9: Principles of Prescribing for Older Adults

On the surface, prescribing for older adults is similar to prescribing for younger adults, requiring understanding of drug indications, dosing, potential adverse reactions, and drug–drug interactions. However, prescribing for older adults is complicated by a variety of factors. Physiologic changes as patients get older result in alterations in drug metabolism and susceptibility to adverse events. The presence of multiple chronic conditions and multiple medications leads to potentially complex drug–drug and drug–disease interactions, as well as the need to balance multiple competing recommendations. Changes in cognitive function, manual dexterity, and social supports complicate adherence to medications, and heterogeneous goals of care require special attention. Because clinical trials that inform many practice guidelines are often conducted in younger patients, there can be ambiguity about the extent to which these evidence-based recommendations apply to older adults. Thus, mastering prescribing for older patients requires expertise not only in technical elements of drug use, but also in synthesizing evidence and biomedical and psychosocial factors into a coordinated plan of care that meets each individual’s unique needs. More details about polypharmacy can be found in Chapter 53, and more details about extrapolating the evidence from clinical research to older patients can be found in Chapter 74.

Pharmacokinetics

Pharmacokinetics refers to how the body handles a drug from the time it is ingested to the time it is excreted. This includes the processes of absorption, distribution, metabolism, and elimination. While each of these processes can vary with age, they are typically more influenced by genetic factors and by an individual’s diseases, environment, and other medications. For most older patients, changes in renal function have the greatest impact on pharmacokinetics.

Absorption

Absorption of drugs is impacted by the size of the absorptive surface, gastric pH, splanchnic blood flow, and gastrointestinal (GI) tract motility. Most of these are relatively unaffected by age, but can be substantially affected by certain diseases and medications. Some medications, including vitamin B₁₂, calcium, and iron, have decreased absorption in older adults as a result of reduced activity of active transport mechanisms.

Distribution

Older patients have an increased fat-to-lean body mass ratio, decreased total-body water, and sometimes decreased serum albumin. Drugs that distribute in fat (eg, diazepam) may thus have a larger volume of distribution. Hydrophilic medications (eg, digoxin) will have a decreased volume of distribution, resulting in higher serum levels. Drugs that bind to serum proteins reach an equilibrium between bound (inactive) and free (active) drug. Use of 2 or more drugs that compete for protein binding (eg, thyroid hormone, digoxin, warfarin, phenytoin) can result in higher levels of free drug, requiring careful monitoring of drug levels and effects. In the case of testosterone, age-associated increases in sex-hormone binding globulin can result in normal serum levels of total testosterone, even while levels of serum free testosterone (the bioactive form) are reduced.

Metabolism

The cytochrome P450 system assists with drug metabolism through oxidation and reduction (known as Phase 1 metabolism). There may be an age-related decline in Phase 1 activity as a result of reductions in hepatic blood flow and liver size. Nonetheless, the cytochrome P450 system is typically far more impacted by genetic polymorphisms that result in some individuals being “fast” or “slow” metabolizers, and by the use of other drugs and foods that can inhibit or induce specific P450 enzymes, resulting in slowed or accelerated drug metabolism. (See the section “Adverse Drug Reactions” below for information on cytochrome P450-mediated drug–drug interactions.) Phase II hepatic metabolism, otherwise known as conjugation, follows phase I metabolism. It typically makes drugs biologically inactive and facilitates their excretion. It is not affected by age.

Excretion

Renal function often decreases with age, involving loss of both glomerular filtration rate and tubular function. Because muscle mass declines in older age, renal function can often be substantially impaired even in the presence of a normal serum creatinine. Thus, estimation of creatinine clearance (or the closely related glomerular filtration rate) is essential. Mathematically complex formulas such as the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) and Modification of Diet in Renal Disease (MDRD) formulas tend to more accurately reflect renal function compared to the Cockroft-Gault equation (shown below). However, each is imperfect and should be interpreted as providing only a rough estimate of renal function. In situations of rapidly changing renal function, neither formula performs well. Nonetheless, it is far better to have (and use) an easy-to-obtain rough estimate of renal function than none at all.

(multiply by 0.85 for women to allow for 15% less muscle mass).

Pharmacodynamics

Pharmacodynamics refers to how the drug affects the body; that is, the physiologic effects exerted by a drug’s action on end-organ receptors. Pharmacodynamics has not been as carefully studied in older adults as pharmacokinetics. Older adults are more sensitive to medications that depress the central nervous system, which can result in delirium, confusion, and agitation. The frequent use of multiple medications in older adults often leads to the simultaneous use of 2 or more drugs that have mutually reinforcing physiologic effects, which can result in harm. Examples include bleeding with simultaneous use of anticoagulant drugs (eg, warfarin) and nonsteroidal antiinflammatory drugs (NSAIDs) or aspirin; and orthostatic hypotension with various blood pressure medications and α blockers.

Adverse Drug Reactions

Epidemiology and Risk Factors

Adverse drug reactions (ADRs) are substantially more common in older than in younger adults. Up to 35% of ambulatory older adults experience 1 or more ADRs annually, 5% to 10% of hospital admissions in older adults are ascribable to ADRs, and 5% or more of older hospitalized patients experience an ADR during their inpatient stay. Simply being old does not meaningfully increase ADR risk. Rather, it is the number of medications taken and the burden of disease (which often, but not always, increases as patients get older) that are the strongest risk factors for ADRs in the outpatient setting.

Several observational studies have failed to find a globally increased risk of ADRs in older adults with impaired functional status and geriatric syndromes. However, the risk of ADRs may increase with specific drugs that can interact with specific impairments. For example, central nervous system-acting drugs may have a particularly high risk of causing ADRs in patients with underlying cognitive impairment.

A note on terminology: “Adverse drug reactions” refers to the unwanted effects of drugs at normal dosage and use. “Adverse drug events” refers to a broader range of potential harms associated with the drug, including overdose, withdrawal reactions from abrupt discontinuation of a drug, and more.

Causes of Adverse Drug Reactions

ADRs are commonly classified into 2 predominant types. Type A ADRs result from expected yet unwanted or exaggerated physiologic effects of the drug. For example, β blockers may cause bradycardia that results in syncope. Type B ADRs, which are less common, result from idiosyncratic effects unrelated to the drug’s usual physiologic targets; for example, anaphylaxis to penicillin.

In older adults, type A ADRs often arise from the interaction of a drug and underlying characteristics of the patient. Medications with a narrow therapeutic index and prolonged half-life cause the most trouble for the older patient. Patients with multiple medications, disease states, and/or subclinical physiologic changes associated with aging can be more susceptible to unwanted effects of a drug.

Drug–drug interactions can lead to ADRs by pharmacokinetic and pharmacodynamic mechanisms. In the former, Drug A inhibits the activity of a cytochrome P450 isoenzyme, resulting in delayed clearance of Drug B which is metabolized by that isoenzyme. This results in excessive tissue levels of Drug B, and resultant adverse effects. Common culprits that inhibit P450 activity include antimicrobials such as ciprofloxacin, fluconazole, and clarithromycin; some selective serotonin reuptake inhibitors; amiodarone; and verapamil and diltiazem. For example, diltiazem inhibits cytochrome P450 isoenzyme 3A4 (CYP3A4). Atorvastatin and several other (but not all) statin medications are metabolized by CYP3A4. Thus, if a patient is taking both diltiazem and atorvastatin, atorvastatin will accumulate because the enzyme that metabolizes it has been rendered less active. Tissue levels of atorvastatin will rise, potentially to the level where they cause substantial toxicity (ie, increase the risk of rhabdomyolysis and liver injury).

Cytochrome P450 isoenzymes can also be induced (“sped up”). This induction results in rapid clearance and thus decreased effectiveness of drugs metabolized by the affected isozyme. Medications that are potent inducers of P450 enzymes include rifampin, barbiturates, carbamazepine, and phenytoin. In general, cytochrome P450-mediated interactions are most important when induction or inhibition is potent (eg, greater than 5-fold change in enzyme activity) and the substrate drug has a narrow therapeutic index (eg, warfarin, sulfonylureas). In contrast, drug interactions that involve weak induction or inhibition and a substrate drug with a wide safety margin are less likely to be clinically relevant.

Use of 2 or more drugs with mutually reinforcing physiologic effects can also result in harms. For example, third-degree heart block may occur in a patient prescribed digoxin and a β blocker, as both suppress conduction of atrial impulses through the atrioventricular node.

Drug–disease interactions occur when an underlying disease state makes a patient more susceptible to the unwanted physiologic effects of a drug. Not every potential interaction results in harm. For example, many patients with mild or moderate chronic obstructive pulmonary disease can tolerate β blockers without adverse effects, although some will develop worse pulmonary symptoms in this setting.

In addition to ADRs, an expanded range of adverse events can occur from misuse of drugs. This can include complications of excessive doses of a drug, failure to prevent or treat disease because of nonadherence to or insufficient dosing of a drug, or withdrawal reactions caused by abrupt discontinuation of a drug to which the body has physiologically adapted (ie, chronic opioids).

Preventing Adverse Drug Reactions and the Role of Monitoring

Fewer than one-quarter of ADRs in ambulatory older adults are a result of clinicians making clearly inappropriate prescribing decisions. In contrast, most ADRs result from drugs that were reasonable to prescribe, and represent the known but unwelcome potential adverse reactions of a given drug. Warfarin and insulin are the most common causes of ADRs severe enough to precipitate an emergency room visit. Nonetheless, these drugs have an important place in the therapeutic armamentarium for older adults. In fact, warfarin is often underprescribed, as for many patients the benefits of preventing a stroke or pulmonary embolism outweigh the risk of hemorrhage.

True prevention is difficult in this setting, as it can be difficult to precisely predict which patients will be helped or harmed by a drug. Yet, many ADRs can be detected and managed early, sparing the patient prolonged symptoms or a cascade of ever-worsening adverse effects (such as untreated orthostasis resulting in a fall with fracture). Monitoring older adults for emerging ADRs thus plays a critical role in reducing the burden of ADRs, yet is often not done well. One important impediment to monitoring is that both patients and physicians may falsely attribute a new symptom to an underlying disease state or “getting old,” rather than recognizing it as an adverse drug reaction. This leads patients to underreport potential ADRs and to physicians not properly diagnosing a symptom as an ADR even when the patient reports it. The mantra “any symptom in an older adult is a medication side effect until proven otherwise” provides a useful reminder to always keep ADRs on the differential diagnosis when evaluating a new or worsened complaint.

Dedicated nurse-led and pharmacist-led programs are effective strategies for ADR monitoring, best exemplified by anticoagulation clinics. Although little data are available to support simple in-office or bedside tools for ADR monitoring, expert opinion suggests that several strategies may be helpful. These include: (a) at the time a drug is prescribed, warning the patient of which adverse reactions to watch for; (b) at the next patient encounter, use a combination of open-ended questions and specific prompts to query for adverse reactions (eg, “Are you having any side effects or problems from Drug X?” followed by specific questions about dangerous and common adverse reactions); and (c) using a similar strategy to query for adverse reactions during annual medication review.

Multiple Medication Use

Epidemiology and Potential Harms and Benefits of Using Multiple Medicines

Nearly 20% of adults age 65 years and older use 10 or more medications. This use of multiple medications is commonly referred to as “polypharmacy.” This term carries a pejorative connotation, in part for good reason. Use of multiple medications substantially increases the risk of drug–drug interactions and of adverse drug events, can impose substantial cost burdens on patients, complicates adherence, and is associated with increased risk of using inappropriate medications. On the other hand, older patients often have multiple chronic conditions that can be substantially helped by medications. In many such patients, the use of multiple drugs is an appropriate therapeutic choice. Thus, while use of multiple medications is a risk factor for medication problems—and should prompt close attention to reducing unnecessary medications—the focus on reducing medications needs to be balanced with the harms to longevity and quality of life that arise from undertreated chronic conditions (see Chapter 53, “Addressing Polypharmacy & Improving Medication Adherence in Older Adults”).

The Prescribing Cascade

One important contributor to multiple medication use is “the prescribing cascade,” in which the adverse effects of one drug are treated with another drug, which itself causes adverse effects that are treated with a third drug, and so forth. This can result from misinterpreting a sign or symptom as the manifestation of an underlying disease process, rather than as an adverse drug effect. As noted above, remembering the mantra “any symptom in an older adult is a medication side effect until proven otherwise” can help guard against potential prescribing cascades. Except in unusual circumstances, it is typically better to withdraw or substitute the offending drug rather than treating its adverse effects with another drug.

Overuse, Misuse, and Underuse of Medications

For many older adults, the question is not whether a patient is taking too many or too few medications, but whether the patient is taking the right medications given the patient’s diseases, preferences, and ability to adhere. Deviations from an optimal regimen can be viewed as problems of overuse (use of a drug where no medication therapy is needed), misuse (use of a drug where a better alternative is available), and underuse (nonuse of a drug that would be beneficial).

Overuse and Misuse

Overuse and misuse of drugs are common. Approximately 20% to 30% of older ambulatory adults use at least 1 drug that consensus criteria recommend avoiding in older patients. Expert review of medication regimens in outpatient, inpatient, and nursing home settings has also identified large proportions of patients taking drugs that are not indicated, ineffective for the condition being treated, or otherwise problematic. It is common for drugs to be continued long after they are no longer necessary. For example, roughly half of patients who are started on a proton pump inhibitor for stress ulcer prophylaxis during a hospital stay are continued on these medications after discharge, for no discernible reason.

Several explicit criteria, commonly called “drugs-to-avoid lists,” have been developed to identify medications and therapeutic situations that are potentially inappropriate for older adults. These tools have proved useful for quality improvement, including flagging instances of such prescribing for special scrutiny and review. Nonetheless, clinical judgment needs to be applied for individual patients, as there are situations in which use of many of these drugs is reasonable.

The Beers criteria of potentially inappropriate medications are shown in Table 9–1. The most frequently cited part of these criteria concern drugs that are potentially inappropriate in any setting. Commonly used medications on this “drugs-to-avoid” list include first-generation antihistamines (eg, diphenhydramine and hydroxyzine), tertiary-amine tricyclic antidepressants, use of benzodiazepines for insomnia, agitation, or delirium, long-acting sulfonylureas (eg, glyburide), and sliding-scale insulin.

The STOPP (Screening Tool of Older Person’s Prescriptions) criteria define an extensive list of specific clinical situations in which drug use is potentially inappropriate (Table 9–2). Examples include use of loop diuretics for ankle edema in the absence of heart failure; selective serotonin reuptake inhibitors in patients with a history of hyponatremia; and use of NSAIDs in patients with heart failure or moderate-to-severe hypertension.

The ACOVE (Assessing Care of Vulnerable Elders) criteria for vulnerable older adults cover a wide range of topics, including a number of criteria related to medication use. ACOVE criteria address not only potentially inappropriate medications but also recommended care practices, including patient education about medications and regular medication review.

Underuse

Although use of inappropriate drugs is commonly discussed in older adults, equally important is the undertreatment of conditions that could be helped by drug therapy. Older adults are less likely to receive indicated medications than their younger counterparts, even after accounting for contraindications to these therapies. Excessive fear of causing adverse events, distraction by other clinical issues, a sense of futility, and subtle ageism likely contribute to this pattern. In addition, treatable conditions are often underdiagnosed in older patients, and symptoms such as pain, fatigue, depressed mood, or orthostasis may be incorrectly attributed to “getting old.”

The START (Screening Tool to Alert doctors to Right Treatment) criteria are consensus criteria that identify potential underuse of beneficial medications in older adults (Table 9–3). As with other explicit criteria, these are intended as a guide but not as a substitute for clinical judgment for individual patients. Examples of drugs which the START criteria recommend should routinely be used include warfarin for chronic atrial fibrillation (in the absence of contraindications); regular inhaled β-agonist or anticholinergic therapy for patients with mild to moderate asthma or chronic obstructive pulmonary disease (COPD); and bisphosphonates in patients on chronic corticosteroid therapy.

The ACOVE criteria also identify instances of potential underuse for a variety of chronic illnesses. This includes not only omissions of specific medication but other recommended care processes, including annual medication review, educating patients about their medications, and monitoring for medication effectiveness and toxicity.

High-Risk Medications

The following medications are often associated with adverse reactions and merit special caution in prescribing.

Warfarin and Other Anticoagulants

The benefits of warfarin for stroke prevention in atrial fibrillation and for the treatment of venous thromboembolism (VTE) outweigh the risk of hemorrhage for most patients, even for patients older than age 80 years and patients with a history of falls. Yet, warfarin is the most common medication implicated in emergency room visits and hospitalizations for ADRs. Safe use requires close monitoring to keep patients in the target anticoagulation range and attention to the increased risk of hemorrhage when used with antiplatelet agents. Close to 700 known medications, supplements, and foods interact with warfarin, either by inhibiting cytochrome P450 enzyme activity, displacing plasma protein binding, affecting vitamin K metabolism, or potentiating bleeding risk through other antithrombotic mechanisms. Antibiotics, antiplatelet agents, and amiodarone are commonly implicated as the source of drug–drug interactions that result in bleeding. Patients receiving warfarin must be monitored closely to keep the international normalized ratio (INR) in target range, as well as for drug–drug interactions whenever a new medication is added.

Despite the risks of warfarin, newer anticoagulant medications have not yet displaced its use. Dabigatran is an oral direct thrombin inhibitor that is approved for stroke prevention in nonvalvular atrial fibrillation and is used outside the United States for VTE prophylaxis after hip or knee replacement surgery. Dabigatran overdose cannot be reversed, and thus, the risk of severe, or even fatal, hemorrhage may make warfarin a more desirable choice. Reduced dosing is recommended in renal impairment. In Canada, doses are reduced for people age 80 years and older, although at this writing the labeling in the United States does not recommend any alteration in dose based on age.

Insulin

Older age is associated with increased risk of drug-induced hypoglycemia. Insulin is the second-most common cause of ADRs that lead to emergency room visits in older adults. Although insulin has a useful place in the treatment of diabetes, caution in prescribing is merited. Special attention should be paid to factors that may increase the risk and consequences of severe hypoglycemia. These risk factors include diminished renal function, use of medications that may interact with insulin’s effects, and impaired cognitive function (which may interfere both with proper use and with the patient’s ability to obtain help if hypoglycemia begins to occur). Long-acting basal insulins (eg, insulin glargine and insulin detemir) are less likely to cause hypoglycemia than neutral protamine Hagedorn (NPH) insulin. Sliding-scale insulin should be avoided, as it increases the risk of hypoglycemia without yielding improved glycemic control. (See also Chapter 42, “Diabetes.”)

Long-acting Sulfonylureas

All sulfonylureas have the potential to cause hypoglycemia. In older adults, the risk of adverse events is particularly high with the long-acting sulfonylureas, including glyburide (also known as glibenclamide) and chlorpropamide. This excess risk is partly a consequence of accumulation of these drugs in patients with diminished drug clearance. If a sulfonylurea is used, a shorter-acting version, such as glipizide, is preferred.

Digoxin

Digoxin toxicity is common, often manifesting as neurologic abnormalities (including fatigue, confusion, or changes in color perception) and/or GI disturbances. Toxic effects including arrhythmias are accentuated in the presence of hypokalemia, which commonly occurs in patients who are also receiving loop diuretics. Impaired renal function and drug–drug interactions often result in elevated serum digoxin levels in older adults, although toxicity can occur even at serum digoxin levels within the normal range. Other agents are typically preferred for management of heart failure and atrial fibrillation with rapid ventricular response, although digoxin may be appropriate in select patients. If used, digoxin should be prescribed at doses ≤0.125 mg/day and patients carefully monitored for serum digoxin levels (aiming for the low-normal range), electrolytes (particularly hypokalemia), and for clinical signs of toxicity. New or worsening neurologic, GI, or cardiac signs or symptoms in a patient taking digoxin should be considered an adverse drug reaction until proven otherwise.

NSAIDs

NSAID-induced peptic ulcer disease and renal impairment occur more commonly in older adults than in younger adults. In addition, these drugs exacerbate hypertension, promote fluid retention in patients with heart failure, and antagonize the cardioprotective effects of aspirin through competitive inhibition of the cyclooxygenase (COX)-1 enzyme. The Beers criteria and pain guidelines from the American Geriatrics Society discourage regular, chronic use of systemic NSAIDs in older adults, preferring acetaminophen, and in many cases opioids, for pain control. NSAIDs are contraindicated in patients with heart failure, renal dysfunction, and those at high risk of peptic ulcer-induced GI bleeding. Risk of this latter complication increases substantially in patients also taking warfarin, selective serotonin reuptake inhibitors (SSRIs), or systemic corticosteroids. If NSAIDs are used for longer than brief episodic use, the following considerations are advised: (a) use at the lowest dose and for the shortest duration possible; (b) coadminister proton pump inhibitors or misoprostol for gastroprotection; (c) maximize the time between taking cardioprotective aspirin and taking an NSAID (ie, take aspirin upon awakening, delay taking NSAIDs until at least 2 hours later); and (d) consider followup in 2–4 weeks after starting an NSAID to evaluate for renal dysfunction, fluid retention, and blood pressure elevation. NSAIDs with a balanced inhibition of COX-1 and COX-2 are preferred in patients who are at risk of cardiovascular disease. Consistent with this, some data suggest that naproxen has among the most favorable cardiovascular risk profiles. Topical NSAIDs, such as topical diclofenac gel, have relatively minimal systemic absorption and thus substantially lower risk of causing systemic toxicity.

Anticholinergics

Drugs that block the action of acetylcholine include sedatives, antihistamines, antidepressants, antipsychotics, bladder and GI antispasmodics, muscle relaxants, and antiemetics (see examples in Table 9–4). The cumulative burden of multiple anticholinergic drugs has been associated with an increased risk of falls, functional decline, and impaired cognition in older persons. If a medication with anticholinergic properties is considered necessary, substitution with a less anticholinergic medication in the same therapeutic category should be attempted when possible.

Opioids

Opioids are useful to treat moderate to severe pain in older persons, but opioids may be underused because of difficulty in diagnosing and assessing pain and concerns about safety and correct use of opioids. Safety concerns include an increased risk of delirium, GI side effects, and ventilatory depression. However, untreated pain can result in delirium, depression, decreased mobility, and impaired sleep. Changes in pharmacokinetics, as well as the increased pharmacodynamic effects of opioids, increase the risk of ADRs. Nonetheless, opioids can often be used safely when keeping these age-related changes in mind. In general, the dosing interval should be the same for an older person of any age, but it is typically appropriate to start with a low dose and titrate up the dose slowly, often termed ‘start low and go slow.’ Drug–drug interactions need to be kept in mind, as many opioids are substrates for P450 enzymes. Finally, tailoring treatment for renal or liver dysfunction may be necessary. For those with substantial renal impairment, morphine should be avoided, and hydromorphone, fentanyl, and methadone may be preferred alternatives. Methadone and codeine should not be used in severe liver impairment, and, in general, opioid dose should be reduced even further, with a longer dosing interval.

Antipsychotics in Dementia

The use of antipsychotics to treat behavioral and psychological symptoms of dementia is associated with increased likelihood of myocardial infarction, stroke, falls, fractures, VTE, and mortality. As a result, FDA warnings, practice guidelines, and initiatives from the Centers for Medicare and Medicaid Services have decreased their use. Older antipsychotics also have significant anticholinergic and extrapyramidal side effects. When possible, behavioral and psychological symptoms of dementia should be treated by nonpharmacologic means. When antipsychotics are deemed necessary for symptoms causing severe distress or harm, benefits and risks should be discussed with a patient’s family or caregiver, the discussion should be clearly documented, and the antipsychotics should be used for a minimum duration of therapy with attempts to taper and discontinue the medication when possible.

Explicit drugs-to-avoid criteria, attention to specific high-risk medications, and an understanding of technical elements of prescribing are important. Yet, these discrete skills address only a small proportion of potentially inappropriate prescribing in older adults. In most situations, close attention to several principles can be the most useful guide to optimizing prescribing decisions for older adults.

Goals of Care

In younger and physically robust adults, there are typically standardized guidelines about the use of common drugs. These guidelines (both formal and informal) are based not only on the risks and benefits of drug therapy for an average patient, but also on the expectation that most people share similar values about what benefits and potential harms are most important to them. In contrast, older patients may have a different profile of benefits and risks than the “average” adult. Moreover, older adults hold widely varying views about what benefits they want their drugs to achieve and what harms are most important for them to avoid. For example, some older adults place great value on extending longevity and preventing future disease, whereas others are more interested in minimizing symptoms (both from their diseases and from drug side effects) and place a lower priority on life extension. Careful elicitation of a patient’s goals of care—and keeping these goals of care in mind when prescribing—can help target therapy to achieve the goals most important to the patient and minimize the unwanted consequences most concerning to them.

Time to Benefit

Medications used to prevent future health events (such as fracture, myocardial infarction, or renal failure) typically have a delayed time to benefit, with a meaningful reduction in risk not achieved until 1-2 or many years after the patient starts taking the drug. In contrast, adverse drug effects typically begin soon after a drug is started. Patients with limited life expectancy may thus spend the final period of their lives exposed to the harms of a drug, without living long enough to reap the benefits. (See tools for estimating life expectancy in Chapter 3, “Goals of Care & Consideration of Prognosis.”) There are limited data on time to benefit for specific drugs in older adults, because most clinical trials of medications exclude older persons and those with multiple comorbid conditions. General estimates of time to benefit include:

• Glycemic control for patients with diabetes—at least 3 years for macrovascular complications (eg, myocardial infarction stroke) and 7 years for microvascular complications (eg, nephropathy, neuropathy)

• Bisphosphonates for osteoporosis—1.5 years to prevent fracture

• HMG-CoA reductase inhibitors (“statins”) for patients with chronic cardiovascular disease—1 to 2 years to prevent cardiovascular events, and more than 3 years to prevent strokes

Attention to Dose

Because older adults are more susceptible to adverse drug effects, it is often helpful to “start low and go slow,” meaning to use a low starting dose and to advance the dose slowly. For many drugs, it is useful to start at half the regular adult starting dose. This can often be accomplished by splitting tablets using an inexpensive pill splitter. Of note, some patients with limited manual dexterity have difficulty splitting pills, and pills with sustained-release delivery mechanisms should not be split.

Careful attention should be placed on renal dysfunction and other characteristics that may result in increased serum levels, and dose escalation should stop at the lowest effective dose. Nevertheless, some older adults require the full dose of a drug, and many older patients are undertreated as a result of clinician reluctance to escalate the dose. This phenomenon is well-documented in the treatment of depression in older adults. Thus, continued dose escalation to the maximum dose is usually advisable if lower doses do not yield the desired effect and the patient is tolerating the drug.

Monitoring

At the time a clinician prescribes a medication, she makes an educated guess that the probability of benefit exceeds the probability of harm, without definitively knowing which beneficial and harmful outcome(s) will occur. Monitoring for benefit and harm as determined by symptoms, signs, and laboratory tests can help determine to what extent a drug is actually helping or harming a patient, and thus plays a key role in individualizing care. Unfortunately, monitoring is often not consistently performed.

Few guidelines are available to guide the frequency of monitoring, either for laboratory values or for signs and symptoms. In the absence of specific evidence-based recommendations, a general approach can be helpful. Patients often underreport ADRs, and clinicians misinterpret these symptoms as markers of an underlying disease. Thus, once the decision is made to prescribe a drug patients should be educated and activated to understand and report medication-related problems (Figure 9–1). Then, at regular intervals the drug should be monitored for potential adverse effects and effectiveness, for patient adherence, and to assess whether the drug is still needed. Following subsequent modifications to the regimen (or not), the cycle repeats again. Although ongoing monitoring is important, in many cases adverse effects and effectiveness become apparent within the first several weeks of use, so particular attention should be paid to monitoring during this time. It is almost always useful to inquire about drug effectiveness, harms, and adherence in the first follow-up visit after a drug is started or changed.

Adherence

Medications are not useful if patients don’t take them. See “Adherence” below.

Discontinuing Medications

Approximately 25% of patients suffer an adverse reaction upon withdrawing a medication, either because of return of the underlying disease or a physiologic withdrawal reaction (such as one might encounter with abrupt discontinuation of an opioid analgesic). To distinguish between drugs that can be stopped abruptly and those that need to be slowly tapered off, a rule of thumb is that if a drug generally requires a titration up from lower to higher doses, it should be titrated down instead of abruptly stopped. Examples of drugs that should not be discontinued abruptly are opioid analgesics, antidepressants, β blockers, and anticonvulsants such as gabapentin. In most cases, the rate of down-titration should match the rate at which one can safely up-titrate the drug. In contrast, drugs that can safely be started at maximum dose can typically be stopped abruptly without risking a physiologic withdrawal reaction. Examples include proton pump inhibitors and NSAIDs. In all cases, patients discontinuing a drug should be monitored for the return of signs or symptoms the drug was used to treat.

Applying Clinical Trial Evidence and Evidence-Based Guidelines

Clinical trials that have established the efficacy of commonly used therapies have largely been conducted in relatively young and otherwise healthy populations. As a result, many have questioned the applicability of these findings to the care of clinically complex older adults. In addition, many clinical practice guidelines provide limited guidance about how their recommendations apply to adults who are frail or in the upper reaches of life span.

Despite this uncertainty, many of these therapies are likely beneficial for the majority of older adults. In some cases older adults may achieve greater benefits from drugs than the younger populations in which trials were conducted. Because older patients typically have a greater risk of developing the outcomes the drug is used to prevent, the relative reduction in risk associated with the drug translates to a greater absolute reduction in risk in older adults. Thus, while older adults can be more likely to suffer harms from drug therapy, in many cases they can have the most to gain. (See also Chapter 74 on extrapolating the evidence from clinical research to older patients).

Prescribing at the End of Life

Prescribing for patients with limited life expectancy typically requires a rebalancing of the benefits and harms of drug therapy. For many patients near the end of life, prevention of long-term outcomes through drugs used to prevent myocardial infarction, fracture, and other such outcomes is less relevant because of delayed time to benefit. Goals of care may shift to prioritizing quality of life and the minimization of medical interventions, and drugs to alleviate symptoms may continue to be highly valued. Difficulty swallowing can also complicate the delivery of oral therapies.

Several consensus recommendations have been published recommending drugs to be avoided in patients with limited life expectancy (typically less than 6 months to 1 year) and/or advanced dementia. Although there is no universal consensus, several groups recommend against routinely using bisphosphonates, cholesterol-lowering therapy, and warfarin. In contrast, bothersome symptoms are often underrecognized and undertreated in the patient in the terminal phase of life, and careful attention should be paid to pain control, constipation, and other symptoms. (See also Chapter 11, “Geriatrics & Palliative Care.”)

Medication adherence refers to a patient actively participating in a therapeutic plan that has been agreed upon by a health care provider and the patient. Persistence refers to a patient taking a therapy for a necessary duration of time. Nonadherence and nonpersistence are prevalent problems, especially in the treatment of chronic, asymptomatic conditions like hypertension or hyperlipidemia. Nonadherence is associated with failure to achieve disease control, misdiagnoses, increased emergency room visits and hospitalizations, increased health care costs, and in some cases, higher mortality. Up to 40% of patients in the United States do not take their medications correctly, and failure to appreciate this could result in overprescription of additional medication. It is thus important to assess adherence regularly during clinical visits. Assessing adherence in a nonjudgmental fashion, such as asking how many doses are missed in a week, is a quick way to gauge adherence. Validated tools could also be used for regular monitoring, including the Morisky Medication Adherence Scale.

Major risk factors for nonadherence in chronic disease therapy include patients’ beliefs that medications are not necessary or are harmful, side effects, cost and copay, and well as increased number of medications. Understanding the reason for nonadherence is essential for designing strategies to improve adherence. For example, if a patient is nonadherent because of high out-of-pocket costs for obtaining the drug, less-expensive drug alternatives or pharmaceutical assistance programs may be useful to consider. If a patient is nonadherent because the patient does not know what the drug is for, or believes that the drug is not providing benefit, educating the patient can be helpful. (Alternatively, if a patient taking a drug used to control symptoms believes the drug is not helping, this may be a sign to try another drug.) Drugs that require dosing 3 or 4 times per day can be difficult to take, and switching patients to a drug regimen that requires only once- or twice-daily dosing has proven to be one of the most successful strategies for improving adherence.

Optimizing prescribing for older adults is a complex process that requires balancing multiple considerations. While seemingly daunting, several strategies can be helpful to unpack this complexity and approach these issues in a systematic manner.

Regular Medication Review

Experts recommend that regular medication review occur at least annually. Patients with multiple medication changes may benefit from more frequent review.

Brown Bag Review

A highly effective technique for medication review is the “brown bag review,” in which the patient is instructed to put all of their medications (including over-the-counter, herbal, and other products) into a bag and bring it to clinic for inspection. In addition to reconciling the medications taken by a patient, this review provides a valuable opportunity to assess patient understanding and adherence. Multiple bottles of the same medicine, prescriptions whose label indicates that they were dispensed many months (or years) ago, and other visual clues can help identify potential problems. For each drug, asking “Do you know what this is for?”; “How do you take this?”; and “Do you sometimes not take it or miss a dose?” (and if yes, inquiring why) can provide valuable clues to improve adherence. Finally, asking the patient “Are you having any problems with this drug” and probing for common and dangerous adverse effects can help elicit previously underreported adverse drug effects.

Critically Reviewing the Medication List

Another important goal of medication review is to provide the clinician an opportunity to think critically and holistically about the medication regimen, rather than the piecemeal approach that often accompanies a typical visit in which the focus is on 1 or 2 specific diseases. This holistic review can include identifying drugs that are no longer needed (ie, overuse), drugs with inadequate or excessive doses or instances where an alternative drug is likely to be safer or more effective (ie, misuse), and omissions of potentially beneficial drugs (ie, underuse). The Medication Appropriateness Index provides a useful list of 10 questions to consider for each drug a patient is taking (Table 9–5).

A helpful strategy for reviewing the medication list is to group medicines by the patient’s diseases or syndromes they are used to treat. This way of organizing medication information can highlight potential problems. If the patient is taking a drug with no corresponding disease on the patient’s problem list, it may be unnecessary. For example, if a patient is taking a proton pump inhibitor and does not have a diagnosis of gastroesophageal reflux disease (GERD) or is not a chronic user of NSAIDs, the drug may be unnecessary. If a patient has a disease with no corresponding medication, this may represent underuse. For example, if an older male has bothersome lower urinary tract symptoms that persist despite lifestyle modifications, a trial of an α-blocker may be warranted. If a patient has poorly controlled disease for which the patient is taking multiple medications, this might indicate suboptimal doses, poor adherence, or a complicating factor. For instance, if an older patient has poorly controlled hypertension and is taking 4 antihypertensive medications, further investigation for nonadherence and potential secondary causes of hypertension may be useful. Finally, careful review of medications can identify other potential problems. For example, if a patient is taking a medication with strong anticholinergic properties, it is worth inquiring about anticholinergic side effects and considering if there is another medicine that can provide the same benefit with less potential for harm.

Interdisciplinary Care

Careful attention to pharmaceutical care for older adults is time-consuming, and is best done using a team approach where feasible. Pharmacists can play an essential role, bringing both content expertise and dedicated time to assessing potential medication problems, reconciling medications, and evaluating and improving patient adherence. The emergence of patient-centered medical home models of care is increasingly integrating pharmacists into primary care clinics, providing opportunities to share workload and expertise. Several other opportunities are also available. Under Medicare Part D (the prescription drug benefit), health plans are required to offer medication therapy management services to high-risk older adults, including comprehensive medication review conducted in-person or by phone at least annually. The scope and impact of these programs is evolving and has yet to be clearly determined. Community pharmacists can work with prescribers and patients, and consultant pharmacists (expert consultants in geriatric pharmacotherapy) can be engaged to provide comprehensive medication review and reconciliation, monitor medications, improve adherence, identify prescription assistance programs to help lower drug costs, and more. In the inpatient setting, hospital pharmacists rounding with teams reduces ADRs, and pharmacists have been widely employed to assist with medication reconciliation at admission and discharge. Nurses can also be valuable partners, helping to reconcile medications, screen for adverse events, and monitor drug effectiveness (eg, through structured protocols for hypertension management, keeping track of overdue laboratory tests for drug monitoring, and so forth).

Although general principles of prescribing apply to different care settings, certain considerations merit special mention in the hospital and nursing home.

Prescribing in the Hospital

Hospitalized older adults are particularly prone to medication misadventures for a number of reasons. Errors in medication reconciliation and communication between providers commonly occur at various stages of transition, including admission, transfer between hospital units, and discharge. Ensuring that patients continue to receive the right medications as they move between providers and locations is critical. Hospital-based pharmacists have been shown to be helpful in improving outcomes related to medication reconciliation during these transitions and should be involved when possible. Medications started for transient reasons during hospital stays are often mistakenly continued at discharge, and may become a permanent fixture on the patient’s medication regimen. Careful attention should be paid to discontinuing proton pump inhibitors that may have been started for stress ulcer prophylaxis (itself a questionable indication), analgesic medications for pain that is no longer present, and so forth.

Clinicians may be tempted to change medication regimens for chronic diseases during hospital stays. In many cases, this temptation should be resisted. Measures of chronic disease control measured during a hospital stay (ie elevated blood pressure and worse lipid profiles) may not be representative of the patient’s usual status, and contextual factors, such as contraindications for a drug, may be known by the primary care provider but not by the inpatient team. Changing multiple medications also increases the risk of adverse drug events. Yet, if important quality gaps are identified, hospital stays can provide an opportune time to rectify these problems if done in consultation with the patient’s primary care clinician.

Sedative/hypnotic drugs or anticholinergic agents such as diphenhydramine are often prescribed “as needed” to hospitalized older adults to aid sleep. These drugs increase the risk of falls and delirium and should be avoided if possible. Instead, environmental interventions such as limiting nighttime vital signs and reducing noise and light stimulation are preferable.

Prescribing in Nursing Homes

Patients in long-term care often have multiple medication conditions, physical frailty and/or cognitive impairment, and use an average of 7 to 8 medications. Prescribing for such patients is complex, and they are at high risk of adverse drug events.

Antipsychotic medications are prescribed to roughly one-quarter to one-third of nursing home residents in the United States, often to manage behavioral problems of dementia. Unfortunately, these drugs confer a substantially increased risk of death in older adults with dementia. Use should be avoided for behavioral management unless nondrug interventions have failed and the patient is a threat to himself or herself or others. Because psychotropic medications have long been prescribed at inappropriately high rates to nursing home patients, federal regulations require that each patient receiving these medicines has ongoing documentation of the reason for treatment, how medication effectiveness and adverse effects will be monitored, and plans for dose reduction or treatment continuation.

Federal regulations also require a pharmacist to conduct monthly medication review of patients in long-term care facilities. Unfortunately, such reviews have at times been found to be deficient, and should not be relied on to catch prescribing problems.

ACOVE-3 Criteria. Introduction to Quality Indicators (explicit criteria to identify potentially inappropriate medication use and to identify potential underuse of medications). http://www.rand.org/health/projects/acove/acove3.html

Age and Ageing. START criteria (in START [screening tool to alert doctors to the right treatment]—an evidence-based screening tool to detect prescribing omissions in elderly patients) (complete article; explicit criteria to identify potential underuse of medications) http://ageing.oxfordjournals.org/content/36/6/632.long

American Geriatrics Society. AGS Beers Criteria 2012 (explicit criteria to identify potentially inappropriate medication use). http://www.americangeriatrics.org/health_care_professionals/clinical_practice/clinical_guidelines_recommendations/2012/

American Pharmacists Association. What is Medication Therapy Management? http://www.pharmacist.com/MTM

American Society of Consultant Pharmacists. Medication Management. http://www.ascp.com

Anticholinergic drug scale. http://www.ncbi.nlm.nih.gov/pubmed/18332297

DailyMed (information from package inserts). http://dailymed.nlm.nih.gov

GlobalRPh. Calculators including renal function online calculator. http://www.globalrph.com/multiple_crcl.htm

Indiana University Department of Medicine. Cytochrome P450 Drug Interaction Table (drug–drug interactions). http://medicine.iupui.edu/clinpharm/DDIs/

Indianapolis Discovery Network for Dementia. Anticholinergic Cognitive Burden Scale. http://www.indydiscoverynetwork.org/AnticholinergicCognitiveBurdenScale.html

Medline Plus. Drugs, Supplements, and Herbal Information (drug information for patients). http://www.nlm.nih.gov/medlineplus/druginformation.html

Morisky Medication Adherence Scale. http://www.acpinternist.org/archives/2009/02/adherence.pdf

National Council on Patient Information and Education (NCPIE). Medication Use Safety Training (MUST) for Seniors. http://www.mustforseniors.org

Proprietary drug–drug interaction programs, including www.epocrates.com (free), and www.lexicomp.com and www.micromedex.com (subscription only)

STOPP Criteria (explicit criteria to identify potentially inappropriate medication use). http://www.ncbi.nlm.nih.gov/pubmed/18218287