Brain, Vol. 122, No. 12, 2309-2319,
December 1999
© 1999 Oxford University Press
Review article |
An autopsy-verified study of the effect of education on degenerative dementia
1 Sección de Neurología, Hospital Severo Ochoa, Leganés, Madrid, Spain, 2 Department of Clinical Neurological Sciences, 3 London Psychiatric Hospital and 4 Department of Pathology, University of Western Ontario, London, Ontario, Canada
Correspondence to:
Professor Vladimir Hachinski, Department of Clinical Neurological Sciences, The University of Western Ontario, 361 Windermere Rd, London, Ontario, Canada N6A 3K7 E-mail: vladimir.hachinski{at}lhsc.on.ca
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
A longitudinal study of the relationship between education and age of onset, rate of progression and cerebral lesion burden in a series of autopsy-confirmed demented patients with clinical and 6-monthly psychometric follow-up and autopsy was carried out. The study was conducted at the London Health Sciences Centre University Campus of the University of Western Ontario on 87 patients with pathologically confirmed Alzheimer's disease (60), dementia with Lewy bodies (11) or dementia with Lewy bodies plus Alzheimer's disease (16). Their educational attainment was classified as below high school, high school or above high school, and was similar to that of the age-adjusted general Ontario population. The age of onset of dementia, age at death, progression of cognitive decline, amount of neurodegenerative changes (senile plaques, neurofibrillary tangles and Lewy bodies) and cerebrovascular lesions (infarcts, lacunar state and white matter rarefaction) were assessed. Less educated patients became demented later and died later, but cognitive function declined at the same rate in all educational groups and there was no difference in the burden of neurodegenerative lesions between them. However, the less educated patients had more cerebrovascular lesions. It can be concluded that higher education does not modify the course of Alzheimer's disease, but lower education relates to the occurrence of cerebral infarcts. Our results suggest that a `brain battering' model related to the higher prevalence of small vascular lesions in less educated individuals may explain their increased risk of dementia described by epidemiological studies better than the prevalent `brain reserve' hypothesis.
education; dementia; vascular lesions; Alzheimer's disease; cerebral infarcts
ESD = Extended Scale for Dementia
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The inverse association between previous education and dementia is one of the most fascinating and debated findings in the field of neurodegenerative diseases (Katzman, 1993
). Several epidemiological studies, from different cultural settings, have shown that the prevalence of dementia is higher in poorly educated individuals (Rocca et al., 1990; Zhang et al., 1990; Dartigues et al., 1991; Canadian Study of Health and Aging, 1994; Stern et al., 1994; Ott et al., 1995; Callahan et al., 1996; Evans et al., 1997; Obadia et al., 1997), although other reliable investigators found no evidence (O'Connor et al., 1991; Beard et al., 1992; Bonaiuto et al., 1995; Cobb et al., 1995) or only partial evidence (Ott et al., 1999) of this association.
The inverse relationship between previous formal education and dementia could be due to educational skills raising scores on the diagnostic tools (O'Connor et al., 1991), but some studies have suggested that this is not the case (Katzman, 1993; Stern et al., 1994; Callahan et al., 1996). A real association of education and reduced risk of dementia could be explained in two ways (Mortimer and Graves, 1993). The `brain reserve' hypothesis states that education has a protective effect because it enhances the cognitive assets above the threshold for dementia, possibly by increasing synaptic density or efficiency (Fratiglioni et al., 1991; Stern et al., 1992, 1994, 1995a, b; Katzman, 1993) or by acquired skills (Snowdon et al., 1989a, b; Fratiglioni et al., 1991; Katzman, 1993; Stern et al., 1995). The neurodegenerative process assumed to affect individuals, regardless of their educational level, would take longer to bring down a larger reserve to the critical threshold. The alternative, which we may call the `brain battering' hypothesis, assumes that individuals with greater educational attainment and associated higher socio-economic status would be exposed to fewer toxins, enjoy a healthier lifestyle and have greater access to quality health care, all of which would tend to spare their brains from lesions contributing to dementing illnesses (Fratiglioni et al., 1991; Hill et al.,1993; Katzman, 1993). Cerebral infarcts may be one of these lesions, whose association with low socio-economic status is well documented (Rocca et al., 1990; Fratiglioni et al., 1991; Ott et al., 1995).
The structural basis of dementia affected by education is controversial. In a Swedish epidemiological study, only alcoholic, unspecified and vascular dementia were influenced by education (Fratiglioni et al., 1991). In a study performed in Shanghai (Zhang et al., 1990) an effect of education was predominant in Alzheimer's disease, but in two population surveys in Appignano (Rocca et al., 1990) and Rotterdam (Ott et al., 1995) it was observed both in Alzheimer's disease and vascular dementia. However, the diagnosis in all these studies was based on clinical impression alone using a clinicopathological series of demented patients. The biological basis of the effect of education on the incidence or prevalence of dementia cannot be addressed. It can, however, be used to examine the dose–response effect of educational level on the age of onset and the rate of progression of the disease, especially the burden of different types of brain lesions and the relationship with the degree of cognitive impairment.
We have examined in a cohort of demented patients followed to autopsy, the effect of educational attainment on: (i) the age of onset of mental impairment and death, (ii) the rate of cognitive decline, and (iii) the type and amount of lesions and their relationship to the degree of dementia.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Setting
The Dementia Study Project of the University of Western Ontario (Merskey et al., 1985
) is a 17-year-old prospective clinicopathological study of dementia, based in the memory clinic of a tertiary hospital with a catchment area of 1 500 000 people. The cohort was recruited through referrals from general practitioners, psychiatrists and neurologists working under the universal coverage rules of the Canadian Health Care System. There is no parallel medical system in this area. Participation in the study was voluntary and implied no costs to the patients or their relatives. Consent for autopsy was requested from patient or relative as appropriate; granting consent was not a condition for participation in the study. The autopsy rate of patients in the Dementia Study was 64% but there were no significant differences in the sex distribution, the age of onset of dementia and the age at death between autopsied and non-autopsied patients, as described elsewhere (Bowler et al., 1998).
Patients
The current study was based on all 95 cases in the Dementia Study Project who had died and were autopsied during the period 1986–93. Our clinicopathological research was focused on Alzheimer's disease and dementia with Lewy bodies; consequently, we excluded six cases with miscellaneous diagnoses (Pick, neuroaxonal dystrophy, Creutzfeldt–Jakob disease, progressive supranuclear palsy and two cases of dementia lacking distinctive histology). No autopsy diagnosis of pure vascular dementia was reached during this period. Two cases for whom information about the level of formal education was missing were also excluded, leaving 87 cases in the analysis.
All patients were Caucasian (38 men, 49 women) and had joined the Dementia Study after informed consent from themselves or their caregivers. The racial distribution is representative of the population of this age in south-western Ontario. On each patient's entry to the study one neurologist (V.H.) reviewed demographic and clinical data, applied a neurological and mental state examination, and performed a routine complementary work-up (haemogram, serum biochemistry, thyroid hormones, luetic serology, EEG and CT). A cognitive examination with the Extended Scale for Dementia (ESD) (Hersch, 1979) in 80 cases and a functional assessment with the London Psychogeriatric Rating Scale (LPRS) (Hersch et al., 1978) in 85 cases were also done. Patients were followed with regular applications of the ESD and LPRS every 6–12 months. All cases were diagnosed as having dementia according to DSM-III criteria (American Psychiatric Association, 1987) and this diagnosis was confirmed during a follow-up of 1–12 years until death. The length of the follow-up was distributed normally with mean of 5.5 ± 2.4 years (SD), median 5 years.
Clinical data
The following data were obtained from the prospectively recorded Dementia Study Project forms: (i) age at the onset of dementia, according to the caregiver's information about the first notice of relevant signs of memory or behaviour disturbances, age at death and duration of disease; this information was available in all cases; (ii) ESD cognitive scores recorded in 80 cases when they joined the Dementia Study (4.4 ± 2.3 years after the onset of dementia) and in many cases also during the last few years before death, as presented in Fig. 1. In all patients who reached the state of severe cognitive decline the score in the ESD was considered to be 0 until death; (iii) severe cognitive decline, defined as the point at which the ESD was no longer applicable. This point was accurately identified in 61 cases. The last recorded ESD in 19 cases was above this level. In the remaining 7 cases ESD data was missing and this point could not be established, although they were not severely demented at entry to the study; and (iv) onset of regular bladder incontinence, determined as described previously (Del Ser et al., 1996). This information was available in 79 cases; three patients did not become incontinent during their life span.
|
Previous formal education
Previous education was graded in three levels: 28 cases below high school education (less than grade 9), 32 cases with high school education (grades 9–13) and 27 cases above high school education (20 cases with non-university further education and seven cases with a university degree). Students normally attend high school from 14 to 18 years of age. The number of years of formal education had been recorded in all but 6 cases, all of whom were professionals, and the number of years was estimated from the ordinary length of education required for graduation in their field. Thirty-one cases had been manual workers, 21 cases specialized manual workers, 9 cases white collar workers, 13 cases technical workers and 13 cases professionals.
Pathological study and diagnostic groups
All pathological procedures were performed at the Division of Neuropathology, Department of Pathology, University Hospital of London, Ontario. All pathological measurements were done by the same researcher (T.D.S.) blinded to the clinical and educational information.
The brain was removed at the autopsy, fixed in formalin and cut into 1 cm thick coronal slices. All macroscopic vascular lesions were mapped on a sheet with 18 diagrams of consecutive brain coronal sections. The brain weight and the approximate volume of macroscopic brain infarcts were recorded.
Histological sections from 10 regions (frontal, parietal, occipital and temporal lobes, hippocampus, amygdala, basal ganglia, thalamus, midbrain and cerebellum) were stained with haematoxylin and eosin, Bielschowsky's method and Congo red. Additional sections were immunolabelled as previously described (Manlow and Munoz, 1992) with an antibody to ubiquitin (Sigma Immunochemicals, Oakville, Ontario, Canada) to facilitate the recognition of cortical Lewy bodies (Dickson et al., 1989) and abnormal neurites in the CA2 sector of the hippocampus (Dickson et al., 1991).
The following pathological variables were assessed in every case.
Semiquantitative variables
Ischaemic lacunar lesions were assessed in the basal ganglia and thalamus and graded in two levels: (i) absent or scarce microscopic, (ii) abundant microscopic or grossly visible. White matter lesions were assessed with a four-point scale in all available histological sections of the centrum semiovale. The average was dichotomized as <2, absent/mild; 
2, moderate/severe. Amyloid angiopathy was rated as absent or present on the basis of Congo red stain.
Quantitative variables
The following lesions were counted in a Bielschowsky's stained hippocampal section (side selected at random): the number of intracellular neurofibrillary tangles and ghost tangles in a 1 mm sector of hippocampal pyramidal layer in CA1, adjacent to CA2, and the number of diffuse plaques, neuritic plaques and neurofibrillary tangles in a 0.5-mm wide strip of neocortex located in the depth of the external lip of the collateral sulcus.
In a parallel section stained with haematoxylin and eosin the number of Lewy bodies was counted in a 1-mm wide strip of paleocortex located in the middle of the parahippocampal gyrus. The number of Lewy bodies was also counted on both sides of the substantia nigra in a section of the mesencephalon stained with haematoxylin and eosin, and averaged.
These quantitative measures were done using a reticule of 0.5 x 0.5 mm and counting all lesions which were partially or totally included in it. Measures were repeated in 20 cases with a test–retest reliability (Pearson's correlation) of 0.91 for plaques in the cortex, 0.94 for tangles in the hippocampus and 0.95 for Lewy bodies in the substantia nigra.
Diagnostic groups
On the basis of histological examination patients were sorted into diagnostic groups defined by the following criteria:
Alzheimer's disease (60 cases).
Abundant neocortical neuritic senile plaques and neurofibrillary tangles (eight or more in 1 mm2 of the temporal neocortex), as described and illustrated by Mirra and colleagues (Mirra et al., 1993), and no cortical Lewy bodies or abnormal ubiquitinated neurites in CA2. There were also some vascular lesions in 29 cases: moderate microscopic lacunes (10 cases) or macroscopic lacunes (1–6 ml in volume, 12 cases) in subcortical regions, moderate infarcts in the left frontotemporal region (30, 27, 20, 20, 10 and 9 ml in volume, 6 cases), marked white matter hypodensities (13 cases) or hippocampal sclerosis, probably of ischaemic cause (5 cases).
Pure dementia with Lewy bodies (11 cases).
Numerous cortical (five or more in 2 mm2 of the parahippocampal cortex) and subcortical Lewy bodies, abnormal ubiquitinated neurites in CA2 and cortical diffuse plaques, with less than one neurofibrillary tangle or four neuritic plaques per mm2 of temporal neocortex. In two cases there were also moderate macroscopic infarcts (6 and 10 ml in volume).
Dementia with Lewy bodies plus Alzheimer's disease (16 cases).
Both numerous neuritic plaques, neurofibrillary tangles (eight or more in 1 mm2 of the temporal neocortex) and Lewy bodies (one or more in 2 mm2 of the parahippocampal cortex) present in the cortex. In four cases there were also moderate microscopic (1 case) or macroscopic infarcts (two cases 3 ml and one case 11 ml in volume).
The whole sample was also divided into two main groups according to the presence or absence of vascular lesions: (i) 52 cases (31 Alzheimer's disease, 9 pure dementia with Lewy bodies, 12 dementia with Lewy bodies plus Alzheimer's disease) with pure degenerative lesions, and (ii) 35 cases (29 Alzheimer's disease, 2 pure dementia with Lewy bodies, 4 dementia with Lewy bodies plus Alzheimer's disease) with degenerative plus vascular lesions. There was no case with pure vascular dementia.
Statistical analysis
All analyses were pre-planned to answer the questions posed in the introduction relating to the dose–response effects of education on dementia. The demographic data, the scores on the ESD (at the time of inclusion in the Dementia Study and every year before death) and the quantitative pathological data were compared with a one-way ANOVA (analysis of variance) between the three educational groups. In order to exclude a confounding cohort effect (Moritz and Petitti, 1993; Cobb et al., 1995; Peterson, 1996) the age of dementia onset and at death were also submitted to another two-way ANOVA, using education and birth cohort as factors. For the study of the relationship of degenerative lesion burden and educational attainment, the effects of age at death, duration of disease and cognitive performance were controlled through the use of ANCOVA (analysis of covariance).
The relationships of the main pathological diagnoses with educational level and with the years of education were analysed with the 
2 test and with a one-way ANOVA, respectively.
The relationships of the presence of vascular lesions (lacunar or white matter lesions, macroscopic infarcts, amyloid angiopathy) with educational level were analysed with the 2 test for linear trend.
Finally, the presence of vascular lesions was introduced together with education in a two-way ANOVA in order to separate the influence of each factor on every demographic and pathological dependent variable. Although adjusting for date of birth would seem desirable, it was not possible to do so because all demographic variables were highly correlated with it (age of onset, r = –0.89; age at death, r = -0.95; age at severe cognitive decline, r = –0.92) No corrections for multiple tests were introduced.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The frequency of cases in every educational level was very similar in this sample (below high school, 32.1%; high school, 36.7%; above high school, 31%) to that of the general population of Ontario aged 65 years and over, according to the 1986 Census of Canada [below high school, 36.1%; high school, 34.3%; above high school, 29.4%;
2(2) = 0.423, n.s.] (Statistics Canada, 1989).
There were no differences between the sexes in the age of onset of dementia (men: 68.4 ± 7.5, women: 68.6 ± 8.3; t = 0.118, n.s.), years of education (men: 11.7 ± 4, women: 10.7 ± 3.7; t = 1.150, n.s.) or level of education (men/women: below high school, 31.5/38.6%; high school, 37/36.7%; above high school, 31.5/30.6%; 2(2) = 0.992, n.s.). Pathological data, other than brain weight, did not show significant differences and therefore sex was not controlled for in the following analyses.
Demographic data
Less educated patients were significantly older at the onset of dementia and at death (Table 1). When the sample was divided into three subgroups by date of birth, the significant effect of education on age of onset (education effect, F = 4.620, P = 0.01; cohort effect, F = 66.4, P < 0.001) and age at death (education effect: F = 4.3115, P = 0.01, cohort effect F = 95.893, P < 0.001) persisted without significant interaction between cohort and education (Fig. 2). Education had no significant effect on the duration of dementia or the age of onset of urinary incontinence (Table 1). Less educated cases were significantly older when cognition became untestable with the ESD (Table 1), but this difference disappeared after controlling for the effect of age of onset (ANCOVA: Fß = 1.120, n.s., indicating a similar effect of the covariable `age of onset' on the independent variable in each group; Fµ = 1.079, n.s., indicating absence of difference between group means corrected for the effects of the covariable `age of onset').
|
|
Cognitive data
The first cognitive scores in the ESD, recorded at an average of 4.4 ± 2.3 years after the onset of dementia, and the scores recorded 5, 4, 3, 2, 1 years or several months before death had a minor tendency to be greater in more educated patients, but the difference did not reach statistical significance at any of these points (Fig. 1
). Adjusting for the initial scores did not alter these results.
Neurodegenerative pathology
There was no significant effect of educational level on the quantitative pathological findings (Table 2). More educated patients showed a slight trend to have greater amounts of neurofibrillary tangles, but the density of neurodegenerative lesions (plaques, tangles and Lewy bodies) was highly variable and the level of statistical significance was not reached. These negative results persisted when the age at death, the duration of the disease and the cognitive performance previous to death were controlled for in the ANCOVAs (data not shown but available on request). These tests have an 80% statistical power to detect an effect size (f) of 0.35. Effect size for an ANOVA test is defined as the standard deviation between groups divided by the standard deviation within groups; unlike the F statistic, f is independent of sample size
|
Vascular pathology
When the pathological diagnoses were grouped in the standard manner, i.e. by the degenerative condition, disregarding minor vascular lesions, their distribution did not vary between the three educational levels [
2(4) = 0.741, n.s.; Fig. 3]. There were no differences in the mean number of years of formal education between groups defined by neurodegenerative pathology (Alzheimer's disease, 10.9 ± 3.5; pure dementia with Lewy bodies, 12.2 ± 3.6; dementia with Lewy bodies plus Alzheimer's disease, 11.4 ± 5.1; ANOVA: F(3) = 0.427, n.s.). However, when patients with and without vascular lesions were compared, significant differences were found. Cases with both degenerative and vascular lesions had lower educational attainment (below high school, 51.4%; high school, 20%; above high school, 28.5%; Fig. 3) and less years of formal education (10.9 ± 3.5) than those with pure degenerative lesions (below high school: 19.2%, high school: 48%, above high school: 32.6%; 2(2) = 11.336, P = 0.003, 12 ± 3.4 years; Student's t(85) = 2.428, P = 0.01).
|
Subcortical lacunar and white matter lesions had a tendency to be more frequent in less educated people, almost reaching the significance level (Table 3
). Macroscopic infarcts and the presence of any vascular lesion were significantly related to low education (Table 3). This statistical trend towards greater vascular pathology in less educated cases was not due to amyloid angiopathy which showed no correlation with education (Table 3).
|
When the education and the presence of vascular lesions were simultaneously analysed in a two-way ANOVA, the former had no effect on any of the demographic variables (Table 4
). Conversely, the presence of vascular lesions was significantly associated with a later age of onset of dementia, incontinence and severe cognitive decline. The most definite parameter, age at death, was delayed by 6.1 years in patients with additional vascular lesions compared with those with degenerative pathology only (P < 0.0001) (Table 4).
|
Cases with Alzheimer-type lesions
All statistical analyses were also done in the subgroup of 76 cases with Alzheimer-type lesions (60 Alzheimer's disease and 16 dementia with Lewy bodies plus Alzheimer's disease). The results were very similar to those of the whole sample and we do not present them in detail. The number of cases with Lewy bodies (25 cases) was too small to be submitted to statistical analysis.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Our study has examined the association of educational attainment with demographic, progression and lesion burden variables. With regard to age of onset, less educated patients became demented significantly later. This finding has been described previously in cross-sectional studies (Filley et al., 1985
; Moritz and Petitti, 1993; Duara et al., 1996; Peterson, 1996) and tentatively attributed to the longer time required by the relatives to become aware of the cognitive decline in a poor social and work environment (Moritz and Petitti, 1993). On the other side of the argument Stern and colleagues, in a longitudinal study of a referral-based cohort of Alzheimer's disease patients, found an increased risk of death in cases with higher educational and occupational attainment (Stern et al., 1995b), although a replication study based on a community cross-sectional survey did not corroborate these findings (Geerlings et al., 1997). Our data on age at death (younger in the more educated group) are congruent with those of Stern and colleagues (Stern et al., 1995b). However, they interpreted their findings as due to delayed manifestation of dementia in these patients (when the underlying pathology is well advanced), but they analysed neither the time of evolution, nor pathological data. Had the onset of dementia been overlooked and artefactually delayed in less educated patients (4 years in our series), or alternatively, retarded in more educated individuals, we would expect a shorter span of the clinically detected illness in the corresponding group, assuming no effect of education on the rate of progression. In our series of patients followed until death, we were able to determine that the duration of dementing illness was the same in the three educational groups. Moreover, the observed cognitive scores in the ESD showed the same trend and took the same time to become severe in all groups. Some short longitudinal studies have found that education influences the rate of cognitive decline in Alzheimer's disease, by using the Mini Mental State Examination (Rasmusson et al., 1996; Christensen et al., 1997) or other tests of crystallized intelligence (Christensen et al., 1997), but more extensive studies (Filley et al., 1985; Mortimer et al., 1992; Bowler et al., 1998) have not detected such an effect. Therefore, our demographic and progression data are consistent with what others have reported, but not with some explanations offered in the literature.
The recruitment rules of clinic-based studies of dementia give rise to a negative correlation between birth year and age of onset of the disease (the sliding window of opportunity effect), as we have reported elsewhere (Bowler et al., 1998). This, in combination with the well-known secular improvement in the spread of education, explains in part the inverse relationship between the age of onset and educational level which we found. However, the persistence of the association after stratification for birth year indicates that this cohort artefact is not the only relevant factor.
The pathogenetic mechanisms underlying the temporal patterns of onset and progression of the illness are best explored by examination of the causative lesions (Kazee et al., 1993). We found that the load of neurodegenerative lesions was unrelated to educational attainment, even after controlling for age at death, duration of dementia and cognitive deterioration. Snowdon and colleagues found a similar lack of inverse association between early linguistic ability (a potential indicator of cultural level) and neurofibrillary tangles and senile plaques (Snowdon et al., 1996).
Conversely, we found that the presence of cerebrovascular lesions is associated with lower education attainment, as well as later ages of onset and death. Our finding that patients with mixed degenerative and vascular lesions were less educated and older is in accordance with population based studies, showing that cerebrovascular lesions and incident stroke (Ferrucci et al., 1996) are more common in the lower socio-economic classes (Rocca et al., 1990; Fratiglioni et al., 1991; Ott et al., 1995) and in late life (Crystal et al., 1993; Skoog et al., 1993; Breteler et al., 1994), and can contribute to the symptomatic onset of dementia (Skoog et al., 1993; Ebly et al., 1994). A recent clinicopathological study by Snowdon and colleagues suggests that small cerebrovascular lesions strongly potentiate the dementing effects of degenerative lesions (Snowdon et al., 1994). It should be remarked that these vascular lesions are usually small and they can remain undetected if the diagnosis is based on clinical data only (Kokmen et al., 1996b), as in many epidemiological surveys lacking neuroimaging (Zhang et al., 1990; Canadian Study of Health and Aging, 1994; Stern et al., 1994; Callahan et al., 1996). These difficulties in diagnosis could be the reason why, in some epidemiological studies, the effect of low education on the prevalence of dementia appears only in vascular dementia (Folstein et al., 1991; Bonaiuto et al., 1995; Cobb et al., 1995), whereas in others it appears also in Alzheimer's disease (Rocca et al., 1990; Zhang et al., 1990; Ott et al., 1995). Several epidemiological studies have considered cerebrovascular lesions as a possible determinant of the higher prevalence of dementia in individuals with low educational attainment (Rocca et al., 1990; Fratiglioni et al., 1991; Cobb et al., 1995; Ott et al., 1995), but they lacked pathological data to confirm their suspicions.
It might seem surprising that the higher prevalence of vascular lesions in less educated individuals would manifest as a delayed onset of dementia and age of death. This is, however, the expected result in a series of demented individuals if we assume that the syndrome of dementia can result from the combined effect of neurodegenerative and vascular lesions (Snowdon et al., 1997; Consortium to Establish a Registry for Alzheimer's Disease, 1998) and that this association is more frequent in older (Consortium to Establish a Registry for Alzheimer's Disease, 1998) and less educated subjects. In fact, any factor increasing the risk of dementia and predominantly affecting older individuals would have contrasting effects on population-based epidemiological studies versus clinicopathological series of demented patients, because the rules of selection are different. On a population basis, patients with the factor will show a greater prevalence of dementia than those without it. In a clinical series, demented patients with the factor will be older than those without it, and thus its presence will appear to delay the onset of dementia. We presume that small cerebrovascular lesions are one such factor. Our study also shows that the onset of urinary incontinence is not related to educational attainment, but rather to the type of pathology, confirming a previous report to this effect (Del Ser et al., 1996).
Our results (particularly the independence of education and neurodegenerative lesion load) do not support the hypothesis that the protective effect of education observed in population-based prevalence studies is due to the greater resistance of better educated individuals to the dementing effect of neurodegenerative lesions, as proposed by the `brain reserve hypothesis'. Rather, it may be that the lower prevalence of vascular lesions among the better educated could explain this observation. This would be supportive, although by no means demonstrative of the `brain battering' hypothesis.
However, several caveats have to be considered. First, the pathological study probes the state of the brain at the terminal stages of disease (>9 years after onset in our series). We cannot address a possible, and plausible, protective effect of education limited to the early stages of disease. Secondly, Mortimer suggested that psychosocial risk factors would have their strongest associations with dementia in late onset Alzheimer's disease (Mortimer, 1988), as was observed in the Shanghai study (Zhang et al., 1990). The average age at death in our sample (78.4 ± 7.2 years) is lower than the average age of living Canadian demented patients (Canadian Study of Health and Aging, 1994), a problem common to many clinicopathological series (Brayne, 1993; Jost and Grossberg, 1995). Perhaps in a series more weighted towards older patients, some effect of education might be found. However, since our study shows that low education, advanced age and cerebral infarcts are associated in demented patients, relationships between education and dementia in the ninth decade may also be determined by the presence of small cerebral vascular lesions. Thirdly, some epidemiological studies have focused on very low education (less than 6 years of schooling) as the risk factor for dementia (Katzman and Kawas, 1994). In fact, a recent study (De Ronchi et al., 1998) which claimed a protective effect for education after controlling for smoking, alcohol consumption and history of hypertension, found no difference between those who had completed 3 years of schooling and those with more lengthy formal education. Since our study included only two cases with no formal education, we cannot comment on this variable. Our results cannot be generalized beyond the conditions associated with westernized, industrialized societies. Vascular risk factors are also associated with early death, including sudden death, and may therefore affect the probability of being included in this sample as well as the length of follow-up, and consequently the risk of progressing to severe cognitive decline.
Furthermore, the relatively small size of our sample raises the risk of a type II error in some of our negative findings. Although the statistical power of the ANOVA in our sample (80%) is appropriate to detect large (f = 0.35) effect sizes (Cohen, 1988), it would be less so for smaller effects. Therefore, we cannot rule out that a small effect of education has been undetected. However, this hypothetical effect would be less than the effects of cerebrovascular lesions, which have been easily identified in this study.
Although the distribution of the educational attainment in our sample was very close to that of elderly people, 65 years and over, in Ontario (Statistics Canada, 1989) we are well aware that a selection bias cannot be excluded (Kokmen et al., 1996b), because a clinicopathological series from a university hospital cannot be representative of the general population (Brayne, 1993). The absence of financial restrictions on access to specialized health care in Canada should reduce referral bias in comparison with other settings. We hope our clinicopathological series will encourage population-based studies to test the hypothesis suggested by our results, i.e that the higher prevalence of dementia among less educated individuals in epidemiological surveys reflects the higher prevalence of small cerebrovascular lesions in this group.
|
Acknowledgments |
|---|
We wish to thank many people who worked in the follow-up of patients included in the Dementia Study of the University of Western Ontario and in the collection of clinical reports. This work was supported by grants from the `Fondo de Investigaciones Sanitarias' from the Ministerio de Salud y Consumo of Spain (93/5545) to T.D.S. and from the Medical Research Council of Canada (MT-12707) to D.G.M., and by Dr Hachinski's Trillium Clinical Scientist Award. D.G.M. also received support from the `Fondo de Investigaciones Sanitarias' of Spain (98/3157).
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
American Psychiatric Association. Diagnostic and statistical manual of mental disorders. DSM-III-R. 3rd ed rev. Washington (DC): American Psychiatric Association; 1987.
Beard CM, Kokmen E, Offord KP, Kurland LT. Lack of association between Alzheimer's disease and education, occupation, marital status, or living arrangement. Neurology 1992; 42: 2063–8.
Bickel H, Cooper B. Incidence and relative risk of dementia in an urban elderly population: findings of a prospective field study. Psychol Med 1994; 24: 179–92.[Web of Science][Medline]
Bonaiuto S, Rocca WA, Lippi A, Giannandrea E, Mele M, Cavarzeran F, et al. Education and occupation as risk factors for dementia: a population-based case-control study. Neuroepidemiology 1995; 14: 101–9.[Web of Science][Medline]
Bowler JV, Munoz DG, Merskey H, Hachinski V. Factors affecting the age of onset and rate of progression of Alzheimer's disease. J Neurol Neurosurg Psychiatry 1998; 65: 184–90.
Brayne C. Clinicopathological studies of the dementias from an epidemiological viewpoint [see comments]. Br J Psychiatry 1993; 162: 439–46. Comment in: Br J Psychiatry 1993; 163: 261–2, Comment in: Br J Psychiatry 1993; 163: 415–6.
Breteler MM, van Swieten JC, Bots ML, Grobbee DE, Claus JJ, van den Hout JH, et al. Cerebral white matter lesions, vascular risk factors, and cognitive function in a population-based study: the Rotterdam Study. Neurology 1994; 44: 1246–52.
Callahan CM, Hall KS, Hui SL, Musick BS, Unverzagt FW, Hendrie HC. Relationship of age, education, and occupation with dementia among a community-based sample of African Americans. Arch Neurol 1996; 53: 134–40.
Canadian Study of Health and Aging. The Canadian Study of Health and Aging: risk factors for Alzheimer's disease in Canada [see comments]. Neurology 1994; 44: 2073–80. Comment in: Neurology 1995; 45: 1635.
Christensen H, Korten AE, Jorm AF, Henderson AS, Jacomb PA, Rodgers B, et al. Education and decline in cognitive performance: compensatory but not protective. Int J Geriatr Psychiatry 1997; 12: 323–30.[Web of Science][Medline]
Cobb JL, Wolf PA, Au R, White R, D'Agostino RB. The effect of education on the incidence of dementia and Alzheimer's disease in the Framingham Study. Neurology 1995; 45: 1707–12.
Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale (NJ): Lawrence Erlbaum; 1988.
Consortium to Establish a Registry for Alzheimer's Disease Demographic and clinical predictors of survival in patients with Alzheimer's disease: CERAD, part XIV.Neurology 1996; 46: 656–60.
Consortium to Establish a Registry for Alzheimer's Disease. Cerebral infarcts in patients with autopsy-proven Alzheimer's disease: CERAD, part XVIII. Neurology 1998; 51: 159–62.
Crystal HA, Dickson DW, Sliwinski MJ, Lipton RB, Grober E, Marks-Nelson H, et al. Pathological markers associated with normal aging and dementia in the elderly. Ann Neurol 1993; 34: 566–73.[Web of Science][Medline]
Dartigues JF, Gagnon M, Michel P, Letenneur L, Commenges D, Barberger-Gateau P, et al. Le programme de recherche PAQUID sur l'épidémiologie de la demence. Méthodes et résultats initiaux. Rev Neurol (Paris) 1991; 147: 225–30.[Medline]
De Ronchi D, Fratiglioni L, Rucci P, Paternico A, Graziani S, Dalmonte E. The effect of education on dementia occurrence in an Italian population with middle to high socioeconomic status. Neurology 1998; 50: 1231–8.
Del Ser T, Munoz D, Hachinski V. Temporal pattern of cognitive decline and incontinence is different in Alzheimer's disease and diffuse Lewy body disease. Neurology 1996; 46: 682–6.
Dickson DW, Crystal H, Mattiace LA, Kress Y, Schwagerl A, Ksiezak-Reding H, et al. Diffuse Lewy body disease: light and electron microscopic immunocytochemistry of senile plaques. Acta Neuropathol (Berl) 1989; 78: 572–84.[Medline]
Dickson DW, Ruan D, Crystal H, Mark MH, Davies P, Kress Y, et al. Hippocampal degeneration differentiates diffuse Lewy body disease (DLBD) from Alzheimer's disease: light and electron microscopic immunocytochemistry of CA2–3 neurites specific to DLBD. Neurology 1991; 41: 1402–9.
Duara R, Barker WW, Lopez-Alberola R, Loewenstein DA, Grau LB, Gilchrist D, et al. Alzheimer's disease: interaction of apolipoproteinE genotype, family history of dementia, gender, education, ethnicity, and age of onset. Neurology 1996; 46: 1575–9.
Ebly EM, Parhad IM, Hogan DB, Fung TS. Prevalence and types of dementia in the very old: results from the Canadian Study of Health and Aging [see comments]. Neurology 1994; 44: 1593–600. Comment in: Neurology 1994; 44: 1563.
Evans DA, Hebert LE, Beckett LA, Scherr PA, Albert MS, Chown MJ, et al. Education and other measures of socioeconomic status and risk of incident Alzheimer disease in a defined population of older persons. Arch Neurol 1997; 54: 1399–405.
Ferrucci L, Guralnik JM, Salive ME, Pahor M, Corti MC, Baroni A, et al. Cognitive impairment and risk of stroke in the older population [see comments]. J Am Geriatr Soc 1996; 44: 237–41. Comment in: J Am Geriatr Soc 1996; 44: 328–3, Comment in: J Am Geriatr Soc 1996; 44: 141011.[Web of Science][Medline]
Filley CM, Brownell HH, Albert ML. Education provides no protection against Alzheimer's disease. Neurology 1985; 35: 1781–4.
Folstein MF, Bassett SS, Anthony JC, Romanoski AJ, Nestadt GR. Dementia: case ascertainment in a community survey. J Gerontol 1991; 46: M132–8.
Fratiglioni L, Grut M, Forsell Y, Viitanen M, Grafstrom M, Holmen K, et al. Prevalence of Alzheimer's disease and other dementias in an elderly urban population: relationship with age, sex and education. Neurology 1991; 41: 1886–92.
Geerlings MI, Deeg DJ, Schmand B, Lindeboom J, Jonker C. Increased risk of mortality in Alzheimer's disease patients with higher education? A replication study. Neurology 1997; 49: 798–802.
Hersch EL. Development and application of the Extended Scale for Dementia. J Am Geriatr Soc 1979; 27: 348–54.[Web of Science][Medline]
Hersch EL, Kral VA, Palmer RB. Clinical value of the London Psychogeriatric Rating Scale. J Am Geriatr Soc 1978; 26: 348–54.[Web of Science][Medline]
Hill LR, Klauber MR, Salmon DP, Yu ES, Liu WT, Zang M, et al. Functional status, education, and the diagnosis of dementia in the Shanghai survey. Neurology 1993; 43: 138–45.
Jost BC, Grossberg GT. The natural history of Alzheimer's disease: a brain bank study [see comments]. J Am Geriatr Soc 1995; 43: 1248–55. Comment in: J Am Geriatr Soc 1995; 43: 1314–5, Comment in: J Am Geriatr Soc 1996; 44: 734.[Web of Science][Medline]
Katzman R. Education and the prevalence of dementia and Alzheimer's disease. [Review]. Neurology 1993; 43: 13–20.
Katzman R, Kawas C. The epidemiology of dementia and Alzheimer disease. In: Terry RD, Katzman R, Bick KL, editors. Alzheimer disease. New York: Raven Press; 1994. p. 105–22.
Kazee AM, Eskin TA, Lapham LW, Gabriel KR, McDaniel KD, Hamill RW. Clinicopathologic correlates in Alzheimer disease: assessment of clinical and pathologic diagnostic criteria [see comments]. Alzheimer Dis Assoc Disord 1993; 7: 152–64. Comment in Alzheimer Dis Assoc Disord 1993; 7: 129–35.[Web of Science][Medline]
Kokmen E, Beard CM, O'Brien PC, Kurland LT. Epidemiology of dementia in Rochester, Minnesota. Mayo Clin Proc 1996a; 71: 275–82.[Abstract]
Kokmen E, Ozsarfati Y, Beard CM, O'Brien PC, Rocca WA. Impact of referral bias on clinical and epidemiological studies of Alzheimer's disease [see comments]. J Clin Epidemiol 1996b; 49: 79–83. Comment in: J Clin Epidemiol 1997; 50: 365–6.[Web of Science][Medline]
Manlow A, Munoz DG. A non-toxic method for the demonstration of gliosis. J Neuropathol Exp Neurol 1992; 51: 298–302.[Web of Science][Medline]
Merskey H, Blume WT, Colhoun EH, Fisman M, Fox AJ, Fox H, et al. Correlative studies in Alzheimer's disease. [Review]. Prog Neuropsychopharmacol Biol Psychiatry 1985; 9: 509–14.[Medline]
Mirra SS, Hart MN, Terry RD. Making the diagnosis of Alzheimer's disease. A primer for practicing pathologists [see comments]. [Review]. Arch Pathol Lab Med 1993; 117: 132–44. Comment in: Arch Pathol Lab Med 1993; 117: 129–31.[Web of Science][Medline]
Moritz DJ, Petitti DB. Association of education with reported age of onset and severity of Alzheimer's disease at presentation: implications for the use of clinical samples [see comments]. Am J Epidemiol 1993; 137: 456–62. Comment in: Am J Epidemiol 1996; 143: 1177.
Mortimer JA. Do psychosocial risk factors contribute to Alzheimer's disease? In: Henderson AS, Henderson JH, editors. Etiology of dementia of Alzheimer's type. Chichester: John Wiley; 1988. p. 39–52.
Mortimer JA, Graves AB. Education and other socioeconomic determinants of dementia and Alzheimer's disease. Neurology 1993; 43 Suppl 4: S39–S44.
Mortimer JA, Ebbitt B, Jun SP, Finch MD. Predictors of cognitive and functional progression in patients with probable Alzheimer's disease. Neurology 1992; 42: 1689–96.
Obadia Y, Rotily M, Degrand-Guillaud A, Guelain J, Ceccaldi M, Severo C, et al. The PREMAP Study: prevalence and risk factors of dementia and clinically diagnosed Alzheimer's disease in Provence, France. Prevalence of Alzheimer's Disease in Provence. Eur J Epidemiol 1997; 13: 247–53.[Web of Science][Medline]
O'Connor DW, Pollitt PA, Treasure FP. The influence of education and social class on the diagnosis of dementia in a community population. Psychol Med 1991; 21: 219–24.[Web of Science][Medline]
Orrell M, Sahakian B. Education and dementia. Research evidence supports the concept `use it or lose it' [editorial; comment]. BMJ 1995; 310: 951–2.
Ott A, Breteler MM, van Harskamp F, Claus JJ, van der Cammen TJ, Grobbee DE, et al. Prevalence of Alzheimer's disease and vascular dementia: association with education. The Rotterdam study [see comments]. BMJ 1995; 310: 970–3. Comment in: BMJ 1995; 310: 951–2, Comment in: BMJ 1995; 311: 125–6.
Ott A, van Rossum CTM, van Harskamp F, van de Mheen H, Hofman A, Breteler MMB. Education and the incidence of dementia in a large population-based study: the Rotterdam Study. Neurology 1999; 52: 663–6.
Peterson B. Re: Association of education with reported age of onset and severity of Alzheimer's disease at presentation: implications for the use of clinical samples [letter; comment]. Am J Epidemiol 1996; 143: 1177. Comment in: Am J Epidemiol 1993; 137: 456–62.
Rasmusson DX, Carson KA, Brookmeyer R, Kawas C, Brandt J. Predicting rate of cognitive decline in probable Alzheimer's disease. Brain Cogn 1996; 31: 133–47.[Web of Science][Medline]
Rocca WA, Bonaiuto S, Lippi A, Luciani P, Turtu F, Cavarzeran F, et al. Prevalence of clinically diagnosed Alzheimer's disease and other dementing disorders: a door-to-door survey in Appignano, Macerata Province, Italy. Neurology 1990; 40: 626–31.
Skoog I, Nilsson L, Palmertz B, Andreasson LA, Svanborg A. A population-based study of dementia in 85-years-olds [see comments]. N Engl J Med 1993; 328: 153–8. Comment in: N Engl J Med 1993; 328: 203–5, Comment in: N Engl J Med 1993; 329: 63–4.
Snowdon DA, Ostwald SK, Kane RL. Education, survival, and independence in elderly Catholic sisters, 1936–1988. Am J Epidemiol 1989a; 130: 999–1012.
Snowdon DA, Ostwald SK, Kane RL, Keenan NL. Years of life with good and poor mental and physical function in the elderly. J Clin Epidemiol 1989b; 42: 1055–66.[Web of Science][Medline]
Snowdon DA, Kemper SJ, Mortimer JA, Greiner LH, Wekstein DR, Markesbery WR. Linguistic ability in early life and cognitive function and Alzheimer's disease in late life. Findings from the Nun Study [see comments]. JAMA 1996; 275: 528–32. Comment in: JAMA 1996; 275: 1879.
Snowdon DA, Greiner LH, Mortimer JA, Riley KP, Greiner PA, Markesbery WR. Brain infarction and the clinical expression of Alzheimer disease. The Nun Study [see comments]. JAMA 1997; 277: 813–7. Comment in: JAMA 1997; 278: 113–4.
Statistics Canada. The nation. Schooling and major field of study. Ottawa: Canadian Government Publishing Centre March; 1989.
Stern Y, Alexander GE, Prohovnik I, Mayeux R. Inverse relationship between education and parietotemporal perfusion deficit in Alzheimer's disease. Ann Neurol 1992; 32: 371–5.[Web of Science][Medline]
Stern Y, Gurland B, Tatemichi TK, Tang MX, Wilder D, Mayeux R. Influence of education and occupation on the incidence of Alzheimer's disease [see comments]. JAMA 1994; 271: 1004–10. Comment in: JAMA 1994; 272: 1405–6.
Stern Y, Alexander GE, Prohovnik I, Stricks L, Link B, Lennon MC, et al. Relationship between lifetime occupation and parietal flow: implications for a reserve against Alzheimer's disease pathology. Neurology 1995a; 45: 55–60.
Stern Y, Tang MX, Denaro J, Mayeux R. Increased risk of mortality in Alzheimer's disease patients with more advanced educational and occupational attainment. Ann Neurol 1995b; 37: 590–5.[Web of Science][Medline]
Zhang MY, Katzman R, Salmon D, Jin H, Cai GJ, Wang ZY, et al. The prevalence of dementia and Alzheimer's disease in Shanghai, China: impact of age, gender, and education. Ann Neurol 1990; 27: 428–37.[Web of Science][Medline]
Received February 8, 1999. Revised June 21, 1999. Accepted June 28, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. Hakansson, S. Rovio, E.-L. Helkala, A.-R. Vilska, B. Winblad, H. Soininen, A. Nissinen, A. H Mohammed, and M. Kivipelto Association between mid-life marital status and cognitive function in later life: population based cohort study BMJ, July 2, 2009; 339(jul02_2): b2462 - b2462. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. W. Small, P. Siddarth, A. C. Burggren, V. Kepe, L. M. Ercoli, K. J. Miller, H. Lavretsky, P. M. Thompson, G. M. Cole, S. C. Huang, et al. Influence of Cognitive Status, Age, and APOE-4 Genetic Risk on Brain FDDNP Positron-Emission Tomography Imaging in Persons Without Dementia Arch Gen Psychiatry, January 1, 2009; 66(1): 81 - 87. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. M. Roe, C. Xiong, E. Grant, J. P. Miller, and J. C. Morris Education and Reported Onset of Symptoms Among Individuals With Alzheimer Disease Arch Neurol, January 1, 2008; 65(1): 108 - 111. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Ngandu, E. von Strauss, E. -L. Helkala, B. Winblad, A. Nissinen, J. Tuomilehto, H. Soininen, and M. Kivipelto Education and dementia: What lies behind the association? Neurology, October 2, 2007; 69(14): 1442 - 1450. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Rundek and D. A. Bennett Cognitive leisure activities, but not watching TV, for future brain benefits Neurology, March 28, 2006; 66(6): 794 - 795. [Full Text] [PDF]
|
|
|
|
 |
|
 N Scarmeas, S M Albert, J J Manly, and Y Stern Education and rates of cognitive decline in incident Alzheimer's disease. J. Neurol. Neurosurg. Psychiatry, March 1, 2006; 77(3): 308 - 316. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. A. Bennett, J. A. Schneider, R. S. Wilson, J. L. Bienias, and S. E. Arnold Education modifies the association of amyloid but not tangles with cognitive function Neurology, September 27, 2005; 65(6): 953 - 955. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. O. Oyegbile, C. Dow, J. Jones, B. Bell, P. Rutecki, R. Sheth, M. Seidenberg, and B. P. Hermann The nature and course of neuropsychological morbidity in chronic temporal lobe epilepsy Neurology, May 25, 2004; 62(10): 1736 - 1742. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Karp, I. Kareholt, C. Qiu, T. Bellander, B. Winblad, and L. Fratiglioni Relation of Education and Occupation-based Socioeconomic Status to Incident Alzheimer's Disease Am. J. Epidemiol., January 15, 2004; 159(2): 175 - 183. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. A. Bennett, R. S. Wilson, J. A. Schneider, D. A. Evans, C. F. Mendes de Leon, S. E. Arnold, L. L. Barnes, and J. L. Bienias Education modifies the relation of AD pathology to level of cognitive function in older persons Neurology, June 24, 2003; 60(12): 1909 - 1915. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Dufouil, A. Alperovitch, and C. Tzourio Influence of education on the relationship between white matter lesions and cognition Neurology, March 11, 2003; 60(5): 831 - 836. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. W Small What we need to know about age related memory loss BMJ, June 22, 2002; 324(7352): 1502 - 1505. [Full Text] [PDF]
|
|
|
|
 |
|
 R. G. Wight, C. S. Aneshensel, and T. E. Seeman Educational Attainment, Continued Learning Experience, and Cognitive Function among Older Men J Aging Health, May 1, 2002; 14(2): 211 - 236. [Abstract] [PDF]
|
|
|
|
|
|
C. Qiu, L. Backman, B. Winblad, H. Aguero-Torres, and L. Fratiglioni The Influence of Education on Clinically Diagnosed Dementia Incidence and Mortality Data From the Kungsholmen Project Arch Neurol, December 1, 2001; 58(12): 2034 - 2039. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Fritsch, M. J. McClendon, K. A. Smyth, A. J. Lerner, C. H. Chen, G. J. Petot, S. M. Debanne, A. Soas, and R. P. Friedland Effects of educational attainment on the clinical expression of Alzheimer's disease: Results from a research registry American Journal of Alzheimer's Disease and Other Dementias, November 1, 2001; 16(6): 369 - 376. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||