence are reported in table 5 table 5. Additional details about clinical studies are reported in supplementary table 3.

Specific aspects of drug-drug interactions in children

An important clinical question is whether some DDIs have clinical characteristics that are specific to children. Although enzymatic systems in children are more similar to those observed in adults than in infants or neonates, the extent of DDIs has been shown to differ for several combinations of drugs used in paediatric patients [11]. We found only 11 clinical studies or case reports that describe DDIs in this population of patients, almost all concerning DDIs between ASMs and antibiotics, and these were generally concordant with findings in adults although no inference can be drawn on the degree of differences for any of the DDIs.

Discussion

There is a growing awareness of the importance of DDIs between ASMs and other drugs as a possible cause of drug-related toxicity or loss of efficacy for both classes of agents [61].

In this review, all DDIs between ASMs and antimicrobials authorized for paediatric patients have been evaluated. Among more than 400 antimicrobials, we eventually restricted the search to the 47 agents that have indications in paediatric patients and potential DDIs with at least one of the 24 ASMs. It is, however, worth noticing that potential DDIs have been confirmed by clinical data in only a few cases and these were often restricted to DDIs involving old-generation ASMs.

One major issue is that more than 60% of anti-infectives were not found in the consulted compendia and consequently no information is available on potential DDIs for these agents. Although we can speculate that the most frequently used drugs have more chance to be included in such compendia, for several antimicrobials, there remains a lack of information.

A further finding is that DDIs predicted on the basis of in vitro studies and reported in drug compendia are not always confirmed by clinical studies. Different hypotheses may be proposed to explain these discrepancies. A negative finding in a clinical study does not necessarily mean that a DDI will not take place in any case as there are several factors that affect the likelihood that a known interaction will occur in a patient, such as drug dose, duration of combined therapy, age, genetic background and underlying diseases. In addition, different mechanisms of interaction can co-occur with uncertain final effects (for example, induction and inhibition of different enzymes that are involved in the metabolism of the affected drug). In the case of antimicrobials, the duration of drug co-administration is also relevant for the detection and assessment of the clinical consequences of a DDI. Enzyme induction requires synthesis of a new enzyme and the full effects of a DDI may be observed after several weeks; conversely, the effects of drug inhibition are evident relatively soon. Clinical consequences of the enzyme induction, often characterized by loss of efficacy, are more frequently observed with drugs that are combined over long periods of time, such as ,for example, the antiviral agents used for HIV treatment. In contrast, signs of drug toxicity, as a consequence of the inhibition of CBZ metabolism by macrolide antibiotics, are observed within a few days [2, 20].

Other sources of concern in the field of pharmacokinetic interactions are derived from the literature. In the case of several drugs, discrepancies in reporting and evaluating severity of interactions have been noted between different compendia [62]. Most often, inconsistences are noted between the effect of a drug on oxidative metabolism or membrane proteins, as described in the SPC and for DDIs reported in the compendia [7].

A first important question is: what is the clinical relevance of DDIs? To what extent are DDIs associated with a failure in resolving clinical effects (seizure relapse or failure to achieve resolution of infection) or toxic effects? Sub-therapeutic antiviral drug levels [63] and a risk of failure in treating viral infection [64] have been reported in patients treated with highly-active antiretroviral therapy combined with strong enzyme-inducing ASMs. These effects were caused by induction of oxidative metabolism of antiviral drugs. As previously described, this DDI is bidirectional because inhibition of CYP3A4 by several antiviral agents causes serious DDIs with ASMs. Here, it may be interesting to note that some DDIs are also caused by induction of cytochromes 2C9 and 2C19 or by induction of glucuronidation. For example, in healthy volunteers treated with lopinavir/ritonavir, dose increments of 200% of LTG were required to achieve concentrations similar to those observed in patients treated with LTG alone [52]. In a patient treated with VPA, combination with lopinavir/ritonavir, zidovudine, and lamivudine resulted in a 48% reduction in VPA levels and recurrence of maniac disease [65]. In both cases, this DDI was probably caused by the inducing properties of ritonavir on glucuronidation.

From an historical point of view, the most known and clinically relevant DDIs are those between several antibiotics and CBZ. Among antibiotics used in children, clarithromycin, chloramphenicol, doxycycline, erythromycin, isoniazid and metronidazole are moderate or strong CYP3A4 inhibitors and CBZ is metabolized mainly by this cytochrome. There are several clinical studies and case reports that show that treatment with one of these antibiotics may cause an increase in CBZ blood levels, that may be even doubled, as reported with erythromycin [66] and clarithromycin [67] with consequent drug toxicity. This DDI is very frequent and often serious because CBZ is widely used, is metabolized almost only by CYP3A4 and exhibits dose-dependent adverse effects (diplopia, ataxia, dizziness) and a narrow therapeutic window. Several other ASMs, such as CLN, MDZ, PER, and to a lesser extent CLB, ETS, LCM and ZNS are metabolized by the same cytochrome and their metabolism may be inhibited. Nonetheless, clinical consequences of such DDIs may be less severe because of a different pattern of dose-dependent adverse effects, a larger therapeutic window, and/or a minor amount of drug metabolized through this pathway.

An example of a particular DDI specific to one ASM is that between VPA and carbapenem antibiotics, which was discovered based on observations from case reports. In this case, enzyme inhibition determines a fall in VPA blood levels which is described at around 60% or more (see supplementary material) and has been associated with seizure recurrence in several cases. Although the exact mechanism of this DDI is not known, it is supposed that the inhibition of the hepatic enzyme acylpeptide hydrolase (an enzyme that hydrolyses VPA glucuronide metabolite back to its active VPA molecule) determines an upregulation of VPA glucuronidation [68]. This mechanism would explain why this DDI takes place rapidly while induction is usually slower. Among carbapenem antibiotics, ertapenem and meropenem have a greater effect than imipenem/cilastatin [35]. A further example of a DDI may be that of the antimycotic, voriconazole. This agent is metabolized by CYP3A4, CYP2C9/10 and CYP2C19, but is also a strong inhibitor of these enzymes. Consequently, in this case, there is also a bidirectional DDI with several ASMs. CBZ, PHT, PB, STP and OXC (to a lesser degree) may decrease the level of voriconazole, leading to possible inefficacy of antimycotic. It has been observed that combination of this agent with PHT (300 mg/day) requires a doubling of voriconazole dose to maintain therapeutic levels [69]. On the other hand, voriconazole inhibits the metabolism of several ASMs (mainly CBD, CBZ, LCM, MDZ, PHT and ZNS). In a clinical study in healthy volunteers, this agent increased Cmax and AUC of PHT by approximately 70% and 80%, respectively [69].

We may speculate that there are thousands of DDIs that have little clinical relevance in the majority of cases, but may cause serious complications in some patients. On the other hand, in some cases, the effect of a DDI may be less evident than predicted because other metabolic pathways may compensate for the effect of the perpetrator. For example, the magnitude of inhibition of MDZ metabolism (metabolized mainly by CYP3A4) by ketoconazole and indinavir (CYP3A4 inhibitors) is partially compensated by an increase in MDZ glucuronidation, which may attenuate the magnitude of the interaction [70]. In other cases, a change in the concentration of an active metabolite, as a consequence of induction or inhibition of the metabolism of the parent drug, may compensate the change in level of the affected drug. This may be the case for induction/inhibition of CBZ metabolism (for example, by the CYP3A4 inducer, rifapentine, or by the CYP3A4 inhibitor, clarithromycin). This may lead to increased or decreased levels of CBZ-epoxide, the active CBZ metabolite. In this case, the net effect of this DDI may not determine a corresponding change in CBZ efficacy or CNS toxicity.

Recommendations for prevention or minimization of the risk of serious complications from DDIs is provided in table 5.

Supplementary data

Summary didactic slides and supplementary tables are available at www.epilepticdisorders.com.

Disclosures

GZ has received speaker's or consultancy fees from Eisai, UCB Pharma and Jazz pharmaceuticals and has served on an advisory board for GW Pharmaceuticals. SL has received speaker's or consultancy fees from Eisai, UCB Pharma, and GW Pharmaceuticals and has served on an advisory board for GW Pharmaceuticals.