Abstract We evaluated dementia symptoms to clarify the character of dementia with Alzheimer's disease (AD) observed in the oldest old patients and that of dementia with early-onset AD. Subjects were consecutive AD inpatients admitted for the first time at age of 90 years and over because of behavioral symptoms (demented nonagenarian group: D90G; n = 18) and those with 24 consecutive inpatients with AD with early-onset (EOG). The Gottfries, Brane and Steen's scale and the Dementia Behavior Disturbance scale were used to evaluate the symptoms and troublesome behaviors. The scores of these scales in D90G and in EOG were compared with those of 26 sex distribution-, severity of dementia-, and disease duration-matched inpatients with AD with late-onset (LOG). Compared with LOG, wakefulness was more impaired and waking up at night was more frequent in D90G, while memory, orientation and inappropriate behaviors were more severe in EOG. These results suggest that the clinical features of dementia in EOG were quantitatively different from those of LOG. In contrast, the clinical feature of dementia of D90G were sleep–wake pattern disturbance and were qualitatively different from those of LOG.
In developed countries including Japan, the numbers of eldery are increasing, especially the oldest old people (85 years and over, or over 90 years and over). Dementia in this generation is important because age is considered to be one of the major risk factors in the dementing process.1 However, there were few studies concerning the clinical features of dementia in the oldest old. In this study, we evaluated the first episodes of behavioral symptoms in demented inpatients with Alzheimer's disease (AD) who were admitted to the psychogeriatric ward at nonagenarian and those with early-onset AD. We compared the behavioral and clinical symptoms with those in inpatients with late-onset AD. The aim of this study was to elucidate the clinical features of dementia in very old people.
SUBJECTS AND METHODS
Subjects were consecutive inpatients with AD according to the criteria of ICD-10 and had the diagnostic criteria for probable or possible AD as developed by the working group of the National Institute of Neurological and Communicative Disorders and Stroke in collaboration with the Alzheimer's Disease and Related Disorders Association. They also were admitted to psychogeriatoric ward of National Shimofusa Hospital, Chiba-City, Chiba Prefecture, Japan from April 1993 to March 2000 due to first episodes of behavioral and psychological symptoms of dementia (BPSD). They had not all yet been admitted to psychiatric wards. Patients with other psychiatric diagnosis prior to the onset of dementia and with a history of cerebral episodes were excluded from the study. We also excluded patients with an active physical illness at the time of study entry and those with a history of alcohol or drug abuse.
All medications except psychotropic medications were administrated steadily over 6 months at the time of study entry.
Eligible inpatients were 18 inpatients with AD who were admitted to our hospital because of BPSD for the first time at age of 90 and over regardless of the onset age (demented nonagenarian group, D90G; onset age in D90G was 65 and over), and 24 inpatients with AD with early onset (early onset group, EOG; onset age was under 64). We compared the symptoms in D90G and in EOG with 26 sex distribution-, severity of dementia-, and disease duration-matched inpatients with AD with late onset (late onset group (LOG); that is, senile dementia of Alzheimer's type, onset age was 65 and over), who were also admitted for first time episode of BPSD during same period. Table 1 shows the demographic data in three subgroups. Written informed consents were obtained from their guardians, and if possible from the patients themselves. This study was approved by the Ethical Committee at the National Shimofusa Hospital, Chiba, Japan.
Table 1. Demographic data of three subgroups (demented nonagenarian group (D90G), Alzheimer's disease with early onset group (EOG) and Alzheimer's disease with late-onset group (LOG))
CDR, Clinical Dementia Rating; HDSR, revised version of Hasegawa dementia scale.
Male : female
Mean age at onset (years)
Mean duration of illness (years)
Mean HDSR score
MEAN CDR score
Mean at age of admission (years)
Assessment of clinical symptoms and behavioral symptoms
The GBS scale2 was used to evaluate the symptoms of dementia and the Dementia Behavior Disturbance (DBD) scale3 to evaluate the troublesome behaviors.
We also assessed demographic data, the severity of dementia using CDR and cognitive functions using the revised version of Hasegawa dementia scale (HDSR).
The comparisons of each item of the GBS and the DBD scales among three subgroups were performed using the Kruskal-Wallis test. Since the GBS scale is composed of 26 items while the DBD scale 28 items, Bonfferoni's correction was used for the correction of multiple comparison (corrected α was 0.0019 in GBS scales and 0.0018 in DBD scales). If these were statistically significant differences, the comparisons of these items of the GBS and the DBD scales between EOG and LOG and between 90DG and LOG were performed using Mann–Whitney U-test. In these cases P < 0.05 was accepted as a statistically significant value. The data were analyzed using the statistical software package SSPS-J (Tokyo, Japan).
With respect to the GBS scale the following were significantly different among three subgroups (α = 0.0019): ‘impaired orientation in space’, ‘impaired personal orientation’, ‘impaired distant memory’, ‘impaired wakefulness’ and ‘confusion’ (Table 2).
Table 2. Comparisons of GBS (Gottfries, Brane and Steen’s) scales among three subgroups (demented nonagenarian group (D90G), Alzheimer's disease with early-onset group (EOG) and Alzheimer's disease with late-onset group (LOG))
Of these five items, ‘impaired wakefulness’ and ‘confusion’ were more prominent in D90G than the were in LOG and ‘impaired distant memory’ was less severe in D90G than in LOG. ‘Impaired orientation in space’, ‘impaired personal orientation’ and ‘impaired distant memory’ were more remarkable in EOG than in LOG (Table 2).
With regard to the DBD scale the following were significantly different among three subgroups (α = 0.0018): ‘loses, misplaces, or hides things’, ‘paces up and down’, ‘dresses inappropriately’, ‘hoards things for no obvious reason’, ‘empties drawers or closets’, ‘gets lost outside’, ‘wanders aimlessly outside or in the house during the day’, ‘wakes up at night for no obvious reason’ and ‘wanders in the house at night’ (Table 3). Although there was a statistically significant difference among three subgroups, the score of ‘dresses inappropriately’ was low. This item was therefore omitted.
Table 3. Comparisons of dementia behavioral disturbance scale (DBD scale) among three subgroups (demented nonagenarian group (D90G), Alzheimer's disease with early-onset group (EOG) and Alzheimer's disease with late-onset group (LOG))
In D90G, ‘wakes up at night for no obvious reason’ and ‘wanders in the house at night’ were more frequent than in LOG. In contrast, the following were less frequent in D90G than in LOG: ‘loses, misplaces, or hides things’, ‘paces up and down’, ‘hoards things for no obvious reason’, ‘empties drawers or closets’, ‘gets lost outside’ and ‘wanders aimlessly outside or in the house during the day’ (Table 3).
In EOG, the following were more frequent than in LOG: ‘loses, misplaces, or hides things’, ‘wakes up at night for no obvious reason’ and ‘wanders in the house at night’ (Table 3).
There was no significant difference of HDSR score and CDR score among the three subgroups, although the mean HDSR score was lower in D90G than those in LOG and EOG.
The oldest old population is remarkably increasing in number. In fact, there were 18 AD inpatients at the age of 90 years and over admitted to the psychogeriatoric ward of the National Shimofusa Hospital for the first time from April 1993 to March 2000 because of BPSD. This meant that about two patients in the oldest old age per year were admitted to psychogeriatoric ward due to first episode BPSD. Therefore, the number of demented inpatients of nonagenarians were about 75% those of demented inpatients with AD with early onset. In the near future, BPSD in demented nonagenarians will be as serious as that in demented inpatients with AD with early onset. Therefore, the problem of dementia in this generation is becoming more and more important. However, there were few studies concerning the clinical features of dementia in the oldest old. The features and the treatments of dementia in very old people should be different from those in younger people if there are any clinical features specific to dementia in very old people. In order to examine this hypothesis, the clinical features of dementia in the oldest old (over 90 years) were investigated.
When they showed as first episodes of BPSD, memory disturbance and disorientation in EOG were more severe than those in LOG. These symptoms are supposed to relate to the intellectual deteriorations in AD. Moreover, wandering and pacing, hoarding and emptying drawers or closets were more frequent in EOG when they showed the first episode of BPSD. These symptoms are thought to be included in wandering behaviors in demented and concomitant troublesome behaviors in demented with ad.4–6 This indicates that as the first troublesome episode, BPSD related to AD pathology were more severe in EOG than in LOG. Therefore the symptoms related to AD pathology might be more pronounced in EOG quantitatively than in LOG when they are shown at the first time.
In contrast, ‘impaired distant memory’ was less severe in D90G than in LOG. ‘Impaired wakefulness’ and ‘confusion’ were more severe and ‘waking up and wandering at night’ was also more frequent as the first episode in D90G than in LOG. Our results are probably compatible with those of Powell.7,8 He suggested that common disturbed pattern of dementia in the oldest old age consisted of preserved awareness of environment, normal participation in conversations and memory impairment with diminished ability to acquire new information. We could contact patients in D90G but not those in LOG although mean HDSR score was lower. This suggests that patients in D90G reserve better cognitive functions than those in LOG. Our results are also compatible with those of Rahkonen et al.9 They reported that delirium was frequently observed in older people especially in the oldest old. In our study, ‘impaired wakefulness’ and ‘confusion’ were more severe and ‘waking up and wandering at night’ was also more frequent in D90G than in LOG. Moreover, the score of ‘sleeps excessively during the day’ was relatively high in D90G and this difference was significant different without Bonfferoni's corrections (P < 0.05). These symptoms seem to have arisen from the sleep–wake pattern having some resemblance to delirium. Our results are partly compatible with those of Furuta et al.10 They reported that behavioral abnormalities, such as excitement, delirium, reversed diurnal rhythm, and wandering were more frequently in very old patients with AD (>85 years old). We showed that ‘paces up and down’, ‘gets lost outside’ and ‘wanders aimlessly outside or in the house during the day’ were less frequent in D90G than in LOG. Mean CDR in D90G in our study was 2.2 (moderate to severe) and mean FAST (Functional Assessment Staging) score in very old patients with AD in their study was 4.77 (mild to moderate). Therefore, this discrepancy might be due to different severity in study patients and we thought the first BPSD episode in D90G was qualitatively different from that in LOG. Moreover, although ‘impaired distant memory’ and ‘impaired personal orientation’ were less in D90G than in LOG, the CDR score was similar in both D90G and LOG. These differences were seen not only in cognitive functions but also in behavioral and psychological symptoms. However, all subjects in our study showed recent memory disturbances and because of intellectual disturbances their ADL functions were disturbed. Therefore, we diagnosed patients in D90G as AD clinically.
What make these differences? In our hospital, seven out of 10 subjects who reached to the neuropathological diagnosis were diagnosed as AD (unpubl. data). This meant that the main pathology in dementia of the oldest old was Alzheimer's type pathology. However, Ravaglia et al. suggested that the aging brain might be susceptible to multiple additive factors (e.g. vascular factor), which could affect cognitive functions.11 We considered that aging, vascular factor as pointed by Ravaglia et al.,11 neuropathological heterogeneity might be one of these factors, which caused different clinical features in the oldest old demented patients with AD.
Jorm et al.1 reported an exponential increase in dementia prevalence with age. However, Ritchie et al. indicated that in spite of the exponential increase of the prevalence rate of dementia up to the age of 85 years, the rate of prevalence decelerated between 80 and 84 and that around the age 95 the prevalence rate of dementia levelled off at about 40%.12,13 Therefore, AD could be regarded as an ‘age-related’ rather than an ‘aging-related’ disorder. In contrast, most of the subjects who survive over 100 are cognitively impaired or demented.7,8,11,14 This indicates that aging might be considered another risk factor precipitating cognitive impairment and that dementia in the oldest old could be different from that of earlier onset. Therefore, AD is considered not to be a physiological aging process but to be a pathological process.15 However, aging is regarded as another risk factor for precipitating cognitive impairment.7 Secondary, patients with a history of cerebral episodes were excluded from the study. However, it could not ruled out that asymptomatic cerebral infraction had a subtle impact on the cognitive, behavioral and psychological functions in the very old brain. Moreover, in oldest old age group, not only AD but also limbic neurofibrillary tangle dementia (LNTD,16,17 or senile dementia of the neurofibrillary tangle type; SD-NET,18,19) appear in neuropathological findings. LNTD was diagnosed as AD in the clinical setting.16–19 In D90G, four subjects reached to neuropathological diagnosis and two of them were diagnosed as AD and one as LNTD (96-year-old man was diagnosed as LNTD whose memory disturbance worsened at the age of 84 and he became depressive and showed delirium and visual hallucination at age of 95. He was admitted to our hospital at age of 96. We could not diagnose him clinically as other than AD). This also made different clinical features in D90G than in LOG and EOG, however, we cannot distinguish LNTD from AD at the present time because in the oldest old memory disturbances invariably appear. Therefore, clinically, we should diagnose these patients as AD. Alternatively, we can elucidate clinical features in LNTD when we evaluate the symptoms in the oldest old patients with clinically diagnosed AD.
Thus, we hypothesized that (i) EOG was caused exclusively by Alzheimer's type changes; (ii) LOG was caused mainly by Alzheimer's type changes and partly by aging process, vascular factor; and (iii) D90G was caused by Alzheimer's type changes as well as by the aging process, vascular factor and LNTD pathology. In fact, impaired orientation in apace, impaired personal orientation and impaired distant memory were milder in D90G than in LOG and were more severe in EOG than in LOG. This meant that EOG was located in the center of AD and D90G was in a different position from that of AD, which was partly common with AD (Fig. 1).
There were some limitations in this study. First is the small sample size. Second, in this study subjects did not necessarily undergo CT scans and they undergo only HDSR as a neuropsychological test. Needless to say, both detailed neuropathological investigation and neuropsychological testing are required to clarify the neuroanatomical basis of functional impairment in the oldest old.20 However, it is often difficult to apply these examinations to the demented, especially to demented patients in the oldest old age because they can sometimes neither think properly nor concentrate on testing. In fact, although the CDR score was similar both in D90G and LOG, the HDSR score was lower in D90G than in LOG. We evaluated clinical symptoms as well as troublesome behaviors in inpatients of D90G, LOG and EOG using several assessment scales. It is true that BPSD are burdensome to caregivers but they are clues that enable us to elucidate the pathophysiology basis of dementia.21 Some authors suggest specific brain dysfunctions other than cognitive impairment were reflected in BPSD.6,22 For this reason we put great emphasis on BPSD. Finally, we stressed not the longitudinal course of these three dementia groups but the first episodes of BPSD in these groups. In fact, in many nonagenarians in this study the first troublesome episodes were the only episodes. Forty per cent of these D90G died soon after the first BPSD appeared. Although we assessed when the first BPSD episodes were shown in spite of longitudinal course, we also should consider the longitudinal course of these dementias.
In spite of these methodological limitations, this study is one of few studies that evaluate quantitatively the behavioral symptoms in the oldest old with AD. On the basis of the results we will be able to propose specific medications to the oldest old presenting with behavioral symptoms. We should assess the BPSD of AD in a larger sample size of the oldest old patients, and the appearance pattern of BPSD in each FAST stage.