Chapter 236. Drug Interactions

The topic of drug interactions is strikingly complex and rapidly evolving. Even experts in this area of pharmacology must rely on print and electronic resources pertaining to drug interactions on a regular basis. With these facts in mind, this chapter focuses on a wide variety of principles, with selected drug interactions utilized to illustrate these principles, as opposed to providing extensive lists of potential drug interactions. Furthermore, there will be an attempt to provide a stratification of risk, putting the greatest emphasis on the drug interactions with the greatest risk to individual patients.

Definitions

The most basic definition of a drug interaction is when two drugs are being administered simultaneously in a patient, and one drug alters the other's serum/tissue levels or mechanism of action. The drug interaction can affect the efficacy or increase the likelihood of adverse effects from one or both drugs. Furthermore, a drug interaction can occur without clinically evident alteration of efficacy or adverse effects. In contrast, drug interactions which induce either a loss of drug efficacy or new adverse effects are known as adverse drug interactions. The topic of drugs interactions is important to physicians in all aspects of medicine regardless if the clinician (a) prescribes the drug “responsible” for a given drug interaction, or (b) the drug previously prescribed by this same clinician is the “victim” of a drug prescribed by another clinician.

Print and Electronic Resources

In reviewing this chapter, the reader is encouraged to strive for a “recognition recall” level of memory; no one can possibly master and retain all important drug interactions. Through learning key principles that assist in attaining the broadest possible understanding of the multitudes of potential drug interactions, a clinician may best be able to interpret and react to new clinician situations involving potential drug interactions. One should always attempt make things “make sense” using the general principles that follow. It is always acceptable to (1) call a drug information pharmacist, (2) call or e-mail an expert in drug interactions, or (3) look up possible interactions in various print and electronic resources. Selected resources for drug interactions are listed in Table 236-1.

P-Glycoprotein

The topic of P-glycoprotein and the role in drug interactions has been relatively recently documented. P-glycoprotein is most common at important points of entry into body or important body structures (gastrointestinal tract, blood brain barrier) as a means to ensure protection against various “toxins”. P-glycoprotein is important for a number of drugs with highly variable absorption, such as digoxin and cyclosporine. Although the topic is not discussed further in this chapter, the interested reader can find additional information in a review by Shapiro and Shear.1

Definitions

All drug interactions can be divided into either pharmacokinetic or pharmacodynamic interactions, which is true regardless of whether they are an adverse drug interaction or an interaction without loss of drug efficacy or the onset of new toxicity.2 Conceptually pharmacokinetics deal with “what the body does to the drug”, while pharmacodynamics deals with “what the drug does to the body”. In the most basic sense, these definitions relate to the normal pharmacology of a given drugs therapeutic effects, as well as in the clinical setting of drug interactions.

Various Subtypes of Pharmacokinetic Drug Interactions

Pharmacokinetic drug interactions are much more common than pharmacodynamic drug interactions. This category of interactions results from alterations in a drug's serum and/or tissue levels. The ADME model (see Sections “Various Subtypes of Pharmacodynamic Interactions” and “Subtypes of Metabolic Drug Interactions”) of studying pharmacokinetics pertains to this category of interactions. The steps from entry of the drug into the body until excretion of the drug and/or its metabolites includes (1) absorption, (2) distribution, (3) metabolism, and (4) excretion. The importance of each of these steps is detailed elsewhere.2 Examples of drug interactions involving each of these steps are included in Table 236-2.

Various Subtypes of Pharmacodynamic Interactions

Pharmacodynamic drug interactions are much less common than pharmacokinetic drug interactions. This category of interactions deals with alterations in the mechanism of either the desired pharmacologic effect or mechanism for an adverse effect. The drug levels are not of central importance in this category of interactions. The two primary subtypes of interactions include (1) agonist—similar “polarity” of the pharmacologic effects for the two interacting drugs, and (2) antagonist—opposite “polarity” of the pharmacologic effects for the two interacting drugs. Both agonist and antagonist interactions can produce either a positive (improved therapeutic response) or negative (toxicity) biologic response. Selected examples to illustrate this innately confusing terminology are listed in Table 236-3. Conceptually the various aspects of drug pharmacodynamics occur between the “distribution” and “metabolism” steps of the pharmacokinetics sequence. An exception to this principle would be for any prodrug that requires metabolic conversion to the active drug form to generate a pharmacologic response. Two examples would be the conversion of valacyclovir (prodrug) to acyclovir (corresponding active drug) and of famciclovir (prodrug) to penciclovir (corresponding active drug).

Subtypes of Metabolic Drug Interactions

The four steps of the “ADME” model for drug pharmacokinetics all have potential drug interactions, with either loss of efficacy or toxicity. By far the greatest number of important drug interactions occurs at the “metabolism” step. Drugs need to be relatively lipophilic to the entire body and travel to the site(s) of intended pharmacologic effect. Subsequently drugs must be converted to relatively hydrophilic metabolites to be excreted in bile or urine. Only a small fraction of drugs are excreted in their active form.

The body has two categories of drug metabolizing enzymes to accomplish the biotransformation of a relatively lipophilic drug form to a more hydrophilic form in order to facilitate renal or biliary excretion. These categories are phase I and phase II metabolic enzymes. It is important to note that these phases do not necessarily occur in a set sequence from phase I to phase II. Phase I enzymes are primarily cytochrome P-450 (CYP) mixed function oxidative enzymes which typically create a somewhat more polar site of attachment for the subsequent phase II (conjugation) enzymes. Examples of the phase I and phase II enzymes which are of greatest importance to drug interactions are listed in eTable 236-3.1 and eTable 236-3.2. Several of these enzymes have polymorphisms of importance to explaining the spectrum of risk between various patients in various ethnic groups.3

Isoforms of CYP of Greatest Importance to Drug Interactions

The importance of cytochrome P-450 (CYP) 3A4 and CYP 2D6 lies in the fact that these isoforms metabolize about 50% and 20%–25% of all drugs, respectively. Due to some overlap of drug substrates that are metabolized by both CYP isoforms, roughly two-thirds of all drugs are metabolized by at least one of these two pathways. Other CYP isoforms of greatest potential importance to drug interactions include CYP 2C9, CYP 2C19, CYP 1A2, and CYP 2E1.

Drug Substrates

A substrate is basically a drug in the process of being metabolized to a more water-soluble form (metabolite) to enable excretion. A substrate may “travel” down one or more than one metabolic pathway, including possibly more than one CYP isoform. The metabolic pathway may be either a CYP isoform, a non-CYP pathway (phase II), or both. Furthermore, a substrate may also be an inhibitor or inducer of the drug's metabolic pathway. An example would be cyclosporine, which is both a substrate and inhibitor of CYP 3A4.

Warfarin has both an S-enantiomer (“super” warfarin—more potent) and R-enantiomer (less potent). The S-enantiomer is metabolized predominantly by CYP 2C9, while CYP 3A4 and CYP 1A2 metabolize the R-enantiomer. Drug interactions involving the S-enantiomer are overall more important because of (a) the greater “potency” of the S-enantiomer, and (b) the fact that it is metabolized only by one pathway (thus, no alternative metabolic pathways), should a CYP 2C9 enzyme inhibitor be utilized in the same patient.

Substrates of greatest interest for three of the most important CYP isoforms are listed in Table 236-4. References listed in Table 236-1 can be consulted for more comprehensive lists of CYP isoforms and pertinent drug substrates.

Substrates Which Are Narrow Therapeutic Index Drugs

A given drug's “therapeutic index” is merely the ratio of the drug dose producing the desired therapeutic response versus the drug dose producing toxicity. A drug with a “narrow” therapeutic index would represent drugs for which the ratio is relatively close to the number “1” regardless of which is the numerator and denominator. Several examples of relatively narrow therapeutic index drugs are listed in Table 236-5.

Enzyme Inhibitors

An enzyme “inhibitor” is a drug that slows/decreases the rate of metabolism of another drug substrate. Visualizing this process in a traffic analogy, the inhibitor blocks the flow of traffic (drug metabolism) on a shared road (CYP isoform or a phase II metabolic pathway). Drug interactions involving enzyme inhibitors are characterized by three generalizations. The net result is (1) immediate, (2) resultant increased substrate drug levels, and (3) resultant drug toxicity. In these interactions, the risk of toxicity is greatest with narrow therapeutic index drugs such as methotrexate, cyclosporine, and warfarin. In addition, drug substrates with just a single metabolic pathway are at greater risk for toxicity than similar drugs with more than one metabolic pathway. In the above traffic analogy, conceptually the backed up traffic can be diverted to an alternative highway if the substrate can be metabolized by another metabolic pathway (CYP isoform or phase II conjugation enzyme).

Two of the strongest inhibitors of CYP 3A4 are erythromycin (macrolide antibiotic) and ketoconazole (azole antifungal agent). In each of these drug groups, there are alternatives that have little or no inhibitory effect on the CYP isoform involved in the potential drug interaction of interest. Examples of variations of enzyme inhibition within a drug group are listed for inhibitors of CYP 3A4, CYP 2D6, and CYP 2C9 in eTable 236-5.1.

Enzyme Inducers

An enzyme “inducer” is a drug that speeds up the metabolism of a drug substrate. The inducer facilitates traffic on the “highway” (metabolic pathway) for drug metabolism. The generalizations involving enzyme inducers result in (1) delayed effects, (2) lowered drug levels, and (3) loss of drug efficacy. Drug interactions involving enzyme induction are delayed somewhat (typically up to a week or more) because they require synthesis of new protein (most notably a synthesis of a given CYP isoform). Conceptually from a traffic analogy standpoint, this synthesis would equate to creating a new lane that once opened would facilitate more rapid and efficient transport of vehicles (drug substrates).

Most CYP enzyme inducers affect more than one CYP isoform. Some of the strongest enzyme inducers include rifampin and the aromatic anticonvulsants phenytoin, phenobarbital, and carbamazepine. Griseofulvin is also an enzyme inducer, albeit a relatively weak inducer. Some clinical scenarios pertinent to dermatology of enzyme induction of significant importance are listed in Table 236-6.

Genetic Polymorphisms3

The term “polymorphism” as it pertains to drug metabolism merely represents a mutation in a phase I or phase II enzyme that results in clinical importance differences of enzyme activity compared to normal. Conceptually “variability” results in a unimodal distribution (a single bell-shaped curve), whereas a genetic polymorphism creates a bimodal distribution (at least two bell-shaped curves) when putting enzyme activity on a graph. Clinically significant polymorphisms involving drug metabolizing enzymes most notably include both phase I enzymes (CYP 2D6 and CYP 2C9) and phase II enzymes (N-acetyl transferase 2 and thiopurine methyltransferase). These polymorphisms can result in (1) drug interactions, (2) altered metabolism of a drug used as monotherapy, (3) cutaneous drug reactions or (4) systemic drug interactions. Table 236-7 summarizes examples of the clinical outcomes in these three categories from the above four drug metabolic enzyme polymorphisms.

Varying Risk of Individual Drug Interactions

Not all drug interactions are “adverse” drug interactions; these pharmacokinetic and pharmacodynamic interactions do not result in an adequate change in drug levels or mechanisms of action to have clinical importance. Furthermore, only a small percentage of drug interactions have a life-threatening potential. At the conclusion of the chapter, twelve of the most potentially serious drug interactions pertinent to dermatology are highlighted. In between these two extremes is where the majority of drug interactions exist. In these clinical settings, the clinician has a great deal of flexibility concerning therapeutic choices. From these three categories, one can derive one additional traffic analogy:

• Adverse drug interactions with life-threatening potential (red light)—do not proceed (e.g., fluconazole and warfarin; the clinician must choose an alternative antifungal agent such as itraconazole or terbinafine).
• Adverse drug interactions of clinical importance, but not life threatening (yellow light)—proceed with caution [see examples in Section “Therapeutic Options with “Proceed with Caution” (Yellow Light) Clinical Scenario”].
• Other drug interactions of no clinical risk due to minor effect of “interaction” on drug levels or mechanisms of action (green light)—can proceed with the drug combination and can reassure the patient.

Therapeutic Options with “Proceed with Caution” (Yellow Light) Clinical Scenario

Typically, there is substantial latitude concerning these clinical scenarios. The clinician may choose to avoid the drug combination altogether if there are suitable drug alternatives or if the disease being treated presents minimal risk to the patient. Alternatively, the clinician may choose one or several of the following options: (1) lower the drug dose—azathioprine and allopurinol, for which a 75% reduction of azathioprine dose is recommended, (2) monitor closely from a clinical standpoint—doxepin and an SSRI antidepressant such as paroxetine or fluoxetine, following the level of sedation closely, and/or (3) appropriate laboratory testing—can order an AM doxepin serum levels when the drug is administered with CYP 2D6 inhibitors such as paroxetine or fluoxetine. Concerning the doxepin/SSRI interaction described above, the clinician may choose to use an SSRI antidepressant alternative that does not inhibit the CYP 2D6 pathway (e.g., citalopram, escitalopram, or sertraline). From a proactive standpoint, baseline genetic testing of either the CYP 2D6 and CYP 2C19 polymorphisms are readily available. This genetic testing may determine in advance the rate of metabolism by the specific CYP isoform in that individual. Possibilities with genetic testing include the following (listed in descending order from the most rapid to the slowest rate of drug metabolism):

• Ultrarapid metabolizer (URM)—uncommon, may involve gene “duplication”.
• Extensive metabolizer (EM)—most common subset, homozygous for the “wild type” (full metabolic activity) gene.
• Intermediate metabolizer (IM)—relatively frequent, heterozygous for the “wild type” gene.
• Poor metabolizer (PM)—also uncommon, homozygous for the gene mutation (low/absent metabolic activity).

Some additional examples of the “proceed with caution” scenarios illustrating these therapeutic options are listed in Table 236-8.

Lessons to Be Learned from Drugs Taken off the Market

There are some important lessons to be learned from drugs interactions that produced life-threatening risk, which led to several drugs being taken off the market. Two issues are worth highlighting concerning the drug interactions, which led to drug removal. In both cases, a known CYP enzyme inhibitor (CYP 3A4 inhibitors) was commonly coprescribed with a CYP 3A4 drug substrate with a narrow therapeutic index. Secondly, in each case, the US Food and Drug Administration published warnings in the form of “Dear Doctor” letters (subsequently known as “Dear Health Care Professional” letters) that in retrospect did not bring about an adequate change of clinician prescribing habits to warrant drug continuation. Several of the potentially life-threatening drug interactions leading to drug withdrawal include the following:

• H1 antihistamines terfenadine and astemizole—a form of ventricular tachycardia (torsades de pointes) seen with (strong CYP 3A4 inhibitors erythromycin or ketoconazole commonly coprescribed with these drugs).
• HMG CoA reductase inhibitor cerivastatin (a “statin”)—rhabdomyolysis (a strong CYP 3A4 inhibitor gemfibrozil commonly coprescribed with this drug).

There are numerous potential drug interactions for the clinician to consider. When trying to pick a starting point to understand adverse drug interactions, it is logical to start with the most potentially serious drug interactions. With this principle in mind, I have compiled a list of a dozen drug interactions which are potentially serious, and thus, of great importance to practicing dermatologists and other clinicians utilizing these drugs (Table 236-9).

With proper awareness of drugs for which potential interactions need to be considered and regular usage of drug interaction databases, clinicians can reduce the incidence of adverse drug interactions. Patient education is likewise essential to minimizing this risk. In the vast majority of cases, a less risky therapeutic alternative is typically available if the print/electronic resources indicate a potential drug interaction. Finally, when in doubt, do not prescribe the drug or drugs in question until a definitive answer on the safety of coprescribing the medications in question can be answered.