Polypharmacy at the End of Life: Why Are So Many Patients Still Taking Statins and Bisphosphonates?

The Medication List Often Outlives the Decision That Created It

A statin started after a risk calculation at age 62 may still be renewed after advanced dementia, recurrent aspiration, metastatic cancer, or severe frailty has fundamentally changed the patient’s clinical horizon. A bisphosphonate started after a fragility fracture may remain on the list even when swallowing is unsafe, renal function is unstable, or several years of therapy have already produced residual skeletal protection. The original prescription may have been entirely reasonable. The failure is not the decision to start. It is the absence of a later decision to reassess.

Medication burden near the end of life is not simply a count of pills. It includes difficulty swallowing, fasting instructions, administration time, laboratory testing, infusion travel, caregiver work, adverse effects, drug interactions, cost, and the cognitive effort of maintaining a regimen. In a retrospective study of 410 adults aged 65 years or older who were hospitalized in their last year of life, the mean number of different medications administered across hospitalizations was 23.8. More than 80% had at least one STOPPFrail-defined potentially inappropriate medication at discharge, and lipid-lowering drugs were among the most frequent categories. [1]

The persistence of preventive therapy is not usually caused by indifference. Clinicians report fear of adverse consequences after stopping, uncertainty about another prescriber’s intent, specialist ownership of the medication, poor exchange of information, limited time, and unclear responsibility for the final decision. Patients and families may hear “stop” as “give up,” particularly when a drug has been described for years as protective. [2]

A structured review can change that pattern. In a randomized trial of 130 adults aged 75 years or older with advanced frailty and polypharmacy, STOPPFrail-guided recommendations reduced medication use by a mean of 2.25 more drugs than usual care and lowered medication costs. The study was too small to establish differences in mortality, falls, hospital presentations, or quality of life. STOPPFrail version 2 therefore emphasizes shared decision-making rather than automatic list-based discontinuation. [3,4]

The Core Question: Will This Patient Live Long Enough to Benefit?

The central comparison is not “benefit versus no benefit.” It is expected benefit within the patient’s likely time horizon versus burden and harm that occur now. Prognosis is uncertain, but uncertainty does not make prognosis irrelevant. A practical review can use broad bands such as days to weeks, months, 1 to 2 years, or several years rather than pretending to know an exact date.

Time to benefit must be interpreted alongside the patient’s baseline risk. A patient with recent acute coronary syndrome is not equivalent to a patient taking a statin solely because of an old primary-prevention calculation. A patient with a recent hip fracture and very high near-term fracture risk is not equivalent to a patient who has completed 7 years of alendronate without an intervening fracture. The same drug name can represent very different clinical purposes.

Five Questions Before Continuing a Long-Term Preventive Medication

Question	Why it matters	Examples
What is the current indication?	The medication may have shifted from active treatment to distant prevention, or the original indication may no longer be present.	Primary versus secondary ASCVD prevention; osteoporosis versus hypercalcemia of malignancy.
When is meaningful benefit expected?	A delayed benefit may not fit a prognosis measured in months.	Primary-prevention statin benefit often requires years; fracture prevention generally requires months or longer.
What burden or harm occurs now?	Immediate burden can outweigh a small delayed benefit.	Dysphagia, fasting instructions, myalgias, renal monitoring, infusion travel, caregiver workload.
What happens if the drug is stopped?	Some medications can be stopped directly; others require a transition plan.	Statin withdrawal is different from denosumab interruption.
Does treatment match the patient’s goals?	The same risk reduction can be valued differently when comfort, alertness, home time, or pill reduction becomes the priority.	Continue a tolerated secondary-prevention statin after a recent event; stop a burdensome primary-prevention drug in the final months.

Statins: Separating Primary Prevention From Secondary Prevention

The 2026 ACC/AHA multisociety dyslipidemia guideline makes the geriatric principle explicit: decisions about lipid-lowering therapy should not be based on chronological age alone. The discussion should include patient priorities, functional status, multimorbidity, frailty, polypharmacy, and life expectancy. For adults aged 75 years or older with life expectancy of at least 2.5 years, initiating a moderate-intensity statin for primary prevention may be reasonable after clinician-patient discussion. For patients with life expectancy of less than 1 year, discontinuing lipid-lowering therapy may be reasonable to avoid unnecessary medication use or adverse effects. [5]

That guidance is intentionally asymmetric. It does not say that everyone older than 75 should start a statin, and it does not say that everyone with a prognosis shorter than 1 year must stop. It creates permission for an individualized decision.

Time to Benefit Is Most Useful in Primary Prevention

A meta-analysis of primary-prevention trials in adults aged 50 to 75 years estimated that treating 100 people for about 2.5 years would prevent one major adverse cardiovascular event. It did not find evidence of a mortality benefit. The estimate is useful for framing delayed benefit, but it should not be directly transplanted to adults older than 75 years, patients with established ASCVD, or people receiving hospice care. [6]

For a patient taking a statin only to reduce the probability of a first event, a prognosis measured in months makes meaningful benefit less likely. For a patient with a recent myocardial infarction, ischemic stroke, symptomatic peripheral artery disease, or recurrent unstable ASCVD, the near-term risk is higher and the rationale for continuation is usually stronger if treatment remains tolerated and goal-concordant. [5]

Two Randomized Trials Addressing Statin Discontinuation in Different Populations

Trial	Population	Practical interpretation
Kutner et al.	381 adults with life expectancy of 1 month to 1 year, functional decline, and no recent active cardiovascular disease.	Sixty-day mortality was not notably different after stopping, but noninferiority was not established. Quality of life improved modestly and cardiovascular events were infrequent. This supports a discussion, not a universal stopping rule. [7]
SAGA/SITE	1,160 adults aged 75 years or older without prior ASCVD who were taking statins for primary prevention; not an end-of-life cohort.	Discontinuation met the prespecified noninferiority criterion for 3-year all-cause mortality. Major cardiovascular events did not differ markedly, LDL rose substantially, and quality-of-life and functional measures did not materially improve. This supports individualization in healthy aging, not automatic discontinuation. [8]

The palliative trial is the more relevant study for patients near the end of life, but its limitations matter. The primary 60-day mortality comparison was not statistically significant, yet the prespecified noninferiority margin was not met. The safest summary is that stopping was not associated with a detected short-term mortality difference in this selected population, while uncertainty remained. [7]

The SAGA/SITE trial answers a different question. Its participants were older adults receiving primary prevention, not patients with a prognosis of months. Stopping did not produce the symptom or quality-of-life improvement that clinicians sometimes assume, and LDL increased. The trial therefore weakens blanket arguments on both sides: continuing every well-tolerated statin is not automatically necessary, but stopping every statin after age 75 is not evidence-based either. [8]

An expert consensus from the National Lipid Association and American Geriatrics Society similarly supports individualized primary-prevention decisions in adults older than 75 years and recognizes deprescribing as reasonable for selected patients with life-limiting disease. [9]

Statin Features That Favor Continuation in Selected Patients

Continuation is more defensible when the patient has a recent or unstable ASCVD event, very high recurrent-event risk, good tolerance, low regimen burden, and a stated preference to preserve preventive treatment. It may also remain reasonable when prognosis is measured in years rather than months and the patient values risk reduction. [5]

Statin Features That Favor Deprescribing in Selected Patients

Deprescribing becomes more defensible when the statin is for primary prevention, life expectancy is less than the expected benefit scope, swallowing or medication burden is substantial, adverse effects or interactions are plausible, and the patient prioritizes comfort or fewer medications. A direct stop was used in the palliative randomized trial; a taper is not generally required for a statin. [7]

Bisphosphonates: Time to Benefit Is Only Half the Story

Bisphosphonate decisions are often mishandled because clinicians apply statin logic without accounting for skeletal retention. The relevant questions are not only how long a new treatment takes to prevent a fracture, but also how long the patient has already been treated, what fracture risk remains, and whether residual effects may persist after discontinuation.

A meta-analysis of randomized trials in postmenopausal women initiating bisphosphonate therapy estimated that approximately 12.4 months of treatment were needed to prevent one nonvertebral fracture per 100 treated. Preventing one hip fracture per 200 treated required about 20.3 months, and preventing one clinical vertebral fracture per 200 required about 12.1 months. These are population averages for treatment initiation, not a stopwatch for an individual patient and not direct evidence about stopping after several years. [10]

Reassessing Treatment Duration and the Possibility of Residual Benefit

The Endocrine Society guideline for postmenopausal osteoporosis recommends reassessing fracture risk after 3 to 5 years of bisphosphonate therapy. Patients who remain at high risk should generally continue, while those at low-to-moderate risk may be considered for a bisphosphonate holiday. The evidence for retained benefit during a holiday is strongest for alendronate and zoledronic acid. [11]

The FLEX extension trial helps explain why. Among postmenopausal women who had already received about 5 years of alendronate, stopping did not significantly increase nonvertebral fracture risk over the extension period. Continued treatment reduced clinically recognized vertebral fractures from 5.3% to 2.4%, suggesting that patients with high vertebral fracture risk may still benefit from continuation. [12]

The HORIZON extension reached a similar risk-stratified conclusion after 3 annual zoledronic acid doses. Bone density and turnover data showed residual effects after discontinuation. Continuing for 6 years reduced morphometric vertebral fractures, while other fracture outcomes did not differ significantly. [13]

For a patient with a short prognosis who has completed several years of therapy and has no recent fracture, stopping may be clinically different from withholding a newly indicated treatment from a high-risk patient. The residual effect can make deprescribing more reasonable in the first scenario. It does not erase near-term fracture risk in the second.

Administration Burden Can Become the Deciding Factor

Oral alendronate must be taken after rising, with 6 to 8 ounces of plain water, at least 30 minutes before food, beverages, or other medications, and the patient must remain upright. Esophageal emptying disorders, inability to sit or stand upright for at least 30 minutes, and hypocalcemia are contraindications. These requirements can become impractical or unsafe in advanced neurologic disease, dysphagia, recurrent aspiration, or profound weakness. [14]

Intravenous zoledronic acid avoids swallowing requirements but creates different burdens. Under the osteoporosis-dose Reclast label, acute renal impairment and creatinine clearance less than 35 mL/min are contraindications; serum creatinine and calculated creatinine clearance are required before each dose, and the infusion must run for at least 15 minutes. Travel, venous access, hydration, monitoring, and post-infusion symptoms may matter more as illness advances. [15]

Long-term bisphosphonate treatment also requires attention to uncommon but serious complications such as osteonecrosis of the jaw and atypical femoral fracture. Those risks should be placed in context rather than used as automatic stopping triggers. [14,15]

Safety Exception: Denosumab Is Not a Bisphosphonate

Denosumab is often grouped with bisphosphonates because both are antiresorptive therapies, but their stopping behavior is fundamentally different. Bisphosphonates bind to bone and may have persistent effects after discontinuation. Denosumab’s effect reverses when doses are delayed, and interruption can lead to rapid bone turnover, bone mineral density loss, and multiple vertebral fractures.

The current Prolia label warns that new vertebral fractures have occurred as early as 7 months after the last injection and directs clinicians to transition patients who discontinue to another antiresorptive therapy. The Endocrine Society also advises against delaying or stopping denosumab without subsequent antiresorptive or other appropriate therapy. [11,16]

A medication review should therefore identify the exact agent before anyone writes “stop osteoporosis drug.” A patient near the end of life may still have a reasonable case for discontinuing denosumab, but the plan must consider the interval since the last dose, prior vertebral fractures, renal function, ability to receive an alternative, prognosis, and goals. Simply allowing the next injection to lapse without a plan is not equivalent to a bisphosphonate holiday.

Another Important Exception: Active Cancer Indications for Bisphosphonate Therapy

Zoledronic acid used for osteoporosis prevention is not the same clinical intervention as zoledronic acid used for hypercalcemia of malignancy, multiple myeloma, or bone metastases. The Endocrine Society recommends an intravenous bisphosphonate or denosumab for adults with hypercalcemia of malignancy, and the oncology zoledronic acid label includes hypercalcemia of malignancy, multiple myeloma, and documented bone metastases from solid tumors receiving antineoplastic therapy. [17,18]

These indications may offer near-term symptom or complication control rather than distant prevention. They still require renal, infusion, adverse-effect, and goals-of-care review, but they should not be stopped by applying an osteoporosis drug-holiday rule.

A Practical Medication Review for Serious Illness

1. Reconstruct the Indication

Do not rely on the medication name alone. Determine whether the statin is for primary or secondary prevention, whether the bisphosphonate is for osteoporosis or active malignancy, and whether the “bone drug” is actually denosumab.

2. Establish the Current Clinical Perspective

Use a broad prognosis band and document the basis: advanced frailty, functional trajectory, cancer status, recurrent hospitalization, nutritional decline, swallowing ability, and the patient’s own understanding. Prognosis should guide the discussion without being presented as certainty.

3. Match the Benefit to the Prognosis

Ask whether the likely benefit is symptom relief, prevention of an event within months, or prevention of an event over years. Primary-prevention statins and newly initiated osteoporosis treatment often have delayed benefits. Active hypercalcemia treatment and prevention after a recent high-risk event may have a shorter and more relevant perspective. [5,6,10,17]

4. Quantify Present Burden

Document pill count, swallowing difficulty, fasting instructions, monitoring, infusion travel, caregiver time, adverse effects, interactions, cost, and whether the regimen interferes with comfort-focused care.

5. Identify Stopping Hazards Before Deprescribing

Statins generally do not require tapering. Prior bisphosphonate therapy may permit a planned holiday after risk reassessment. Denosumab requires an explicit transition strategy. Active cancer therapy requires coordination with oncology or the responsible palliative clinician. [7,11-18]

6. Make the Decision With the Patient or Surrogate

A deprescribing recommendation is stronger when it is tied to the patient’s stated priorities. Some patients will choose fewer medications even if a small future benefit is possible. Others will accept a low-burden preventive drug because avoiding stroke or fracture remains a dominant goal.

7. Document Ownership and Follow-Up

The note should state the indication, prevention category, prognosis band, expected time to benefit, present burden, stopping risk, patient preference, final decision, and who will revisit the plan. Clear ownership prevents the medication from being restarted automatically at the next transition of care.

Decision Cues by Medication Situation

Medication situation	Factors that strengthen continuation	Factors that strengthen deprescribing or deferral
Primary-prevention statin	Prognosis measured in years, high estimated ASCVD risk, good tolerance, low burden, patient preference for prevention.	Prognosis less than expected benefit reach, major pill or swallowing burden, adverse effects or interactions, comfort-focused goals. [5,6,8,9]
Secondary-prevention statin	Recent or unstable ASCVD, recurrent events, high short-term risk, good tolerance, prevention remains a priority.	Remote stable event, profound frailty, prognosis measured in months, substantial burden, adverse effects, patient preference to stop. [5,7]
Oral bisphosphonate for osteoporosis	Recent fragility fracture, very high vertebral or hip fracture risk, safe administration, benefit purview fits prognosis.	Several years completed, low-to-moderate current risk, dysphagia, inability to remain upright, aspiration concern, prognosis shorter than likely benefit. [10-14]
IV zoledronic acid for osteoporosis	High fracture risk, acceptable renal function, feasible infusion, prognosis long enough to benefit.	Residual benefit after prior doses, low-to-moderate risk, renal contraindication, infusion burden, short prognosis. [11,13,15]
Zoledronic acid for active cancer	Hypercalcemia of malignancy, myeloma, or bone metastasis where near-term complication control is goal-concordant.	Active indication absent, burdens exceed likely benefit, or therapy conflicts with goals after oncology review. [17,18]
Denosumab	High fracture risk and ability to continue on schedule; or an appropriate transition has not yet been arranged.	Discontinuation may be appropriate only with a deliberate transition or risk-management plan; do not simply miss the next dose. [11,16]

How to Explain Deprescribing Without Signaling Abandonment

The language used matters. “This medication is no longer worth it” can sound dismissive. A better explanation preserves the logic of the original prescription while acknowledging that the patient’s situation has changed:

“This medicine was reasonable when it was started. Your health priorities and the time it would need to help have changed. Stopping it is not stopping care. It is focusing care on treatments that can help now.”

For a family worried that prevention is being withdrawn, the clinician can make the tradeoff explicit:

“We are not stopping medicines because of age. We are reviewing which treatments are likely to help within the time ahead, which create burden now, and which could cause harm if stopped abruptly.”

When uncertainty is real, say so. The goal is not to convert uncertainty into a confident stopping order. It is to make uncertainty visible, compare it with the patient’s priorities, and assign responsibility for the final decision.

Acknowledging the Limitations of the Current Evidence

Direct randomized evidence for deprescribing at the end of life remains limited. The Kutner trial is the principal statin discontinuation trial in a life-limiting illness population, but it was open-label, heterogeneous, and excluded recent active cardiovascular disease. The 2026 SAGA/SITE trial is larger and more current but studied healthier older adults receiving primary prevention, not patients in hospice or the final months of life. [7,8]

Bisphosphonate time-to-benefit estimates come from postmenopausal osteoporosis trials and mainly address starting treatment. Extension trials inform residual effects after years of alendronate or zoledronic acid but do not directly test deprescribing in advanced frailty or palliative care. [10,12,13]

Prognostic uncertainty also limits precision. A time-to-benefit framework should inform shared decisions, not create a rigid life-expectancy cutoff. Clinical judgment remains necessary for recent cardiovascular events, recent fractures, severe vertebral disease, active cancer complications, and denosumab transition planning.

Future Directions and Unanswered Questions in End-of-Life Deprescribing Research

Research should evaluate class-specific deprescribing strategies in patients with advanced frailty, dementia, organ failure, and cancer. Studies need to include outcomes that matter to patients and caregivers, not only mortality or disease-specific endpoints. They should also distinguish primary prevention, secondary prevention, and active treatment, since the evidence base and clinical implications differ substantially across those categories.

Trials should also test how best to communicate medication changes across hospitals, primary care, specialists, hospice teams, pharmacies, patients, and families. The fragmentation of care transitions remains a practical barrier to safe deprescribing, and implementation research is needed to determine which communication strategies reduce unintended continuation, duplication, or abrupt withdrawal.

Conclusion: Deprescribing as Disciplined Reassessment, Not Blanket Withdrawal

Statins and bisphosphonates remain on end-of-life medication lists because prescriptions renew more easily than goals are revisited. The appropriate response is not a blanket ban on preventive therapy. It is a disciplined reassessment of indication, time to benefit, prognosis, immediate burden, stopping hazards, and patient priorities.

For statins, primary and secondary prevention must be separated, and recent cardiovascular instability matters. For bisphosphonates, prior treatment duration and residual skeletal effects matter. Denosumab requires a transition plan, and oncology bisphosphonate indications require their own analysis. Deprescribing is good care when it removes a treatment unlikely to help within the patient’s purview without creating a new avoidable risk.

 

References

1. Curtin D, O’Mahony D, Gallagher P. Drug consumption and futile medication prescribing in the last year of life: an observational study. Age Ageing. 2018;47(5):749-753. DOI. PubMed. PMID: 29688246.
2. Abou J, Crutzen S, Tromp V, et al. Barriers and enablers of healthcare providers to deprescribe cardiometabolic medication in older patients: a focus group study. Drugs Aging. 2022;39(3):209-221. DOI. PubMed. PMID: 35187614.
3. Curtin D, Jennings E, Daunt R, et al. Deprescribing in older people approaching end of life: a randomized controlled trial using STOPPFrail criteria. J Am Geriatr Soc. 2020;68(4):762-769. DOI. PubMed. PMID: 31868920.
4. Curtin D, Gallagher P, O’Mahony D. Deprescribing in older people approaching end-of-life: development and validation of STOPPFrail version 2. Age Ageing. 2021;50(2):465-471. DOI. PubMed. PMID: 32997135.
5. Blumenthal RS, Morris PB, Gaudino M, et al. 2026 ACC/AHA/AACVPR/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA guideline on the management of dyslipidemia: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. J Am Coll Cardiol. 2026;87(19):2624-2757. DOI. PubMed. PMID: 41824590. AHA guideline resources.
6. Yourman LC, Cenzer IS, Boscardin WJ, et al. Evaluation of time to benefit of statins for the primary prevention of cardiovascular events in adults aged 50 to 75 years: a meta-analysis. JAMA Intern Med. 2021;181(2):179-185. DOI. PubMed. PMID: 33196766.
7. Kutner JS, Blatchford PJ, Taylor DH Jr, et al. Safety and benefit of discontinuing statin therapy in the setting of advanced, life-limiting illness: a randomized clinical trial. JAMA Intern Med. 2015;175(5):691-700. DOI. PubMed. PMID: 25798575.
8. Bonnet F, Poulizac P, Rousselot N, et al. Discontinuation of statins for primary prevention of atherosclerotic cardiovascular disease in adults aged 75 years or older (SAGA/SITE): a multicentre, open-label, pragmatic, non-inferiority randomised trial. Lancet Healthy Longev. Published online August 11, 2026;100884. DOI. PubMed. PMID: 42580354.
9. Bittner V, Linnebur SA, Dixon DL, et al. Managing hypercholesterolemia in adults older than 75 years without a history of atherosclerotic cardiovascular disease: an expert clinical consensus from the National Lipid Association and the American Geriatrics Society. J Am Geriatr Soc. 2025;73(6):1674-1696. DOI. PubMed. PMID: 40207842.
10. Deardorff WJ, Cenzer I, Nguyen B, Lee SJ. Time to benefit of bisphosphonate therapy for the prevention of fractures among postmenopausal women with osteoporosis: a meta-analysis of randomized clinical trials. JAMA Intern Med. 2022;182(1):33-41. DOI. PubMed. PMID: 34807231.
11. Eastell R, Rosen CJ, Black DM, Cheung AM, Murad MH, Shoback D. Pharmacological management of osteoporosis in postmenopausal women: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab. 2019;104(5):1595-1622. DOI. PubMed. PMID: 30907953. Current guideline resource.
12. Black DM, Schwartz AV, Ensrud KE, et al. Effects of continuing or stopping alendronate after 5 years of treatment: the Fracture Intervention Trial Long-term Extension (FLEX): a randomized trial. JAMA. 2006;296(24):2927-2938. DOI. PubMed. PMID: 17190893.
13. Black DM, Reid IR, Boonen S, et al. The effect of 3 versus 6 years of zoledronic acid treatment of osteoporosis: a randomized extension to the HORIZON-Pivotal Fracture Trial. J Bone Miner Res. 2012;27(2):243-254. DOI. PubMed. PMID: 22161728.
14. DailyMed. Alendronate sodium tablet, prescribing information. U.S. National Library of Medicine. Official label. Accessed August 20, 2026.
15. DailyMed. Reclast (zoledronic acid) injection, prescribing information. U.S. National Library of Medicine. Official label. Accessed August 20, 2026.
16. DailyMed. Prolia (denosumab) injection, prescribing information. U.S. National Library of Medicine. Official label. Accessed August 20, 2026.
17. El-Hajj Fuleihan G, Clines GA, Hu MI, et al. Treatment of hypercalcemia of malignancy in adults: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab. 2023;108(3):507-528. DOI. PubMed. PMID: 36545746. Official guideline resource.
18. DailyMed. Zoledronic acid injection, prescribing information for hypercalcemia of malignancy, multiple myeloma, and bone metastases of solid tumors. U.S. National Library of Medicine. Official label. Accessed August 20, 2026.