Prevalence of Alzheimer's disease, vascular dementia and dementia with Lewy bodies in a Japanese population


Correspondence address: DrTatsuoYamada Department of Internal Medicine and Health Care, Fukuoka University, School of Medicine, 7-45-1 Nanakuma, Johnan-ku, Fukuoka 814-0180, Japan. Email:


Abstract We studied the prevalence of dementing disorders in a rural town of Japan (Amino-cho), using a door-to-door two-phase design. Of the 170 persons screened as having cognitive impairment, 142 cases were diagnosed as having dementia. The prevalence (cases/100 aged 65 years older) was 3.8 for all types of dementia, 2.1 for Alzheimer's disease (AD), 1.0 for vascular dementia (VD) and 0.7 for other types of dementia. Among other types of dementia, there were four male patients with dementia with Lewy bodies (prevalence: 0.1), but no patients with frontotemporal lobar degeneration. The overall prevalence was higher in women for AD, while that of VD was the same in both sexes. With results similar to many previous studies in Western countries and some recent surveys in Japan, the present sudy clearly showed that AD is more prevalent than VD.


Research on differences in the rates of dementia in various populations is one approach to the identification of risk factors. In the past, several surveys have indicated that 1–13% of persons over 65 years have a dementing illness.1,2 The prevalences of two major subtypes of dementia, Alzheimer's disease (AD) and vascular dementia (VD), have been reported in various nationalities. Alzheimer's disease is the major subtype in many Western countries, while VD has often been reported to be dominant in Japan.3–8 However, some recent epidemiological studies in Japan found the tendency towards more AD than VD.9–11 In the past, diagnosis criteria and classification methods have not been well standardized, so that the different prevalence ratios reported may reflect methodological differences.

Dementia with Lewy bodies (DLB) is increasingly being recognized as a common cause of dementia in elderly people. It is now considered as probably the second most common neurodegenerative demented disorder.12,13 To date, no epidemiological data have been reported for the prevalence rate of DLB. We investigated the prevalence of dementing disorders in a rural area in Japan by door-to-door survey, and found a higher AD/VD ratio than reported in earlier studies. We also estimated the prevalence of DLB as 0.1.


The epidemiological study was done in 1998 for the entire population of Amino-cho in Kyoto prefecture, Japan. Public health nurses, working as the permanent care provider, had kept detailed information about the physical and mental health of the entire town for many years. To be included in the study, subjects were required to be living and to legally reside in the town on the prevalence day, 1 January 1998. The population of Amino-cho in 1998 was 16 765 (8076 males and 8689 females). The number of elderly people over 65 years old was 3715 (1503 males and 2212 females), or 22.2% of the total population. We examined the prevalence rate of dementia in the elderly.

In phase 1, a brief screening examination for all the people over 65 years old were administered by public health nurses in the town. The screening examination included an interview for both subjects and their family which surveyed activities of daily living, psychological and medical symptoms, and the medical history of the subject. In phase 2, all cases who showed cognitive impairment in phase 1 were examined by neurologists. An assessment of medical history, a physical examination, a comprehensive cognitive evaluation using the Mini-Mental State Examination (MMSE), and routine laboratory tests were done. Brain computed tomographs or magnetic resonance images (MRI) obtained in several hospitals were also used for diagnosis. The diagnosis of dementia was defined using DSM-III-R criteria.14 Patients who satisfied the criteria recommended by the National Institute of Neurological and Communication Disorders and Stroke–Alzheimer's disease and Related Disorders Association (NINCDS– ARDRA)15 were diagnosed as having probable AD. Patients who satisfied the criteria recommended by the National Institute of Neurological Disorders and Stroke–Association Internationale pour la Rech-erche et l'Enseignement en Neuroscience (NINDS– AIREN)16 were diagnosed as having probable VD. Patients who fulfilled the criteria by the consensus guideline for the clinical diagnosis of probable and possible DLB12 were diagnosed as having probable DLB. Patients who fulfilled the criteria of frontotemporal lobar degeneration (FTLD)17 were diagnosed as having FTLD. We excluded cases of cognitive decline secondary to depression and other psychiatric diseases through a psychiatric interview and medical history. For subjects who could not be examined directly, we collected information from family members, the public nurse, and physicians in the town. Prevalence was calculated for dementia of all types and for specific dementing disorders.


The study population included 3715 subjects aged 65–99 years old residing in Amino-cho on the prevalence day. Phase 1 study done for about 1 month (January 1998) picked up 170 cases as having cognitive impairment. The number of subjects who were diagnosed as having dementia by phase 2 study was 142, yielding a prevalence for all dementias of 3.8 cases/100. The age-specific prevalence of dementia increased relatively steeply with advancing age for women (Table 1). However, for men, the incidence between 90–94 years dropped for a reason that is unknown (Table 1). Particularly high incidence was shown in those over 95 years (Table 1).

Table 1.  Age- and sex-specific prevalence (cases/100) of dementia of all types, and overall sex-specific prevalence by type of dementia in Amino-cho, Japan, 1 January 1998
 MenWomenBoth sexes
at risk
at risk
at risk
Dementia of all types
65–69 years45770.755890.8911660.8
70–74 years94082.245310.8139391.4
75–79 years71953.6124412.7196363.0
80–84 years91655.5253167.9344817.1
85–89 years1210611.32623311.23833911.2
90–94 years2464.3199120.92113715.3
95–99 years3650.051145.581747.1
65–99 years2015031.35922132.77937152.1
65–99 years1515031.02322131.03837151.0
Other dementia
65–99 years1115030.71422130.62537150.7

The overall prevalence was 2.1 for AD, 1.0 for VD and 0.7 for other dementias (Table 1). The prevalence of AD was two times higher in women than in men, while that of VD was the same in both sexes. Among other dementia (Table 1), seven patients were diagnosed as having mixed dementia and four were diagnosed as having DLB. All the patients with DLB were men. Two patients with normal pressure hydrocephalus, three patients with Parkinson's disease and one patient with progressive supranuclear palsy were also included in the category of other dementias. Six patients were not able to be diagnosed clearly and were classified as unknown dementia. We found no patient who met the criteria for FTLD.

The severity of dementia measured by MMSE is shown in Fig. 1. In AD, the greatest number of patients were at a moderate stage (MMSE score: 10–19). In VD, however, the highest number of patients had severe symptoms.

Figure 1.

Score of Mini-Mental State Examination in each type of dementia (bsl00023, AD (n = 52); ▪, VD (n = 27); □, mixed (n = 7).


This study showed that the prevalence of all types of dementia in the elderly population over 65 years was 3.8 in a rural community of Japan. As shown in Table 2, this prevalence rate was similar to previous reports in small Japanese communities, but somewhat lower than that found in most of these surveys.5,6,8,9,18 Although it is generally believed that overall dementia rates are generally similar among nations, many previous papers suggested significant regional differences in type, with the Japanese, Chinese and Russian studies reporting higher rates of VD than AD, while studies in most Western nations showed either no significant difference or significantly higher rates of AD than VD.4 Most Japanese surveys done in last years have reported AD/VD ratios of less than 1.18 However, several recent studies in Japan showed higher ratios of AD/VD, similar to those in Western countries.9,11 In Daisen-cho, the frequency of AD was lower than that of VD in 1980,8 but in 1990 it was higher than that of VD.9 The Hisayama study in 1985 showed a AD/VD ratio of 0.47,6 but this had risen to 0.8 in 1992.19 Furthermore, a very recent study in an extremely rural community in Japan found a ratio of 1.2 with very high prevalence (8.5).11 The present study also revealed that the ratio was 2.1, the highest among previous reports, but very similar to that found for Kaokaoping in southern Taiwan.20 Also in a study for older Japanese-American men in Hawaii, the ratio was 1.5.21 As described by Kosaka et al.,22 this evidence also suggests that AD is the most predominant type of dementia in Japan.

Table 2.  Prevalence of dementia and AD/VD ratio in the population aged over 65 years of smaller areas in Japan, Hawaii (Japanese American) and Taiwan
Survey yearSitePopulation at riskPrevalence (%)AD (%)VD (%)AD/VD ratioReference
  • * 

    Subjects aged 71 to 93 years, NA: not available.

1998This study37153.855.626.82.1

In the 7 year follow-up Hisayama study, the recent prevalence of VD and the incidence of stroke was found to have decreased in men, which suggests that the incidence of VD may decrease as the risk of stroke is decreased.19 Prevalence estimates of AD rise exponentially with age.7,18 In Japan, the average life expectancy is securely prolonged. The increasing AD/VD ratio might be due to the above-mentioned reasons.

According to clinicopathological studies, DLB now ranks second to AD as a cause of dementia among the elderly.12,13 In the present study, we diagnosed only four males as having DLB, giving a prevalence of 0.1. The male predominance of DLB is consistent with earlier reports.23 Possible inclusion of DLB cases in clinically diagnosed AD, VD and Parkinson's disease has been suspected by Kosaka.24 Therefore, development of some methods, such as those using biological markers, for accurate clinical diagnosis of DLB is needed in such a community-based study.


This work was supported in part by Grants-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan.