Case Studies in Geriatric Emergencies: Polypharmacy

Emerg Med 39(6):10, 2007

By Michael Stern, MD

After being found unconscious in her apartment by a neighbor, an elderly woman presents to an emergency department complaining of abdominal pain and cough. The author analyzes the multiple medications that precipitated this incident.

With advances in health education, technology, and pharmacotherapy, people over age 65, the fastest-growing segment of the U.S. population, are living longer, in many cases with multiple chronic medical conditions such as heart disease, diabetes, and arthritis. As a result, they are more likely to be taking numerous medications, often prescribed by different physicians. In addition, physiologic changes associated with aging alter the pharmacokinetics and pharmacodynamics of drug metabolism, which, in turn, affects potential drug toxicities. The adverse drug reactions and toxic interactions that may accompany polypharmacy represent a growing problem for the elderly.

In this article, I will review the case of an elderly woman who suffered a fall and loss of consciousness as a result of a medication-induced gastrointestinal bleed. Epidemiology, age-related contributing factors, risk factors, and management goals for adverse drug reactions in the setting of polypharmacy in the elderly will be discussed.

PATIENT PRESENTATION

A 78-year-old woman was brought in by EMS to our emergency department on a long board with a cervical collar in place after having been found unconscious on the floor of her apartment by a neighbor. Her vital signs were stable in the field, EMS reported. She had been weak but alert and oriented to person and place and conversant throughout transport to the emergency department.

The patient was placed on a cardiac monitor, given two liters of oxygen via nasal cannula, and fully exposed. Her initial vital signs were: heart rate, 94; blood pressure, 116/78; respiratory rate, 22; temperature, 97.7°F. Her oxygen saturation was 96% on room air; her glucose level, 85 mg/dl; and her Glasgow Coma Scale score, 15. The primary survey revealed no airway, breathing, or circulation abnormalities. Two 18-gauge intravenous lines were established. A full set of admission labs and a type and cross were drawn. An initial bolus of 500 ml of normal saline was administered, followed by a normal saline infusion drip at 100 ml/hr. The other line was heparin-locked.

Because the patient was hemodynamically stable and conversant, she was able to provide an account of what happened and a brief medical history. She reported feeling abdominal pain and nausea after eating dinner the night before. Before going to bed, she had a bowel movement that was black. After leaving the bathroom, she became light-headed and experienced palpitations.

She did not remember falling but recalled waking up in the middle of the night on her living room carpet. She stated she was unable to get up by herself and felt profound weakness. She was unaware of how long she lay on the floor before her neighbor arrived in the morning and called EMS.

The patient complained of dull epigastric abdominal pain, mild nausea, and dizziness. She denied any headache or head trauma, neck pain, visual symptoms, or motor or sensory changes, and reported no chest pain, shortness of breath, vomiting, diarrhea, back pain, or urinary symptoms. She stated she had had a cold with a productive cough and congestion for four days, for which her internist had prescribed levofloxacin. She denied fever or chills. She stated she had been taking ibuprofen for her body aches and an over-the-counter (OTC) elixir for her cough for three days, as well as a liquid antacid for her upset stomach after dinner the night before. She took diazepam before getting ready for bed.

She had long-standing hypertension, coronary artery disease, atrial fibrillation, congestive heart failure, and osteoarthritis. She had had a cholecystectomy more than 10 years ago. Her medications included metoprolol, digoxin, warfarin, furosemide, atorvastatin, baby aspirin, celecoxib, paroxetine, and diazepam as needed for sleep, which her psychiatrist had prescribed. She had no known drug allergies. Her last colonoscopy was more than five years ago. She denied any history of gastritis or ulcer disease. She was a retired librarian who had been living alone since her husband died three years ago. Occasionally, she said, she would drink wine with dinner.

Physical examination revealed an elderly woman who appeared pale and weak but in no apparent distress. There was no evidence of head trauma. Her HEENT exam was normal except for mildly dry mucous membranes. Her neck was nontender in the cervical collar. The lung exam revealed mild rhonchi bilaterally but no rales or wheezing. Her heart exam revealed an irregularly irregular rhythm without a murmur; the abdominal exam revealed normal bowel sounds and mild epigastric tenderness on palpation but no rebound or guarding. Her pelvis was stable and nontender, and her back and genitourinary exams were normal. The rectal exam revealed good tone and dark black stool that was positive for blood on hemoccult testing.

Her extremities were unremarkable and her skin was warm and dry. She had a normal neurovascular exam distally. Her neurologic exam revealed her to be alert and oriented to person and place but not to time. There were no focal neurologic findings. Although hesitant and slow to respond, she answered questions appropriately and followed commands.

A 12-lead ECG revealed atrial fibrillation at 96 bpm and left ventricular hypertrophy. The characteristic ST-T segment finding of digoxin was present. There were no ischemic changes. A troponin-I, brain natriuretic peptide (BNP), and digoxin level were added to the admission labs. A portable chest x-ray and anteroposterior pelvis x-ray were negative for any pathology. A bedside FAST (focused abdominal sonography for trauma) exam revealed no bleeding.

The patient’s vital signs and lung exam were reassessed, without change. She was given another 500-ml bolus of normal saline. Her blood pressure rose to 124/84. Boluses of famotidine 20 mg and esomeprazole 40 mg were administered intravenously, followed by an esomeprazole drip.

The patient was transported on cardiac monitoring to the radiography suite for a noncontrast head computed tomography (CT) scan and cervical spine CT. The head CT revealed no bleeding, mass, or mid-line shift, and the cervical spine was negative for any fracture or dislocation.

Lab test results revealed a normal white blood cell count and platelet count, but a hematocrit of 29.4, well down from a hematocrit of 38 taken two months earlier. The chemistry profile revealed normal electrolyte levels except for a mild hyponatremia with a sodium level of 131 mEq/L. Blood urea nitrogen/creatinine revealed a pre-renal azotemia with a ratio of 39/1.7. Her INR was therapeutic. Troponin-I was negative, BNP was 248 pg/ml, and her digoxin level was mildly supratherapeutic at 4.2 mcg/ml.

She was given a transfusion of two units of packed red blood cells, with 20 mg of furosemide administered intravenously between units. Her blood pressure climbed to 136/90. She was admitted to the medicine step-down unit for management of a presumed upper gastrointestinal bleed. She remained stable, requiring no further transfusions. Her mental status returned to complete orientation. An esophagogastroduodenoscopy performed later that day revealed a small bleeding gastric ulcer, likely secondary to a combination of nonsteroidal anti-inflammatory drugs and aspirin, and exacerbated by her baseline coumadin, which was likely supra-therapeutic because of the levofloxacin antibiotic.

She had a negative colonoscopy. A myocardial infarction was ruled out and she remained hemodynamically stable throughout her hospitalization. Her pre-renal azotemia resolved and her creatinine returned to her baseline 1.2 mg/dl. She was discharged on hospital day 3 with a prescription for lansoprazole and famotidine. She was instructed to stop taking the celecoxib and aspirin for now, as well as the ibuprofen. She was advised to contact her internist and psychiatrist regarding possible medication changes that might decrease the risk for future adverse events. In addition, she was instructed to call her doctor before she took any OTC medications.

DISCUSSION

As the elderly population in the United States continues to grow, their consumption of medications is increasing. Currently, the elderly take approximately 40% of prescribed drugs, roughly twice the percentage that younger adults take. It is projected that this number will increase to 50% by 2020. The average community-dwelling older adult takes 4.5 prescription drugs and 2.1 OTC medications.

Not surprisingly, this population suffers a disproportionately high percentage of adverse drug reactions (ADRs), at least twice that of younger persons. Evidence suggests that physicians’ prescribing practices are playing a big role in this. One large community-based study found that nearly 25% of persons aged 65 years or older were taking 1 of 20 drugs considered to be inappropriate for this population by an expert panel.

The consequences of polypharmacy are far-reaching. Between 3% and 28% of hospital admissions can be attributed to drug-related adverse events. And fatal ADRs would be the fifth leading cause of death in the United States if they were classified as a distinct category. The estimated cost of drug-related morbidity and mortality in this country is more than $136 billion.

EFFECTS OF PHYSIOLOGIC CHANGES

There are several contributing factors related to the normal physiologic changes of aging that put the geriatric population at greater risk for drug-related adverse events. A decreased functional reserve capacity in the major organ systems can cause physiologic decompensation in older patients in response to certain medications, especially those affecting the heart, brain, and kidneys, in the same way that acute illness or trauma can alter an elderly person’s tenuous homeostasis. For example, because cardiac output diminishes with age, the heart relies increasingly on the inotropic support provided by endogenous catecholamines. As a result, giving a beta blocker to an elderly hypertensive patient may precipitate congestive heart failure. Nonsteroidal anti-inflammatory drugs (NSAIDs) may cause acute renal failure in older patients with borderline renal function and a dependence on prostacyclin-mediated renal afferent arteriolar vasodilatation to maintain glomerular blood flow. In the case presented here, the patient’s renal function had decreased from her recent baseline, most likely due to the combination of mild dehydration and the ibuprofen she was taking for her myalgias.

The gastrointestinal system is also affected by aging. A decrease in gastric bicarbonate secretion, blood flow, and mucosal function, as well as delayed gastric emptying time, all contribute to a loss of stomach protection and an increased risk of gastritis or ulcer formation. Overuse of NSAIDs or aspirin, or the concomitant use of both, places an elderly patient at significant risk for a life-threatening gastrointestinal bleed, as illustrated by the patient presented here.

Aging affects both pharmacokinetics (bioavailability, distribution, and clearance) and pharmacodynamics. Although the above-mentioned changes in the gastrointestinal system play a role in the decreasing bioavailability of most drugs, it is drug interactions that play a more important role. In the case above, the levofloxacin that the patient was prescribed for her upper respiratory infection was likely not as well absorbed because of the magnesium-based antacid she was taking. Also, the antihistamine in her cough syrup had anticholinergic effects that delay gastric emptying further and can decrease the absorption of other medications, in addition to exacerbating the harmful effects of NSAIDs and aspirin on the stomach. In addition, the interaction of coumadin and ibuprofen almost certainly increased the patient’s bleeding risk. And the concomitant fluoroquinolone antibiotic likely prolonged the coumadin’s effect due to an unknown mechanism, perhaps secondary to a decrease in vitamin K production and altered coumadin metabolism.

ALTERED DRUG DISTRIBUTION

Age-related changes in body composition affect drug distribution. After a drug is administered, the volume and rate of distribution determine its peak concentration, which frequently plays a role in toxicity and adverse effects. The volume of distribution is determined by solubility characteristics in lipids and water, the degree of plasma protein binding, and the extent of tissue uptake and binding. Blood flow through different tissues and organs determines drug distribution rates. Because fat is not as well perfused as lean tissues and solid organs, drug distribution in adipose tissue occurs more slowly.

With aging, lean body mass declines to approximately 65% of ideal body weight in the average patient over 65 years old. Drugs such as aminoglycosides, procainamide, digoxin, and coumadin (as in our patient) are distributed primarily in lean tissues and consequently have reduced distribution volumes and increased serum concentrations. Therefore, dosages often need to be adjusted downward to avoid toxicity. Other drugs, such as phenytoin, benzodiazepines, and barbiturates, have increased volumes of distribution in the elderly as a result of their higher proportion of adipose tissue. This effectively prolongs their duration of action, leading to potential side effects such as increased sedation.

Aging also decreases serum protein production, especially of albumin. In addition, malnutrition, which is more common in elderly patients (particularly those who are institutionalized), is an important cause of decreased serum albumin. Protein binding sites are diminished, and competitive inhibition for these sites can cause displacement of one drug by another, resulting in increased serum levels of the displaced drug. In our case presentation, for example, aspirin increases the unbound fraction of coumadin.

CHANGES IN DRUG CLEARANCE

Hepatic metabolism and renal elimination are the main factors in drug clearance. Aging is associated with decreased hepatic blood flow, which can alter the clearance of drugs that undergo rapid first-pass hepatic metabolism, such as beta blockers, calcium channel blockers, and narcotics, thereby increasing their potential toxicity. An age-related decline in functional hepatocyte number and enzyme activity affects the clearance of other drugs, such as phenytoin. Also, aging affects the nonsynthetic hepatic biotransformation reactions (oxidative and hydrolytic) more readily than synthetic enzymatic reactions, such as conjugation. Therefore, the diazepam used by our patient is more likely to cause increased sedation because it undergoes oxidative metabolism, unlike lorazepam, which is conjugated by the liver.

An age-related decline in renal blood flow and functional renal mass diminishes the glomerular filtration rate and creatinine clearance, thus affecting subsequent drug elimination by the kidneys. Drugs with narrow therapeutic windows, such as digoxin and the aminoglycosides, need to be administered with diminished renal function in mind in order to avoid potential toxicity with deleterious results. Yet, because creatinine production decreases with a decline in lean body mass that parallels the reduction in creatinine clearance associated with aging, the serum creatinine level should not be used as a marker of renal function in the elderly. The Cockroft-Gault formula (creatinine clearance = [(140 – age) x weight(kg)]/[72 x creatinine(mg/dl) x 0.85 for women]) is recommended because of the correction for age and weight.

Responses to certain medications are altered with aging, resulting in potentially significant consequences. The elderly appear to have an age-related decrease in both receptors and post-receptor function for cardiac medications, including calcium channel blockers and parasympathetic agonists and antagonists. These changes may be the basis for the altered baroreflex response to hypotension often seen in elderly patients. Falls in the elderly may be a significant sequela of these changes. Older patients also appear to be hypersensitive to central nervous system-mediated medications, such as benzodiazepines, antidepressants, and antihistamines. Consequently, these drugs can cause frequent and often severe adverse effects, such as delirium, agitation, somnolence, depression, and even worsening dementia.

The patient in the case presented here may have been more likely to fall as a result of her cognitive reaction to the diazepam and the antihistamine-containing cough remedy, in conjunction with a blunted cardiac response to the relative hypotension caused by the gastrointestinal bleed. However, symptoms leading up to a fall are often not recognized by physicians as potentially due to medication side effects or synergies.

RISK FACTORS

There are many factors that place the elderly at increased risk for medication-related illness and complications. Because elderly patients often see different physicians for multiple medical problems and take many medications, sometimes with drug interactions or synergies, a dangerous cycle that has been described as the “prescribing cascade” can develop. When an adverse drug reaction is misinterpreted as a new medical condition requiring treatment, the original problem is only confounded or exacerbated by the additional medication and the cycle is perpetuated. The end result is frequently a trip to the emergency department.

Several studies have examined the risk factors for ADRs in elderly outpatients: namely, polypharmacy (taking more than five medications), more than two chronic medical problems, prior history of an ADR, dementia, renal insufficiency (creatinine clearance below 50 ml/min), advanced age (more than 85 years), and multiple prescribers. In addition, some of the causes of noncompliance, such as problems with memory, misperception of the need for medications, fear of medication toxicity, prolonged length of therapy, labeling confusion, complex scheduling, poor eyesight, cost, use of multiple pharmacies, living alone, and lack of assistance or monitoring in medication administration, may be considered risk factors for ADRs. The interaction of prescribed medications with OTC drugs and nonconventional medications cannot be underestimated, especially given their increased utilization by the elderly. In this case, the patient’s use of ibuprofen, an antacid, and a cough syrup, in conjunction with her prescribed medications may have contributed to her gastrointestinal bleed and fall.

INAPPROPRIATE MEDICATIONS

An expert panel of thought leaders in geriatrics, clinical pharmacology, and psychopharmacology developed explicit consensus criteria for determining potentially inappropriate medications in the elderly (see table below). The original criteria developed by Beers in 1997 were targeted to the frail nursing home patient. The updated Beers Criteria are applicable to the general elderly population and should be consulted when dealing with polypharmacy in the elderly. For example, analgesics (including NSAIDs), certain anxiolytic/sedative/hypnotics (such as diazepam and other long-acting benzondiazepines), and first-generation antihistamines (such as diphenhydramine, because of its anticholinergic properties) have all been deemed inappropriate for the elderly.

However, it is important to keep in mind the heterogeneity of the elderly (the concept of chronologic versus physiologic age) when prescribing medications. Medication choices must be patient- and context-driven, an often difficult balancing act between benefit and risk.

CARE MODEL

The geriatric emergency care model is particularly relevant to pharmacologic issues in the elderly. Instead of prescribing medications for each disease symptom, a more holistic and pragmatic approach should be taken that balances therapeutic interventions with realistic expectations of outcome. In essence, a risk-benefit analysis that incorporates the patient’s functional status, comorbidities, and quality of life, as well as the potential for increased risk of deleterious adverse drug reactions, should be conducted when caring for the elderly patient.

Clear therapeutic endpoints need to be established at the beginning of treatment. An awareness of the age-related changes that affect functional reserve capacity and drug metabolism must be maintained. Medications should be started at low doses and increased slowly in small increments. Be mindful of the signs and symptoms of ADRs and toxicity and always maintain a high index of suspicion for drug-related events as the possible etiology of clinical presentations of illness or trauma. Effective patient communication and education concerning medications can prevent significant morbidity and even mortality in elderly patients.

back to top

Suggested Reading Beers MH: Explicit criteria for determining potentially inappropriate medication use by the elderly. An update. Arch Intern Med 157(14):1531, 1997. Blanda MP: Pharmacologic issues in geriatric emergency medicine. Emerg Med Clin North Am 24(2):449, 2006. Department of Health and Human Services, Administration on Aging (AoA). A Profile of Older Americans: 2003. Web site. Available at http://www.aoa.gov/prof/statistics/profile/2003/2003profile.pdf. Accessed on May 21, 2007. Evans R, et al.: Pharmacology and aging. In Sanders AB (ed): Emergency Care of the Elder Person, Beverly Cracom Publications, 1996. Gurwitz JH, et al.: Incidence and preventability of adverse drug events in nursing homes. Am J Med 109(2):87, 2000. Meldon SW et al.: Geriatric Emergency Medicine, 1st ed, McGraw-Hill Professional, 2004.

Printable version of this article

E-mail this article