Individual variation in drug responses

Individual variation in drug responses has been a recurrent theme throughout the early chapters of this text. We noted that, because of individual variation, we must tailor drug therapy to each patient. In this chapter, we discuss the major factors that can cause one patient to respond to drugs differently than another. With this information, you will be better prepared to reduce individual variation in drug responses, thereby maximizing the benefits of treatment and reducing the potential for harm. Much of this chapter is a review of information presented in previous chapters.

The impact of kidney disease is illustrated in Figure 8–1, which shows the decline in plasma levels of kanamycin (an antibiotic) following injection into two patients, one with healthy kidneys and one with renal failure. (Elimination of kanamycin is exclusively renal.) As indicated, kanamycin levels fall off rapidly in the patient with good kidney function. In this patient, the drug’s half-life is brief—only 1.5 hours. In contrast, drug levels decline very slowly in the patient with renal failure. Because of kidney disease, the half-life of kanamycin has increased by nearly 17-fold—from 1.5 hours to 25 hours. Under these conditions, if dosage is not reduced, kanamycin will quickly accumulate to dangerous levels.

Figure 8–1 Effect of renal failure on kanamycin half-life.
Kanamycin was administered at time “0” to two patients, one with healthy kidneys and one with renal failure. Note that drug levels declined very rapidly in the patient with healthy kidneys and extremely slowly in the patient with renal failure, indicating that renal failure greatly reduced the capacity to remove this drug from the body. (T_1/2 = half-life.)

Liver disease

Like kidney disease, liver disease can cause drugs to accumulate. Recall that the liver is the major site of drug metabolism. Hence, if liver function declines, rates of metabolism will decline too, and drug levels will climb. Accordingly, to prevent accumulation to toxic levels, dosage must be reduced if liver disease develops. Of course, this guideline applies only to those drugs that are eliminated primarily by the liver, not to drugs that are eliminated by the kidneys and other nonhepatic routes.

Acid-base imbalance

By altering pH partitioning (see Chapter 4), changes in acid-base status can alter the absorption, distribution, metabolism, and excretion of drugs.

Figure 8–2 illustrates the impact of altered acid-base status on drug distribution. Specifically, it shows the results of altered acid-base status on the distribution of phenobarbital (a weak acid) in a dog. The upper curve shows plasma levels of phenobarbital. The lower curve shows plasma pH. Acid-base status was altered by having the dog inhale a mixture of gas rich in carbon dioxide (CO₂), thereby causing respiratory acidosis. In the figure, acidosis is indicated by the drop in plasma pH. Note that the decline in pH is associated with a parallel drop in levels of phenobarbital. Upon discontinuation of CO₂, plasma pH returned to normal and phenobarbital levels moved upward.

Figure 8–2 Altered drug distribution in response to altered plasma pH.
*Lower curve,* Plasma (extracellular) pH. Note the decline in pH in response to inhalation of CO₂. *Upper curve,* Plasma levels of phenobarbital. Note the decline in plasma drug levels during the period of extracellular acidosis. This decline results from the redistribution of phenobarbital into cells. (See text for details.) (Redrawn from Waddell WJ, Butler TC: The distribution and excretion of phenobarbital. J Clin Invest 36:1217, 1957.)

Why did acidosis alter plasma levels of phenobarbital? Recall that, because of pH partitioning, if there is a difference in pH on two sides of a membrane, a drug will accumulate on the side where the pH most favors its ionization. Hence, because acidic drugs ionize in alkaline media, acidic drugs will accumulate on the alkaline side of the membrane. Conversely, basic drugs will accumulate on the acidic side. Since phenobarbital is a weak acid, it tends to accumulate in alkaline environments. Accordingly, when the dog inhaled CO₂, causing extracellular pH to decline, phenobarbital left the plasma and entered cells, where the environment was less acidic (more alkaline) than in plasma. When CO₂ administration ceased and plasma pH returned to normal, the pH partitioning effect caused phenobarbital to leave cells and re-enter the blood, causing blood levels to rise.

Altered electrolyte status

Electrolytes (eg, potassium, sodium, calcium, magnesium) have important roles in cell physiology. Consequently, when electrolyte levels become disturbed, multiple cellular processes can be disrupted. Excitable tissues (nerves and muscles) are especially sensitive to alterations in electrolyte status. Given that disturbances in electrolyte balance can have widespread effects on cell physiology, we might expect that electrolyte imbalances would cause profound and widespread effects on responses to drugs. However, this does not seem to be the case; examples in which electrolyte changes have a significant impact on drug responses are rare.

Perhaps the most important example of an altered drug effect occurring in response to electrolyte imbalance involves digoxin, a drug for heart disease. The most serious toxicity of digoxin is production of potentially fatal dysrhythmias. The tendency of digoxin to disturb cardiac rhythm is related to levels of potassium: When potassium levels are depressed, the ability of digoxin to induce dysrhythmias is greatly increased. Accordingly, all patients receiving digoxin must undergo regular measurement of serum potassium to ensure that levels remain within a safe range. Digoxin toxicity and its relationship to potassium levels are discussed at length in Chapter 48.

Tolerance

Tolerance can be defined as decreased responsiveness to a drug as a result of repeated drug administration. Patients who are tolerant to a drug require higher doses to produce effects equivalent to those that could be achieved with lower doses before tolerance developed. There are three categories of drug tolerance: (1) pharmacodynamic tolerance, (2) metabolic tolerance, and (3) tachyphylaxis.

Pharmacodynamic tolerance

The term pharmacodynamic tolerance refers to the familiar type of tolerance associated with long-term administration of drugs such as morphine and heroin. The person who is pharmacodynamically tolerant requires increased drug levels to produce effects that could formerly be elicited at lower drug levels. Put another way, in the presence of pharmacodynamic tolerance, the minimum effective concentration (MEC) of a drug is abnormally high. Pharmacodynamic tolerance is the result of adaptive processes that occur in response to chronic receptor occupation.

Metabolic tolerance

Metabolic tolerance is defined as tolerance resulting from accelerated drug metabolism. This form of tolerance is brought about by the ability of certain drugs (eg, barbiturates) to induce synthesis of hepatic drug-metabolizing enzymes, thereby causing rates of drug metabolism to increase. Because of increased metabolism, dosage must be increased to maintain therapeutic drug levels. Unlike pharmacodynamic tolerance, which causes the MEC to increase, metabolic tolerance does not affect the MEC.

The experiment summarized in Table 8–1 demonstrates the development of metabolic tolerance in response to repeated administration of pentobarbital, a central nervous system depressant. The study employed two groups of rabbits, a control group and an experimental group. Rabbits in the experimental group were pretreated with pentobarbital for 3 days (60 mg/kg/day subQ) and then given an IV challenging dose (30 mg/kg) of the same drug. Drug effect (sleeping time) and plasma drug levels were then measured. The control rabbits received the challenging dose of pentobarbital but did not receive any pretreatment. As indicated in Table 8–1, the challenging dose of pentobarbital had less effect on the pretreated rabbits than on the control animals. Specifically, whereas the control rabbits slept an average of 67 minutes, the pretreated rabbits slept only 30 minutes—less than half the sleeping time seen in controls.

TABLE 8–1

Development of Metabolic Tolerance as a Result of Repeated Pentobarbital Administration

	Type of Pretreatment
Results	None	Pentobarbital
Sleeping time (minutes)	67 ± 4	30 ± 7
Pentobarbital half-life in plasma (minutes)	79 ± 3	26 ± 2
Plasma level of pentobarbital upon awakening (mcg/mL)	9.9 ± 1.4	7.9 ± 0.6

Data from Remmer H: Drugs as activators of drug enzymes. In Brodie BB, Erdos EG (eds): Metabolic Factors Controlling Duration of Drug Action (Proceedings of First International Pharmacological Meeting, Vol 6). New York: Macmillan, 1962:235. (See text for details.)

Why was pentobarbital less effective in the pretreated animals? The data on half-life suggest an answer. As shown in the table, the half-life of pentobarbital was much shorter in the experimental group than in the control group. Since pentobarbital is eliminated primarily by hepatic metabolism, the reduced half-life indicates accelerated metabolism. This increase in metabolism, which was brought on by pentobarbital pretreatment, explains why the experimental rabbits were more tolerant than the controls.

You might ask, “How do we know that the experimental rabbits had not developed pharmacodynamic tolerance?” The answer lies in the plasma drug levels when the rabbits awoke. In the pretreated rabbits, the waking drug levels were slightly below the waking drug levels in the control group. Had the experimental animals developed pharmacodynamic tolerance, they would have required an increase in drug concentration to maintain sleep. Hence, if pharmacodynamic tolerance were present, drug levels would have been abnormally high at the time of awakening, rather than reduced.

Tachyphylaxis

Tachyphylaxis is a form of tolerance that can be defined as a reduction in drug responsiveness brought on by repeated dosing over a short time. Hence, unlike pharmacodynamic and metabolic tolerance, which take days to develop, tachyphylaxis occurs quickly. Tachyphylaxis is not a common mechanism of drug tolerance.

Transdermal nitroglycerin provides a good example of tachyphylaxis. When nitroglycerin is administered using a transdermal patch, effects are lost in less than 24 hours (if the patch is left in place around the clock). As discussed in Chapter 51, the loss of effect results from depletion of a cofactor required for nitroglycerin to act. When nitroglycerin is administered on an intermittent schedule, rather than continuously, the cofactor can be replenished between doses and no loss of effect occurs.

Placebo effect

A placebo is a preparation that is devoid of intrinsic pharmacologic activity. Hence, any response that a patient may have to a placebo is based solely on his or her psychologic reaction to the idea of taking a medication and not to any direct physiologic or biochemical action of the placebo itself. The primary use of the placebo is as a control preparation during clinical trials.

In pharmacology, the placebo effect is defined as that component of a drug response that is caused by psychologic factors and not by the biochemical or physiologic properties of the drug. Although it is impossible to assess with precision the contribution that psychologic factors make to the overall response to any particular drug, it is widely believed that, with practically all medications, some fraction of the total response results from a placebo effect. Although placebo effects are determined by psychologic factors and not physiologic ones, the presence of a placebo response does not imply that a patient’s original pathology was “all in the head.”

Not all placebo responses are beneficial; placebo responses can also be negative. If a patient believes that a medication is going to be effective, then placebo responses are likely to help promote recovery. Conversely, if a patient is convinced that a particular medication is ineffective or perhaps even harmful, then placebo effects are likely to detract from his or her progress.

Because the placebo effect depends on the patient’s attitude toward medicine, fostering a positive attitude may help promote beneficial effects. In this regard, it is desirable that all members of the healthcare team present the patient with an optimistic (but realistic) assessment of the effects that therapy is likely to produce. It is also important that members of the team be consistent with one another; the beneficial placebo responses may well be decreased if, for example, nurses on the day shift repeatedly reassure a patient about the likely benefits of his or her regimen, while nurses on the night shift express pessimism about those same drugs.

Until recently, the power of the placebo effect was unquestioned by most clinicians and researchers. However, evidence now suggests that responses to placebos may be much smaller than previously believed (Box 8–1).

BOX 8–1 SPECIAL INTEREST TOPIC

HAS THE PLACEBO LOST ITS EFFECT?

In 1955, H. K. Beecher wrote his famous paper—“The Powerful Placebo”¹—which was heralded as solid proof for the long-held (but largely unsubstantiated) belief that placebos can effectively relieve symptoms in many patients. This widely cited paper had gone unchallenged until 2001, when two Danish scientists—Hróbjartsson and Gøtzsche—wrote their own paper on the subject, titled “Is the Placebo Powerless?”² From their research, the Danes concluded that, at least in the context of clinical trials, placebo treatment has little or no measurable effect. Who’s right? Let’s consider both papers and see if we can decide.

Beecher analyzed the data from 15 placebo-controlled clinical trials. In all of these trials, patients in the placebo groups were evaluated at baseline, treated with placebo for a prescribed time, and then re-evaluated. Beecher then looked to see if improvement took place between baseline and the end of the treatment period. Based on his analysis, he concluded “It is evident that placebos have a high degree of effectiveness, decided improvement . . . being produced in 35.2% of cases.” Pretty impressive. Unfortunately, there’s a flaw: How do we know the placebos produced the benefits? Perhaps 35.2% of the patients would have improved with no treatment, owing simply to the natural course of their disease or to other factors. After all, many people do get better on their own—without doctors, drugs, placebos, or anything else. Furthermore, although Beecher claims to have selected the 15 papers at random, this seems improbable in that seven of them were his own.

To address questions left open by Beecher, Hróbjartsson and Gøtzsche took a different approach. First, they analyzed data from 114 published trials—not just 15. More than 8500 patients were involved. More importantly, in all of these trials, placebo treatment was compared with no treatment. That is, in each trial, some subjects received placebo treatment and some received no treatment. (Of course, in most [112] of the trials, there was a third group of subjects who received an active treatment.) The trials involved 40 clinical conditions, including anemia, asthma, hypertension, hyperglycemia, epilepsy, Parkinson’s disease, schizophrenia, depression, smoking, and pain. In 38 of the trials, the measured outcomes were objective (eg, reduction in blood pressure, increase in red blood cell count), and in 76 the outcomes were subjective (eg, improvement in mood, reduction of pain). Of the 114 trials, 45 evaluated pharmacologic interventions, 26 evaluated physical interventions, and 43 evaluated psychologic interventions. The type of placebo employed was matched to the active treatment: For the pharmacologic studies, typical placebo treatment consisted of giving a lactose pill; for the physical studies (eg, evaluating the effect of transcutaneous electrical nerve stimulation on pain), typical placebo treatment consisted of performing the procedure but with the equipment turned off; and for the psychologic studies (eg, evaluating the effect of psychotherapy on depression), typical placebo treatment consisted of nondirectional, neutral discussion between the patient and the treatment provider.

What did the analysis reveal? In trials with objective outcomes, placebo treatment had virtually no measurable effect: outcomes in patients receiving placebo treatment were identical to those in patients receiving no treatment at all. However, in some trials with subjective outcomes, placebo treatment did have a convincing effect—but it was small, and limited primarily to studies of pain. In their conclusion, the authors stated, “We found little evidence that placebos in general have powerful clinical effects,” although they go on to say they did find “significant effects of placebo . . . for the treatment of pain.” In addition, they concede that their analysis does “leave open the question of whether placebo effects in clinical practice might differ from placebo effects among research subjects.”

Is this the end of the story? Is the placebo effect really just a myth? Well, we really can’t say. Yes, the Danish study, which was far superior to Beecher’s, failed to reveal a powerful effect of placebo treatment. However, this does not prove there is no placebo effect. Rather, it may simply indicate that we can’t readily measure a placebo effect in clinical trials. There are some good arguments supporting this possibility: