The Prevention of Dementia with Antihypertensive Treatment: New Evidence from the Systolic Hypertension in Europe (Syst-Eur)

ORIGINAL INVESTIGATION The Prevention of Dementia With Antihypertensive Treatment New Evidence From the Systolic Hypertension in Europe (Syst-Eur) Study

Franc¸oise Forette, MD; Marie-Laure Seux, MD; Jan A. Staessen, MD, PhD; Lutgarde Thijs, MSc; Marija-Ruta Babarskiene, MD; Speranta Babeanu, MD; Alfredo Bossini, MD; Robert Fagard, MD; Blas Gil-Extremera, MD; Tovio Laks, MD; Zhanna Kobalava, MD; Cinzia Sarti, MD; Jaakko Tuomilehto, MD; Hannu Vanhanen, MD; John Webster, MD; Yair Yodfat, MD; Willem H. Birkenha¨ger, MD; for the Syst-Eur Investigators

Background: After the double-blind, placebo- tolic/diastolic blood pressure was 7.0/3.2 mm Hg higher controlled Systolic Hypertension in Europe (Syst-Eur) in the 1417 control patients than in the 1485 subjects trial ended in February 1997, randomized patients were randomized to active treatment. At the last examina- offered active study medication for a further period of tion, the blood pressure difference was still 4.2/2.9 observation. mm Hg; 48.1%, 26.4%, and 11.4% of the control patients were taking nitrendipine, enalapril, and/or hydro- Objective: To refine the estimates of the long-term ef- chlorothiazide, whereas in the active treatment group fects of antihypertensive therapy on the incidence of de- these proportions were 70.2%, 35.4%, and 18.4%, re- mentia. spectively. Compared with the controls, long-term antihypertensive therapy reduced the risk of dementia by Methods: Eligible patients had no dementia and were 55%, from 7.4 to 3.3 cases per 1000 patient-years (43 vs at least 60 years old. Their systolic blood pressure at en- 21 cases, PϽ.001). After adjustment for sex, age, educa- try was 160 to 219 mm Hg, with diastolic blood pres- tion, and entry blood pressure, the relative hazard rate as- sure below 95 mm Hg. Antihypertensive therapy was sociated with the use of nitrendipine was 0.38 (95% con- started immediately after randomization in the active treat- fidence interval, 0.23-0.64; PϽ.001). Treatment of 1000 ment group, but only after termination of the double- patients for 5 years can prevent 20 cases of dementia (95% blind trial in the control patients. Treatment consisted confidence interval, 7-33). of nitrendipine (10-40 mg/d), with the possible addition of enalapril maleate (5-20 mg/d), hydrochlorothia- Conclusion: The extended follow-up of Syst-Eur pa- zide (12.5-25 mg/d), or both add-on drugs. tients reinforces the evidence that blood pressure– lowering therapy initiated with a long-acting dihydropyri- Results: Median follow-up increased from 2.0 years in dine protects against dementia in older patients with the double-blind trial to 3.9 years overall. The inci- systolic hypertension. dence of dementia doubled from 32 to 64 cases, 41 of whom had Alzheimer disease. Throughout follow-up, sys- Arch Intern Med. 2002;162:2046-2052

YPERTENSION IS associ- incident cases in the two treatment groups ated with increased risk combined was only 32. of both vascular demen- On February 14, 1997, the Syst-Eur tia and Alzheimer dis- trial was terminated early, because at the ease.1,2 In view of the second interim analysis significant ben- increasing longevity of populations efit was reached with respect to stroke, the H 6 worldwide, prevention of dementia has primary end point. However, for a vari- turned into a major public health chal- ety of ethical and scientific reasons,7 lenge.3 In the patients randomized in the we decided to extend the Syst-Eur trial double-blind Systolic Hypertension in Eu- into an open-label, active-treatment fol- rope (Syst-Eur) trial, we observed that an- low-up study in the same population and tihypertensive therapy, compared with pla- based on the original active study medi- cebo, reduced the incidence of dementia cation (Syst-Eur 2). To review our initial by 50%, from 7.7 to 3.8 cases per 1000 pa- estimates of benefit,4 we also continued the Author affiliations and a list of 4 8 the principal investigators of tient-years (21 vs 11 cases, P=.05). How- Vascular Dementia Project. In the pres- 5 the Systolic Hypertension in ever, Collins and MacMahon suggested ent article, we have updated the inci- Europe Trial appear at the end that our estimates of benefit were of lim- dence rates of dementia in Syst-Eur pa- of this article. ited value, largely because the number of tients, whose blood pressure was actively

(REPRINTED) ARCH INTERN MED/ VOL 162, OCT 14, 2002 WWW.ARCHINTERNMED.COM 2046

Downloaded From: https://jamanetwork.com/ on 10/02/2021 lowered from randomization compared with those who received active treatment only since the start of Syst- 3228 Patients Randomized in Dementia Study Eur 2. 68 Patients Excluded SUBJECTS AND METHODS • 9 Had Dementia at Baseline • 36 MMSE Scores at Baseline Not Available Ethics committees of the University of Leuven, Leuven, Bel- • 7 Had More Than 5 MMSE Items gium, and the participating centers approved the protocol of Not Answered at Baseline • 16 Had Baseline MMSE Score ≤23 the Syst-Eur study. Eligible patients had no dementia, were at and Did Not Undergo DSM-III-R least 60 years old, and had a sitting systolic blood pressure rang- ing from 160 to 219 mm Hg, with diastolic blood pressure 6,9 below 95 mm Hg. The patients were randomized to active 3160 Considered for Analysis treatment or placebo. Active treatment started with the dihydropyridine calcium channel blocker nitrendipine (10-40

mg/d), which could be combined with or replaced by enala- 258 Follow-up Examinations pril maleate (5-20 mg/d), hydrochlorothiazide (12.5-25 Planned but Not Yet Available mg/d), or both second-line drugs. In the placebo group, matching placebos were used in a similar way. However, after the

double-blind trial had stopped, patients originally allocated pla- 2902 Included in Analysis cebo were offered the same active treatment sequence as the other patients. The study drugs were stepwise titrated and/or combined to reduce the sitting systolic blood pressure by 20 mm Hg or more to a level below 150 mm Hg. 1417 Originally Randomized to Placebo 1485 Originally Randomized to Syst-Eur investigators opting to take part in the dementia Active Treatment project4,8 had to screen all their patients for cognitive impairment at baseline and at annual follow-up visits, using the 10 Figure 1. Study profile. MMSE indicates Mini-Mental State Examination; Mini-Mental State Examination (MMSE). Sequential diag- DSM-III-R, Diagnostic and Statistical Manual of Mental Disorders, Revised nostic procedures to establish the presence and cause of de- Third Edition. mentia were started if the MMSE score was 23 or less,11 if symptoms reported by the patients or their relatives or clinical signs suggested cognitive impairment, or if for any reason a RESULTS planned MMSE questionnaire had not been administered. The diagnosis of dementia relied on the criteria of the Diagnostic At 106 centers in 19 European countries, 3228 patients and Statistical Manual of Mental Disorders, Revised Third Edi- were enrolled. However, 9 patients already had demen- tion (DSM-III-R),12 which when Syst-Eur began in 1988 was tia at baseline, whereas in 59 patients cognitive impair- the generally accepted standard.13 If the DSM-III-R criteria12 ment could not be excluded (Figure 1). On December confirmed the diagnosis of dementia, the modified ischemic 1, 2000, a follow-up examination was planned but had 14 score, including brain imaging by computed tomography not yet taken place in 258 patients. Thus, the number of (CT), served to differentiate vascular from degenerative dis- patients included in the present analysis was 2902 (Fig- ease. If a CT scan could not be performed, the Hachinski ure 1). At randomization, the patients of the former pla- score15 replaced the modified ischemic score to establish the cause of dementia. A blinded review board validated all cases cebo (n=1417) and active treatment (n=1485) groups of dementia, and an independent blinded neuroradiologist re- had similar characteristics. Median age at randomiza- viewed the CT scans. To evaluate the functional status of the tion was 68 years (range, 60-92 years). Mean±SD body patients, the Index of Independence in Activities of Daily Liv- mass index (calculated as weight in kilograms divided ing (ADL score) was recorded at randomization and thereafter by the square of height in meters) averaged 26.6±3.2 in at yearly intervals.16 984 men and 27.3±4.4 in 1918 women. A total of 764 Stroke was the primary end point in the Syst-Eur trial.6 patients (26.3%) had previous cardiovascular complica- This disorder was defined as a neurologic deficit with symp- tions. In addition, 17 patients randomized to placebo toms continuing for more than 24 hours or leading to death (1.2%) and 20 of the active treatment group (1.3%) had with no other cause than vascular. The Endpoint Committee, a history of stroke. Mean±SD age when leaving school which was unaware of a patient’s treatment status, ascertained all stroke cases by reviewing the patients’ files and other source was 16.7±4.5 years. documents, by requesting detailed written information from The number of patient-years of follow-up was 5849 the investigators, or by both approaches. Transient ischemic in the former placebo group and 6359 in the patients ini- attack was defined as focal cerebral dysfunction lasting for less tially allocated active treatment. Overall, median fol- than 24 hours. Physicians who were unaware of the treatment low-up from randomization was 3.9 years (interquartile group status checked the diagnosis of transient ischemic at- range [IQR], 2.8-5.6 years). Table 1 shows treatment tack at the coordinating office.6 status at the last follow-up visit by randomization group. Database management and statistical analysis were per- The proportion of patients who remained untreated was formed with SAS statistical software, version 8.01 (SAS Insti- larger in the control group than in the patients allocated tute Inc, Cary, NC). Means and medians were compared by the active treatment (25.2% vs 3.0%, PϽ.001), because 277 standard normal z test and Wilcoxon 2-sample test and proportions by the ␹2 statistic. The incidence of dementia was control patients (19.5%) were taking only double-blind analyzed by means of Kaplan-Meier survival function esti- placebo medication. At the last visit, blood pressure of mates and the log-rank test. Covariables associated with the the control patients averaged 156.1±12.0 mm Hg sys- risk of dementia were identified using single and multiple Cox tolic and 82.5±6.0 mm Hg diastolic. In the patients ran- regression.) domized to active treatment, these levels were 149.1±9.7

(REPRINTED) ARCH INTERN MED/ VOL 162, OCT 14, 2002 WWW.ARCHINTERNMED.COM 2047

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Table 1. Number of Patients at Last Follow-up Visit* Table 2. Origin of Dementia

Control Group Active Group Control Active All Variable (n = 1417) (n = 1485) Variable Group Group Participants Study phase No. of incident cases Double-blind follow-up in Syst-Eur 1 287 292 All causes 43 21 64 Open follow-up in Syst-Eur 1 104 58 Alzheimer dementia 29 12 41 Open follow-up in Syst-Eur 2 1012 1126 Mixed or vascular dementia* 12 7 19 Nonsupervised open follow-up 14 9 Origin unknown 2 2 4 Antihypertensive treatment Rate, per 1000 patient-years Active study medication only 751 1201 All causes 7.4 3.3 5.2 Other antihypertensive drugs 264 192 Alzheimer dementia 5.0 1.9 3.4 Double-blind placebo treatment 277 0 Mixed or vascular dementia* 2.1 1.1 1.6 Untreated 80 44 Origin unknown 0.3 0.3 0.3 Treatment unknown 45 48 Antihypertensive drugs† *The cause of dementia was likely to be vascular in 4 control patients and Nitrendipine only 372 627 3 patients randomized to active treatment. Nitrendipine 681 1042 Enalapril maleate 374 526 Hydrochlorothiazide 162 273 forms in 55 patients, copies of medical records made available by the investigator in 6 cases, or both an *Syst-Eur indicates Systolic Hypertension in Europe. †Because many patients were undergoing combined drug treatment, MMSE score of 23 or less and a written confirmation of numbers do not total. the diagnosis by the investigator in 3 cases. Of the patients with dementia, 41 developed Alzheimer disease and 19 had mixed or vascular dementia. The etiologic 180 Placebo diagnosis depended on the modified ischemic score14 in Active Treatment 1417 51 patients, of whom 47 also underwent CT imaging of Systolic the brain, or on the Hachinski score15 in 9 cases. We 170 could not ascertain the cause of dementia in 4 patients. At diagnosis, median follow-up since randomization

160 1417 was 3.8 years (IQR, 2.7-5.9 years) in the control group 1417 381 and 3.4 years (IQR, 1.7-5.2 years) in the patients ran- 1417 515 1415 1016 742 domized to active treatment. Of the incident cases, 12 150 were male and 52 were female. The sex-specific rates were 2.9 and 6.5 cases per 1000 patient-years, respectively (difference, 55%; 95% CI, 16%-76%; P=.02). 140 Median age at diagnosis was 79 years. Before dementia developed, 3 patients had experienced a stroke (control,

85 Diastolic 2; active treatment, 1), whereas 3 other patients (con- 1485 trol, 2; active treatment, 1) had a history of transient Blood Pressure, mm Hg ischemic attack. The overall incidence of dementia was 5.2 cases per 80 1485 1000 patient-years (Table 2). Forty-three cases of de- 1485 1485 1485 mentia occurred in the control group, and 21 cases were 1072 758 observed in the patients randomized to active treat- 552 75 410 ment. Thus, the incidence of dementia decreased by 55% 012536 4 78 (95% CI, 24%-73%) with long-term active treatment com- Time Since Randomization, y pared with the control group (3.3 vs 7.4 cases per 1000 patient-years, PϽ.001, Figure 3). The incidence of Alz- Figure 2. Average sitting systolic and diastolic blood pressure at randomization and during follow-up. For each mean value, the number of heimer disease and mixed or vascular dementia was re- patients is given. duced (Table 2). At the rate observed in the control group, 1000 patients would have to be treated for 5 years to prevent 20 new cases (95% CI, 7-33 cases). The median mm Hg and 79.4±6.1 mm Hg, respectively. During the MMSE score at baseline was 29 (IQR, 27-30) in both treat- entire follow-up, the net baseline-subtracted between- ment groups (P=.94). Subsequent changes in the MMSE group differences (PϽ.001) were 7.0 mm Hg systolic (95% score were also similar in the control and active treat- confidence interval [CI], 6.2-7.8 mm Hg) and 3.2 mm ment groups (Table 3). Hg diastolic (95% CI, 2.8-3.6 mm Hg). At 8 years, these At the last visit, ADL scores had been recorded for differences remained significant (PϽ.01, Figure 2), av- 60 patients with dementia (control, 41; active treat- eraging 4.2 mm Hg systolic (95% CI, 1.5-6.9 mm Hg) and ment, 19) and 2815 participants without dementia (con- 2.9 mm Hg diastolic (95% CI, 1.5-4.4 mm Hg). trol, 1364; active treatment, 1451). The median score was There were 64 incident cases of dementia 6 (IQR, 6-6; range, 0-6). Within the groups of patients (Table 2). The diagnosis relied on the DSM-III-R with and without dementia, there were no differences at

(REPRINTED) ARCH INTERN MED/ VOL 162, OCT 14, 2002 WWW.ARCHINTERNMED.COM 2048

Downloaded From: https://jamanetwork.com/ on 10/02/2021 10 Placebo Table 3. Changes in Mini-Mental State Examination Scores Active Treatment From Randomization*

8 Time From Control Group Active Group Randomization (n = 1263) (n = 1327) P Value 1 0.16 ± 1.53 0.10 ± 1.44 .28 6 –55% 2 0.15 ± 1.69 0.17 ± 1.64 .75 P < .001 3 0.14 ± 1.85 0.17 ± 1.82 .73

4 *Intention-to-treat analysis with last observation carried forward. A total of 155 patients randomized to placebo and 159 allocated active treatment Cases per 100 Patients whose baseline Mini-Mental State Examination scores were not administered during the run-in period but shortly after randomization were excluded from 2 analysis. Values are given as mean ± SD.

0 the risk of dementia. The crude and adjusted relative hazard rates associated with an increase in the daily dose by one tablet (20 mg) were 0.37 (95% CI, 0.23- 02468 Ͻ Time Since Randomization, y 0.61; P .001) and 0.48 (95% CI, 0.29-0.78; P=.003), respectively. Figure 3. Culmulative rate of dementia by treatment group. COMMENT the last visit between the 2 treatment arms in the distributions of the ADL scores (PϾ.68). However, at both en- In comparison with our earlier report,4 the present analy- try and the last follow-up visit, the distributions of the sis provides a more precise estimate of the potential of ADL scores were shifted toward lower values in patients antihypertensive treatment to prevent dementia, be- with dementia compared with those without dementia. cause the number of incident cases and the years of fol- At entry, a score of less than 6 (maximum) was ob- low-up doubled. Antihypertensive therapy starting with served in 5% of the patients with dementia but only in the dihydropyridine calcium channel blocker nitren- 1% of the participants without dementia (PϽ.001). At dipine reduced the incidence of dementia by 55%. By in- the last visit, these proportions were 42% and 4%, reference, treating 1000 patients for 5 years can prevent spectively (PϽ.001). 20 cases of dementia. Thus, the present analysis demon- In stepwise Cox regression, the risk of dementia in- strates a high degree of consistency with our previous creased with age and diastolic blood pressure at base- estimate of 19 prevented cases per 5000 patient-years of line. It was lower in more educated patients and men. antihypertensive treatment.4 The mutually adjusted relative hazard rates were 2.10 We observed that long-term antihypertensive treat- (95% CI, 1.76-2.50; PϽ.001) for a 5-year increase in age, ment decreased the incidence of Alzheimer disease and 1.39 (95% CI, 1.08-1.81; P=.01) for a 5–mm Hg in- vascular or mixed dementia. Our initial working hypoth- crease in diastolic pressure, 0.88 (95% CI, 0.82-0.95; esis was that treatment of hypertension would protect PϽ.001) for each additional year of education, and 0.53 mainly against vascular dementia.4,8 However, recent pro- (95% CI, 0.27-1.03; P=.06) for men compared with spective studies17-20 suggest that hypertension and, more women. After adjustment for these covariables, the rela- generally, all risk factors involved in arteriosclerosis21 may tive hazard rate associated with initial randomization to contribute to the incidence of degenerative dementias as active treatment was 0.43 (95% CI, 0.25-0.74), which was well. There is a growing awareness1,2,22-24 that the dis- statistically significant (P=.003), whereas the hazard rate tinction between Alzheimer disease and vascular demen- for systolic blood pressure at baseline was nonsignifi- tia is less clear than initially envisaged, both conditions cant (1.07 for a 10–mm Hg increase; 95% CI, 0.84-1.36; sharing similar mechanisms and lesions albeit to differ- P=.58). Pulse pressure did not significantly predict the ent degrees. In practical terms, it is therefore expedient risk of dementia. The unadjusted and adjusted relative in prospective clinical trials, such as Syst-Eur, to con- hazard rates were 1.15 (95% CI, 0.93-1.42; P=.20) and sider dementia as one clinical entity. In keeping with 0.95 (95% CI, 0.75-1.21; P=.70), respectively. some25 but not all26,27 studies, men had a lower risk of In further Cox regression analyses, we investigated dementia than women. The Rotterdam Study27 showed the effects of nitrendipine on the incidence of dementia. no sex differences in the incidence of dementia up to a If we introduced intake of nitrendipine as a time- very old age. After 90 years of age, the incidence of Alz- dependent covariable in the Cox model, the crude rela- heimer disease was higher in women than men, whereas tive hazard rate was 0.30 (95% CI, 0.18-0.50; PϽ.001). at all ages vascular dementia occurred more frequently After adjustment for sex, age treated as time-dependent in men than women.27 covariable, years of education, and diastolic blood pres- Three longitudinal observational studies28-30 showed sure at entry, the hazard rate was 0.38 (95% CI, 0.23- significant reductions in the incidence of dementia in rou- 0.64; PϽ.001). We obtained similar results if we used tinely treated vs nontreated hypertensive patients. In con- the average daily dose of nitrendipine before the onset trast, the Systolic Hypertension in the Elderly Pro- of cognitive impairment as an independent predictor of gram31 and the Medical Research Council trial in older

(REPRINTED) ARCH INTERN MED/ VOL 162, OCT 14, 2002 WWW.ARCHINTERNMED.COM 2049

Downloaded From: https://jamanetwork.com/ on 10/02/2021 adults20,32 failed to demonstrate any effect on cognition, dine compared with hydralazine.41 In vascular and de- despite differences in systolic/diastolic blood pressures generative dementias, the MMSE scores decreased less between patients allocated placebo or active treat- with the calcium channel blocker nimodipine comment,33 which averaged 11.5/4.1 mm Hg31 and 10.6/5.6 pared with placebo.38 The aging brain loses its ability to mm Hg,34 respectively. In both trials,31,34 diuretics or regulate intracellular calcium, leading to a cascade of cel- ␤-blockers were the first-line antihypertensive agents. Re- lular impairments and ultimately to cell death.39,40 Alter- cently, the double-blind, placebo-controlled Perindo- ations in calcium homeostasis are involved in the aging pril Protection Against Recurrent Stroke Study trial35 dem- process of the brain and in the neuropathology of Alz- onstrated that among patients with a history of stroke heimer disease.42-44 In patients with degenerative demen- or transient ischemic attack, blood pressure–lowering tia, ␤-amyloid may raise the concentration of intraneu- treatment reduced the risk of recurrent stroke by 28%. ronal free calcium and through this mechanism may Of 6105 patients, 58% were randomized to dual treat- sensitize the brain to neurotoxins, such as proinflamma- ment with perindopril erbumine and indapamide vs tory substances or pro-oxidants.43 The hypothesis of a matching placebos, whereas the remainder were allo- possible central nervous action of dihydropyridines is also cated monotherapy with perindopril or matching pla- supported by the observation that these drugs cross the cebo.36 Unexpectedly, a prespecified subgroup analysis blood-brain barrier45,46 and reduce the turnover of mono- revealed marked heterogeneity of treatment effect sizes amine neurotransmitters,45 of which many are deficient for stroke risk between participants who received com- in degenerative dementias.39 Nitrendipine binding in the bination therapy with perindopril plus indapamide and rat brain also occurs mainly at those sites that are pri- those who received single-drug therapy with perindo- marily affected by Alzheimer disease, such as the super- pril alone. Combination therapy reduced blood pres- ficial cortex, thalamus, and hippocampus, and not in areas sure by 12/5 mm Hg and the risk of stroke by 43%, with with low synaptic density.47 similar benefits in hypertensive and nonhypertensive pa- Selective recruitment of relatively healthy patients tients.35 Treatment with perindopril alone lowered blood who accepted long-term follow-up in a double-blind trial pressure by 5/3 mm Hg but did not affect stroke recur- is likely to explain the high MMSE scores at entry into rence. The 95% CI of the relative risk reduction ranged the Syst-Eur trial. During follow-up, the changes in the from −19% to 23%. With regard to cognitive function, MMSE scores were similar in the control and active treat- only stroke-related dementia was reduced by approxi- ment groups and the incidence of dementia was low (7.4 mately 50% in the group undergoing combined treat- cases per 1000 patient-years in the control group) in ment with perindopril and indapamide. In the present comparison with population-based studies.3,13,48,49 In- study, benefit was predominantly attributable to the pre- deed, among elderly patients (Ն65 years) enrolled in the vention of degenerative dementia rather than dementia EURODEM Study,25 528 cases of dementia occurred dur- occurring in association with cerebrovascular events, such ing 28768 person-years of follow-up (rate, 18.4 cases per as stroke or transient ischemic attack. 1000 person-years). In the present study, patients in the Some proponents of the use of the older drug classes control group were also given blood pressure–lowering speculated that in the Systolic Hypertension in the El- therapy after termination of the double-blind trial. Early derly Program study31 differential dropout from the pla- cognitive disorder may lead to less compliance to therapy. cebo and active treatment groups biased the cognitive and The direction of all these biases would be to reduce the functional evaluations toward a null effect.37 Neverthe- protective effect of the intervention. Thus, our current less, the observation that in randomized clinical trials findings, however favorable, probably underestimate the blood pressure reduction induced by thiazides,31,32 potential of preventing dementia by blood pressure– ␤-blockers,31,32 or the angiotensin II–converting en- lowering therapy. zyme inhibitor perindopril given in monotherapy35 failed In conclusion, the reduction by 55% of the inci- to protect against cognitive impairment32 or demen- dence of dementia by antihypertensive drug treatment tia31,35 again raises the issue of the mechanism by which based on the dihydropyridine calcium channel blocker antihypertensive treatment may prevent dementia. In the nitrendipine as the first-line drug may have important present analysis, the between-group difference in blood public health implications. In view of the increasing lon- pressure during the whole period of follow-up averaged gevity of populations worldwide, dementia is a leading 7/3 mm Hg and persisted until 8 years after randomiza- cause of disability among all races and continents.3 At tion, when it was still 4/3 mm Hg. Thus, blood pressure the rate observed in the Syst-Eur control group, treating lowering might explain the prevention of stroke. On the 1000 hypertensive patients for 5 years can prevent 20 cases other hand, in the present study, the dihydropyridine cal- of dementia, a benefit that is likely to be even larger in cium channel blocker nitrendipine was the mainstay of unselected hypertensive patients who experience a higher active treatment, and its use was associated with a dose- risk of dementia.25 Our present findings are in line with dependent reduction in the probability of dementia. In- those of recent overviews50,51 of the actively controlled deed, the risk declined by approximately 50% for every outcome trials in hypertension, which suggested that tablet of 20 mg taken per day. calcium channel blockers might provide better protec- Several reports38-40 suggest that calcium channel tion against stroke than treatment based on diuretics blockers may confer specific neuroprotection. In spon- and ␤-blockers. In view of the growing body of evi- taneously hypertensive rats, nicardipine counteracted the dence1,2,22-24 that vascular factors increase the risk of de- neurodegenerative effects of high blood pressure over and generative dementia, we already called for a prospective above the antihypertensive action of this dihydropyri- face-to-face comparison of a long-acting calcium chan-

(REPRINTED) ARCH INTERN MED/ VOL 162, OCT 14, 2002 WWW.ARCHINTERNMED.COM 2050

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Clinical Investigators Participating in the Dementia Project

Belgium: H. Celis, MD; J. Claus, MD; E. De Graef, MD; H. Dieu, MD; P. Gilbert, MD; J. Gremling, MD; W, Pelemans, MD. Bulgaria: M. Grigorov, MD; K. Janculova, MD; E. Lilov, MD; C. Nachev, MD; T. Tchernev, MD; T. N. Vasileva, MD. Czech Republic: J. Filipovsky, MD. Estonia: T. Laks, MD. Finland: M. Alaluoto, MD; R. Antikainen, MD; M. Haapio, MD; T. Haka- ma¨ki, MD; K. Halonen, MD; M. Ja¨a¨skivi, MD; P. Kivinen, MD; P. P. Kohonen-Jalonen, MD; P. Kuusisto, MD; A. Latva- Nevala, MD; E. Lehmus, MD; E. Lehtoma¨ki, MD; A. Lehtonen, MD; C. Sarti, MD; R. Tilvis, MD; H. Tuomilehto, MD; E. Rahter, MD; E. Ruotsalainen, MD; H. Vanhanen, MD; O. Va¨nska¨, MD; M. Vietaniemi, MD; S. Vinni, MD; H. Virtanen, MD; H. Wallinheimo, MD. France: P. Berger, MD; F. Bezot, MD; L. Boucher, MD; J. Boussac, MD; A. Campagne, MD; C. Copere, MD; R. Corlier, MD; E. De Saint Lorette, MD; G. Donnarel, MD; M. Escande, MD; G. Etchegaray, MD; H. Feuillette, MD; Y. M. Flores, MD; F. Forette, MD; G. Franc¸ois, MD; X. Girerd, MD; M. Gre´goire, MD; S. Houdrie-Pavie, MD; J. R. Israel, MD; J. B. Leblond, MD; M. Masieri, MD; C. H. Mercier, MD; G. Meridjen, MD; Y. Ollivier, MD; I. Pe´rilliat, MD; L. M. Pommier, MD; E. Quignard, MD; P. Romejko, MD; M. Safar, MD; J. M. Vigne, MD. Greece: A. D. Efstratopoulos, MD; S. Voyaki, MD. Ireland: L. Bradley, MD; J. Duggan, MD; E. T. O’Brien, MD. Israel: J. Fidel, MD; A. Goldhaber, MD; I. Moran, MD; J. R. Viskoper, MD; Y. Yodfat, MD. Italy: A. Bossini, MD; G. Maiorano, MD; P. Palatini, MD; C. Pasotti, MD; A. Pirrelli, MD; L. Terzoli, MD. Lithuania: M. R. Babarskiene, MD. Netherlands: A. J. Man In’t Veld, MD; H. Stom, MD; A. H. van den Meiracker, MD; J. M. J. Van Der Cammen, MD; N. Vogel, MD; J. Woittiez, MD. Poland: J. Kocemba, MD. Portugal: M. Carrageta, MD; G. Leiria, MD. Romania: S. Babeanu, PhD; V. Bogdaneanu, MD; L. Serban, MD. Russian Federa- tion: G. G. Arabidze, MD (deceased); D. R. Ivleva, MD; V. Moisseyev, MD. Slovakia: Z. Gerova, MD. Spain: B. Gil-Extremera, MD; A. Maldonado-Martin, MD; R. Marin, MD; F. Vega, MD; R. Navarro, MD; V. Cuesta, MD; J. O. Martinez, MD; A. R. Bataro, MD; J. O. Pujadas, MD; J. Mora-Macia, MD; J. L. Rodicio, MD; L. M. Ruilope, MD. United Kingdom: M. Beevers, RN; C. Davidson, MD; P. Gunawardena, MD; A. O’Brien, MD; J. C. Petrie, MD; J. Webster, MD; P. R. Wilkinson, MD.

Committees and Coordination Coordinators of Dementia Project: F. Forette, MD; M. L. Seux, MD; T. Strasser, MD (deceased). Dementia Project Review Board: A. Alpe´rovitch, MD; F. Boller, MD; M. M. Dubs, MD; F. Forette, MD; D. Furet, MD; M. L. Seux, MD; L. Traykoff, MD. Trial Coordinators: R. Fagard, MD, PhD; J. A. Staessen, MD, PhD. Data Monitoring Committee: C. J. Bulpitt, MD, PhD; A. E. Fletcher, PhD; J. A. Staessen, MD, PhD; L. Thijs, MSc. Endpoint Committee: P. W. de Leeuw, MD; R. Fagard, MD; G. Leonetti, MD; J. C. Petrie, MD; H. Vanhanen, MD (associated member for Finland). Ethics Committee: W. H. Birkenha¨ger, MD; C. T. Dol- lery, MD; R. Fagard, MD. Publication Committee: W. H. Birkenha¨ger, MD; C. J. Bulpitt, MD; J. A. Staessen, MD, PhD; A. Zan- chetti, MD. Steering Committee: P. De Cort, MD; R. Fagard, MD; F. Forette, MD; K. Kawecka-Jaszcz, MD; G. Leonetti, MD; C. Nachev, MD; E. T. O’Brien, MD; J. Rodicio, MD; J. Rosenfeld, MD; J. Tuomilehto, MD; J. Webster, MD; Y. Yodfat, MD. Coordination of General Practices: H. Celis, MD, in collaboration with J. Heyrman, MD; G. Stibbe, MD; M. Van den Haute, MD; Y. Yodfat, MD. Coordinating Office: N. Ausloos; H. Celis, MD; E. Den Hond, PhD; L. De Pauw, RN; P. Drent, RN; R. Fagard, MD; H. Fan; J. Gasowski, MD; T. Kuznetsova, MD; T. Nawrot, MSc; J. A. Staessen, MD, PhD; L. Thijs, MSc; Y. Toremans; S. Van Hulle, RN; J. G. Wang, MD; R. Wolfs.

nel blocker and a diuretic, the current reference drug for erdeen Royal Hospitals, University of Aberdeen, Aberdeen, the initiation of antihypertensive treatment,52 in the pre- Scotland (Dr Webster); Department of Family Medicine, Ha- vention of cognitive impairment. This would resolve the dassah Medical School, Hebrew University of Jerusalem, current uncertainty,53,54 which weighed against the hu- Jerusalem, Israel (Dr Yodfat); and Erasmus University, man suffering and societal and economical losses caused Rotterdam, the Netherlands (Dr Birkenha¨ger). by dementia3 is simply unacceptable. The Syst-Eur Vascular Dementia Project was a con- certed action of the European Union’s Biomed Research Pro- Accepted for publication April 3, 2002. gramme (Brussels, Belgium) and conducted under the aus- From the Department of Geriatrics, Hoˆpital Broca, Uni- pices of the Fondation Nationale de Ge´rontologie (Paris, versity of Paris V, Paris, France (Drs Forette and Seux); Study France). The Syst-Eur trial was performed in consultation Coordinating Centre, Department of Molecular and Cardio- with the World Health Organization (Geneva, Switzer- vascular Research, University of Leuven, Leuven, Belgium (Drs land), International Society of Hypertension, European Staessen and Fagard and Ms Thijs); Institute of Cardiology, Society of Hypertension, and World Hypertension League. Kaunas, Lithuania (Dr Babarskiene); National Institute Ana Bayer AG (Wuppertal, Germany) supported the Syst-Eur Aslan, Bucharest, Romania (Dr Babeanu); Centro Iperten- trial. Bayer AG and Merck Sharpe & Dohme (West Point, sione Policlinico Umberto I, Rome, Italy (Dr Bossini); Fac- Pa) donated the study medication. The Belgian National Re- ultad de Medicina, Universidad de Granada, Granada, Spain search Fund (Brussels), Specia S.A. (Paris), and INSERM (Dr Gil-Extremera); Tallinn Central Hospital, Tallinn, Es- (Institut National de la Sante´ et de la Recherche Me´dicale, tonia (Dr Laks); Department of Internal Diseases, Munici- Paris) provided additional grants in support of the vascu- pal Clinical Hospital, Moscow, Russian Federation (Dr Ko- lar dementia project. balava); Department of Epidemiology and Health Promotion, Corresponding author: Jan A. Staessen, MD, PhD, National Public Health Institute, and Department of Public Studie Coo¨rdinatie Centrum, Laboratorium Hypertensie, Health, University of Helsinki, Helsinki, Finland (Drs Sarti Gebouw Onderwijs en Navorsing, Campus Gasthuisberg, and Tuomilehto); Department of Medicine, Helsinki Univer- Herestraat 49, B-3000 Leuven, Belgium (e-mail: jan sity Central Hospital, Helsinki, Finland (Dr Vanhanen); Ab- [email protected]).

(REPRINTED) ARCH INTERN MED/ VOL 162, OCT 14, 2002 WWW.ARCHINTERNMED.COM 2051

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Reprints: Franc¸oise Forette, MD, Department of Geri- 29. Tzourio C, Dufouil C, Ducimetie`reP,Alpe´rovitch A, for the EVA Study Group. Cognitive decline in individuals with high blood pressure: a longitudinal study in atrics, Hoˆpital Broca, CHU Cochin, University of Paris V, the elderly. Neurology. 1999;53:1948-1952. 54-56 rue Pascal, 75013 Paris, France (e-mail: francoise 30. in’t Veld BA, Ruitenberg A, Hofman A, Stricker BH, Breteler MM. Antihyperten- [email protected]). sive drugs and incidence of dementia: the Rotterdam Study. Neurobiol Aging. 2001;22:407-412. 31. SHEP Cooperative Research Group. Prevention of stroke by antihypertensive drug REFERENCES treatment in older persons with isolated systolic hypertension: final results of the Systolic Hypertension in the Elderly Program (SHEP). JAMA. 1991;265:3255- 3264. 1. Rigaud AS, Seux ML, Staessen JA, Birkenha¨ger WH, Forette F. Cerebral compli- 32. Prince MJ, Bird AS, Blizard RA, Mann AH. Is the cognitive function of older pa- cations of hypertension. J Hum Hypertens. 2000;14:605-616. tients affected by antihypertensive treatment? results from 54 months of the Medi- 2. Birkenha¨ger WH, Forette F, Seux ML, Wang JG, Staessen JA. Blood pressure, cal Research Council’s treatment trial of hypertension in older adults. BMJ. 1996; cognitive functions, and prevention of dementias in older patients with hyper- 312:801-805. tension. Arch Intern Med. 2001;161:152-156. 33. Staessen JA, Gasowski J, Wang JG, et al. Risks of untreated and treated isolated 3. Fratiglioni L, De Ronchi D, Agu¨ero-Torres H. Worldwide prevalence and inci- systolic hypertension in the elderly: meta-analysis of outcome trials. Lancet. 2000; dence of dementia. Drugs Aging. 1999;15:365-375. 355:865-872. 4. Forette F, Seux ML, Staessen JA, et al. Prevention of dementia in randomised 34. Medical Research Council Working Party. Medical Research Council trial of treat- double-blind placebo-controlled Systolic Hypertension in Europe (Syst-Eur) trial. ment of hypertension in older adults: principal results. BMJ. 1992;304:405- Lancet. 1998;352:1347-1351. 412. 5. Collins R, MacMahon S. Reliable assessment of the effects of treatment on mor- 35. PROGRESS Collaborative Group. Randomised trial of a perindopril-based blood tality and major morbidity, I: clinical trials. Lancet. 2001;357:373-380. pressure lowering regimen among 6105 individuals with prior stroke or tran- 6. Staessen JA, Fagard R, Thijs L, et al. Randomised double-blind comparison of sient ischaemic attack. Lancet. 2001;358:1033-1041. placebo and active treatment for older patients with isolated systolic hyperten- 36. PROGRESS Management Committee. PROGRESS—Perindopril Protection Against sion. Lancet. 1997;350:757-764. Recurrent Stroke Study: characteristics of the study population at baseline. J Hy- 7. Gasowski J, Staessen JA, Celis H, et al. Systolic Hypertension in Europe (Syst- pertens. 1999;17:1647-1655. Eur) trial phase 2: objectives, protocol and initial progress. J Hum Hypertens. 37. Di Bari M, Pahor M, Franse LV, et al. Dementia and disability outcomes in large 1999;13:135-145. hypertension trials: lessons learned from the Systolic Hypertension in the El- 8. Forette F, Amery A, Staessen J, et al. Is prevention of vascular dementia pos- derly Program (SHEP) trial. Am J Epidemiol. 2001;153:72-78. sible? the Syst-Eur Vascular Dementia Project. Aging Clin Exp Res. 1991;3:373- 38. Morich FJ, Bieber F, Lewis JM, et al. Nimodipine in the treatment of probable 382. Alzheimer’s disease: result of two multicentre trials. Clin Drug Invest. 1996;11: 9. Amery A, Birkenha¨ger W, Bulpitt CJ, et al. Syst-Eur: a multicentre trial on the 185-195. treatment of isolated systolic hypertension in the elderly: objectives, protocol, 39. Parneti L, Senin U, Mecocci P. Cognitive enhancement therapy for Alzheimer’s and organization. Aging Clin Exp Res. 1991;3:287-302. disease: the way forward. Drugs. 1997;53:752-768. 10. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”: a practical method 40. Mason RP, Leeds PR, Jacob RF, et al. Inhibition of excessive neuronal apopto- for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975; sis by the calcium antagonist amlodipine and antioxidants in cerebellar granule 12:189-198. cells. J Neurochem. 1999;72:1448-1456. 11. Gagnon M, Letenneur L, Dartigues JF, et al. The validity of the Mini-Mental State 41. Amenta F, Strocchi P, Sabbatini M. Vascular and neuronal hypertensive brain Examination as a screening instrument for cognitive impairment and dementia damage: protective effect of treatment with nicardipine. J Hypertens. 1996;14 in French elderly community residents. Neuroepidemiology. 1990;3:143-150. (suppl 3):S29-S36. 12. American Psychiatric Association. Diagnostic and Statistical Manual of Mental 42. Zipfel GJ, Lee JM, Choi DW. Reducing calcium overload in the ischemic brain. Disorders, Revised Third Edition. Washington, DC: American Psychiatric Asso- N Engl J Med. 1999;341:1543-1544. ciation; 1987. 43. Pascale A, Etcheberrigaray R. Calcium alterations in Alzheimer’s disease: patho- 13. Launer LJ. Overview of incidence studies of dementia conducted in Europe. physiology, models and therapeutic opportunities. Pharmacol Res. 1999;39:81- Neuroepidemiology. 1992;11(suppl 1):2-13. 88. 14. Loeb C, Gandolfo C. Diagnostic evaluation of degenerative and vascular demen- 44. Thibault O, Porter NM, Chen KC. Calcium dysregulation in neuronal aging and tia. Stroke. 1983;14:399-401. Alzheimer’s disease: history and new directions. Cell Calcium. 1998;25:417- 15. Hachinski VC, Lliff LD, Zilkha E, et al. Cerebral blood flow in dementia. Arch 433. Neurol. 1975;32:632-637. 45. Hong YL. The relationship between calcium antagonist–induced hypotension and 16. Katz S, Ford AB, Moskowitz RW, Jackson BA, Jaffe MW. Studies of illness in the central monoaminergic system in spontaneously hypertensive rats. Int J Clin Phar- aged. JAMA. 1963;185:914-919. macol Ther Toxicol. 1987;11:589-597. 17. Hofman A, Ott A, Breteler MMB, et al. Atherosclerosis, apolipoprotein E, and preva- 46. Murzenok PP, Huang BS, Leenen FHH. Sympathoinhibition by central and pe- lence of dementia and Alzheimer’s disease in the Rotterdam Study. Lancet. 1997; ripheral infusion of nifedipine in spontaneously hypertensive rats. Hyperten- 349:151-154. sion. 2000;35:631-636. 18. Breteler MMB, Claus JJ, Grobbee DE, Hofman A. Cardiovascular disease and dis- 47. Gould RJ, Murphy KMM, Snyder SH. Autoradiographic localization of calcium tribution of cognitive function in elderly people: the Rotterdam study. BMJ. 1994; channel antagonist receptors in rat brain with [3H]nitrendipine. Brain Res. 1985; 308:1604-1608. 330:217-223. 19. Launer LJ, Webster Ross G, Petrovitch H, et al. Midlife blood pressure and de- 48. Letenneur L, Commenges RD, Dartigues JF, Barberger-Gateau P. Incidence of mentia: the Honolulu-Asia Aging Study. Neurobiol Aging. 2000;21:49-55. dementia and Alzheimer’s disease in elderly community residents of south- 20. Prince M, Cullen M, Mann A. Risk factors for Alzheimer’s disease and dementia: western France. Int J Epidemiol. 1994;23:1256-1261. a case-control study based on the MRC elderly hypertension trial. Neurology. 49. Ott A, Breteler MMB, van Harskamp F, Grobbee DE, Hofman A. The incidence of 1994;44:97-104. dementia in the Rotterdam study. In: Iqbal K, Winblad B, Nishimura T, Takeda K, 21. Breteler MMB. Vascular risk factors for Alzheimer’s disease: an epidemiologic Winieski H, eds. Alzheimer’s Disease: Biology, Diagnosis and Therapeutics. New perspective. Neurobiol Aging. 2000;21:153-160. York, NY: John Wiley & Sons; 1997:1-10. 22. Pasquier F, Leys D. Why are stroke patients prone to develop dementia? J Neu- 50. Blood Pressure Lowering Treatment Trialists’ Collaboration. Effect of ACE in- rol. 1997;244:135-142. hibitors, calcium antagonists, and other blood pressure-lowering drugs: results 23. Skoog I, Marcusson J, Blennow K. It’s getting better all the time. Lancet. 1998; of prospectively designed overviews of randomised trials. Lancet. 2000;355: 352(suppl 4):SIV4. 1955-1964. 24. Hachinski V. Vascular dementia: the need for a new approach. J Intern Med. 1997; 51. Staessen JA, Wang JG, Thijs L. Cardiovascular prevention and blood pressure 242:277-279. reduction: a meta-analysis. Lancet. 2001;358:1305-1315. 25. Andersen K, Launer LJ, Dewey ME, et al. Gender differences in the incidence of 52. The Joint National Committee on Prevention, Detection, Evaluation, and Treat- AD and vascular dementia: the EURODEM Studies. Neurology. 1999;53:1992- ment of High Blood Pressure. The sixth report of the Joint National Committee 1997. on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 26. Skoog I, Nilsson L, Palmertz B, Andreasson LA, Svanborg A. A population- VI). Arch Intern Med. 1997;157:2413-2446. based study of dementia in 85-year-olds. N Engl J Med. 1993;328:153-158. 53. Pahor M, Psaty BM, Alderman MH, et al. Health outcomes associated with cal- 27. Ruitenberg A, Ott A, van Swieten JC, Hofman A, Breteler MB. Incidence of de- cium antagonists compared with other first-line antihypertensive therapies. Lan- mentia: does gender make a difference? Neurobiol Aging. 2001;22:575-580. cet. 2000;356:1949-1954. 28. Guo Z, Fratiglioni L, Zhu L, Fastbom J, Winblad B, Viitanen M. Occurrence and 54. Maxwell CJ, Hogan DB, Ebly EM. Calcium-channel blockers and cognitive func- progression of dementia in a community population aged 75 years and older: tion in elderly people: results from the Canadian Study of Health and Aging. CMAJ. relationship with medication use. Arch Neurol. 1999;56:991-996. 1999;161:501-506.

(REPRINTED) ARCH INTERN MED/ VOL 162, OCT 14, 2002 WWW.ARCHINTERNMED.COM 2052

Downloaded From: https://jamanetwork.com/ on 10/02/2021 Although this is by no means a rig- This study was supported in part by 3. Liu YH, Yang XP, Sharov VG, et al. Effects of angiotensin-converting enzyme inhibitors and an- orously controlled study, it should grant P50HL55001 from the Na- giotensin II type 1 receptor antagonists in rats help alleviate the concerns of phy- tional Heart, Lung, and Blood Insti- with heart failure: role of kinins and angioten- sicians who may be reluctant to use tute, Bethesda, Md. sin II type 2 receptors. J Clin Invest. 1997;99: 1926-1935. an ARB in such patients, despite an- 4. Gohlke P, Pees C, Unger T. AT2 receptor stimu- ticipated benefits. lation increases aortic cyclic GMP in SHRSP by 1. Gavras I. Bradykinin-mediated effects of ACE in- a kinin-dependent mechanism. Hypertension. hibition. Kidney Int. 1992;42:1020-1029. 1998;31:349-355. Irene Gavras, MD 2. Griendling KK, Lassegue B, Alexander RW. An- 5. Lacourciere Y, Brunner H, Irwin R, et al. Effects giotensin receptors and their therapeutic impli- of modulators of the renin-angiotensin- Haralambos Gavras, MD, FRCP cations. Annu Rev Pharmacol Toxicol. 1996;36: aldosterone system on cough: Losartan Cough Boston, Mass 281-306. Study Group. J Hypertens. 1994;12:1387-1393.

Correction

Error in Figure. In the Original Investigation by Forette et al titled “The Pre- vention of Dementia With Antihypertensive Treatment: New Evidence From the Systolic Hypertension in Europe (Syst-Eur) Study,” published in the October 14, 2002, issue of the ARCHIVES (2002;162:2046-2052), an error occurred in Figure 2 on page 2048. In the key to that figure, patients randomized to placebo should have been indicated with open circles and those allocated to active treatment, with closed circles. The corrected figure is reprinted here. The journal regrets the error.

180 Placebo Active Treatment 1417 Systolic 170

160 1417 1417 1417 381 742 515 1415 1016 150

140

Blood Pressure, mm Hg 85 1485 Diastolic

80 1485 1485 1485 1485 1072 758 552 75 410

01234 5 6 7 8 Time Since Randomization, y

Figure 2. Average sitting systolic and diastolic blood pressure at randomization and during follow-up. For each mean value, the number of patients is given.

(REPRINTED) ARCH INTERN MED/ VOL 163, JAN 27, 2003 WWW.ARCHINTERNMED.COM 241

Downloaded From: https://jamanetwork.com/ on 10/02/2021