Older adults frequently experience adverse drug events in the intensive care unit (ICU), due to age-related pharmacokinetic and pharmacodynamic changes, polypharmacy, and frequent transitions of care.
Adverse drug events can be reduced in the geriatric population by individualizing dosing, avoiding potentially inappropriate medications when possible, recognizing prescribing cascades, reconciling medications at ICU discharge, and incorporating a critical care pharmacist on the ICU team.
The effects of age on drug absorption, volume of distribution, and drug metabolism and clearance must be considered when optimizing medication therapy in the ICU setting.
Important dosing considerations must be employed when nonintravenous routes of administration are used in critically ill older adults.
Extremes in body weight are common in older adults and should be considered when optimizing pharmacotherapy in this population.
Kidney function and the use of strategies such as renal replacement therapy (RRT) to support this function must be carefully evaluated and considered when optimizing drug therapy in older ICU adults.
Older adults usually require far lower doses of opioids and sedatives and experience more adverse events with their use.
Nonpharmacologic interventions such as early mobilization and the avoidance of delirium-causing medications (e.g., benzodiazepines) are more effective strategies to reduce the burden of delirium in older adults than initiating antipsychotic therapy.
A number of pharmacodynamic and pharmacokinetic factors should be considered when optimizing anti-infective regimens in critically ill older adults.
Body weight, kidney function, and the availability of a reversal agent should be considered when implementing anticoagulant therapy in the geriatric ICU population.
Older adults (≥65 years) comprise an ever-increasing proportion of patients admitted to the ICU . Very old adults (≥80 years) now make up more than 25 percent of the patients admitted to most ICUs. An ever-increasing proportion of older adults is chronically critically ill and frequently move between the long-term acute care hospital (LTACHs) and ICU settings . Drug therapy plays a key role in improving outcomes after critical illness. With the average older adult admitted to an ICU administered 30 different medications throughout their stay, critical care clinicians are faced with making multiple drug-related decisions on a daily basis . A number of different age-related effects increase the risk for medication-associated adverse events and drug interactions. Polypharmacy is a common consequence of critical illness among older adults and is associated with deleterious outcomes and increased costs . This chapter reviews key concepts regarding medication choice, dosing, and monitoring in critically ill older adults and provides ICU clinicians with a number of strategies to improve the medication-related outcomes of geriatric patients under their care.
Many older adults have multiple chronic conditions that require multiple medications and, on average, are prescribed 12 different prescription medications . Adverse drug events account for more than 30 percent of all hospital admissions among older adults, and a drug-related cause is a frequent reason for ICU admission [6,7]. Use of herbal and over-the-counter (OTC) medications has doubled among older adults in the past 10 years; 10 percent will experience a major herbal/OTC–prescription drug interaction each year . Prescribing cascades (i.e., a new medication is initiated to treat the side effects of another) remain an important driver of polypharmacy in the ICU setting.
Safe and effective medication use is crucial to ensure optimal patient care and ICU outcome. Critical illness–associated acute organ dysfunction will affect drug absorption, clearance, and response and increase adverse events . Data generated in non-ICU settings should not be extrapolated to elderly ICU patients, in whom acuity of illness is higher, goals of therapy are different and change frequently with patient pathophysiologic condition, the duration of medication use is generally short, and monitoring practices are more rigorous . There is a dearth of rigorous data to guide prescribing decisions for critically ill patients, which results in “off label” use for nearly half the medications ultimately prescribed .
Several drugs often used to treat critical illness are listed in the Beers List of low-benefit and/or high-risk medications – a list of medication deemed by geriatricians as “preferably avoided” in older adults in chronic healthcare settings . Potentially inappropriate medications (PIMs) are frequently prescribed in the ICU setting ; the number of PIMs prescribed directly correlates with the duration of ICU stay . Factors shown to increase the risk for a PIM being continued at ICU discharge include the number of PIMs prescribed during the ICU admission, admission to a surgical (versus medical) service, and discharge to another facility rather than home .
Transitions of care are frequent among older adults given the high proportion of this cohort that reside in long-term care facilities. They undergo frequent periods of decompensation, or suffer a new acute illness that requires management in the ICU. The post–intensive care syndrome (PICS), a new or worsening decrement in mental, cognitive, or physical health following critical illness persisting beyond the acute hospitalization, occurs frequently in older adults who survive their ICU admission . Medications that potentiate glucose dysregulation and delirium or worsen ICU-acquired weakness may worsen PICS .
Pharmacokinetic and Pharmacodynamic Alterations in Older Adults
Pharmacokinetics refers to the discipline that describes the absorption, distribution, metabolism, and elimination of drugs (i.e., what the body does to the drug). Pharmacodynamics is the study of the relationship between the concentration of a drug and the response obtained in a patient (i.e., what the drug does to the body). Several significant alterations develop in pharmacokinetic and pharmacodynamic parameters with aging and in the geriatric critically ill patient that must be considered in order to maximize outcomes and minimize drug toxicity and adverse effects.
While splanchnic blood flow, gastric emptying time, and small intestine absorptive capacity are decreased in the elderly, drug absorption, for the most part, remains relatively unchanged in this population . Nevertheless, a few key concepts should be considered. Delayed gastric emptying may be clinically significant for medications for which a rapid onset is desired (e.g., opioids). Because of reductions in liver mass and blood flow, the first-pass metabolism is lower in the elderly, and the bioavailability of medications that undergo extensive first-pass metabolism (e.g., labetalol, propranolol) may be higher.
Volume of distribution is a mathematical concept that refers to the compartments into which a drug disperses. Significant age-related changes occur to body composition that have an impact on volume of distribution and hence drug concentration. Age-related reductions in total body water and increases in total body fat may lead to increased concentrations (due to a smaller volume of distribution) for hydrophilic drugs such as aminoglycosides and digoxin . In contrast, lipophilic drugs (e.g., diazepam) will have a larger volume of distribution, and loading doses may need to be modified. However, clinicians should be cautious in increasing doses in elderly patients, given that age-related reductions in clearance may result in drug accumulation and prolonged effects.
Age-related changes in protein binding are of minimal clinical significance in most patients, but in the setting of critical illness, they can be profound. Acidic drugs bind to albumin, which is often decreased in patients with burns, liver disease, sepsis, uremia, and trauma. Low albumin levels lead to reductions in drug-protein binding, which cause an increase in the unbound fraction of pharmacologically active drugs such as phenytoin and warfarin. Basic drugs such as morphine bind to alpha-1-acid glycoprotein, which is an acute-phase reactant and is often increased in patients with renal failure, burns, infections, and myocardial infarction and in those who have undergone recent surgery. Protein binding in these settings is elevated, and unbound drug concentrations are decreased.
Drug metabolism depends on both the amount of blood flow to the liver and the liver’s ability to extract the drug from the bloodstream. Both of these processes are affected by age. Hepatic blood flow decreases by 30 percent between the ages of 30 and 75; thus the metabolism of drugs that rely on flow-dependent hepatic clearance (i.e., high-extraction drugs) may be negatively affected (e.g., labetalol, morphine, verapamil) . Drugs that depend on enzymatic function for clearance (as opposed to hepatic blood flow) are considered low-extraction drugs (e.g., haloperidol, diazepam, phenytoin). These agents undergo either phase I reactions (i.e., oxidation, reduction, hydrolysis) or phase II reactions (i.e., glucuronidation, acetylation, sulfation). Phase I reactions are much more sensitive to age, and the clearance of drugs that are metabolized via these mechanisms may be reduced (e.g., diazepam, midazolam). Phase II reactions, in contrast, are not impaired in older adults, and the clearance of these agents is not reduced in an age-dependent fashion.
A common clearance pathway for many drugs used in the ICU setting is the cytochrome P-450 (CYP-450) system. The efficiency of the CYP-450 system may be affected by both advanced age and critical illness. However, not all CYP isoforms are equally affected by aging. Although wide variability has been noted, clearance appears to be lower for substrates of CYP-1A2 and CYP-2C19, decreased or unchanged for substrates of CYP-3A4 and CYP-2C9, and unchanged for substrates of CYP-2D6 .
Increasing age is associated with several structural and functional changes in the kidney that result in decreased glomerular function and altered renal elimination. After age 30, there is a linear decline in glomerular filtration by approximately 8 ml/min per decade of life . By age 70, a 30 to 50 percent loss of functioning glomeruli is observed . This is likely due to decreases in renal mass, loss of functional nephrons, and diminished renal artery perfusion.
Aging is associated with several pharmacodynamic changes that can alter the therapeutic response and lead to adverse drug reactions. These changes can be due to altered receptor density, receptor affinity, signal transduction (i.e., ability of the cells to respond to receptor occupation), or homeostatic mechanisms . Examples describing the pharmacodynamic variability of commonly used drugs in the ICU are presented in Table 3.1 [23–25].
One in seven Medicare beneficiaries experienced an adverse event when hospitalized in 2008. Between 2007 and 2009, almost 100,000 emergency hospitalizations in the United States were due to adverse drug events (ADEs) in adults 65 years of age and older. Nearly half were age 80 and older . ADEs increase mortality, length of hospitalization, and healthcare costs, yet only 50 percent are reported in US hospitals. Adverse drug reactions or medication errors are only considered ADEs if there is documented harm to the patient (Figure 3.1) . The medication classes most frequently associated with the ADEs in the ICU are drugs with CNS activities (e.g., opioids, sedatives), antimicrobials, drugs with cardiovascular effects, and anticoagulants.
Older adults with complex medical conditions, many of whom require an admission to the ICU during their hospitalization, are at particularly high risk for errors during care transitions and experiencing ADEs secondary to polypharmacy and drug-drug interactions  (Table 3.2). During these transitions, patients are most vulnerable to medication errors, either through medication additions or omissions, that may lead to ADEs. Strategies to improve continuity of care and minimize gaps in care, such as better communication between clinicians, can reduce medication errors and improve patient outcomes .
Medications are frequently administered to critically ill geriatric patients by nonintravenous routes (e.g., oral/enteral, subcutaneous, transdermal, inhalational) when an IV formulation is not available, to prolong the duration of effect, to decrease monitoring requirements, to facilitate ICU discharge, and to decrease drug costs . In general, bioavailability is nearly always lower when drugs are administered by nonintravenous routes. In addition, there can be considerable variability with the rate and extent of absorption, time of onset, magnitude of effect, and duration of action. These differences can have a substantial impact on treatment effect and clinical outcome, so several important factors should be considered (Table 3.3).
An ever-increasing proportion of older adults is obese. Obesity is disproportionately present among patients in the ICU and may be associated with increased ICU mortality . The pharmacokinetic parameters that are most important to consider when dosing medication in older, obese, critically ill patients are volume of distribution and clearance . In general, drugs with a small volume of distribution are typically hydrophilic (e.g., aminoglycosides) with little distribution into adipose tissue . In contrast, drugs with a larger volume of distribution are often more lipophilic, and thus distribution into adipose tissue and other body compartments is extensive. Studies describing the relationship between obesity and clearance have produced mixed results. Some studies demonstrate an increase in clearance (likely due to increased kidney size and blood flow in obesity), whereas others have shown no difference.
There are limited data describing how drugs should be dosed in the obese critically ill patient and even fewer that are specific to geriatrics. Nevertheless, when crafting a drug dosing regimen for the obese, critically ill geriatric patient, the clinician should first assess the degree of obesity present in the individual patient. With mild to moderate forms of obesity (e.g., body mass index [BMI] = 25–39 kg/m2), published dosing recommendations are usually appropriate. With more extreme forms of obesity (e.g., BMI ≥ 40 kg/m2), drug dosing becomes more complicated because these patients were often excluded from formal pharmacokinetic dosing studies [30,31]. In these settings, clinicians should seek clinical trials where dosing has been evaluated in morbid obesity. If these trials do not exist, then pharmacokinetic trials conducted in obese individuals should be sought, and the presence of dose proportionality should be evaluated. Dose proportionality suggests that as weight increases, pharmacokinetic parameters such as volume of distribution and clearance also increase by the same ratio. If dose proportionality exists, then the clinician must weigh the benefits versus risks of using total body weight (for weight-based dosing) or use a dose on the higher end of the dosing range (for non-weight-based dosing). If dose proportionality does not exist, then either lean or adjusted body weight should be used. Generally speaking, there are few drugs that are renally eliminated that display properties of dose proportionality. Other principles for drug dosing in obese, critically ill geriatric patients are listed in Table 3.4.
At the other extreme, older adults may have a low body mass, particularly when patients have a chronic condition that has left them nutritionally depleted. Data describing drug dosing in patients with low body mass are limited. In general, standard drug doses or doses on the lower end of the dosing range should be appropriate. Caution should be exercised with fixed doses of anticoagulant medications (e.g., low-molecular-weight heparin) because standard doses commonly used for prophylaxis may actually achieve therapeutic anticoagulation levels.
Renal insufficiency is common in the critically ill older adults given the clear correlation between increasing age and reduced kidney function. One study identified an age of 65 years or older as an independent risk factor for acute kidney injury (odds ratio [OR] = 1.5; 95 percent confidence interval [CI] = 1.16–1.92) . The most common method to estimate kidney function for the purposes of drug dosing in the ICU is the Cockcroft-Gault equation. However, the validity of this equation in the elderly ICU patient is poor, largely because it relies primarily on the serum creatinine concentration. Creatinine is not very sensitive to changes in the glomerular filtration rate, making the timely recognition of acute kidney insufficiency difficult. Moreover, the serum creatinine is heavily influenced by body muscle mass, which is usually diminished in most elderly patients. This can lead to false assumptions on creatinine clearance if traditional ranges of normal are used to interpret the serum creatinine concentration. To adjust for this issue, some institutions will round low serum creatinine values to an arbitrary value of 1 mg/dl. However, research has demonstrated that this practice will underestimate creatinine clearance and should be avoided because it may lead to subtherapeutic medication doses [33,34].
Many older ICU patients with renal insufficiency will require renal replacement therapy (RRT). Drug clearance with RRT depends on various factors, such as the drug’s volume of distribution, degree of protein binding, molecular weight, and the duration and intensity of RRT. It is essential that clinicians understand local treatment practices, especially dialysis flow rates, because this will markedly affect drug removal. Discrepancies between local practices and published dosing recommendations may lead to either an over- or underestimation of extracorporeal drug removal and thus increase the likelihood for treatment failure or an adverse drug reaction.
Several resources exist to assist with drug dosing in RRT, but substantial variability exists in the dose recommended by each reference [35–37]. Many of these recommended doses are based on studies that used older dialyzer membranes that have poor permeability and lower effluent rates or were extrapolated from pharmacokinetic data in noncritically ill patients or those with chronic kidney disease. This is important because patients with chronic kidney disease also have compromised nonrenal clearance mechanisms, whereas in patients with acute kidney injury, nonrenal clearance mechanisms may be preserved. This has the potential to lead to the under dosing of medications that has been demonstrated with antibiotics such as imipenem, meropenem, and vancomycin . In fact, several reports have demonstrated an inability to reach pharmacodynamic goals with doses commonly recommended for RRT . Principles of drug dosing in elderly patients with acute kidney injury (AKI) or RRT are listed in Table 3.5.
Drug dosing in hepatic dysfunction is challenging because of the general inability to estimate liver function. Although tests exist to detect hepatocellular changes (e.g., aspartate aminotransferase [AST], alanine aminotransferase [ALT]) and evaluate synthetic function (e.g., international normalized ration [INR]), none of these reflect the ability of the liver to metabolize drugs. Some medications can be adjusted based on the Child-Pugh score, but specific recommendations for commonly used-ICU medications using this strategy are limited. Nevertheless, dosing adjustments are not typically suggested until hepatic dysfunction becomes moderately (class B) or severely (class C) impaired .
Pharmacokinetic changes that occur with hepatic dysfunction are related to reductions in drug metabolism, plasma protein synthesis, and/or liver blood flow. Hepatic drug clearance is typically classified based on the dependence on hepatic blood flow (i.e., high-extraction drugs) or on the activity of drug-metabolizing enzymes (i.e., low-extraction drugs). Drugs that have a high degree of hepatic extraction are expected to have an increased bioavailability and decreased clearance in conditions associated with decreased hepatic blood flow. For medications that are administered orally, both the initial dose and maintenance doses should be reduced. For medications administered intravenously, only a reduction in the maintenance dose is necessary. For low-extraction drugs, metabolism is more so dependent on the degree of binding to albumin and the activity of metabolizing enzymes (e.g., CYP-450). This is further complicated by the fact that all CYP-450 enzymes do not uniformly decrease at the same severity of liver disease. Instead, a sequential progressive model has been described . According to this model, patients with mild hepatic dysfunction will experience a decrease in CYP-2C19 activity, but CYP-1A2, CYP-2D6, and CYP-2E1 will be sustained. As the level of hepatic dysfunction becomes more severe, a sequential reduction in CYP-1A2 activity, followed by CYP-2D6 activity, is noted. CYP-2E1 activity remains relatively preserved until hepatic decompensation or hepatorenal syndrome develops. Principles of drug dosing in hepatic dysfunction are listed in Table 3.6.
Drug/Disease-Specific Pharmacotherapy Issues
Pain and discomfort are prevalent in critically ill older adults and thus require around-the-clock evaluation and treatment when they are present. Self-reporting of pain is considered the reference standard for pain assessment in this population. However, if a patient is nonverbal, the 2013 Pain, Agitation, and Delirium (PAD) Guidelines recommend the use of the Behavioral Pain Scale or the Critical Care Pain Observational Tool in the ICU . Most ICU patients can be successfully treated with intermittent opioid therapy. Continuous opioid infusions may be required in some patients particularly when an analgesia-sedative approach to sedation management is being used. Older adults are at particularly high risk for developing sedative-associated coma and delirium particularly with benzodiazepine use. Increasing evidence suggests that using an opioid-first approach is safer and is associated with improved patient outcomes . Despite strong evidence that patients who remain in a lightly sedated state have improved outcomes (e.g., less delirium, a shorter duration of mechanical ventilation, reduced posttraumatic stress disorder), over sedation remains a major concern in many ICUs. Deep sedation, particularly coma, should be avoided in all older adults unless it is a goal of therapy, given that even short periods spent in this state are associated with increased mortality [42,43].
The ABCDEF bundle (assess, prevent, and manage pain; both spontaneous awakening trials and spontaneous breathing trials; choice of drugs; delirium: assess, prevent, and manage; early mobility and exercise; and family engagement and empowerment) has been shown to improve both short- and long-term outcome in older adults and help return patients back to their pre-ICU functional and cognitive status [2,42,44]. It should be emphasized that the appropriate choice of analgesic and sedative therapy (if needed) is only one factor affecting an older adult’s cognitive and functional status after leaving the ICU.
Consideration of potential adverse events is the most important criterion when selecting opioid and sedative therapy in older adults given that these agents are usually administered at higher doses and for more prolonged periods than in non-ICU settings . Adverse drug events with these agents are common in the ICU given that they are usually administered at far higher doses and for longer periods of time than outside the ICU. Furthermore, critically ill patients have a higher prevalence of end-organ dysfunction (e.g., renal, hepatic) that may result in higher drug concentrations than in patients on the floor . Factors such as altered postreceptor binding, downregulation of receptors, and brain dysfunction may dramatically alter the response of ICU patients to these agents . Cardiac dysfunction may increase the risk for dysrhythmias and hypotension. Adjuvants in the injectable formulations of sedatives (e.g., propylene glycol in parenteral lorazepam) may result in additional toxic effects.
Fentanyl is a synthetic opioid that is preferred over morphine given that it is associated with less hypotension and bronchospasm and its clearance is not affected by renal dysfunction . Fentanyl has been reported to cause muscle rigidity and bradycardia. Fentanyl patches should be avoided for acute analgesia because the time to reach peak effect is delayed by up to 24 hours after patch application, and a prolonged drug effect is seen after patch removal. Hydromorphone has a half-life of 2 to 3 hours and also undergoes glucuronidation similar to morphine. However, the hydromorphone-3-glucuronide metabolite that is produced is inactive, making hydromorphone the opioid of choice for use in patients with end-stage renal disease. Ketamine and other opioid-sparing analgesics such as acetaminophen are being increasingly administered in the ICU to reduce opioid administration and the potential side effects associated with opioid use (e.g. constipation, respiratory depression, risk of future addiction) .
Compared with benzodiazepines (e.g., lorazepam and midazolam), propofol and dexmedetomidine are associated with faster neurologic recovery and a shorter duration of mechanical ventilation after discontinuation . Moreover, given that benzodiazepine use is a well-established risk factor for delirium [48,49], the 2013 PAD Guidelines recommend that propofol or dexmedetomidine be used judiciously in older adults who require continuous sedation .
Propofol is an intravenous general anesthetic agent that has a rapid onset and offset of action and thus provides clinicians with a sedative option that is far more titratable than other sedative options. Propofol’s duration of effect is longer in the very old elderly. The key safety concerns with propofol are bradycardia, hypotension, hypertriglyceridemia, and the propofol-associated infusion syndrome (PRIS). Propofol should not be bolused nor administered at a dose greater than 60 µg/kg per minute in the elderly. Serum triglyceride levels should be checked at least twice weekly. Propofol should be discontinued in a patient who exhibits sudden hypotension, metabolic acidosis, and cardiac failure given that these are the earlier and more common manifestations of PRIS.
Dexmedetomidine is a centrally acting alpha-2 agonist that has sedative and analgesic properties but no effect on respiratory drive . Older adults are more likely to experience bradycardia and/or hypotension with its use, particularly patients with severe congestive heart failure, and thus it should be initiated at a dose of no more than 0.2 µg/kg per hour and titrated upward with care. Dexmedetomidine can be particularly effective in older adults with agitated delirium or in those who are admitted with severe alcohol withdrawal.