Suspected donepezil withdrawal syndrome in an older woman with Alzheimer disease of abrupt cessation: A case report

Abstract

BACKGROUND

Acetylcholinesterase inhibitors such as donepezil are frequently discontinued in older adults with dementia when adverse events emerge, when perceived benefit declines, or during transitions to institutional care. Abrupt discontinuation may be followed by cognitive or behavioural deterioration, although the clinical phenotype and causal mechanisms remain incompletely defined.

CASE SUMMARY

An 89-year-old woman with a 10-year history of Alzheimer disease had received donepezil 10 mg/day for more than 5 years. Two weeks before nursing-home admission, reduced spontaneous speech and functional decline were already noted, with Hasegawa Dementia Scale-Revised declining from 16/30 to 8/30. Donepezil was abruptly stopped on day 0 at admission. Forty-eight hours later, she developed acute agitation, verbal aggression, insomnia and new gait instability, with a positive confusion assessment method assessment on day 2-3. Infection, major metabolic disturbance and acute structural cerebrovascular disease were not identified. Donepezil 5 mg/day was restarted on day 7, while as-needed quetiapine 12.5 mg was also used during the early post-rechallenge period. Agitation and verbal aggression decreased within 5 days, quetiapine was discontinued, sleep normalized by day 12, and by day 21 she was eating independently and adapting to the facility. The clinical interpretation remained provisional because several alternative or contributing explanations could not be fully excluded.

CONCLUSION

The temporal sequence and improvement after donepezil rechallenge are compatible with suspected donepezil withdrawal, but the diagnosis remains provisional and causality is uncertain. Natural dementia progression, nursing-home relocation, occult pain, unrecorded over-the-counter drugs or supplements, non-convulsive seizures, and concomitant quetiapine are alternative or contributing explanations. The 48-hour onset does not imply complete drug elimination because donepezil has a long elimination half-life; rather, it is compatible with an early clinical response to a falling drug exposure, although this mechanism cannot be demonstrated in the present case. Abrupt discontinuation should therefore be approached cautiously and monitored clinically.

Key Words: Donepezil; Withdrawal syndrome; Delirium; Behavioral and psychological symptoms of dementia; Nursing-home admission; Older adults; Case report

Core Tip: Abrupt discontinuation of long-term donepezil was followed 48 hours later by agitation, insomnia and gait instability in an 89-year-old woman who had taken 10 mg/day for more than five years. Symptoms improved after rechallenge with 5 mg/day, but interpretation is limited by pre-existing cognitive decline, nursing-home relocation, concomitant quetiapine and other unmeasured factors. The case illustrates why suspected withdrawal may be considered in the differential diagnosis without assuming that all post-discontinuation deterioration is drug-related.

INTRODUCTION

Acetylcholinesterase inhibitors (AChE-Is) are established symptomatic treatments for Alzheimer disease and may be continued into moderate or severe stages in selected patients[1]. Decisions to discontinue them are variable and may be influenced by adverse effects, perceived lack of benefit, disease progression, caregiver preference or changes in care setting[2]. In the DOMINO-AD randomized trial, continuation of donepezil over 52 weeks was associated with better cognitive and functional outcomes than discontinuation in community-dwelling patients with moderate-to-severe Alzheimer disease[3]. A subsequent secondary and post-hoc analysis found that discontinuation was associated with a higher risk of nursing-home placement during the first year, but not during the subsequent 3 years[4]. Observational nursing-home data have also reported worsening of behavioural and mood symptoms after cholinesterase-inhibitor discontinuation[5].

Case reports and small series have described acute cognitive or behavioural deterioration after donepezil discontinuation, sometimes improving after re-initiation[6-9]. Deliberate withdrawal studies in dementia with Lewy bodies (DLB) and Parkinson disease dementia (PDD) likewise documented deterioration after withdrawal[10,11]. These reports do not establish a uniform withdrawal syndrome, and the populations, withdrawal procedures and outcome measures differ substantially. In contrast, a prospective observational study of institutionalised patients with advanced dementia found no significant difference in cognitive, behavioural or functional outcomes 3 months after clinically selected cholinesterase-inhibitor discontinuation[12], illustrating the heterogeneity of available evidence.

Despite these observations, descriptions of acute deterioration after abrupt donepezil cessation in community-dwelling older women with Alzheimer disease remain limited. The present case is particularly relevant because symptoms emerged 48 hours after discontinuation at a care transition, but important alternative explanations were also present. We therefore describe the clinical sequence, correct previously reported case details, discuss the pharmacokinetic plausibility of early symptom onset, and explicitly limit causal interpretation.

CASE PRESENTATION

Chief complaints

Sudden onset of agitation, verbal aggression toward family members, insomnia and gait instability beginning two days after nursing-home admission in an 89-year-old woman with Alzheimer disease.

History of present illness

The patient had been followed at the outpatient clinic of a primary-care physician for a 10-year history of progressive cognitive decline. Donepezil had been initiated more than five years earlier at 3 mg/day and titrated to 10 mg/day, at which dose she had remained stable for years. At the most recent assessment before the pre-admission decline, Hasegawa Dementia Scale-Revised (HDS-R) was 16/30 and Mini-Mental State Examination was 14/30; ambulation was independent and no falls had been reported. Concomitant medication was amlodipine 5 mg/day for hypertension, with stable home blood-pressure readings.

Two weeks before admission, the family and caregivers noted gradual withdrawal, reduced spontaneous speech and increasing reliance on prompting for activities of daily living; HDS-R at the pre-admission review was 8/30. Because the family could no longer sustain home care, a nursing-home bed was arranged. On day 0, defined as the day of nursing-home admission and abrupt discontinuation of donepezil 10 mg/day, the prescribing physician stopped donepezil because it was considered “no longer necessary” in advanced dementia; amlodipine 5 mg/day was continued. This pre-admission cognitive and functional decline is an important potential confounder and is distinct from the acute syndrome that followed discontinuation.

On day 2, 48 hours after the last donepezil dose, the patient suddenly became verbally aggressive toward her daughter over minor issues, refused to sit for meals, remained awake much of the night, and developed new gait unsteadiness. The family initially interpreted the change as progression of dementia. She was referred to the emergency department and evaluated on day 3. Head computed tomography (CT) and routine laboratory testing were unremarkable, and she was assessed as having delirium and returned to community care. The abrupt behavioural change therefore occurred after both a recent decline in cognition and a major environmental transition, neither of which can be separated completely from the medication change.

From day 3 through day 6 she remained agitated and sleepless at home. On day 7, the medication history was re-checked and the temporal relationship to donepezil cessation was recognized. Donepezil 5 mg/day was restarted as a clinical rechallenge. As-needed quetiapine 12.5 mg was also prescribed for behavioural control when the patient or caregivers were at risk and was used almost daily during the first week. Because quetiapine was administered during the early post-rechallenge period, its sedative contribution cannot be separated completely from the effect of donepezil re-initiation. By day 12 she was sleeping through the night; by day 21 she was eating independently and adapting to the facility (Figure 1 and Table 1).

Figure 1 Clinical course timeline (day 0 = abrupt discontinuation of donepezil 10 mg/day at nursing-home admission). Dark-gray bar: Donepezil 10 mg before discontinuation; light-gray bar: Donepezil 5 mg after rechallenge on day 7. Medium-gray bar: As-needed quetiapine 125 mg during the symptomatic/rechallenge period. Grey bars indicate symptom periods reconstructed from the clinical record. Confusion assessment method was positive on day 2 to 3. The timeline is aligned with Table 2; symptom intervals are based on retrospective chart documentation and are not standardized severity measurements. The figure has been converted to grayscale to improve accessibility for readers with color-vision deficiency. Confusion assessment method; PRN: Pro re nata.

Table 1 Day-by-day clinical timeline from the pre-admission decline through 12-week follow-up.

Day	Location	Donepezil/other medication	CAM/cognition	Key symptoms or events	Interpretation
-14 to -1	Home	Donepezil 10 mg/day; amlodipine 5 mg/day	HDS-R 16/30 previously; 8/30 at pre-admission review	Reduced spontaneous speech, withdrawal and increased prompting for ADLs	Pre-existing cognitive/functional decline
0	Nursing-home admission	Donepezil 10 mg/day stopped abruptly; amlodipine continued	No acute CAM assessment documented	Major environmental transition	Start of medication exposure change
1	Nursing home/home	No donepezil	Not documented	No detailed standardized assessment available	Early post-discontinuation period
2	Home	No donepezil	CAM-positive syndrome recognized over day 2-3	Acute agitation, verbal aggression, refusal to sit for meals, insomnia, new gait instability	48-hour temporal association
3	Emergency department	No donepezil	CAM positive; distractible, disorganized thinking	CT and routine laboratory work-up unremarkable; diagnosed clinically as delirium	Major infection/metabolic/structural causes not identified
4-6	Home	No donepezil	Not standardized	Persistent agitation and sleeplessness	Ongoing symptoms before rechallenge
7	Nursing-home review	Donepezil 5 mg/day restarted; quetiapine 125 mg PRN added	No standardized severity scale	Withdrawal suspected; quetiapine used almost daily during first week	Rechallenge, but confounded by rescue medication
8-11	Nursing home	Donepezil 5 mg/day; PRN quetiapine	Not standardized	Agitation and verbal aggression progressively decreased	Improvement during combined rechallenge/adaptation period
12	Nursing home	Donepezil 5 mg/day; quetiapine discontinued	Not standardized	Sleeping through the night	Early clinical improvement
21	Nursing home	Donepezil 5 mg/day	Not standardized	Eating independently and adapting to facility	Functional recovery
84	Nursing home	Donepezil 5 mg/day	No recurrence reported	12-week follow-up without recurrence	Longer-term follow-up

Days are grouped when separate daily documentation was unavailable, to avoid implying precision that was not supported by the retrospective record. CAM: Confusion assessment method; HDS-R: Hasegawa Dementia Scale-Revised; CT: Computed tomography; ADLs: Activities of Daily Living; PRN: Pro re nata.

History of past illness

Hypertension, controlled on amlodipine 5 mg/day (no other cardiovascular comorbidities). No prior stroke, seizures, head trauma, Parkinsonism, rapid eye movement sleep behavior disorder or visual hallucinations. No history of depression, anxiety or other psychiatric disorder. No prior episodes of delirium under anesthesia or during intercurrent illness.

Personal and family history

Non-smoker, no alcohol use. Lives with her daughter and had received visiting home-care support. Family history was negative for early-onset dementia or other neurodegenerative disease.

Physical examination

On emergency department evaluation on day 3, vital signs were normal (body temperature 36.7 °C, blood pressure 132/78 mmHg, heart rate 76 bpm, SpO₂ 96% on room air). The patient was alert but distractible, with psychomotor agitation and intermittent verbal aggression; she met confusion assessment method (CAM) criteria for delirium (acute onset/fluctuating course, inattention and disorganized thinking)[13]. Gait was wide-based and unsteady; no focal neurological deficits and no new rigidity, bradykinesia or tremor were present. No fever, dysuria, cough or newly documented pain was present. The CAM established delirium as a clinical syndrome but did not provide a standardized measure of delirium severity.

Laboratory examinations

Complete blood count, electrolytes, blood urea nitrogen/creatinine, C-reactive protein, liver enzymes, thyroid-stimulating hormone, vitamin B₁₂, folate, ammonia and urinalysis were all within reference ranges. Blood and urine cultures were not obtained (no clinical evidence of infection).

Imaging examinations

Non-contrast head CT performed on day 3 showed only age-appropriate cortical atrophy, with no acute hemorrhage, mass effect, territorial infarct or new vascular lesion. No electroencephalography (EEG) was performed, so non-convulsive seizure activity could not be excluded. Serum cholinesterase or other peripheral cholinergic biomarkers were not measured. No systematic pain assessment was documented, and the complete list of over-the-counter medications and dietary supplements around admission was not systematically recorded. These omissions limit the exclusion of non-drug contributors to agitation, insomnia and gait instability.

MULTIDISCIPLINARY EXPERT CONSULTATION

After donepezil was discontinued, nursing staff involved in the patient's care reported a perceived change in personality. Consequently, the staff discussed the matter and ultimately decided to resume the medication at half the previous dosage.

FINAL DIAGNOSIS

Suspected donepezil withdrawal syndrome in an 89-year-old woman with Alzheimer disease, temporally associated with abrupt cessation of a long-term 10 mg/day regimen at nursing-home admission. The diagnosis is clinical and remains provisional because of pre-existing dementia progression, environmental transition and other unmeasured confounders.

TREATMENT

Donepezil was restarted at 5 mg/day on day 7 as a rechallenge. As-needed quetiapine 12.5 mg was administered during the first week for episodes of agitation with risk of harm to the patient or caregivers. No intravenous or oral benzodiazepines, anticholinergic agents or new opioids were introduced.

OUTCOME AND FOLLOW-UP

Within five days of rechallenge, psychomotor agitation and verbal aggression decreased and the as-needed quetiapine could be discontinued. By day 12 the patient was sleeping through the night, and by day 21 she was eating independently and adapting to the facility. No recurrence was noted during 12 weeks of follow-up while she remained on donepezil 5 mg/day. Because improvement followed both donepezil rechallenge and the period of environmental adaptation, the outcome cannot be attributed to donepezil alone.

DISCUSSION

This case has several features compatible with an acetylcholinesterase-inhibitor withdrawal reaction: Long-term donepezil was abruptly stopped, acute behavioural symptoms emerged within 48 hours, and symptoms improved after re-initiation[6-11]. However, the temporal association and rechallenge response are supportive rather than diagnostic. The Naranjo algorithm was developed for adverse drug reactions rather than withdrawal phenomena[14]. Because several Naranjo items are inherently mismatched to a withdrawal event, we removed the previous numerical score and replaced it with a structured clinical causality assessment (Table 2). The term “suspected” is therefore used to reflect a provisional clinical interpretation rather than a validated probability score.

Table 2 Structured clinical causality assessment for the suspected donepezil withdrawal syndrome.

Domain	Evidence in the present case	Interpretation
Temporality	Donepezil 10 mg/day was stopped on day 0; acute agitation, insomnia, gait instability and CAM-positive delirium emerged on day 2-3	Strong temporal association, but not specific for withdrawal
Dechallenge	Symptoms began after the drug had already been stopped; therefore conventional dechallenge cannot be assessed	Not applicable to the usual Naranjo framework
Rechallenge	Donepezil 5 mg/day was restarted on day 7; agitation and verbal aggression decreased within 5 days	Supportive, but improvement overlapped with quetiapine use and environmental adaptation
Dose relationship	Improvement occurred after restarting at 5 mg/day rather than the previous 10 mg/day dose	Compatible with recovery after re-exposure, but not a validated dose-response finding
Alternative explanations	Pre-admission HDS-R decline, nursing-home relocation, occult pain, unrecorded OTC drugs/supplements, possible non-convulsive seizures, and concomitant quetiapine were not fully excluded	Causal certainty is limited
Objective evidence	CAM-positive delirium was documented; no plasma donepezil concentration or cholinergic biomarker was obtained	Supports the presence of delirium, but does not establish drug causality
Pharmacokinetic plausibility	Donepezil has a long terminal half-life of approximately 70 hours	A 48-hour onset does not require complete drug elimination; an early response to declining exposure is plausible but unproven
Overall assessment	The sequence is compatible with suspected donepezil withdrawal	Clinical causality remains uncertain; the term “suspected” is used to reflect a provisional clinical interpretation rather than a validated probability score

The Naranjo Adverse Drug Reaction Probability Scale was designed for adverse drug reactions rather than withdrawal phenomena. We therefore removed the previous numerical Naranjo score and present a structured clinical causality assessment instead. CAM: Confusion assessment method; HDS-R: Hasegawa Dementia Scale-Revised.

The 48-hour onset also warrants pharmacokinetic consideration. Donepezil has a long terminal elimination half-life of approximately 70 hours, and the half-life may be even longer in older adults[11]. Thus, 48 hours after the last dose, substantial drug exposure would be expected to remain; the clinical syndrome cannot be explained simply by complete elimination of donepezil. Nevertheless, withdrawal symptoms need not coincide with complete elimination. A plausible interpretation is that the abrupt reduction from chronic steady-state exposure produced a clinically relevant change in cholinergic stimulation before the drug was fully cleared. This remains an inference because no plasma donepezil concentration or cholinergic biomarker was measured. The short latency is nevertheless within the range described in previous clinical reports[6-10]. The proposed interpretation should also be distinguished from a demonstrated cholinergic rebound or receptor-adaptation phenomenon, and from withdrawal-related neurotoxicity. In this case, neither receptor-level adaptation nor neurotoxicity was directly assessed; therefore, the mechanism should be regarded as a pharmacological hypothesis rather than an established explanation.

The differential diagnosis of acute mental-status and behavioural change in this setting is broad. Infection-associated delirium was considered less likely because the patient was afebrile and had normal complete blood count, C-reactive protein and urinalysis; major metabolic derangement was not identified; and head CT showed no acute structural cerebrovascular event. No new benzodiazepine, anticholinergic or opioid agent was introduced. However, several alternatives remain incompletely assessed. First, natural progression of Alzheimer disease was already evident before admission, with HDS-R declining from 16/30 to 8/30 over approximately two weeks. Second, nursing-home relocation and an unfamiliar environment coincided with the medication change and could have contributed to agitation, insomnia or delirium. Third, occult pain was not assessed systematically. Fourth, no EEG was obtained, so non-convulsive seizures were not excluded. Fifth, over-the-counter medications and supplements were not comprehensively documented. Finally, quetiapine was administered during the early rechallenge period and may have contributed to subsequent behavioural improvement. These factors prevent attribution of the entire clinical picture to donepezil withdrawal.

Three practical implications follow. First, abrupt discontinuation of long-term donepezil should be considered carefully and, when discontinuation is clinically appropriate, monitored for subsequent cognitive or behavioural deterioration[1,2,12]. Second, rechallenge may provide useful clinical information when withdrawal is suspected, but improvement after rechallenge should be interpreted cautiously when sedating rescue medication is given concurrently. Third, the decision to discontinue an AChE-I should distinguish expected disease progression from potentially reversible deterioration after medication change; the available evidence is heterogeneous and does not support a universal assumption that discontinuation will cause withdrawal symptoms[1,3-5,12].

Previously reported cases and withdrawal studies are summarized in Table 3. Singh and Dudley described two patients with discontinuation syndrome after donepezil cessation[7]. Bidzan and Bidzan described a 62-year-old woman with Alzheimer disease who had taken donepezil for approximately one year; her family physician stopped the drug after attributing low-grade fever and digestive complaints to it, after which dementia symptoms worsened and delirium developed, resolving after donepezil was re-prescribed[8]. Greiman et al[8] reported rapid cognitive and functional decline after abrupt switching and discontinuation in a patient with DLB. Minett et al[9] observed loss of treatment gains and acute cognitive and behavioural deterioration after sudden withdrawal in patients with DLB and PDD. These reports support clinical awareness but are heterogeneous in diagnosis, withdrawal procedure and outcome measurement. Conversely, the prospective institutional study by García-García and Calleja-Hernández[12] found no significant cognitive, behavioural or functional deterioration at 3 months after clinically selected discontinuation in advanced dementia.

Table 3 Selected published reports of cognitive or behavioural deterioration after cholinesterase-inhibitor discontinuation, compared with the present case.

Ref.	Age/sex	Drug/daily dose	Duration before stop	Onset after stop	Outcome after restart/withdrawal
Singh and Dudley[6]	77 male/79 female	Donepezil 10 mg/5 mg	Approximately 1 year/3 years	Within days	Improved within days after donepezil 5 mg
Bidzan and Bidzan[7]	62 female (Alzheimer disease)	Donepezil	Approximately 1 year	After discontinuation; delirium developed	Delirium resolved and mental state stabilized after restart
Greiman et al[8]	83 female (DLB)	Donepezil 10 mg; abrupt switch/discontinuation	Not clearly stated in report	Rapid cognitive/functional decline	Improved after restoration of cholinesterase-inhibitor therapy
Minett et al[9]	DLB/PDD; 19 patients	Donepezil up to 10 mg/day	20 week treatment; 6-week withdrawal	During withdrawal	Loss of cognitive treatment gains and behavioural deterioration; improvement after recommencement
Present case	89 female (Alzheimer disease)	Donepezil 10 mg	> 5 years	2 days	CAM-positive delirium/behavioural syndrome; improvement after 5 mg restart

DLB: Dementia with Lewy bodies; PDD: Parkinson disease dementia; CAM: Confusion assessment method.

The present case has important limitations. First, symptom severity was based on retrospective chart review and was not anchored to prospective standardized instruments such as the Neuropsychiatric Inventory or Cohen-Mansfield Agitation Inventory[15]. Second, gait instability was described subjectively without a standardized gait or fall-risk assessment. Third, delirium was identified using the CAM but was not quantified with a delirium-severity instrument. Fourth, the marked HDS-R decline from 16/30 to 8/30 before admission indicates that natural dementia progression was already occurring. Fifth, nursing-home transition may have contributed to Behavioral and Psychological Symptoms of Dementia or delirium. Sixth, quetiapine was used during the early rechallenge period, so its contribution to clinical improvement cannot be excluded. Seventh, no EEG, serum cholinesterase/cholinergic biomarker measurement or systematic pain assessment was performed, and over-the-counter medications and supplements were not comprehensively documented. These limitations mean that the temporal sequence supports, but does not establish, a causal diagnosis of donepezil withdrawal. Individual daily sleep and eating measures, CAM item scores and formal fall-risk documentation were not prospectively collected; where daily information was unavailable, the timeline uses grouped intervals or explicitly states that the assessment was not documented.

In the context of medication discontinuation during care transitions or end-of-life care, this case illustrates the importance of considering withdrawal among the differential diagnoses when acute behavioural symptoms emerge after stopping an AChE-I[1,12]. Rechallenge may be informative, but concurrent symptomatic treatment, environmental adaptation and pre-existing disease progression can confound interpretation. The case therefore supports careful monitoring rather than definitive attribution of post-discontinuation deterioration to withdrawal.

CONCLUSION

Abrupt discontinuation of long-term, full-dose donepezil was temporally associated with a CAM-positive behavioural syndrome that emerged within 48 hours in an 89-year-old woman with Alzheimer disease. Symptoms improved after rechallenge with donepezil 5 mg/day, but natural dementia progression, nursing-home transition, concomitant quetiapine and other unmeasured factors limit causal certainty. The clinical diagnosis of suspected donepezil withdrawal therefore remains provisional. The 48-hour onset is pharmacokinetically compatible with a clinical response during declining exposure despite substantial residual donepezil because of its long half-life. Clinicians may consider withdrawal in the differential diagnosis after abrupt AChE-I discontinuation while evaluating alternative causes and avoiding unplanned abrupt cessation when clinically feasible.