Vliet, P. van
Citation
Vliet, P. van. (2010, November 10). Determinants of cognitive function in old age. Retrieved from https://hdl.handle.net/1887/16134
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/16134
Note: To cite this publication please use the final published version (if applicable).
Chapter 6
Plasma levels of apolipoprotein E and cognitive function in old age
S.P. Mooijaart, P. van Vliet, D. van Heemst, P.C.N. Rensen, J.F.P. Berbée, J. Jolles, A.J.M. de Craen, R.G.J. Westendorp
Ann N Y Acad Sci 2007;1100:140-147
Abstract Introduction
The relationship between structural variants of the apolipoprotein E gene, APOE ε2/
ε3/ε4, and dementia is well established, whereas the relationship of plasma apoE levels with dementia is less clear. Plasma apoE levels are under tight genetic control but vary widely within the various genotypes indicating that the APOE ε2/ε3/ε4 locus explains only a small fraction of this variation. Here we studied the association of plasma apolipoprotein E (apoE) levels with cognitive function in the elderly population at large.
Materials and methods
Within the Leiden 85-plus Study, a prospective population based study of subjects aged 85 years, we measured plasma apoE level and genotype at baseline. During a five-year follow-up period, cognitive function was annually assessed using the Mini Mental State Examination (MMSE) and a standardized neuropsychological test battery.
Results
Among ε3ε3 carriers (n=324), high plasma apoE levels associated with impaired global cognitive function (-1.10 points change in MMSE score per one standard deviation increase of plasma apoE level, p=0.001), as well as lower attention (p=0.064), speed and memory function (all p<0.05). Adjustment for cardiovascular risk factors and exclusion of all subjects who suffered a stroke did not materially change the associations. Similar estimates were obtained in ε3ε4 carriers (n=100), but not in ε2ε3 carriers (n=90).
Discussion
In old age, in non-ε2-allele carriers, high plasma apoE levels are associated with cognitive impairments, independent of genotype, cardiovascular risk factors, and stroke.
6
Introduction
The relationship between structural variants of the apolipoprotein E gene, APOE ε2/
ε3/ε4, and dementia is well established, whereas the relationship of plasma apoE levels with dementia is less clear. Two APOE gene polymorphisms constitute the ε2/ε3/ε4-alleles that encode three structurally different isoforms apoE2, apoE3, and apoE4. Numerous studies have consistently shown that, compared to carriers of the APOE ε3-allele, subjects with the ε4-allele have an increased risk of developing dementia 1,2, whereas ε2-allele carriers are at a lower risk 3.
The apoE2/3/4 isoforms have different affi nity to receptors that regulate their clearance and as a result the isoforms circulate at different plasma levels 4. Plasma apoE levels are under tight genetic control 5 but the APOE locus itself explains only a small fraction of this variation 6 and plasma apoE levels thus vary within genotypes as well 7. Recently we showed that, for each APOE genotype separately, plasma apoE levels strongly associate with an increased risk of mortality from cardiovascular causes 8. The association of plasma apoE levels with cognitive function in the elderly population at large has not been reported yet.
Here, we studied whether, for each APOE genotype separately, plasma apoE levels associate with cognitive function. Therefore, in the Leiden 85-plus Study, a population based prospective follow up study, we determined APOE ε2/ε3/ε4 genotypes, and plasma apoE levels at baseline. Annual assessment of cognitive function was obtained over a follow up period of fi ve years.
Materials and methods Participants
Between September 1, 1997, and September 1, 1999, a total of 705 inhabitants of the community of Leiden, the Netherlands, reached the age of 85 years. Among these 85-year-old persons, we initiated a follow up study to investigate determinants of successful aging. There were no selection criteria on health or demographic characteristics. Fourteen inhabitants died before they could be enrolled. The response rate was 87%; a total of 599 subjects (397 women and 202 men) participated 9. There were no signifi cant differences in various demographic characteristics between the 599 respondents and the source population. Of the 599 participants in the cohort, 38 refused to provide a blood sample, yielding a total number of 561 participants for
the present study. Subjects were visited within one month after their 85th birthday at their home for face-to-face interviews and neuropsychological testing. Subjects were revisited annually until age 90 years. The Medical Ethical Committee of the Leiden University Medical Centre approved the study, and informed consent was obtained from all subjects.
APOE genotypes
For genotyping two TaqMan assays (Applied Biosystems) were used. For the single nucleotide polymorphism (SNP) in codon 112 primers were sense 5’-GCTGGGCGCGGACAT-3’ and antisense 5’-CACCTCGCCGCGGTACT-3’
and probes were 5’-CGGCCGCGCACGTCC-3’ labelled with FAM and 5’-AGGCGGCCGCACACGTC-3’ labelled with VIC. For the SNP in codon 158 an Assay-On-Demand (Applied Biosystems) was available with assay ID C___904973_10. The assays were run on a 7900HT (Applied Biosystems) according to manufacturer’s specifications, with the following modifications: Eurogentec qPCR core kit was used according to standard specifications; half the concentrations of primers and probes were used; the number of PCR cycles was 50. Fluorescence intensities were measured after the PCR and genotypes were indicated by the Sequence Detection Software version 2.0 (Applied Biosystems).
Plasma parameters
Blood samples were collected early in the morning, although not fasting. Baseline plasma apoE levels were determined using a human apoE-specific sandwich ELISA essentially as described in detail elsewhere 8.
Plasma levels of total cholesterol, HDL cholesterol, and triglycerides were analyzed at baseline on fully automated computerized analyzers (Hitachi 747 and 911; Hitachi, Ltd, Tokyo, Japan). The level of LDL cholesterol was estimated by the Friedewald equation (LDL cholesterol [mmol/L] = total cholesterol – HDL cholesterol – [triglycerides/2.2]), whereby subjects with a triglycerides concentration higher than 443 mg/dL (5 mmol/L) were excluded (n = 5).
Cognitive function
Global cognitive functioning was assessed with the Mini-Mental State Examination (MMSE) 10. Possible scores on the MMSE range from 0 to 30 points, with lower scores indicating worse global cognitive functioning. Attention was measured using the third chart of the 40-item Stroop Test. Outcome was the time needed to name the ink colour of 40 incongruous printed names of colours 11. Higher scores indicated worse attention. Processing speed was assessed with the Letter Digit Coding Test
6
(LDT). Outcome was the total number of correct digits according to a code key in 60 seconds. Lower scores indicated a slower speed. Immediate and delayed recall memory were determined with the 12 Picture Learning Test (PLT-i and PLT-d).
In this test, 12 pictures were presented, and the subject was asked to recall the presented pictures. This procedure was carried out three times. After twenty minutes, the participants were asked to recall the presented pictures again. Outcome for the immediate recall memory was the total number of correctly recalled pictures during the three procedures, and possible scores ranged from 0 to 36 pictures. Outcome for the delayed recall memory was the total number of correctly recalled pictures after 20 minutes, and possible scores ranged from 0 to 12 pictures 12. Lower scores indicated worse memory. Parallel versions for the processing speed and memory tests, using identical procedures, but with different items were used to prevent learning effects. These fi ve neuropsychological tests have been used in the assessment of large samples of elderly subjects and have been shown reliable and sensitive in the detection of small differences in cognitive function 11,13.
Global cognitive function was assessed in all participants. Attention, processing speed, immediate and delayed recall memory were not administered in subjects with a MMSE score below 19 points because of a lack of reliability and validity of these tests in subjects with severe cognitive impairment. The percentage of participants with a MMSE score below 19 points increased from 17% at age 85 years (92 out of 546 participants) to 27% at age 90 years (72 out of 265 participants).
Ischemic cerebrovascular disease
The presence of stroke in the medical history and the incidence of stroke during the follow up period were assessed by annually interviewing treating physicians of all subjects.
Cardiovascular risk factors
Subjects were classifi ed as having diabetes when they met at least one of the following criteria: 1) history of type 2 diabetes obtained from the general practitioner or the subject’s treating physician; 2) use of sulfonylureas, biguanides, or insulin, based on information obtained from the subject’s pharmacist; or 3) non-fasting glucose of 11.1 mmol/L or higher. Subjects were classifi ed as having hypertension when they met at least one of the following criteria: 1) history of hypertension obtained from the general practitioner or the subjects’ treating physician; 2) mean systolic blood pressure of 165 mmHg or higher or diastolic blood pressure of 95 mmHg or higher.
Of all subjects length and weight was measured at baseline. Body mass index (BMI) was calculated from these measurements.
Level of education
Level of education of each subject was determined by the number of years a subject went to school. Information was obtained at the first visit using a questionnaire.
Low education was defined by six or less than six years of schooling, whereas high education was defined by seven or more years of schooling.
Statistical analyses
Plasma levels of total and HDL- and LDL-cholesterol were distributed normally and are reported as means and standard deviations. Plasma levels of apoE and triglycerides were not normally distributed and are presented as medians and interquartile range to assess central tendency. The association between plasma apoE levels and the course of cognitive function during follow up was analysed with linear mixed models. The flexibility of mixed models makes them the preferred choice for the analysis of repeated-measures data. Mixed models use all available data during follow up, can properly account for correlation between repeated measurements, and can handle missing data more appropriately than traditional models 14. Plasma apoE levels were put in the model as a continuous variable, and dichotomized around the median. In an additional analysis standard deviation scores of log-transformed plasma apoE levels were used. All models were adjusted for gender and level of education. In an additional analysis the association between plasma apoE levels and cognition was adjusted for cardiovascular risk factors using prevalence of hypertension and diabetes mellitus, body mass index, and plasma levels of triglycerides, HDL-, and LDL-cholesterol as covariates in the linear mixed model. All calculations were performed using SPSS software (version 12.0.1, SPSS Inc, Chicago, Ill).
6
Results
Figure 1. Annual follow-up visits of 546 participants from age 85 years through 90 years within the Leiden 85-plus Study.
From the 561 participants who were eligible for the present study, APOE genotyping failed in thirteen and measurement of plasma apoE level in two subjects. Figure 1 shows the annual follow up visits of the 546 remaining participants from age 85 through 90 years. Table 1 lists the baseline characteristics of the 546 participants.
When all participants were analyzed, plasma apoE levels were inversely correlated with scores on the MMSE (p<0.001). At baseline subjects with plasma apoE levels above the median scored 1.29 points lower (p=0.015) on the MMSE compared to those with plasma apoE levels below the median. However, these crude analyses are confounded by the various apoE isoforms. Compared to mean plasma apoE levels of APOE ε3ε3 carriers (mean plasma apoE level: 56,8 mg/l; SD: 31,2), ε4 carriers had 8.2 mg/l (p=0.012) lower plasma apoE levels, whereas ε2 carriers had 22.0 mg/l (p=0.001) higher plasma apoE levels. As the ε4 allele carriers are at an increased risk of cognitive impairment, whereas carriers of the ε2 allele are relatively protected, the correlation between plasma apoE levels and MMSE score is thus underestimated.
When the analysis was adjusted for APOE genotypes, subjects with plasma apoE levels above the median scored 1.79 points lower (p=0.001) on the MMSE when compared to those with plasma apoE levels below median.
546 subjects
491 subjects
425 subjects
374 subjects
318 subjects
42 died
60 died
47 died
50 died
48 died
13 refused
6 refused
4 refused
6 refused
5 refused 265 subjects
85 yr
86 yr
87 yr
88 yr
89 yr
90 yr
Table 1. Baseline characteristics of the study participants.
Total number (n) 546
Men 181 (33%)
Institutionalized 97 (18%)
Low level of educationa 354 (65%)
History of stroke 55 (10%)
APOE genotype
e2e2 4 (1%)
e2e3 90 (17%)
e2e4 13 (2%)
e3e3 324 (59%)
e3e4 100 (18%)
e4e4 15 (3%)
Plasma apoE levels (in mg/L)b 50.2 (35.2 – 72.0)
Plasma lipid levels
Total cholesterol (in mmol/L)c 5.71 (1.13)
LDL cholesterol (in mmol/L)c 3.68 (0.98)
HDL cholesterol (in mmol/L)c 1.32 (0.40)
Triglycerides (in mmol/L)b 1.34 (1.01 – 1.95)
Cognitive function
Mini-Mental State Examination (points)b 26 (22-28)
Stroop Test (seconds)b 75 (60-98)
Letter Digit Coding Test (digits)b 16 (12-21)
Picture Learning Test Immediate Recall (pictures)b 25 (20-28) Picture Learning Test Delayed recall (pictures)b 9 (7-11)
a Level of education was dichotomised by six years of schooling.
b Not normally distributed continuous data are presented as medians with interquartile ranges.
c Normally distributed continuous data are presented as means with standard deviation.
To overcome confounding by the various isoforms we also analysed the association between plasma apoE levels and cognitive function in groups of subjects with the same APOE ε2/ε3/ε4 genotype. These analyses were performed in the three largest genotype groups ε2ε3 (n=90), ε3ε3 (n=324) and ε3ε4 (n=100). The genotype groups ε2ε2 (n=4), ε2ε4 (n=13) and ε4ε4 (n=15) were considered too small to yield informative results. Figure 2 shows global cognitive function during the follow up period from age 85 through 90 years separately for those with plasma apoE levels above and below the median. Among ε2ε3 carriers, mean MMSE scores during follow-up were not significantly different between those with high and those with
6
Figure 2. MMSE scores during fi ve year follow-up dependent on plasma apoE levels in subjects with the APOE ε2ε3, ε3ε3, and ε3ε4 genotype.
Plasma apoE levels were dichotomised around the median within each genotype group. Dots represent estimated mean (and standard error) scores using linear mixed models adjusted for sex, and level of education. Number of subjects present at age 90 years for each genotype in the group with low and the group with high plasma apoE levels respectively: e2e3: 27, 25; e3e3: 101, 61; e3e4: 25, 18.
ε2ε3
20 21 22 23 24 25 26 27 28
MMSE (points)
p = 0.711
N=90
high apoE
low apoE
ε3ε3
18 19 20 21 22 23 24 25 26 27 28
MMSE (points)
p = 0.001
N=324 low apoE
high apoE
ε3ε4
14 16 18 20 22 24 26 28
85 86 87 88 89 90
age in years
MMSE (points)
p = 0.018
N=100 low apoE
high apoE
low plasma apoE levels (p=0.711). Among ε3ε3- and ε3ε4 carriers, those with high plasma apoE levels had significantly lower mean MMSE scores during follow-up (p=0.001 and p=0.018 respectively).
When the association of plasma apoE levels and cognitive function was analyzed on a continuous scale, per standard deviation increase of log-transformed plasma apoE level, the MMSE score changed with -0.77 points (p=0.272), -1.10 points (p=0.001), and -2.32 points (p=0.001) for carriers of the ε2ε3-, ε3ε3-, and ε3ε4 genotype respectively. Adjustment for classical cardiovascular risk factors (plasma levels of lipids, body mass index, prevalence of diabetes mellitus, and prevalence of hypertension) did not materially change the observed associations. The adjusted effect sizes were -1.06 points (p=0.190) in ε2ε3 carriers, -1.45 points (p<0.001) in ε3ε3 carriers, and -2.41 points (p=0.001) in ε3ε4 carriers. When this analysis was repeated, now restricting the analysis to only those subjects who did not have a history of stroke at age 85 years and who did not suffer a stroke during the five-year follow-up the effect sizes were -0.09 points (p=0.897), -1.12 points (p=0.002), and -2.06 points (p=0.007) respectively.
To investigate the association of plasma apoE levels and cognitive function in more detail we analyzed data from an additional cognitive test battery including the domains of attention, processing speed and memory (immediate and delayed). These additional tests were not performed in subjects with a MMSE score below 19 points.
Table 2 shows that in ε2ε3 carriers higher plasma apoE levels did not associate with worse performance on the cognitive tests. In ε3ε3 carriers, an increase in plasma apoE level associated with worse performance in attention (p=0.064), processing speed (p=0.033), immediate recall memory (p=0.013), and delayed recall memory (p=0.015). In ε3ε4 carriers similar trends were observed in the scores for attention, but not for processing speed, whereas the effect size on the memory tests was more pronounced. Estimates remained unchanged after adjustment for cardiovascular risk factors.
6
Table 2. Test scores on specifi c domains of cognitive function during follow-up dependent on plasma apoE levels stratifi ed for subjects with the APOE ε2ε3, ε3ε3, and ε3ε4 genotype. GenotypeNChange in test score per 1 standard deviation rise in plasma apoE level (p-value) Stroop (seconds needed to perform taskp-valueLDT (number of correctly coded letters) p-valuePLT-I (number of pictures recalled) p-valuePLT-d (number of pictures recalled)
p-value ε2ε383+1.18 (3.37)0.728+0.35 (0.79)0.657+0.25 (0.62)0.685+0.20 (0.31)0.512 ε3ε3278+3.22 (1.73)0.064-0.82 (0.38)0.033-0.79 (0.32)0.013-0.36 (0.15)0.015 ε3ε475-1.15 (4.32)0.791-0.83 (0.82)0.320-1.07 (0.84)0.208-0.72 (0.41)0.080 Legend. Estimates represent the change in test score per standard deviation increase in log-transformed plasma apoE level. Higher test scores on Stroop and lower test scores on LDT, PLT-i, and PLT-d represent worse cognitive function. Change in test scores (and standard error) were calculated using linear mixed models adjusted for sex, and level of education.
Discussion
The main finding of our study is that in old age, in non-ε2-allele carriers, high plasma apoE levels are associated with lower cognitive function, independent of plasma levels of lipids and other cardiovascular risk factors. Clinical strokes explained only a small proportion of the lower cognitive function during follow-up.
Our study is the first prospective study on the relation between plasma apoE levels and cognition that has been performed in the population at large. The association of plasma apoE levels with cognitive decline has already been studied providing inconclusive results. In the apoEurope Study 15 and the Rotterdam Study 16 subjects with Alzheimer’s disease (AD) had lower plasma apoE levels compared to control subjects. When adjusted for APOE genotype, age, and gender this effect remained the same in the apoEurope Study, but disappeared largely in the Rotterdam Study. In contrast, results from a study by Taddei et al showed higher plasma apoE levels in subjects with AD compared to control subjects in both ε4- and non-ε4 carriers 17. Yet two other studies, performed by Scacchi et al and Folin et al, showed no association of plasma apoE levels with AD 18,19. There are a number of possible explanations for these variable outcomes. First, these studies did not consistently stratify their analyses for genotype groups, thereby allowing the distorting effect of the isoforms of apoE. Second, in these studies cases were AD patients, representing a group of subjects with moderate to severely impaired cognitive function, thereby leaving out subjects with minor impaired cognitive function. As in this population based study we have included individuals with a wide variety of cognitive function, it gave us the opportunity to detect more subtle differences in cognitive function.
Within the population under study, plasma apoE levels have been shown to associate strongly with cardiovascular mortality 8, suggesting that high plasma apoE levels associate with vascular pathologies. In the brain, strokes occur as the consequence of cardiovascular disease and contribute to poor cognition. To rule out the possible influencing effect of strokes we analyzed the data with the exclusion of all clinical strokes. This only partly explained a small part of the observed association. There are a number of possible mechanisms how plasma apoE levels, which are genetically determined, may affect cognition other than stroke. First, plasma apoE may cause vascular pathologies in cerebro other than stroke that contribute to cognitive decline in old age. In autopsies, vascular pathologies have been observed in a large portion of brains and associated with cognitive function 20. These vascular pathologies included haemorrhages, infarctions, lacunes as well as small vessel disease. Second,
6
pro-infl ammatory effects of systemic apoE may infl ict damage in the brain. Recently apoE has been shown to stimulate systemic immune responses by mediating lipid antigen presentation after binding and internalization by antigen presenting cells through the LDL-receptor 21. Peripheral cytokines, produced by the pro- infl ammatory effect of apoE, have been shown to penetrate the blood brain barrier
22 and considerable evidence has accumulated that infl ammation plays a pivotal role in AD 23. Third, systemically produced apoE may cross the blood-brain barrier and infl ict damage to the brain locally. It has been shown that cerebral apoE is locally produced by astrocytes and that cerebrospinal fl uid levels of apoE do not correlate with plasma apoE levels, implying that systemically produced apoE does not cross the blood brain barrier 24,25. However, others have shown that the blood brain barrier is prone to change both structurally and functionally with increasing age 26 and it has been shown that subjects with dementia have a blood brain barrier dysfunction
27. Therefore, systemically produced apoE may cross the dysfunctional blood brain barrier and interfere with pathological processes in cerebro that lead to cognitive decline. For instance, it has been shown that apoE facilitates beta-amyloid deposition in the brain in mouse-models 28-30.
In subjects with the APOE ε2ε3 genotype high plasma apoE levels did not associate with an impaired cognitive function. ApoE2 has far lower affi nity for LDL-receptors compared to apoE3 and apoE4 31, and apoE exerts its pro-infl ammatory effect and the regulation of the lipid metabolism, at least in part, through the LDL-receptor. The effect of apoE2 in lipid antigen presentation was also lower when compared to apoE3
21. Taken together, these observations suggest a genotype specifi c effect of plasma apoE levels, which adds to the reasoning that high levels of plasma apoE3 and apoE4 are causal to cognitive decline, and not an epiphenomenon of other mechanisms.
A possible limitation of our study is that this fi nding cannot directly be extrapolated beyond this age group, especially because cognitive decline is an age-specifi c disorder. However, our study population of 85-years-old subjects represents 15% of the men and 36% of the women from the birth cohort 1912-1914 (offi cial data from the Dutch Bureau of Statistics, www.cbs.nl). Apparently a substantial portion of the total population reaches this age-category and an even larger proportion will reach it in the future. Another possible limitation is that mortality of subjects in this study is dependent on plasma apoE levels, whereas the linear mixed models assumes the missing of data to be independent of the determinant. However, mortality is higher in subjects with high plasma apoE levels, who also have a lower cognitive function. This may have resulted in an underestimation of cognitive decline in subjects with high
plasma apoE levels. A third possible limitation is that the observed number of clinical strokes may have been an underestimation of the total number of cerebrovascular accidents. Minor strokes may not have been recognised as such by participants and health care workers. Imaging of the brain, such as MRI, could help to identify more sub-clinical strokes, but are not available in this population based study. As the effect of excluding the clinical, and thus the most severe, strokes was minimal, we favour the hypothesis that cerebrovascular accidents, including those that may not have been detected clinically, cannot completely explain for the observed lower cognitive function.
A strong point of our study is that the data come from a population based study without inclusion criteria on health and demographics. Therefore our results can be extrapolated to other populations in the same age range. Another strong point is that numbers were large, allowing stratification for the three largest genotype groups, thereby eliminating the potentially distorting effect of structural changes in the three isoforms of apoE. A third strong point is that the used neuropsychological tests have proven to be highly sensitive in the detection of small differences in cognitive function in elderly populations 11,32,33. Scores on these tests can be considered a marker of brain function and likewise as a marker of anatomical and functional changes in the aging brains of the population under study.
In conclusion, we found that, in non-ε2-allele carriers, high plasma apoE levels are associated with lower cognitive function, independent of genotype, plasma levels of lipids and other cardiovascular risk factors. This finding suggests that apart from the well accepted association of apoE structural variation with cognitive function, apoE quantitative variation is an important and independent indicator of cognitive decline in old age.
6
References
1. Corder, EH, Saunders, AM, Strittmatter, WJ, et al. Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer’s disease in late onset families. Science. 1993; 261:921-923.
2. Raber, J, Huang, Y, and Ashford, JW. ApoE genotype accounts for the vast majority of AD risk and AD pathology. Neurobiol Aging. 2004; 25:641-650.
3. Corder, EH, Saunders, AM, Risch, NJ, et al. Protective effect of apolipoprotein E type 2 allele for late onset Alzheimer disease. Nat Genet. 1994; 7:180-184.
4. Utermann, G, Pruin, N, and Steinmetz, A. Polymorphism of apolipoprotein E. III. Effect of a single polymorphic gene locus on plasma lipid levels in man. Clin Genet. 1979; 15:63-72.
5. Beekman, M, Heijmans, BT, Martin, NG, et al. Heritabilities of apolipoprotein and lipid levels in three countries. Twin Res. 2002; 5:87-97.
6. Beekman, M, Posthuma, D, Heijmans, BT, et al. Combined association and linkage analysis applied to the APOE locus. Genet Epidemiol. 2004; 26:328-337.
7. Neale, MC, de Knijff, P, Havekes, LM, et al. ApoE polymorphism accounts for only part of the genetic variation in quantitative ApoE levels. Genet Epidemiol. 2000; 18:331-340.
8. Mooijaart, SP, Berbee, JF, van Heemst, D, et al. ApoE plasma levels and risk of cardiovascular mortality in old age. PLoS Med. 2006; 3:e176.
9. Bootsma van der Wiel, A, van Exel, E, de Craen, AJ, et al. A high response is not essential to prevent selection bias: results from the Leiden 85-plus study. J Clin Epidemiol. 2002; 55:1119- 1125.
10. Folstein, MF, Folstein, SE, and McHugh, PR. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975; 12:189-198.
11. Houx, PJ, Shepherd, J, Blauw, GJ, et al. Testing cognitive function in elderly populations: the PROSPER study. PROspective Study of Pravastatin in the Elderly at Risk. J Neurol Neurosurg Psychiatry. 2002; 73:385-389.
12. van der Elst, W, van Boxtel, MP, van Breukelen, GJ, et al. Rey’s verbal learning test: normative data for 1855 healthy participants aged 24-81 years and the infl uence of age, sex, education, and mode of presentation. J Int Neuropsychol Soc. 2005; 11:290-302.
13. Moller, JT, Cluitmans, P, Rasmussen, LS, et al. Long-term postoperative cognitive dysfunction in the elderly ISPOCD1 study. ISPOCD investigators. International Study of Post-Operative Cognitive Dysfunction. Lancet. 1998; 351:857-861.
14. Gueorguieva, R and Krystal, JH. Move over ANOVA: progress in analyzing repeated-measures data and its refl ection in papers published in the Archives of General Psychiatry. Arch Gen Psychiatry. 2004; 61:310-317.
15. Siest, G, Bertrand, P, Qin, B, et al. Apolipoprotein E polymorphism and serum concentration in Alzheimer’s disease in nine European centres: the ApoEurope study. ApoEurope group. Clin Chem Lab Med. 2000; 38:721-730.
16. Slooter, AJ, de Knijff, P, Hofman, A, et al. Serum apolipoprotein E level is not increased in Alzheimer’s disease: the Rotterdam study. Neurosci Lett. 1998; 248:21-24.
17. Taddei, K, Clarnette, R, Gandy, SE, et al. Increased plasma apolipoprotein E (apoE) levels in Alzheimer’s disease. Neurosci Lett. 1997; 223:29-32.
18. Scacchi, R, Gambina, G, Ruggeri, M, et al. Plasma levels of apolipoprotein E and genetic markers in elderly patients with Alzheimer’s disease. Neurosci Lett. 1999; 259:33-36.
19. Folin, M, Baiguera, S, Conconi, MT, et al. Apolipoprotein E as vascular risk factor in neurodegenerative dementia. Int J Mol Med. 2004; 14:609-613.
20. Neuropathology Group. MRC CFAS. Pathological correlates of late-onset dementia in a multicentre, community-based population in England and Wales. Neuropathology Group of the Medical Research Council Cognitive Function and Ageing Study (MRC CFAS). Lancet. 2001;
357:169-175.
21. van den Elzen, P, Garg, S, Leon, L, et al. Apolipoprotein-mediated pathways of lipid antigen presentation. Nature. 2005; 437:906-910.
22. Wilson, CJ, Finch, CE, and Cohen, HJ. Cytokines and cognition—the case for a head-to-toe inflammatory paradigm. J Am Geriatr Soc. 2002; 50:2041-2056.
23. Tuppo, EE and Arias, HR. The role of inflammation in Alzheimer’s disease. Int J Biochem Cell Biol. 2005; 37:289-305.
24. Yamauchi, K, Tozuka, M, Nakabayashi, T, et al. Apolipoprotein E in cerebrospinal fluid: relation to phenotype and plasma apolipoprotein E concentrations. Clin Chem. 1999; 45:497-504.
25. Carlsson, J, Armstrong, VW, Reiber, H, et al. Clinical relevance of the quantification of apolipoprotein E in cerebrospinal fluid. Clin Chim Acta. 1991; 196:167-176.
26. Shah, GN and Mooradian, AD. Age-related changes in the blood-brain barrier. Exp Gerontol.
1997; 32:501-519.
27. Skoog, I, Wallin, A, Fredman, P, et al. A population study on blood-brain barrier function in 85-year-olds: relation to Alzheimer’s disease and vascular dementia. Neurology. 1998; 50:966- 971.
28. Holtzman, DM, Fagan, AM, Mackey, B, et al. Apolipoprotein E facilitates neuritic and cerebrovascular plaque formation in an Alzheimer’s disease model. Ann Neurol. 2000; 47:739- 747.
29. Bales, KR, Verina, T, Cummins, DJ, et al. Apolipoprotein E is essential for amyloid deposition in the APP(V717F) transgenic mouse model of Alzheimer’s disease. Proc Natl Acad Sci U S A. 1999;
96:15233-15238.
30. Bales, KR, Verina, T, Dodel, RC, et al. Lack of apolipoprotein E dramatically reduces amyloid beta-peptide deposition. Nat Genet. 1997; 17:263-264.
31. Lund-Katz, S, Wehrli, S, Zaiou, M, et al. Effects of polymorphism on the microenvironment of the LDL receptor-binding region of human apoE. J Lipid Res. 2001; 42:894-901.
32. de Groot, JC, de Leeuw, FE, Oudkerk, M, et al. Cerebral white matter lesions and cognitive function: the Rotterdam Scan Study. Ann Neurol. 2000; 47:145-151.
33. van Boxtel, MP, Menheere, PP, Bekers, O, et al. Thyroid function, depressed mood, and cognitive performance in older individuals: the Maastricht Aging Study. Psychoneuroendocrinology. 2004;
29:891-898.
