Clinical pharmacology in the geriatric patient

Authors

  • Sarah N. Hilmer,

    Corresponding author
    1. Departments of Clinical Pharmacology and Aged Care and Rehabilitation Medicine, Royal North Shore Hospital and the University of Sydney, St Leonards, NSW 2065, Australia
      *Correspondence and reprints: shilmer@med.usyd.edu.au
    Search for more papers by this author
  • Andrew J. McLachlan,

    1. Centre for Education and Research on Ageing (CERA) Concord RG Hospital and the University of Sydney, Concord, NSW 2139, Australia
    2. Faculty of Pharmacy, University of Sydney, Sydney, NSW 2006, Australia
    Search for more papers by this author
  • David G. Le Couteur

    1. Centre for Education and Research on Ageing (CERA) Concord RG Hospital and the University of Sydney, Concord, NSW 2139, Australia
    Search for more papers by this author

*Correspondence and reprints: shilmer@med.usyd.edu.au

Abstract

Geriatric patients are a subset of older people with multiple comorbidities that usually have significant functional implications. Geriatric patients have impaired homeostasis and wide inter-individual variability. Comprehensive geriatric assessment captures the complexity of the problems that characterize frail older patients and can be used to guide management, including prescribing. Prescribing for geriatric patients requires an understanding of the efficacy of the medication in frail older people, assessment of the risk of adverse drug events, discussion of the harm:benefit ratio with the patient, a decision about the dose regime and careful monitoring of the patient's response. This requires evaluation of evidence from clinical trials, application of the evidence to frail older people through an understanding of changes in pharmacokinetics and pharmacodynamics, and attention to medication management issues. Given that most disease occurs in older people, and that older people are the major recipients of drug therapy in the Western world, increased research and a better evidence base is essential to guide clinicians who manage geriatric patients.

The demand for expertise in clinical pharmacology and geriatric medicine grew exponentially at the end of the last century, reflecting the dramatic increase in medication usage and the ageing population. The interplay between these specialties is critical for modern prescribing because older people are the major users of medications and their responses to medications are highly variable [1,2].

The basic concepts of modern pharmacokinetics and pharmacodynamics were developed in the first half of the 20th century. By the 1970s it was widely recognized that disease states and extremes of age introduce considerable variability in both pharmacokinetic and pharmacodynamic responses [3]. Even so, there is still a very limited evidence base underpinning geriatric prescribing and the complexities of geriatric pharmacology often appear to be underappreciated both in the design of clinical trials and prescribing of medications [2,4]. In this review, we attempt to draw together the principles of geriatric medicine and clinical pharmacology in order to guide prescribing in older people.

Comprehensive assessment – a keystone of geriatric medicine

Regulatory bodies usually consider older people to be those over 65 years of age and as such this definition includes an extremely diverse group of people. Geriatric patients are a subset of frail older people with multiple comorbidities that usually have significant functional implications. Frailty is a poorly defined but increasingly studied condition characterized by high susceptibility to disease, impending decline in physical function and high risk of death [5]. The frailty syndrome includes an excessive reduction of lean body mass, a reduction in walking performance and mobility, and poor endurance associated with a perception of exhaustion and fatigue [5].

Clinicians should develop comprehensive care plans – including decisions on medication use – for frail older patients by integrating comprehensive information on all factors that can affect health status: disability, cognition, comorbidities, social role, psychological state and the availability of services and carers. This multidimensional approach is supported by clinical trials demonstrating that Comprehensive Geriatric Assessment has positive effects on health, functional status and mortality both in the acute hospital setting [6] and in the community [7]. Comprehensive geriatric assessment captures the complexity of problems that characterize frail older persons [8].

General principles of prescribing to geriatric patients

Comprehensive geriatric assessment and the recognition of frailty can assist the clinician in designing effective, multidisciplinary management plans. Broad functional outcomes are usually the major therapeutic goal of such treatment plans in geriatric patients, rather than the specific disease-based outcomes typically investigated in clinical trials. This approach also facilitates assessment of the risk and benefit of prescribing a medication for a particular condition in the context of comorbidity and disability, predicts likely changes in pharmacokinetics and pharmacodynamics, and gives information on what assistance the patient may require to adhere to the optimal medication regime [4].

Appropriate medical management requires a consideration of all possible treatment options for a patient based on the available evidence including non-pharmacological management options. In Australia, the Quality Use of Medicines Framework has identified three key steps in prescribing: (1) decide what the best treatment is (i.e. use non-pharmacological management options first); (2) select medicines wisely (based on the suitability of the patient); and (3) use medicines based on the best evidence (the right dose and duration) (http://www.health.gov.au/internet/wcms/publishing.nsf/Content/nmp-quality.htm). In addition in geriatric patients, prescribing requires a detailed knowledge of the deficits in the evidence base and an appreciation of age-related changes in drug disposition, pharmacodynamic responses and the high prevalence of adverse drug reactions. The following steps broadly guide prescribing in geriatric patients [4]:

  • 1Determine the evidence for efficacy in older subjects. This requires an analysis of clinical trial data with a focus on whether optimal and meaningful outcomes were achieved in older people. In the absence of such evidence, extrapolation of clinical trial data from younger patients or an understanding of disease processes might be of some value in determining efficacy and safety;
  • 2Determine the likelihood of adverse drug events in older subjects. In general adverse drug reaction data are poorly described by clinical trials and especially so in subgroups of older patients. Thus the clinician usually will need to rely on data from post-marketing surveillance. Allowance should be made for the increased prevalence of adverse drug reactions in older people, which is exacerbated as patients receive multiple medications for the management of different medical conditions;
  • 3Discuss the harm:benefit analysis with the patient. Patient autonomy is an important medical ethical principle often underemphasized in prescribing guidelines. Participation of older patients in treatment decisions including medicines presents challenges, not the least of which may be a divergence between the goals of the patient and the prescriber [9];
  • 4Decide on the dose regime. There are many age-related changes in the disposition of and response to medications. However, clinical trial evidence for any requirement for dose adjustment is limited. Furthermore, selection of dose form may be an important issue to achieve optimal drug delivery in a convenient manner;
  • 5Monitor the patient very carefully. The paucity of clinical trial data in frail older patients and the marked increase in the prevalence of adverse drug reactions necessitate close monitoring of the patient. Functional and quality of life outcomes may be more relevant to the older patient than individual primary disease outcomes investigated in clinical trials.

Evidence for efficacy in geriatric patients

Older adults’ definition of successful ageing is multidimensional, encompassing physical, functional, psychological and social health [10]. Ideally, therapeutics should aim to meet these goals, but outcomes in clinical trials rarely address such a wide spectrum of issues. Furthermore endpoints that are relevant to older people such as independence, falls, cognitive impairment and physical function are difficult to measure in clinical trials and rarely assessed as adverse drug reactions.

Physical function is an important outcome for older people, and has been shown to predict disability, nursing home admission and mortality [11]. Benzodiazepine exposure in community dwelling older people has been associated with poorer self-reported functional status [12] and greater decline in objective physical performance tests over 4 years [13]. Potentially inappropriate medication use in older people is currently guided predominantly by expert consensus statements such as the recently updated Beers criteria [14]. However, this exposure does not correlate with decline in self-reported functional status in community dwelling older people [15] or with health outcomes in hospitalized older people [16].

In some studies, angiotensin-converting enzyme (ACE) inhibitors, 3-hydroxy-3-methylglutaryl (HMG) CoA reductase inhibitors and testosterone have been associated with delayed functional decline in older people [17]. However, ACE inhibitors have also been associated with impaired balance, a risk factor for falls [18], HMG CoA reductase inhibitors with myopathy [19], disability [20] and a trend towards increased falls [21], and testosterone with behavioural changes [22]. Treatment of hypertension with antihypertensive medications has been inconsistently associated with prevention of cognitive decline in older people [23–25].

One of the most difficult aspects of interpreting current clinical trials relates to analysing the data obtained in older subgroups. For example the majority of people with heart failure are over the age of 75 years. Yet in this specific subgroup there is no statistically significant effect on outcomes with beta blockers [26] or ACE inhibitors [27]. On one hand, this may be a statistical quirk related to power and subgroup analysis. On the other hand, it is quite plausible that these drugs have no effect in geriatric patients because of changes in comorbidity, life expectancy, disease pathogenesis and pharmacological responses. Similarly, there are well-performed, large trials indicating that bisphosphonates might be ineffective in preventing hip fractures in women over the age of 80 years [28]. Such issues can only be resolved by conducting randomized clinical trials in geriatric subjects, despite the technical difficulties in performing trials in older people. Published trials performed in the very old are rare [29], but have been increasing in recent times [30].

Even when clinical data show statistically significant efficacy in older people, the results still have to be interpreted for clinical relevance. Cholinesterase inhibitors are widely recommended for the symptomatic treatment of moderately severe Alzheimer's disease and the trials were performed in older patients. Clinically perceptible improvement was measured in clinical trials using the CIBIC-Plus scale and was increased in treated subjects (OR 1.56, 95% CI 1.32–1.85). From the clinician's point of view, if he or she treated 100 subjects with Alzheimer's dementia, 24 would improve but in 17 of these patients the improvement would be secondary to the placebo effect. However, 15 subjects (over and above the placebo response) would report side-effects particularly nausea, abdominal pain and diarrhoea and 10 additional subjects would withdraw from therapy because of adverse reactions [31]. Furthermore, recent trials showed an increase in mortality with galantamine therapy in mild cognitive impairment [32]. Understandably, whether such drugs are useful in clinical practice is fiercely debated.

Adverse drug reactions

The prevalence of adverse drug reactions is increased in older people [33–41], and reactions are generally more severe [33,41]. It has been reported that adverse drug reactions are the fourth to sixth greatest cause of death [42]. Between 5% and 10% of hospital admissions amongst older people are related to adverse drug reactions [43–46] and for every dollar spent on medications in nursing facilities for older people, $1.33 is subsequently required for the treatment of drug-related morbidity and mortality [47]. The rates of hospital admissions for adverse drug reactions amongst older people, particularly reactions to cardiovascular medications, have been steadily increasing over the last decade [48].

Even so ageing may not be an independent risk factor for adverse drug reactions but merely a marker for comorbidity, altered pharmacokinetics and the use of multiple medications [36,46,49]. Of all the factors that are most consistently associated with adverse drug reactions, polypharmacy has been considered the most important [33] and indeed, some studies that have used multivariate analysis report that the association between old age and adverse drug reactions is the result of the confounding association between age and polypharmacy [36]. However, it is clear that age-related changes in pharmacodynamics and pharmacokinetics contribute the risk of adverse medicine events [33,40].

Most adverse drug reactions causing admission of older people to hospital are classified as type A reactions and hence are predictable and potentially preventable [40]. In a review and meta-analysis of hospitalizations caused by adverse drug reactions it was concluded that older people are four times more likely to be admitted to hospital as a result of an adverse drug reaction (16.6% vs. 4.1%) and are more likely to have preventable adverse drug reactions (88% vs. 24%). In another study of people over 75 years, 30.4% of admissions were secondary to adverse drug reactions of which one half were considered preventable [50].

However, it is also clear that ageing itself is associated with increased risk of adverse drug reactions to specific classes of drugs that may be independent of polypharmacy and altered pharmacology [34]. The association between old age and non-steroidal anti-inflammatory drug (NSAID)-induced adverse effects has become a major issue recently, particularly with the introduction and widespread use of cyclooxygenase-2 selective NSAID agents. The incidence of upper gastrointestinal haemorrhage or perforation increases substantially with old age in subjects taking NSAIDs [51]. In subjects over 70 years of age, the number needed to harm each year to produce an upper gastrointestinal haemorrhage or perforation is approximately 50 [51]. In addition, older people exposed to NSAIDs have a 1.7-fold increased chance of subsequent antihypertensive therapy [52,53] and an increased prevalence of renal impairment [54]. Even rare and probably idiosyncratic adverse reactions, such as interstitial nephritis and hepatitis associated with H2-receptor antagonists are primarily an issue for older people [55].

The association between falls and medication use has been substantiated by epidemiological and observational studies [56–59]. A systematic review examining the relationship between psychotropic drugs and falls in older people, found that the odds ratio for one or more falls was 1.73 (95% CI 1.52–1.97) for exposure to any psychotropic medication, and there was little difference between the different classes of psychotropic agents with respect to risk [59]. In this regard the newer generation psychotropic drugs have not lived up to expectations related to falls in older people. Selective serotonin inhibitors and newer generation antipyschotic agents appear to have at least an equivalent propensity to falls and fractures in older people as do the older tricyclic antidepressants [60] and first-generation antipsychotic agents [61]. The association between some medications and falls may be partially explained by increased mobility of patients receiving these therapies. The significant problem of hip fracture has been associated with the use of barbiturates [62], benzodiazepines [63], tricyclic antidepressants [64], antipsychotics [64] and selective serotonin reuptake inhibitors [60] in elderly patients. In a systematic review, we estimated that the risk of hip fracture was increased by 50–110% in older subjects receiving benzodiazepines and that up to 10% of hip fractures were directly attributable to benzodiazepine usage [65]. Lower limb muscle weakness has been associated with an OR of 1.76 (95% CI 1.31–2.37) for any fall and 3.06 (95% CI 1.41–5.04) for recurrent falls [66]. However, while not well described, the risk of falls does not appear to be significantly increased by exposure to medications associated with myopathy, such as HMG CoA inhibitors and corticosteroids [21]. The inconsistent association between the number of medications that the patient is exposed to and falls risk will be discussed under polypharmacy.

Confusion is frequently secondary to exposure to medications in older people. In the hospital setting, medications have been reported to be the cause of delirium in 11–30% of cases [67]. Benzodiazepine exposure in community dwelling older people has been associated with memory impairment [68]. Exposure to anticholinergic medications and high serum anticholinergic activity, a measure of peripheral blood anticholinergic burden, have been associated with decreased Mini-Mental State Examination scores in community dwelling older people [69], with non-progressive mild cognitive impairment [70] and with the presence of delirium in older medical inpatients [71].

Polypharmacy, defined as the use of five or more medications, occurs in 20–40% of older people [72–74] and has been linked to poor health outcomes [75]. The risk factors for polypharmacy include old age, comorbidity, recent hospitalization, female gender, depression, number of treating doctors [76] and practitioner characteristics [76].

In addition to using multiple prescribed medications, older people are also major users of complementary and alternative medicines and may not report these without prompting [77–79]. A longitudinal series of surveys of the use of complementary and alternative medicines conducted in South Australia [80] found that 37% of people over the age of 65 years regularly used such medicines. Older people are at great risk of adverse effects and herb–drug interactions when using complementary and alternative medicines [81]. About a third of older patients are at risk of interactions between complementary and alternative medicines and their prescription medicines [77,82] especially some of the herbal medicines promoted for use in the elderly [81,83].

Part of the risk of polypharmacy may be the unintentional practice of prescribing additional drugs for the adverse effects of other drugs –‘prescribing cascade’ [84] or ‘double-dipping’ [85]. For example, the odds ratio for starting antihypertensive treatment for recent users of NSAIDs compared with non-users was 1.66 (95% CI 1.54–1.80) and the risk was dose-dependent [52]. In older patients, the risk of commencing levodopa was increased in subjects taking metoclopramide (odds ratio 3.09, 95% CI 2.25–4.26) [86].

The harm associated with polypharmacy includes increased risks of adverse drug reactions, drug interactions, increased costs and errors in patient adherence to therapy. The prevalence of adverse drug reactions increases with the number of prescribed drugs. The risk of a definite adverse drug reaction is increased in subjects of any age taking four or more medications (odds ratio 2.94, 95% CI 2.38–3.6) [36] and the prevalence of any adverse drug reaction in hospital inpatients was 18.6% for those taking one to five drugs compared with 81.4% among those on six or more medications [37]. The risk of falling and of recurrent falls was doubled in those subjects taking four or more medications [58]. However, a recent study reported that falls are associated with increasing numbers of chronic diseases rather than with polypharmacy [87]. Some of the hazards attributed to polypharmacy may be related to the underlying comorbidities for which the medications are prescribed.

It is important to determine the potential benefits of polypharmacy in particular settings before dismissing it as entirely inappropriate. Clinical trials tend to exclude subjects with comorbidity and polypharmacy. If subjects with multiple conditions are studied, often the intervention and outcome focus on a single disorder. For example, there is good evidence for the benefit of polypharmacy from clinical trials in subjects with diabetes mellitus. The use of antihyperlipidemic agents, antihypertensives, antiplatelet agents and ACE inhibitors have all been shown to have considerable mortality benefits in these diabetic patients [88,89].

Medication withdrawal

There is evidence for the benefit of reducing exposure to some classes of medications in older people. In a randomized placebo-controlled trial of withdrawal of psychotropic medications in older subjects taking, on average, 5–6.5 medications each, it was found that there was a 76% reduction in falls over 44 weeks (odds ratio 0.34, 95% CI 0.16–0.74) [90]. There have been at least four trials investigating the effect of withdrawal of antihypertensive medications in older people. Overall, 40% of subjects remain normotensive, particularly if weight loss and salt restriction are implemented [91]. One study of 333 elderly (70–84 years) hypertensive patients found that antihypertensive therapy could be withdrawn for up to 5 years in 20% of subjects. During the state of ‘no treatment’ subjects had lower total mortality risk than the matched general population of the treated group [92].

Patient satisfaction is often increased when polypharmacy is reduced [93,94]. In a retrospective study of drug withdrawal in older subjects, 238 medications were ceased in 124 subjects for a variety of clinician-based reasons. There were no clinical consequences for nearly three quarters of the medications that had been ceased [95]. However, the risk of postoperative complications was increased in a retrospective analysis of subjects who had medications withdrawn in the peri-operative period [96].

Medication withdrawal is difficult to implement. Prescriber feedback and pharmacist-led medication reviews have been tried [97,98]. General practitioners have also been encouraged to withdraw medications in their older patients with polypharmacy [97]. Most drugs can be stopped without major withdrawal effects but it should be noted that acute withdrawal of benzodiazepines [99] and anticonvulsants can be associated with seizures, beta-blockers with tachycardia and exacerbation of ischaemic heart disease, antidepressants with a well defined withdrawal syndrome [100] and levodopa with neuroleptic malignant syndrome [101].

Although there is little evidence supporting the approach to medication withdrawal, in general a stepwise approach is recommended, with weaning of psychotropic and cardiovascular medications. Research by geriatricians and clinical pharmacologists is needed to develop more sophisticated evidence-based prescribing guidelines than simply counting drugs, to ascertain which medications at what doses improve relevant functional outcomes, which are detrimental, and which can be safely withdrawn.

Medication management in geriatric patients

Continuity of prescribing

Obtaining an accurate medication history and reviewing all of a patient's medications is an essential component of geriatric assessment. Geriatric patients have multiple comorbidities and thus may have their medications prescribed by several doctors. These patients are also frequent users of acute hospitals where medications may be reviewed and altered. Medication histories obtained from older people are often inaccurate and reconciliation of the history from multiple sources is required [102]. Trials of interventions to improve the transfer of medication information from the hospital to the community, such as medication cards [103] or a pharmacist to co-ordinate medication-management [104] have provided some benefits but results remain disappointing. Pharmacist-led home review of medications has also been studied [98]. While the role of electronic data management is currently being investigated, at present the most useful and accurate record of patients’ medications is the ‘plastic bag’ containing all of their medications [102].

Adherence

Typical adherence rates for prescribed medications are about 50% [105] and do not vary significantly with age [106]. Adherence to medications is difficult to measure, poorly recognized, and complex [107]. Barriers to compliance that are common in older people include chronic conditions, polypharmacy, complex regimens, a higher prevalence of adverse drug reactions, unaffordable drug costs, cognitive impairment, visual impairment, manual dexterity problems and dysphagia [106,108]. Dysphagia often results in the need to crush (or reformulate) solid oral doses, which is not appropriate for all dose forms, especially modified release formulations, and new formulations appropriate for older people with swallowing disorders will need to be developed [109–112]. Strategies to improve adherence, which are often multi-factorial and complex, have generally only provided minimal benefit [113].

Monitoring

The ageing process, especially frailty, diminishes homeostatic responses and the capacity to deal effectively with any physiological perturbation, including initiation of a new medication and changes in dose regimes [2]. Therefore, it is imperative that such subjects are closely monitored both for evidence of efficacy (the medication should be ceased if it is ineffective) or adverse drug reactions. Adverse drug reactions present with a wide variety of symptoms and are the ‘great imitators’ of medicine in the 21st century. Medications need to be reviewed regularly in response to rapid changes in clinical status including hospitalization, new diseases, mediation load and transition into a palliative phase.

Age-related changes in pharmacokinetics

Once a decision has been made to prescribe a medication to an older person, a dosage regime needs to be constructed. This is often based upon consideration of age-related changes in pharmacokinetics, although it must be conceded that there are few clinical trials that have addressed the effects of altering dose on safety and efficacy outcomes. Furthermore, it has been argued that inter-individual variability in pharmacokinetics (from all causes) may be far greater than age-related variability [114].

There are many age-related changes in the processes of drug absorption from the gastrointestinal tract, plasma protein binding and drug distribution; however, the implications for drug dosage are minimal [2].

The hepatic metabolism of many drugs is reduced in elderly patients [115]. The extent is variable between drug groups but usually represents a 30–50% reduction in clearance of drugs cleared by phase I hepatic metabolism [115,116]. This appears to be secondary to age-related changes in hepatic blood flow, liver mass and the hepatic endothelium rather than ageing changes in drug metabolizing enzymes or their expression [2,114,117,118]. Age-related reductions in hepatic clearance increase the bioavailability of drugs with a significant first pass effect and reduce the clearance of hepatically metabolized drugs, probably increasing the risk of type A dose-related adverse drug reactions. However, there will also be a decrease in the activation of some pro-drugs, causing reduced or delayed efficacy in elderly patients.

Phase II metabolism via conjugation pathways appears to be maintained in healthy older people but reduced with frailty. Conjugation of paracetamol [119] is reduced in frail but not fit older people. Similarly, esterase activity is reduced in frail but not fit older people [120].

Hepatic transporters have recently been recognized as important determinants of drug disposition, for both uptake and biliary excretion. Variability in the expression and/or function of these transporters has been described with genetic polymorphisms and disease. There are no studies on the expression of transporters in aged human livers, but there is evidence that hepatic expression of P-glycoprotein is increased in old rats, which, if functional, should result in increased biliary excretion [121].

While most drug metabolism occurs in the liver, enzymes in the wall of the gastrointestinal tract, such as CYP450 contribute to the pre-systemic metabolism of a number of drugs. Age-related changes in the expression of these enzymes in the gastrointestinal tract are not well described. Data from animal studies provide some evidence that CYP3A expression is maintained in the intestinal mucosa of old rats [122].

It has been accepted that there is a marked age-related reduction in creatinine clearance in older people, even in the presence of normal serum creatinine concentrations. The Cockcroft Gault equation [123] is often used to estimate the creatinine clearance in older people in order to adjust the maintenance dose of renally excreted drugs that have narrow therapeutic indices, such as aminoglycosides, digoxin and lithium. However, the Cockcroft Gault equation was derived from men being investigated for renal disease. A review of recent studies of healthy older people has shown that in the absence of renal disease, glomerular filtration rate is reasonably maintained into old age [2]. Practitioners have recently started to estimate renal function using the Modification of Diet in Renal Disease equation, which also includes age, and shows a better correlation with accurately measured glomerular filtration rate than creatinine clearance [124]. However, it has not been validated in extremes of age or for adjusting doses of renally excreted drugs [125]. The few studies on the effects of healthy ageing on lithium, gentamicin [126] and digoxin [127] pharmacokinetics have not shown any dramatic reduction in renal clearance independent of changes in renal function. While many studies have identified an effect of age on the clearance of such drugs in patients [128], this is likely to reflect the high prevalence of renal disease in older people, as well as the high prevalence of polypharmacy and potential drug interactions affecting renal excretion.

Rather than relying on generalizations about ageing changes in liver and renal function to define dosage, regulatory authorities now require pharmacokinetic data from older people. In many cases, such data include only relatively ‘young’ old people (e.g. older than 65 years) so that data on the very elderly or frail patient, where altered kinetics are likely to have a major impact, are not available.

Age-related changes in pharmacodynamics

Age-related changes in effector system function result in age-related changes in pharmacodynamics. End-organ response is affected by physiological changes that occur with increasing age in the absence of pathology. Of particular interest to pharmacologists are the age-related changes in calcium channels and beta-adrenergic receptors, with implications for the clinical use of their agonists and antagonists [1]. For example, the beta-adrenergic response decreases with increasing age, and controlling for plasma concentration, the bradycardic response to labetolol is decreased in older people [129].

Older people have been shown to be more sensitive to sedating effects of some medications due to changes in both pharmacokinetics and pharmacodynamics, e.g. they lose consciousness at a dose and lower plasma concentration of propofol at the effector site than younger people [130] and have increased sensitivity to sedation with benzodiazepines such as triazolam [131]. Age-related changes in the autonomic nervous system predispose older people to postural hypotension [132], which may be further exacerbated by medications with anticholinergic effects and antihypertensives [133].

Application of principles of geriatric pharmacology to specific pharmacotherapies

Treatment of cardiac failure with beta blockers

Appropriate treatment of cardiac failure in geriatric patients highlights the hazards of extrapolating evidence from younger adults to geriatric patients. For example, there is substantial clinical trial evidence for the use of beta-blockers in patients with cardiac failure receiving an ACE inhibitor and a recent subgroup analysis of the clinical trials for metoprolol concluded ‘the time has come to overcome the barriers that physicians perceive to beta-blocker treatment and to provide it to the large number of elderly patients with heart failure’. However, in subjects over 75 years (n = 490) the relative risk for primary outcome of hospitalization and all cause mortality was not statistically significant [0.79 (95% CI 0.55–1.14)] nor was total mortality [0.71 (95% CI 0.42–1.19)] [134]. Similarly in the SENIORS study it was concluded that nebivolol is ‘an effective and well tolerated treatment for heart failure in the elderly’ yet the relative risk for the primary outcome (hospitalization and all cause mortality) in the older cohort aged more than 75.2 years was not significant [0.92 (95% CI 0.72–1.12)] [135].

There are many possible reasons for the lack of benefit of beta-blockers in older people with heart failure. Outcomes like hospitalization and mortality are less likely to be influenced by management of a single disease, such as heart failure, in an older person with more comorbidities than a younger person. Pharmacodynamically, with both increasing age and heart failure, there is decreasing responsiveness of beta receptors, with different changes in the second messenger mechanisms [136], and this could contribute to lack of effect of beta-blockers in patients with both old age and heart failure. With increasing age, diastolic failure accounts for a higher proportion of the population with heart failure [137] and even in younger people, the evidence for beta-blockers in heart failure is predominantly in patients with systolic failure. Interestingly, caloric restriction, which delays the ageing process, has been shown to reduce the incidence diastolic heart failure [138]. This suggests that management of cardiac failure in older people may require interventions directed at the ageing process, rather than blanket application of interventions that have only been shown to be efficacious in young- and middle-aged adults.

Cancer chemotherapy

The incidence of most cancers increases with age and prescription of chemotherapy in the geriatric patient requires consideration of the principles of both medical ethics and geriatric pharmacology. There is limited evidence from clinical trials to support the use of chemotherapy in older subjects with cancer, particularly those with comorbidities. Older people are more prone to toxicity from chemotherapeutic agents, due to age-related changes in both pharmacokinetics and pharmacodynamics [8,139]. A comprehensive geriatric assessment approach, including identification of the frailty syndrome, may assist oncologists in identification of older patients who are likely to develop severe toxicity and severe side effects in response to aggressive treatment [8].

Paracetamol for chronic pain

Chronic pain is common in older people and tends to be musculoskeletal [140]. A study of adults aged 65 years and older living in retirement facilities in the USA with persistent pain found that 61% had used paracetamol and management with paracetamol was rated more than moderately helpful in 40% of cases [141]. Studies of pain relief in older people must consider both age-related differences in pain perception and reporting, which are not yet well defined [142], and meaningful outcomes for older people, such as mobility.

Unintentional overdose with paracetamol is accounts for approximately half of the cases of acute liver failure from paracetamol toxicity and cases associated with chronic paracetamol use occur more often in older patients [143]. The increased risk of paracetamol toxicity from chronic use in older people is partly explained by the reduced hepatic clearance. The 30–40% reduction in liver weight in normal ageing must be accounted for in dose calculations. Recognition of frailty is critical, as paracetamol conjugation per unit volume of liver is normal in healthy older people but reduced in the frail [119]. Fasting may also enhance paracetamol hepatotoxicity [144] and malnutrition is common in geriatric patients, affecting up to 15% of community-dwelling older people [145], 23–62% of hospitalized patients in acute wards [146], 50% of older patients in subacute wards [147] and up to 85% of nursing home residents [148]. Toxicology studies in rats do not show an increase in paracetamol hepatotoxicity with increasing age [149]. The clinical increased risk of paracetamol hepatotoxicity in older people is likely to be related to dosing that does not account for decreased liver volume with age, and to frailty and malnutrition.

Warfarin for stroke prevention in atrial fibrillation

The use of warfarin in geriatric patients with atrial fibrillation requires careful analysis of the potential benefits and risks for the individual. On meta-analysis of clinical trials, the benefits of warfarin are well established [150]. Both atrial fibrillation and stroke increase with age, and the association between stroke and atrial fibrillation is maintained in older people [151]. Therefore, the benefits of warfarin for stroke prevention are potentially greater in older people with atrial fibrillation. However, SPAF II [152], a randomized controlled trial to compare aspirin and warfarin for atrial fibrillation specifically in patients over 75 years, found the rate of all stroke (ischaemic or haemorrhagic) with residual deficit was 4.3% per year with aspirin and 4.6% per year with warfarin (RR 1.1). Furthermore, the outcome of anticoagulation in older patients with atrial fibrillation in terms of quality adjusted life years is either negligible or negative [153,154].

Patients aged over 75 years who are prescribed warfarin for atrial fibrillation have a significantly greater risk of major haemorrhage than those less than 75 years [152]. For those over 75 years, the annual rate of major haemorrhage was 4.2% for patients in a randomized trial [152] and 10.0% for frail older people in an observational study [155]. Some of the high risk of haemorrhage with warfarin in geriatric patients may be related to difficulty maintaining optimal levels of anticoagulation [156]. Achieving therapeutic anticoagulation is challenged in older people by compliance with dosing and monitoring requirements, particularly in those with cognitive impairment, increased risk of drug–drug interactions with increased prevalence of polypharmacy, increased inter-individual variability, increased hospitalizations [157] and irregular food and alcohol intake. However, major bleeds occur at lower international normalized ratios with increasing age [158]. The increased risk of haemorrhage in older people may be related to vascular rigidity and endothelial dysfunction, despite an increase in coagulation system proteins and platelet activation with increasing age [159]. Older people have a higher prevalence of gastrointestinal pathology, predisposing them to gastrointestinal bleeding, and are at increased risk of falls, which can cause severe injury in the presence of anticoagulation, particularly from intracerebral bleeds [160].

Warfarin has also recently been associated with osteoporotic fractures [161], which are a functionally important adverse event in older people.

The decision on whether to prescribe warfarin for stroke prevention in an individual geriatric patient with atrial fibrillation depends on the complexity of the individual. A multidisciplinary intervention to optimize antithrombotic use in older patients with atrial fibrillation that captured patient complexity [162] found that the intervention was associated with fewer patients receiving warfarin, after having been assessed inappropriate candidates. Prescribing biases may be introduced by clinical experience. While a prescriber's experience of an adverse bleeding event in a patient on warfarin for atrial fibrillation will decrease subsequent prescribing of warfarin, treating a patient who has a stroke while not anticoagulated does not increase warfarin prescribing for subsequent patients [163]. The prescriber needs to calculate the risks and benefits of anticoagulation for the individual geriatric patient, discuss these with the patient, whose preferences may differ widely from clinical practice guidelines [164], when possible obtain informed consent for the decision, and review the decision frequently.

Conclusions

Most disease occurs in older people, particularly geriatric patients and the frail elderly. These older patients have the greatest potential to benefit from medications, however, observational studies indicate an increased prevalence of adverse drug reactions and some clinical trials have failed to establish therapeutic benefits seen in younger adult subjects. Thus the decision to prescribe medications to geriatric patients requires individual analysis of harm and benefit rather than broad application of prescribing guidelines. To do this the clinician must have a thorough knowledge of the magnitude of the reported risks and benefits of the medications.

Much of geriatric medicine is concerned with the recognition and management of adverse drug reactions and frequently the major intervention is withdrawal of medications rather than the prescription of new medications. This practice may generate short-term gains in terms of function but the long-term consequences of medication withdrawal have not been investigated extensively.

Age-related changes in pharmacokinetics and pharmacodynamics, polypharmacy and numerous comorbidities further contribute to the complexity of drug therapy.

Given that older people are the major recipients of drug therapy in the Western world, increased research and a better evidence base is an imperative to guide clinicians who manage geriatric patients and frail older people.

Acknowledgements

This review was supported by grants from National Health and Medical Research Council of Australia, Healthy Ageing Research Programme of the NHMRC, Ageing and Alzheimer's Research Foundation, Geoff and Elaine Penney Ageing Research Unit.

Ancillary