Managing polypharmacy: Walking the fine line between help and harm.  Feb. 2002

Steven F. Werder, DO
Assistant professor
Departments of psychiatry and internal medicine
Sheldon H. Preskorn, MD
Professor and chairman
Department of psychiatry and behavioral sciences
University of Kansas School of Medicine
Wichita

Drug combinations often represent ‘uncontrolled experiments,’ with unknown potential for toxic effects. Yet, combination therapies often are used in managing psychiatric disorders. These authors provide practical tools for prescribing multiple drugs as safely as possible.
“Do no harm” is the first rule of medicine, yet 106,000 Americans die each year from properly prescribed and correctly taken medications.1 In some cases, the cause of death is known and can be attributed to a drug-drug interaction. The likelihood of death or hospitalization is directly proportional to the number of medications a patient is taking, even after controlling for underlying diseases.2
In psychiatry, it is not unusual for us to prescribe more than one psychotropic agent to manage a patient’s symptoms:
Patients with affective and psychotic disorders are commonly prescribed combinations of antipsychotics, mood stabilizers, antidepressants (often from more than one class), anxiolytics, antihistamines, and anticholinergics.
Patients with posttraumatic stress disorder may take selective serotonin reuptake inhibitors, buspirone, trazodone, antipsychotics, mood stabilizers, benzodiazepines, beta blockers, and opiates.
Multiple-drug regimens are used in treating other medical and psychiatric disorders, including chronic pain, fibromyalgia, chronic fatigue syndrome, sleep disorders, and epilepsy.
The greater the number of drugs used, the greater the likelihood that adverse events are emerging and are being treated, sometimes while being mistaken for patient psychopathology. As a prescriber, you are in a unique position to recognize and prevent interactions that can occur when patients are treated with two or more medications. This article defines polypharmacy, describes its consequences, prevalence, and risk factors, and offers an eight-step strategy with two mnemonics to help you avoid adverse events when prescribing multiple-drug regimens.
What is polypharmacy?
Many definitions have been used to describe polypharmacy (Box 1).3-6 The most common definition is the use of five or more drugs at the same time in the same patient.7 Although polypharmacy often has a pejorative connotation, using five or more drugs may be therapeutic or contratherapeutic.
Therapeutic polypharmacy occurs, for example, when expert panels or researchers in carefully controlled clinical trials recommend using multiple medications to treat specific diseases. For example, the five-drug combination of isoniazid, rifampin, ethambutol, pyrazinamide, and pyridoxine is therapeutic in initial tuberculosis treatment. More is better in this case because four antibiotics are needed to prevent the development of multiple drug-resistant Mycobacterium tuberculosis, and adding pyridoxine prevents isoniazid-induced neurotoxicity. This example illustrates two prescribing principles:
using multiple drugs can help achieve an intended therapeutic goal
adding one drug can prevent a known side effect of another drug.
Another example is the therapeutic management of congestive heart failure, in which five drug classes—an angiotensin-converting enzyme (ACE) inhibitor, a diuretic, a digitalis glycoside, a beta blocker, and an aldosterone antagonist—are used in various combinations. All play a role in improving cardiac function and reducing morbidity and mortality.
Using combination drug therapy can also generate cost benefits, such as by adding a drug to delay or inhibit the metabolism of an expensive principal drug. For example, adding diltiazem—a cytochrome P450 (CYP) 3A4 inhibitor—to cyclosporine—which is metabolized by CYP 3A4 enzymes—reduces the dosage of cyclosporine needed to achieve a desired serum level, thereby reducing the cost of this drug. (Some have abandoned this strategy because of cyclosporine’s narrow therapeutic index.)
Contratherapeutic polypharmacy occurs when a patient taking multiple drugs experiences an unexpected or unintended adverse outcome.
Settings for polypharmacy
Polypharmacy occurs in five principal prescribing situations:
treatment of symptoms
treatment of multiple illnesses
treatment of phasic illnesses, such as many affective, anxiety, seizure, and neurodegenerative disorders
preventing or treating adverse effects of other drugs
attempting to accelerate the onset of action or augment the effects of a preceding drug.
As described above, diseases such as tuberculosis and congestive heart failure, with well-understood causes and pathophysiologies, are often treated with multiple therapeutic drug combinations. However, the causes of many psychiatric disorders and syndromes are less well-understood, which makes prescribing drug combinations more difficult. It may be that treating less well-understood diseases is a risk factor for contratherapeutic polypharmacy.
Most individuals who are prescribed five or more drugs are taking unique drug combinations.8 These heterogeneous regimens represent “an uncontrolled experiment,” with effects that cannot be predicted from studies in the literature.9 Tables 1, 2, and 3 describe how contratherapeutic polypharmacy may occur with combinations of any number of drugs, whether five or more by the classic definition or only two. For example, contratherapeutic polypharmacy may occur when a patient is given the mood-stabilizing drugs valproate and carbamazepine (CBZ) at the same time.10 Here is why this combination may be dangerous:
Carbamazepine is oxidized by arene oxidase to CBZ 10,11-epoxide, which is hydrolyzed by epoxide hydrolase to CBZ 10,11-dihydroxide. The metabolite CBZ 10,11-epoxide has both therapeutic and toxic effects.
In monotherapy, the ratio of carbamazepine to CBZ 10,11-epoxide is 10:1, with CBZ 10,11-epoxide having a shorter half-life than carbamazepine.
However, when carbamazepine and valproate are taken as co-pharmacy, valproate blocks the hydrolysis of CBZ 10,11-epoxide by inhibiting epoxide hydrolase, so that the ratio of carbamazepine to CBZ 10,11-epoxide becomes 2:1. Higher concentrations of the epoxide metabolite contribute to neurotoxicity.
Other examples of potentially dangerous drug combinations include those associated with torsades de pointes, which may occur with certain combinations of antihistamines, antidepressants, antipsychotics, antivirals, antibacterials, antifungals, antiarrhythmics, and promotility agents.
Drug-drug interactions
In a drug-drug interaction, the presence of one drug alters the nature, magnitude, or duration of the effect of a given dose of another drug; the interaction may be either therapeutic or adverse, depending on the desired effect. A drug-drug interaction may be intended or unintended and is determined by pharmacokinetics and pharmacodynamics rather than by therapeutic class.
Most available drug information describes the effects of individual drugs used alone (monopharmacy). Information on how one drug interacts with another (co-pharmacy) is more difficult to come by. A recent literature search using broad criteria for drug-drug interactions uncovered 4,277 indexed articles. Another search, this time using narrow criteria, produced only 316 articles, suggesting that systematic studies regarding drug-drug interactions are few.
However, if you understand the pharmacodynamics and pharmacokinetics that rule co-pharmacy, then you can apply this knowledge to more complex drug-drug interactions involving contratherapeutic polypharmacy.
How drug effects are determined. The nature and magnitude of a drug’s effect are determined by its site of action and its binding affinity, concentration, and action at that site.11 This relationship can be represented by the formula:
effect = potency at the site of action
X concentration at the site of action
Potency at the site of action is determined by the binding affinity for the drug and the degree to which the receptor is stimulated or blocked, thus activating or inhibiting transmembrane and intracellular messengers (pharmacodynamics). Concentration at the site of action is determined by absorption, metabolism, distribution, and elimination (pharmacokinetics). Thus, the above model can be represented mathematically by:
effect = pharmacodynamics X pharmacokinetics
These factors determine a drug’s usual effect in the usual patient on the usual dosage, which is the goal of most clinical trials. However, all patients are not “usual,” because of inter-individual differences due to genetics, gender, age, environment, social habits such as smoking, intercurrent diseases affecting organ function, and concomitant drug therapy. Thus, when we take these factors into account, the first mathematical equation becomes:
effect = potency at the site of action X concentration
at site of action X inter-individual variance
In other words, the clinical response equals the drug’s potency at the site of action times the drug’s concentration at the site of action times the patient’s underlying biology. Likewise, when we consider variability among patients, the second equation becomes:
effect = pharmacodynamics X pharmacokinetics
X inter-individual variance
This addition to the equation explains how inter-individual variability can shift the dose-response curve to produce a greater or lesser effect than that which would be expected in the “usual” patient taking the prescribed dosage.
Inter-individual variance. The metabolism of dextromethorphan illustrates the effect of inter-individual variance. After a single dose, about 93% of Caucasians develop relatively lower dextromethorphan:dextrophan ratios, and about 7% develop relatively higher ratios. This difference defines patients who are pharmacogenetically CYP 2D6 extensive metabolizers versus those who are not.
Similarly, drugs sometimes cause biological variance, which predisposes to a drug-drug interaction. For example, the literature is replete with case reports and case series reporting that a substantial CYP 2D6 inhibitor—such as fluoxetine—blocks the metabolism of drugs that are principally metabolized by CYP 2D6. If the drug being metabolized has a narrow therapeutic index—such as amitriptyline—the resultant increase in its serum level can cause serious cardio- and neurotoxicity, including arrhythmias, delirium, seizures, coma, and death.12
In such cases, a CYP 2D6 inhibitor converts the phenotype from a CYP 2D6 extensive metabolizer into a CYP 2D6 poor metabolizer. Hence, the clinician must consider how a specific patient may differ from the usual patient when selecting and dosing a drug. The difference may be genetic or acquired, as in this example.
The following equation explains how dose is related to drug concentration, which takes into account the drug’s pharmacokinetics:
drug concentration = dosing rate (mg/day)
÷ clearance (ml/min)
In other words, the concentration achieved in a specific patient is determined by the dosage relative to the patient’s ability to clear the drug from the body.
Consequences, prevalence of polypharmacy
Polypharmacy increases patients’ risk for many ill effects, including incidence and severity of adverse events, drug-drug interactions, medication errors, hospitalizations, morbidity, mortality, and direct and indirect costs. At least 12 reports and studies have been published showing the association between polypharmacy and death,2,13-23 and in some of these reports the association is present even after controlling for underlying diseases.
The prevalence of polypharmacy varies by country and population. In Denmark, for example, the prevalence of polypharmacy is approximately 1.2%,6 compared with approximately 7% in the United States.24 Nearly one-half (46%) of all elderly persons admitted to U.S. hospitals may be taking seven or more medications.25 Polypharmacy is especially problematic in patients age 65 and older (Box 2),26-31 in whom the top five preventable threats to health are congestive heart failure, breast cancer, hypertension, pneumonia, and adverse drug events.32 Although older persons make up less than 15% of the population, they take the greatest number and quantity of medications, purchase 40% of all nonprescription medications, and use 33% of all retail prescriptions.30
Psychiatric disorders including schizophrenia, bipolar disorder, depression, personality disorders, and substance abuse place patients at higher risk for polypharmacy, as do certain demographic, psychosocial, medication, medical, and neurologic factors (Table 4). Other factors that increase the risk for polypharmacy include:
institutional factors (recent hospitalization, admission to a surgical ward, nursing home placement, home health care, increased number of pharmacies used, increased number of clinics attended, client-centered psychiatric treatment compared with non-client-centered psychiatric treatment)
provider factors (visit to a physician, treatment by general practitioners compared with specialists, increased number of providers, undocumented rationale or diagnosis supporting multiple medication use)
having medical insurance.
Steps to avoiding polypharmacy
By identifying polypharmacy’s risk factors, we may decrease its associated morbidity, mortality, and cost. Steps to follow while prescribing—as represented by the mnemonics SAIL33 and TIDE—may help you avoid polypharmacy’s negative consequences.
SAIL. Keep the drug regimen as simple as possible. Aim for once-daily or twice daily dosing. Try to simplify complex drug regimens by discontinuing any drug that does not achieve its defined therapeutic goal. For diseases and syndromes with less clear-cut causes, subtracting drugs from a complicated regimen may be more therapeutic than adding another drug. Try to treat multiple symptoms and syndromes with a single drug that may have multiple beneficial effects, rather than treating each symptom or syndrome with individual drugs.
Understand the potential adverse effects of each drug and potential drug-drug interactions. Whenever practical, choose drugs with broad rather than narrow therapeutic indices.
Each prescribed drug should have a clear indication and a well-defined therapeutic goal. Prescribe using evidence-based medicine as much as is practical.
List the name and dosage of each drug in the patient’s chart, and provide this information to the patient.33 Consider adopting computer data entry and feedback procedures, which have been shown to decrease polypharmacy34 and drug-drug interactions.35
TIDE. In the busy medical practice, writing a prescription signals to the patient that his or her time with the doctor is almost finished. Allow time to address medication issues.
Apply the understanding of individual variability, pharmacokinetics, and pharmacodynamics when prescribing. Review with the patient all prescription and nonprescription drugs and dietary supplements being taken.
Be careful to avoid potentially dangerous drug-drug interactions, especially those associated with serious adverse events such as torsades de pointes.
Educate patients regarding drug and non-drug treatments. Explain potential adverse effects of each drug and potential drug-drug interactions.

References
1. Lazarou J, Pomeranz BH, Corey PN. Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies. JAMA 1998;279(15):1200-5.
2. Incalzi RA, Gemma A, Capparella O, et al. Predicting mortality and length of stay of geriatric patients in an acute care general hospital. J Gerontol 1992;47(2):M35-9.
3. Berube MS, Neely DJ, DeVinne, PB. American Heritage Dictionary (2nd College ed). Boston: Houghton Mifflin Co, 1982.
4. Friend DG. Polypharmacy: multiple-ingredient and shotgun prescriptions. N Engl J Med 1959;260(20):1015-8.
5. Sheppard C, Collins L, Fiorentino D, Fracchia J, Merlis S. Polypharmacy in psychiatric treatment. I. Incidence at a state hospital. Curr Ther Res Clin Exp 1969;(12):765-74.
6. Bjerrum L, Rosholm JU, Hallas J, Kragstrup J. Methods for estimating the occurrence of polypharmacy by means of a prescription database. Eur J Clin Pharmacol 1997;53(1):7-11.
7. Werder SF. Polypharmacy: definitions and risk factors (grand rounds). University of Kansas School of Medicine-Wichita, Department of Psychiatry and Behavioral Sciences. Via Christi Regional Medical Center, St. Joseph Campus: Dec 12, 2000.
8. Bjerrum L, Sogaard J, Hallas J, Kragstrup J. Polypharmacy: correlations with sex, age and drug regimen. A prescription database study. Eur J Clin Pharmacol 1998;54(3):197-202.
9. Bjerrum L. Pharmacoepidemiological Studies of Polypharmacy: Methodological issues, population estimates, and influence of practice patterns (PhD thesis). Odense University Faculty of Health Sciences, Department of clinical pharmacology and research unit of general practice. Denmark; 1998. Available at http://www.sdu.dk/ health/IPH/genpract/staff/ lbjerrum/PHD/PHD.HTM. Accessed Jan. 9, 2003.
10. Wilder BJ. Pharmacokinetics of valproate and carbamazepine. J Clin Psychopharmacol 1992;12(1 suppl):64S-68S.
11. Preskorn SH. The rational basis for the development and use of newer antidepressants. In: Outpatient management of depression: a guide for the practitioner (2nd ed). Caddo, OK: Professional Publications, Inc; 1999:57-103.
12. Vaughan DA. Interaction of fluoxetine with tricyclic antidepressants. Am J Psychiatry 1988;145(11):1478.
13. Meeker JE, Reynolds PC. Postmortem tissue methamphetamine concentrations following selegiline administration. J Anal Toxicol 1990;14(5):330-1.
14. Sallee FR, DeVane CL, Ferrell RE. Fluoxetine-related death in a child with cytochrome P-450 2D6 genetic deficiency. J Child Adolesc Psychopharmacol 2000;10(1):27-34.
15. Ellis RJ, Mayo MS, Bodensteiner DM. Ciprofloxacin-warfarin coagulopathy: a case series. Am J Hematol 2000;63(1):28-31.
16. Konig SA, Siemes H, Blaker F, et al. Severe hepatotoxicity during valproate therapy: an update and report of eight new fatalities. Epilepsia 1994;35(5):1005-15.
17. Fattinger K, Roos M, Vergeres P, et al. Epidemiology of drug exposure and adverse drug reactions in two Swiss departments of internal medicine. Br J Clin Pharmacol 2000;49(2):158-67.
18. Ebbesen J, Buajordet I, Erikssen J, Svaar H, Brors O, Hilberg T. Drugs as a cause
of death. A prospective quality assurance project in a department of medicine (Norwegian). Tidsskr Nor Laegeforen 1995;115(19):2369-72.
19. Alarcon T, Barcena A, Gonzalez-Montalvo JI, Penalosa C, Salgado A. Factors predictive of outcome on admission to an acute geriatric ward. Age Ageing 1999;28(5):429-32.
20. Smith NK, Albazzaz MK. A prospective study of urinary retention and risk of death after proximal femoral fracture. Age Ageing 1996;25(2):150-4.
21. Pulska T, Pahkala K, Laippala P, Kivela SL. Six-year survival of depressed elderly Finns: a community study. Int J Geriatr Psychiatry 1997;12(9):942-50.
22. Waddington JL, Youssef HA, Kinsella A. Mortality in schizophrenia. Antipsychotic