Fluctuating confusion in a 68-year-old on an anticholinergic cocktail
A 68-year-old retired teacher is brought to the emergency department by his daughter after three days of fluctuating confusion. He has hypertension, osteoarthritis, and an overactive bladder. Current medications: hydrochlorothiazide, ibuprofen, oxybutynin 5 mg twice daily, diphenhydramine 25 mg nightly for sleep, and oxycodone 5 mg as needed since a knee injection four days ago. He lives alone and managed his own finances until last week. The daughter reports that he asked the same question five times in an hour and did not recognize her for a few minutes the previous evening; he is worse at night. On examination he is afebrile, oriented to person only, intermittently drowsy, and scores 22/30 on the MMSE. The daughter adds that over the past year he has occasionally forgotten appointments but always recalled them when reminded, and he has never been disoriented before. The resident writes "dementia, likely Alzheimer's type" in the chart and plans outpatient follow-up.
1. Why is the resident's diagnosis premature, and what in this history should drive the immediate workup?
Read the worked answer
The core distinction is delirium versus dementia. Delirium requires an acute onset with a fluctuating course, inattention, and altered level of consciousness; this man deteriorated over three days, is intermittently drowsy, and is worse at night. Dementia is a chronic, progressive decline in which arousal and attention are usually preserved until late. The resident anchored on the daughter's mention of a year of occasional forgetfulness - which, by her own account, resolved with cueing - and never screened for delirium. On the Confusion Assessment Method, acute onset/fluctuation plus inattention, with either disorganized thinking or altered consciousness, is a positive screen.
Medication burden is the strongest lead. Diphenhydramine and oxybutynin both carry high anticholinergic load and appear on lists of drugs to avoid in older adults; together they are a textbook delirium precipitant. Oxycodone started four days ago adds a second hit, and a thiazide with an NSAID raises the risk of hyponatremia and renal impairment.
Immediate steps: medication review, vitals, glucose, electrolytes including sodium, renal function, urinalysis and culture, oxygen saturation, and pain assessment; treat as delirium until proven otherwise and reassess cognition only after recovery. Documenting a dementia diagnosis now risks a self-fulfilling label and misses a reversible cause.
2. Two weeks later, after the anticholinergic drugs are stopped and a urinary tract infection treated, he is alert and oriented but misplaces keys and forgets new acquaintances' names. Free recall is 4/10 but improves to 9/10 with semantic cueing, and recognition is intact. What does this pattern suggest about the locus of the memory problem, and what does it argue against?
Read the worked answer
The pattern points to retrieval failure rather than storage failure. On a cueing paradigm of the Grober–Buschke type, the key comparison is free recall versus recall after semantic cueing, and then recognition. A large cueing benefit with preserved recognition indicates that the material was encoded and stored but not strategically accessed - a profile associated with frontal/executive or subcortical dysfunction, attentional and organizational problems, depression, or medication effects.
What it argues against. Probable Alzheimer's disease typically shows a storage deficit: poor free recall, limited benefit from cueing, and impaired recognition, reflecting medial temporal damage. That is not the profile here.
How far to take it. Not far on a single assessment. The comparison assumes comparable encoding opportunity, and practice, education, fatigue, and anxiety all shift scores. A person can also have mild cognitive impairment with a non-amnestic or executive pattern, and vascular contributions are common. The defensible move is a longitudinal plan: repeat testing with alternate forms in 6–12 months, collateral report of instrumental activities, vascular risk factor management, and explicit criteria for mild cognitive impairment rather than an immediate neurodegenerative label.
3. The family asks whether donepezil should be started now, as a relative's physician did. What is the appropriate response, and what would change it?
Read the worked answer
No - not on this evidence. Cholinesterase inhibitors are indicated for symptomatic treatment of mild-to-moderate Alzheimer's disease dementia, and the diagnosis here is not established. Two independent problems block the prescription.
First, there is no confirmed dementia syndrome; there is a fluctuating delirium that has resolved with an unmasked, milder memory complaint of uncertain significance. Second, even if he meets criteria for mild cognitive impairment, the trial evidence does not support cholinesterase inhibitors for that state: randomized trials in mild cognitive impairment did not show a clinically meaningful delay in progression to dementia, while cholinergic side effects (nausea, bradycardia, syncope, GI bleeding risk) were increased.
What would change the answer: a longitudinal decline corroborated by an informant, functional impairment in instrumental activities, a profile consistent with Alzheimer's pathophysiology (for example a storage-pattern memory deficit, or supportive biomarker findings where appropriate), and exclusion of reversible and vascular contributors. Even then, the conversation should include modest average benefit, shared decision-making, and a plan to reassess tolerability and goals. Prescribing now would medicalize a reversible episode and substitute a drug for an etiologic workup.