# Polypharmacy and Deprescribing

## Introduction

Polypharmacy, commonly defined as the concurrent use of five or more medications, is one of the most prevalent and consequential clinical challenges in geriatric medicine. Hyperpolypharmacy, defined as the use of ten or more medications, represents the extreme end of the spectrum. The prevalence of polypharmacy is striking: 30 to 40 percent of community-dwelling elderly adults take five or more medications, and 50 to 70 percent of nursing home residents meet the threshold. Among adults aged 65 and older with multimorbidity, the average medication burden is eight to ten drugs.

Polypharmacy is independently associated with adverse drug reactions, falls, cognitive impairment, functional decline, hospitalization, and mortality. Adverse drug reactions account for 5 to 10 percent of hospital admissions among elderly patients, and 30 to 50 percent of these events are preventable. Deprescribing, defined as the systematic process of identifying and discontinuing medications that are causing or may cause harm, or that are no longer providing benefit, is a core competency in geriatric medicine and represents one of the most impactful interventions a clinician can make for the health of an older patient.

## Consequences of Polypharmacy

### Adverse Drug Reactions

The incidence of adverse drug reactions increases exponentially with the number of medications: approximately 13 percent with two medications versus 82 percent with seven or more. Drug-drug interactions follow a similar exponential curve, with the probability of a clinically significant interaction rising from 6 percent with two drugs to 50 percent with five drugs to essentially 100 percent with eight or more drugs. Drug-disease interactions add a further layer of risk, with common examples including beta-blockers masking hypoglycemia symptoms in patients with diabetes, anticholinergic medications worsening cognitive function in patients with dementia, and non-steroidal anti-inflammatory drugs exacerbating heart failure.

Prescribing cascades represent a particularly insidious consequence of polypharmacy in which an adverse drug reaction is misidentified as a new medical condition, leading to the prescription of an additional medication, which may in turn produce further adverse effects. Classic examples include NSAID-induced hypertension leading to prescription of an antihypertensive, cholinesterase inhibitor-induced urinary incontinence leading to prescription of oxybutynin (an anticholinergic that then worsens the cognition the cholinesterase inhibitor was meant to preserve), and calcium channel blocker-induced peripheral edema leading to prescription of a diuretic.

### Medication-Related Geriatric Syndromes

Polypharmacy drives several of the core geriatric syndromes. Falls are potentiated by sedatives, antihypertensives, psychotropic medications, and opioids, with each CNS-active medication increasing fall risk by a factor of 1.5 to 2. Cognitive impairment results from exposure to anticholinergic medications, benzodiazepines, opioids, and antihistamines. Fox and colleagues demonstrated in 2011 that a cumulative anticholinergic burden score of 3 or greater was associated with measurable cognitive decline over two years. Delirium is precipitated by anticholinergics, benzodiazepines, opioids, corticosteroids, and fluoroquinolones. Urinary incontinence can result from diuretics, alpha-blockers, cholinesterase inhibitors, and sedatives. Orthostatic hypotension is caused by antihypertensives, alpha-blockers, tricyclic antidepressants, antipsychotics, and antiparkinsonian medications.

### Inappropriate Medication Use

Forty to 50 percent of community-dwelling elderly adults take at least one potentially inappropriate medication, and an estimated 20 percent of Medicare spending on medications goes toward drugs that are potentially unnecessary or harmful. This represents an enormous opportunity for harm reduction and cost savings through systematic medication review.

## Screening Tools for Potentially Inappropriate Medications

### Beers Criteria (AGS, updated 2023)

The Beers Criteria, developed and maintained by the American Geriatrics Society, constitute an expert consensus list of potentially inappropriate medications for older adults, organized into five categories. The first category encompasses medications to avoid in most older adults, including first-generation antihistamines, benzodiazepines, non-COX-2 selective NSAIDs used long-term, and sliding-scale insulin regimens. The second category identifies medications to avoid in the presence of specific diseases or syndromes. The third category lists medications to use with caution, such as aspirin for primary cardiovascular prevention and dabigatran in patients with creatinine clearance below 30 mL/min. The fourth category identifies drug-drug interactions to avoid, such as the concurrent use of three or more CNS-active drugs or the combination of warfarin with NSAIDs. The fifth category addresses medications requiring dose adjustment based on renal function.

Key Beers Criteria medications that every geriatrician should know include benzodiazepines (increased fall risk, cognitive impairment, delirium), long-acting sulfonylureas such as glyburide (prolonged hypoglycemia), first-generation antihistamines such as diphenhydramine (strongly anticholinergic), proton pump inhibitors used beyond eight weeks without a clear ongoing indication (associated with Clostridioides difficile infection, fractures, and hypomagnesemia), meperidine (neurotoxic metabolite normeperidine, seizure risk), and sliding-scale insulin used as the sole insulin regimen (reactive dosing with increased hypoglycemia risk).

### STOPP/START Criteria (O'Mahony et al., updated 2023)

The STOPP/START criteria, developed in Europe, provide a complementary approach organized by physiological system. The STOPP component (Screening Tool of Older Persons' Prescriptions) identifies potentially inappropriate medications, while the START component (Screening Tool to Alert to Right Treatment) identifies potential prescribing omissions where beneficial medications are not being provided. This dual focus on both overuse and underuse represents a significant advantage over the Beers Criteria, which focus primarily on overuse. Examples of START-identified omissions include failure to prescribe a statin after myocardial infarction or stroke in patients with reasonable life expectancy, failure to prescribe vitamin D in patients with osteoporosis, and failure to prescribe an ACE inhibitor or ARB in heart failure with reduced ejection fraction.

### Medication Appropriateness Index (MAI)

The Medication Appropriateness Index is a ten-item implicit assessment applied to each individual medication, evaluating indication, effectiveness, dosage, directions, drug interactions, duration, duplication, cost, practicality, and whether untreated conditions are present. Although more time-intensive than explicit criteria, the MAI provides a thorough evaluation that is particularly valuable for complex cases.

<image>A side-by-side comparison infographic of the Beers Criteria and STOPP/START Criteria. On the left, show the Beers Criteria organized by its five categories with 2-3 example medications in each category and their associated risks. On the right, show the STOPP criteria organized by physiologic system (cardiovascular, CNS, GI, endocrine, musculoskeletal) with example medications, and the START criteria with example omissions that should be prescribed. In the center, include a Venn diagram showing overlap and unique contributions of each tool. At the bottom, include a comparison table with rows for: origin, update frequency, organization method, strengths, limitations, and recommended use setting. Highlight that STOPP/START identifies both overuse and underuse while Beers focuses primarily on overuse.</image>

## The Deprescribing Process

### Definition and Principles

Deprescribing is the planned and supervised process of dose reduction or discontinuation of medications that are causing or may cause harm, or that are no longer providing benefit within the context of a patient's current clinical status, goals, and remaining life expectancy. It is fundamentally distinct from underprescribing; rather, it represents the optimization of the medication regimen to align with the individual patient's needs and priorities. Shared decision-making is essential throughout the process, as patients and caregivers must be engaged as partners in understanding the rationale for medication changes and in monitoring for effects of those changes.

### Systematic Deprescribing Framework (Scott et al., 2015)

The deprescribing process follows a systematic five-step framework. The first step is to ascertain all medications the patient is currently taking, which requires reconciliation across all sources including pharmacy records, brown bag reviews (in which the patient brings all medication containers to the visit), caregiver input, over-the-counter medications, supplements, and as-needed medications. The second step is to identify potentially inappropriate medications using the Beers Criteria, STOPP/START criteria, clinical judgment, and consideration of patient goals. The third step is to determine whether each identified medication can be ceased, taking into account its current indication, ongoing relevance, the benefit-harm balance in the current clinical context, and the patient's remaining life expectancy. The fourth step is to plan the withdrawal regimen, including tapering schedules for medications that require gradual discontinuation and a monitoring plan for each change. The fifth step is to monitor and support the patient, with follow-up within four to eight weeks to assess for withdrawal symptoms or return of the original condition.

### Priority Medications for Deprescribing

#### Proton Pump Inhibitors

Proton pump inhibitors are among the highest-yield targets for deprescribing, as 40 to 65 percent of PPI prescriptions lack a clear ongoing indication. The D-PPI trial demonstrated that step-down or discontinuation was successful in 66 percent of patients without symptom relapse. The recommended approach is to reduce the dose by 50 percent for two to four weeks, then switch to an H2 receptor antagonist or as-needed PPI use, then discontinue entirely. Patients should be warned about rebound acid hypersecretion, which peaks at approximately two weeks and resolves within four to eight weeks. PPIs should be maintained in patients with Barrett esophagus, severe erosive esophagitis (Los Angeles grade C or D), or ongoing NSAID or antiplatelet therapy with high gastrointestinal bleeding risk.

#### Benzodiazepines and Z-drugs

Deprescribing of benzodiazepines is feasible in more than 60 percent of elderly users when attempted with a gradual taper. The recommended approach is to reduce the dose by 10 to 25 percent every two to four weeks, with conversion from a short-acting agent to a long-acting agent such as diazepam if needed to facilitate a smoother taper. Total taper duration is typically six to twelve weeks, with longer tapers for individuals with prolonged use histories. Adjunct strategies during the taper include cognitive behavioral therapy for insomnia (CBT-I) and melatonin 2 mg in a slow-release formulation. The EMPOWER cluster randomized controlled trial demonstrated that patient education through a brochure alone achieved 27 percent discontinuation at six months, illustrating that informed patients are often willing and able to reduce benzodiazepine use. Withdrawal symptoms including rebound insomnia, anxiety, tremor, and rarely seizures (which are extremely uncommon with gradual tapering) should be anticipated and managed.

#### Antipsychotics in Dementia

Antipsychotic deprescribing should be attempted every three to six months per CMS guidelines. The DART-AD trial demonstrated that the majority of nursing home residents did not relapse upon discontinuation, while continued antipsychotic use was associated with a two-fold increase in mortality. The recommended approach is a gradual dose reduction of 25 to 50 percent every two weeks with close monitoring for behavioral recurrence.

#### Statins

Statin deprescribing should be considered in patients with limited life expectancy (less than one to two years), advanced dementia, significant functional decline, or primary prevention use in the very elderly (over 80 years). The ECSTATIC trial demonstrated that statin discontinuation in patients aged 70 and older with polypharmacy was non-inferior at two years. The STOPPFrail criteria recommend discontinuation for primary prevention in patients with limited life expectancy. Statins should be maintained in patients with recent acute coronary syndrome (within one year) or known significant atherosclerotic disease with reasonable remaining life expectancy.

#### Antihypertensives

Dose reduction of antihypertensive medications should be considered when systolic blood pressure is consistently below 120 mmHg in frail elderly patients (who are at increased risk of falls, syncope, and acute kidney injury), when symptomatic orthostatic hypotension is present, or in advanced dementia with limited life expectancy. The OPTIMISE trial demonstrated that antihypertensive reduction in patients aged 80 and older with polypharmacy was non-inferior for cardiovascular outcomes at twelve weeks. The approach should involve deprescribing one agent at a time, prioritizing reduction of agents most strongly associated with falls, such as alpha-blockers and centrally acting agents.

#### Cholinesterase Inhibitors

Discontinuation of cholinesterase inhibitors should be considered in advanced dementia (FAST stage 7, non-verbal, immobile) when care goals are comfort-focused. However, the DOMINO-AD trial demonstrated that discontinuation was associated with faster cognitive and functional decline, and this finding must be discussed with family members as part of the decision-making process. If discontinuation is trialed, the medication should be stopped for six to eight weeks with reassessment of cognitive and functional status, and restarted if a meaningful decline is observed.

<image>A step-by-step deprescribing algorithm flowchart. Start at the top with "Comprehensive Medication Review" showing a brown bag review icon. Step 1: "List all medications (prescribed, OTC, supplements, PRN)" with example count (e.g., 12 medications). Step 2: "Screen with Beers/STOPP criteria and assess anticholinergic burden (ACB scale)." Step 3: Decision diamond "Does each medication have a current valid indication?" — if No, "Consider discontinuation." Step 4: "For each indicated medication, assess benefit vs. harm in context of patient's goals, life expectancy, and functional status" with a balance scale graphic. Step 5: "Prioritize medications to deprescribe" — show a ranked list with highest priority items (benzodiazepines, PPIs without indication, anticholinergics, antipsychotics in dementia). Step 6: "Plan gradual taper with monitoring schedule" showing timeline. Step 7: "Follow-up at 4-8 weeks — assess for withdrawal, symptom recurrence, improved outcomes." Include a feedback loop arrow from Step 7 back to Step 1 for ongoing optimization.</image>

## Special Considerations

### Anticholinergic Burden

Cumulative anticholinergic exposure is a major driver of cognitive impairment, delirium, and falls in older adults. The Anticholinergic Cognitive Burden (ACB) scale assigns scores of 1 (possible anticholinergic activity), 2 (definite anticholinergic activity), or 3 (definite anticholinergic activity with established cognitive effects) to individual medications. Common high-burden medications scoring 3 on the ACB scale include oxybutynin, paroxetine, amitriptyline, diphenhydramine, hydroxyzine, chlorpheniramine, and tolterodine. The goal is to maintain a total ACB score below 3. Substitution with lower-anticholinergic alternatives is a practical deprescribing strategy: oxybutynin can be replaced with mirabegron or vibegron, paroxetine with sertraline, amitriptyline with duloxetine, and diphenhydramine with melatonin.

| High-ACB Medication (Score 3) | Indication | Lower-ACB Alternative | ACB Score of Alternative |
|-------------------------------|------------|----------------------|--------------------------|
| Oxybutynin | Overactive bladder | Mirabegron or vibegron | 0 |
| Paroxetine | Depression | Sertraline | 0 |
| Amitriptyline | Pain, depression | Duloxetine | 0 |
| Diphenhydramine | Insomnia, allergy | Melatonin (insomnia); loratadine (allergy) | 0 |
| Hydroxyzine | Anxiety, pruritus | Buspirone (anxiety); cetirizine (pruritus) | 0–1 |
| Tolterodine | Overactive bladder | Mirabegron or vibegron | 0 |
| Chlorpheniramine | Allergy | Loratadine or cetirizine | 0 |

### Deprescribing in End of Life

The STOPPFrail criteria are specifically designed for frail elderly patients with limited life expectancy. Medications often appropriate to discontinue in this context include statins, vitamins, bisphosphonates, memantine, and screening-related medications. Medications that should generally be continued include analgesics, antiemetics, anxiolytics used for symptom control, anticonvulsants for active seizure disorders, and any medication providing immediate symptom relief. The OncPal deprescribing guideline provides analogous recommendations for cancer patients receiving palliative care.

### Patient and Caregiver Communication

Effective communication is essential for successful deprescribing. Patient-friendly language should be used, such as "We want to make sure each medication is helping more than it could hurt." Clinicians should address the common fear of "giving up" by framing deprescribing as optimization and improvement of care rather than abandonment. Shared decision-making aids including the EMPOWER brochure and medication review worksheets support patient engagement in the process. Clear documentation of the rationale for each change in the medical record supports continuity of care.

## Key Clinical Pearls

- Every medication encounter should include the question: "Can anything be stopped?" — deprescribing is an ongoing process, not a one-time event
- Prescribing cascades (ADR → new diagnosis → new drug) are the most preventable cause of polypharmacy — always consider whether a new symptom is a drug side effect
- Anticholinergic burden (ACB score ≥3) is directly associated with cognitive decline, falls, and delirium — calculate and minimize it in every patient
- PPIs, benzodiazepines, antipsychotics in dementia, and statins for primary prevention are the highest-yield deprescribing targets
- Benzodiazepine deprescribing is feasible in >60% of elderly users — patient education alone (EMPOWER) achieves 27% discontinuation
- Never abruptly discontinue: benzodiazepines (seizures), beta-blockers (rebound tachycardia), corticosteroids (adrenal crisis), SSRIs (discontinuation syndrome), opioids (withdrawal), or clonidine (rebound hypertension)

| Medication Class | Taper Strategy | Duration | Risk of Abrupt Discontinuation | Key Evidence |
|-----------------|----------------|----------|-------------------------------|--------------|
| PPIs | 50% dose x 2–4 wk → H2RA or PRN → stop | 4–8 weeks | Rebound acid hypersecretion | D-PPI trial: 66% success |
| Benzodiazepines | 10–25% reduction q2–4 wk | 6–12 weeks | Rebound insomnia, anxiety, seizures | EMPOWER: 27% discontinuation with education alone |
| Antipsychotics (dementia) | 25–50% reduction q2 wk | 4–8 weeks | Behavioral relapse in ~30% | DART-AD: continued use doubles mortality |
| Statins (primary prevention) | Discontinue directly | Immediate | No withdrawal risk | ECSTATIC: non-inferior at 2 years |
| Antihypertensives | Remove one agent at a time | 2–4 weeks per agent | Rebound hypertension (clonidine) | OPTIMISE: non-inferior at 12 weeks |
| Cholinesterase inhibitors | Stop and observe 6–8 wk | 6–8 weeks | Cognitive/functional decline possible | DOMINO-AD: discontinuation accelerated decline |

- Deprescribing saves lives — DART-AD showed continued antipsychotic use in dementia doubles mortality

## References
1. American Geriatrics Society 2023 Updated AGS Beers Criteria for potentially inappropriate medication use in older adults. *J Am Geriatr Soc*. 2023;71(7):2052-2081.
2. O'Mahony D, Cherubini A, Guiteras AR, et al. STOPP/START criteria for potentially inappropriate prescribing in older people: version 3. *Eur Geriatr Med*. 2023;14:625-632.
3. Scott IA, Hilmer SN, Reeve E, et al. Reducing inappropriate polypharmacy: the process of deprescribing. *JAMA Intern Med*. 2015;175(5):827-834.
4. Martin P, Tamblyn R, Bhatt S, et al. A consumer-targeted, pharmacist-led, educational intervention (EMPOWER): a cluster randomized controlled trial. *JAMA Intern Med*. 2018;178(12):1611-1621.
5. Reeve E, Shakib S, Hendrix I, Roberts MS, Wiese MD. Review of deprescribing processes and development of an evidence-based, patient-centred deprescribing process. *Br J Clin Pharmacol*. 2014;78(4):738-747.