Improvement in delusions and hallucinations in patients with dementia with Lewy bodies upon administration of yokukansan, a traditional Japanese medicine


Dr Kenji Kosaka MD PhD, Medical Care Court Clinic, 3-23-25 Edanishi, Aoba-ku, Yokohama, Kanagawa 225-0014, Japan. Email:


Background:  This multicentre open-label trial examined the efficacy and safety of the traditional Japanese medicine, or Kampo medicine, yokukansan (YKS), for behavioural and psychological symptoms of dementia (BPSD) in patients with dementia with Lewy bodies.

Methods:  Sixty-three dementia with Lewy bodies patients with probable BPSD (M : W, 30 : 33; mean age, 78.2 ± 5.8 years) were enrolled and treated with YKS for 4 weeks.

Results:  Significant improvements in Neuropsychiatric Inventory scores (mean decrease, 12.5 points; P < 0.001) and Zarit Burden Interview-Japanese edition tests (mean decrease, 3.6 points; P= 0.024) were observed. In patients who consented to an assessment after 2 weeks of treatment, a time-dependent significant improvement was observed in the Neuropsychiatric Inventory score (n= 23; mean decrease, 14.4; P < 0.001), each subscale, including delusions and hallucinations, the Zarit Burden Interview-Japanese edition (n= 22; mean decrease, 8.2; P < 0.01) and the behavioural pathology in Alzheimer's disease insomnia subscale. The Mini-Mental State Examination and the Disability Assessment for Dementia (DAD) showed no significant change. Adverse events were observed in 11 (18%) patients. Three patients (5%) discontinued YKS due to adverse reactions, namely, spasticity and exacerbation of BPSD, edema, and nausea. Hypokalaemia (<3.5 mEq/L) was present in four patients (6%) at the study endpoint. Worsening of extrapyramidal symptoms was not observed.

Conclusion:  YKS improved BPSD in dementia with Lewy bodies patients and caregiver burden scores without deterioration in cognitive function. YKS is useful for the treatment of delusions and hallucinations in BPSD.


Dementia with Lewy bodies (DLB) is the second most common cause of dementia. It is clinically characterized by three symptoms: progressive, fluctuating cognitive decline, visual hallucinations and parkinsonism.1 Behavioural and psychological symptoms of dementia (BPSD), especially vivid, fresh visual hallucinations, are commonly seen in DLB. The hallucinations cause great distress to DLB patients, and BPSD place a significant burden on caregivers and are a major reason for the early institutionalization of dementia patients.2 Moreover, hallucinations are consistently correlated with poor performance of activities of daily living (ADL).3 Pharmacological interventions including antipsychotics and acetylcholine esterase inhibitors have been investigated,4,5 but drug-induced extrapyramidal symptoms (EPS) and other adverse events are not uncommon consequences in DLB patients.

In a single-blinded randomized controlled trial (RCT) in 2005, Iwasaki et al. reported that the traditional Japanese medicine, Kampo medicine, yokukansan (YKS; yi-gan san in Chinese) improved BPSD in dementia patients, including those with DLB, with no severe adverse events.6 They also reported a 15-patient case series on DLB in which hallucinations and other BPSD were successfully improved with YKS treatment.7 For this study, we designed an expanded multicentre case series that investigated the efficacy and safety of YKS in DLB patients. Some of the study results were reported in a letter published in the Journal of the American Geriatrics Society (available online, The efficacy and safety of YKS are described in greater detail in this paper.


Subjects were recruited from 15 hospitals in Japan. Based on data from a previous report, the required sample size was calculated to be 28 with an α-error of 0.05 and power of 0.8.7 At baseline, each patient underwent a uniform evaluation, including a medical history, physical and neurological examination, and brain magnetic resonance imaging. In addition, a Mini-Mental State Examination (MMSE) was performed to assess cognitive function,8 and the Disability Assessment for Dementia (DAD) was used to evaluate ADL.9 BPSD was evaluated using the Neuropsychiatric Inventory (NPI),10 and the burden on caregivers was evaluated with the Japanese version of the Zarit Burden Interview (J-ZBI).11 Both the NPI and J-ZBI were conducted three times, at baseline, after 2 weeks of treatment and at the end of the study (after 4 weeks of treatment). Other scales and blood examinations including serum potassium concentration, were conducted at baseline and at study termination.

Based upon the study inclusion and exclusion criteria (Table 1), 63 DLB patients (30 men and 33 women; mean age, 78.2 ± 5.8 years) were enrolled. Written informed consent was obtained from participants or their families after a detailed explanation of the study was provided. Fluctuations in cognitive function were evaluated with a modified Neuropsychiatric Inventory (NPI-11).15 Sleep disturbances were evaluated with a subscale of behavioural pathology in Alzheimer's disease.16 All changes in outcome measures between baseline and measurement points were compared using a signed rank-sum test. EPS were assessed using the Barnes Akathisia Scale.

Table 1. Entry criteria
  1. PD, PDD and DLB should be understood within the spectrum of Lewy body disease.13 The DLB Consortium and the DLB/PDD Working Group have indicated that PDD and DLB may be understood within the spectrum of Lewy body disease.12,14 In the consortium on DLB guidelines, revised in 2005, the term ‘Lewy body disease’ was used as the generic term including DLB, PD and PDD.12 Therefore, patients with probable DLB or PDD were included in our study. BPSD, behavioural and psychological symptoms of dementia; DLB, dementia with Lewy bodies; NPI, Neuropsychiatric Inventory; PD, Parkinson's disease; PDD, Parkinson's disease with dementia; YKS, yokukansan.

Inclusion criteria
(1) Diagnosed as probable DLB according to the international consensus criteria.1,12
(2) Concomitant BPSD with NPI score ≥4 on at least one of the subscales in the NPI.
(3) Serum potassium concentration (sK) measured within the previous 6 weeks was within the normal limit for each hospital. If not yet measured, the patient voluntarily agreed to have sK measured.
(4) Age from 55 to 90 years, and of either sex.
(5) If an outpatient, family members were able to take care of the patient.
(6) If an inpatient, the patient must have been admitted for more than 2 weeks, and a caregiver for the patient able to be arranged.
Exclusion criteria
(1) Major medical illness such as neoplastic disease, acute inflammation, or any other disease that would be likely to prevent completion of this study.
(2) Another type of dementia (such as Alzheimer's disease, vascular dementia, frontotemporal dementia) as the principal diagnosis.
(3) Psychosis due to other disorders such as schizophrenia and/or depression.
(4) Delirium due to drugs, alcohol, metabolic intoxication or inflammation.
(5) Patients unable to take YKS orally.
(6) Use of neuroleptics, anti-anxiety drugs, anti-epileptic drugs, antidepressants or herbal remedies.
(7) Patients on antiparkinsonian drugs, hypnotics and/or donepezil hydrochloride, where the dosage of these medicine had been modified in the past 4 weeks. Participants who had taken these medications continuously for more than 4 weeks were not excluded.
(8) Patients whose attending doctors prohibited participation.

The occurrence of any adverse events was carefully noted and incidence rates were calculated. Patients who withdrew from YKS treatment and/or observation were considered to have dropped out. Results were expressed as mean ± SD. P < 0.05 was considered to be statistically significant, except for the NPI, for which multiple observations were adjusted by using the Bonferroni method. All statistical analyses were performed by a trained statistician using SAS 9.1.3 (SAS Institute Inc., Cary, NC, USA).

YKS extract was provided by Tsumura & Co. (Tokyo, Japan). The preparation method has already been described in previous reports.6,7 YKS contains a mixture of dried herbs, 4-g Atractylodis lanceae rhizoma, 4-g Poria, 3-g Cnidii rhizoma, 3-g Angelicae radix, 2-g Bupleuri radix, 1.5-g Glycyrrhizae radix, and 3-g Uncariae uncis cum ramulus. These herbs are registered in the Pharmacopoeia of Japan v. 15 (available online, Each participant received 2.5-g YKS powder (1.08-g extract) three times a day for 4 weeks. The production and supply processes of YKS comply with Good Manufacturing Practices for Kampo products and have been approved by the Japanese Ministry of Health, Labour and Welfare.

This study was carried out in compliance with the ethical principles embodied in the Helsinki Declaration (2000). Written informed consent was obtained from each patient (if capable of giving consent) or the patient's proxy consent provider prior to participation in this study. The study protocol was approved by the institutional review board of each centre and registered with the University Hospital Medical Information Network clinical trial registry (UMIN000001511,


Of 63 registered participants, three did not receive YKS treatment because of the onset of cholecystitis, an injury, and the taste of YKS, respectively. These three patients were excluded from any analysis. One patient did not come back to the clinic after the first interview, and we could not make contact with him. After the end of treatment, it was discovered that five patients were using drugs prohibited under the exclusion criteria. These six patients were excluded from the efficacy analysis but included in the safety analysis. Therefore, efficacy was analyzed in 54 participants while safety was analyzed in 60. Three patients (11%) withdrew because of adverse events and one patient withdrew of his own volition. In total, 54 participants completed 4 weeks of treatment. The number of participants with complete data for each scale was 52 for NPI, Hoehn and Yahr score and DAD, 53 for MMSE, and 51 for J-ZBI.

Significant improvements were observed in NPI total scores (mean decrease, 12.5 points; P < 0.001) (Table 2, Fig. 1a) and the J-ZBI (mean decrease, 3.6 points; P= 0.024) (Table 2, Fig. 2a).

Table 2. Symptomatology according to rating scale scores at weeks 0, 2 and 4
Rating scale  n Mean ± SDDifference P-value n Mean ± SDDifference P-value
  1. P < 0.005, Baseline vs Endpoint, signed rank-sum test, adjusted significance level by Bonferroni correction for NPI subscales. DAD, Disability Assessment for Dementia; J-ZBI, Zarit Burden Interview-Japanese edition; MMSE, Mini-Mental State Examination; NPI, Neuropsychiatric Inventory.

NPI totalWeek 05430.5 ± 18.52324.8 ± 13.4
Week 22313.8 ± 10.5−11.00.0000
Week 45216.9 ± 16.2−12.50.00002310.4 ± 13.4−14.40.0000
NPI delusionsWeek 0397.5 ± 3.3177.3 ± 2.9
Week 2172.9 ± 2.0−4.40.0000
Week 4374.4 ± 3.8−3.10.0000172.4 ± 2.7−4.90.0000
NPI hallucinationsWeek 0518.0 ± 3.5227.4 ± 3.7
Week 2223.1 ± 3.2−4.30.0000
Week 4493.9 ± 3.5−4.00.0000221.9 ± 1.9−5.50.0000
NPI agitation/aggressionWeek 0295.4 ± 3.564.8 ± 3.0
Week 262.3 ± 2.3−2.50.2813
Week 4273.0 ± 2.7−2.40.005663.5 ± 3.6−1.30.7500
NPI dysphoriaWeek 0303.8 ± 2.6174.0 ± 3.0
Week 2173.1 ± 3.0−0.90.1802
Week 4291.5 ± 1.5−2.30.0000171.5 ± 1.7−2.50.0032
NPI anxietyWeek 0414.3 ± 3.4194.0 ± 3.1
Week 2192.8 ± 2.2−1.20.0976
Week 4392.6 ± 3.1−1.70.0009192.2 ± 3.0−1.80.0266
NPI euphoriaWeek 082.8 ± 2.450.8 ± 1.3
Week 251.2 ±
Week 471.1 ± 1.1−1.40.125050.8 ± 1.301.000
NPI apathyWeek 0315.3 ± 2.9153.1 ± 2.6
Week 2153.2 ±
Week 4303.8 ± 3.3−1.40.0241152.8 ± 2.7−0.30.7480
NPI disinhibitionWeek 0124.7 ± 3.122.0 ± 2.8
Week 220.0 ± 0.0−2.01.000
Week 4102.4 ± 3.1−2.10.031321.0 ± 1.4−1.01.000
NPI irritability/labilityWeek 0254.9 ± 3.6112.8 ± 1.6
Week 2111.3 ± 1.4−1.50.1094
Week 4231.6 ± 2.3−2.90.0005111.0 ± 2.4−1.80.0977
NPI aberrant motor activityWeek 0255.4 ± 3.5102.4 ± 2.3
Week 2101.0 ± 1.2−1.40.0938
Week 4233.5 ± 3.9−1.80.0166101.1 ± 1.6−1.30.3594
MMSEWeek 05418.0 ± 7.02319.3 ± 6.4
Week 45319.2 ± ±
DAD (%)Week 05356.0 ± 31.62361.2 ± 27.8
Week 45257.4 ± ±
J-ZBIWeek 05233.1 ± 17.92233.1 ± 17.1
Week 22225.0 ± 17.8−8.00.0004
Week 45129.8 ± 18.7−3.60.02442224.9 ± 16.8−8.20.0053
Serum potassium level (mEq/L)Week 0554.2 ± 0.4234.2 ± 0.3
Week 4503.9 ± 0.5−0.20.0084223.9 ± 0.4−0.20.0049
Figure 1.

Changes in Neuropsychiatric Inventory (NPI) total scores in dementia with Lewy bodies patients. (a) Fifty-two participants were evaluated by NPI. The NPI was administered twice, at week 0 and after week 4. (b) Twenty-three of 52 patients consented to be assessed three times, at week 0 and after weeks 2 and 4. ***P < 0.001, signed rank-sum test.

Figure 2.

Changes in Zarit Burden Interview-Japanese edition (J-ZBI) total scores in dementia with Lewy bodies patients. (a) Fifty-one participants were evaluated by J-ZBI. The J-ZBI was administered twice, at week 0 and after week 4. (b) Twenty-two of 51 patients consented to be assessed three times, at week 0 and after weeks 2 and 4. ***P < 0.001, **P < 0.01, *P < 0.05, signed rank-sum test.

In participants who consented to be assessed three times, NPI scores (n= 23, a mean decrease of 14.4, P < 0.001), J-ZBI (n= 22; mean decrease, 8.2; P < 0.01) and the behavioural pathology in Alzheimer's disease insomnia subscale (P= 0.011) were significantly improved after 2 weeks of treatment and at the end of the study (Table 2, Figs 1b,2b). NPI subscales, including delusions, hallucinations and dysphoria, were also significantly improved at the end of the study (Table 2). MMSE (mean increase, 0.7 points; P= 0.124) and DAD (mean increase, 3.3 points; P= 0.182) showed no significant changes. EPS listed in the Barnes Akathisia Scale were not observed. The serum potassium concentration decreased slightly but significantly from 4.18 ± 0.29 mEq/L to 3.93 ± 0.44 mEq/L. Four patients showed hypokalaemia levels lower than 3.5 mEq/L at the end of study. Spasticity and worsening of BPSD were encountered in a single patient who recovered after discontinuation of YKS. Worsening hypotension was recorded in one patient. Other adverse events were oedema, gastrointestinal dysfunction, spasticity, and exacerbation of delusions and hallucinations.


Behavioural and psychological symptoms have a serious impact on the quality of life of dementia patients, as well as on their caregivers. Recent trials have demonstrated that acetylcholine esterase inhibitors might be somewhat beneficial but not completely effective.15 Use of neuroleptics is also common in treating patients with BPSD, but they have the drawback of being associated with EPS in elderly patients.5

Though the present study was only a preliminary case series, it demonstrates a significant improvement of BPSD, especially delusions and hallucinations, in DLB patients given YKS. The burden imposed on caregivers was also reduced. In evaluation of DAD, ADL were not significantly changed. In our short-term observation, MMSE slightly increased. No other report has found cognitive improvement with YKS in dementia patients. The MMSE change in the present study might be attributed to the cognitive fluctuation that is characteristic of DLB. Nevertheless, our data suggest that YKS did not have a deleterious effect on cognitive function, a finding that is in agreement with previous YKS studies. Because DLB is associated with cognitive fluctuations, the clinical significance of any change in MMSE needs to be evaluated carefully.

In the present study a single case showed exacerbation of BPSD, but previous studies have not reported this in association with YKS. Hypokalaemia due to Glycyrrhizae radix should not be ignored, and serum potassium concentration should be monitored during YKS treatment.17 Although these adverse events were observed, the dropout rate was far less than with the neuroleptics formerly administered for dementia.18–20 In the present study, EPS were not encountered. There have been no reports of EPS with the use of YKS in previous studies.6,7 The efficacy of YKS for BPSD, especially for hallucinations, is superior to cholinesterase inhibitors.15,21,22 In the present study, 10 participants had already used donepezil, but their BPSD had not been sufficiently controlled. YKS was effective in these patients.

The mechanisms whereby YKS has an effect on psychosomatic and neurological symptoms have been previously published. YKS inhibits the 2,5-dimethoxy-4-iodoamphetamine-induced head-twitch response, decreases expression of 5-hydroxytryptamine 2A receptors in the prefrontal cortex,23 possesses a 5-hydroxytryptamine 1A partial agonistic effect,24 and has an inhibitory effect on glutamate-mediated excitotoxicity.25,26 YKS may affect these neurotransmitters and receptors in a multifaceted manner. The herbal ingredients and their components have the following pharmacological effects, which may be responsible for the clinical effects seen with the administration of YKS. An aqueous extract of the hooks and stems of Uncaria sinensis (Oliv.) Havil., Uncariae uncus cum ramulus, one of the herbs in YKS, protects against glutamate-induced neuronal death in cultured cerebellar granule cells,27 and oxyindole alkaloids, such as isorhynchophylline, isocorynoxeine and rhynchophylline, and indole alkaloids, such as hirsuteine and hirsutine, are the active components of Uncariae.28 Rhynchophylline and isorhynchophylline show antagonistic effects on N-methyl-D-aspartate receptors.29 Geissoschizine methyl ether, corynantheine and dihydrocorynantheine obtained from Uncariae uncis cum ramulus were found to be partial agonists for 5-hydroxytryptamine receptors.30 Glycyrrhizin, one of the main components of Glycyrrhizae radix, and its metabolite, 18 beta-glycyrrhetinic acid, are likely responsible for amelioration of dysfunction of glutamate transport in astrocytes.31

The present study demonstrates that YKS improved BPSD in DLB patients and reduced the burden on caregivers without compromising cognitive function or ADL. Although some adverse events were observed, the dropout rate was low and no worsening of EPS was observed.

There were some limitations to the present study. First, it was not an RCT. Because there are fluctuations in DLB, the results of a study conducted in a single group might have been biased by the spontaneous healing process. With the sociocultural background in Japan, it was difficult to carry out an RCT on DLB patients. Since the original case series reported in 2005, YKS has been administered widely within Japan to DLB patients with appropriate indications (e.g. insomnia).6 Additional evidence on YKS is needed as soon as possible, and expanded RCT should be performed before this treatment is officially introduced as a medical treatment overseas. Second, YKS placebo is now nearly ready for use, but it was not available at the start of our trial. Though the present study is a preliminary one, it supports the initiation of an expanded RCT to evaluate the effects and safety of YKS as a candidate for the treatment of DLB.


This study was financially supported by Tsumura & Co., the pharmaceutical company that provided YKS.