| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
a Aging Research Centre, University of Southern Denmark, Odense University
b Department of Geriatrics G, Odense University Hospital, Denmark
c Epidemiology, Institute of Public Health, University of Southern Denmark, Odense University
Karen Andersen-Ranberg, Aging Research Center, Institute of Public Health, University of Southern Denmark, Odense University, Sdr. Boulevard 23A, DK-5000 Odense C, Denmark E-mail: karanberg{at}health.sdu.dk.
Decision Editor: Toni C. Antonucci, PhD
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
IT is widely recognized that the prevalence of dementia increases with advancing age and that this increase shows an exponential pattern, with a doubling of the prevalence every 5 years to a level higher than 20% at the age of 85 (
Hofman et al. 1991
;
Jorm, Korten, and Henderson 1987). Extrapolation into very old ages has been used to predict the burden of dementia in the population of oldest persons, that is, those aged 85 years and older.
Jorm and colleagues 1987 predicted a prevalence of 62% at the age of 95; others have predicted that the prevalence of dementia levels off at 45% at the same age (
Ritchie and Kildea 1995;
Ritchie, Kildea, and Robine 1992;
Wernicke and Reischies 1994). The latter could be caused by a differential mortality in nondemented and demented persons, as suggested by
Johansson and Zarit 1995. They also reported a lower annual incidence rate from age 86 to 94 as compared with the incidence in a comparable but slightly younger cohort observed from age 84 to 92. In contrast, in a recent meta-analysis
Jorm and Jolley 1998 showed that the incidence of dementia rises exponentially at least to the age of 90 years.
Recently, reliable demographic data from the past 30 years in developed countries have shown a leveling off of mortality rates of persons aged 80 and older (
Vaupel et al. 1998). This phenomenon may also be seen in connection with the prevalence of dementia, as proposed by
Ritchie and Kildea 1995. They found a modified logistic pattern for the prevalence of late life dementia rather than an exponential pattern, which would predict that nearly all extremely old people, such as centenarians, would be demented. Because the proportion of oldest old persons has increased dramatically within the last 30 years, and probably will continue to increase in the future (
Vaupel et al. 1998), there is a need to estimate both the incidence and the prevalence of dementia in extremely old individuals.
To our knowledge, there are no incidence studies on centenarians. However, in several recent prevalence studies (
Asada et al. 1996;
Ebly, Parhad, Hogan, and Fung 1994;
Ivan 1990;
Olsen, Jeune, and Andersen-Ranberg 1996;
Powell 1994;
Samuelsson et al. 1997;
Sobel et al. 1995)—all using diagnostic criteria from the revised third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R;
American Psychiatric Association 1987)—prevalence rates show a huge variation, ranging from about 27% to 85%. This diversity may be explained by differences in sample sizes and methods. Also, the high mortality among centenarians makes it difficult to estimate a point prevalence in a representative sample. It is also difficult to evaluate cognitive function in these very old individuals, because physical impairment, including poor vision and hearing, is associated with relative isolation and apathy, which can mimic a cognitive decline, and makes cognitive tests difficult to administer. Only few centenarian studies discuss these problems in detail (
Ivan 1990;
Powell 1994;
Samuelsson et al. 1997). However, despite the variations between the different studies, it seems that a certain proportion of centenarians are cognitively intact, although the exact magnitude of this proportion and the proportion of moderately to severely demented are not known.
Knowledge regarding the subtypes and causes of dementia in advanced age is also inconclusive. Most studies have reported Alzheimer's disease (AD) to be the predominant dementia subtype. In studies of 85-year-olds,
Ebly and colleagues 1994 found that AD accounted for 75%, and a vascular etiology alone for 13%, of all dementia cases. Recent studies have indicated a larger cerebrovascular involvement in the development of dementia, especially the presence of hypertension earlier in life (
Hofman et al. 1997;
Skoog et al. 1996;
Skoog, Nilsson, Palmertz, Andreasson, and Svanborg 1993). In brain autopsy material,
Snowdon and colleagues 1997 showed that persons with both brain infarcts and neuropathologic lesions compatible with AD had poorer cognitive function and higher prevalence of dementia compared with persons with neuropathologic lesions alone.
The proportion of centenarians reported to have AD ranges from very low (
Ivan 1990;
Powell 1994) to very high (
Asada et al. 1996;
Sobel et al. 1995). However, none of these studies based their diagnosis of AD on the strict criteria of National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA;
McKhann et al. 1984); at best, only a modified version of these criteria has been used (
Asada et al. 1996;
Sobel et al. 1995). Two centenarian studies (
Asada et al. 1996;
Sobel et al. 1995) based on different criteria (the National Institute of Neurological and Communicative Disorders and Stroke-Association Internationale pour la Recherche et l'Enseignement en Neurosciences [NINCDS-AIREN;
Roman et al. 1993] and the International Classification of Diseases [ICD-10;
World Health Organization 1992]) both reported a prevalence of vascular dementia (VaD) of around 3%. A prevalence of mixed type of dementia (MIX), that is, possible AD with a vascular component, at 6% has been reported in a single centenarian study (
Asada et al. 1996). Both the divergence in AD prevalence and the lacking prevalence of VaD and MIX in some studies illustrate the difficulties in assessing dementia subtypes in extremely old people with concomitant high comorbidity, which would highly interfere with the current criteria for AD and VaD.
We launched the Longitudinal Study of Danish Centenarians to elucidate the health and morbidity of people at the extreme end of a human life span. In this article we address the issue of mental health by assessing the prevalence of dementia and its relation to other diseases. Additionally, we discuss the difficulties in assessing cognitive function in extremely old people, such as centenarians.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
Procedure and Measures
With the exception of the 5 participants who refused to have a personal visit, all the centenarians were interviewed and examined by the same geriatrician (K.A.-R.) and geriatric nurse (L.V.) at their place of residence, including institutions. The interviews were conducted within 3 months of the participants' 100th birthday with the exception of 3 centenarians who were interviewed within 101, 113, and 146 days, respectively. The sociodemographic interview included items on family background, education, occupation, and lifestyle. The assessment of activities of daily living (ADLs) was based on an instrument previously used and validated in Denmark (
Avlund, Kreiner, and Schultz-Larsen 1996), including the items of Katz's ADL index (
Katz, Ford, Moskowitz, Jackson, and Jaffe 1963) and the items on instrumental activities of daily living (IADL) of
Lawton and Brody 1969. This information was collected from both participants and caregivers on all 207 subjects and represents the minimum amount of available data for the entire sample. We screened cognitive function using the Mini-Mental State Examination (MMSE;
Folstein, Folstein, and McHugh 1975). A proxy to each of the 207 participants was interviewed regarding their short- and long-term memory and ability to recognize other people, find their way home, recall the time of the year, understand messages, and make decisions.
Health was assessed by a thorough interview with the centenarians (when possible) and/or proxies. The interview included specific questions regarding diseases associated with dementia, for example, neurological disease, cerebrovascular and cardiovascular disease, vitamin B12 and folic acid deficiencies, and thyroid disease. At the end of the interview the centenarians had a physical examination including a neurological examination, blood pressure measurements (average of three measurements), and electrocardiography. Hypertension was defined as blood pressure higher than 160/100 mm Hg. The diagnoses of vitamin B12 and/or folic acid deficiency were based on medical records and actual medication, and the diagnoses of hypothyroidism and diabetes were based on both blood examination and medical history. A neurologist confirmed the diagnosis of Parkinson's disease according to international criteria (
Calne, Snow, and Lee 1992). Hearing problems could mostly be overcome by the use of the centenarian's personal hearing aid, but in some cases we had to use a special hearing device (Hear-It, Original; Logia, Mejeribakken 8, DK–3540 Lynge, Denmark) with headphones and microphone as a supplement to improve communication.
Additional health information was retrieved on all 207 participants from their general practitioners' medical records, the Danish Cancer Register (in existence since 1943), and the National Discharge Register (in existence since 1977), which register all Danish hospital discharge diagnoses and surgical treatments according to World Health Organization (WHO) classifications. Both registries offer nationwide coverage. Diseases diagnosed and treated by doctors, and not self-reported diseases, have been used for the present analysis.
The diagnosis of dementia was made according to WHO's (1992) criteria of ICD-10, which include the pattern of daily activities. The severity was rated according to the criteria in the Clinical Dementia Rating (CDR;
Hughes, Berg, Danziger, Coben, and Martin 1982).
We assessed every participant in detail regarding the following items, which used to operationalize the ICD-10 criteria:
The specific domains of memory, orientation, and judgment + problem solving in the CDR were rated by the same markers used for the ICD-10 criteria. The more complex domain of community affairs was assessed with information on participants' ability to make decisions on shopping and financial affairs, and the domain of home and hobbies was assessed through information on the centenarians' interest in reading or hearing the newspaper or books read aloud, housework, gardening, or hobbies. When rating according to the CDR, we excluded the evaluation of job functioning. As in the diagnosis of dementia, a distinct evaluation was made in each case of whether physical handicaps were interfering with the IADL items (e.g., shopping, managing financial affairs, cooking, doing the laundry, looking up a phone number, socializing, reading books, or playing solitaire).
In our dementia rating we found it appropriate to add an additional category, probably no dementia, to the existing five levels of the CDR (healthy, questionable, mild, moderate, and severe dementia). This category consisted of all centenarians whose sensory deficits and/or unwillingness to perform all the items in the MMSE made it difficult to rate them as nondemented according to the strict criteria. However, all other information indicated normal cognitive function (e.g., recalling during the interview). Centenarians on whom we did not have enough information to make a solid evaluation of cognitive function were placed in a separate category, termed not classified.
Because the criteria of NINCDS-ADRDA (
McKhann et al. 1984) for AD and NINCDS-AIREN (
Roman et al. 1993) for VaD could not be applied to our data, we tried another approach to identify possible etiologies of dementia. First, on the basis of the physical examination together with information from medical records, all individuals who were suffering from diseases defined in ICD-10 to have a potential influence on cognitive function (e.g., hypothyroidism, vitamin B12 or folic acid deficiency, and Parkinson's disease) were identified. Demented participants with these diagnoses were classified as having possible other dementia (pOD). Second, we tried to identify those centenarians in whom a cerebro- or cardiovascular disease could be identified on the basis of the information in the medical records and the findings in the physical and neurological examinations. No brain imaging was done.
For the 56 centenarians who refused to participate in the study, a brief evaluation of cognitive status (possible nondementia or suspected dementia) was possible at the time of the initial contact with the centenarian or the proxy. We used either the impression they made at the time of contact or information from proxies, that is, family members or caregivers, to rate the cognitive status of the centenarians.
The Scientific-Ethical Committee of the Counties of Funen and Vejle approved the Longitudinal Study of Danish Centenarians. We used chi-square to test for differences between groups. A significance level of .05 was chosen.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
Seventy-five percent (n = 156) of the centenarians were tested by means of the MMSE (Table 2 ). About one fifth (22%) had a score of 24 or more, and a slightly higher proportion (29%) had scores below 17. More than 90% of the 45 centenarians who scored 24 points or more in the MMSE were clinically assessed as having no dementia or probably no dementia. The remaining 4 were diagnosed as having either questionable dementia or mild dementia. A wider variation in the CDR was seen among those scoring 23–18 points and 17–10 points; among those scoring 9–0 points, more than 75% were clinically evaluated as having moderate to severe dementia.
The difficulties in applying the MMSE are shown in Table 3 . Of the 156 participants who were examined, less than half (46%) went through a complete examination. In the remaining 54% the major reason for incomplete examination was visual impairment. Six participants were totally deaf and could not participate in the MMSE. The special hearing aid Hear-It improved communication for 27 participants.
|
In all, 13 (12%) of 105 demented centenarians could be identified as belonging to the group of pOD (Table 4 ), although they could have had a cerebro- or cardiovascular disease, too. One centenarian had both hypothyroidism and vitamin B12 deficiency, and 1 had Parkinson's disease (mildly demented).
|
|
The crude evaluation of suspected dementia/possible no dementia in the nonparticipants showed that 23 (41%) could be evaluated as possible no dementia, 17 (30%) as suspected dementia, and 16 (29%) as unclassifiable/unknown. Among the 13 deceased the level of cognitive function prior to death was unknown. If all the unclassifiable persons among participants and nonparticipants and the deceased were demented (i.e., worst case scenario), the prevalence of dementia in the total population would be 58.7%, which was not significantly different from the estimated prevalence among the participants.
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Our findings are compatible with the representative study of
Sobel and colleagues 1995, who found that among 179 Finns aged 100 years or more the proportion of moderately to severely demented centenarians was 33%, similar to our findings (34%), whereas 79 (44%) could be classified as nondemented according to DSM-III-R criteria. Also, our proportion of questionable dementia together with mild dementia (24%) is comparable to the Finnish study, which showed 23% of participants with cognitive decline or mild dementia.
The higher proportions (58–85%) of demented participants found in most centenarian studies (
Asada et al. 1996;
Ebly et al. 1994;
Ivan 1990;
Powell 1994) are difficult to explain, because such proportions would imply a continued increase in incidence and no differential mortality between demented and nondemented persons. However, because differential mortality probably does exist, disfavoring demented persons, we therefore would expect to find lower prevalences of dementia in very old people, a finding that would be accentuated by the study methods used in many of these previous studies, that is, selection of relatively healthy survivors (
Ivan 1990), time lag between identification and examination during which the frailest die, and problems of age validation (
Asada et al. 1996). Additionally, few details are reported regarding the basis of the diagnosis, apart from the use of DSM-III-R criteria, and the difficulties in applying such criteria to participants with sensory impairments. Only one study (
Samuelsson et al. 1997) reported a surprisingly low prevalence of dementia of only 27%. As in our study, Samuelsson and colleagues used sampling methods to minimize selection bias, that is, a population-based, dynamic cohort of 100-year-olds. Furthermore, their dementia diagnosis was based on an extensive neuropsychological assessment, which was not done in other reported centenarian studies, including ours. However, because the participation rate in the Swedish study was lower than in our study (61% vs 75%), we have tried to recalculate the Swedish data to make a better comparison: If we include Swedish refusals and deceased into a worst case scenario, the prevalence of dementia in Swedish centenarians rises to 54%, which is not significantly different from the worst case scenario in our study (58%).
A neuropsychological assessment would probably be the optimal way to evaluate cognitive functioning. But the difficulties in performing such assessments in extremely old persons are illustrated by the fact that the whole neuropsychological assessment battery could be performed on only approximately half of the Swedish centenarians (
Samuelsson et al. 1997). Furthermore, in cases of severe problems with hearing and vision, dementia diagnosis could not be based on the DSM-III-R criteria. These difficulties were also seen in our study, where only one third of the participants could complete the whole MMSE adequately, and the rest were only examined partly due to different deficits or could not or would not go through the examination.
The MMSE is not an adequate test in centenarians, as illustrated in this study. Others have shown how deficits in vision affect MMSE scores among community-dwelling centenarians (
Holtsberg, Poon, Noble, and Martin 1995). In our study impaired vision interfered with the tasks of drawing two pentagons (n = 52) and writing a sentence (n = 34) in the 84 centenarians who did not complete the MMSE. The task of reading an order ("Close your eyes!") could be overcome, in most cases, by showing the command in big letters, with the exception of 30 participants, for whom we had to omit the task.
Hearing disability can partly be overcome by the use of hearing aids, which most of the Danish centenarians have. Although hearing aids improved communication, poor hearing could still be a problem. For example, if the centenarian did not hear the three words the first time in the learning and recalling test of the MMSE we repeated them, but this might have created a learning effect, which could have biased the result. This may have also been a problem in the item concerning reiterating a sentence. In some cases the hearing disability was so severe that we had to give up further questioning.
Impaired physical capability, which is common among extremely aged persons such as centenarians (
Andersen-Ranberg et al. 1999), makes it difficult to evaluate the ability and degree of maintaining an independent life. Caregiving necessitated by physical disability may obscure a higher level of cognitive function, especially in the groups of probably no dementia and questionable dementia. But we achieved full information regarding ADLs (
Avlund et al. 1996;
Katz et al. 1963) and IADLs (
Lawton and Brody 1969) as well as proxy information regarding intellectual abilities, which enabled us to distinguish between disability due to cognitive or physical impairment. In this study only the MMSE and a complete physical examination could not be performed on all participants. However, for most of the participants with missing MMSE data, other variables could provide us with information to assess cognitive functions.
The large differences in the reported prevalence of dementia are probably due to the use of different methods both in sampling and in assessment, as well as the diagnostic criteria. The use of a dynamic population-based cohort is essential in minimizing selection bias in studies of dementia in extremely aged people. Cognitive function tests and neuropsychological assessment that can evaluate participants with sensory impairment at the extreme end of life must be developed for future studies regarding cognition. In our opinion the optimal way to conduct such studies is to use a two-phase strategy, keeping the number of investigators visiting the participant to a minimum. A clinician (e.g., a geriatrician) and a neuropsychologist should make the first visit together; the same neuropsychologist could conduct the second interview. It is important that the participants and proxies not have to meet more than two investigators, even if the study is longitudinal. A two-phase strategy would also minimize fatigue in the participants, which can otherwise become a major problem and interfere negatively with the results in an extensive and clinical examination. Unfortunately, in a population study like ours, we did not have the capacity required to carry through such a two-phase strategy.
We characterized the type of dementia by applying the ICD-10 criteria of dementia and using medical history and previous diagnoses, as well as results from the clinical examination. It would have been preferable to have brain imaging done in order to make a proper evaluation of cerebrovascular disease. However, this was not feasible mainly because of the centenarians' or their relatives' negative attitude to transportation to the hospital. On the other hand, our collection of registered diagnoses from hospitals and general practitioners' files, together with the objective clinical examination and results from a blood test, allowed us to carefully identify and assess a possible vascular component or other somatic disease relevant to the pathogenesis of dementia.
In 13 centenarians we found a history of systemic disorders or Parkinson's disease resulting in a 12% prevalence of pOD. This is nearly the same as in the Finnish study (9%;
Sobel et al. 1995). The use of the group of pOD is mainly due to exclusion criteria from NINCDS-ADRDA and NINDS-AIREN. They could very well have had AD or VaD.
It is even more complicated for us to differentiate between AD and VaD. The NINCDS-ADRDA criteria for diagnosing AD were not applicable to our data because the criteria require onset of disease between the ages of 40 and 90, and such information was not sufficiently valid in our material. We could not apply the NINDS-AIREN criteria for diagnosing VaD because we lacked brain imaging data. Instead, we evaluated the prevalence of different cardio- and cerebrovascular diseases known as risk factors in the development of VaD. In half of the demented population we found a cerebro- or cardiovascular disease, which might indicate at least a coexisting vascular component. The diseases were mainly stroke/TCI, atrial fibrillation, and hypertension.
The findings of
Snowdon and colleagues 1997 suggest that both neuropathologic lesions of AD and cerebrovascular lesions are necessary in the development of poorer cognitive function and dementia among oldest old persons. This might explain why in our material, with the exception of hypertension, we could not show any differences in the prevalences of the different cardiovascular and cerebrovascular diseases between nondemented and demented persons. Maybe the demented centenarians had neuropathologic lesions of AD in addition to cerebrovascular lesions, whereas the nondemented had only cerebrovascular lesions. Only postmortem studies of centenarian brains can give us the answer. The neuropathologic criteria for AD used in younger elderly persons might not be applicable in centenarians, as suggested by others (
Giannakopoulos, Hof, Giannakopoulos, et al. 1995;
Giannakopoulos, Hof, Vallet, et al. 1995;
Hauw et al. 1986;
Mizutani and Shimada 1992).
Hypertension surprisingly turned out to be more common in nondemented than in demented centenarians. In this cross-sectional study we can suggest two explanations for this finding: Either hypertension protects against dementia, or blood pressure falls in demented persons. The first explanation is somewhat contradictory to the hypothesis of cerebrovascular disease's being a risk factor in the development of dementia. In contrast, the latter explanation is supported by the findings of
Skoog and colleagues 1996 in their 15-year follow-up survey of 70-year-old people, in which they found a decline in blood pressure in the years before onset of dementia.
Although the causes of dementia are still unknown, the findings of
Skoog and colleagues 1993,
Skoog and colleagues 1996,
Hofman and colleagues 1997, and
Snowdon and colleagues 1997 indicate that treatment of risk factors in the development of cerebrovascular disease could play an important role in lowering the risk of becoming demented, whatever the subtype. In the last decade the beneficial effects of treating hypertension even in highly aged persons have been recognized, and more people are being treated. If the abovementioned findings are true, this may influence the prevalence of dementia in old people, so that we will not find the same high prevalence of dementia in the growing proportion of very old people in the future.
In conclusion, our data show that dementia is a common, but not inevitable, phenomenon in extremely aged people such as centenarians, thus implying that the prevalence of this disease does not reach 100% at the extreme end of the human life span.
|
Acknowledgments |
|---|
Received for publication March 2, 1999. Accepted for publication November 4, 1999.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. F. Calvert Jr., J. Hollander-Rodriguez, J. Kaye, and M. Leahy Dementia-free survival among centenarians: an evidence-based review. J. Gerontol. A Biol. Sci. Med. Sci., September 1, 2006; 61(9): 951 - 956. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Gondo, N. Hirose, Y. Arai, H. Inagaki, Y. Masui, K. Yamamura, K.-i. Shimizu, M. Takayama, Y. Ebihara, S. Nakazawa, et al. Functional status of centenarians in Tokyo, Japan: developing better phenotypes of exceptional longevity. J. Gerontol. A Biol. Sci. Med. Sci., March 1, 2006; 61(3): 305 - 310. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. T Perls The different paths to 100 Am. J. Clinical Nutrition, February 1, 2006; 83(2): 484S - 487S. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. C. Antonucci Introduction to Special Section on Centenarians and Dementia J. Gerontol. B. Psychol. Sci. Soc. Sci., May 1, 2001; 56(3): P133 - P133. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
