# The principles of drug use in palliative medicine

Chapter:
The principles of drug use in palliative medicine
DOI:
10.1093/med/9780198570295.003.0047

The control of distressing symptoms is central to the practice of palliative medicine. It is the foundation of ‘whole patient’ care in that it is not possible to deal with psychological, social, or spiritual concerns if patients have uncontrolled physical symptoms. The management of these symptoms is largely based on drug treatment, which means that effective symptom control requires some understanding of clinical pharmacology. The purpose of this chapter is to describe those principles of clinical pharmacology that are relevant to day-to-day practice in palliative medicine.

Unfortunately, symptom management is not a simple exercise of targeting a particular symptom with a specific drug. Patients with advanced disease are a vulnerable population and environmental and psychological factors have a variable but potentially very great influence on physical well-being. Sometimes, the response to drug treatment may be unpredictable for these reasons.

One hazard for palliative medicine physicians is the use of drugs for unlicensed indications or by unlicensed routes of administration, which is common in pain and symptom management. A quarter of prescriptions, affecting up to two-thirds of inpatients in specialist palliative care units in the United Kingdom, may fall into this category( 1 , 2 ). This highlights the obligation of prescribers to understand the principles of drug use, which should guide their practice and the specific limitations of the data available.

There are other complicating factors in this patient population. Patients are predominantly elderly with many co-morbidities and often with multi-system dysfunction. Polypharmacy is almost invariable, and the potential for modified or abnormal responses to drugs and drug interactions is considerable. Iatrogenic problems are common in that the prescription of one symptomatic remedy, for example, an opioid analgesic, will invariably cause other symptoms, in this case constipation, drowsiness, and possibly nausea, and dry mouth. A laxative, antiemetic, psychostimulant, and artificial saliva may all be added to the treatment regimen as a result. The skill of the palliative medicine physician is to deal effectively with each symptom without imposing a greater burden on the patient due to unwanted or intolerable drug effects or too complex a drug regimen. The principles of effective symptom control must be kept in mind: make a diagnosis of the underlying mechanism or cause of each symptom, individualize the treatment, and keep the regimen as simple as possible.

## Clinical pharmacology

Clinical pharmacology may be broadly divided into pharmacokinetics (‘what the body does to the drug’) and pharmacodynamics (‘what the drug does to the body’). It is often assumed that these are theoretical and rather esoteric disciplines that have little direct ‘clinical’ relevance. As we shall demonstrate, this assumption is wrong. For example, disordered handling of drugs by the body as a result of disease is common. Drugs are often used in special formulations; either modified release orally administered agents or novel forms designed for administration by other routes, and palliative care patients commonly require parenteral administration of drugs at some stage of their illness. These everyday clinical situations cannot be managed properly without some pharmacokinetic knowledge.

Pharmacodynamics is about drug action in man and, like pharmocokinetics, its measurement has become highly sophisticated; for example, in the use of imaging techniques to monitor biochemical changes in target organs. However, one of the challenges of drug use in palliative care is that often outcome measures are not easy to define and measure. In palliative care, drugs are not used to cure underlying disease but to improve comfort. Subjective symptoms such as pain, nausea, or depression are less easy to quantify than biochemical changes in the brain or even tumour size or serum calcium (although cortical responses to pain are now the subject of research because of developments in brain imaging), yet these are common targets for drug treatment in palliative medicine. Knowledge of the basic modes of action of drugs will underpin the logical selection and use of the most appropriate remedies for these symptoms.

### Pharmacokinetics

Pharmacokinetics is often described under the headings of absorption, distribution, metabolism, and excretion of drugs. A number of other terms are used also to describe the way in which the body handles drugs. These terms can be defined and modelled mathematically; however, the practising clinician should not be put off by the mathematical formulae often applied to the subject. An appreciation of the physiological, pharmacological, and pathophysiological factors that influence the ultimate concentration of a drug in the blood and eventually at its site of action is invaluable and does not need an understanding of complex mathematical models.

This section will define commonly used kinetic terms and identify, where possible, factors contributing to the variability of each. In order to illustrate the terminology, some pharmacokinetic parameters for frequently used drugs will be described at the end of each section.

### Half-life (t½)

This is perhaps the most well-known and commonly used pharmacokinetic parameter. It is a measure of the rate at which a particular process takes place. For example, the elimination half-life is a measure of the time taken for half the drug in the body to be removed. However, the process is often complex, for example, the decay in drug concentration following intravenous administration may comprise several exponentials, representing movement of drug inwards and outwards of body compartments and also the elimination of the drug from the body. It is generally the elimination half-life, which correlates most closely with the duration of action of the drug, though this is not always the case. Technical difficulties in defining a drug’s half-life, for example, complex assay procedures, may account for variation in quoted figures.

The elimination half-life is a dependent variable; dependent on volume of distribution and clearance. In the simplest kinetic model, half-life (t½) equals volume of distribution (Vd) multiplied by a constant (k) divided by clearance (Cl): $Display mathematics$

This means that half-life is prolonged as distribution volume increases and clearance reduces. These terms are defined in the following paragraphs.

Drugs with a long half-life may accumulate over a prolonged period of time and build up to toxic levels. For example, this may happen with methadone, which on regular dosing has a half-life of 20 to 40 h(3 ).

The elimination half-life of benzodiazepine hypnotics to some extent predicts the potential of each drug to cause residual or hangover effects the next day. Temazepam (t½ 8–10 h) and triazolam (t½ 2–3 h) are much less likely to cause daytime drowsiness than flurazepam (which has an active metabolite desalkylflurazepam with a t½ of 2–4 days).

### Apparent volume of distribution (Vd)

The apparent volume of distribution is defined as the volume into which all the drug in the body would need to be distributed to achieve the blood concentration. For drugs which are taken up into fat stores or muscle, the volume may be many times body size. Thus, the apparent volume of distribution is not a real volume that can be described in anatomical terms but a mathematical concept providing a convenient way of expressing how a drug is distributed in the body; this is why it is termed ‘apparent’. Similarly, a ‘compartment’ does not have an anatomical equivalent but is a theoretical space, often shown diagramatically as a box.

The volume of distribution is important as a determinant of half-life and is also of theoretical importance in the calculation of the loading dose of a drug where one is needed. Alteration in body composition and in the physicochemical environment of the body causes changes in distribution volume. Such changes for individual drugs in specific conditions may be quite difficult to predict. For example, emaciation may reduce the distribution volume of many centrally acting agents and lead to an enhanced effect after single doses. If a drug is normally highly bound to plasma protein, distribution volume may increase when plasma proteins are reduced because more drug is available for tissue-binding sites. However, because the proportion of active (unbound) drug is increased, a lower plasma concentration of drug will produce a given therapeutic effect. Thus, the net result is that the effects of these changes will tend to cancel each other out.

Digoxin is rapidly and extensively taken up by body tissues, particularly skeletal and cardiac muscle and in a 70-kg man has a volume of distribution of some 490 litres. In contrast, morphine is relatively hydrophilic and taken up in tissue stores less; it has a volume of distribution of 140 litres. In both of these examples the ‘apparent’ volume of distribution is greater than the size of the body.

### Clearance

Clearance is defined as the volume of blood which is completely cleared of the drug in a unit of time. It is usually measured in ml/min or litre/h. It therefore most closely reflects the efficiency of the elimination process. It is a major determinant of half-life and of the steady-state drug concentration. This is because at steady state, the amount of blood being cleared in the interval between doses will contain an amount of drug equivalent to the dose (assuming 100 per cent bioavailability). The two major organs of elimination are the liver and kidneys, both of which are susceptible to pharmacological and pathophysiological sources of variability. Clearance can be measured experimentally by plotting blood concentrations of the drug over time after a dose has been administered. Clearance (Cl) can then be calculated as: $Display mathematics$

This is illustrated in Fig. 8.1.

Figure 8.1
Plasma concentrations of a typical drug after oral administration. The shaded area is the ‘area under the curve’ of plasma concentration versus time and can be calculated mathematically, and from this the clearance can be derived.

Most drugs are cleared from the bloodstream either by metabolism in the liver, or by excretion by the kidneys, or a combination of both. Total blood clearance determines steady-state concentration and is the sum of all clearance mechanisms. It is often claimed that if one organ of elimination is compromised by disease, the other can compensate by allowing increased excretion of the drug; however, this does not prevent a rise in drug level in the body, which is determined by total clearance.

The aim of any dosing regimen in an individual patient is to achieve a concentration of drug in the blood that is high enough to give the intended effect without producing adverse effects (Fig. 8.2). This concentration can never be completely steady as peaks will occur at the point of maximum drug absorption after administration, and troughs will occur immediately before each dose (Fig. 8.3). Steady state is said to have been achieved when all trough and peak concentrations do not vary. The degree of swing between peak and trough concentrations is determined by the drug’s elimination half-life and the frequency of drug administration; for instance, where a drug has a short half-life and the difference between peak and trough concentrations is small, then dosing should be frequent. The actual steady-state concentrations achieved in an individual are dependent on the dose administered and the bioavailability and clearance of the drug from the blood in that patient.

Figure 8.2
Plasma concentrations of a typical drug after oral administration of a single dose. The aim in chronic dosing is to maintain plasma concentrations within the therapeutic range.

Figure 8.3
Steady-state plasma concentration of a short-acting drug showing peaks and troughs after each dose.

It is a common misconception that steady-state drug concentration is dependent on the size of the body. In fact the volume into which the drug is distributed plays no part. The reason why smaller people and children often require reduced doses is because they have smaller organs of elimination and therefore lower clearance rates. This is pertinent in the context of palliative care where it cannot be predicted that a patient whose body is wasted may need to be given a smaller dose. The elimination process itself may be unimpaired.

### Time to reach steady-state plasma concentration

Whilst the actual concentration of drug in the blood is independent of its half-life, the time it takes to reach that level is entirely and solely dependent on this parameter. As a rough guide, if the same dose of drug is given at a constant time interval it takes about four half-lives to reach 95 per cent of the steady-state plasma concentration. This applies only to drugs whose elimination is governed by ‘first-order kinetics’. Fortunately, this comprises the vast majority of drugs, with phenytoin being the notable exception. During a first-order process, a constant proportion of remaining drug is eliminated in a unit of time. Indeed, that is why the term half-life can be applied; half of the drug in the body is removed in the half-life interval.

A first-order process is independent of the concentration of drug; as the concentration increases, the proportion eliminated is the same per unit of time whilst the actual amount of drug eliminated increases. For a first-order process, therefore, the amount of drug in the body must build up until the amount of drug being eliminated during the dose interval equates with the dose. Since the time to reach steady state depends on half-life, when half-life is very long, it may be necessary to give a loading-dose initially, and then to follow with a reduced maintenance dose.

A zero-order process is dependent on concentration and the amount eliminated per unit of time is fixed. Phenytoin is initially metabolized according to first-order kinetics but the enzyme system responsible has limited capacity. Once this system is ‘saturated’, the metabolism of phenytoin continues as a zero-order process.

The elimination half-life (t½) of morphine is 2–4 h. Thus, when morphine is administered at regular 4-hourly intervals, steady state will be 95 per cent achieved after 16 h or so. In contrast, the half-life of digoxin is about 2 days so that it would take at least 8 to 10 days to achieve steady state. It is usual to give a loading dose of digoxin and then maintain patients on a single daily dose.

### Bioavailability

The bioavailability of a drug is the percentage of administered drug that gains access unchanged to the systemic circulation. Bioavailability is of most clinical relevance after oral administration. It can be measured in individual patients by comparing the area under the plasma concentration time curves after oral and parenteral administration allowing for any difference in the dose administered and assuming no change in the rate of elimination. Many factors may contribute to a reduction of bioavailability below 100 per cent. The extent of drug absorption may be susceptible to changes in gastrointestinal function. For some drugs, such as digoxin and phenytoin, variation in the pharmaceutical formulation may account for significant differences in bioavailability. More important, however, is the extent of pre-systemic (or first pass) metabolism (Fig. 8.4).

Figure 8.4
‘First pass’ metabolism may take place in the liver, and sometimes also in the intestinal mucosa, before the absorbed drug enters the systemic circulation.

Once absorbed from the gastrointestinal tract, all drugs must pass via the portal venous system through the liver. If within that organ there is an avid system of enzymes for metabolizing the particular drug, then a percentage will be biotransformed before passing through the liver to the systemic circulation. For some drugs, the amount extracted during one pass through the liver, the extraction ratio, may be as high as 80 per cent, giving an oral bioavailability of only 20 per cent. The steady-state blood concentration of such drugs after oral administration may exhibit high variability; minor changes in the extraction ratio will be reflected in large percentage changes in bioavailability. Thus, interactions with other drugs which induce or inhibit hepatic enzymes or changes in hepatic function due to disease may have a profound effect on such drug levels after oral administration but relatively little effect when the drug is given parenterally. In patients with chronic liver disease or hepatic metastases blood may be ‘shunted’ from portal to systemic vessels. The drug may thus bypass hepatic enzymes, the pre-systemic metabolism will be reduced, and the bioavailability may be considerably increased. Consequently, much higher levels of drug may build up after oral administration if the enhanced bioavailability is not taken into account. The effect of pre-systemic metabolism on drug bioavailability must also be considered when calculating oral doses during conversion from parenteral regimens. The bioavailability of some drugs may also be reduced by intraluminal degradation in the gastrointestinal tract or metabolism by enzymes within the intestinal wall.

### Drug absorption

In the main, the absorption of drugs is a passive process along a concentration gradient across a lipid cell membrane. As long as the drug is in solution and has a degree of lipid solubility, there is sufficient surface area for diffusion, and the drug remains in contact with the absorptive areas for long enough, then problems should not arise. Most drugs are absorbed where the greatest surface area is available, that is in the small bowel. A reduced rate of absorption may therefore occur if there is a delay in emptying of the stomach. This might arise as part of a pathological process or due to pharmacological agents that slow gastric motility, such as drugs with anticholinergic effects or opioid analgesics. Drugs must have the physicochemical characteristics to facilitate dissolution in the gut, but once presented to the vast surface area of the small bowel, their potential for full absorption should be easily achieved. Only the most severe structural gastrointestinal disease will cause problems.

Many drugs are now formulated as modified-release preparations. For them to achieve their expected absorption profile they may need to remain in the small bowel for a prolonged period of time. Usually such formulations have been tested only under ideal conditions in healthy volunteers. In patients with increased gastrointestinal transit there is a risk that a modified-release drug will be propelled past the absorptive zone of the gut before all of the drug has been released, resulting in therapeutic failure.

Within the gut there is some potential for drug interaction, which results in reduced bioavailability. Most examples are well known and involve loose chemical binding between two drugs within the gut lumen. For example, cholestyramine binds many drugs; iron salts and tetracycline bind to each other, and sucralfate binds phenytoin. Other less obvious drug interactions may occur, involving interference with absorption. Some broad-spectrum antibiotics decrease the effectiveness of the oral contraceptive; the mechanism may be increased gastrointestinal transit caused by the antibiotic, leading to reduced absorption of the contraceptive.

Absorption and bioavailability are not the same. Morphine, for example, is more or less completely absorbed (i.e. 100 per cent). However, it undergoes extensive pre-systemic metabolism, mainly in the liver but possibly also in the wall of the gastrointestinal tract. The bioavailability of morphine is thus 20 to 30 per cent.

### Drug metabolism

Drug biotransformation takes place mainly in the liver and contributes both to the rate of elimination of drug and its bioavailability. The rate at which metabolism proceeds usually determines the clearance; however, where removal is particularly rapid (high extraction ratio) the rate of delivery of drug to the liver, rather than the rate of metabolism, may determine clearance (flow-dependent kinetics). For such drugs, if liver blood flow is markedly reduced, drug accumulation will result.

The biochemical processes of drug metabolism are complex. Two phases of metabolism are usually described, involving initial oxidation or hydrolysis (phase I) followed by conjugation (phase II), but this concept can be misleading. All of the reactions involve the production of products which are more polar and, therefore, more water-soluble and amenable to excretion by the kidney. In some circumstances, phase II reactions may take place without a prior phase I reaction. Phase I reactions involve oxidation, reduction, hydrolysis, hydration, dethioacetylation, and isomerization. Such reactions may prepare the drug molecule for a phase II reaction by producing or uncovering a chemically reactive group, which then forms the substrate for a phase II reaction.

Of the phase I reactions, oxidation involving the ‘mixed-function oxidase system’ is the most important and its behaviour is best understood. This system of enzymes is based in hepatic microsomes and requires molecular oxygen, NADPH and cytochrome P450, and NADPH-cytochrome P450 reductase. Amongst the reactions catalysed by the mixed-function oxidase system are aromatic hydroxylation, aliphatic hydroxylation, epoxidation, N-dealkylation, O-dealkylation, oxidative deamination, N-oxidation, S-oxidation, and alcohol oxidation. Not all oxidative processes are carried out by this system; alcohol dehydrogenation is performed by a non-microsomally located enzyme which is responsible for the major pathway for alcohol detoxification (in non-enzyme induced subjects).

Phase II reactions mostly involve conjugation; glucuronidation, glycosylation, sulphation, methylation, and acetylation or conjugation with gluthatione or certain amino acids. An appreciation of these processes is necessary in developing a scientific approach to dose management.

### Pharmacodynamics

Drugs produce their effects on the body by combining with receptors, by modifying enzyme processes, or by direct chemical or physical actions.

### Receptors, agonists, and antagonists

Receptors are specialized areas of the cell membrane which are highly specific for certain drug or hormone molecules. A drug that combines with a receptor to ‘activate’ it is called an agonist; this terminology derived initially from the actions of hormones and neurotransmitters. The term agonist refers to a drug that binds to receptors to induce changes in the cell which stimulate physiological activity. Some drugs can combine with receptors without initiating any change in cell function. Such drugs are called competitive antagonists because they interfere with the action of agonists by blocking the receptor sites. Non-competitive antagonists do not compete for the same receptor as the agonist but block the effect of the agonist in some other way.

A partial agonist is a drug with low intrinsic activity (efficacy) so that its dose–response curve exhibits a ceiling effect at less than the maximal effect produced by a full agonist. The difference between a partial agonist and a full agonist is thus a difference in efficacy (Fig. 8.5). This should not be confused with potency, which is a measure of the amount of drug required to produce a given effect, and is a measure also of affinity for receptors; the more potent a drug, the greater its affinity for receptors. Thus a drug may be a partial agonist (less effective) but still more potent than a full agonist. This is the case with buprenorphine, which has limited efficacy compared with morphine but greater potency (0.3 mg intramuscular buprenorphine ≡ 10 mg intramuscular morphine(4 ). However, because it is more potent it has greater affinity for μ opioid receptors and can displace morphine from them. In this way it can act as an ‘antagonist’ of morphine by reducing the overall μ opioid effect (see Chapter 10.1.6). Buprenorphine can therefore be classified as an ‘agonist antagonist’. Other opioid analgesics, such as pentazocine, are also classified as ‘agonist antagonists’ but have a different profile. These drugs have both agonist and antagonist effects at receptors, but at different receptors (see following paragraphs).

Figure 8.5
Log dose–response curves for two full agonist drugs A and B, and a partial agonist C. Drug B is more potent than drug A but no more effective. Drug A is more effective than drug C but has similar potency.

There are similar examples in other therapeutic areas, and they are likely to increase as new receptors and receptor subtypes are identified. For example, metoclopramide has been regarded primarily as a dopamine receptor blocker. This is true at low doses, however, at high doses, metoclopramide blocks (antagonizes) 5HT3 receptors and is a more effective antiemetic. Metoclopramide also has a prokinetic effect on the upper gastrointestinal tract and this is mediated through an agonist effect at 5HT4 receptors, leading to an enhancement of the effects of acetylcholine release in the gut. Metoclopramide is thus both an antagonist and an agonist at different serotonin receptors.

Morphine is an agonist at μ opioid receptors. Buprenorphine is a partial agonist at μ receptors and can, in certain circumstances, reverse (antagonize) the effects of morphine. Pentazocine is a weak competitive antagonist at μ receptors (so may also antagonize the effects of morphine but by a different mechanism) and is a partial agonist at the κ opioid receptor. Naloxone is an antagonist at the μ opioid receptor and will block the effects of μ agonists and partial agonists (but with varying efficiency).

### Drugs which alter enzyme activity

Drugs affecting enzyme processes may have diverse therapeutic applications but many act by being inhibitors of enzyme actions. For example, non-steroidal anti-inflammatory drugs block the effect of the enzyme cyclo-oxygenase and thereby interfere with the synthesis of prostaglandins; this is believed to be the basis for their anti-inflammatory activity. Monoamine oxidase inhibitor antidepressants interfere with the degradation of monoamine neurotransmitters thus enhancing their effect in central synapses. Angiotensin-converting enzyme inhibitors block the conversion of angiotensin I to angiotensin II by inhibiting the relevant enzyme and are effective in the treatment of hypertension and cardiac failure.

### Drugs which have a direct chemical or physical action

Antacids are an example of drugs with a direct chemical action; they are bases which neutralize gastric acid. Drugs with a physical mode of action include the bulk laxatives such as ispaghula husk. Whilst the mode of action of such drugs seems less complex than that of drugs that interact with receptors, the same attention to detail in their use and an individualized approach are necessary to maximize their benefits and reduce potential adverse effects.

### Tolerance, drug dependence, and drug resistance

#### Tolerance

Tolerance refers to the phenomenon of decreasing response to a drug as a consequence of its continued use. It is manifest by a shift to the right in the dose–response curve; an increased dose is required to achieve a similar effect (Fig. 8.6). In contrast, sensitization refers to the phenomenon of increasing response to a drug, as a consequence of its continued use. It is manifest by a shift to the left in the dose–response curve; a decreased dose is required to achieve a similar effect (Fig. 8.6). Sensitization is a relatively uncommon phenomenon.

Figure 8.6
Graphical representation of drug tolerance and sensitization.

Tolerance may be due to:

• An alteration in the pharmacokinetic profile of a drug (pharmacokinetic tolerance). For example, tolerance to barbiturates has been linked to induction of hepatic microsomal enzymes, which results in an increased metabolism of the barbiturate.

• An alteration in the pharmacodynamic profile of the drug (pharmacodynamic tolerance). Tolerance to opioids has been linked to uncoupling of opioid/G protein receptor complexes, which results in a decreased effect of the opioid.

Tolerance occurs at a variable time after initiation of the drug. Moreover, it may occur to some, or all, of the effects of the drug. For example, tolerance to the analgesic effects of opioids is uncommon in clinical practice. Indeed, increases in opioid requirement are invariably related to progression of disease, rather than to the development of tolerance(5 ). However, tolerance to some of the adverse effects of opioids is common in clinical practice. Thus, sedation and nausea and vomiting usually settle within a few days, or a week or two. Cross-tolerance refers to a decreased response to a drug as a consequence of the use of a drug with a similar structure or function. It may occur to some or all of the effects of the drugs. The extent of cross-tolerance between opioid drugs is very variable. The practice of opioid switching relies on incomplete cross-tolerance between opioids. Thus, opioid switching will only be successful if cross-tolerance to the analgesic effects is less than cross-tolerance to the adverse effects(6 ).

#### Drug dependence

In pharmacological terms, drug dependence has been divided into two types(7 ):

• Psychological dependence—characterized by a compulsion to continue taking the drug. This is a pathological response to the drug.

• Physical dependence—characterized by a withdrawal syndrome if the drug is not taken (or if an antagonist drug is taken). This is a normal physiological response to the drug.

The relationship between the two types of drug dependence is variable: patients may develop psychological dependence or physical dependence, or a combination of psychological and physical dependence. Opioid drug misusers invariably develop a combination of psychological and physical dependence.

Psychological dependence is associated with a positive (perceived) effect from taking the drug. Drugs that produce a rapid effect are more likely to induce psychological dependence than drugs producing a delayed effect. For example, a drug given intravenously is more likely to induce psychological dependence than the same drug given orally. The positive effects of the drug are often reinforced by the rituals and environment associated with drug-taking. Indeed, the treatment of psychological dependence involves measures aimed at reducing these reinforcing factor(7 ).

Physical dependence is associated with a negative effect from not taking the drug, i.e. a withdrawal syndrome. The features of the withdrawal syndrome are usually the opposite of the features of the drug’s acute effect. The treatment of physical dependence involves trying to reduce the impact of this reinforcing factors(7 ). A variety of different ‘detoxification’ regimens have been used, including gradual reduction of the original drug, gradual reduction of a substitute drug, and symptomatic treatment of the withdrawal syndrome. Many patients require a combination of different strategies in order to overcome their drug dependence.

Patients receiving opioids for cancer pain rarely develop psychological dependence, although they may more often develop physical dependence to the opioid(8 ). The best opioid for the patient’s clinical situation should be chosen, whatever its potential for abuse; a patient with difficult breakthrough pain will require an opioid with a rapid onset of action, for example, even if the patient has a history of drug misuse (see Chapter 10.1.6).

#### Drug resistance

Resistance refers to the phenomenon of lack of responsiveness to a drug. The expression is primarily applied to antimicrobial drugs and cytotoxic chemotherapy, but could be used for other classes of drug. Drug resistance has been sub-classified into: (1) primary or intrinsic resistance—this type of resistance develops de novo; and (2) secondary or acquired resistance—this type of resistance develops after exposure to the drug. Drug resistance may involve a single drug, several drugs from the same class (cross-resistance), or several drugs from different classes (also described as cross-resistance). Cross-resistance is invariably due to a common molecular mechanism.

Drug resistance is primarily related to the genetic profile of the cell. Resistance may occur in response to spontaneous mutation or, in the case of antimicrobial agents, to transfer of genetic material (involving plasmids, bacteriophages, or other mechanisms). The molecular mechanisms underlying drug resistance are shown in Table 8.1. Continuing drug usage encourages the development of resistance by suppressing the growth of sensitive cells, thereby encouraging the growth of resistant cells.

Table 8.1 Molecular mechanisms of drug resistance.

Molecular mechanisms of drug resistance

Specific example

Drug inactivated

Resistance to penicillin may result from the production of β lactamase

Drug prevented from entering into the cell

Melphalan resistance has been linked to deactivation of various active transport systems (e.g. system L)

Drug prevented from remaining in the cell

Fluconazole resistance has been linked to activation of various active transport systems (e.g. major facilitator efflux pump)

Pro-drug not activated

Resistance to 5-fluorouracil may result from the abnormalities of several enzymes in its metabolic pathway

Alteration of drug target

Amphotericin resistance has been linked to alteration of cell membrane

Overproduction of drug target

Methotrexate resistance has been linked to overproduction of dihydrofolate reductase

Effect of drug counteracted

Resistance to cisplatin may result from repair of damaged DNA

Antimicrobial drug resistance is a major clinical problem. Indeed, the Standing Medical Advisory Committee Sub-Group on Antimicrobial Resistance (UK) has described the current situation as ‘looking into the abyss’(9 ). Antimicrobial drug resistance involves not only antibacterial agents, but also antifungal and antiviral agents. Infections caused by resistant organisms generally result in increased morbidity, increased mortality, and increased use of resources(9 ).

There is relatively little information about the impact of antimicrobial resistance in palliative care, however one study from the United Kingdom reported that methicillin-resistant Staphylococcus aureus (MRSA) carriage was low. Only 6 per cent of patients transferred into one large hospice from environments not known to have MRSA, and 7 per cent of patients transferred from hospital wards known to have MRSA were carriers, suggesting the burden of conventional eradication regimens in this patient group needs to be considered carefully(10 ). Infections in hospices are associated with increased physical and psychological morbidity and increased use of resources(11 ). Moreover, this study reported that infection control measures secondary to MRSA infections resulted in major operational problems for the hospices concerned.

Antimicrobial drugs should generally only be prescribed for cases of proven infection, and not for prophylaxis of infection. Narrow spectrum drugs should be employed, and these should be prescribed in relatively high doses, for relatively short courses(9 ). Other strategies that are relevant include the use of appropriate infection-control measures (e.g. hand washing).

Cytotoxic chemotherapy drug resistance is a perennial clinical problem(12 ). Many tumours are primarily resistant, whilst many other tumours become secondarily resistant.

### Routes of administration of drugs in palliative care

A variety of different routes of drug administration are used in palliative care. The choice of route will depend on a combination of patient, drug, and organizational factors, including availability of drug formulations and financial and human resources. Moreover, the choice of route may vary over the course of the patient’s illness.

#### Oral route

The oral route is the principal route of administration of drugs in palliative care. The advantages of this route are that it is simple, non-invasive, acceptable to patients, and maintains patient control. Moreover, the majority of drugs used in palliative care are available in oral forms, and often in different oral formulations (liquid and solid). The disadvantages of this route are that there are numerous factors that can affect the absorption and bioavailability of the drug.

#### Modified-release formulations for oral administration

There are two theoretical benefits of modified-release oral preparations; they prolong the absorption time and thereby extend the overall duration of action of a short-acting drug (Fig. 8.7) and they attenuate peak plasma concentrations of a drug where such peak concentrations could be associated with adverse effects. The most common examples in palliative medicine are modified-release opioids. Most orally administered opioids are short-acting drugs with duration of analgesia of about 4 h. In order to maintain control of chronic pain they have to be given six times a day. Modified-release opioid tablets have a 12- or 24-h duration of effect. Reduction in the frequency of dosing has important benefits in terms of patient acceptance and compliance. The theoretical reduction in adverse effects has not been demonstrated in practice with modified-release oral opioids.

Figure 8.7
Typical plasma concentration profile for a normal-release and modified-release formulation of the same drug.

The plasma concentration profile of a modified-release preparation is different from that of a normal-release preparation; the time to peak plasma concentration is delayed and the peak is attenuated. This has implications for the use of such preparations; in general, modified-release formulations are designed for maintenance treatment. Some drugs need to be rapidly absorbed and have a relatively short duration of action if they are to achieve their intended therapeutic effect without producing significant unwanted effects. This applies, for example, to analgesics used in the treatment of acute pain. It is inappropriate, therefore, to use modified-release preparations in such situations.

Modified-release formulations prolong the absorption phase, but do not change the elimination process. For example, the elimination half-life of morphine (2–4 h) is the same whether or not a normal- or modified-release formulation is used(13 ).

#### Oral transmucosal route (sublingual, buccal)

In theory, the oral transmucosal route offers certain advantages over the oral route, particularly, increased speed of drug absorption and increased drug bioavailability for drugs that undergo first-pass hepatic metabolism. The advent of oral transmucosal and buccal preparations of opioids, usually fentanyl, has allowed more rapid management of breakthough pain than can be achieved by normal release oral preparations( 14 , 15 ). The buccal route has been used to treat epileptic seizures in children(16 ), and, anecdotally, this approach has also been successful for the management of seizures and terminal agitation in palliative care patients.

#### Nasogastric/enteral feeding tubes

The decision to use this route is dependent on a number of factors, including the availability of a suitable formulation of the drug, the compatibility of the drug and enteral feed if it is being administered, and practical issues such as the position and size of the feeding tube(17 ).

Ideally, liquid drug formulations should be used with feeding tubes. However, solid drug formulations can be used, though it is necessary to dissolve them or crush and suspend them in water. Occasionally, parenteral drug formulations are used in place of oral formulations. The appropriateness of crushing solid drug formulations, and of using parenteral formulations, should be verified with a reliable source of drug information.

Modified-release formulations should not be crushed, since this can affect their physical characteristics and therefore their pharmacokinetic profile, usually resulting in an increased rate of absorption or ‘dose dumping’. Dose dumping may result in decreased efficacy, decreased tolerability, or sometimes even serious drug toxicity. Enteric-coated tablets should not be crushed unless the end of the feeding tube lies beyond the stomach.

Drugs should always be administered separately from the enteral feed, and the feeding tube should be flushed with water before and after the drug is given. Similarly, if several drugs are to be administered, the tube should be flushed before and after each drug. In the case of drugs that do not interact with enteral feed, the feed can be stopped immediately before and restarted immediately after the drug has been administered.

Modified-release granules of morphine are available, which can be suspended in water and given through nasogastric and enteral feeding tubes. Particular care is needed with flushing when using these preparations to avoid them blocking the tube.

#### Rectal route

The rectal route may be an important route for administration of drugs in palliative care, depending on the care setting. Its main advantage is that it is simple, does not require additional equipment, and non-professional carers can be taught to administer drugs via this route. Its main disadvantage is that it is not acceptable to many patients and may be inconvenient for very ill patients. A variety of drugs can be given rectally; some are formulated specifically for rectal use and, sometimes, the oral or parenteral formulation is used.

The absorption of drugs from the rectum can be very variable. Absorption is limited by the small surface area, the presence of faeces, defaecation, and involuntary expulsion of the drug. The venous drainage of the lower and middle part of the rectum is the inferior vena cava and of the upper part of the rectum the portal vein, with significant anastomoses between these two venous systems. Therefore drugs given rectally are subject to some first-pass hepatic metabolism. It is recommended that drugs are inserted one finger length into the rectum and that 10 ml of warm water is administered also to ensure dissolution if the rectum is dry(18 ). Alcohol/glycol containing formulations should be avoided since they can cause rectal irritation.

#### Subcutaneous route

The subcutaneous route is the most commonly used parenteral route in palliative care. The advantages of the subcutaneous route are that it is simple and acceptable to patients and subcutaneous injections are less painful than intramuscular injections. A subcutaneous cannula can be inserted which allows repeated administration of intermittent boluses of drugs, or administration of a continuous subcutaneous infusion, thus reducing the need for repeated skin puncture. Whilst a number of drugs may be given subcutaneously, most will not be licensed for this route of administration.

Subcutaneous infusions can be given using infusion devices available in most hospitals, or, more commonly, using small battery-powered syringe drivers. A recent survey from the United Kingdom reported that a variety of different combinations of drugs were being given by continuous infusion; however, the stability/compatibility of only half of the drug combinations in use was known(19 ). The compatibility of drug combinations is dependent on a number of factors, including the drugs used, the concentration, the diluent, and other factors such as temperature and UV light. Databases of compatible drug combinations are now available on the Internet (http://www.pallcare.info and http://palliativedrugs.com/pdi.html). It should be noted that the absence of precipitation within a drug mixture is not synonymous with compatibility between the drugs in that mixture(20 ).

Continuous subcutaneous infusions produce stable blood levels of drugs, which makes them suitable for the control of ongoing symptoms. Intermittent subcutaneous bolus injections of drugs are usually rapidly absorbed, which makes this route suitable for the emergency treatment of symptoms. However, the absorption of drugs may be affected if the cutaneous blood flow is compromised. Drugs given via the subcutaneous route, as for other parenteral routes, are not susceptible to first-pass hepatic metabolism and therefore tend to have a high bioavailability, generally near 100 per cent.

#### Intramuscular route

The intramuscular route is used infrequently in palliative care. It is generally used when the subcutaneous route is contraindicated, such as when the drug is irritant to the skin, when the drug is of large volume, and when the cutaneous blood supply is compromised.

#### Intravenous route

The intravenous route is used infrequently in palliative care. It is generally used in emergency situations to obtain a rapid therapeutic response and in circumstances where the other parenteral routes are contraindicated, such as where there is a bleeding diathesis or the peripheral blood supply is compromised. However, if a patient has an indwelling central venous catheter, it is often used rather than instituting another parenteral route.

#### Dermal route

Few drugs are reliably absorbed through the skin. Absorption is limited by the physical characteristics of the epidermis and is increased if the epidermis is damaged or destroyed. The absorption of drugs is also dependent on cutaneous hydration and blood flow—poor hydration and blood flow leading to decreased absorption. Drugs that are absorbed through the skin are not susceptible to first-pass hepatic metabolism.

Drugs are usually applied to the skin in order to achieve a local effect. Capsaicin cream is used for non-malignant neuropathic pain, and lidocaine patches are now licensed for postherpetic neuralgia; both have been used for other causes of neuropathic pain in palliative care, but experience is limited. Parenteral formulations of diamorphine and morphine (in suitable bases) have been used to treat localized pain secondary to non-malignant ulceration, tumour infiltration, and tumour fungation(21 ). However, drugs can also be applied to the skin in order to achieve a systemic effect. For example, specific transdermal formulations of fentanyl and buprenorphine are available for chronic pain and are used extensively in some countries. In a European Association for Palliative Care survey of over 3000 patients in 21 countries, transdermal fentanyl accounted for 14 per cent of all opioid prescriptions(22 ) (see Chapter 10.1.6).

#### Pulmonary route

In palliative care, the pulmonary route is used primarily to administer drugs to the lungs. Drugs are given either as a fine powder or an aerosol administered via an inhaler, with or without a holding chamber or ‘spacer’, or a nebulizer. The addition of a spacer improves the efficiency of inhalers. A systematic review concluded that bronchodilators administered via an inhaler and spacer were as effective as bronchodilators administered via a nebulizer in acute asthma(23 ). The pulmonary route of administration of drugs is discussed in more detail in Chapter 11.1.

#### Spinal route

The spinal route refers to administration of drugs either into the epidural space (epidural administration), or into the subarachnoid space (intrathecal administration). The spinal route of administration of drugs is discussed in detail in Chapter 10.1.9.

#### Other routes

Other routes of administration that have been used in palliative care include intranasal, intracavitary, and regional application of drugs. The intranasal route can be used to deliver drugs systemically, as well as locally, for instance in the management of breakthrough pain(24 ). Fentanyl(25 ), sufentanil and diamorphine(26 ), have been used intranasally, although their use has largely been overtaken by the availability of oral transmucosal delivery systems. There have been several reports of the successful use of intracavitary local anaesthetics for localized pain; local anaesthetic administered into the pleural cavity has been used to manage chest wall pain(27 ), whilst local anaesthetic administered into the joint space has been used for bone pain secondary to pathological fracture(28 ). Similarly, there have been several reports of the successful use of regional (perineural) application of local anaesthetics for neuropathic pain(29 ).

### Variability in drug response

Wide variability in the rates of drug metabolism occurs between individuals as a result of genetic factors, pathological processes, concurrent medication, and ageing, and creates the major obstacle to matching dose to patients’ requirements.

Both pharmacokinetic and pharmacodynamic factors may be responsible for therapeutic failure or adverse effects and the mechanism of each may relate to drug interaction, disease, genetics, or the effect of old age. The study of variability in drug response encompasses the whole of the science of clinical pharmacology. In this section we will merely highlight important principles and provide examples most relevant to palliative medicine.

#### Pharmacogenetics

Wide inter-individual variation in the rates of metabolism of drugs has been observed for many years and a combination of genetic and environmental factors is assumed to be responsible. Our understanding of the influence of genetic factors has increased rapidly with the advent of modern molecular biological techniques. For example, the rate of elimination of isoniazid within populations has been recognized to have a bimodal distribution(30 ). This drug is N-acetylated and the rate of drug acetylation is under the control of two autosomal alleles, R for fast acetylation and r for slow (R being dominant and r recessive)(31 ). ‘Fast acetylators’ of isoniazid may be more susceptible to hepatic damage caused by the drug and at the same time may be at risk of under dosage with other agents which are acetylated(32 ). The high incidence of the recessive trait, which determines slow acetylation, may indicate a selective advantage for slow acetylators, which is not related to drug metabolism. There is variation in the proportion of fast and slow acetylators in different populations and ethnic groups. In most European groups, about 40 per cent, and in the United States, 45 per cent, are fast acetylators, but 80 to 90 per cent of Asian populations and nearly 100 per cent of Canadian Eskimos are fast acetylators(33 ). Numerous drugs are metabolized by acetylation and the effects of genetic polymorphism have not been studied for all of these drugs. Hydralazine, some sulphonamides, and some benzodiazepines are worthy of mention but within the field of palliative medicine, where, in general, dose is titrated to patients’ needs, there do not appear to be significant clinical implications of this phenomenon.

A genetic component to drug oxidation was first documented when rates of metabolism of antipyrine were shown to have a greater degree of concordance in identical rather than fraternal twins(34 ). This was difficult to interpret because of the lack of any observed correlation in the rates of oxidation of different drugs between individuals. However, it was subsequently demonstrated that people who were slow oxidizers of debrisoquine were also slow metabolizers of other drugs such as sparteine, phenformin, phenytoin, metoprolol, nortriptyline, and others(35 ). We now know that the enzyme system involved, cytochrome P450, can be classified into a number of sub-families each of which is probably under genetic and environmental control. The most widely studied is the debrisoquine 4-hydroxylation phenotype, which is under the control of cytochrome P450 2D6 (CYP2D6). About 90 per cent of the population are extensive metabolizers of debrisoquine and family studies have indicated this is a dominant trait( 36 , 37 ). Using kinetic tests, it has been shown that different racial groups exhibit different proportions of poor and extensive metabolizers. Egyptians show the lowest incidence of debrisoquine poor metabolizer status (about 1 per cent), whereas West Africans show the highest (about 13 per cent) and Caucasians have an intermediate incidence. CYP2D6 is involved in the methylation of codeine and failures of this drug to produce analgesia have been attributed to the inability to metabolize codeine to the active moiety, morphine(38 ).

There is evidence of genetic control in the metabolism of mephenytoin in which another isoform of cytochrome P450 is involved (2C19). About 3 per cent of Caucasians and 18 per cent of Japanese people are poor metabolizers of mephenytoin(39 ). There is a single case report of a CYP3A5 polymorphism leading to rapid methadone metabolism in a patient with a long history of drug misuse(40 ). CYP1A2, responsible for the metabolism of theophylline amongst many other drugs, is easily induced by alcohol and flavonoids found in the diet, especially cruciferous vegetables and polycyclic aromatic hydrocarbons found in barbecued food. The variation between genetically determined poor and extensive metabolism is wide and it is important therefore that the doses of susceptible drugs should be managed carefully. Recent studies in palliative care populations suggest that opioid receptor polymorphism(41 ) and COMT-polymorphism(42 ) may influence patients’ requirements for opioids (see Chapter 10.1.6).

There are pharmacodynamic examples of genetic variation in drug response. These include resistance to the effects of warfarin due to increased sensitivity to vitamin K, haemolysis in patients with glucose-6-phosphate dehydrogenase deficiency in response to drugs such as sulphonamides or nitrofurantoin, and flushing in response to alcohol in patients taking chlorpropamide(43 ).

#### Disease states

A myriad of conditions, both acute and chronic, can affect the response to drugs through both pharmacokinetic and pharmacodynamic mechanisms. From the pharmacokinetic viewpoint, diseases of the two most important organs of elimination, the kidneys and the liver, are most important.

Excretion of drugs by the kidneys depends upon either filtration of drug unbound to plasma protein at the glomerulus or the active transport systems, which secrete drug at the renal tubule. The reduction in the renal clearance of any drug closely follows renal function as measured by creatinine clearance. The consequences of renal disease therefore depend upon the extent to which renal clearance contributes to total drug clearance and how critical drug concentration is in terms of toxicity. Information about the need for dose adjustment in patients with renal impairment is usually readily available in prescribing information sources. A more complex situation exists for drugs metabolized in the liver. The relevance to palliative medicine is much greater because centrally-active drugs tend to be those requiring metabolism simply because they are lipid soluble; a necessary property for penetration into the brain. The metabolic process converts them into more polar, water-soluble metabolites. The liver is also often affected in malignant disease, not least because it is a common site for metastases.

In chronic liver disease, intrahepatic and extrahepatic anastomoses (‘shunts’) exist between the portal and systemic circulation. This means that drugs which normally have a low bioavailability because of pre-systemic metabolism in the liver can bypass that metabolic process and thus achieve greatly increased oral bioavailability. The bioavailability of modified release morphine and tramadol are markedly increased in patients with both primary and secondary liver tumours, and clearance was reduced and elimination impaired in one study of patients with primary liver tumours( 44 , 45 ). Increased bioavailability has been shown to occur also with pentazocine, pethidine(46 ), chlormethiazole(47 ), and many other drugs not used in palliative medicine, and may be expected to occur with methadone(48 ), metoclopramide(49 ), and others. Reduced metabolite formation of several drugs in the absence of a reduction in systemic clearance in patients with hepatic metastases suggests that such shunting can also take place within tumour deposits inside the liver(50 ). This has implications for the oral dosing of drugs with normally low bioavailability in malignant disease, particularly for patients with primary liver tumours where lengthening of dose interval as well as dose reduction may be necessary(45 ).

In chronic liver disease, the total mass of functioning hepatocytes is reduced and therefore it is not surprising that drug metabolism is generally impaired. However, normal inter-individual variation in drug metabolism rates is so wide that an effect may only be evident when liver disease is severe. The level of serum albumin has been shown to correlate as closely as any parameter to the degree of pharmacokinetic disturbance(51 ). The situation is further complicated by an apparent differential effect on the enzyme system involved. Thus, glucuronidation seems to be relatively protected from the effects of liver disease compared to oxidation and demethylation and there is also evidence of a differential effect on the subfamilies of cytochrome P450(50 ).

Hepatic drug metabolizing capacity is not just susceptible to local effects within that organ but widespread pathophysiological changes may affect drug clearance rates, for example, acute febrile illnesses and endocrine abnormalities may impair drug metabolism.

Some drugs rely on conversion within the liver to their active moiety, for example, prednisone, methylprednisone, and many of the angiotensin-converting enzyme inhibitors. This may render some drugs less effective than others in the presence of liver disease and influence the choice of drug in this situation(50 ).

In malignant disease of the liver there appears to be no overall loss of functional hepatocytes so that systemic clearance is usually unimpaired(52 ). Some more detailed studies of cytochrome P450 subfamilies have emerged; it has been shown that CYP1A2 and CYP2E1 are decreased in cirrhosis but not in hepatocellular carcinoma(53 ) and there is evidence that cytochrome P450s of the 2C subfamily may actually be up-regulated (i.e. have increased activity) in patients with carcinoma(54 ).

Not only renal and hepatic failure but also other organ failures can cause changes in distribution volume either through reducing plasma proteins to which drugs bind or through a qualitative change in the binding sites. Tissue binding may also be affected and the changes in body composition in relation to organ failure, and cachexia may change a drug’s distribution volume according to whether it is distributed mainly in water or in lipid tissue. Although poorly studied, the kinetic profile of many drugs is likely to be abnormal in the presence of advanced malignant disease.

In disease states, pharmacodynamic mechanisms may cause altered response. Increased sensitivity to centrally acting agents and those acting on the cardiovascular system is common. Particular care is needed when both pharmacodynamic and pharmacokinetic changes are potentially occurring in the same patient, for example, when using drugs with sedating properties in patients with hepatic impairment or respiratory insufficiency.

#### Ageing

Many of the changes discussed relevant to disease processes can be applied to the elderly, for after the age of about 65 years there is a gradual decline in renal and hepatic function. Body composition changes so that there is an increase in lipid in relation to total body weight and plasma albumin gradually declines. However, it should be emphasized that changes in pharmacokinetics are quite small and are only detectable in group studies. The most notable change associated with ageing is an increase in variability so that no assumptions can be made about reduced doses being required. Titration of dose to patients’ requirements must be carried out more carefully in the aged.

The effect of the ageing process on hepatic drug metabolism has been extensively studied. Total hepatic mass decreases with age, and much of the documented decline in drug metabolizing capacity has been attributed to the reduction in total functional hepatocyte mass(55 ). It appears from studies with antipyrine and other agents that only microsomally-located enzyme systems decline in activity with age but the results of studies are inconsistent(56 ), and some authors estimate that only 3 per cent of total variance in drug metabolic rate could be attributed to ageing(57 ). A further source of debate is the response of the cytochrome P450 system to enzyme inhibitors and inducers with obvious implications for drug interaction. There appears little doubt that cimetidine, and presumably other enzyme inhibitors, have a similar effect in the elderly compared to the young(58 ). Earlier studies had suggested that microsomal enzymes in the elderly were incapable of being induced(59 ) (see following paragraphs) but this suggestion is now completely refuted(60 ) and it is clear that elderly patients are at risk of drug interaction through these mechanisms in the same way as the young.

Reduced cardiovascular and other homeostatic mechanisms and reduced central nervous system function in the aged make this group susceptible to excess effects from diuretics, blood pressure-lowering agents, and central nervous system depressants. The prescription of prochlorperazine for the symptom of dizziness illustrates the potential hazards, for example. The dizziness is usually due to age-related postural hypotension. However, the α-adrenergic blocking properties of the phenothiazine may cause vasodilation and worsen the symptom, the dopamine receptor blocking effect can precipitate parkinsonism and the sedating effect intellectual impairment.

#### Drug interaction

Patients with palliative care needs may already be receiving drugs for a variety of conditions. Some may still be required but others may not be necessary but may be continued because of possible adverse impacts of discontinuing them, physical or psychological. If drugs to relieve symptoms are added, this creates an enormous potential for drug interaction. Interaction is adverse if it causes therapeutic failure or toxicity from any one drug. It may be regarded, therefore, as simply another source of variability in drug response. Remembering all the possible drug interactions is virtually impossible; so frequent consultation with prescribing information is important. However, knowledge of the underlying mechanisms of drug interaction can put the prescriber on guard.

In broad terms, drug interactions are either pharmacokinetic or pharmacodynamic. Kinetic interaction results in a change in the total body exposure to the drug reflected in a change in blood concentration. The effect is entirely predictable from that change. However, the existence of a pharmacokinetic interaction cannot be predicted easily; every drug has to be studied before its potential for kinetic interaction can be recognized.

The site of pharmacodynamic interaction is the receptor and so it follows that the consequences of such interactions are common to pharmacological groupings of drugs and are therefore to some extent predictable from a working knowledge of each drug group. Kinetic interaction arises through alterations in the rate and extent of absorption and changes in metabolism (both pre-systemic and elimination), distribution, and renal excretion. Drugs such as metoclopramide, anticholinergics(61 ), and opioid analgesics(62 ), which alter the rate of gastric emptying, will affect the speed of absorption of other agents. The resultant effects may not be easily predictable. For example, food increases the bioavailability of morphine(63 ) and metoclopramide increases morphine’s rate of absorption and sedative effects(64 ). Some drugs bind others in the gastrointestinal tract and affect their bioavailability. Certain antacid preparations, iron salts, and cholestyramine are the worst offenders and care is necessary in the use of these drugs concurrently with others.

Of least importance are interactions due to changes in distribution. This is because volume of distribution is not a determinant of steady-state concentration, although an acute change could cause a temporary effect. In the past, much was made of interactions between drugs as a result of plasma protein-binding displacement but it is now known this is rarely a problem. Even if a drug is very heavily bound to protein, displacement will result in only a temporary rise in the free (unbound) fraction because of immediate compensatory mechanisms. There will be wider distribution throughout the distribution volume and the first-order nature of the elimination process results in increased removal of drug. Most clinically significant interactions previously ascribed to protein-binding displacement have now been explained by enzyme inhibition(65 ).

Drug interactions resulting from changes in the rate of metabolism by the liver will result in both changes in bioavailability for those drugs with a significant first-pass effect, and decreased clearance. Steady-state concentrations of drug may be profoundly affected. A number of drugs (particularly the barbiturates, carbamazepine, phenytoin, and rifampicin) are capable of inducing the mixed function oxidase and glucuronidase enzyme systems in the liver. The process involves hypertrophy of the endoplasmic reticulum and takes some weeks to be fully achieved. There is a myriad of substrates for this interaction, amongst them warfarin, corticosteroids, the oral contraceptive, and anticonvulsant drugs. Serum methadone levels can be reduced by the concurrent use of carbamazepine, phenobarbitone, and phenytoin(66 ) and by rifampicin(67 ). The oestrogen component of the oral contraceptive has been shown to double the clearance of morphine by induction of glucuronyl transferase, suggesting the need for increased doses in patients on oestrogens(68 ). The plasma clearance of dexamethasone can be trebled in patients receiving concurrent phenytoin, phenobarbitone, or rifampicin, and presumably other enzyme inducers, leading to a reduced bioavailability and shortened half-life(69 ). This may be important clinically in the care of patients with malignancy.

Our understanding of the process of inhibition of the mixed function oxidase system by drugs increased dramatically with the identification of the subfamilies of cytochrome P450. Previously, there was no explanation as to why one drug might reduce the clearance of a second but not that of a third. It is now clear that some drugs inhibit specific sub-families whilst others, such as cimetidine, are capable of inhibiting all forms of cytochrome P450; this is why cimetidine causes interactions with so many other agents. Cimetidine has been reported to precipitate apnoea in patients taking methadone(70 ) and to have had a similar effect in a patient taking morphine(71 ); however, others have found only a small and insignificant effect on morphine kinetics(72 ). Other H2 receptor blocking drugs have no enzyme inhibitory effect and the proton pump inhibitor omeprazole has a modest and rather unpredictable effect.

Clomipramine and amitriptyline increase the bioavailability and reduce the clearance of morphine and enhance the analgesic effect. The mechanism seems to be both kinetic, through enzyme inhibition, and pharmacodynamic, through the antidepressants’ analgesic effects(73 ). Enzyme inhibition tends to occur rapidly, so the full effect is seen after four or five of the newly prolonged half-lives. The list of drugs that cause interaction through enzyme inhibition is long but non-steroidal anti-inflammatory drugs, especially azapropazone(74 ), and some of the opioid analgesics, such as dextropropoxyphene(75 ), and tramadol(76 ), are implicated.

Interactions between drugs and foods due to hepatic enzyme induction and inhibition are now recognized. Flavonoids in cabbage and broccoli, for example, may induce some cytochrome P450 sub-families whereas a component of grapefruit juice can inhibit them. These interactions may reach clinical significance when large quantities of these products are taken.

The most important drug interactions in the kidney involve competition between agents for active tubular secretion. Active tubular secretion is used by organic acids, and the most frequent interactions are caused by the loop diuretics and some non-steroidal anti-inflammatory drugs. The renal excretion of methotrexate may be inhibited by some non-steroidal anti-inflammatory drugs through this mechanism(77 ). Although renal excretion of some drugs is pH dependent, in general this has minor implications in normal therapeutics. There are a few exceptions; for example, methadone’s renal clearance is considerably enhanced by concurrent use of urinary acidifiers such as acetazolamide(78 ). Whilst acetazolamide is used rarely now, this interaction may be relevant with other carbonic anhydrase inhibitors, although it has not been reported.

In palliative medicine, the most important pharmacodynamic interactions involve intracerebral mechanisms. Drugs that cause sedation also have the potential to cause confusion. Not only will the effects of central nervous system depressants be additive, but if the ionic or metabolic environment is deranged by other drugs, such as diuretics, the problem may be compounded. For example, phenothiazines increase the risk of respiratory depression, sedation, and hypotension(79 ).

### Polypharmacy

Polypharmacy is endemic in palliative medicine. A survey of 385 patients three weeks after referral to a palliative care service found that the median number of drugs per patient was five, with a maximum of 11(80 ). These numbers are similar to those reported in other elderly patient populations, particularly those admitted to hospital with drug-induced illness(81 ). There are several causes of polypharmacy(82 ). Patients with advanced cancer invariably suffer several symptoms, many of which will be amenable to drug therapy. These symptoms may justify the use of combinations of drugs, which may lead to polypharmacy, even in leading specialist centres(80 ). The overriding principle should be to avoid unnecessary duplicate prescribing and to be aware of the potential consequences of multiple drug use. Clearly, the more drugs employed at any one time the greater the likelihood of drug interaction. Sometimes, such interactions will be unpredictable, but often the adverse consequences of polypharmacy are both predictable and avoidable. The increasing availability of clinical guidelines and drug formularies should improve prescribing habits and encourage the use of the most simple effective regimens possible.

The adverse effects of one symptomatic remedy may be self-limiting, both in terms of severity and duration, so that any additional treatment introduced to deal with these adverse effects should be reviewed and stopped if possible. Nausea associated with morphine, for example, is usually an initiation side effect, if it occurs at all, and may need specific treatment only for a few days. Trials of therapy for specific symptoms should be encouraged and the treatment changed if it is ineffective rather than being continued in conjunction with a new drug.

Apart from avoiding duplicate prescribing of similar drugs, other important principles of good practice should reduce the tendency towards polypharmacy. The drug regimen should be reviewed regularly and potentially redundant treatments identified. Both patient and physician may need to be persuaded of the benefit of stopping drugs. These conversations can be difficult, since suggesting drugs taken to reduce long-term harm could be stopped requires the physician to talk about the patient’s shortened prognosis.

### Use of drugs for unlicensed indications

Legislation regarding the manufacture, marketing, and medical use of therapeutic drugs varies throughout the world. Pharmaceutical companies can only market drugs that have an appropriate marketing authorization or product licence. The conditions for licensed use of a drug will be laid out in the Summary of Product Characteristics, such as indications for use, patient populations, formulations, and dosages of the drug. In contrast, doctors are permitted to prescribe drugs beyond these specific criteria. This practice is known as ‘unlicensed’, ‘off licence’, or ‘off label’ use.

Table 8.2 Factors associated with adherence to drug treatment.

 Patient-related factors: ♦ Personality♦ Health care beliefs• General• Drug therapy♦ Physical impairment♦ Cognitive impairment♦ Social support mechanisms♦ Impact of disease or symptom Drug-related factors: ♦ Efficacy of drug♦ Tolerability of drug♦ Drug regimen• Formulation• Dose• Frequency• Duration of treatment♦ Concomitant drug treatment Prescriber-related factors: ♦ Information supplied♦ Monitoring provided♦ Prescriber–patient relationship

Unlicensed use of drugs is common in palliative medicine and in other fragile populations, such as children and the elderly, because the clinical research necessary for licensing is difficult to perform in these patients. In observational studies from two inpatient palliative care units in the United Kingdom about 15 per cent of all prescribed drugs were being used for an unlicensed indication(1 ), and 12 per cent of all prescribed drugs were being given via an unlicensed route, i.e. subcutaneously(2 ). Indeed, a more recent study reports that 40 per cent of all medication prescribed for the five most common symptoms in one unit was unlicensed, and that staff were often unaware they were prescribing outside the product licence(83 ).

The use of unlicensed drugs has implications for clinical governance and medical litigation. Thus, the decision to use an unlicensed drug should be influenced by the:

• risk:benefit ratio for the patient;

• strength of evidence for the use of the drug;

• availability of alternative pharmacological therapies;

• availability of alternative non-pharmacological therapies;

• practice of other palliative medicine physicians.

It has been recommended that consent should be obtained before starting the drug, that the reasons for prescribing the drug are recorded in the clinical notes, and that these reasons are conveyed to other members of the health-care team(84 ). It is important to stress, however, that prescribing of drugs off label is not poor practice or ‘experimental’; the reasons no licence will have been obtained for the use of the drug in the given population will therefore need to be explained. However, a recent questionnaire survey suggested 69 per cent of paediatricians would not seek informed consent for the unlicensed use of a drug in a child(85 ), and it is still not routine practice to obtain consent before starting an unlicensed drug in palliative care in the United Kingdom(86 ). The reasons given for this include the impracticality of obtaining consent, the potential impact on compliance, and the possibility of causing distress to patients and carers. In clinical practice there seems to be a preference for not burdening patients, carers, or parents with the information, or informing them that a drug is being used off label, rather than seeking informed consent.

### Sources of drug information

There are a large number of different sources of information on the drugs used in palliative medicine, including journals, textbooks, national pharmacopoeias, drug information units, drug regulatory authorities, and pharmaceutical companies. Many sources of information are generic in nature, and provide limited information specifically about the uses of drugs in palliative medicine. However, some sources are particular to palliative medicine, such as the Palliative Care Formulary(87 ) and the Hospice and Palliative Care Formulary, USA. The UK Palliative Care Formulary is modelled on the British National Formulary and provides similar information such as formulations, cautions and side effects, with additional information about the uses of the drug in palliative medicine.

An increasingly important source of information is the Internet. The Internet has a number of advantages, especially ease of access, but also a number of disadvantages, particularly the lack of peer review of information(88 ). The Health On the Net Foundation (http://www.hon.ch/) was set up in response to the latter problem. It provides impartial information about the quality of health-care websites; it also produces a voluntary code of conduct for the developers of health-care websites. Table 8.3 gives a list of some useful drug resources on the Internet. These websites are maintained by respected organizations, are freely accessible to health-care professionals, and have links to other relevant websites.

Table 8.3 Selected Internet sources of drug information.

Website content

http://www.bnf.org/

British National Formulary online

http://www.usp.org/USPNF/

The United States Pharmacopeia: National Formulary; public pharmacopoeial standards

The UK electronic Medicines Compendium (eMC) provides electronic Summaries of Product Characteristics (SPCs) and Patient Information Leaflets (PILs)

http://www.palliativedrugs.com/

UK Palliative Care Formulary online

http://www.pallcare.info/

Palliative Care Matters: a generic UK website with syringe driver drug compatibility database

http://nccam.nih.gov/

National Center for Complementary and Alternative Medicine, National Institute of Health, USA.

Clinical pharmacists can also be useful sources of drug information. A study from Australia reported that a clinical pharmacist identified potential problems with the drug regimen of 13 per cent of hospice inpatients, and that changes in the drug regimen resulted in improvements in care in the majority of instances(89 ). The problems identified included inappropriate drug dose or frequency, and potential drug interactions; the problems were predominantly related to drugs being used for general medical, rather than for palliative care, indications.

## Conclusions

The skilful use of drugs to palliate symptoms is essential to the practice of palliative medicine. Individualization of drug and dose and simplicity are fundamental whatever type of treatment is being prescribed. An understanding of the principles outlined in this chapter should facilitate day-to-day management of clinical problems, improve the risk:benefit ratio of drugs used in symptom control, and ultimately contribute to improving the quality of life of patients with advanced disease.

## References

1. Atkinson, C.V. and Kirkham, S.R. (1999). Unlicensed uses for medication in a palliative care unit. Palliative Medicine, 13, 145–52.
Find This Resource

2. Todd, J. and Davies, A. (1999). Use of unlicensed medication in palliative medicine. Palliative Medicine, 13, 446.
Find This Resource

3. Ettinger, D.S., Vitale, P.J., and Trump, D.L. (1979). Important clinical pharmacologic considerations in the use of methadone in cancer patients. Cancer Treatment Reports, 63, 457–9.
Find This Resource

4. Bullingham, R.E.S., McQuay, H., Dwyer, D. et al. (1981). Sublingual buprenorphine used post-operatively: clinical observations and preliminary pharmacokinetic analysis. British Journal of Clinical Pharmacology, 12, 117–22.
Find This Resource

5. Collin, E., Poulain., Gauvain-Piquard, A., Petit, G., Pichard-Leandri, E. (1993). Is disease progression the major factor in morphine ‘tolerance’ in cancer pain treatment? Pain, 55, 319–26.
Find This Resource

6. Fallon, M. (1997). Opioid rotation: does it have a role? Palliative Medicine, 11, 177–8.
Find This Resource

7. Littleton, J. (1997). Drug dependence and drugs of abuse. In Integrated Pharmacology (eds. C.P. Page, M.J. Curtis, M.C. Sutter, et al.), pp. 539–51. London: Mosby.
Find This Resource

8. Kanner, R.M. and Foley, K. (1981). Patterns of narcotic use in a cancer pain clinic. Annals of New York Academy of Science, 362, 161–72.
Find This Resource

9. Standing Medical Advisory Committee Sub-Group on Antimicrobial Resistance. (1998). The Path of Least Resistance, London: Department of Health.
Find This Resource

10. Ali, S., Sykes, N., Flock, P. et al. (2005). An investigation of MRSA infection in a hospice. Palliative Medicine, 19, 188–96.
Find This Resource

11. Prentice, W., Dunlop, R., Armes, P.J. et al. (1998). Methicillin resistant Staphylococcus aureus infection in palliative care. Palliative Medicine, 12, 443–9.
Find This Resource

12. Souhami, R.L., Tannock, I., Hohenberger, P. et al. (eds.) (2002). Oxford Textbook of Oncology, 2nd edition. Oxford: Oxford University Press.
Find This Resource

13. Savarese, J., Goldenheim, P.D., Thomas, G.B. et al. (1986). Steady-state pharmacokinetics of controlled release oral morphine sulphate in healthy subjects. Clinical Pharmacokinetics, 11, 505–10.
Find This Resource

14. Hanks, G.W., Nugent, M., Higgs, C.M. et al. OTFC Multicentre Study Group. (2004). Oral transmucosal fentanyl citrate in the management of breakthrough pain in cancer: an open, multicentre, dose-titration and long-term use study. Palliative Medicine, 18, 698–704.
Find This Resource

15. Portenoy, R.K., Taylor, D., Messina, J. et al. (2006). A randomised, placebo-controlled study of fentanyl buccal tablet for breakthrough pain in opioid-treated patients with cancer. Clinical Journal of Pain, 22, 805–11.
Find This Resource

16. Scott, R.C., Besag, F.M., Neville, B.G. (1999). Buccal midazolam and rectal diazepam for treatment of prolonged seizures in childhood and adolescence: a randomised trial. Lancet, 353, 623–6.
Find This Resource

17. Gilbar, P.J. (1999). A guide to enteral drug administration in palliative care. Journal of Pain and Symptom Management, 17, 197–207.
Find This Resource

18. Warren, D.E. (1996). Practical use of rectal medications in palliative care. Journal of Pain and Symptom Management, 11, 378–87.
Find This Resource

19. Wilcock, A., Jacob, J.K., Charlesworth, S. et al. (2006). Drugs given by a syringe driver: a prospective multicentre survey of palliative care services in the UK. Palliative Medicine, 20, 661–64.
Find This Resource

20. Grassby, P.F. and Hutchings, L. (1997). Drug combinations in syringe drivers: the compatibility and stability of diamorphine with cyclizine and haloperidol. Palliative Medicine, 11, 217–24.
Find This Resource

21. Krajnik, M., Zylicz, Z., Finlay, I. et al. (1999). Potential use of opioids in palliative care – report of 6 cases. Pain, 80, 121–5.
Find This Resource

22. Klepstad, P., Kaasa, S., Cherny, N. et al. and the Research Steering Committee of the EAPC. (2005). Pain and pain treatments in European palliative care units. A cross sectional survey from the European Association for Palliative Care Research Network. Palliative Medicine, 19, 477–84.
Find This Resource

23. Cates, C.J., Crilly, J.A., and Rowe, B.H. (2006). Holding chambers (spacers) versus nebulisers for beta-agonist treatment of acute asthma Cochrane Database of Systematic Reviews, 19, CD000052.
Find This Resource

24. Zepetella, G. (2000). An assessment of the safety, efficacy, and acceptability of intranasal fentanyl citrate in the management of cancer-related breakthrough pain: a pilot study. Journal of Pain and Symptom Management, 20, 253.
Find This Resource

25. Borland, M., Jacobs, I., King, B. et al. (2007). A randomized controlled trial comparing intranasal fentanyl to intravenous morphine for managing acute pain in children in the emergency department. Annals of Emergency Medicine, 49, 335–40.
Find This Resource

26. Kendall, J.M., Reeves, B.C., Latter, V.S. Nasal Diamorphine Trial Group. (2001). Multicentre randomised controlled trial of nasal diamorphine for analgesia in children and teenagers with clinical fractures. British Medical Journal, 322, 261–5.
Find This Resource

27. Amesbury, B., O’Riordan, J., and Dolin, S. (1999). The use of intrapleural analgesia using bupivicaine for pain relief in advanced cancer. Palliative Medicine, 13, 153–8.
Find This Resource

28. Shabat, S., Stern, A., Kollender, Y. et al. (2001). Continous intra-articular patient-controlled analgesia in a cancer patient with pathological hip fracture. A case report. Acta Orthopaedica Belgica, 67, 304–6.
Find This Resource

29. Wang, M.Y., Teitelbaum, G.P., Loskota, W.J. et al. (2000). Brachial plexus catheter reservoir for the treatment of upper-extremity cancer pain: a technical case report. Neurosurgery, 46, 1009–12.
Find This Resource

30. Evans, D.A.P., Manley, K.A., and McKusick, V.A. (1960). Genetic control of isoniazid metabolism in man. British Medical Journal, 2, 485–91.
Find This Resource

31. Sim, E. and Hickman, D. (1991). Polymorphism in human N-acetyltransferase - the case for the missing allele. Trends In Pharmacological Science, 12, 1211–13.
Find This Resource

32. Ellard, G.A. and Gammon, P.T. (1977). Acetylator phenotyping of tuberculosis patients using matrix isoniazid on sulphadimidine and its prognostic significance for treatment with several intermittent isoniazid-containing regimens. British Journal of Clinical Pharmacology, 4, 5–14.
Find This Resource

33. Lunde, P.K.M., Frislid, K., Hansteen, V. (1977). Disease and acetylation polymorphism. Clinical Pharmacokinetics, 2, 182.
Find This Resource

34. Vesell, E.S. Pharmacogenetics. (1975). Biochemical Pharmacology, 24, 445–50.
Find This Resource

35. Steiner, E., Iselius, L., Alvan, G. et al. (1985). A family study of genetic and environmental factors determining polymorphic hydroxylation of debrisoquine in man. Clinical Pharmacology and Therapeutics, 38, 394–401.
Find This Resource

36. Eichelbaum, M. and Gross, A.S. (1990). The genetic polymorphism of debrisoquine/sparteine metabolism - clinical aspects. Clinical Pharmacology and Therapeutics, 46, 377–94.
Find This Resource

37. Gonzalez, F.J. and Meyer, U.A. (1991). Molecular genetics of the debrisoquine/sparteine polymorphism. Clinical Pharmacology and Therapeutics, 50, 233–8.
Find This Resource

38. Chen, Z.R., Somogyi, A.A., Reynolds, G. et al. (1991). Disposition and metabolism of codeine after single and chronic doses in one poor and seven extensive metabolisers. British Journal of Clinical Pharmacology, 31, 381–390.
Find This Resource

39. Gibson, G.G. and Skett, P. (1994). Introduction to Drug Metabolism. 2nd edition. Glasgow: Blackie Academic and Professional.
Find This Resource

40. De Fazio, S., Gallelli, L., De Siena, A. et al. (2008). Role of CYP3A5 in abnormal clearance of methadone. Annals of Pharmacotherapy, 42, 893–7.
Find This Resource

41. Klepstad, P., Rakvag, T.T., Kaasa, S. et al. (2004). The 118 A > G polymorphism in the human mu-opioid receptor gene may increase morphine requirements in patients with pain caused by malignant disease. Acta Anaesthesiologica Scandinavica, 48, 1232–1239.
Find This Resource

42. Rakvag, T.T., Klepstad, P., Baar, C. et al. (2005). The Val158Met polymorphism of the human catechol-O-methyltransferase (COMT) gene may influence morphine requirements in cancer pain patients. Pain, 116, 73–78.
Find This Resource

43. Anonymous. (1999). Pharmacogenetics. In A Textbook of Clinical Pharmacology (eds. J.M.Ritter, L.D. Lewis and T.G.K. Mant), 4th edition, pp. 108–18. London: Arnold.
Find This Resource

44. Kotb, H.I., El-Kady, S.A., Emara, S.E. et al. (2005). Pharmacokinetics of controlled release morphine (MST) in patients with liver carcinoma. British Journal of Anaesthesia, 94, 95–9.
Find This Resource

45. Kotb, H.I., Fouad, I.A., Fares, K.M. et al. (2008). Pharmacokinetics of tramadol in patients with liver cancer. Journal of Opioid Management, 4, 99–104.
Find This Resource

46. Pond, S.M., Tong, T., Benowitz, N.L. et al. (1980). Enhanced bioavailability of pethidine and pentazocine in patients with cirrhosis of the liver. Australian and New Zealand Journal of Medicine, 10, 515.
Find This Resource

47. Pentikainen, P.J., Neuvonen, P.J., and Jostell, K.G. (1980). Pharmacokinetics of chlormethiazole in healthy volunteers and patients with cirrhosis of the liver. European Journal of Pharmacology, 17, 275.
Find This Resource

48. Inturrusi, C.E. and Verebely, K. (1972). Disposition of methadone in man after a single oral dose. Clinical Pharmacology and Therapeutics, 13, 923–30.
Find This Resource

49. Bateman, D.N. (1983). Clinical pharmacokinetics of metoclopramide. Clinical Pharmacokinetics, 8, 523–9.
Find This Resource

50. Morgan, D.J. and McLean, A.J. (1995). Clinical pharmacokinetic and pharmacodynamic considerations in patients with liver disease. Clinical Pharmacokinetics, 29, 370–91.
Find This Resource

51. Homeida, M., Jackson, L., and Roberts, C.J.C. (1978). Decreased first pass metabolism of labetalol in chronic liver disease. British Medical Journal, 2, 1048.
Find This Resource

52. Robertz-Vaupel, G.M., Lindecken, K.D., Edeki, T. et al. (1992). Disposition of antipyrine inpatients with extensive metastatic liver disease. European Journal of Clinical Pharmacology, 42, 465–9.
Find This Resource

53. Guengerich, F.P. and Turvy, C.G. (1991). Comparison of levels of several human microsomal cytochrome P450 enzymes and epoxide hydrolase in normal and disease states using immunochemical analysis of surgical liver samples. Journal of Pharmacology and Experimental Therapeutics, 256, 1189–91.
Find This Resource

54. Murray, M. (1992). P450 enzymes: inhibition mechanisms, genetic regulation and effects of liver disease. Clinical Pharmacokinetics, 23, 132–46.
Find This Resource

55. Swift, C.G., Homeida, M., Halliwell, M. et al. (1985). Antipyrine disposition and liver size in the elderly. European Journal of Clinical Pharmacology, 20, 119–28.
Find This Resource

56. Durnas, C., Loi, C.M., and Cusack, B.J. (1990). Hepatic drug metabolism and aging. Clinical Pharmacokinetics, 19, 359–89.
Find This Resource

57. Vestal, R.E., Norris, A.H., Tobin, J.D. et al. (1975). Antipyrine metabolism in man: influence of age, alcohol, caffeine and smoking. Clinical Pharmacology and Therapeutics, 18, 425–32.
Find This Resource

58. Feely, J., Pareira, I., Guy, E. et al. (1984). Factors affecting the response to inhibition of drug metabolism by cimetidine - dose response and sensitivity of elderly and induced subjects. British Journal of Clinical Pharmacology, 17, 77–81.
Find This Resource

59. Salem, S.A.M., Rajjayabun, P., Shepherd, A.M.M. et al. (1978). Reduced induction of drug metabolism in the elderly. Age and Ageing, 7, 68–73.
Find This Resource

60. Pearson, M.W. and Roberts, C.J.C. (1984). Drug induction of hepatic enzymes in the elderly. Age and Ageing, 13, 313–16.
Find This Resource

61. Nimmo, J., Heading, R.C., Tothill, P., et al. (1973). Pharmacological modification of gastric emptying: effects of propantheline and metoclopramide on paracetamol absorption. British Medical Journal, 1, 587–9.
Find This Resource

62. Nimmo, W.S., Heading, R.C., Wilson, J. et al. (1975). Inhibition of gastric emptying and drug absorption by narcotic analgesics. British Journal of Clinical Pharmacology, 2, 509–13.
Find This Resource

63. Gourlay, G.K., Plummer, J.L., Cherry, D.A. et al. (1989). Influence of a high fat meal on the absorption of morphine from oral solutions. Clinical Pharmacology and Therapeutics, 46, 463–8.
Find This Resource

64. Manara, A.R., Shelley, M.P., Quinn, K. et al. (1988). The effect of metoclopramide on the absorption of oral controlled release morphine. British Journal of Clinical Pharmacology, 25, 518–21.
Find This Resource

65. Kristensen, M.B. (1983). Drug interaction and clinical pharmacokinetics. In Handbook of Clinical Pharmacokinetics. (eds. M. Gibaldi and L. Prescott), pp. 242–64. Auckland: ADIS Health Science Press.
Find This Resource

66. Bell, J., Seves, V., Bowren, P. et al. (1988). The use of serum methadone levels in patients receiving methadone maintenance. Clinical Pharmacology and Therapeutics, 43, 623–9.
Find This Resource

67. Bending, M.R. and Skacel, P.O. (1977). Rifampicin and methadone withdrawal. Lancet, 1, 1211.
Find This Resource

68. Watson, K.J.R., Ghabrial, H., Mashford, M.L. et al. (1986). The oral contraceptive pill increases morphine clearance but does not increase hepatic blood flow. Gastroenterology, 90, 1779.
Find This Resource

69. Anonymous. (1991). Dexamethasone. In Therapeutic Drugs (eds. C. Dollery, A.R. Boobis, D. Burley et al.), pp. D44–50. Edinburgh: Churchill Livingstone.
Find This Resource

70. Dawson, G.W. and Vestal, R.E. (1984). Cimetidine inhibits the in vitro N-demethylation of methadone. Research Communications in Chemical Pathology and Pharmacology, 46, 301–4.
Find This Resource

71. Fine, A. and Churchill, D.N. Potential lethal interaction of cimetidine and morphine. (1981). Canadian Medical Association Journal, 124, 1434.
Find This Resource

72. Lam, A.M. and Clement, J.L. (1984). Effect of cimetidine pre-medication on morphine-induced ventilatory depression. Canadian Anaesthetic Society Journal, 31, 36–43.
Find This Resource

73. Ventafridda, V., Ripamonti, C., De Conno, F. et al. (1987). Antidepressants increase bioavailability of morphine in cancer patients. Lancet, 1, 1204.
Find This Resource

74. Roberts, C.J.C., Daneshmend, T.K., Macfarlane, D. et al. (1981). Anti-convulsant intoxication precipitated by azapropazone. Postgraduate Medical Journal, 57, 191.
Find This Resource

75. Orme, M. and Breckenridge, A. (1976). Warfarin and distalgesic interaction. British Medical Journal, 1, 200.
Find This Resource

Find This Resource

77. Daly, H.M., Scott, G.L., Boyle, J. et al. (1986). Methotrexate toxicity precipitated by azapropazone. British Journal of Dermatology, 114, 733–5.
Find This Resource

78. Bellward, G.D., Warren, D.M., Howald, W. et al. (1977). Methadone maintenance: effect of urinary pH on renal clearance in chronic high and low doses. Clinical Pharmacology and Therapeutics, 22, 92–9.
Find This Resource

79. Grothe, D.R., Ereshefsky, L., Jann, M.W. et al. (1986). Clinical implication of the neuroleptic–opioid interaction. Drug Intelligence in Clinical Pharmacology, 20, 75–7.
Find This Resource

80. Twycross, R.G., Bergl, S., John, S. et al. (1994). Monitoring drug use in palliative care. Palliative Medicine, 8, 137–43.
Find This Resource

81. Colt, A.G. and Shapiro, A.P. (1989). Drug induced illness as a cause for admission to a community hospital. Journal of the American Geriatric Society, 37, 323–6.
Find This Resource

82. Kroenke, K. (1985). Polypharmacy. Causes, consequences, and cure. The American Journal of Medicine, 79, 149–52.
Find This Resource

83. Verhagen, C.C., Niezink, A.G., Engels, Y.Y. et al. (2008). Off-label use of drugs in pain medicine and palliative care: an algorithm for the assessment of its safe and legal prescription. Pain Practice, 8, 157–163.
Find This Resource

84. Ferner, R.E. (1996). Prescribing licensed medicines for unlicensed indications. Prescribers’ Journal, 36, 73–8.
Find This Resource

85. McLay, J.S., Tanaka, M., Ekins-Daukes, S. et al. (2006). A prospective questionnaire assessment of attitudes and experiences of off label prescribing among hospital based paediatricians. Archives of Disease in Childhood, 91, 584–7.
Find This Resource

86. Pavis, H. and Wilcock, A. (2001). Prescribing of drugs for use outside their licence in palliative care: survey of specialists in the United Kingdom. British Medical Journal, 323, 484–5.
Find This Resource

87. Twycross, R., Wilcock, A., Charlesworth, S. et al. (2002). Palliative Care Formulary, 2nd edition. Oxford: Radcliffe Medical Press.
Find This Resource

88. Pereira, J. and Bruera, E. (1998). The Internet as a resource for palliative care and hospice: a review and proposals. Journal of Pain and Symptom Management, 16, 59–68.
Find This Resource

89. Lucas, C., Glare, P.A., and Sykes, J.V. (1997). Contribution of a liaison clinical pharmacist to an inpatient palliative care unit. Palliative Medicine, 11, 209–16.
Find This Resource