What it is:Geriatric pharmacotherapy isn't a disease state, it's a lens. The same drug, at the same dose, behaves differently in a 78-year-old than a 38-year-old because absorption, distribution, metabolism, excretion, and receptor sensitivity all shift with age, usually without any single lab value telling you so.
The core problem:Older adults carry more chronic disease, take more drugs, and tolerate errors worse than any other population, yet they're the group most often excluded from the trials those drugs were tested in. Medication-related problems (MRPs) are common, costly, and mostly preventable.
What you do about it:Assume nothing is "just aging" until you've ruled out the drug list. Start low, go slow, reassess constantly, and be just as willing to stop a drug as to start one.
Simonson's framing: "It is an art of no little importance to administer medicines properly, but it is an art of much greater and more difficult acquisition to know when to suspend or altogether omit them."(Philippe Pinel, 1745-1826). Two and a half centuries later, deprescribing is still the harder skill to teach.
None of these changes are diseases. They're the expected trajectory of every organ system, and they're why "normal" adult dosing can be wrong in an otherwise healthy 80-year-old.
| System | What happens |
|---|---|
| Renal | ↓ GFR, ↓ renal blood flow, ↓ functioning nephrons, ↓ tubular secretion, ↓ renal mass |
| Sensory | Presbyopia (can't focus near), ↓ night vision, presbycusis (high-pitch hearing loss), ↓ smell and taste |
| Musculoskeletal | ↓ bone mass (osteopenia), ↓ muscle mass, joint stiffening, altered gait and posture |
| Skin/hair | Thinning stratum corneum, fewer melanocytes, thinner subcutaneous fat, atrophied sweat glands |
| Body composition | ↓ lean muscle, ↑ fat mass; height loss of roughly 0.4 in per decade after 40, more after 70 |
GFR falls roughly 25-50% between ages 25 and 90, though the rate varies a lot by patient. Since creatinine production also falls with lost muscle mass, serum creatinine alone can look deceptively normal in a frail older adult with real renal impairment. Estimate CrCl, don't eyeball SCr.
| Phase | Change | Why it matters |
|---|---|---|
| Absorption | Passive diffusion and bioavailability are largely unchanged for most drugs. Active transport drops, so actively absorbed nutrients (calcium, iron, B12) absorb less. Reduced first-pass metabolism can raise bioavailability of some drugs and lower it for some prodrugs. | Absorption is the PK phase that changes the least. Don't blame it first. |
| Distribution | ↓ total body water and ↑ body fat. Water-soluble drugs (digoxin, morphine, lithium) get a smaller Vdand higher plasma concentration per dose. Fat-soluble drugs, including most CNS/psychiatric drugs, get a larger Vdand longer terminal half-life. Serum albumin tends to fall, so more drug circulates unbound (active) for highly protein-bound agents. | This is why a benzodiazepine or antipsychotic can linger for days in an older adult even after a single dose. |
| Metabolism | Liver shrinks and hepatic blood flow falls. Capacity-limited (Phase I, oxidative) metabolism is affected more than Phase II (conjugation). Drugs with high hepatic extraction lose clearance and gain half-life. | When picking between two similar drugs, the one cleared by Phase II glucuronidation (like lorazepam) is often gentler than one needing Phase I oxidation. |
| Excretion | Renal clearance falls in step with GFR, extending half-life for renally cleared drugs and their active metabolites. | This is the phase with the most direct, most predictable dosing consequence. Always estimate CrCl before dosing renally cleared drugs. |
Digoxinis water-soluble and distributes mainly into lean tissue. Less muscle mass in an older adult means a smaller volume of distribution, so the same loading dose produces a higher peak concentration. Combine that with reduced renal clearance of a narrow-therapeutic-index drug and you can see why digoxin toxicity skews heavily toward older patients.
PD changes are harder to quantify than PK changes and can go either direction, enhancing or blunting a drug's effect. The pattern seen most consistently is heightened CNS sensitivity: sedation, confusion, and falls appear at doses a younger adult would tolerate easily. Anticholinergic toxicity is also more pronounced, and baroreceptor reflexes blunt with age, making orthostatic hypotension more common and more dangerous (a fall risk, not just a lab abnormality).
Simonson's one-line summary of this whole section: "Start low and go slow."It sounds obvious, but it's the organizing principle behind essentially every geriatric dosing decision that follows in this document.
The classic textbook presentation of a disease is frequently absent in older adults. Confusion, lethargy, and functional decline often stand in for the pain or fever you'd expect. Missing this is a major driver of delayed diagnosis.
| Disease | How it actually shows up |
|---|---|
| Acute MI | Only ~50% present with chest pain. Look instead for weakness, confusion, syncope, or abdominal pain. ECG findings are unchanged from younger patients. |
| Heart failure | Instead of classic dyspnea, expect hypoxic symptoms, lethargy, restlessness, and confusion. |
| GI bleed | ~10% mortality, but presentation is nonspecific: altered mental status or syncope with hemodynamic collapse. Abdominal pain is often absent entirely. |
| Upper respiratory infection | Lethargy, confusion, anorexia, and decompensation of an existing condition. Fever and productive cough may or may not show up. |
| UTI | Dysuria, fever, and flank pain may be absent. More common: incontinence, confusion, abdominal pain, nausea/vomiting, azotemia. |
New confusion, forgetfulness, gait instability, or incontinence in an older adult gets written off as "just aging" or "just dementia" far too often. Before accepting that explanation, rule out an acute illness andthe medication list. Simonson quoted this directly: don't attribute new symptoms to aging without first asking whether they're an adverse drug effect.
Before you can judge whether a regimen is appropriate, you need to know what the patient can actually do for themselves.
| ADLs (basic self-care) | IADLs (independent living) |
|---|---|
| Feeding · continence (bladder/bowel) · walking independently · toileting independently | Using the telephone · shopping · preparing food · housekeeping · laundry · transportation · handling medications· handling finances |
"Handling medications" is an IADL, not an ADL.A patient can be fully continent and mobile and still be unable to safely self-administer a 12-drug regimen. Nursing facility residents need heavy ADL assistance (96% need help bathing, 58% eating); assisted living residents need much less, which is exactly why medication regimen review requirements differ between the two settings.
Both are lists of potentially inappropriate medications (PIMs), but "potentially" is doing real work: PIM ≠ definitely inappropriate. They're a starting point for judgment, not a rulebook to apply blindly.
| Beers Criteria (AGS) | STOPP Criteria | |
|---|---|---|
| Organization | 5 tables, organized by organ system/therapeutic category | 11 sections, organized by physiologic system |
| Covers | PIMs for all older adults · PIMs by drug-disease interaction · PIMs to use with caution · high-risk drug-drug interactions to avoid · dose adjustments by renal function | Drugs to avoid, drug-drug interactions, renal dosing considerations, drug-disease interactions |
| Unique feature | Explicitly flags "avoid ≥3 concurrent CNS-active drugs" to reduce fall risk | Includes a non-drug-specific section: no evidence-based indication, wrong duration, therapeutic duplication |
Ten questions to ask about every single medication on the list, every time you review it:
Beers and STOPP tell you which drugsare risky in general. The MAI is a processyou apply to one patient's actual list. You need both: a drug can pass every Beers screen and still fail the MAI if it's duplicative or the dose was never reassessed.
Hepler and Strand defined a medication-related problem (MRP) as any event or circumstance involving drug therapy that actually or potentially interferes with an optimal outcome. In the elderly, MRPs aren't rare: they cause or worsen falls, cognitive loss, dehydration, incontinence, and depression, and drive hospitalization, nursing facility placement, and death.
MRPs cost an estimated $200 billion annuallyin the US and contribute to over 200,000 deaths a year. If adverse drug effects were tracked as a single disease, they'd rank as roughly the 5th leading cause of deathin the country. Nearly a third of ambulatory ADEs and half of nursing-facility ADEs are preventable, and most preventable errors happen at the prescribing or monitoring step, not dispensing.
| # | Type | Real example from practice |
|---|---|---|
| 1 | Needs therapy for an untreated condition | Undiagnosed osteoporosis never started on a bisphosphonate |
| 2 | Wrong drug for the condition | Look-alike/sound-alike dispensing error: hydroxyzine ordered, hydralazinedispensed, patient became hypotensive and syncopal. Why TALL-man lettering (hydrOXYzine vs hydRALazine) exists. |
| 3 | Correct drug, dose too low | Underdosed pain control or antidepressant therapy, common because clinicians under-titrate in older patients out of caution |
| 4 | Correct drug, dose too high | Methotrexate 7.5 mg ordered dailyinstead of weekly for RA. The resident developed mouth sores, then severe leukopenia, then fatal septicemia. Weekly methotrexate is effective; daily dosing at that strength is lethal within weeks. |
| 5 | Not receiving the prescribed drug | Fosamax (alendronate) 70 mg weekly with bioavailability of only 0.64% under ideal conditions, dropping ~60% further if taken with coffee or OJ, and becoming negligible if taken after breakfast. Poor administration technique = the patient effectively isn't getting the drug. |
| 6 | Taking a drug with no valid indication | Haloperidol and omeprazole started during a hospitalization for delirium/stress ulcer prophylaxis, then carried forward indefinitely after discharge with no one revisiting whether either was still needed |
| 7 | Adverse drug reaction | Confusion and cardiac rhythm disturbances traced to OTC Tylenol PM(diphenhydramine + acetaminophen). The diphenhydramine, a Beers-listed anticholinergic, was the culprit, and the patient didn't consider it "a medication" because it was OTC. |
| 8 | Drug-drug, drug-food, or drug-lab interaction | Stable warfarin patient's INR crashed after he started eating large amounts of spinach for its "health benefits." Vitamin K antagonizes warfarin's inhibition of factors II, VII, IX, X. |
A 75-year-old on a routine inhaled corticosteroid developed bilateral Achilles tendon ruptures within 2 weeks of starting a fluoroquinolone for a skin infection. Baseline tendinopathy risk with a quinolone alone is ~0.14-0.4%, roughly 1.5-2.7x higher after age 60, but concurrent corticosteroid use in patients 60+ raises that risk about 47-fold.Median onset is 6 days; it can occur from hours after the first dose to months after the drug is stopped.
More chronic conditions, more drugs, more high-risk drugs, more prior ADRs, more prescribers who don't talk to each other, physical/cognitive barriers to taking drugs correctly, and financial or access barriers all stack on top of the PK/PD changes already covered above.
There's no single "treatment algorithm" for geriatrics the way there is for a single disease. The algorithm is a way of thinking that gets applied to every prescription, every refill, and every new symptom.
Underuse, overuse, intelligent non-compliance(patient deliberately skips a dose because of a side effect they never reported), premature discontinuation, and simply not filling the prescription. "Intelligent non-compliance" is worth specifically asking about, since patients often assume you already know why they stopped.
These aren't a complete geriatric dosing reference, they're the specific examples that keep showing up as teaching cases because the standard adult dose causes real harm.
| Drug / class | Standard adult dosing | Geriatric approach | Why |
|---|---|---|---|
| Sedative-hypnotics | |||
| Zolpidem | 10 mg (men) / 5 mg (women) at bedtime | 5 mg for all older adults, and avoid outright per Beers when possible | Slower hepatic clearance across the board; falls and next-day cognitive impairment risk regardless of sex |
| Benzodiazepines, if unavoidable | Varies | Lorazepam preferredover diazepam, alprazolam, or clonazepam | No active metabolites and an intermediate half-life; less accumulation than long-acting agents like diazepam |
| Doxepin for insomnia | 25-50 mg (antidepressant dosing) | ≤6 mgas a sleep aid | At antidepressant doses it's a strongly anticholinergic Beers-listed drug; at ≤6 mg it's selective enough to avoid that burden |
| Anxiolytics | |||
| Alprazolam (GAD) | 0.25-0.5 mg TID | 0.25 mg 2-3x/day | Reduced clearance in older adults produces higher plasma concentrations at the same dose |
| Narrow therapeutic index | |||
| Digoxin | Weight-based loading | Lower loading and maintenance doses; confirm renal function first | Smaller Vd from reduced lean mass plus reduced renal clearance = higher peak and trough concentrations |
| Methotrexate (RA) | Low-dose weekly regimen | Confirm "weekly" explicitly with every new order; daily dosing at RA strength is fatal | Daily administration at what should be a weekly dose causes cumulative myelosuppression and death within weeks |
Alendronate's oral bioavailability is only 0.64% under perfect conditions (full glass of water, empty stomach, upright 30 minutes, nothing else by mouth for 30 minutes) and drops another ~60% with coffee or OJ, becoming negligible after breakfast. A "correctly dosed" bisphosphonate taken incorrectly is functionally an MRP type 5: not receiving the drug at all.
You won't be expected to recite all ~100 Beers entries, but a handful come up constantly in case-based questions and in real geriatric patients:
The common thread across nearly the entire high-yield Beers list: anticholinergic burden and CNS depression.If you can name the mechanism, you can usually guess whether a drug is on the list even without memorizing it.
| Disease | Drug/class to avoid |
|---|---|
| Heart failure | NSAIDs, COX-2 inhibitors, thiazolidinediones, dronedarone |
| HFrEF specifically | Non-dihydropyridine CCBs (diltiazem, verapamil) |
| Syncope | Acetylcholinesterase inhibitors, peripheral nonselective α1-blockers, tertiary TCAs, antipsychotics |
| Delirium | Anticholinergics, antipsychotics, benzodiazepines, corticosteroids, H2 blockers, z-hypnotics |
| Cognitive impairment | Anticholinergics, antipsychotics, benzodiazepines, nonbenzodiazepine hypnotics, z-hypnotics |
| History of falls | Antiepileptics, antipsychotics, benzodiazepines, antidepressants, z-hypnotics, opioids |
| Parkinson disease | Antipsychotics exceptquetiapine, clozapine, pimavanserin; metoclopramide, prochlorperazine, promethazine |
| Peptic ulcer disease | Aspirin ≥325 mg/day, NSAIDs |
| CKD stage IV+ | NSAIDs |
| Urinary incontinence (women) | Peripheral α1-blockers |
Notice how many rows point back to the same few drug classes (anticholinergics, antipsychotics, benzodiazepines, NSAIDs). Learn those four classes cold and most of this table falls out for free.
If a patient is already dependent on a benzodiazepine, stopping abruptly is its own MRP, not a fix. The withdrawal itself can worsen the anxiety or insomnia you're trying to treat, and in a case example an abrupt lorazepam discontinuation directly preceded a return visit for worsened symptoms.
Don't forget the rest of the regimen while you're focused on the benzodiazepine:in the teaching case, correcting an iatrogenically over-replaced levothyroxine dose (causing subclinical hyperthyroidism) and correcting an iron deficiency driving restless legs syndrome were just as important to fixing the patient's sleep as anything done to the sedative-hypnotic.
Deprescribing is the systematic, planned, and supervised process of stopping or reducing medications when existing or potential harms outweigh existing or potential benefits. It is not "cutting corners," it's the second half of good prescribing.
| Benefits | Barriers |
|---|---|
| ↓ medication burden · ↑ cognition (less delirium) · ↑ mobility · ↓ falls · ↑ quality of life · financial savings | Pressure to prescribe · clinical complexity · fragmented care across multiple prescribers · reluctance to stop a drug someone else started · incomplete patient information · ambiguous or shifting care goals · time pressure · fear of withdrawal effects · simply not knowing how |
The drug classes with the most-developed deprescribing algorithms: proton pump inhibitors, antihyperglycemics, antipsychotics, benzodiazepines, cholinesterase inhibitors/memantine, and opioids. PPIs deserve special mention: they're started appropriately for stress ulcer prophylaxis in the hospital and then frequently never stopped, and long-term use carries fracture and C. difficilerisk that most patients and prescribers underweight.
"It's a lot easier to start a medication than it is to stop one."Indications come and go; medications, left unattended, tend to stay. Ongoing medication regimen review exists specifically to keep matching the drug list back to a current, active indication.
Where a patient lives changes both the acuity you should expect and the pharmacist's regulatory role.
| Setting | Scale (US) | Pharmacist's role |
|---|---|---|
| Nursing facility | ~14,700 facilities, 1.6M licensed beds, 1.2M residents | Federally mandated monthly Medication Regimen Review (MRR)for every resident, plus committee work (QAPI, psychotropic drug review) and survey compliance support |
| Assisted living / residential care | ~32,200 facilities, 1.3M beds, ~1M residents | Similar population and polypharmacy burden to NF, but with less-skilled staff and MRR requirements that vary by state rather than federal mandate |
| Home health care | ~11,500 agencies, 3.3M individuals served annually | Largely untapped for pharmacy; major opportunity in preventing re-hospitalization |
| Hospice | ~5,800 agencies, 1.8M served annually | Comfort-focused care where deprescribing shifts from "consider it" to the default assumption |
Both populations look similar on paper (elderly, multiple diagnoses, increasing acuity, multiple prescriptions, ADL assistance needs), but nursing facility residents need dramatically more hands-on assistance (96% need help bathing vs 62% in AL; 58% need help eating vs 20% in AL) and are under federal MRR mandate, while AL regulation is state-dependent and inconsistent. Don't assume "assisted living" means low-acuity or low medication risk.
Cost context(useful for counseling conversations about site of care): a semi-private nursing facility room runs roughly $115,000/year, private room $130,000/year; assisted living averages about $74,000/year; non-medical in-home caregiving runs about $80,000/year. Costs vary enormously by state (a private NF room can run from roughly $90,000/year to well over $380,000/year).
| Parameter | When | Watching for |
|---|---|---|
| Estimated CrCl/GFR | Baseline and at least annually, more often if acutely ill or on new renally cleared drugs | Renally cleared drug accumulation; don't trust SCr alone given reduced muscle mass |
| Full medication list, including OTC/supplements | Every visit, and formally every month in nursing facilities (MRR) | Duplication, expired indications, Beers-listed agents, new drug-drug or drug-disease interactions |
| Cognition and mental status | Baseline and with any new confusion, sedation, or behavioral change | Anticholinergic burden, delirium, over-sedation; don't default to "it's just aging" |
| Falls history and orthostatics | Every visit, especially after starting/titrating CNS-active or antihypertensive drugs | Blunted baroreceptor response, ≥3 concurrent CNS-active drugs |
| Functional status (ADL/IADL) | Periodically and after any hospitalization | New inability to self-manage medications; may require simplifying the regimen or caregiver involvement |
| Adherence pattern | Every refill | Under-use, over-use, intelligent non-compliance, cost barriers |
| Narrow-TI drug levels (digoxin, etc.) | Steady state, and with any renal or dose change | Smaller Vd and reduced clearance both push levels up |