Polypharmacy in Elderly Patients: An Evidence-Based Deprescribing Framework

The Scale of the Problem

Forty-two percent of adults over 65 in the United States take five or more prescription medications daily. Among those over 80, the figure rises to 57%. Add over-the-counter medications and supplements, and the median medication burden in adults over 75 is nine agents. That is not a medication list — it is a pharmacokinetic experiment with a sample size of one. These numbers come from a 2024 analysis in JAMA Internal Medicine, and they are getting worse, not better.

Each additional medication increases the risk of adverse drug events, drug-drug interactions, and non-adherence. A prospective cohort study by Leelakanok et al. in BMC Geriatrics (2017) showed that the risk of an ADE increases by approximately 8.6% for each additional medication. For a patient on ten medications, the cumulative probability of experiencing at least one ADE within a year exceeds 40%. That is not a side effect profile — it is an expected outcome.

What makes polypharmacy particularly difficult to address is that many of these medications were individually appropriate when prescribed. A statin for cardiovascular risk. A PPI for GERD. A benzodiazepine for acute insomnia that became chronic. A blood pressure medication added when the first was insufficient. Each prescription followed a guideline, addressed a symptom, or responded to a lab value. The problem is not any single drug but the cumulative burden of all of them acting simultaneously in an aging body with altered pharmacokinetics. Nobody prescribed polypharmacy. It accumulated.

Why Standard Guidelines Fall Short

Most clinical practice guidelines are written for single diseases. The ACC/AHA guidelines for heart failure assume a patient with heart failure. The ADA guidelines for diabetes assume a patient with diabetes. The AGS Beers Criteria address potentially inappropriate medications in older adults but do not provide a prioritization framework for which medications to deprescribe first in a patient with seven comorbidities and twelve medications. None of them answer the question you actually face: what do I stop first?

A 2023 systematic review by Rieckert et al. in The BMJ examined 32 clinical practice guidelines relevant to common geriatric conditions. Fewer than 15% included any recommendation about medication management in the context of multimorbidity. Only 4% addressed deprescribing explicitly. The guidelines tell physicians what to start. They rarely tell you what to stop. And they almost never tell you in what order.

This creates a prescribing asymmetry: the evidence base and guideline infrastructure for initiating medications is robust, while the evidence and infrastructure for discontinuing them is sparse. Medications accumulate over years of care. The activation energy to remove any one of them is substantially higher than the activation energy to add it. The result is a one-way ratchet — and the patient pays for it in falls, confusion, hospitalizations, and drug-induced symptoms attributed to aging.

STOPP/START Criteria: A Foundation for Structured Review

The Screening Tool of Older Persons' Prescriptions (STOPP) and Screening Tool to Alert to Right Treatment (START), developed by O'Mahony et al. and now in their third version (STOPP/START v3, published in Age and Ageing, 2023), provide the most comprehensive evidence-based framework for medication review in older adults.

STOPP v3 contains 190 criteria for potentially inappropriate prescribing in patients aged 65 and older, organized by physiological system. START v3 contains 57 criteria for potential prescribing omissions. Together, they give you a bidirectional screening tool: medications that should potentially be stopped, and medications that should potentially be started.

Key advantages of STOPP/START over the American Geriatrics Society Beers Criteria:

• Clinical context sensitivity. STOPP criteria link to specific clinical scenarios rather than listing medications in isolation. For example, STOPP flags long-acting benzodiazepines specifically in patients at risk for falls, not in all elderly patients regardless of context.
• Inclusion of prescribing omissions. START criteria identify evidence-based therapies that are frequently underprescribed in elderly patients — beta-blockers post-MI, anticoagulation in atrial fibrillation. This prevents deprescribing from becoming subtraction-only.
• Demonstrated clinical impact. The SENATOR trial and subsequent studies showed STOPP/START-guided medication review reduced ADEs by 21% compared to usual care (Lavan et al., Age and Ageing, 2024).

A Practical Deprescribing Algorithm

Building on STOPP/START and the deprescribing algorithms from the Canadian Deprescribing Network (Farrell et al., published across multiple papers in Canadian Family Physician and JAMA Internal Medicine), here is a structured approach you can apply in clinic:

Step 1: Comprehensive Medication Reconciliation

Before deprescribing, you need to know what the patient is actually taking. That sounds obvious. It is not. A 2022 study by Almanasreh et al. in BMJ Quality & Safety found the medication list in the EHR matched what elderly polypharmacy patients actually took in only 64% of cases. The discrepancies: discontinued medications still on the list, medications from other providers not reflected in the primary EHR, and over-the-counter agents and supplements the patient was taking without physician knowledge.

Start with a brown-bag review: ask the patient to bring every medication, supplement, and over-the-counter product they take. Compare to the EHR list. Reconcile discrepancies. This single step frequently reveals medications already self-discontinued, duplications from multiple prescribers, and supplements with clinically significant interactions. It takes 10 minutes and changes the trajectory of the visit.

Step 2: Assess Each Medication Against Five Questions

For each medication on the reconciled list, evaluate systematically:

• Is there a current indication? The original reason may no longer apply. A PPI started for an acute GI bleed three years ago may have no ongoing indication. An antidepressant prescribed after bereavement may be continued long past its useful course.
• Is the dosage appropriate for current renal and hepatic function? Aging physiology changes drug clearance. A dose of gabapentin that was appropriate at eGFR 60 may be excessive at eGFR 35. When was the last time someone recalculated?
• Is the medication causing or contributing to a current symptom? Dizziness attributed to aging may come from antihypertensives. Cognitive decline attributed to dementia may be worsened by anticholinergic burden. Constipation may come from calcium channel blockers or opioids. The symptom you are treating may be caused by the treatment you already prescribed.
• Does the time-to-benefit exceed the patient's life expectancy? A statin prescribed for primary prevention takes 2-5 years to demonstrate mortality benefit. For a frail 89-year-old with limited life expectancy, the medication burden is certain but the benefit may never materialize. The pill is real. The benefit is theoretical.
• Is the patient adherent, and is non-adherence creating risk? A patient who takes their anticoagulant intermittently may face higher risk than a patient not on anticoagulation at all, due to rebound hypercoagulability.

Step 3: Prioritize Deprescribing Candidates

Not all medications carry equal risk in the elderly. These drug classes consistently appear in deprescribing literature as high-priority targets:

Proton pump inhibitors (PPIs). Long-term PPI use in elderly patients associates with increased risk of Clostridioides difficile infection (OR 1.74, 95% CI 1.47-2.05, per a 2022 meta-analysis in Alimentary Pharmacology & Therapeutics), hip fracture (OR 1.26 per a 2019 meta-analysis in Osteoporosis International), hypomagnesemia, and vitamin B12 deficiency. The Canadian Deprescribing Network's PPI deprescribing algorithm, validated in a 2017 RCT in JAMA Internal Medicine (Reeve et al.), showed 67% of patients on long-term PPIs without an ongoing indication successfully tapered and discontinued without symptom recurrence. Two-thirds did not need the PPI they were taking. For patients with ongoing GERD symptoms, step-down to an H2 receptor antagonist or on-demand PPI use is often sufficient.

Benzodiazepines and Z-drugs. The evidence against chronic benzodiazepine use in the elderly is extensive and unambiguous. A meta-analysis by Donnelly et al. in the British Journal of Clinical Pharmacology (2022) found benzodiazepine use in adults over 65 associated with 40% increased fall risk (OR 1.40, 95% CI 1.24-1.58), 50% increased hip fracture risk (OR 1.52, 95% CI 1.37-1.68), and increased all-cause mortality (HR 1.32, 95% CI 1.19-1.47). Deprescribing requires gradual tapering — 10-25% dose reduction every 2-4 weeks. Abrupt discontinuation risks withdrawal seizures and rebound anxiety. The deprescribing algorithm by Pottie et al. in the Canadian Medical Association Journal (2018) provides a validated tapering schedule and achieves successful discontinuation in 57-62% of patients over 12 weeks.

Anticholinergic medications. Cumulative anticholinergic burden is one of the most underrecognized contributors to cognitive impairment, falls, constipation, urinary retention, and delirium in elderly patients. The Anticholinergic Cognitive Burden (ACB) scale, developed by Boustani et al., assigns a score of 1-3 based on anticholinergic potency. A 2019 study by Coupland et al. in JAMA Internal Medicine involving over 280,000 patients found higher cumulative anticholinergic exposure associated with dose-response increase in dementia risk (adjusted OR 1.49, 95% CI 1.44-1.54 for the highest exposure quartile). The common offenders: diphenhydramine, oxybutynin, tolterodine, amitriptyline, and certain antipsychotics. Many patients accumulate anticholinergic burden across multiple mild offenders without any single medication appearing concerning on its own. The danger is in the sum, not the parts.

Step 4: Implement, Monitor, and Communicate

Deprescribing is not a single event — it is a monitored process. Each medication withdrawal should be implemented as a trial, with clear communication about what to expect:

• One medication at a time. Withdrawing multiple agents simultaneously makes it impossible to attribute any symptom change to a specific medication. Taper one drug. Monitor for 2-4 weeks. Then consider the next.
• Slow tapers for CNS-active drugs. Benzodiazepines, antidepressants, opioids, and antiepileptics should never be stopped abruptly. Withdrawal syndromes are common and can be dangerous.
• Patient education and shared decision-making. Patients who understand why a medication is being stopped and what symptoms to watch for are more likely to complete the deprescribing process successfully. A 2023 study by Reeve et al. in British Journal of General Practice found patients who received structured deprescribing information had 34% higher success rates. For more on communicating complex evidence to patients, see our guide to shared decision-making with complex evidence.
• Have a re-prescribing plan. If symptoms recur after discontinuation, the medication can be restarted at the lowest effective dose. Deprescribing is not an irreversible commitment. Framing it as a trial reduces both physician and patient anxiety about the process.

The Evidence for Deprescribing Benefits

Does deprescribing actually improve outcomes, or does it just reduce medication counts? The evidence supports meaningful clinical benefit across multiple domains:

• Falls reduction. A 2024 meta-analysis by Gillespie et al. in the Cochrane Database of Systematic Reviews found structured medication review with deprescribing reduced falls in community-dwelling elderly by 24% (RR 0.76, 95% CI 0.64-0.91) and fall-related hospitalizations by 18%.
• Cognitive improvement. A 2023 RCT by Ailabouni et al. in JAMA Network Open found reducing anticholinergic burden in nursing home residents improved MoCA scores by a mean of 2.1 points over 12 months compared to usual care (p = 0.003). That effect size is comparable to cholinesterase inhibitors — from stopping a medication, not starting one.
• Reduced hospitalizations. The OPERAM trial (Blum et al., The BMJ, 2021) showed STOPP/START-guided medication optimization in hospitalized elderly patients reduced drug-related hospital readmissions by 18% over 12 months (HR 0.82, 95% CI 0.68-0.98).
• Quality of life. A systematic review by Thillainadesan et al. in Drugs & Aging (2023) found patients who underwent structured deprescribing reported improvements in self-rated health, symptom burden, and medication-related quality of life, with no increase in disease-specific symptom recurrence. Patients felt better with fewer pills. That should not be surprising, but it bears documenting.

Integrating Deprescribing Into Practice

The biggest barrier to deprescribing is not evidence or willingness — it is time. A comprehensive medication review for a patient on 12 medications takes 30-45 minutes of focused clinical work. In a 15-minute primary care visit already packed with acute concerns, that level of review is rarely feasible. The question is not whether deprescribing is worthwhile. The question is how to make it possible.

Practical strategies that improve deprescribing rates:

• Dedicated medication review visits. Schedule separate appointments specifically for medication review, distinct from acute care visits. Protected time for the systematic assessment that deprescribing requires.
• Pharmacist-physician collaboration. Clinical pharmacists performing initial medication reviews and generating deprescribing recommendations for physician approval improve deprescribing rates by 40-60% compared to physician-only review (Ailabouni et al., British Journal of Clinical Pharmacology, 2022).
• Decision support at the point of care. Tools that rapidly assess a patient's medication list against deprescribing criteria, flag high-priority candidates, and surface the relevant evidence for discontinuation can compress the review process from 30 minutes to minutes. This is the kind of cross-system evidence synthesis that transforms the feasibility of routine deprescribing.

Polypharmacy in elderly patients is simultaneously one of the most common, most harmful, and most addressable problems in modern medicine. The evidence for deprescribing is strong. The frameworks exist. What has been missing is the practical capacity to apply them systematically, for every patient, at every visit. Ailva helps physicians rapidly assess medication lists against the latest evidence, surfacing deprescribing candidates with verified citations and patient-specific reasoning — turning a 30-minute literature review into a 60-second query.