Occurrence and Progression of Dementia in a Community Population Aged 75 Years and Older: Relationship of Antihypertensive Medic

ORIGINAL CONTRIBUTION Occurrence and Progression of Dementia in a Community Population Aged 75 Years and Older Relationship of Antihypertensive Medication Use

Zhenchao Guo, MD, PhD; Laura Fratiglioni, MD, PhD; Li Zhu, MD; Johan Fastbom, MD, PhD; Bengt Winblad, MD, PhD; Matti Viitanen, MD, PhD

Objective: To examine whether antihypertensive medi- (adjusted relative risk, 0.7; 95% confidence interval, 0.6- cation use can affect the occurrence and progression of 1.0; P = .03). Furthermore, subjects taking diuretics dementia. (n = 484) had an adjusted relative risk of 0.7 (95% confidence interval, 0.5-1.0; P = .02) for all dementia, and Subjects and Methods: In a community cohort of subjects taking diuretic monotherapy (n = 345) had an 1810 persons aged 75 years and older, 225 prevalent adjusted relative risk of 0.6 (95% confidence interval, 0.4- cases of dementia were detected. Among the 1301 per- 0.9; P = .006). The use of other antihypertensive medi- sons without dementia, 224 incident cases of dementia cation (calcium antagonists or ␤-blockers), however, was were identified during an average period of 3 years. related to a reduced risk of Alzheimer disease (adjusted Among the 225 prevalent cases of dementia, 79 were relative risk, 0.6; 95% confidence interval, 0.3-1.2) only suitable for the analysis of cognitive decline. Information in the subpopulation with a higher baseline blood pres- on drug use was collected for the 2 weeks preceding the sure (n = 458). Patients with dementia at baseline who baseline interview. were not taking diuretics had a 2-fold faster rate of decline in the score on the Mini-Mental State Examination Results: Subjects taking antihypertensive medication than those taking diuretics. (n = 651, 83.9% of whom took diuretics) had a lower prevalence of dementia than those not taking antihyper- Conclusion: The use of diuretics may protect against de- tensive medication (PϽ.001). Subjects without demen- mentia in elderly persons. tia who were taking antihypertensive medication at baseline (n = 584) had a reduced incidence of dementia Arch Neurol. 1999;56:991-996

F DEMENTIA with a vascular com- in decreasing the risk of dementia in the ponent comprises nearly half of general population,14 although several all cases of dementia in persons clinical trials15,16 have examined the ef- aged 85 years and older,1 a sub- fects of these drugs on cognition. stantial portion of cases of de- We examine whether antihyperten- Imentia may be prevented in this age group. sive medication use affects the occur- Among many vascular factors or diseases rence of dementia and cognitive decline that may be related to dementia,2 hyper- in patients with dementia in a community- tension has been suggested3 as the most based cohort aged 75 years and older. important vascular risk factor. In fact, hypertension may not only increase the risk RESULTS of vascular dementia (VaD) by causing is- chemic stroke and other cerebrovascular The mean age at baseline of 1810 partici- events or lesions,4-10 it may also be in- pants in the Kungsholmen Project was 82.5 volved in the pathogenesis of Alzheimer years (range, 75-101 years), 76.1% were disease (AD).11 This hypothesis is sup- women, and 53.0% had less than 8 years ported by a recent study12 showing that of education (only 2 persons had less than brain infarction increases the clinical ex- 4 years of education). The mean (SD) sys- pression of AD and a report13 indicating tolic and diastolic blood pressures were From the Department of that cerebral infarction in AD is associ- 154 (22) and 81 (11) mm Hg, respec- Geriatric Medicine, Karolinska ated with severe amyloid angiopathy and tively. The prevalence of heart disease, Institute and Stockholm hypertension. Despite these findings, no stroke, and dementia was 17.6%, 8.5%, and Gerontology Research Center, data are available regarding the possible 12.4%, respectively. The baseline mean Stockholm, Sweden. effects of antihypertensive medication use MMSE score was 24.5 (6.2).

ARCH NEUROL / VOL 56, AUG 1999 991

©1999 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 SUBJECTS AND METHODS functional, and medical aspects and also included an assess- ment of the performance of activities of daily living. The physicians examined the physical and neurologic conditions, as STUDY SUBJECTS well as the cognitive and psychiatric states. For the cognitive examination, memory functions were explored by ask- Data for this study came from the baseline and first fol- ing facts of general knowledge and past personal informa- low-up examinations of the Kungsholmen Project,17 which tion, language functions by assessing subjects’ object- is a longitudinal study of aging and dementia. The study was naming ability and comprehension, abstract thinking by approved by the ethics committee of the Karolinska Insti- assessing subjects’ problem-solving ability and analysis of prov- tute, Stockholm, Sweden. The study population included all erbs, praxis function by examining simple motor activities inhabitants of the Kungsholmen district of Stockholm who (dressing and pantomime), and visuospatial orientation by on October 1, 1987, were aged 75 years or older. At base- assessing subjects’ performance on copying figures. Psycho- line, 1810 subjects were administered the Mini-Mental State logical tests included the MMSE, episodic and primary Examination (MMSE),18 and then subjects with MMSE scores memory tasks such as free recall and the recognition of ran- of less than 24 (n = 314) and a random sample of subjects with dom words, and digit span tests. Diagnostic criteria of the Di- MMSE scores of 24 or higher (n = 354) were extensively ex- agnostic and Statistical Manual of Mental Disorders, Revised, amined.19 By the 2-phase design, which included a screen- Third Edition,21 were used to define dementia. The diagnosis ing phase and a clinical examination phase, 225 prevalent cases of dementia was given after consensus among 3 indepen- of dementia were detected (216 derived from subjects whose dent physicians (L.F., B.W., and M.V.). dementia was detected during screening and 9 from the ex- Dementia severity was determined according to the amined random sample of subjects whose screening results Clinical Dementia Rating Scale,22 with some modifica- were normal). One hundred ten persons refused to partici- tions.23 For analysis, we treated severity of the disease as a pate in the baseline clinical examination, and 174 persons category variable by artificially coding questionable, mild, (2 with a mental disorder) refused to participate in the fol- moderate, and severe dementia as 1, 2, 3, and 4, respec- low-up examination or were unavailable during the fol- tively. No information regarding drug use was used in the low-up period. Of the 1301 subjects without dementia who diagnosis of dementia. For the dead subjects during the fol- were eligible for the follow-up evaluation,20 314 died before low-up period, however, information from medical rec- the follow-up examination. The remaining subjects (n = 987) ords and death certificates was used to make dementia di- were administered a comprehensive clinical examination be- agnoses. Details of the clinical examination and diagnostic tween November 1990 and April 1992. procedures have been reported elsewhere.19,20

DIAGNOSING DEMENTIA BASELINE DATA COLLECTION

The same protocol was used to gather the information for The baseline interview was conducted by trained nurses from the diagnosis of dementia at baseline and at follow-up. It October 1987 to December 1989. Information on drug use included the collection of family and personal histories by was collected for the 2 weeks preceding the baseline inter- nurses, clinical examination by physicians, and the ad- view.24 Both prescription and nonprescription drug use were ministration of psychological tests by trained personnel. inquired about; drug containers and drug prescription forms The interview by nurses involved social, occupational, were inspected to verify this information. Drug use was

PREVALENCE OF DEMENTIA INCIDENCE OF DEMENTIA

The baseline characteristics of the study population ac- Among the 1301 persons without dementia who partici- cording to the use of antihypertensive medication are pated in the follow-up evaluation, 987 (75.9%) received shown in Table 1. Subjects taking diuretics were older a comprehensive clinical examination, and 199 were di- and had a higher systolic blood pressure; there was also agnosed as having dementia. Among the 314 subjects who a higher proportion with less than 8 years of education died during the follow-up interval, 25 were diagnosed and with a history of heart disease or stroke. They were as having dementia, using data from medical records and more likely to be female than those not taking antihy- death certificates. The mean follow-up of the cohort was pertensive medication. 36.7 months, with a maximum of 63.1 months. At baseline, 225 cases of dementia were detected, Among the 1301 subjects, 584 (44.9%) took at least giving an overall prevalence of 12.4%. Compared with 1 type of antihypertensive drug at baseline. Subjects tak- those not taking antihypertensive medication, subjects ing antihypertensive medication had an adjusted RR of taking diuretics had a higher score on the MMSE (P = .006) 0.7 (95% CI, 0.6-1.0; P = .03) for dementia. and a lower prevalence of dementia (PϽ.004). Logistic At baseline, 484 (37.2%) persons took diuretics. Sub- regression analysis showed that subjects taking diuret- jects taking diuretics were older (age [mean ± SD], ics had an odds ratio of 0.4 (95% confidence interval [CI], 82.6 ± 5.1 years vs 81.7 ± 4.9 years, respectively), more 0.3-0.6) for dementia, after adjustment for age, sex, edu- likely to be female (81.4% vs 70.9%), and had a higher cation, systolic blood pressure, heart disease, and stroke. prevalence of heart disease (26.9% vs 8.1%) and stroke Subjects taking other antihypertensive drugs but no di- (9.5% vs 4.9%) than those not taking antihypertensive uretics also had a lower prevalence of dementia, with an medication. There was no difference in the MMSE scores odds ratio of 0.3 (95% CI, 0.1-0.8). between those taking diuretics and those not taking an-

ARCH NEUROL / VOL 56, AUG 1999 992

©1999 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 classified according to the Anatomical Therapeutic Chemi- confounders in the association between the use of antihy- cal classification system.25 Antihypertensive drugs included pertensive medication and prevalent dementia. all medicines potentially used for lowering blood pressure The incidence of dementia was calculated by dividing (Anatomical Therapeutic Chemical codes C02, C03, and C07). the number of cases by the number of person-years of follow- For analysis, we first divided the whole sample into 3 groups: up. The follow-up time for persons without dementia was de- the group using diuretics, regardless of other drug use; the termined from the date of the baseline interview to the date group using other antihypertensive drugs, no diuretics; and of the follow-up examination or death. For the persons with the group not using any antihypertensive drugs. A portion dementia, half of this time was assumed. We used Cox pro- of diuretic users took other antihypertensive drugs at the same portional hazards regression analysis to estimate the relative time. We also checked the subjects receiving diuretic mono- risk (RR) of dementia developing in relation to the use of an- therapy separately. The drugs that were used by more than tihypertensive medication. We performed the analyses in the 20 subjects included diuretics (bendroflumethiazide, hydro- entire study population (n = 1301), in the subpopulation with chlorothiazide, furosemide, spironolactone, and amiloride hy- a baseline MMSE score of greater than 23 (n = 1212), in the drochloride), calcium channel antagonists (verapamil hy- subpopulation who participated in the follow-up clinical ex- drochloride, nifedipine, and diltiazem hydrochloride), and amination (n = 987), and in the subpopulation with a base- ␤-adrenergic blocking agents (metoprolol tartrate, proprano- line systolic blood pressure of greater than 160 mm Hg or a lol hydrochloride, alprenolol hydrochloride, and atenolol). diastolic blood pressure of greater than 95 mm Hg (n = 458). Among them, bendroflumethiazide, furosemide, amiloride, When the age at the onset of dementia was used as the time- and verapamil were used by more than 100 subjects. When scale in the Cox proportional hazards regression model, as the subject was unable to provide reliable information, as as- suggested by Korn et al,27 similar results were obtained. There- sessed by the subject’s response to the first few items of the fore, the results from the models in which the follow-up time questionnaire and to all items of the MMSE, a proxy respon- was used are presented here. dent (relative, caregiver, or other) was used. The rate of decline in the score on the MMSE among The arterial blood pressure (systolic: Korotkoff phase patients with dementia at baseline was calculated using the 1, and diastolic: phase 5) was measured with a mercury following formula: (baseline score − follow-up score) sphygmomanometer with the subject in a sitting position ÷ baseline score ÷ years of follow-up. A multiple stepwise after a 5-minute rest.26 Information regarding the medical linear regression analysis was used to identify the signifi- history for each participant was obtained from the com- cant predictors for the rate of cognitive decline. We also puterized inpatient register, which covers all hospitals in performed other analyses by establishing several general- the Stockholm area. We treated heart disease (coronary heart ized models. None of these models identified more predic- disease, cardiac dysrhythmia, and heart failure) and stroke tors than the linear model at the same significant level. as 2 potential confounders. Age, sex, education, systolic (or diastolic) blood pressure, heart disease, and stroke were considered as covari- DATA ANALYSIS ates in all the analyses. Although we reanalyzed our data taking into account the use of estrogens, nonsteroidal anti- The difference of the score on the MMSE and the preva- inflammatory drugs, aspirin, vitamins E and C, and apoli- lence of dementia between drug use groups were exam- poprotein E genotypes, we did not present the results in ined, respectively, by the Student t test and the ␹2 test. Lo- this study because none of these factors changed the re- gistic regression analysis was used to adjust for potential sults substantially.

tihypertensive medication (26.5 ± 2.9 vs 26.6 ± 2.7, re- subjects receiving diuretic monotherapy (n = 122) had spectively). Subjects taking diuretics had a significantly an adjusted RR of 0.7 (95% CI, 0.5-1.2) for dementia, reduced incidence of dementia (P = .02) (Table 2). The and subjects taking other antihypertensive drugs reduced risk of dementia was more pronounced among (n = 94, 42 subjects taking calcium channel antagonists those receiving diuretic monotherapy. and 43 taking ␤-blockers; 19 also with diuretics) had an In the subpopulation with a baseline MMSE score adjusted RR of 0.6 (95% CI, 0.3-1.2). of greater than 23 (n = 1212), subjects taking diuretics had an adjusted RR of 0.7 (95% CI, 0.5-0.9) for COGNITIVE DECLINE AMONG PERSONS dementia, and subjects receiving diuretic mono- WITH DEMENTIA therapy had an adjusted RR of 0.6 (95% CI, 0.4-0.8). In addition, the inclusion of baseline MMSE scores in Among the 225 subjects with prevalent dementia, 115 the models produced similar results regarding the use died before the follow-up examination. Among the 110 of diuretics. persons who participated in the follow-up evaluation, 28 In the subpopulation who participated in the fol- with a baseline MMSE score of less than 6 were ex- low-up clinical examination (n = 987, excluding those who cluded from further analyses. We also excluded 3 per- died during the period of follow-up), subjects taking di- sons who did not use diuretics but who took other an- uretics had an adjusted RR of 0.7 (95% CI, 0.5-0.96), and tihypertensive drugs because the number was too small subjects receiving diuretic monotherapy had an ad- to be analyzed separately. There was no difference in the justed RR of 0.6 (95% CI, 0.4-0.9). baseline MMSE score between those taking and those not In the subpopulation with a baseline systolic blood taking diuretics. Subjects not taking diuretics had a de- pressure of greater than 160 mm Hg or a diastolic cline of 17% every year from the baseline score, which blood pressure of greater than 95 mm Hg (n = 458), was 2 times more than the decline of those taking di-

ARCH NEUROL / VOL 56, AUG 1999 993

©1999 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 Table 1. Baseline Characteristics of the Table 3. Characteristics and Mini-Mental State Examination 1810 Participants in the Kungsholmen Project, Score of 79 Patients With Dementia at Baseline by Use of Antihypertensive Medication* of the Kungsholmen Project, by Diuretic Use*

Use of Antihypertensive Medication Diuretic No Antihypertensive Use† Medication Use Diuretics Other None Characteristics (n = 23) (n = 56) Variable (n = 651)† (n = 125)‡ (n = 1034) Age, mean ± SD, y 85.3 ± 5.9 84.1 ± 5.5 Age, y 83.1 ± 5.2 81.0 ± 4.3 82.3 ± 5.2 Female sex 17 (74) 45 (80) Female sex, % 81.4 72.8 73.2 Education Ͻ8 y 17 (74) 36 (64) Education Ͻ8 y, % 57.5 41.6 51.5 Heart disease 11 (48) 5 (9) Heart disease, % 30.1 19.2 9.6 Stroke 6 (26) 13 (23) Stroke, % 10.1 8.8 7.4 Blood pressure, mean ± SD, mm Hg Blood pressure, mm Hg§ Systolic 157 ± 26 144 ± 27 Systolic 156 ± 23 160 ± 25 153 ± 22 Diastolic 82 ± 18 77 ± 11 Diastolic 81 ± 11 82 ± 12 80 ± 11 Follow-up, mean ± SD, mo 34.5 ± 8.1 34.1 ± 6.3 Score on the MMSE 25.1 ± 5.2 26.5 ± 3.6 23.9 ± 7.1 Severity of the dementia‡ Dementia, % 9.4 4.8 15.3 Questionable 7 (30) 7 (13) Mild 7 (30) 21 (37) *Data are given as mean ± SD except where noted. MMSE indicates Moderate 9 (39) 28 (50) Mini-Mental State Examination. Mini-Mental State Examination, †Of this group, 70% were receiving diuretic monotherapy; 14.4% were mean ± SD also taking calcium channel antagonists, and 10.6% were taking -blockers. ␤ Baseline score 17.1 ± 5.5 16.8 ± 5.1 ‡Of this group, 52% were taking calcium channel antagonists, and 41.6% Follow-up score 13.4 ± 6.2 9.5 ± 6.8 were taking ␤-blockers. §Values were missing for 63 persons. Rate of decline§ 0.08 ± 0.13 0.17 ± 0.13

*Data are given as number (percentage) except where noted. †Eighteen persons were taking diuretic monotherapy; 3 were also taking Table 2. Use of Antihypertensive Medication calcium channel antagonists, and 2 were also taking ␤-blockers. and Incident Rate of Dementia in the First Follow-up ‡According to the Clinical Dementia Rating Scale. of the Kungsholmen Project §Calculated as follows: (baseline score − follow-up score) ÷ (baseline score) ÷ (years of follow-up). Antihypertensive No. of Incident Rate Relative Risk Medication Use Dementia per 1000 (95% Confidence sons aged 75 years and older in which several factors, in- at Baseline Cases Person-Years Interval)* cluding possible protectors for AD reported in the lit- Diuretics (n = 484)† 73 56.7 0.7 (0.5-1.0) erature,28 were considered. Diuretic monotherapy 48 52.5 0.6 (0.4-0.9) (n = 345) Other (n = 100)‡ 17 56.7 0.9 (0.5-1.5) POSSIBLE MECHANISMS None (n = 717) 134 66.2 1.0 (Reference) The antihypertensive effect of diuretics, especially in the *Adjusted for age, sex, education, systolic blood pressure, heart disease, elderly, has been proved.29 It is, therefore, conceivable and stroke. that taking diuretics may reduce the risk of dementia or †Of this group, 15.1% were also taking calcium channel antagonists, and 11.0% were also taking ␤-blockers. slow the progress of dementia by reducing the number ‡Of this group, 50% were taking calcium channel antagonists, and 45% of cerebrovascular events or silent vascular lesions in the were taking ␤-blockers. brain. We demonstrated that the use of diuretics was related to a 30% reduced risk of dementia in the general uretics (Table 3). By using a multiple stepwise linear population. The result is similar to that of a meta- regression analysis, the use of diuretics was found to be analysis30 that reported a 34% reduced risk of major ce- inversely related to the rate of cognitive decline (regres- rebrovascular events with low-dose diuretic therapy in sion coefficient = −0.07, P = .04). hypertensive patients. Hypertension-related silent vascular lesions in the brain, such as white matter changes, 31 COMMENT are common in persons with dementia (both AD and VaD), although they can also occur with normal aging. MAIN FINDINGS In addition, a recent study32 showed an increased incidence of senile plaques and neurofibrillary tangles in per- The use of diuretics was related to a lower prevalence of sons with hypertension. Controlling hypertension may dementia, a reduced incidence of dementia, and slower prevent the development and progression of these lesions. cognitive decline among patients with dementia. A re- Increasing evidence33 suggests that cerebrovascu- duced risk of dementia associated with the use of other lar dysfunction and damage may also be involved in the antihypertensive drugs (calcium channel antagonists or pathogenesis of AD. A recent study34 found that amy- ␤-blockers), however, was observed only in the sub- loid ␤-protein, the major component of amyloid plaques population with a baseline systolic blood pressure of in the AD brain,35,36 enhanced vasoconstriction and re- greater than 160 mm Hg or a diastolic blood pressure of sistance to the relaxation of endothelial cells in intact rat greater than 95 mm Hg. The results were derived from a aorta. In vitro, this toxicity of amyloid ␤-protein can be population-based epidemiological study of 1810 per- partially prevented by the use of verapamil (a com-

ARCH NEUROL / VOL 56, AUG 1999 994

©1999 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 monly used calcium channel antagonist),37 probably which contribute more to the development of dementia. through verapamil’s antioxidative activities. Diuretics, es- Nonetheless, more than 50% of AD cases may have a vas- pecially thiazides, also have direct vasorelaxant effects cular component.2 In this study, we have not intended on vessel wall,38-40 which may contribute to their anti- to differentiate AD and VaD because overt cerebrovas- hypertensive properties and may have implications in re- cular events are included in the criteria for VaD, and we ducing this toxic effect of amyloid ␤-protein. have not identified silent cerebrovascular events or le- A faster rate of decline of cognition among those not sions. Nonetheless, AD consisted of more than 70% of taking antihypertensive medication could be due to the incident dementia cases according to our clinical diag- poorer control of blood pressure during the follow-up. nosis. Excluding other dementias (secondary and un- Furthermore, other unknown mechanisms may explain specified dementias) did not change the results. the possible protection of diuretics. This should be explored further. CONCLUSIONS

LIMITATIONS OF THE STUDY To our knowledge, this is the first large-scale prospec- tive study that examined the relationship between anti- Information on the duration of antihypertensive medi- hypertensive medication use and dementia in a general cation use was not available. The classification of drug population. Taking diuretics appears to protect against use was made only at baseline. Persons taking diuretics dementia. The main mechanism is probably that diuret- at baseline may take other antihypertensive drugs later, ics can significantly reduce the number of cerebrovas- such as calcium channel antagonists. Unlike a clinical trial, cular events or lesions, which are now thought to be our study is not designed specifically to assess the ef- important in the pathogenesis of both AD and VaD; fects of antihypertensive medication. The group using an- other mechanisms are, however, possible. Because of tihypertensive medication in our cohort differs inher- the nature of an observational study, the findings need ently from the group not using antihypertensive to be replicated in further studies, particularly clinical medication in a number of factors, such as heart dis- trials. ease, stroke, and hypertension, that may be related to the risk of dementia. Generally, these limitations may de- Accepted for publication October 27, 1998. crease the power to detect a given association or lead to This study was supported by grants from the Swedish an underestimation of the association. These limita- Medical Research Council, the Swedish Council for Social tions may particularly affect the results regarding the use Research, the Swedish Municipal Pension Institute, the Na- of other antihypertensive medication because this group tional Corporation of Swedish Pharmacies’ Fund for Re- was relatively small. search and Studies in Health Economics and Social Phar- 41,42 In cross-sectional studies, patients with demen- maceutics, the Torsten and Ragnar So¨derbergs Foundation, tia often have a lower blood pressure level than those with- the SHMF (Stiftelsen Hja¨lp till Medicinsk Forskning) Foun- out dementia. Therefore, current use of antihyperten- dation, and the Gun and Bertil Stohne Foundation. sive medication could be lower in those with clinically We thank Margaret Maytan, MD, for revision of the manifest dementia. This may explain the much lower RRs manuscript and our colleagues at the Kungsholmen Project, of dementia associated with the use of antihypertensive Stockholm, Sweden, for their collaboration in collecting and medication estimated from prevalence data than those managing data. from incidence data. In addition, subjects with cogni- Reprints: Zhenchao Guo, MD, PhD, Department of Ge- tive impairment may underestimate their use of medi- riatric Medicine, Karolinska Institute and Stockholm Ger- cations, although we tried to get information from proxy ontology Research Center, Box 6401, S-113 82, Stock- informants on drug use of the subjects with possible cog- holm, Sweden. nitive dysfunction. There was no difference, however, in the use of cardiac glycosides between those with dementia (18.2%) and those without dementia (19.2%). This REFERENCES may suggest that the recall bias is minimal in this study. In addition, similar results were obtained when subjects 1. Skoog I, Nilsson L, Palmertz B, Andreasson LA, Svanborg A. A population- with a baseline MMSE score of less than 24 were excluded. based study of dementia in 85-year-olds. N Engl J Med. 1993;328:153-158. Although we cannot definitely rule out the possibil- 2. O’Brien MD. How does cerebrovascular disease cause dementia? Dementia. 1994; 5:133-136. ity that our findings are the results of combined effects of 3. Forette F, Boller F. Hypertension and the risk of dementia in the elderly. Am J antihypertensive medication and lifestyle modifications for Med. 1991;90(suppl 3A):14S-19S. hypertension control, it is unlikely that a fat-modified diet 4. Hachinski V. Preventable senility: a call for action against the vascular demen- can fully explain the slower progression of dementia as- tias. Lancet. 1992;340:645-648. 5. Scheinberg P. Dementia due to vascular disease: a multifactorial disorder. Stroke. sociated with the use of diuretics found in this study. In- 1988;19:1291-1299. deed, similar RRs of dementia associated with antihyper- 6. Gorelick PB. Status of risk factors for dementia associated with stroke. Stroke. tensive medication use were obtained in the subpopulation 1997;28:459-463. with a baseline systolic blood pressure of greater than 160 7. Roman GC. Senile dementia of the Binswanger type: a vascular form of demen- mm Hg or a diastolic blood pressure of greater than 95 tia in the elderly. JAMA. 1987;258:1782-1788. 8. Guo Z, Viitanen M, Fratiglioni L, Winblad B. Blood pressure and dementia in the mm Hg. elderly: epidemiologic perspectives. Biomed Pharmacother. 1997;51:68-73. The coexistence of AD and vascular changes in dis- 9. Geldmacher DS, Whitehouse PJ. Evaluation of dementia. N Engl J Med. 1996; ease is common, and there is often difficulty in deciding 335:330-336.

ARCH NEUROL / VOL 56, AUG 1999 995

©1999 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 10. Yoshitake T, Kiyohara Y, Kato I, et al. Incidence and risk factors of vascular de- Uppsala, Sweden: Nordic Council on Medicines; 1985. NLN publication 16. mentia and Alzheimer’s disease in a defined elderly Japanese population: the Hi- 26. Guo Z, Fratiglioni L, Winblad B, Viitanen M. Blood pressure and performance on sayama Study. Neurology. 1995;45:1161-1168. the Mini-Mental State Examination in the very old: cross-sectional and longitu- 11. Skoog I, Lernfelt B, Landahl S, et al. 15-Year longitudinal study of blood pres- dinal data from the Kungsholmen Project. Am J Epidemiol. 1997;145:1106- sure and dementia. Lancet. 1996;347:1141-1145. 1113. 12. Snowdon DA, Greiner LH, Mortimer JA, Riley KP, Greiner PA, Markesbery WR. 27. Korn EL, Graubard BI, Midthune D. Time-to-event analysis of longitudinal fol- Brain infarction and the clinical expression of Alzheimer’s disease: the Nun Study. low-up of a survey: choice of the time-scale. Am J Epidemiol. 1997;145:72-80. JAMA. 1997;277:813-817. 28. Davis KL. Future therapeutic approaches to Alzheimer’s disease. J Clin Psychia- 13. Olichney JM, Hansen LA, Hofstetter CR, Grundman M, Katzman R, Thal LJ. Ce- try. 1998;59(suppl 11):14-16. rebral infarction in Alzheimer’s disease is associated with severe amyloid angi- 29. Kaplan NM, Gifford RW Jr. Choice of initial therapy for hypertension. JAMA. 1996; opathy and hypertension. Arch Neurol. 1995;52:702-708. 275:1577-1580. 14. Viitanen M, Guo Z. Are cognitive function and blood pressure related? Drugs Ag- 30. Psaty BM, Smith NL, Siscovick DS, et al. Health outcomes associated with an- ing. 1997;11:165-169. tihypertensive therapies used as first-line agents: a systematic review and meta- 15. Prince MJ, Bird AS, Blizard RA, Mann AH. Is the cognitive function of older patients analysis. JAMA. 1997;277:739-745. affected by antihypertensive treatment? results from 54 months of the Medical Re- 31. Pantoni L, Garcia JH. The significance of cerebral white matter abnormalities 100 search Council’s trial of hypertension in older adults. BMJ. 1996;312:801-805. years after Binswanger’s report: a review. Stroke. 1995;26:1293-1301. 16. Applegate WB, Pressel S, Wittes J, et al. Impact of the treatment of isolated sys- 32. Sparks DL, Scheff W, Liu H, et al. Increased incidence of neurofibrillary tangles tolic hypertension on behavioral variables: results from the Systolic Hyperten- (NFT) in non-demented individuals with hypertension. J Neurol Sci. 1995;131: sion in the Elderly Program. Arch Intern Med. 1994;154:2154-2160. 162-169. 17. Fratiglioni L, Viitanen M, Ba¨ckman L, Sandman PO, Winblad B. Occurrence of 33. Kalaria R. Cerebral vessels in aging and Alzheimer’s disease. Pharmacol Ther. dementia in advanced age: the study design of the Kungsholmen Project. Neu- 1996;72:193-214. roepidemiology. 1992;11(suppl 1):29-36. 34. Thomas T, Thomas G, McLendon C, Sutton T, Mullan M. ␤-Amyloid-mediated 18. Folstein MF, Folstein SE, McHugh PR. “Mini-Mental State”: a practical method vasoactivity and vascular endothelial damage. Nature. 1996;380:168-171. for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975; 35. Rosenberg RN. A causal role for amyloid in Alzheimer’s disease: the ending of 12:189-198. the beginning. Neurology. 1993;43:851-856. 19. Fratiglioni L, Grut M, Forsell Y, et al. Prevalence of Alzheimer’s disease and other 36. Selkoe DJ. Amyloid beta-protein and the genetics of Alzheimer’s disease. J Biol dementias in an elderly urban population: relationship with age, sex, and edu- Chem. 1996;271:18295-18298. cation. Neurology. 1991;41:1886-1892. 37. Suo Z, Fang C, Crawford F, Mullan M. Superoxide free radical and intracellular

20. Fratiglioni L, Viitanen M, von Strauss E, Tontodonati V, Herlitz A, Winblad B. Very calcium mediate A ␤1-42 induced endothelial toxicity. Brain Res. 1997;762:144- old women at highest risk of dementia and Alzheimer’s disease: incidence data 152. from the Kungsholmen Project, Stockholm. Neurology. 1997;48:132-138. 38. Kool MJ, Lustermans FA, Breed JG, et al. The influence of perindopril and the 21. American Psychiatric Association. Diagnostic and Statistical Manual of Mental diuretic combination amiloride + hydrochlorothiazide on the vessel wall proper- Disorders, Revised, Third Edition. Washington, DC: American Psychiatric Asso- ties of large arteries in hypertensive patients. J Hypertens. 1995;13:839-848. ciation; 1987. 39. Ka¨ho¨nen M, Ma¨kynen H, Arvola P, Wuorela H, Po¨rsti I. Arterial function after tri- 22. Hughes CP, Berg L, Danziger WL, Coben LA, Martin RL. A new clinical scale for chlormethiazide therapy in spontaneously hypertensive rats. J Pharmacol Exp the staging of dementia. Br J Psychiatry. 1982;140:566-572. Ther. 1995;272:1223-1230. 23. Forsell Y, Fratiglioni L, Grut M, Viitanen M, Winblad B. Clinical staging of de- 40. Abrahams Z, Tan LLY, Pang MYM, Abrahams B, Tan MM, Wright JM. Demon- mentia in a population survey: comparison of DSM-III-R and the Washington stration of an in vitro direct vascular relaxant effect of diuretics in the presence University Clinical Dementia Rating Scale. Acta Psychiatr Scand. 1992;86:49- of plasma. J Hypertens. 1996;14:381-388. 54. 41. Guo Z, Viitanen M, Fratiglioni L, Winblad B. Low blood pressure and dementia in 24. Wills P, Fastbom J, Claesson CB, Cornelius C, Thorslund M, Winblad B. Use of elderly people: the Kungsholmen Project. BMJ. 1996;312:805-808. cardiovascular drugs in an older Swedish population. J Am Geriatr Soc. 1996; 42. Hogan DB, Ebly EM, Rockwood K. Weight, blood pressure, osmolarity, and glu- 44:54-60. cose levels across various stages of Alzheimer’s disease and vascular demen- 25. Nordic Council on Medicines. Guidelines for ATC Classification [in Swedish]. tia. Dement Geriatr Cogn Disord. 1997;8:147-151.

ARCH NEUROL / VOL 56, AUG 1999 996