Unique Aspects of Antimicrobial Use in Older Adults
- Department of Pharmacy, Wake Forest University Baptist Medical Center, Winston-Salem, North Carolina
- Reprints or correspondence: Dr. John C. Williamson, Dept. of Pharmacy, Wake Forest University Baptist Medical Center, Medical Center Blvd., Winston-Salem, NC 27157 (johnwill{at}wfubmc.edu).
Abstract
Elderly persons consume a significant proportion of health care resources. Antimicrobials are just one class among many pharmaceuticals that are prescribed more frequently to elderly patients than to younger patients. There are unique aspects of antimicrobial use in elderly persons that make prescribing complicated and monitoring unpredictable. Physiologic changes associated with aging result in altered pharmacokinetics, and accurate estimates of renal function cannot be made with standard methods. Together, these qualities make antimicrobial dosing difficult. Because of a higher prevalence of other chronic diseases, there is a greater propensity for polypharmacy and a resulting risk of an adverse event or a significant drug interaction. Lastly, irrespective of altered pharmacokinetics, adverse effects of many antimicrobials are more common in elderly persons, which introduces an added dimension to ensuring safety with antimicrobial therapy.
In addition to medicines used for other conditions, an 83-year-old woman takes verapamil and hydrochlorothiazide for hypertension. She goes to her podiatrist because of a small ulcer on the bottom of her foot. The podiatrist realizes that the patient is economically disadvantaged and prescribes erythromycin to treat the ulcer. One week later, the woman reports to her primary care physician with dizziness, weakness, hypotension, and bradycardia. The primary care physician stops treatment with verapamil and erythromycin. After a washout period of a few days, verapamil therapy is successfully restarted without complications. The adverse event occurred because both verapamil and erythromycin are substrates and inhibitors of the cytochrome P450 3A4 enzyme; a 2-way interaction exists, resulting in potentially toxic concentrations of these drugs. Several factors contribute to such unexpected or unfortunate outcomes. The risk of polypharmacy, the involvement of multiple health care professionals capable of prescribing medication, the likelihood of an adverse drug event, and the chance of a harmful drug interaction are all increased in elderly persons.
General Principles of Prescribing Antimicrobials to Elderly Persons
Risk of an adverse drug event, a harmful drug interaction, and the role of older adults as a reservoir for resistant pathogens all justify a thorough consideration of the patient's true need for antimicrobial therapy. Guidelines exist that list minimum criteria for initiation of antibiotic therapy, particularly in the long-term care setting [1]. The guidelines emphasize the importance of initiating therapy only when there is a clear potential clinical benefit and discourage the use of antimicrobials in ways that promote resistance. Empirical broad-spectrum antibiotics should be narrowed when a pathogen is identified. Such targeted therapies minimize the chance of resistance and may limit the scope of potential adverse events.
Many practitioners have lived by the credo “start low, go slow” when initiating or titrating medical therapy in an elderly patient. Unfortunately, antimicrobial therapy in older individuals is influenced by more than just physiologic changes. Comorbidities and immunosenescence, the dysregulated immune function associated with aging, predisposes elderly persons to suboptimal therapeutic efficacy. A study of levofloxacin for treatment of nosocomial pneumonia found that advanced age adversely affects the probability of both clinical and microbiological success [2]. Only the pharmacodynamic parameter area under the curve (AUC):MIC ratio was more predictive of microbiological response than was age. Although it has been proven that older patients are more likely to have medication-related adverse events, a finding that supports the “start low, go slow” credo, this approach must be balanced with aggressiveness in dosing to achieve optimal pharmacodynamics and sufficient clinical outcomes in the face of immunosenescence, comorbidities, and limited functional reserve.
Pharmacokinetic Considerations for Older Individuals
The physiological changes that occur with aging affect the pharmacokinetic parameters of many medicines, including antimicrobials. The magnitude of these alterations can vary greatly between individuals. Table 1 lists many of the common physiologic and resulting pharmacokinetic changes that occur in elderly persons.
Table 1
Physiologic and pharmacokinetic changes in elderly persons.
Absorption
Age-related changes occur that reduce gastric acid secretion. Dysfunctional gastric mucosal parietal cells secrete less acid, resulting in hypochlorhydria or achlorhydria. Antimicrobials, such as itraconazole, ketoconazole, sulfonamides, dapsone, pyrimethamine, and atazanavir, exhibit reduced absorption in this setting. Dapsone, specifically, is insoluble at a neutral pH. Clinical failures of dapsone to prevent Pneumocystis jiroveci pneumonia have been reported for patients receiving concomitant didanosine chewable tablets. The proposed reason for the failures is the increased gastric pH resulting from the buffer in didanosine [5]. Increased absorption of acid-labile antimicrobials, such as erythromycins or penicillins, is also possible, resulting in higher-than-expected concentrations and potential adverse effects [6]. Lastly, drugs prescribed for underlying gastroesophageal reflux disease or for peptic ulcer disease may interact with antimicrobials by means of a similar mechanism.
There is a significant reduction in splanchnic blood flow among elderly patients aged 60–73 years, and it is likely that similar changes are present in gastrointestinal blood flow [7]. Reduced surface area of the small intestine and increased gastric emptying time have also been described [8]. In patients with gastroparesis, which is often associated with underlying diabetes, less ciprofloxacin is absorbed, compared with healthy volunteers. In diabetic patients with gastroparesis, the AUC ratio of a single 400-mg intravenous dose of ciprofloxacin and a 500-mg oral dose was 0.84 (90% CI, 0.68–0.98). This is in contrast to an AUC ratio of 1.04 (90% CI, 0.94–1.17) reported for healthy volunteers [9].
Distribution
The amount of lean and adipose tissues, the total body water, and the body's protein-binding capacity all contribute to interpatient variability in the volume of distribution. The proportion of adipose tissue relative to total body weight is estimated to increase 18%-33% in elderly men and 33%-48% in elderly women [10]. With advancing age, both intracellular and extracellular total body water can decrease by as much as 15% [6]. These changes affect the volume of distribution and the half-life of corresponding lipid-soluble and water-soluble antimicrobials. Elderly patients may also take diuretics chronically, contributing to decreased extracellular water levels and the risk of higher serum concentrations.
Cardiac output decreases by ∼1% yearly after 30 years of age. Generally, blood flow is maintained at near-normal levels for skeletal muscles and for coronary and cerebral arteries, but blood flow to the liver, kidneys, and gastrointestinal tract is decreased. A drug dependent on hepatic blood flow will have a slower rate of clearance. Similarly, a corresponding decrease in renal blood flow reduces the elimination of drugs cleared via renal excretion, such as aminoglycosides, β-lactams, glycopeptides, and daptomycin. The rate of daptomycin clearance was reduced by 35% and the AUC was increased by 58% in a sample comprised of elderly volunteers; however, no dosage reduction is recommended [11]. As with other antimicrobials, the incidence of many adverse events was greater among elderly persons in trials of daptomycin for skin and skin-structure infection (Cubist Pharmaceuticals, unpublished data).
Pharmacokinetic changes are possible because of altered plasma protein levels in elderly persons. In addition, concurrent use of a larger number of medications by elderly persons increases the risk that another medicine will displace the antimicrobial from its protein binding site. For example, sulfadiazine-unbound concentrations are significantly higher in elderly patients receiving concurrent medications than in elderly patients receiving no concurrent medication [12].
Metabolism
Very little research has been performed to fully characterize metabolism alterations in elderly persons. Metabolism is expected to be reduced because hepatic blood flow, liver mass, and intrinsic enzyme activity (phase 1 metabolism) are all lower in older persons, compared with young adults [10]. Cytochrome P-450 substrates, such as macrolide antibiotics, azole antifungals, and antiretrovirals, may exhibit a longer half-life in elderly persons.
Renal Elimination
Parameters of renal function, such as glomerular filtration rate, renal blood flow, and creatinine clearance, at 20 years of age decrease by one-third to one-half by 90 years of age [13, 14]. In addition, there is a trend toward an increasing rate of decrease with advancing age [13, 14]. In young patients with stable renal function, serum creatinine level serves as a predictor of renal function; however, it is not as reliable in elderly persons. Creatinine is a product of muscle breakdown and is typically lower in elderly persons because of decreased muscle mass. So, normal serum creatinine levels in elderly persons have the potential to be misleading. Dosage adjustments, such as reducing the dose or increasing the dosing interval, should be considered for elderly persons when prescribing antimicrobials that are renally eliminated.
The Cockcroft and Gault equation is one of many calculations used to predict creatinine clearance [15]. Pharmaceutical industry dosing recommendations for renally eliminated antimicrobials are based on this or similar methods of estimation. Such equations are not as accurate for estimating creatinine clearance in elderly persons, and overestimation can occur [16]. In elderly patients with a very low creatinine level (i.e., <1 mg/dL), many clinicians elect to substitute the actual creatinine level with 1 mg/dL in calculations performed using the Cockcroft and Gault equation. This reduces the risk of overestimating creatinine clearance. Likewise, cautious prescribing is warranted for elderly patients with elevated serum creatinine levels. A creatinine level of 1.5 mg/dL may suggest a much more significant decrease in renal function in an elderly patient, compared with a younger patient.
Recently, the equation developed from the Modification of Diet in Renal Disease (MDRD) study was found to estimate renal function more precisely than the Cockcroft and Gault equation [17]. Figure 1 allows comparison of the 2 equations. Because of its greater complexity, the MDRD equation is not an easy, routine calculation. However, for patients with a potentially misleading serum creatinine level, such as elderly persons, it may provide a more accurate assessment of creatinine clearance.
For example, a 70-year-old white man admitted to the hospital with pneumonia has the following clinical and laboratory parameters: ideal body weight, 78 kg; serum creatinine level, 1.4 mg/dL; blood urea nitrogen level, 68 mg/dL; and albumin level, 2.3 mg/dL. The Cockcroft and Gault and MDRD equations provide estimates of 54 and 37 mL/min, respectively, for this hypothetical patient. If product-specific recommendations for dosing are followed, it is possible that these estimates would lead to different prescribed doses. If levofloxacin is prescribed, a dose of 500 mg every 24 h is appropriate according to the Cockcroft and Gault estimate, whereas a dose of 250 mg every 24 h is appropriate according to the MDRD estimate [18]. Sanaka et al. [19] developed an equation that incorporates albumin, similar to the MDRD equation [17]. In a small population of elderly patients with muscle atrophy, the Sanaka equation predicted creatinine clearance more accurately than did the Cockcroft and Gault equation. Regardless of which method is used to estimate renal function, careful dosing is required to maintain efficacy and minimize toxicity due to reductions in elimination and clearance that occur with increasing age.
Impact of Polypharmacy
Polypharmacy is most commonly defined as the administration of ⩾5 medications. On the basis of this definition, polypharmacy has been documented at a rate of 39% among elderly persons in the community [20]. Polypharmacy has a direct correlation with the number of concomitant disease states and with increasing numbers of visits to physician offices each year [20]. Thus, it is not too surprising that elderly persons are more likely to experience polypharmacy than are younger patients. Recent research suggests that polypharmacy may even be increasing among elderly persons. [21].
There are a number of undesirable consequences of polypharmacy in elderly persons. In a retrospective chart review of emergency department visits by patients ⩾65 years of age, 10.6% of all visits were attributable to adverse drug events (ADEs) [22]. ADEs were more frequent among patients who were taking >5 medications. The antibiotic class was one of the major drug classes implicated in this population, accounting for 16.7% of all ADEs and 25% of definite or probable ADEs. In a similar review of deaths among patients admitted to an internal medicine service, 18.2% of the deaths were either directly (48.1%) or indirectly (51.9%) associated with ⩾1 medicine [23]. The average age of the patients with a fatal ADE was 72 years, and the median number of medicines in this group was higher than that for patients without a fatal ADE (5 vs. 4). Finally, a case-control study of nursing home residents found that use of antimicrobials is a greater independent risk factor for an ADE than is use of drugs from many other classes of medication, including antipsychotics or antidepressants [24]. This study also found an association between an increasing number of medicines and the risk of ADE.
Polypharmacy is associated with a significantly increased risk of a harmful drug interaction, a consequence not too unrelated or distant from the risk of an ADE. Indeed, the result of many drug interactions is an ADE. The case patient described in the introduction demonstrates the potential clinical significance of a drug interaction. The outcome of this case could have been much worse. A recent study determined that the risk of sudden cardiac death was 5 times greater for patients receiving the combination of erythromycin and either verapamil or diltiazem [25]. Table 2 lists many of the common drug interactions encountered with antimicrobial use in elderly persons.
Table 2
Drug interactions among antimicrobials prescribed to elderly persons.
Antimicrobial-Induced Adverse Events in Elderly Persons
It is well recognized that ADEs occur more frequently among older patients. Polypharmacy, comorbidities, and difficulty with adherence to therapy, as well as age-related changes in pharmacokinetics and pharmacodynamics, contribute significantly to the higher incidence. Antimicrobials are routinely listed among drug classes deemed to confer a high-risk for ADEs. This section addresses many of the notable ADEs that have a significant association with old age. See table 3 for a more comprehensive list.
Table 3
Common antimicrobial-induced adverse events in elderly persons.
Insufficient adjustments to the dosage for patients with renal dysfunction remains a significant contributor to antimicrobial-induced adverse events. A study of antibiotic dosage errors involving 1044 hospitalized patients >80 years of age revealed an overall rate of dosing error of 34% [29]. Third-generation cephalosporins and gentamicin were most commonly associated with dosing errors among the antibiotics studied, with error rates of 50% and 65%, respectively. In addition, most of the elderly patients in the study required dosage adjustments, particularly those >85 years of age. This study highlights the importance of assessing renal insufficiency in elderly persons as a means to reduce the risk for adverse events.
ADEs due to antiviral treatments for influenza are more commonly associated with the matrix 2 (M2) protein inhibitors, amantadine and rimantadine. CNS-related adverse events have been reported in up to 41% of elderly patients receiving amantadine [30]. Symptoms such as hyperexcitability, slurred speech, insomnia, lightheadedness, difficulty concentrating, nervousness, anxiety, confusion, psychosis, hallucinations, and seizures are reported more frequently for amantadine therapy than for rimantadine treatment and are more likely to occur with concurrent use of antihistamines or anticholinergic agents [31]. Such adverse events may predispose to falls and contribute further to morbidity and mortality among elderly persons [30]. To minimize the risk of toxicity, most authorities recommend a reduction in the dose of both amantadine and rimantadine for patients ⩾65 years old, from twice-daily to single-daily dosing [32]. If abnormal renal function exists, even greater dose reductions are appropriate (e.g., every 48 h or one time only). The neuraminidase inhibitors oseltamivir and zanamivir are less likely to cause CNS-related adverse events. Zanamivir may be inappropriate for some elderly individuals whose diminished dexterity prevents them from operating the dry-powder inhaler.
Fluoroquinolones are reportedly well tolerated. However, elderly patients who have several comorbidities and are receiving multiple medications are often excluded from phase III studies, including those involving fluoroquinolones. Depending on the exclusion criteria, elderly patients included in these studies may be considered to be “healthy” patients, and tolerance to the study drug may appear to be equivalent to the comparator. Nicholson et al. [33] reported results from a large phase IV clinical trial that assessed the safety and efficacy of gatifloxacin in elderly persons. Comorbidities such as diabetes mellitus and cardiovascular disease were not part of the exclusion criteria. The incidence of ADEs among patients 65–79 years old was similar to the incidence among those aged <65 years, but patients ⩾80 years of age were more likely to experience an ADE. Gastrointestinal events, such as nausea, vomiting, and diarrhea, were more common among older patients. Some CNS effects were also more common among patients aged ⩾80 years, including dizziness and hallucinations.
Additionally, these investigators discovered a rate of hyperglycemia among very old patients that was higher than that among patients aged <80 years. Other reports of both hypoglycemia and hyperglycemia associated with fluoroquinolone therapy have appeared in the literature [34, 35]. Many of these cases occurred in elderly patients with undiagnosed or borderline diabetes or with age-related renal dysfunction that led to higher than usual fluoroquinolone concentrations. When possible, careful monitoring of blood glucose levels is recommended during fluoroquinolone therapy for patients with glucose homeostasis problems or for patients with underlying renal dysfunction.
Although the incidence is rare, cases of torsades de pointes have been reported in several elderly patients after treatment with fluoroquinolones [36]. Predisposing factors include hypokalemia, hypomagnesemia, preexisting heart disease, concomitant QT interval-prolonging agents (e.g., class IA or III antiarrhythmics), and renal insufficiency.
Older age has been associated with an increased risk of hepatotoxicity during treatment with antituberculosis agents. In a study of 131 subjects taking isoniazid plus rifampin, 38% of patients <60 years of age developed increased aminotransferase levels, compared with 18% of patients ⩾60 years of age (P = .02) [37]. Two additional studies document similar results, not only with hepatotoxicity, but with other serious adverse events as well [38, 39]. Age of at least 60 years was significantly associated with hepatitis (hazard ratio, 7.7; 95% CI, 1.5–40) and dyspepsia (hazard ratio, 6.4; 95% CI, 1.2–36). Interestingly, pyrazinamide was associated with the highest incidence of any serious adverse effect and was equivalent to isoniazid with respect to hepatotoxicity [38].
Advanced age is documented as one of many risk factors for aminoglycoside-associated nephrotoxicity [40–42]. It is rare that an elderly person will need treatment that requires dosing >2 times/day. Many older patients with a normal serum creatinine level will require a longer interval (e.g., 24 or 48 h), especially patients ⩾80 years of age. Furthermore, the “once-daily” or “extended interval” method of dosing may not be appropriate for elderly patients [43]. Certainly, monitoring of drug concentrations should be performed to more accurately assess drug clearance and to make appropriate dosage adjustments.
Vestibular toxicity due to aminoglycoside use can be particularly devastating, especially in elderly persons whose comorbidities or concurrent medications may already predispose to falls. Risk factors for aminoglycoside induced vestibular toxicity include age, renal dysfunction, high daily or cumulative doses, and concurrent ototoxic medications, such as loop diuretics or vancomycin [44]. In a meta-analysis of 3 randomized, controlled trials, only age was a risk factor for auditory (cochlear) toxicity after adjustments for confounding variables [45]. Logistic regression analysis revealed an increasing probability of ototoxicity with advancing age, from 3% among patients aged 14 years to 26% among those aged 90 years. During receipt of aminoglycosides, audiometry may prove useful by detecting high-frequency hearing loss, which usually precedes clinical hearing loss. If long-term aminoglycoside therapy is prescribed, audiometry should be performed at baseline and every 2 weeks during therapy. Increasing age and high concentrations are also associated with hearing loss and/or tinnitus due to macrolide (or azalide) antibiotics and vancomycin. In such cases, ototoxicity is generally associated with maximally recommended doses, but it is usually reversible [46–49].
Summary
Many factors, both patient related and treatment related, contribute to the pitfalls of antimicrobial use in elderly persons. Altered physiology leads to altered pharmacokinetics, so the most appropriate dose in an elderly person may not be the dose that is typically used in the general population. The risk of polypharmacy in elderly persons requires a review of the patient's medication list to assess potential drug interactions. However, this step is not without pitfalls, because of the potential for multiple prescribers and an incomplete medication list. Even if all precautions are taken, it may be inevitable that an adverse event occurs. For these reasons, clinicians are wise to prescribe judiciously to prevent unnecessary treatments and to monitor closely any antimicrobial therapy, however benign it may seem.
Acknowledgments
Potential conflicts of interest. J.C.W. has recently received research funding by Merck and is a member of a speakers' bureau for Elan Pharmaceuticals. C.M.F. and H.L.C.: no conflicts.
- Received August 19, 2004.
- Accepted November 28, 2004.
- © 2005 by the Infectious Diseases Society of America
References
This Article
- AbstractFree
- » Full Text (HTML)Free
- Full Text (PDF)Free
Classifications
-
- Aging and Infectious Diseases
Services
Citing Articles
Google Scholar
PubMed
Related Content
Share
- Email this article
-
CiteULike
Connotea
Delicious
Facebook
Google+
Mendeley
Twitter
Published on behalf of
IDSA Members: For your free access to this journal, log in via the IDSA members area.
The Journal
Impact Factor: 8.186
5-Yr impact factor: 7.898
For The Media
For Authors
- Services for authors
- Self-archiving Policy
- IDSA Conflict of Interest policy
- Instructions for Authors
- ICMJE Form
- Submit Now!
Open access options for authors visit Oxford Open
This journal enables compliance with the NIH Public Access Policy
Corporate Services
Alerting Services
Related IDSA Publications
