Cytochrome P450 (CYP) enzymes may be termed an "overnight discovery" a billion years in the making (Table 7.1). Only recently have we begun to understand the important role these enzymes play in determining a patient's response to pharmacotherapy. The inhibition of specific CYP enzymes is also the major distinguishing characteristic among SSRIs. Such inhibition produces the potential for specific pharmacokinetic drug interactions between specific SSRIs and concomitantly administered drugs dependent on specific CYP enzymes for their elimination. This section will provide the background information of how our understanding of these enzymes has evolved and their significance relative to the optimal care of patients.

Our knowledge of oxidative metabolism in the body and the role played by CYP enzymes has been relatively short, but has rapidly expanded, particularly over the last 10 years (Table 7.2). It was only about 100 years ago that drug metabolism was generally accepted to occur in the body. The pigments we now recognize to be multiple, different enzymes were identified approximately 50 years ago and named cytochrome P450 (CYP) due to their ability to absorb light at a frequency of 450 nm. The first gene to code for a specific CYP enzyme was isolated approximately 10 years ago. Due to the advances made possible by molecular biology, we can now study the effects of specific CYP enzymes on specific drugs and vice versa. We are currently "backfilling" our knowledge concerning which CYP enzymes are responsible for the metabolism of drugs.

Although our knowledge is relatively recent, the ancestral CYP enzyme from which all of the current ones evolved came into existence over 1 billion years ago, which underscores the biological importance of these enzymes to organisms, including man.^106,107 They are heme-containing monoxygensase enzymes responsible for much of the oxidative metabolism occurring in the body.¹⁰⁵

Once the genes that code for specific CYP enzymes were isolated, they could be used to produce the enzymes in purified form - their amino acid sequence could then be determined. This has now been accomplished for all of the human CYP enzymes.

TABLE 7.2 — History of Our Knowledge of Biotransforming Enzymes Knowledge Time Concept of entgiftung (detoxification)    was appreciated Correlation of some drug effects with    in vitro metabolizing enzyme activities Induction of such enzymes possible Membrane bound cytochrome-CO complex    discovered, named "cytochrome P450,"    and biochemically characterized Three hundred substances identified to    induce their metabolism or that of    foreign and endogenous substances Role of mRNA and protein synthesis    identified in induction process and    purification of multiple P450 proteins    accomplished, making polyclonal and    monoclonal antibody studies possible Identification of specific P450 genes using    DNA cloning techniques End of the 19th century 1920s-40s 1940s-50s 1950s-60s Mid-1960s 1970s Mid-1980s to present Reference: 218

TABLE 7.3 — Human CYP Enzymes as Classified by Family, Subfamily and Gene* 1A1 1A2 2A6 2A7 2B6 2C8 2C9 2C18 2C19 2D6 2E1 2F1 3A3/4 3A5 3A7 4A9 4B1 4F2 4F3 7 11A1 11B1 11B2 17 19 21A2 27 * Key to classification: 1) The first Arabic numeral represents the family, 2) the alphabetic letter represents the subfamily, and 3) the second Arabic numeral represents the individual gene within the subfamily. From reference: 231

TABLE 7.4 — Two General Classes of CYP Enzymes Families 1, 2, 3, 4 “Xenobiotic” CYP Enzymes • Evolved from steroidogenic CYP enzymes during the period of plant-animal differentiation • Co-evolved with transferase enzymes • Degrade dietary toxins produced by plants • Biotransforms drugs to permit elimination Families 7, 11, 17, 19, 21, 27 “Steroidogenic” CYP Enzymes • For membrane integrity of single cell organisms • For hormonal mediators of the development of differentiated organisms • With rigid substrate and product specificity Data from references: 105, 115, 116, 183-185, 187

TABLE 7.5 — What Are The Functions of CYP Enzymes? • Steroidogenic: conversion of consumed substances into biologically needed constituents:     – Steroids     – Bile acids     – Cholesterol     – Prostaglandin biosynthesis • Xenobiotic: detoxification of consumed substances:     – Toxins     – Carcinogens     – Drugs     – Mutagens Data from references: 105, 115, 116, 183-185, 187

FIGURE 7.1 — Drug Metabolism by CYP Enzymes Modified from reference: 282

Using this information, a classification system has been developed based on the degree of structural similarities between different enzymes (ie, amino acid sequence homology).¹⁸⁷ The rationale for such a classification system is that the more similar the structure, the closer the enzymes are both phylogenetically and functionally (ie, function follows structure). The CYP enzymes are grouped into families designated by the first number. All enzymes in a family have at least 40% amino acid sequence homology. They are further grouped into subfamilies designated by an alphabet letter. All enzymes in the same subfamily have at least 55% amino acid sequence homology. The last number designates the gene that codes for a specific enzyme. Table 7.3 shows all of the human CYP enzymes grouped into families and subfamilies.

These enzymes are divided into two major groups (Table 7.4):

• "Steroidogenic" CYP enzymes
• "Xenobiotic" CYP enzymes

The former are phylogenetically older, occurring in
even single cell organisms. They are located in the mitochondria of cells and are responsible for the synthesis of steroids and other substances necessary for the maintenance of cell wall integrity (Table 7.5). Substantial impairment in the functional integrity of the steroidogenic enzymes (eg, deficiency due to genetic mutation) is incompatible with life.

The xenobiotic CYP enzymes evolved from the steroidogenic CYP enzymes over 1 billion years ago (Table 7.4). These enzymes are located in the smooth endoplasmic reticulum of cells (Figure 7.1) and appear to have evolved during the era of plant-animal differentiation.^107,186 The term "xenobiotic" refers to the ability of these enzymes to metabolize foreign (ie, xeno) biological substances. These enzymes allowed animals who possessed them to metabolize plant toxins before they could enter the animal's systemic circulation and cause damage. Thus, these enzymes conveyed survival advantage to the animals that possessed them.

Since drugs originally came from plants and resemble plant toxins, they can be metabolized by the same enzymes. In fact, these enzymes determine in part what substances can become drugs. If a mechanism did not exist to eliminate a substance, then that substance could not be used as a drug.

As illustrated in Figure 7.1, many drugs have a principal CYP enzyme that is responsible for the bulk, if not all, of its metabolism. In this illustration, Drug A is principally metabolized by CYP 1A2, Drug B by CYP 2D6, and Drug D by CYP 3A3/4. Drug C is an example of a drug that can be metabolized by either CYP 2D6 or 1A2.

This latter situation has caused some confusion in that it has been erroneously stated that "another enzyme could take over for an inhibited enzyme." However, these enzymes do not change their enzymatic activity or affinity for a substrate based on whether the principal enzyme is inhibited. Instead, a second enzyme comes into play when the concentration of a substrate (eg, drug) such as Drug C has reached a sufficient level in the body that this pathway becomes meaningful for biotransformation and subsequent elimination.

The genetic material for these enzymes is carried in every cell in the body and expressed in multiple cells in the body. For example, CYP 2D6 is found in the brain where it is spatially linked to the dopamine uptake transporter pump.¹³⁹ However, our knowledge of the precise role of these enzymes in organs other than the liver and the bowel wall remains rudimentary.¹⁴⁷ In the liver and bowel wall, these enzymes are responsible for the bulk of phase I or oxidative metabolism of xenobiotics including dietary toxins, carcinogens, mutagens, and more recently, drugs. As mentioned above, their existence has permitted the development and use of medications in the treatment of patients (Table 7.5).

What Is Oxidative Metabolism?

Oxidative metabolism involves the conversion of a substance into a more polar species (ie, a metabolite) by the insertion or incorporation of atmospheric oxygen into the molecule (Figure 7.2). In some instances, the oxidized product is the final metabolite. In other instances, it is an intermediate metabolite and undergoes further biotransformation. CYP enzymes mediate a number of different biochemical reactions (Figure 7.3). The final result of each of these biotransformations is to produce a polar metabolite that can be eliminated in urine or feces. Often the final step is conjugation of the metabolite at a polar site with a moiety such as glucuronic acid (ie, phase II metabolism). The resulting conjugated product is water soluble and can be eliminated in the urine. This sequence from phase I to phase II metabolism and eventual elimination is illustrated in Figure 7.4.

Why Do We Have Such Enzymes?

FIGURE 7.2 — CYP Enzyme Reaction Cycle Reference: 180

FIGURE 7.3 — Diverse Monooxygenase Activities of CYP Enzymes The oxygen derived from atmospheric oxygen is denoted by boldface.

As mentioned above, the development of specific CYP enzymes had, at one time, survival value for animals by metabolizing dietary toxins before they could enter the systemic circulation. Whether these enzymes still serve such a role for man is unclear. By facilitating the biotransformation of dietary toxins, carcinogens and mutagens, these enzymes determine the total cumulative exposure to such agents in terms of both absolute concentration achieved and duration of exposure once such chemical entities are ingested.^114,117 This fact has raised the possibility that deficiency in these enzymes might be a risk factor in the development of specific diseases caused by environmental exposure to such chemicals.^53-55 For example, an active area of research is whether cigarette-induced increase in the functional activity of CYP 1A2 may explain the increased risk of some cancers in cigarette smokers;^163,181 CYP 1A1/2 transforms some procarcinogens into carcinogens.^50,121,193 Another example is the theory that individuals genetically deficient in CYP 2D6 may have an increased risk of early-onset Parkinson's disease because they lacked this enzymatic barrier in the liver to the entry of an environmental dopamine neurotoxin such as MPTP.^17,91 The primary approach in this area of research has been to study the relative risk of different illnesses as a function of genetically determined deficiency in specific enzymes.⁵⁴

What Is the Role of These Enzymes in Determining Response to Drug Therapy?

The functional state of these enzymes is the major determinant of the rate of biotransformation of the drug into metabolites and, hence, a major determinant of the elimination rate (ie, clearance) of specific drugs. As such, their activity determines the concentration of the drug and/or its metabolites that will be achieved in the body as a function of the dose administered (Figure 7.5). Their activity also determines the relative ratio of the administered drug to its metabolites. This ratio may also be important because the metabolite(s) may be biologically active and that activity may differ in clinically important ways from that of the parent drug (eg, clomipramine and desmethylclomipramine) (see Section 3).

FIGURE 7.4 — Phases of Xenobiotic Metabolism Y = moiety such as glucuronic acid

FIGURE 7.5 — Relationship Between Dosing Rate, Clearance, Steady-state Drug Concentration and Clinical Response CSS = Steady-state drug concentration

To put this matter in perspective, recall that the concentration(s) of the parent drug and/or its metabolites is one of the 3 variables that determines the magnitude and/or the nature of the patient's response to the drug treatment (Figure 7.6). Obviously, the drug must affect a site of action (SOA) in a specific way (ie, act as an agonist or antagonist at the SOA) to produce the desired effect. However, the magnitude of the effect is dependent on the concentration of the drug achieved at that SOA. In addition to these 2 variables, biological variance among different patients can shift the dose-response (ie, concentration-response) curve. For example, a drug such as propanolol slows heart rate by its negative chronotropic effect mediated by its mechanism of action (MOA) (ie, the blockade of B-adrenergic receptors). The higher the concentration of propanolol, the greater the degree of B-adrenergic receptor blockade, and the slower the heart rate. A patient with an underlying cardiac disease may experience a substantially greater effect at the same concentration (ie, their dose- or concentration-response curve has been shifted to the left due to their underlying cardiac condition).

FIGURE 7.6 — Relationship of Pharmacodynamics, Pharmacokinetics and Biological Variance in Determining Overall Result of Drug Treatment

Since the magnitude and nature of the patient's response is concentration-dependent, the functional activity of these enzymes can make the difference between therapeutic success and failure. Such therapeutic failure may take the form of either suboptimal response or toxicity. A suboptimal response can result from the development of an inadequate concentration at the SOA due to unusually rapid clearance. Toxicity can result from the development of excessively high concentrations due to unusually slow clearance. In essence, a change in the elimination rate (ie, clearance) of a drug is the mirror image of a change in its dosing rate (Figure 7.5). Parenthetically, clearance is the amount of drug cleared per unit of time. The reason is the concentration of drug achieved in the body is directly proportional to the dosing rate and is inversely proportional to its elimination rate.

This inverse relationship is illustrated in Figure 7.5. Due to this relationship, a change in the functional activity of the enzymes mediating the biotransformation of a drug necessary for its eventual elimination typically produces an effect the opposite of what would occur with a change in the dosing rate. For example, a decrease in functional activity in the CYP enzymes will lead to a decrease in drug clearance and an increase in drug accumulation (ie, the same net effect as will occur with a dose increase). Conversely, an increase in functional activity of the enzyme would typically result in a decrease in drug accumulation (ie, a net effect similar to a dose decrease).

How Do These Enzymes Relate to Drug-Drug Interactions?

Drugs can increase or decrease the functional activity of CYP enzymes. Induction of the gene responsible for the production of the enzyme increases its rate of synthesis, thus increasing the cellular content and activity of the induced CYP enzyme.145,206 Since induction involves protein synthesis, there is a delay in both its onset and its offset relative to starting and stopping the inducer. Therefore, the full effect may not be apparent for several weeks after the inducer has been started and the effect will take a similar period of time to fully dissipate after the inducer has been stopped and the rate of enzyme production has returned to baseline.

Drug-induced inhibition is usually competitive and occurs immediately.²⁷⁸ However, the magnitude of the inhibition is a function of the concentration of the inhibitor (see Section 8). Thus, the half-life of the inhibitor will determine how long it must be administered before the full effect is achieved and, conversely, how long after its discontinuation the inhibition will persist. For this reason the full effect of fluoxetine, and particularly its long-lived active metabolite norfluoxetine, on several CYP enzymes can take several weeks to be achieved and can persist for several weeks after its discontinuation.^112,219

How Do Such Drug-Drug Interactions Present Clinically?

There are two types of drug-drug interactions: pharmacodynamic and pharmacokinetic (Table 7.6).²²⁴ In the former, both drugs are acting via their MOA to produce a physiological response and where the effect of one drug on its SOA amplifies or diminishes the response produced by the effect of the other drug on its SOA. For this reason, a pharmacodynamic drug interaction is likely to produce a qualitative change in the patient's response (ie, a change in the nature of the response). Examples of a qualitative change include the central serotonin syndrome that can occur when an SSRI and a monoamine oxidase inhibitor (MAOI) are taken together. Due to the frequently dramatic presentations of pharmacodynamic drug interactions, they are often readily identified and, hence, have received considerable attention in the medical literature.

In contrast, pharmacokinetic interactions are more likely to present as a quantitative (ie, more or less) rather than a qualitative change in the patient's response (Table 7.7). These types of interactions occur when the second drug affects the pharmacokinetics of the first drug and alters the concentration of the first drug at its SOA. In contrast to pharmacodynamic drug interactions, these interactions do not involve the addition of another effect on another SOA to the treatment equation. For this reason, a pharmacokinetic drug interaction often produces the same result as would occur with a dose change and, thus, is something that can be expected when only using the affected drug. In other words, the result is a known effect of the drug whose pharmacokinetics were altered by the addition of a second drug.

TABLE 7.7 — Pharmacokinetic Interactions: How Do They Present Clinically? Generally quantitatively (ie, magnitude of the response) rather than qualitatively (ie, nature of the response) Rather than an unexpected response, such an interaction typically produces an expected response but at an unexpected dose (eg, failure to respond, adverse effect) This outcome is easily attributed to the patient being either “sensitive” or “resistant” to the affected drug May also present as an erratic response which can be dismissed as being due to noncompliance

FIGURE 7.7 — Dose-Response Curves with Risperidone Extrapyramidal Adverse Effects Data from reference: 169

FIGURE 7.8 — Dose-response Curves for Seizure Risk Clozapine1 Bupropion2 Data from references: 160, 272

However, the effect is occurring at an unexpected dose due to the change in the drug's clearance. What the physician may not realize is that adding the second drug has, in essence, changed the effective dose of the first drug by altering its clearance (Figure 7.5). Thus, the patient may appear unusually sensitive to the first drug (ie, develop a known dose-dependent adverse effect of the first drug on a dose that is usually well tolerated) or nonresponsive to the first drug (ie, failure to optimally respond to what is usually an effective dose). Thus, pharmacokinetic drug interactions may be dismissed as being due to idiosyncrasies on the part of the patient rather than being recognized for what they are.

Some examples may be useful to illustrate this concept. Risperidone is a recently marketed antipsychotic with some unique features. Figure 7.7 shows the dose-dependent nature of its antipsychotic efficacy and its risk of causing extrapyramidal adverse effects. As the dose of risperidone goes up, there is initially an increase in its antipsychotic efficacy; but somewhere between 8 to 10 mg/day, its antipsychotic efficacy either plateaus or actually begins to decrease.¹⁶⁹ However, the incidence and severity of extrapyramidal adverse effects continue to go up with dose increases. For these reasons, 6 mg/day is the recommended optimal dose. If a concomitantly administered drug decreases the clearance of risperidone, the consequence can be the same as increasing its dose. The result will be a reduction in its efficacy and an increase in the incidence of extrapyramidal adverse effects. If the clinician does not know the dose had been effectively changed by adding the inhibitor, s/he may simply conclude that the patient is resistant to the beneficial effects and sensitive to the adverse effects of risperidone. This drug is metabolized by the CYP enzyme 2D6.

The consequences can also be more serious than loss of efficacy and increased extrapyramidal side effects. Figure 7.8 shows the seizure risk due to clozapine as a function of its daily dose. If a drug inhibits the clearance of clozapine, the consequences will be similar to increasing its dose. If the patient is on 300 mg/day and the clearance rate is reduced by 50%, this will be comparable to doubling the dose. With such a functional dose increase, the patient's risk of experiencing a seizure will go from 1.5% to 5.0% (ie, over a threefold increase in risk).

Even though such an increase in the relative risk of this potentially serious adverse effect for a given patient is substantial, it will be virtually impossible to detect in clinical practice. Since the clinician knows that some patients will have seizures even on daily doses of 300 mg/day, s/he may simply conclude that the patient is "sensitive" to this adverse effect.

To prove that such a pharmacokinetic drug interaction causes the increased risk by functionally increasing the dose would require a formal study of patients randomly assigned to clozapine with and without the inhibitor. Given the relatively infrequent nature of this adverse effect, even at doses of 600 to 900 mg/day, the study would have to involve several hundred patients to be assured of having the statistical power to detect the difference in the predicted seizure risk for the two conditions (ie, 1.5% risk versus a 5% risk). Such a study will likely never be done for ethical and economic reasons. Hence, physicians will need to make decisions based on an understanding of these principles. Of note, several CYP enzymes have been implicated in the metabolism of clozapine, including CYP 1A2, 3A3/4²⁰⁸and possibly 2D6,²⁵⁵ although there is conflicting data for this CYP enzyme. Also, fluvoxamine and fluoxetine have been reported to cause moderate increases in clozapine levels,^137,300 but it is not known which enzymes they are affecting to produce this increase.

Bupropion, which is an effective antidepressant with several advantageous features,^214,220 has a dose-dependent risk of seizures (Figure 7.8). For that reason, this drug has only been marketed in the U.S.; nonetheless, it provides another good illustration of this point. Bupropion undergoes extensive biotransformation to active metabolites, but the enzymes mediating those transformations have not been established.²¹⁴ Inhibition of bupropion's clearance or the clearance of its active metabolites can increase the seizure risk just as occurs with a dose increase. Limited case data indicate that fluoxetine can produce substantial elevation of two metabolites of bupropion, although the precise mechanism for this effect has not been established.²¹⁴ Again, a large scale study of hundreds of patients would be needed to statistically test the hypothesized increased seizure risk with such a combination.

Despite the general rule being quantitative changes in response, pharmacokinetic drug interactions can sometimes produce a qualitative change in the patient's response. Two examples are provided in Table 7.8. Since qualitative changes are easier to recognize in clinical practice than are quantitative changes, these interactions are better known.

In the first type, the drug whose clearance is changed has effects on multiple SOAs in the body in a concentration-dependent manner.^221,225 Tricyclic antidepressants (TCAs) are excellent examples (Figure 7.9). At low concentrations, they block histamine and acetylcholine receptors, producing effects mediated by those actions. At therapeutic concentrations, TCAs inhibit the norepinephrine and serotonin uptake pumps believed to mediate their antidepressant efficacy. At toxic concentrations, TCAs inhibit fast sodium channels; hence, they can slow intracardiac conduction leading to serious, and even fatal, conduction disturbances. This same MOA is probably responsible for their central nervous system (CNS) toxicity (ie, delirium, seizures and coma).

If a drug inhibits the clearance of TCAs, it will have the same effect as increasing the dose of the TCA. The response to such a pharmacokinetic drug interaction may be a fatality rather than an antidepressant response.²¹⁶ That is what is meant by a change in the quality (ie, nature) of the response (eg, seizure or arrhythmia, as opposed to antidepressant efficacy) as opposed to the quantity of the response (eg, severity of dry mouth). The principal enzyme responsible for the clearance of TCAs particularly secondary amine TCAs (eg, desipramine, nortriptyline) is the CYP enzyme 2D6 which mediates the ring hydroxylation of these drugs. This conclusion is based on extensive and varied types of pharmacokinetic data.^{25,40,42,45,46,48,147,195,223,262} The N-demethylation pathway, which is more prominent for tertiary amine TCAs (eg, amitriptyline, imipramine) than for secondary amine TCAs, although still not the principal route of elimination, is mediated by several CYP enzymes including CYP 1A2, 2C19 and 3A3/4.^{42,46,58,161,168,192,202,256,257}

Pharmacokinetic drug interactions can also present qualitatively when the inhibition of the enzyme significantly changes the ratio of the parent drug to a metabolite which is pharmacologically active in a substantially different way than the parent drug. An example is the inhibition of the conversion of terfenadine to its acid metabolite. That conversion is mediated by CYP 3A3/4 and occurs in the bowel wall and liver during the absorption of terfenadine.²⁹⁵ This conversion is typically so efficient and complete that virtually all of the terfenadine is converted to its acid metabolite before entry into the systemic circulation. This metabolite is an active antihistaminic agent and is relatively devoid of effects on intracardiac conduction. Therefore, the use of terfenadine generally does not carry a risk of adverse effects on intracardiac conduction. However, terfenadine itself does inhibit intracardiac conduction and can cause arrhythmias.²⁵⁸ If the conversion of terfenadine to its acid metabolite is substantially blocked during absorption, then terfenadine will enter the systemic circulation. If its concentrations are sufficiently high, cardiac arrhythmias can result. Thus, coadministration of terfenadine with a potent inhibitor of CYP 3A3/4 can have serious, and even life-threatening, adverse consequences.²⁹⁷ Ketoconazole related antifungal agents and several macrolide antibiotics (eg, erythromycin) can substantially inhibit this enzyme.^{29,126,127,279} Fluvoxamine, norfluoxetine and the non-SSRI antidepressant, nefazodone, also inhibit this enzyme, but are substantially less potent than ketoconazole in this regard.^144,276

TABLE 7.8 — Examples of Metabolically Mediated Pharmacokinetic Drug-drug Interactions and Their Typical Clinical Presentations* Type Presentation Affected Drug Causative Drug CYP Enzyme Buildup of drug levels due to inhibition of clearance Increase in incidence and/or severity of expected dose-dependent adverse effects • Carbamazepine • Dextromethorphan • Phenytoin • Propanolol • Theophylline • Tricyclic antidepressants • Tricyclic antidepressants • Warfarin • Erythromycin • Paroxetine • Fluoxetine • Cimetidine • Fluvoxamine • Fluoxetine • Fluvoxamine • Fluvoxamine • 3A3/4 • 2D6 • 2C9/10 • 1A2, 3A3/4 • 1A2 • 2D6 • 1A2, 3A3/4 • 1A2 Reduction of drug levels due to induction of clearance Loss of efficacy • Disopyramide • Oral contraceptives • Theophylline • Propanolol • Quinidine • Theophylline • Warfarin • Phenytoin • Carbamazepine • Phenytoin • Rifampin • Phenobarbital • Phenytoin • Secobarbital • 3A3/4 • 3A3/4 • 1A2 • 3A3/4 • 3A3/4 • 1A2 • 1A2 Blockade of the production of an active metabolite Loss of efficacy • Codeine • Paroxetine • 2D6 Increased accumulation of an unusual toxic parent drug or metabolite Unexpected toxicity based on the usual pharmacology of the drug • Astemizole • Terfenadine • Ketoconazole • 3A3/4 * The above listing is in no way an exhaustive list of “causative” or “affected” drugs for either a type of interaction or for a specific CYP enzyme, but simply representative examples. Reference: 224

FIGURE 7.9 — Multiple Concentration: Response Curves of Tertiary TCAs Reference: 221

Where Are We Going With This Knowledge?

FIGURE 7.10 — How Knowledge of Drug-metabolizing Enzymes Will Simplify Understanding of Pharmacokinetic Interactions

* Could be inhibition or induction.

Two sets of knowledge are now being developed as illustrated in Figure 7.10. The first is which drugs affect which CYP enzymes, either by induction or inhibition. The second is which drugs are metabolized by a specific CYP enzyme (Table 7.9). With these two sets of knowledge, we can determine whether the coadministration of one drug is likely to affect the biotransformation of another drug. We can also determine whether the second drug is likely to alter the overall clearance (ie, elimination rate) of the first drug and hence its accumulation on a given dose and whether the second drug is likely to affect the relative ratio of specific metabolites to the parent drug by inducing or inhibiting a specific biotransformation pathway. In addition to the examples already given, the coadministration of an inhibitor or an inducer might increase the production of what is normally a minor metabolite to a substantial extent. This metabolite may mediate clinically important effects. For example, the increased risk of hepatoxicity when valproate is coadministered with another anticonvulsant (ie, carbamazepine or phenytoin) may have been due to the increased production of the 4-ene metabolite of valproate via the induction of the CYP enzyme that mediates the conversion.¹⁶⁹

TABLE 7.9—Drugs Metabolized by CYP Enzymes* CYP 1A2 Antidepressants—amitriptyline, clomipramine, imipramine Antipsychotics—clozapine B-Blockers—propanolol Miscellaneous—caffeine, paracetamol, theophylline, R-warfarin CYP 2C9/10 phenytoin, S-warfarin, tolbutamide CYP 2C19 Antidepressants—citalopram, clomipramine, imipramine Barbiturates—hexobarbital, mephobarbital, S-mephenytoin Benzodiazepines—diazepam B-Blockers—propranolol CYP 2D6 Antiarrhythmics—encainide, flecainide, mexiletine, propafenone Antipsychotics—haloperidol, perphenazine, risperidone, thioridazine B-Blockers—alprenolol, bufarolol, metoprolol, propranolol, timolol Miscellaneous—debrisoquin, 4-hydroamphetamine, perhexiline, phenformin, sparteine Opiates—codeine, dextromethorphan, ethylmorphine SSRIs—fluoxetine, N-desmethylcitalopram, paroxetine TCAs—amitriptyline, clomipramine, desipramine, imipramine, N-desmethylclomipramine, clomipramine, nortriptyline, trimipramine Other Antidepressants—venlafaxine, the mCPP metabolite of nefazodone and trazodone CYP 3A3/4 Analgesics—acetaminophen, alfentanil, codeine, dextromethorphan Antiarrhythmics—amiodarone, disopyramide, lidocaine, propafenone, quinidine Anticonvulsants—carbamazepine, ethosuximde Antidepressants—amitriptyline, clomipramine, imipramine, nefazodone, sertraline, O-desmethylvenlafaxine Antiestrogens—docetaxel, paclitaxel, tamoxifen Antihistamines—astemizole, loratadine, terfenadine Antipsychotics—clozapine Benzodiazepines—alprazolam, clonazepam, diazepam, midazolam, triazolam Calcium Channel Blockers—diltiazem, felodipine, nicardipine, nifedipine, niludipine, minodipine, nisoldipine, nitrendipine, verapamil Immunosuppressants—cyclosporine, tacrolimus (FK506—macrolide) Local Anesthetics—cocaine, lidocaine Macrolide Antibiotics—clarithromycin, erythromycin, triacetyloleandomycin Steroids—androstendione, cortisol, dihydroepiandrosterone 3-sulfate, dexamethasone, estradiol, ethinylestradiol, progesterone, testosterone Miscellaneous—benzphetamine, cisapride, dapsone, lovastatin, omeprazole (sulfonation) * Major pathway for elimination of tricylcic antidepressants is ring hydroxylation; N-desmethylation, a minor pathway, is mediated by several CYP enzymes. Adapted from references: 39, 119, 144

TABLE 7.10 — Genetically Determined CYP Enzyme Deficiency CYP Orientals Blacks Whites 2C19 20% Not known 3% 2D6 1% 4% 8% Data from references: 104, 152

In the past, this type of pharmacokinetic drug interaction could not be anticipated, but had to occur and then be identified clinically after the fact. Early pharmacokinetic drug interaction studies were done mainly on the basis of what drugs were likely to be used together rather than because there was a substantial reason to suspect a clinically meaningful drug-drug interaction. Physicians had to remember these types of interactions without having any organizing principles to aid them.

Due to improved knowledge of CYP enzymes, this situation has substantially changed. Now studies can be done by deductive reasoning using the 3 sets of information illustrated in Figure 7.10. From a research standpoint, this knowledge will permit the anticipation of clinically meaningful interactions which can then be confirmed by focused formal pharmacokinetic drug interactions studies done under clinically relevant conditions. From a clinical standpoint, this knowledge will allow physicians to anticipate such an interaction and make appropriate drug and dose selections to avoid undesired consequences.

This knowledge is also being used to determine what drugs to develop. New candidate drugs are now tested for how they are metabolized and whether they affect specific CYP enzymes. A new candidate drug may not be developed based on the results of such tests if it is principally metabolized by a CYP enzyme that is inducible or inhibited by many commonly used drugs or is genetically deficient in a sizable percentage of the population (Table 7.10). Also, a new drug may not be developed if it induces or inhibits CYP enzymes responsible for the metabolism of many other drugs.

If the unique efficacy of the new drug outweighs these considerations, then the information about how it is metabolized and its affects on the CYP enzymes will be used to help guide its safe and effective use in clinical practice. For example, this information can be entered into a readily accessible computer database; the physician can enter the patient's current treatment into a computer and the new drug which is going to be added to the regimen. Using such an information system, the physician can quickly determine:

• Whether an interaction is likely
• The nature of the interaction, including its potential seriousness
• What adjustment should be made in terms of dosage
• What drug can be substituted for the proposed new drug to avoid the interaction

Patients can also be genotyped and phenotyped for specific CYP enzyme activity, and that information can also be entered into the computer database along with their current treatment regimen and any proposed changes. This additional information will further increase the accuracy of the predictions and recommendations. Obviously, the development of this knowledge has substantial implications for improving patient care in terms of safety, tolerability and efficacy. Those implications extend well beyond any single class of medications, but nonetheless, have specific implications now for the SSRI class of antidepressants as discussed in Section 8.