First-Line Overactive Bladder Therapy in the Older Woman: Weighing the Anticholinergic Signal
Two patients share the same first prescribing decision for overactive bladder — but one variable, present in only one of them, changes which risk actually deserves to drive the choice.
Renata S., 68, has tutored adults for their GED two evenings a week at the community college since retiring from a payroll department eleven years ago, and was ready to give it up entirely last winter when she started needing three bathroom trips per class instead of one. She has hypertension of about fifteen years' standing, treated with lisinopril and amlodipine and reasonably controlled at home — an average of 132/78 across the log she brought in, not the labile, poorly-controlled pattern that would make any new drug's blood-pressure effect an immediate emergency, but controlled enough that her physician would rather not add a variable that could quietly erode it. She has no other chronic conditions, has never smoked, and lives alone since her husband's death four years ago, managing her own household and finances without assistance. A Mini-Cog screen in clinic today, done as routine baseline before starting any new medication in a patient over 65, was entirely normal.
Her voiding diary shows nine daytime episodes and two nocturnal, with urgency severe enough that she has had two leakage episodes reaching the classroom parking lot before she could get inside — embarrassing enough that she has started parking closer and skipping her usual coffee beforehand, a workaround that hasn't actually solved the underlying urgency. The reflex first step — an antimuscarinic — carries an observational association with incident dementia that has now replicated across several independent cohorts using different databases and methods (Gray 2015; Coupland 2019), a consistency that argues against pure chance even though none of the underlying studies can rule out confounding by indication: urinary symptoms are themselves recognized as an early, nonspecific feature of prodromal dementia in some patients, which means the same undiagnosed process could in principle be driving both the prescription and a diagnosis years later, with no drug effect required at all. Her own hypertension adds a second, more concrete and immediately testable consideration on the other side of the ledger: mirabegron, the older of the two beta-3 agonists, carries its own labeled blood-pressure signal, which would need monitoring in a patient already treated for hypertension rather than sidestepping one risk by walking into another.
Solifenacin first. Head-to-head and network meta-analysis data still put antimuscarinics slightly ahead on efficacy against beta-3 agonists, and I'm not willing to hand a patient a less effective drug over an association that has never been shown to be causal. Both Gray and Coupland are observational — urinary symptoms are themselves an early, non-specific feature of prodromal dementia in some patients, which means the same underlying process could plausibly drive both the prescription and the diagnosis years apart, with no drug effect required at all.
If her cognitive screen ever turns abnormal, that's the moment to revisit the class, not before.
The confounding-by-indication argument would carry more weight if this were one study. It's not — Gray's prospective cohort and Coupland's independent nested case-control, using entirely different populations and designs, both find the same dose-response gradient: more cumulative anticholinergic exposure, more dementia risk. That kind of convergence across methods is exactly what starts to outweigh a single-study confounding objection, even short of a randomized trial that could prove causation outright.
I'd still start a beta-3 agent in her, not because her screen today means anything is wrong, but because if she needs this drug for years, today is the cheapest day to avoid the exposure entirely.
I don't think the dementia question is actually what should decide this visit — her hypertension is. If we're choosing a beta-3 agent for her regardless of which side of the efficacy argument wins, mirabegron and vibegron are not interchangeable in a treated hypertensive: mirabegron's label carries a real blood-pressure elevation signal, while EMPOWUR's own cardiovascular safety data found vibegron's rates of hypertension no different from placebo. Whichever way the class debate resolves, if a beta-3 agent is going in her chart, it should be vibegron specifically, not mirabegron by default.
Agreed within the visit: vibegron 75mg daily, home BP log continued at its current cadence, voiding diary repeat in six weeks.
Solifenacin remains available; the anticholinergic question would need revisiting explicitly, not defaulted back to as the obvious next step.
No adjustment needed to the OAB regimen itself — vibegron was chosen specifically to keep this variable out of the picture.
Not agreed: whether the anticholinergic-dementia association is real and causal, or a persistent confound that happens to replicate. The geriatrician would apply the same beta-3-first logic to any woman over 65 regardless of her blood pressure; the urogynecologist still considers antimuscarinics reasonable first-line therapy for most patients and views today's choice as driven by her hypertension specifically, not by the dementia literature.
T.O., 74, was a hospital pharmacy technician for over thirty years before retiring, and now spends most of her time helping raise her granddaughter three days a week while her daughter works. Eight months ago her primary care physician flagged word-finding difficulty and short-term memory lapses significant enough for formal cognitive testing; a neuropsychological evaluation confirmed mild cognitive impairment, amnestic subtype, with her MoCA scoring 22/30. She lives independently and still drives locally, though her daughter has started reviewing her medication box weekly since the diagnosis, and a neurologist started donepezil at that visit, which she has tolerated without issue. She has type 2 diabetes and hyperlipidemia, both stable on metformin and atorvastatin for over a decade, and otherwise no history of falls, syncope, or gait disturbance before the two episodes below.
Her overactive bladder symptoms — urgency, frequency, and two nighttime falls in the past three months while rushing to the bathroom in the dark — predate the cognitive diagnosis by roughly two years and have worsened since. Both falls occurred without injury, caught by furniture rather than the floor, but her daughter is now openly worried about a third with less luck attached to it. The clinical question here is sharper than the same first-drug decision in a cognitively normal woman: the anticholinergic-dementia literature (Gray 2015; Coupland 2019) was built on populations without dementia at baseline, reasoning forward about a future risk that hadn't happened yet. T.O. already carries the outcome those studies were trying to predict, which removes the confounding-by-indication defense that still has real force in a patient whose cognition is presently intact — there is no unresolved question here of whether her urinary symptoms might be an early marker of a diagnosis that hasn't happened yet, because it already has, eight months on record.
I don't think a cognitive diagnosis by itself should take our most effective drug class off the table. Undertreated OAB has its own real cost in a patient like her — two falls already, both nocturia-related, and a third is exactly the kind of event that ends independent living. A documented MCI diagnosis is a reason to monitor more closely, not necessarily a reason to start with a less effective agent.
I take the fall risk seriously, but the reasoning that makes antimuscarinics defensible in a cognitively normal patient doesn't transfer here. In Renata's case, the honest objection to the dementia literature is that we can't rule out her urinary symptoms and a future diagnosis sharing a common cause we haven't identified yet. That objection has no purchase in T.O. — she already carries the diagnosis. There's no future outcome left to confound; adding anticholinergic burden on top of an existing amnestic MCI diagnosis is not a probabilistic bet the way it was for Renata, it's compounding a risk in the population the original studies actually enrolled.
The fall risk is real, but it's an argument for treating her OAB effectively, not an argument for which class does it — a beta-3 agent treats the urgency just as directly.
Agreeing to avoid the antimuscarinic class doesn't finish this. Her chart lists donepezil, which her neurologist added at diagnosis — and donepezil is metabolized substantially through CYP2D6. Mirabegron is a moderate CYP2D6 inhibitor; using it here risks elevating her donepezil exposure in a way that has nothing to do with the bladder question at all. Vibegron lacks meaningful CYP2D6 activity. The 'avoid the anticholinergic' instinct is right, but which beta-3 agent gets chosen still has real pharmacologic stakes specific to her.
I'd flag this to her neurologist regardless of which drug we start, simply so it's on record that the interaction was checked.
Agreed: vibegron 75mg daily, nightlight and clear bathroom path installed as a fall-prevention measure independent of the drug choice, and a note sent to her neurologist documenting the donepezil interaction check.
Not agreed, and stated plainly rather than smoothed over: how much weight the real, already-occurring functional harm (two falls) should carry against a longer-horizon, still-probabilistic cognitive risk that may never advance further in her case regardless of today's drug choice.