ALZHEIMER S disease (AD) is the most prevalent form

Transcription

1 Journal of Gerontology: MEDICAL SCIENCES 2003, Vol. 58A, No. 4, Copyright 2003 by The Gerontological Society of America Association of Alzheimer s Disease Onset With Ginkgo Biloba and Other Symptomatic Cognitive Treatments in a Population of Women Aged 75 Years and Older From the EPIDOS Study Sandrine Andrieu, 1,2 Sophie Gillette, 1,2 Karine Amouyal, 1 Fati Nourhashemi, 1,2 Emma Reynish, 1 Pierre Jean Ousset, 1 Jean Louis Albarede, 1 Bruno Vellas, 1 and Hélène Grandjean 2 1 Department of Internal Medicine and Clinical Gerontology, Acute Unit for Alzheimer s Patients, Toulouse, France. 2 Institut National de la Santé et de la Recherche Médicale (INSERM) U558, Toulouse, France. Background. Peripheral C4A treatment (cerebral and peripheral vasotherapeutics) and especially Ginkgo biloba extracts are prescribed for a number of symptoms, particularly memory impairment, in elderly patients. It is postulated that because of its pharmacological actions, this treatment could prevent the decline of cognitive function, but no studies have been published to date to test its efficacy in prevention of Alzheimer s disease. The potential association between use of C4A treatments, in particular EGb 761 (standardized Ginkgo biloba extracts), and dementia of the Alzheimer type was investigated. Methods. A case-control study was nested in a cohort of 1462 community-dwelling elderly women aged over 75 years. Sixty-nine women with Alzheimer-type dementia were compared with 345 paired women whose cognitive function remained normal. This study involved women whose cognitive function was evaluated at baseline by use of Pfeiffer s test and whose medication history was taken. The onset of cognitive impairment was investigated over a 7-year follow-up period. In order to study the factors associated with the onset of dementia, the data concerning women with a score of >8 on Pfeiffer s test at inclusion, indicating normal cognitive function, were analyzed. Results. A multivariate analysis including potential confounding factors showed that fewer women who developed Alzheimer s dementia had been prescribed C4A treatment (including EGb 761) for at least 2 years (odds ratio , 95% confidence interval , p 5.018). Figures for EGb 761 alone were similar but did not reach statistical significance (odds ratio , 95% confidence interval , p 5.22). Conclusion. These results suggest that C4A treatment may reduce the risk of developing Alzheimer s dementia in elderly women. The potential preventive effect of C4A treatments, including EGb 761, requires further examination. To establish a causal relationship, these findings have to be confirmed with prospective studies. ALZHEIMER S disease (AD) is the most prevalent form of dementia, affecting 5% of people over 65 years old and nearly 30% of people over 80 (1). It is essential to develop treatments that are effective in preventing the onset of the disease. C4A treatments (cerebral and peripheral vasotherapeutics), including Ginkgo biloba extracts, are widely used in Europe and in the United States, particularly for memory impairment in elderly persons. These medications are well tolerated and could have potential for preventing cognitive decline. Research in this field is particularly important today because of the relationship between cognitive decline and increased risk of AD in elderly people (2). A number of studies have investigated the efficacy of standardized Ginkgo biloba extracts (EGb 761) in patients with altered cognitive function or AD. In a review of the literature, Oken and colleagues identified four trials of rigorous design (double blind, placebo controlled). After meta-analysis, they concluded that treatment with EGb 761 had a moderate but significant effect on cognitive function (3). More recently, two double-blind, randomized, placebocontrolled clinical trials reported conflicting results concerning the efficacy of EGb 761 in the treatment of AD (4,5). In addition to its vasodilator properties (6), Egb 761 is also an antioxidant. Oxidative stress is well recognized as one of the mechanisms involved in the development of AD (7,8), and because EGb 761 has antioxidant properties, it could have a preventive effect on the onset of this disease. Prospective large-scale long-term randomized studies are necessary to establish this effect. Prior to planning such studies, we had the opportunity to examine this hypothesis in an ongoing study, the EPIDOS (EPIDemiology of OSteoporosis) study. In this multicenter study, women (>75 years) were assessed cognitively at baseline and were then followed for 4 years, during which the history of their prescribed drugs was recorded. At the end of the EPIDOS study the cognitive function of the participants from one center, Toulouse, was reassessed, allowing us to perform a case-control study in this cohort. METHODS The EPIDOS Study The EPIDOS study was a prospective multicenter study with the aim of investigating the risk factors for femoral 372

2 C4A TREATMENT AND RISK OF ALZHEIMER S DISEASE 373 Table 1. Peripheral Treatments Coded as C4A 1. Ginkgo biloba extracts 2. Other C4A drugs Naftidrofuryl Vinburnine Raubasine-Almitrine Nicergoline Ergot alkaloids Piribedil Ifenprodil Buflomedil Cinnarizine Pentoxifylline Vincamine Moxisylyte Cyclandelate Papaverine neck fracture. From January 1992 to January 1994, 7598 ambulatory women aged 75 years and over in 5 French cities (Amiens, Lyon, Montpellier, Paris, and Toulouse) volunteered to participate in this study. The protocol has been described in detail elsewhere (9). At baseline, women were assessed for physical and functional capacities. Autonomy was assessed by three items concerning basic activities of everyday life (dressing, washing, and mobility) and Lawton s Instrumental Activities of Daily Living (IADL) scale (10). Cognitive function was determined by Pfeiffer s test (short portable mental status questionnaire), which yields scores ranging from 0 to 10, with 10 representing the best performance (11). This test has good sensitivity and specificity for detecting cases of dementia compared with diagnosis by a clinician (12). Demographic data (age, education, family situation, and income), medical history, life habits (social support, activities, lifestyle, accommodation, etc.), and treatments were recorded. As a way to assess treatment, subjects were asked to bring their prescribed medications to the visit. All peripheral treatments coded as C4A in the European Pharmaceutical Marketing Research Association (EphMRA) classification, from which the Anatomical Therapeutic Chemical (ATC) system originated ( The study was approved by the local ethics committees, and all participating women signed informed consent forms. Every 4 months, participants completed a questionnaire about falls, fractures and other new conditions. At yearly intervals, they also listed the medications they were taking. Toulouse Cohort Study TheToulousecohort(n51462)wasfollowed-upattheUniversity Hospital and all cases of dementia of the Alzheimer type were recorded. All participants were invited to take part in an additional follow-up study. Those who had given informed consent were assessed either at home or at the Department of Internal Medicine and Clinical Geriatrics. During this visit, demographic data and information on medical conditions and prescribed drugs were recorded. Instrumental activities of daily living (IADL) and cognitive function were reassessed. Cognitive function was tested by use of Pfeiffer s test (11), Folstein s Mini-Mental State Examination (13), and the test from Grober and colleagues (14). This information was analyzed independently by a geriatrician and a neurologist from the Department of Internal Medicine and Clinical Geriatrics, who had expertise in AD and were unaware of medication taken. Diagnostic and Statistical Manual, 4th edition (DSM-IV) criteria were used to establish a clinical diagnosis of dementia. AD was diagnosed according to the criteria of the NINCDS-ADRDA work group (15). When the two assessors disagreed on a diagnosis, a joint decision was reached after additional information was obtained from the general practitioner. Subjects were then classified into four groups: normal cognitive function, mild cognitive impairment according to Petersen s criteria (16), AD, and other types of dementia. Initially, 1462 women were included in the EPIDOS study in Toulouse. During the initial 4-year follow-up, 72 women were diagnosed with dementia of the Alzheimer type. Of the 642 women who were followed up for 7 years, 450 were considered as normal, 38 had dementia of the Alzheimer type, and 154 had other disturbances of cognitive function (mild cognitive impairment or another type of dementia, including mixed dementia). Thus, at the end of the study, data on cognitive status were available for 714 women (48.8% of the initial cohort), of whom 110 had dementia of the Alzheimer type. Of the other 748 women, whose cognitive status remained undetermined, 25.8% (n 5 193) died during follow-up, 54.7% (n 5 414) were lost to follow-up, and 18.9% (n 5 141) withdrew from the follow-up study. Statistical Analysis Analysis was performed with SAS software (SAS Version 8.0; SAS Institute, Inc., Cary, NC). Bivariate analyses were based on conventional tests comparing means for quantitative variables (analysis of variance), and frequency distributions for qualitative variables (chi-square test or Fisher exact test, according to sample size). Comparisons were made between AD patients and a control group consisting of five controls for one case. For all cases, medication history was available from six follow-up consultations: baseline (T0), each year of follow-up study (T1, T2, T3, and T4), and at the final examination (T7). For the cases diagnosed as AD during follow-up and for their five corresponding controls, only data preceding the diagnosis were used. In order to study the factors associated with the onset of dementia, only data for women with a score >8 on Pfeiffer s test at baseline (17,18), who were thus considered as cognitively normal, were used. The analysis included 69 women with AD of the 110 women who were diagnosed with AD, and 345 controls with normal cognitive evaluation. To account for potential confounding factors, the variables statistically related to AD and those related to exposure to C4A treatment, with a value of p,.25, were included in the logistic regression model, with Alzheimer s dementia as a dependent variable. The model was fitted by using the Hosmer Lemeshow test (19). RESULTS Those patients lost to follow-up were initially in worse health than those included in the study. However, there were

3 374 ANDRIEU ET AL. Table 2. Characteristics of Women With and Without ATD Baseline Parameters Non-ATD no differences concerning C4A prescription between the 748 women lost to follow-up and the 714 women for whom follow-up data were available (38.8% vs 39.0%). Exposure to C4A Medications In the study population, 50.7% (n 5 210) of women reported having taken a C4A medication (Table 1); 32.9% (n 5 69) of this subgroup had been given EGb 761, 55.7% (n 5 117) had been given another C4A medication, and 11.4% (n 5 24) had been given both types of treatment. Factors Associated With AD Women diagnosed with AD during follow-up were considerably older at inclusion than nondemented subjects ( years vs years; p 5.02). They generally had lower financial status and level of education, and poorer perception of their health with difficulty performing activities of daily living (ADL) (Tables 2 and 3). They also performed slightly less well on the Pfeiffer test ( vs ; p 5.03). Analysis of the medication histories of patients with AD showed they were less likely to have taken C4A treatments for three consecutive periods, but there were no other significant differences regarding other treatments (aspirin, calcium channel blockers, ATD p OR 95% CI Age.0292 <80 y 247 (71.6) 40 (57.9) (22.9) 20 (29.0) yr 19 (5.5) 9 (13.1) End of primary school certificate No (,5 y) 29 (8.4) 11 (15.9) 1 Yes (.5 y) 316 (91.6) 58 (84.1) Income (euros/mo) (50.1) 17 (24.6) (23.2) 26 (37.7) ,457 9 (2.6) 4 (5.8) Unknown 83 (24.1) 22 (31.9) Perceived health Good 284 (85.5) 46 (66.7) 1 Poor 48 (14.5) 23 (33.3) Marital status.3125 Married 100 (29.0) 14 (20.3) 1 Widowed 190 (55.1) 41 (59.4) Divorced 20 (5.8) 7 (10.1) Single 35 (10.1) 7 (10.1) Washing, dressing, & walking Independent 329 (95.4) 60 (87.0) 1 Dependent 16 (4.6) 9 (13.0) IADL 0 incapacity 315 (91.3) 56 (81.2) 1 >1 incapacity 30 (8.7) 13 (18.8) Depression No 309 (89.8) 59 (85.5) 1 Yes 35 (10.2) 10 (14.5) Hearing problem No 158 (45.8) 37 (53.6) 1 Yes 187 (54.2) 32 (46.4) Notes: ATD 5 Alzheimer s-type dementia; OR 5 odds ratio; CI 5 confidence interval; IADL 5 instrumental activities of daily living. Non-ATD = 345 (83.33%); ATD 5 69 (16.67%). mineral supplements, or nonsteroidal anti-inflammatory drugs). Factors Associated With C4A Treatment Women who had received C4A treatment were older at inclusion ( years vs years; p 5.01), but they were otherwise not significantly different from the untreated group (Table 4). Logistic Regression Table 5 shows the results of the final model. Three variables were excluded, using the stepwise process model (ADL, IADL, and education level). Logistic regression showed that women with dementia appeared to have been regularly exposed to C4A treatment less often than had others (odds ratio , 95% confidence interval [CI] , p 5.018). In an additional analysis, the variable C4A exposure was replaced with EGb or Ginkgo exposure. The results were similar but did not reach statistical significance with odds ratio , 95% CI , and p 5.59; odds ratio , 95% CI , and p 5.41; and odds ratio , 95% CI , and p 5.22 for at least one exposure, two consecutive exposures, and at least three consecutive exposures, respectively.

4 C4A TREATMENT AND RISK OF ALZHEIMER S DISEASE 375 Table 3. More Characteristics of Women With and Without ATD Treatment Non-ATD DISCUSSION Alzheimer s dementia is a neurodegenerative disease with long-term evolution before definitive lesions. With an unprecedented demographic change in the Western population, prevention and effective treatment of age-dependent disease is a public health priority. Many elderly people worried about cognitive impairment are requesting therapy, and millions of people in Europe and the United States are regularly taking Ginkgo biloba extracts. The value of this treatment deserves study. Our objective was to investigate the potential association between use of C4A treatments, in particular EGb 761 (standardized Ginkgo biloba extracts), and Alzheimer s dementia onset in elderly women. We found that women regularly taking C4A treatment were three times more numerous in the group without dementia than in the group with dementia. The odds ratio for EGb 761 was similar but did not reach statistical significance. Women participating in the EPIDOS study were volunteers and therefore not representative of the general population. Cognitive status was evaluated in only 50% of the initial population, so we could not study the relative risk of AD associated with C4A prescription. However, our study was prospective and could precisely determine the cognitive ATD p OR 95% CI Estrogens No 342 (99.1) 68 (98.6) 1 Yes 3 (0.9) 1 (1.4) Mineral supplements.8006 Not exposed 249 (72.2) 50 (72.5) 1 Nonconsecutive exposure(s) 77 (22.3) 14 (20.3) consecutive exposures 12 (3.5) 4 (5.8) >3 consecutive exposures 7 (2.0) 1 (1.4) Calcium antagonist.3634 Not exposed 242 (70.1) 49 (71.0) 1 Nonconsecutive exposure(s) 23 (6.7) 8 (11.6) consecutive exposures 31 (9.0) 6 (8.7) >3 consecutive exposures 49 (14.2) 6 (8.7) Aspirin.9078 Not exposed 232 (67.2) 44 (63.8) 1 Nonconsecutive exposure(s) 62 (18.0) 15 (21.7) consecutive exposures 25 (7.2) 5 (7.2) >3 consecutive exposures 26 (7.5) 5 (7.2) NSAIDs.1811 Not exposed 270 (78.3) 51 (73.9) 1 Nonconsecutive exposure(s) 43 (12.5) 11 (15.9) consecutive exposures 14 (4.1) 6 (8.7) >3 consecutive exposures 18 (5.2) 1 (1.5) C4A treatment.0289 Not exposed 169 (49.0) 35 (50.7) 1 Nonconsecutive exposure(s) 57 (16.5) 16 (23.2) consecutive exposures 39 (11.3) 12 (17.4) >3 consecutive exposures 80 (23.2) 6 (8.7) EGb or Ginkgo.5254 Not exposed 264 (76.5) 57 (82.6) 1 Nonconsecutive exposure(s) 37 (10.7) 8 (11.6) consecutive exposures 20 (5.8) 2 (2.9) >3 consecutive exposures 24 (7.0) 2 (2.9) Notes: ATD5 Alzheimer s-type dementia; OR 5 odds ratio; CI 5 confidence interval; NSAIDs 5 nonsteroidal anti-inflammatory drugs; C4A 5 cerebral and peripheral vasotherapeutics; EGb 5 standardized Ginkgo biloba extracts. status of the women (normal or AD). Therefore, a casecontrol study nested in the cohort was possible. One of the strong points of the study is the fact that the population had cognitive assessments at the outset by the Pfeiffer test and only those with scores of 8 or more were included. Because C4A treatments are physician prescribed, we could be sure that all patients received the standardized extract of Ginkgo biloba. The medication history was recorded each year, so we assumed patients who had three consecutive recordings of C4A medication received the treatment long term. It was this group of patients who showed significant results. However, in the early stages of dementia, women could have forgotten whether they had taken specific medication. We postulate that they were not more likely to have forgotten their C4A treatment than their other treatments, and yet there was no difference in the two groups for any other type of medication (aspirin, calcium channel blockers, mineral supplements, nonsteroidal anti-inflammatory drugs). Epidemiological studies should help to identify risk factors for AD. Our study has once again identified age attained as an independent risk factor. It has not, however, shown any relationship between aspirin or nonsteroidal anti-inflammatory treatment and risk for AD. This may be due to the relatively infrequent use of these drugs in the study population.

5 376 ANDRIEU ET AL. Table 4. Bivariate Analysis: Characteristics of Women According to Exposure to C4A Treatment at the Time of Diagnosis Baseline Parameters Not Exposed Conclusion In conclusion, our study including 414 patients over 75 years old showed that women with diagnosed dementia had been less frequently exposed to chronic use of C4A treatment. This effect remained significant after adjustment for potential confounding effects. Although these results should be interpreted with caution because of the study design, the evidence for the protective effects of other treatments was not stronger than that shown for C4A treatment. Primary prevention trials should be undertaken to confirm these findings. In fact, large interventional studies are ongoing with EGb 761 in North America and Europe to determine the preventive effect of these treatments, but results will not be available for 6 to 7 years. A recent study with Ginkgo extract treatment has shown no effects on memory (20); however, this study was done in younger people (68.5 years) with no cognitive complaints and with only 6 weeks of treatment. The effect of Ginkgo on oxidative stress and on vascular function will require much longer time review to be effective. Following our long-term observational study, some long-term placebo-controlled studies are on the way over a 1-year period to prevent dementia in older elderly persons (more than 75 years) with cognitive complaints. 1 Exposure C4A Treatments 2 Exposures >3 Exposures Age.01 <80 y 154 (75.5) 43 (58.9) 37 (72.6) 53 (61.6) (21.6) 21 (28.8) 10 (19.6) 24 (27.9).85 y 6 (2.9) 9 (12.3) 4 (7.8) 9 (10.5) End of primary school certificate.20 No 16 (7.8) 12 (16.4) 5 (9.8) 7 (8.1) Yes 188 (92.2) 61 (83.6) 46 (90.2) 79 (91.9) Income euros/mo) (44.6) 29 (39.7) 23 (45.1) 47 (54.7) (26.5) 23 (31.5) 12 (23.5) 17 (19.8),457 6 (2.9) 3 (4.1) 0 (0.0) 4 (4.7) Unknown 53 (26.0) 18 (24.7) 16 (31.4) 18 (20.9) Perceived health.25 Good 163 (82.7) 54 (76.1) 39 (79.6) 74 (88.1) Poor 34 (17.3) 17 (23.9) 10 (20.4) 10 (11.9) Marital status.42 Married 63 (30.9) 17 (23.3) 15 (29.4) 19 (22.1) Widowed 110 (53.9) 43 (58.9) 32 (62.7) 46 (53.5) Divorced 13 (6.4) 5 (6.8) 1 (2.0) 8 (9.3) Single 18 (8.8) 8 (11.0) 3 (5.9) 13 (15.1) Washing, dressing, & walking.17 Independent 194 (95.1) 67 (91.8) 45 (88.2) 83 (96.5) Dependent 10 (4.9) 6 (8.2) 6 (11.8) 3 (3.5) Binary IADL.97 0 incapacity 183 (89.7) 65 (89.0) 45 (88.2) 78 (90.7) >1 incapacity 21 (10.3) 8 (11.0) 6 (11.8) 8 (9.3) Hearing problem.07 No 108 (52.9) 33 (45.2) 23 (45.1) 31 (36.1) Yes 96 (47.1) 40 (54.8) 28 (54.9) 55 (63.9) Depression.18 No 177 (86.8) 69 (94.5) 47 (94.0) 75 (87.2) Yes 27 (13.2) 4 (5.5) 3 (6.0) 11 (12.8) Notes: C4A5 cerebral and peripheral vasotherapeutics; not exposed, n (49.3%); 1 exposure, n 5 73 (17.6%); 2 exposures, n 5 51 (12.3%); >3 exposures, n 5 86 (20.8%). IADL 5 instrumental activities of daily living. However, a preliminary study of the effect on memory of 6 weeks of treatment with Ginkgo gave negative results (20). If the protective effect of C4A treatment is confirmed, this class of medication is a strong contender for widespread Table 5. Multivariate Analysis (Logistic Regression) Model Variable to be Explained p OR 95% CI Pfeiffer test in Age (y) Income (ref:.915 euros) , Unknown Perceived health (poor vs good) Hearing problem C4A treatment (ref: not exposed).0698 Nonconsecutive exposure(s) consecutive exposures >3 consecutive exposures Adequacy of the model (Hosmer Lemeshow test).7175 Notes: For variable to be explained, dementia of the Alzheimer s type (n 5 69) is versus dementia not of the Alzheimer s type (n 5 332). OR 5 odds ratio; CI 5 confidence interval. p

6 C4A TREATMENT AND RISK OF ALZHEIMER S DISEASE 377 preventive treatment use because of its minimal side effects compared with others such as nonsteroidal anti-inflammatory drugs or estrogens. ACKNOWLEDGMENTS This work was part of the EPIDOS study supported by an Institut Nationale de la Santé et de la Recherche Médicale (INSERM)/MSD- Chibret contract. Additional follow-up was supported by an INSERM/ Beaufour Ipsen contract. EPIDOS study participants included coordinators G. Breart, P. Dargent-Molina, P. J. Meunier, A. M. Scott, D. Hans, and P. D. Delmas, and principal investigators C. Baudoin and J. L. Sebert (Amiens), M. C. Chapuy and A. M. Scott (Lyon), F. Favier and C. Marcelli (Montpellier), C. J. Menkes, C. Cormier, and E. Hausherr (Paris), and H. Grandjean and C. Ribot (Toulouse). The authors thank Ms. Christelle Cantet for her collaboration in this study. Address correspondence and to Sandrine Andrieu, MD, PhD, Department of Internal Medicine and Clinical Gerontology, Acute Unit for Alzheimer s Patients, 170 Chemin de Casselardit, Toulouse Cedex 03, France. sandrieu@cict.fr REFERENCES 1. Small GW, Rabins PV, Barry PP, et al. Diagnosis and treatment of Alzheimer disease and related disorders. Consensus statement of the American Association for Geriatric Psychiatry, the Alzheimer s Association, and the American Geriatrics Society. JAMA. 1997;278: Geerlings MI, Jonker C, Bouter LM, Ader HJ, Schmand B. Association between memory complaints and incident Alzheimer s disease in elderly people with normal baseline cognition. Am J Psychiatr. 1999; 156: Oken BS, Storzbach DM, Kaye JA. The efficacy of Ginkgo biloba on cognitive function in Alzheimer disease. Arch Neurol. 1998;55: Le Bars PL, Kieser M, Itil KZ. A 26-week analysis of a doubleblind, placebo-controlled trial of the Ginkgo biloba extract EGb 761 in dementia. Dement Geriatr Cogn Disord. 2000;11: Van Dongen MC, van Rossum E, Kessels AG, Sielhorst HJ, Knipschild PG. The efficacy of Ginkgo for elderly people with dementia and age-associated memory impairment: new results of a randomized clinical trial [in process citation]. J Am Geriatr Soc. 2000;48: Dealaflotte S, Auguet M, DeFeudis FV, et al. Endothelium-dependent relaxations of rabbit isolated aorta produced by carbachol and by Ginkgo biloba extract. Biomed Biochim Acta. 1984;43:S212 S Smith MA, Perry G, Richey PL, et al. Oxidative damage in Alzheimer s. Nature. 1996;382: Halliwell B. Free radicals, antioxidants, and human disease: curiosity, cause, or consequence? Lancet. 1994;344: Dargent-Molina P, Favier F, Grandjean H, et al. Fall-related factors and risk of hip fracture: the EPIDOS prospective study. Lancet. 1996; 348: Lawton M, Brody E. Assessment of older people: self-maintaining and instrumental activities of daily living. Gerontologist. 1969;9: Pfeiffer E. A short portable mental status questionnaire for the assessment of organic brain deficit in elderly patients. J Am Geriatr Soc. 1975;23: Fillenbaum G, Heyman A, Williams K, Prosnitz B, Burchett B. Sensitivity and specificity of standardized screens of cognitive impairment and dementia among elderly black and white community residents. J Clin Epidemiol. 1990;43: Folstein M, Folstein S, McHugh P. Mini-Mental State. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12: Grober E, Buschke H, Crystal H, Bang S, Dresner R. Screening for dementia by memory testing. Neurology. 1988;38: McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer s disease: report of the NINCDS- ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer s Disease. Neurology. 1984;34: Petersen RC, Smith GE, Waring SC, Ivnik RJ, Tangalos EG, Kokmen E. Mild cognitive impairment: clinical characterization and outcome. Arch Neurol. 1999;56: Rozzini R, Ferrucci L, Losonczy K, Havlik RJ, Guralnik JM. Protective effect of chronic NSAID use on cognitive decline in older persons. J Am Geriatr Soc. 1996;44: Erkinjuntti T, Sulkava R, Wikstrom J, Autio L. Short Portable Mental Status Questionnaire as a screening test for dementia and delirium among the elderly. J Am Geriatr Soc. 1987;35: Hosmer D, Lemeshow S. Applied Logistic Regression. New York: John Wiley & Sons; Solomon PR, Adams F, Silver A, Zimmer J, DeVeaux R. Ginkgo for memory enhancement: a randomized controlled trial. JAMA. 2002; 288: Received May 15, 2002 Accepted October 18, 2002