The processes of absorption, distribution, metabolism, and excretion—collectively termed drug disposition—determine the concentration of drug delivered to target effector molecules.
ABSORPTION AND BIOAVAILABILITY
When a drug is administered orally, subcutaneously, intramuscularly, rectally, sublingually, or directly into desired sites of action, the amount of drug actually entering the systemic circulation may be less than with the intravenous route (Fig. 5-2A). The fraction of drug available to the systemic circulation by other routes is termed bioavailability. Bioavailability may be <100% for two main reasons: (1) absorption is reduced, or (2) the drug undergoes metabolism or elimination prior to entering the systemic circulation. Occasionally, the administered drug formulation is inconsistent or has degraded with time; for example, the anticoagulant dabigatran degrades rapidly (over weeks) once exposed to air, so the amount administered may be less than prescribed.
When a drug is administered by a nonintravenous route, the peak concentration occurs later and is lower than after the same dose given by rapid intravenous injection, reflecting absorption from the site of administration (Fig. 5-2). The extent of absorption may be reduced because a drug is incompletely released from its dosage form, undergoes destruction at its site of administration, or has physicochemical properties such as insolubility that prevent complete absorption from its site of administration. Slow absorption rates are deliberately designed into “slow-release” or “sustained-release” drug formulations in order to minimize variation in plasma concentrations during the interval between doses.
Idealized time-plasma concentration curves after a single dose of drug. A. The time course of drug concentration after an instantaneous IV bolus or an oral dose in the one-compartment model shown. The area under the time-concentration curve is clearly less with the oral drug than the IV, indicating incomplete bioavailability. Note that despite this incomplete bioavailability, concentration after the oral dose can be higher than after the IV dose at some time points. The inset shows that the decline of concentrations over time is linear on a log-linear plot, characteristic of first-order elimination, and that oral and IV drugs have the same elimination (parallel) time course. B. The decline of central compartment concentration when drug is distributed both to and from a peripheral compartment and eliminated from the central compartment. The rapid initial decline of concentration reflects not drug elimination but distribution.
When a drug is administered orally, it must traverse the intestinal epithelium, the portal venous system, and the liver prior to entering the systemic circulation (Fig. 5-3). Once a drug enters the enterocyte, it may undergo metabolism, be transported into the portal vein, or be excreted back into the intestinal lumen. Both excretion into the intestinal lumen and metabolism decrease systemic bioavailability. Once a drug passes this enterocyte barrier, it may also be taken up into the hepatocyte, where bioavailability can be further limited by metabolism or excretion into the bile. This elimination in intestine and liver, which reduces the amount of drug delivered to the systemic circulation, is termed presystemic elimination, presystemic extraction, or first-pass elimination.
Mechanism of presystemic clearance. After drug enters the enterocyte, it can undergo metabolism, excretion into the intestinal lumen, or transport into the portal vein. Similarly, the hepatocyte may accomplish metabolism and biliary excretion prior to the entry of drug and metabolites to the systemic circulation. (Adapted by permission from DM Roden, in DP Zipes, J Jalife [eds]: Cardiac Electrophysiology: From Cell to Bedside, 4th ed. Philadelphia, Saunders, 2003. Copyright 2003 with permission from Elsevier.)
Drug movement across the membrane of any cell, including enterocytes and hepatocytes, is a combination of passive diffusion and active transport, mediated by specific drug uptake and efflux molecules. One widely studied drug transport molecule is P-glycoprotein, the product of the MDR1 gene. P-glycoprotein is expressed on the apical aspect of the enterocyte and on the canalicular aspect of the hepatocyte (Fig. 5-3). In both locations, it serves as an efflux pump, limiting availability of drug to the systemic circulation. P-glycoprotein–mediated drug efflux from cerebral capillaries limits drug brain penetration and is an important component of the blood-brain barrier.
Drug metabolism generates compounds that are usually more polar and, hence, more readily excreted than parent drug. Metabolism takes place predominantly in the liver but can occur at other sites such as kidney, intestinal epithelium, lung, and plasma. “Phase I” metabolism involves chemical modification, most often oxidation accomplished by members of the cytochrome P450 (CYP) monooxygenase superfamily. CYPs that are especially important for drug metabolism are presented in Table 5-1, and each drug may be a substrate for one or more of these enzymes. “Phase II” metabolism involves conjugation of specific endogenous compounds to drugs or their metabolites. The enzymes that accomplish phase II reactions include glucuronyl-, acetyl-, sulfo-, and methyltransferases. Drug metabolites may exert important pharmacologic activity, as discussed further below.
Clinical Implications of Altered Bioavailability
Some drugs undergo near-complete presystemic metabolism and, thus, cannot be administered orally. Nitroglycerin cannot be used orally because it is completely extracted prior to reaching the systemic circulation. The drug is, therefore, used by the sublingual or transdermal routes, which bypass presystemic metabolism.
Some drugs with very extensive presystemic metabolism can still be administered by the oral route, using much higher doses than those required intravenously. Thus, a typical intravenous dose of verapamil is 1–5 mg, compared to the usual single oral dose of 40–120 mg. Administration of low-dose aspirin can result in exposure of cyclooxygenase in platelets in the portal vein to the drug, but systemic sparing because of first-pass aspirin deacylation in the liver. This is an example of presystemic metabolism being exploited to therapeutic advantage.
Most pharmacokinetic processes, such as elimination, are first-order; that is, the rate of the process depends on the amount of drug present. Elimination can occasionally be zero-order (fixed amount eliminated per unit time), and this can be clinically important (see “Principles of Dose Selection”). In the simplest pharmacokinetic model (Fig. 5-2A), a drug bolus (D) is administered instantaneously to a central compartment, from which drug elimination occurs as a first-order process. Occasionally, central and other compartments correspond to physiologic spaces (e.g., plasma volume), whereas in others they are simply mathematical functions used to describe drug disposition. The first-order nature of drug elimination leads directly to the relationship describing drug concentration (C) at any time (t) following the bolus:
where Vc is the volume of the compartment into which drug is delivered and t1/2 is elimination half-life. As a consequence of this relationship, a plot of the logarithm of concentration versus time is a straight line (Fig. 5-2A, inset). Half-life is the time required for 50% of a first-order process to be complete. Thus, 50% of drug elimination is achieved after one drug-elimination half-life, 75% after two, 87.5% after three, etc. In practice, first-order processes such as elimination are near-complete after four–five half-lives.
In some cases, drug is removed from the central compartment not only by elimination but also by distribution into peripheral compartments. In this case, the plot of plasma concentration versus time after a bolus may demonstrate two (or more) exponential components (Fig. 5-2B). In general, the initial rapid drop in drug concentration represents not elimination but drug distribution into and out of peripheral tissues (also first-order processes), while the slower component represents drug elimination; the initial precipitous decline is usually evident with administration by intravenous but not by other routes. Drug concentrations at peripheral sites are determined by a balance between drug distribution to and redistribution from those sites, as well as by elimination. Once distribution is near-complete (four–five distribution half-lives), plasma and tissue concentrations decline in parallel.
Clinical Implications of Half-Life Measurements
The elimination half-life not only determines the time required for drug concentrations to fall to near-immeasurable levels after a single bolus, it is also the sole determinant of the time required for steady-state plasma concentrations to be achieved after any change in drug dosing (Fig. 5-4). This applies to the initiation of chronic drug therapy (whether by multiple oral doses or by continuous intravenous infusion), a change in chronic drug dose or dosing interval, or discontinuation of drug.
Drug accumulation to steady state. In this simulation, drug was administered (arrows) at intervals = 50% of the elimination half-life. Steady state is achieved during initiation of therapy after ∼5 elimination half-lives, or 10 doses. A loading dose did not alter the eventual steady state achieved. A doubling of the dose resulted in a doubling of the steady state but the same time course of accumulation. Once steady state is achieved, a change in dose (increase, decrease, or drug discontinuation) results in a new steady state in ∼5 elimination half-lives. (Adapted by permission from DM Roden, in DP Zipes, J Jalife [eds]: Cardiac Electrophysiology: From Cell to Bedside, 4th ed. Philadelphia, Saunders, 2003. Copyright 2003 with permission from Elsevier.)
Steady state describes the situation during chronic drug administration when the amount of drug administered per unit time equals drug eliminated per unit time. With a continuous intravenous infusion, plasma concentrations at steady state are stable, while with chronic oral drug administration, plasma concentrations vary during the dosing interval but the time-concentration profile between dosing intervals is stable (Fig. 5-4).
In a typical 70-kg human, plasma volume is ∼3 L, blood volume is ∼5.5 L, and extracellular water outside the vasculature is ∼20 L. The volume of distribution of drugs extensively bound to plasma proteins but not to tissue components approaches plasma volume; warfarin is one such example. By contrast, for drugs highly bound to tissues, the volume of distribution can be far greater than any physiologic space. For example, the volume of distribution of digoxin and tricyclic antidepressants is hundreds of liters, obviously exceeding total-body volume. Such drugs are not readily removed by dialysis, an important consideration in overdose.
TABLE 5-1Molecular Pathways Mediating Drug Disposition |Favorite Table|Download (.pdf) TABLE 5-1Molecular Pathways Mediating Drug Disposition
|Molecule ||Substratesa ||Inhibitorsa |
|CYP3A ||Calcium channel blockers ||Amiodarone |
| ||Antiarrhythmics (lidocaine, quinidine, mexiletine) ||Ketoconazole, itraconazole |
| ||HMG-CoA reductase inhibitors (“statins”; see text) ||Erythromycin, clarithromycin |
| ||Cyclosporine, tacrolimus ||Ritonavir |
| ||Indinavir, saquinavir, ritonavir || |
|CYP2D6b ||Timolol, metoprolol, carvedilol ||Quinidine (even at ultra-low doses) |
| ||Phenformin ||Tricyclic antidepressants |
| ||Codeine ||Fluoxetine, paroxetine |
| ||Propafenone, flecainide || |
| ||Tricyclic antidepressants || |
| ||Fluoxetine, paroxetine || |
|CYP2C9b ||Warfarin ||Amiodarone |
| ||Phenytoin ||Fluconazole |
| ||Glipizide ||Phenytoin |
| ||Losartan || |
|CYP2C19b ||Omeprazole ||Omeprazole |
| ||Mephenytoin || |
| ||Clopidogrel || |
|CYP2B6b ||Efavirenz || |
|Thiopurine S-methyltransferaseb ||6-Mercaptopurine, azathioprine || |
|N-acetyltransferaseb ||Isoniazid || |
| ||Procainamide || |
| ||Hydralazine || |
| ||Some sulfonamides || |
|UGT1A1b ||Irinotecan || |
|Pseudocholinesteraseb ||Succinylcholine || |
|P-glycoprotein ||Digoxin ||Quinidine |
| ||HIV protease inhibitors ||Amiodarone |
| ||Many CYP3A substrates ||Verapamil |
| || ||Cyclosporine |
| || ||Itraconazole |
| || ||Erythromycin |
|SLCO1B1b ||Simvastatins and some other statins || |
Clinical Implications of Drug Distribution
In some cases, pharmacologic effects require drug distribution to peripheral sites. In this instance, the time course of drug delivery to and removal from these sites determines the time course of drug effects; anesthetic uptake into the central nervous system (CNS) is an example.
For some drugs, the indication may be so urgent that administration of “loading” dosages is required to achieve rapid elevations of drug concentration and therapeutic effects earlier than with chronic maintenance therapy (Fig. 5-4). Nevertheless, the time required for true steady state to be achieved is still determined only by the elimination half-life.
Rate of Intravenous Administration
Although the simulations in Fig. 5-2 use a single intravenous bolus, this is usually inappropriate in practice because side effects related to transiently very high concentrations can result. Rather, drugs are more usually administered orally or as a slower intravenous infusion. Some drugs are so predictably lethal when infused too rapidly that special precautions should be taken to prevent accidental boluses. For example, solutions of potassium for intravenous administration >20 mEq/L should be avoided in all but the most exceptional and carefully monitored circumstances. This minimizes the possibility of cardiac arrest due to accidental increases in infusion rates of more concentrated solutions.
Transiently high drug concentrations after rapid intravenous administration can occasionally be used to advantage. The use of midazolam for intravenous sedation, for example, depends upon its rapid uptake by the brain during the distribution phase to produce sedation quickly, with subsequent egress from the brain during the redistribution of the drug as equilibrium is achieved.
Similarly, adenosine must be administered as a rapid bolus in the treatment of reentrant supraventricular tachycardias (Chap. 276) to prevent elimination by very rapid (t1/2 of seconds) uptake into erythrocytes and endothelial cells before the drug can reach its clinical site of action, the atrioventricular node.
Clinical Implications of Altered Protein Binding
Many drugs circulate in the plasma partly bound to plasma proteins. Since only unbound (free) drug can distribute to sites of pharmacologic action, drug response is related to the free rather than the total circulating plasma drug concentration. In chronic kidney or liver disease, protein binding may be decreased and thus drug actions increased. In some situations (myocardial infarction, infection, surgery), acute phase reactants transiently increase drug binding and thus decrease efficacy. These changes assume the greatest clinical importance for drugs that are highly protein-bound since even a small change in protein binding can result in large changes in free drug; for example, a decrease in binding from 99% to 98% doubles the free drug concentration from 1% to 2%. For some drugs (e.g., phenytoin), monitoring free rather than total drug concentrations can be useful.
Drug elimination reduces the amount of drug in the body over time. An important approach to quantifying this reduction is to consider that drug concentrations at the beginning and end of a time period are unchanged and that a specific volume of the body has been “cleared” of the drug during that time period. This defines clearance as volume/time. Clearance includes both drug metabolism and excretion.
Clinical Implications of Altered Clearance
While elimination half-life determines the time required to achieve steady-state plasma concentration (Css), the magnitude of that steady state is determined by clearance (Cl) and dose alone. For a drug administered as an intravenous infusion, this relationship is:
When drug is administered orally, the average plasma concentration within a dosing interval (Cavg,ss) replaces Css, and the dosage (dose per unit time) must be increased if bioavailability (F) is less than 1:
Genetic variants, drug interactions, or diseases that reduce the activity of drug-metabolizing enzymes or excretory mechanisms lead to decreased clearance and, hence, a requirement for downward dose adjustment to avoid toxicity. Conversely, some drug interactions and genetic variants increase the function of drug elimination pathways, and hence, increased drug dosage is necessary to maintain a therapeutic effect.
Metabolites may produce effects similar to, overlapping with, or distinct from those of the parent drug. Accumulation of the major metabolite of procainamide, N-acetylprocainamide (NAPA), likely accounts for marked QT prolongation and torsades des pointes ventricular tachycardia (Chap. 276) during therapy with procainamide. Neurotoxicity during therapy with the opioid analgesic meperidine is likely due to accumulation of normeperidine, especially in renal disease.
Prodrugs are inactive compounds that require metabolism to generate active metabolites that mediate the drug effects. Examples include many angiotensin-converting enzyme (ACE) inhibitors, the angiotensin receptor blocker losartan, the antineoplastic irinotecan, the anti-estrogen tamoxifen, the analgesic codeine (whose active metabolite morphine probably underlies the opioid effect during codeine administration), and the antiplatelet drug clopidogrel. Drug metabolism has also been implicated in bioactivation of procarcinogens and in generation of reactive metabolites that mediate certain adverse drug effects (e.g., acetaminophen hepatotoxicity, discussed below).
THE CONCEPT OF HIGH-RISK PHARMACOKINETICS
When plasma concentrations of active drug depend exclusively on a single metabolic pathway, any condition that inhibits that pathway (be it disease-related, genetic, or due to a drug interaction) can lead to dramatic changes in drug concentrations and marked variability in drug action. This problem of high-risk pharmacokinetics is especially pronounced in two settings. First, variability in bioactivation of a prodrug can lead to striking variability in drug action; examples include decreased CYP2D6 activity, which prevents analgesia by codeine, and decreased CYP2C19 activity, which reduces the antiplatelet effects of clopidogrel. The second setting is drug elimination that relies on a single pathway. In this case, inhibition of the elimination pathway by genetic variants or by administration of inhibiting drugs leads to marked elevation of drug concentration and, for drugs with a narrow therapeutic window, an increased likelihood of dose-related toxicity. Individuals with loss-of-function alleles in CYP2C9, responsible for metabolism of the active S-enantiomer of warfarin, appear to be at increased risk for bleeding. When drugs undergo elimination by multiple-drug metabolizing or excretory pathways, absence of one pathway (due to a genetic variant or drug interaction) is much less likely to have a large impact on drug concentrations or drug actions.
PRINCIPLES OF PHARMACODYNAMICS
For drugs used in the urgent treatment of acute symptoms, little or no delay is anticipated (or desired) between the drug-target interaction and the development of a clinical effect. Examples of such acute situations include vascular thrombosis, shock, or status epilepticus.
For many conditions, however, the indication for therapy is less urgent, and a delay between the interaction of a drug with its pharmacologic target(s) and a clinical effect is clinically acceptable. Common pharmacokinetic mechanisms that can contribute to such a delay include slow elimination (resulting in slow accumulation to steady state), uptake into peripheral compartments, or accumulation of active metabolites. Another common explanation for such a delay is that the clinical effect develops as a downstream consequence of the initial molecular effect the drug produces. Thus, administration of a proton pump inhibitor or an H2-receptor blocker produces an immediate increase in gastric pH but ulcer healing that is delayed. Cancer chemotherapy similarly produces delayed therapeutic effects.
Drug Effects May Be Disease Specific
A drug may produce no action or a different spectrum of actions in unaffected individuals compared to patients with underlying disease. Further, concomitant disease can complicate interpretation of response to drug therapy, especially adverse effects. For example, high doses of anticonvulsants such as phenytoin may cause neurologic symptoms, which may be confused with the underlying neurologic disease. Similarly, increasing dyspnea in a patient with chronic lung disease receiving amiodarone therapy could be due to drug, underlying disease, or an intercurrent cardiopulmonary problem. Thus, the presence of chronic lung disease may argue against the use of amiodarone.
While drugs interact with specific molecular receptors, drug effects may vary over time, even if stable drug and metabolite concentrations are maintained. The drug-receptor interaction occurs in a complex biologic milieu that can vary to modulate the drug effect. For example, ion channel blockade by drugs, an important anticonvulsant and antiarrhythmic effect, is often modulated by membrane potential, itself a function of factors such as extracellular potassium or local ischemia. Receptors may be up- or downregulated by disease or by the drug itself. For example, β-adrenergic blockers upregulate β-receptor density during chronic therapy. While this effect does not usually result in resistance to the therapeutic effect of the drugs, it may produce severe agonist-mediated effects (such as hypertension or tachycardia) if the blocking drug is abruptly withdrawn.
PRINCIPLES OF DOSE SELECTION
The desired goal of therapy with any drug is to maximize the likelihood of a beneficial effect while minimizing the risk of adverse effects. Previous experience with the drug, in controlled clinical trials or in postmarketing use, defines the relationships between dose or plasma concentration and these dual effects (Fig. 5-1) and has important implications for initiation of drug therapy:
The target drug effect should be defined when drug treatment is started. With some drugs, the desired effect may be difficult to measure objectively, or the onset of efficacy can be delayed for weeks or months; drugs used in the treatment of cancer and psychiatric disease are examples. Sometimes a drug is used to treat a symptom, such as pain or palpitations, and here it is the patient who will report whether the selected dose is effective. In yet other settings, such as anticoagulation or hypertension, the desired response can be repeatedly and objectively assessed by simple clinical or laboratory tests.
The nature of anticipated toxicity often dictates the starting dose. If side effects are minor, it may be acceptable to start chronic therapy at a dose highly likely to achieve efficacy and down-titrate if side effects occur. However, this approach is rarely, if ever, justified if the anticipated toxicity is serious or life-threatening; in this circumstance, it is more appropriate to initiate therapy with the lowest dose that may produce a desired effect. In cancer chemotherapy, it is common practice to use maximum-tolerated doses.
The above considerations do not apply if these relationships between dose and effects cannot be defined. This is especially relevant to some adverse drug effects (discussed in further detail below) whose development are not readily related to drug dose.
If a drug dose does not achieve its desired effect, a dosage increase is justified only if toxicity is absent and the likelihood of serious toxicity is small.
Assuming the diagnosis is correct and the correct drug is prescribed, explanations for failure of efficacy include drug interactions, noncompliance, or unexpectedly low drug dosage due to administration of expired or degraded drug. These are situations in which measurement of plasma drug concentrations, if available, can be especially useful. Noncompliance is an especially frequent problem in the long-term treatment of diseases such as hypertension and epilepsy, occurring in ≥25% of patients in therapeutic environments in which no special effort is made to involve patients in the responsibility for their own health. Multidrug regimens with multiple doses per day are especially prone to noncompliance.
Monitoring response to therapy, by physiologic measures or by plasma concentration measurements, requires an understanding of the relationships between plasma concentration and anticipated effects. For example, measurement of QT interval is used during treatment with sotalol or dofetilide to avoid marked QT prolongation that can herald serious arrhythmias. In this setting, evaluating the electrocardiogram at the time of anticipated peak plasma concentration and effect (e.g., 1–2 h postdose at steady state) is most appropriate. Maintained high vancomycin levels carry a risk of nephrotoxicity, so dosages should be adjusted on the basis of plasma concentrations measured at trough (predose). Similarly, for dose adjustment of other drugs (e.g., anticonvulsants), concentration should be measured at its lowest during the dosing interval, just prior to a dose at steady state (Fig. 5-4), to ensure a maintained therapeutic effect.
Concentration of Drugs in Plasma as a Guide to Therapy
Factors such as interactions with other drugs, disease-induced alterations in elimination and distribution, and genetic variation in drug disposition combine to yield a wide range of plasma levels in patients given the same dose. Hence, if a predictable relationship can be established between plasma drug concentration and beneficial or adverse drug effect, measurement of plasma levels can provide a valuable tool to guide selection of an optimal dose, especially when there is a narrow range between the plasma levels yielding therapeutic and adverse effects. Monitoring is commonly used with certain types of drugs including many anticonvulsants, antirejection agents, antiarrhythmics, and antibiotics. By contrast, if no such relationship can be established (e.g., if drug access to important sites of action outside plasma is highly variable), monitoring plasma concentration may not provide an accurate guide to therapy (Fig. 5-5A).
A. The efflux pump P-glycoprotein excludes drugs from the endothelium of capillaries in the brain and so constitutes a key element of the blood-brain barrier. Thus, reduced P-glycoprotein function (e.g., due to drug interactions or genetically determined variability in gene transcription) increases penetration of substrate drugs into the brain, even when plasma concentrations are unchanged. B. The graph shows an effect of a β1-receptor polymorphism on receptor function in vitro. Patients with the hypofunctional variant (red) may display lesser heart-rate slowing or blood pressure lowering on exposure to a receptor blocking agent.
The common situation of first-order elimination implies that average, maximum, and minimum steady-state concentrations are related linearly to the dosing rate. Accordingly, the maintenance dose may be adjusted on the basis of the ratio between the desired and measured concentrations at steady state; for example, if a doubling of the steady-state plasma concentration is desired, the dose should be doubled. This does not apply to drugs eliminated by zero-order kinetics (fixed amount per unit time), where small dosage increases will produce disproportionate increases in plasma concentration; examples include phenytoin and theophylline.
An increase in dosage is usually best achieved by changing the drug dose but not the dosing interval (e.g., by giving 200 mg every 8 h instead of 100 mg every 8 h). However, this approach is acceptable only if the resulting maximum concentration is not toxic and the trough value does not fall below the minimum effective concentration for an undesirable period of time. Alternatively, the steady state may be changed by altering the frequency of intermittent dosing but not the size of each dose. In this case, the magnitude of the fluctuations around the average steady-state level will change—the shorter the dosing interval, the smaller the difference between peak and trough levels.