Get Newsletter
AlzRisk Risk Factor Discussion
Back

Risk Factor:
  (BP, DBP, diastolic, diastolic blood pressure, hypertension, hypotension, pulse pressure, SBP, systolic, systolic blood pressure)
Risk Factor Type: Chronic disease, Metabolic
Current Understanding:
The tables below present inconsistent data on the relationship between blood pressure and AD and total dementia. There is a suggestion of an age-dependent relationship, i.e., hypertension may be harmful in midlife and protective in late-life. However, few studies have specifically considered the association between midlife blood pressure and AD, and it is possible that that bias (in particular reverse causation or selection bias) might account for any protective association in late-life. Further research is necessary and should focus on the impact of blood pressure in midlife, on potential modifying effect of anti-hypertensive drug use, and on quantifying the potential that selection bias or reverse causation might account for the observed associations. Despite the uncertainty regarding the relationship between blood pressure and risk of AD, the benefits of blood pressure control on cardiovascular risk are sufficient to justify current treatment standards. For a discussion of the putative mechanisms by which blood pressure may influence AD risk and commentary on interpreting the findings below in a broader context, please see the Discussion. A longer review and discussion can be found in the published review and meta-analysis, Power MC, Weuve J, Gagne JJ, McQueen MB, Viswanathan A, Blacker D. The association between blood pressure and incident Alzheimer disease: a systematic review and meta-analysis (Epidemiology 2011;22:646-659).
Literature Extraction: Search strategy  * New *
Last Search Completed: 01 November 2011 - Last content update released on 1 Nov 2012.

Risk Factor Overview

Cite as:

Power MC, Weuve J, Gagne JJ, McQueen MB, Viswanathan A, Blacker D. "Blood Pressure." The AlzRisk Database. Alzheimer Research Forum. Available at: http://www.alzrisk.org. Accessed [date of access]*.


***

INTRODUCTION

The tables in the Risk Factor Overview summarize a number of reports on the association between blood pressure and Alzheimer disease (AD). Taken as a whole, these reports do not suggest a consistent relationship between blood pressure and AD. However, the relationship may differ by age of exposure. There is a suggestion that hypertension in midlife may be associated with increased risk of AD, while hypertension in late life may be associated with reduced risk of AD (or alternately, hypotension in late life is associated with increased risk of AD). As discussed below, part of the difficulty in interpreting this body of literature lies in the potential that this pattern is attributable to other factors, including reverse causation and selection bias. Regardless, it is important to remember that blood pressure management in midlife is justified by its demonstrated benefits on cardiovascular health.

Please see the published review and meta-analysis, Power MC, Weuve J, Gagne JJ, McQueen MB, Viswanathan A, Blacker D. The association between blood pressure and incident Alzheimer disease: a systematic review and meta-analysis. Epidemiology (in press), for a more detailed discussion of many of the issues discussed here.

POTENTIAL MECHANISMS OF ACTION

Blood pressure is hypothesized to influence the development of clinical AD through several different mechanisms. Perhaps the best supported of these is through its influence on cerebrovascular pathology, which has been shown in multiple studies to increase the likelihood of manifesting clinical dementia for a given level of underlying AD pathology (e.g., Diaz-Ruiz et al., 2009; Esiri et al., 1999; Snowdon et al., 1997; Schneider et al., 2004). There may also be direct effects on AD pathology itself, as there is evidence that hypertension is correlated with higher levels of neurofibrilary tangles and neuritic plaques in the brain (e.g., Hoffman et al., 2009; Petrovitch et al., 2000; Sparks et al., 1995). For example, hypertension may lead to cerebral ischemia, which may potentiate aggregation or deposition of amyloid-beta peptides (Aß) through mechanisms such as increased Aß production (Banati et al., 1995; Shi et al., 2000) or slowed Aß clearance (Weller et al., 2002). Additionally, hypertension and cerebrovascular disease may be associated with greater degree of hippocampal (e.g., den Heijer et al., 2005; Firbank et al., 1991; Korf et al., 2004; O’Sullivan et al. 2009) and cortical atrophy (Jouvent et al. 2009). The mechanism underlying enhanced atrophy is unclear, but may be related to the effect of vascular disease on cortical neuronal apoptosis (Viswanathan et al., 2006).


METHODOLOGICAL ISSUES

Exposure

The tables report on hypertension (directory assessed and self-reported histories during varying times over adulthood) and both systolic and diastolic blood pressure. Each of these measures has inherent advantages and disadvantages in terms of the relevance of the type of blood pressure and blood-pressure-related pathology they capture in relation to AD risk.

Type of blood pressure assessed. Generally, reports focused on one of four measurements of blood pressure: history of hypertension, hypertension at baseline, systolic blood pressure, and diastolic blood pressure. However, some reports that reported findings for “hypertension” did not provide information on the timing of hypertension, which may be important given the weak evidence for an age-dependent association between AD and measures of blood pressure.

Method of exposure assessment. Reports varied widely in how blood pressure was measured. Those considering systolic or diastolic blood pressure generally relied on direct measurement on a single occasion. Those considering current hypertension typically relied on self-report (sometimes augmented by review of current medications) or direct measurement (using standard but not always identical thresholds, e.g., systolic >140 mm Hg, diastolic >90 mm Hg). Finally, reports considering whether participants had a history of hypertension at enrollment typically relied on self-report. All three methods may introduce measurement error. For example, given day-to-day variation in blood pressure, direct measurement of blood pressure on a single day may not reflect “typical” blood pressure for a given individual during that period of life. Similarly, methods relying on self-report or medication use may not accurately classify persons as hypertensive or having a history of hypertension due to poor recall of diagnosis or medication use, or use of antihypertensive medications for another condition. However, any errors in the measurement of blood pressure or hypertension should be random relative to AD diagnosis given that these studies all assessed blood pressure prior to AD diagnosis (our database includes only prospective studies). (See “Reverse Causation” below for a potential violation of this assumption). Although this measurement “noise” would not be expected to create a spurious association, it might contribute to our inability to draw firm conclusions about the relationship between blood pressure and AD.

Timing of exposure assessment. Another major issue is the timing of blood pressure measurement, as there may be a critical window for exposure. If blood pressure at a specific time of life (e.g., mid-life) and AD are related, only those studies reporting on the relationship between AD and blood pressure measurements taken during this period would be expected to show an association. Therefore, we cannot rule out a true age-dependent effect of blood pressure on AD given that there is some suggestion of a difference in the direction and magnitude of the association across age at blood pressure measurement.

Design and Analysis

Bias in outcome assessment. Information on blood pressure is often used in the differential diagnosis of AD. If participants with higher blood pressure who actually have AD were wrongly classified as having vascular dementia, this could induce a downward bias, leading to a reduction or reversal of a true adverse effect of blood pressure on AD.

How the exposure is modeled. Reports using measures of systolic or diastolic blood pressure chose to use this information in one of two ways: either they categorized blood pressure measurements and examined the relationship between AD and being in a particular category of blood pressure or they computed the association between AD and an x-unit increase in blood pressure, assuming a log-linear relationship. Categorization of blood pressure allows investigation of a non-linear relationship across a range of blood pressure measurements. However, the choice of category cut-points was inconsistent across reports, making comparison difficult, and the choice of category cut points may obscure a true relationship if one is truly present. For example categorization of blood pressure into “severely hypertensive,” “moderately hypertensive” and “not hypertensive” would make it difficult to observe a relationship between blood pressure and AD in which both hypertension and hypotension increase risk of AD relative to normal blood pressure (i.e., a U-shaped association), because individuals with normal blood pressure and those with hypotension would be included in the same category. Similarly, the assumption of a log-linear relationship would be inappropriate for characterizing this U-shaped relationship, and, more generally, would be suboptimal whenever the relationship was not log-linear.

Selection. All observational studies of risk factors for AD among older adults potentially suffer from two potential sources of bias due to non-participation. First, those who choose and are able to enroll in the study may be different from those who do not. Second, those who continue to participate after initial enrollment may be different from those who die or are otherwise unable or unwilling to continue participation. Although selection may be less of a concern (or at least quantifiable) in well-designed cohort studies than in other types of studies, initial participation and willingness to continue follow-up is still affected by overall health status and other factors.

If participation is related to both blood pressure and AD (or factors associated with each), then the reported associations are likely to be biased. Poor cardiovascular health can reduce participation and hypertension is related to both cardiovascular morbidity and mortality (e.g., Chobanian et al., 2003; Miura et al., 2001; Psaty et al., 2001). Poor or declining cognitive function also predicts both mortality and non-participation (e.g., Bassuk et al., 2000; Euser et al., 2008). Thus, nonparticipation is related to both exposure and outcome, which, in this case, would lead to an underestimation of the association—or even potentially a reversal of the association. Specifically, differential selection of this nature could lead to conservative estimates of the adverse effect of midlife hypertension and the appearance of a protective effect of late-life hypertension (and an adverse effect of hypotension) on AD risk.

Effect Modification. Use of antihypertensive medications to control hypertension, either at the time of blood pressure assessment or thereafter, may modify a true effect of blood pressure on AD. For example, one report (Launer et al., 2000) found a stronger association between measures of midlife blood pressure and AD among those who had never been treated with antihypertensive medications during the study period than among those who had undergone treatment. However, few reports have considered the impact of antihypertensive medication use (and in fact several use it as a way to detect a diagnosis of hypertension), which may account for some of the inconsistency across studies. In addition, there is some suggestion that certain antihypertensive medications may have an independent effect on AD or cognitive impairment distinct from their effect on blood pressure (e.g. Fournier et al. 2009; Tzourio et al. 1999; Wang et al. 2007). However, further investigation is required to confirm this effect and determine which classes of antihypertensive agents, if any, have this independent effect.

Reverse Causation. In order to establish a causal relationship between blood pressure and initiation of AD, the measurement of blood pressure must precede the onset of the disease. It is generally accepted that AD pathology begins several years prior to AD diagnosis. While measures of midlife blood pressure were generally taken over 10 years prior to AD diagnosis, measures of late-life blood pressure were typically taken within approximately 5 years of diagnosis. Therefore, the suggestion of an inverse association between late-life blood pressure and AD risk may not reflect a true protective effect of hypertension on risk of AD at older ages. Instead, AD pathogenesis may influence blood pressure regulation, leading to reduction in blood pressure in the presence of preclinical AD (Qiu et al., 2004). On the other hand, at least up to a point, higher blood pressure in late life may be a sign of overall health and, in particular, of sufficient cardiac output to ensure adequate perfusion, suggesting that the protective association may be real.


RESULTS FROM OTHER LINES OF RESEARCH

Other studies have considered the relationship between blood pressure and cognitive decline or total dementia. Interestingly, there also appears to be an age-dependent association between blood pressure and cognitive function or dementia (for review, see Qiu et al., 2004). As expected given the established association between cognition and cerebrovascular pathology or disease, studies of midlife blood pressure and total dementia or cognitive decline consistently suggest increases in both of these outcomes with midlife hypertension. The association between late-life blood pressure and cognitive decline or total dementia is less consistent. A majority of studies considering the association between late-life blood pressure and cognitive impairment or decline observed either an increased risk of these cognitive outcomes with late-life hypertension or a U-shaped relationship, with increased risk for both hypertensive and hypotensive participants. As with measures of mid-life blood pressure, this pattern may be due to the impact of hypertension on vascular-related cognitive impairment. However, this pattern is not present in studies considering the association between late-life blood pressure and total dementia, perhaps because AD is the most common form of dementia. As many of the studies considering the association between late-life blood pressure and total dementia are conducted in the same cohorts as those considering the association between late-life blood pressure and AD, the potential that bias accounts for the suggestion of an inverse association may explain the findings for total dementia as well.

Several randomized placebo-controlled clinical trials have considered the relationship between pharmacologic treatment of hypertension and cognition or dementia. Most, (Forette et al. 1998; Forette et al. 2002; Lithell et al. 2003; Peters et al. 2008; Tzourio et al. 2003) but not all, (Applegate et al. 1994; Prince et al., 1996) reported an inverse association between antihypertensive drug use and the development of dementia or cognitive decline. However, it is difficult to state that this evidence supports an adverse effect of hypertension on AD. On the one hand, such trials may better evaluate whether treatment of hypertension, and by extension, hypertension itself, influences AD onset and progression given that these studies may be less susceptible to selection bias (the trials were restricted to hypertensive individuals, random assignment of treatment dissociates treatment from enrollment probability, and short follow-up limits the potential for loss to follow-up associated with treatment). Alternately, certain antihypertensive drugs may have an effect on risk of dementia that is independent of their effect on blood pressure as previously discussed, or, as these trials generally considered only the broader outcomes of dementia or cognitive decline, the observed results may also be attributable to the effect of pharmacologic treatment of blood pressure on vascular cognitive impairment, rather than on AD onset or progression.

The association between antihypertensive drug use and cognitive decline, dementia, or AD has also been considered in several large, prospective epidemiologic studies. Collectively, the evidence from the epidemiologic literature also suggests a protective effect of both midlife and late-life antihypertensive drug use on risk of dementia or AD (Haag et al. 2009; Li et al. 2011). However, as these studies generally had follow-up times of less than 5 years, it is possible that these associations are attributable to reverse causation; persons with prodromal AD may experience a decline in blood pressure, making antihypertensive medication use less likely. Alternately, as previously discussed, antihypertensive drug use may have an effect on risk of dementia or AD that is independent of their effect on hypertension. Finally, there is emerging evidence that blood pressure may be related to the rate of cognitive decline in patients diagnosed with AD (Mielke et al. 2007). Again, this may be attributable to either an additive or synergistic effect of blood pressure-related cerebrovascular pathology on cognitive impairment in those with AD.


DISCUSSION AND RECOMMENDATIONS

Overall, these data fail to indicate a consistent relationship between blood pressure and AD. While the data suggest that there may be an age-dependent relationship between blood pressure and AD, the current data are limited and are therefore insufficient to draw firm conclusions. Furthermore, the suggestion of an inverse relationship between late-life blood pressure and AD may be attributable to reverse causation or non-participation. It will be critical for future research to focus on midlife measures of blood pressure, to consider the potential impact of anti-hypertensive drug use, and to attempt to account for the potential impact of differential participation on observed findings. In the meantime, the overall benefits of blood pressure control are sufficient to justify its use on cardiovascular grounds alone, and the findings on cognition and total dementia suggest there may be additional benefits, at least for treatment in midlife.


REFERENCES CITED

Applegate WB, Pressel S, Wittes J, Luhr J, Shekelle RB, Camel GH, Greenlick MR, Hadley E, Moye L, Perry HM, Jr., et al. Impact of the treatment of isolated systolic hypertension on behavioral variables. Results from the systolic hypertension in the elderly program. Arch Intern Med 1994;154(19):2154-60.

Banati RB, Gehrmann J, Wiessner C, Hossmann KA, Kreutzberg GW. Glial expression of the beta-amyloid precursor protein (APP) in global ischemia. J Cereb Blood Flow Metab. 1995;15(4):647-54.

Bassuk SS, Wypij D, Berkman LF. Cognitive impairment and mortality in the community-dwelling elderly. Am J Epidemiol. Apr 1 2000;151(7):676-688.

Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL, Jr., Jones DW, Materson BJ, Oparil S, Wright JT, Jr., Roccella EJ. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. JAMA. May 21 2003;289(19):2560-2572.

den Heijer T, Launer LJ, Prins ND, van Dijk EJ, Vermeer SE, Hofman A, Koudstaal PJ, Breteler MM. Association between blood pressure, white matter lesions, and atrophy of the medial temporal lobe. Neurology. Jan 25 2005;64(2):263-267.

Diaz-Ruiz C, Wang J, Ksiezak-Reding H, Ho L, Qian X, Humala N, Thomas S, Martinez-Martin P, Pasinetti GM. Role of hypertension in aggravating AB neuropathology of AD type and tau-mediated motor impairment. Cardiovascular Psychiatry and Neurology. Sept 17 2009: Article ID 107286, 9 pages.

Esiri MM, Nagy Z, Smith MZ, Barnetson L, Smith AD. Cerebrovascular disease and threshold for dementia in the early stages of Alzheimer's disease. Lancet. Sep 11 1999;354(9182):919-920.

Euser SM, Schram MT, Hofman A, Westendorp RG, Breteler MM. Measuring cognitive function with age: the influence of selection by health and survival. Epidemiology. May 2008;19(3):440-447.

Firbank MJ, Wiseman RM, Burton EJ, Saxby BK, O'Brien JT, Ford GA. Brain atrophy and white matter hyperintensity change in older adults and relationship to blood pressure. Brain atrophy, WMH change and blood pressure. J Neurol. Jun 2007;254(6):713-721.

Forette F, Seux ML, Staessen JA, Thijs L, Babarskiene MR, Babeanu S, Bossini A, Fagard R, Gil-Extremera B, Laks T, Kobalava Z, Sarti C, Tuomilehto J, Vanhanen H, Webster J, Yodfat Y, Birkenhager WH. The prevention of dementia with antihypertensive treatment: new evidence from the Systolic Hypertension in Europe (Syst-Eur) study. Arch Intern Med 2002;162(18):2046-52.

Forette F, Seux ML, Staessen JA, Thijs L, Birkenhager WH, Babarskiene MR, Babeanu S, Bossini A, Gil-Extremera B, Girerd X, Laks T, Lilov E, Moisseyev V, Tuomilehto J, Vanhanen H, Webster J, Yodfat Y, Fagard R. Prevention of dementia in randomised double-blind placebo-controlled Systolic Hypertension in Europe (Syst-Eur) trial. Lancet 1998;352(9137):1347-51.

Fournier A, Oprisiu-Fournier R, Serot JM, Godefroy O, Achard JM, Faure S, Mazouz H, Temmar M, Albu A, Bordet R, Hanon O, Gueyffier F, Wang J, Black S, Sato N. Prevention of dementia by antihypertensive drugs: how AT1-receptor-blockers and dihydropyridines better prevent dementia inhypertensive patients than thiazides and ACE-inhibitors. Expert Rev Neurother. Sept 2009;9(9):1413-31.

Haag MD, Hofman A, Koudstaal PJ, Breteler MM, Stricker BH. Duration of antihypertensive drug use and risk of dementia: A prospective cohort study. Neurology 2009;72(20):1727-34.

Hoffman LB, Schmeidler J, Lesser GT, Beeri MS, Purohit DP, Grossman HT, Haroutunian V. Less Alzheimer disease neuropathology in medicated hypertensive than nonhypertensive persons. Neurology. May 19 2009;72(20):1720-1726.

Jouvent E, Viswanathan A, Chabriat H. Cerebral atrophy in cerebrovascular disorders. J Neuroimaging. 2010;20(3):213-8.

Korf ES, White LR, Scheltens P, Launer LJ. Midlife blood pressure and the risk of hippocampal atrophy: the Honolulu Asia Aging Study. Hypertension. Jul 2004;44(1):29-34.

Launer LJ, Ross GW, Petrovitch H, Masaki K, Foley D, White LR, Havlik RJ. Midlife blood pressure and dementia: the Honolulu-Asia aging study. Neurobiol Aging. Jan-Feb 2000;21(1):49-55.

Li J, Wang YJ, Zhang M, Xu ZQ, Gao CY, Fang CQ, Yan JC, Zhou HD. Vascular risk factors promote conversion from mild cognitive impairment to Alzheimer disease. Neurology. 2011; in press.

Lithell H, Hansson L, Skoog I, Elmfeldt D, Hofman A, Olofsson B, Trenkwalder P, Zanchetti A. The Study on Cognition and Prognosis in the Elderly (SCOPE): principal results of a randomized double-blind intervention trial. J Hypertens 2003;21(5):875-86.

Mielke MM, Rosenberg PB, Tschanz J, Cook L, Corcoran C, Hayden KM, Norton M, Rabins PV, Green RC, Welsh-Bohmer KA, Breitner JC, Munger R, Lyketsos CG. Vascular factors predict rate of progression in Alzheimer disease. Neurology 2007;69(19):1850-8.

Miura K, Daviglus ML, Dyer AR, Liu K, Garside DB, Stamler J, Greenland P. Relationship of blood pressure to 25-year mortality due to coronary heart disease, cardiovascular diseases, and all causes in young adult men: the Chicago Heart Association Detection Project in Industry. Arch Intern Med. Jun 25 2001;161(12):1501-1508.

O'Sullivan M, Ngo E, Viswanathan A, Jouvent E, Gschwendtner A, Saemann PG, Duering M, Pachai C, Bousser MG, Chabriat H, Dichgans M. Hippocampal volume is an independent predictor of cognitive performance in CADASIL. Neurobiol Aging. 2009;30(6):890-7.

Peters R, Beckett N, Forette F, Tuomilehto J, Clarke R, Ritchie C, Waldman A, Walton I, Poulter R, Ma S, Comsa M, Burch L, Fletcher A, Bulpitt C. Incident dementia and blood pressure lowering in the Hypertension in the Very Elderly Trial cognitive function assessment (HYVET-COG): a double-blind, placebo controlled trial. Lancet Neurol 2008;7(8):683-9

Petrovitch H, White LR, Izmirilian G, Ross GW, Havlik RJ, Markesbery W, Nelson J, Davis DG, Hardman J, Foley DJ, Launer LJ. Midlife blood pressure and neuritic plaques, neurofibrillary tangles, and brain weight at death: the HAAS. Honolulu-Asia aging Study. Neurobiol Aging. Jan-Feb 2000;21(1):57-62.

Power MP, Weuve J, Gagne JJ, McQueen MB, Viswanathan A, Blacker D. The association between blood pressure and incident Alzheimer disease: a systematic review and meta-analysis. Epidemiology. In press.

Prince MJ, Bird AS, Blizard RA, Mann AH. Is the cognitive function of older patients affected by antihypertensive treatment? Results from 54 months of the Medical Research Council's trial of hypertension in older adults. BMJ 1996;312(7034):801-5.

Psaty BM, Furberg CD, Kuller LH, Cushman M, Savage PJ, Levine D, O'Leary DH, Bryan RN, Anderson M, Lumley T. Association between blood pressure level and the risk of myocardial infarction, stroke, and total mortality: the cardiovascular health study. Arch Intern Med. May 14 2001;161(9):1183-1192.

Qiu C, von Strauss E, Winblad B, Fratiglioni L. Decline in blood pressure over time and risk of dementia: a longitudinal study from the Kungsholmen project. Stroke. Aug 2004;35(8):1810-1815.

Qiu C, Winblad B, Fratiglioni L. The age-dependent relation of blood pressure to cognitive function and dementia. Lancet Neurol. Aug 2005;4(8):487-499.

Schneider JA, Wilson RS, Bienias JL, Evans DA, Bennett DA. Cerebral infarctions and the likelihood of dementia from Alzheimer disease pathology. Neurology. Apr 13 2004;62(7):1148-1155.

Shi J, Yang SH, Stubley L, Day AL, Simpkins JW. Hypoperfusion induces overexpression of beta-amyloid precursor protein mRNA in a focal ischemic rodent model. Brain Res. 2000;853(1):1-4.

Snowdon DA, Greiner LH, Mortimer JA, Riley KP, Greiner PA, Markesbery WR. Brain infarction and the clinical expression of Alzheimer disease. The Nun Study. JAMA. Mar 12 1997;277(10):813-817.

Sparks DL, Scheff SW, Liu H, Landers TM, Coyne CM, Hunsaker JC, 3rd. Increased incidence of neurofibrillary tangles (NFT) in non-demented individuals with hypertension. J Neurol Sci. Aug 1995;131(2):162-169.

Tzourio C, Anderson C, Chapman N, Woodward M, Neal B, MacMahon S, Chalmers J. Effects of blood pressure lowering with perindopril and indapamide therapy on dementia and cognitive decline in patients with cerebrovascular disease. Arch Intern Med 2003;163(9):1069-75.

Tzourio C, Dufouil C, Ducimetiere P, Alperovitch A. Cognitive decline in individuals with high blood pressure: a longitudinal study in the elderly. EVA Study Group. Epidemiology of Vascular Aging. Neurology. 1999;53(9):1948-52.

Viswanathan A, Gray F, Bousser MG, Baudrimont M, Chabriat H. Cortical neuronal apoptosis in CADASIL. Stroke. 2006;37(11):2690-5.

Wang J, Ho L, Chen L, Zhao Z, Zhao W, Qian X, Humala N, Seror I, Bartholomew S, Rosendorff C, Pasinetti GM. Valsartan lowers brain ß-amyloid protein levels and improves spatial learning in a mouse model of Alzheimer disease. J Clin Invest. Nov 1 2007;117(11):3393-3402.

Weller RO, Yow HY, Preston SD, Mazanti I, Nicoll JA. Cerebrovascular disease is a major factor in the failure of elimination of Abeta from the aging human brain: implications for therapy of Alzheimer's disease. Ann N Y Acad Sci. 2002;977:162-8.