Donepezil-induced sleep spindle in a patient with dementia with Lewy bodies: a case report


Masaki Nishida MD, Department of Psychiatry, Jichi Medical University, 3311-1 Yakushiji, Shimotsuke, Tochigi 329-0498, Japan. Email:;


Here, we report a case of rapid eye movement sleep behavioural disorder in an elderly patient with dementia with Lewy bodies. Pretreatment polysomnography revealed atonia during rapid eye movement sleep, absence of sleep spindles and loss of slow-wave sleep. Administration of donepezil, an acetylcholinesterase inhibitor, markedly improved delusional symptoms and cognitive function. Pretreatment polysomnography performed after donepezil administration revealed a considerable number of sleep spindles. The effects of cholinergic modulation induced by donepezil seemed to cause remarkable improvement in mental status, incorporating associated with sleep spindles generated by the thalamocortical circuit involved in this patient.


Dementia with Lewy bodies (DLB) is characterized by movement disorders with rapid eye movement (REM) sleep behavioural disorder (RBD) at the onset of symptoms, accompanied by visual hallucinations and cognitive fluctuations. Donepezil, an acetylcholine (Ach) esterase inhibitor, is gaining acceptance for treatment of cognitive impairment, as well as for behavioural and psychiatric complications. Though its effects on psychotic symptoms are known, there have been no reports on the physiological influence of donepezil on sleep architecture in DLB patients. Here we report a DLB patient in whom donepezil increased spindle activity.


An 80-year-old woman presented with a 15-year history of night screaming and a 10-year history of parkinsonism. A neurologist initiated amantadine hydrochloride and levodopa treatment 8 years prior to presentation.

The patient began to display depressive symptoms 3 years before she presented at out clinic. When she attempted suicide, her neurologist consulted us. Her chief complaints were anxiety, depressive mood, memory loss and night screaming, which was thought to be a symptom of RBD. Persecutory delusions and visual and auditory hallucinations were also observed. Quetiapine was initiated at 25 mg daily and then increased to 75 mg daily. Although her depressive symptoms remitted with quetiapine, her psychotic symptoms continued for 2 months. The addition of 2.5-mg olanzapine caused severe drowsiness, and the patient was eventually was hospitalized in the psychiatry department.

On admission, she had persecutory delusions about her family and auditory hallucinations. She screamed loudly every night in the ward. She was prescribed levodopa therapy at her previous dose (300 mg daily).

We performed several diagnostic examinations. Single photon emission computed tomography with the tracer 99mTc-ethyl cysteinate dimer demonstrated remarkably reduced perfusion in the occipital region, though other brain regions retained substantial perfusion. Cardiac uptake of [123I]-metaiodobenzylguanidine showed considerable reduction. The Mini Mental State Examination (MMSE) revealed the deterioration of visual conceptualization, as well as the deterioration of delayed recall. Based on these findings, we diagnosed the case as probable DLB.1

We performed polysomnography (PSG) 9 days after admission to assess the patient's RBD. The patient slept 2 nights from 2100 (lights off) to 0600 (lights on) in the sleep laboratory of the Tokyo Medical and Dental University Hospital (Tokyo, Japan). The first night served as an adaptation night for recording. PSG recording was performed according to standardized techniques, using digital electroencephalography, electromyography and electrooculography.

Sleep epochs on the PSG record were scored according to standard criteria.2 The signals were displayed on a computer monitor and rated visually, epoch by epoch, as non-REM (NREM) sleep stages 1–4, REM sleep, awake, or movement. Slow wave sleep consisted of stage 3 and stage 4 NREM sleep. With respect to spindles, we defined the spindles as rhythmic oscillations ranging between 12 and 16 Hz that last for 0.5–1.5 s.

After the first PSG recording, treatment with donepezil 3 mg was initiated and then increased to 5 mg 1 week later. After taking donepezil, not only did the patient's depressive and psychotic symptoms gradually diminish, but her night screaming also diminished slightly. Follow-up PSG was performed 23 days after donepezil initiation. She was discharged to her family after the second PSG.

The Ethics Committee of Tokyo Medical and Dental University approved this case study, and written informed consent was obtained from the subject.


Table 1 shows the patient's cognitive and sleep changes after donepezil administration. The patient scored 24 on the MMSE and 12 on the Pittsburgh Sleep Quality Index on admission. MMSE improved to 28 after donepezil administration. Her complaints of insomnia diminished, with an improvement from 12 to 9 on the Pittsburgh Sleep Quality Index.

Table 1. Change of mental status and sleep parameters after donepezil
  1. Change of mental status and sleep parameters after donepezil.. Duration of sleep stage in minutes, total sleep time (TST), and each sleep stage percentage are represented. MMSE, Mini Mental Scale Examination; PSQI, Pittsburgh Sleep Quality Index; REM, rapid eye movement sleep; REM latency, latency in minutes to REM sleep onset; SWS, slow wave sleep (Stage 3 and Stage 4); Sleep efficiency, (total sleep time/total time in bed) × 100.

Stage 1452.582.2%256.557.1%
Stage 2325.8%154.534.4%
Sleep efficiency90.9%79.3%
Sleep latency1.54.5
REM latency125.594.5

Figure 1 shows hypnograms of the PSG recordings before and after the administration of donepezil. In the initial PSG recording, REM sleep without atonia (RWA) was detected on PSG, coinciding with the patient's night screaming episodes. On the initial PSG, no sleep spindle was observed throughout the night.

Figure 1.

Comparison of patient hypnograms through donepezil treatment. Upper, hypnogram prior to administrating donepezil. Lower, after administration of donepezil (5 mg daily). REM, rapid eye movement; SWS, slow wave sleep.

In contrast, the follow-up PSG revealed increased sleep spindles in NREM sleep. Prolongation of sleep latency was compared with the initial PSG. The amount of slow wave sleep decreased substantially. The duration of REM sleep did not change significantly, whereas RWA was slightly attenuated compared to the initial PSG.


Previous research has revealed that REM sleep is prolonged by administration of Ach esterase inhibitor in healthy controls.3 Cholinergic neurons in the pedunculopontine (PPT) and lateroposterior tegmental nucleus are critical in the control of REM sleep,4 so it is reasonable to postulate that donepezil would prolong REM sleep by strengthening levels of extracellular Ach. However, the duration of REM sleep did not increase with donepezil, indicating that the neurochemical mechanism that modulates the Ach system in DLB differs from that in normal subjects.

Cholinergic neurons in the PPT regulate anterior horn neurons via gamma-aminobutyric acid in the ventrolateral reticular formation.5 In an animal study, Ach stimulation to the basal forebrain enhanced wakefulness and induced muscle atonia.6 The improvement in this case may be attributed to increasing extracellular Ach, which enhanced muscle atonia and reduced RWA.

In addition to Ach modulation, this case may also indicate the involvement of monoamine neurons. Muscle atonia may also be generated by noradrenaline neuronal suppression of anterior horn neurons via the peri-locus coeruleus-α.7 Monoamine neurotransmitter is reportedly decreased because of α-synuclein aggregation in the brain stem of DLB. It is possible that monoamine neurons, which are usually active during NREM sleep, were indirectly activated by an Ach esterase inhibitor in this case, improving the patient's symptoms.

We diagnosed DLB in this case regardless of the improvement in the patient's MMSE score, because she maintained high function on the job until the onset of disease. Radiological findings and the decline of visual conceptualization reinforced the diagnosis of DLB, in accordance with the diagnosis criteria.

To our knowledge, there have been no reports of PSG recording in DLB cases. In Parkinson's disease, which often manifests with RBD, Gagnon et al. reported that RBD and RWA were frequent in Parkinson's disease, as shown by PSG recordings.8 Moreover, previous studies have failed to report specific findings within NREM sleep among Parkinson's cases. As such, more sleep studies need to be undertaken in neurodegenerative diseases to elucidate neural mechanisms.

This patient insisted that she had experienced subjective improvement in the Pittsburgh Sleep Quality Index because of the administration of donepezil, although there were objectively detectable deteriorations such as worsened sleep efficiency and prolonged sleep latency. This discrepancy may be attributed to the improvement in her depressive symptoms, which may have caused her to underestimate her sleep quality.

Hornung et al. reported that, in addition to REM sleep, donepezil increases stage 2 NREM sleep in healthy elderly subjects.9 Others postulate that the sleep spindle is generated when excitement of the thalamic reticular neurons is transmitted to the cortical pyramidal cells via the thalamocortical neurons.10 The PPT associated with REM sleep initiation projects neurons to a broad area and involves non-specific thalamic nuclei, including a part of the ascending reticular activating system.11 The Ach esterase agent physostigmine increases the sleep spindle,12 as the thalamocortical mechanisms underlying spindle activity are thought to enhance cortical memory integration.13 We hypothesize that augmented Ach tone in PPT caused by donepezil may have induced the sleep spindle in this case.

When cholinesterase inhibitors are used to stimulate central cholinergic transmission, the natural circadian fluctuations of central cholinergic transmission may have a considerable effect. In addition, interfering with nocturnal cholinergic activity is likely to induce memory problems and sleep disorders.14

Enhanced learning function due to increased Ach in the nucleus basalis of Meynert may account for this patient's cognitive improvement. Ach activity during sleep mediates sleep-dependent learning and memory consolidation.12 Recent research has postulated that sleep spindles are a neurophysiological marker of synaptic potentiation,15,16 with corresponding electrophysiological frequencies that are more commonly associated with long-term potentiation than with long-term depression.13 We suggest that sleep spindles play a critical role in memory processing involving cholinergic modulation of the thalamocortical neural circuit.