The Geriatric Depression Scale is a self-administered depression scale, which was developed as a basic screening measure for depression in older adults. By "yes" or "no" answers, scores permit to classify patients into groups of "severely depressed" (score of 21 to 30), "moderately depressed" (score of 11 to 20) and "normal" (score of 0 to10). It takes10 to 15 minutes to administer.

Subjects completed a verbal fluency test. Higher scores represent higher levels of verbal fluency. Minimum score=0 and Maximum score= N/A.

Letter fluency: this task consists of enouncing as many words as possible that begin with a given letter of the alphabet. Participants are not allowed to use proper names.

Categorical fluency: in this task participants are asked to list as many words as possible that belong to a given semantic category (e.g. animals, fruits, towns) Each condition foresees 60 sec of word generation time. The score corresponds to the number of words correctly given. The verbal fluency task measures semantic storage and executive retrieval functions.

MMSE is a brief screening instrument used to assess cognitive function in elderly participants. It assesses orientation, memory, attention, ability to name objects, follow verbal and written commands, write a sentence, and copy figures. Total score ranges from 0 to 30, with a lower score indicating greater disease severity.

Clock drawing test is a visuo-constructive task, where subjects are asked to draw the face of a clock in a pre-drawn circle and then to draw in the arms to denote 16:45 (a quarter to five). The drawing can then be evaluated by a quantitative scoring method, which is based on the degree of completion of the drawing. The scoring system ranges from 0 to 6 with higher scores reflecting a greater number of errors and more impairment.

Cube drawing: Subjects are asked to draw a cube by heart. In case of failure, a model of a cube is given to the subjects to copy.

The score system ranges from 0 (worse score) to 6 (best score). Score calculation is following: 1 point by face with 4 sides, 2 points for each face where each angle should be respected.

Free and Cued Selective Reminding Test (FCSRT) Assessment of verbal episodic memory. By this test performances in free recalls, cued recalls and in a recognition task can be analysed, because the process of encoding is controlled.

Subjects are asked to remember a list of 16 words. Three tasks of free and cued recalls, as well as one recognition task and one delayed recall give the scores.

Total recall is obtained by the addition of cued recalls to free recalls. Maximum score is 48 for immediate: 16 words X 3 corresponding to immediate free recall + immediate cued recall + immediate recognition test.

Maximum score is 64 (better score) when delayed recall : 16 words X 4 . The minimum score is 0 (worse).

Subjects are asked to memorize two short stories, one consisting of 24, the other one of 26 information units. After the stories have been read aloud by the investigator, subjects are asked to enounce all items of information they can remember (free recall). Correctly reported items are added for each story. This test applies analytical capacities, as well as auditive and verbal synthesis, working memory and episodic memory.

Score range from 0(worst) to 75 (better).

Wechsler Adult Intelligence Scale (WAIS) In this test subjects are asked to say how two seemingly dissimilar items might in fact be similar (14 item couples). This test involves especially abstract thinking and concept capacities.

Cognitive Tests-Age-Adjusted Wechsler Adult Intelligence Scale (WAIS): 19 tests on similarities. The items 1 to 5 are graded from 1 (good) to 0 (bad), and the items 6 to 19 are graded from 2 (good) to 0 (bad). Item 6 and 7 can be repeated. At the end, the worst total score is 0, the best total score is 33.

TMT is a neuropsychological test of visual attention and task switching. The task requires a subject to 'connect-the-dots' of 25 consecutive numbers (1,2,3,etc.) on a sheet of paper or computer screen. The goal of the subject is to finish the test as quickly as possible and the time taken to complete the test is used as the primary performance metric (in seconds). The maximum time allowed is 300 seconds. A negative change score indicates improvement.

First, in the TMT A, the subject has to connect numbers increasingly as fast as possible.

The TMT B requires the subject to connect and letters in an alternating pattern (1-A-2-B-3-C, etc.) in as little time as possible.

The Diagnostic and Statistical Manual of Mental Disorders: 4th Edition of the American Psychiatric Association (DSM-IV, 1994) also outlines diagnostic criteria for dementia of the Alzheimer's type that are generally consistent with the NINCDS-ADRDA criteria.

National Institute of Neurological and Communicative Diseases and Stroke / Alzheimer's Disease and Related Disorders Association (NINCDS/ADRDA)

The most widely accepted diagnostic criteria for probable AD are those offered by the National Institute of Neurological and Communicative Disorders and Stroke and by the Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA; McKhann et al., 1984). These criteria include the presence of dementia established by clinical examination and confirmed by neuropsychological testing. The dementia is described as involving multiple, progressive cognitive deficits in older persons in the absence of disturbances of consciousness, presence of psychoactive substances, or any other medical, neurological, or psychiatric conditions that might in and of themselves account for these progressive deficits.

CDR is a structured interview to collect information regarding subject's memory in a standard way from both the patient and the helper. Scores are calculated using below scale; q CDR=No dementia (score: 0), q CDR=Very mild dementia (score: 0.5), q CDR=Mild dementia (score: 1), q CDR=Moderate dementia (score: 2) and q CDR=Severe dementia (score: 3)

The Geriatric Depression Scale is a self-administered depression scale, which was developed as a basic screening measure for depression in older adults. By "yes" or "no" answers, scores permit to classify patients into groups of "severely depressed" (score of 21 to 30), "moderately depressed" (score of 11 to 20) and "normal" (score of 0 to10). It takes10 to 15 minutes to administer.

Subjects completed a verbal fluency test. Higher scores represent higher levels of verbal fluency. Minimum score=0 and Maximum score= N/A.

Letter fluency: this task consists of enouncing as many words as possible that begin with a given letter of the alphabet. Participants are not allowed to use proper names.

Categorical fluency: in this task participants are asked to list as many words as possible that belong to a given semantic category (e.g. animals, fruits, towns) Each condition foresees 60 sec of word generation time. The score corresponds to the number of words correctly given. The verbal fluency task measures semantic storage and executive retrieval functions.

MMSE is a brief screening instrument used to assess cognitive function in elderly participants. It assesses orientation, memory, attention, ability to name objects, follow verbal and written commands, write a sentence, and copy figures. Total score ranges from 0 to 30, with a lower score indicating greater disease severity.

Clock drawing test is a visuo-constructive task, where subjects are asked to draw the face of a clock in a pre-drawn circle and then to draw in the arms to denote 16:45 (a quarter to five). The drawing can then be evaluated by a quantitative scoring method, which is based on the degree of completion of the drawing. The scoring system ranges from 0 to 6 with higher scores reflecting a greater number of errors and more impairment.

Cube drawing: Subjects are asked to draw a cube by heart. In case of failure, a model of a cube is given to the subjects to copy.

The score system ranges from 0 (worse score) to 6 (best score). Score calculation is following: 1 point by face with 4 sides, 2 points for each face where each angle should be respected.

Free and Cued Selective Reminding Test (FCSRT) Assessment of verbal episodic memory. By this test performances in free recalls, cued recalls and in a recognition task can be analysed, because the process of encoding is controlled.

Subjects are asked to remember a list of 16 words. Three tasks of free and cued recalls, as well as one recognition task and one delayed recall give the scores.

Total recall is obtained by the addition of cued recalls to free recalls. Maximum score is 48 for immediate: 16 words X 3 corresponding to immediate free recall + immediate cued recall + immediate recognition test.

Maximum score is 64 (better score) when delayed recall : 16 words X 4 . The minimum score is 0 (worse).

Subjects are asked to memorize two short stories, one consisting of 24, the other one of 26 information units. After the stories have been read aloud by the investigator, subjects are asked to enounce all items of information they can remember (free recall). Correctly reported items are added for each story. This test applies analytical capacities, as well as auditive and verbal synthesis, working memory and episodic memory.

Score range from 0(worst) to 75 (better).

Wechsler Adult Intelligence Scale (WAIS) In this test subjects are asked to say how two seemingly dissimilar items might in fact be similar (14 item couples). This test involves especially abstract thinking and concept capacities.

Cognitive Tests-Age-Adjusted Wechsler Adult Intelligence Scale (WAIS): 19 tests on similarities. The items 1 to 5 are graded from 1 (good) to 0 (bad), and the items 6 to 19 are graded from 2 (good) to 0 (bad). Item 6 and 7 can be repeated. At the end, the worst total score is 0, the best total score is 33.

TMT is a neuropsychological test of visual attention and task switching. The task requires a subject to 'connect-the-dots' of 25 consecutive numbers (1,2,3,etc.) on a sheet of paper or computer screen. The goal of the subject is to finish the test as quickly as possible and the time taken to complete the test is used as the primary performance metric (in seconds). The maximum time allowed is 300 seconds. A negative change score indicates improvement.

First, in the TMT A, the subject has to connect numbers increasingly as fast as possible.The TMT B requires the subject to connect and letters in an alternating pattern (1-A-2-B-3-C, etc.) in as little time as possible.

The Diagnostic and Statistical Manual of Mental Disorders: 4th Edition of the American Psychiatric Association (DSM-IV, 1994) also outlines diagnostic criteria for dementia of the Alzheimer's type that are generally consistent with the NINCDS-ADRDA criteria.

The most widely accepted diagnostic criteria for probable AD are those offered by the National Institute of Neurological and Communicative Disorders and Stroke and by the Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA; McKhann et al., 1984). These criteria include the presence of dementia established by clinical examination and confirmed by neuropsychological testing. The dementia is described as involving multiple, progressive cognitive deficits in older persons in the absence of disturbances of consciousness, presence of psychoactive substances, or any other medical, neurological, or psychiatric conditions that might in and of themselves account for these progressive deficits.

The relationship of an adverse event to the study medication will be classified according to the following criteria:

Related : reports including good reasons and sufficient information (e.g. temporal relationship, dose-response relationship, pharmacology, positive de-challenge and/or re-challenge) to assume a causal relationship with the study drug in the sense that it is plausible, conceivable, or likely

Not related: reports including good reasons and sufficient information (e.g. no temporal relationship and/or attributable to concurrent disease or other drugs) to rule out a causal relationship with the study drug.

Evolution of brain morphology (degree of cortical atrophy) after 18 months with EGb761, using MRI voxel based morphometry

Total of all reporting groups

CDR is a structured interview to collect information regarding subject's memory in a standard way from both the patient and the helper. Scores are calculated using below scale; q CDR=No dementia (score: 0), q CDR=Very mild dementia (score: 0.5), q CDR=Mild dementia (score: 1), q CDR=Moderate dementia (score: 2) and q CDR=Severe dementia (score: 3).

ITT population. N=Number of subjects with assessment.

The Geriatric Depression Scale is a self-administered depression scale, which was developed as a basic screening measure for depression in older adults. By "yes" or "no" answers, scores permit to classify patients into groups of "severely depressed" (score of 21 to 30), "moderately depressed" (score of 11 to 20) and "normal" (score of 0 to10). It takes10 to 15 minutes to administer.

ITT population. N=Number of subjects with assessment.

Subjects completed a verbal fluency test. Higher scores represent higher levels of verbal fluency. Minimum score=0 and Maximum score= N/A.

Letter fluency: this task consists of enouncing as many words as possible that begin with a given letter of the alphabet. Participants are not allowed to use proper names.

Categorical fluency: in this task participants are asked to list as many words as possible that belong to a given semantic category (e.g. animals, fruits, towns) Each condition foresees 60 sec of word generation time. The score corresponds to the number of words correctly given. The verbal fluency task measures semantic storage and executive retrieval functions.

ITT population. N=Number of subjects with assessment.

Following statistical parametric mapping (SPM) analyses were performed by the Commissariat à l'Energie Atomique (CEA):

• The comparison between treatment groups separately for each group of elderly subjects (CNE and MC groups only)
• The comparison between the two groups of elderly subjects (CNE and MC groups only) by treatment group

FDG PET demonstrates reductions in the cerebral glucose metabolism that may occur a few years before the overt clinical manifestation of disease.

SUVBSA2 = [Brain radioactivity (Bq/cc)] / [Injected dose (MBq)/BSA2] x [Blood glucose (g/l)]

BSA2(m^2) = 0.007184 x Height (cm)^0.35 x weight (kg)^0.80

Standardized Uptake Value (SUV) Body Surface Area (BSA)

Intention to treat (ITT) population: All treated CNE and MC patients

MMSE is a brief screening instrument used to assess cognitive function in elderly participants. It assesses orientation, memory, attention, ability to name objects, follow verbal and written commands, write a sentence, and copy figures. Total score ranges from 0 to 30, with a lower score indicating greater disease severity.

ITT population. N=Number of subjects with assessment.

Clock drawing test is a visuo-constructive task, where subjects are asked to draw the face of a clock in a pre-drawn circle and then to draw in the arms to denote 16:45 (a quarter to five). The drawing can then be evaluated by a quantitative scoring method, which is based on the degree of completion of the drawing. The scoring system ranges from 0 to 6 with higher scores reflecting a greater number of errors and more impairment.

ITT population. N=Number of subjects with assessment.

Cube drawing: Subjects are asked to draw a cube by heart. In case of failure, a model of a cube is given to the subjects to copy.

The score system ranges from 0 (worse score) to 6 (best score). Score calculation is following: 1 point by face with 4 sides, 2 points for each face where each angle should be respected.

ITT population. N=Number of subjects with assessment.

Free and Cued Selective Reminding Test (FCSRT) Assessment of verbal episodic memory. By this test performances in free recalls, cued recalls and in a recognition task can be analysed, because the process of encoding is controlled.

Subjects are asked to remember a list of 16 words. Three tasks of free and cued recalls, as well as one recognition task and one delayed recall give the scores.

Total recall is obtained by the addition of cued recalls to free recalls. Maximum score is 48 for immediate: 16 words X 3 corresponding to immediate free recall + immediate cued recall + immediate recognition test.

Maximum score is 64 (better score) when delayed recall : 16 words X 4 . The minimum score is 0 (worse).

ITT population. N=Number of subjects with assessment.

Subjects are asked to memorize two short stories, one consisting of 24, the other one of 26 information units. After the stories have been read aloud by the investigator, subjects are asked to enounce all items of information they can remember (free recall). Correctly reported items are added for each story. This test applies analytical capacities, as well as auditive and verbal synthesis, working memory and episodic memory.

Score range from 0(worst) to 75 (better).

ITT population. N=Number of subjects with assessment.

Wechsler Adult Intelligence Scale (WAIS) In this test subjects are asked to say how two seemingly dissimilar items might in fact be similar (14 item couples). This test involves especially abstract thinking and concept capacities.

Cognitive Tests-Age-Adjusted Wechsler Adult Intelligence Scale (WAIS): 19 tests on similarities. The items 1 to 5 are graded from 1 (good) to 0 (bad), and the items 6 to 19 are graded from 2 (good) to 0 (bad). Item 6 and 7 can be repeated. At the end, the worst total score is 0, the best total score is 33.

ITT population. N=Number of subjects with assessment.

TMT is a neuropsychological test of visual attention and task switching. The task requires a subject to 'connect-the-dots' of 25 consecutive numbers (1,2,3,etc.) on a sheet of paper or computer screen. The goal of the subject is to finish the test as quickly as possible and the time taken to complete the test is used as the primary performance metric (in seconds). The maximum time allowed is 300 seconds. A negative change score indicates improvement.

First, in the TMT A, the subject has to connect numbers increasingly as fast as possible.

The TMT B requires the subject to connect and letters in an alternating pattern (1-A-2-B-3-C, etc.) in as little time as possible.

ITT population. N=Number of subjects with assessment.

The Diagnostic and Statistical Manual of Mental Disorders: 4th Edition of the American Psychiatric Association (DSM-IV, 1994) also outlines diagnostic criteria for dementia of the Alzheimer's type that are generally consistent with the NINCDS-ADRDA criteria.

National Institute of Neurological and Communicative Diseases and Stroke / Alzheimer's Disease and Related Disorders Association (NINCDS/ADRDA)

ITT population. N=Number of subjects with assessment

The most widely accepted diagnostic criteria for probable AD are those offered by the National Institute of Neurological and Communicative Disorders and Stroke and by the Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA; McKhann et al., 1984). These criteria include the presence of dementia established by clinical examination and confirmed by neuropsychological testing. The dementia is described as involving multiple, progressive cognitive deficits in older persons in the absence of disturbances of consciousness, presence of psychoactive substances, or any other medical, neurological, or psychiatric conditions that might in and of themselves account for these progressive deficits.

ITT population. N=Number of subjects with assessment.

CDR is a structured interview to collect information regarding subject's memory in a standard way from both the patient and the helper. Scores are calculated using below scale; q CDR=No dementia (score: 0), q CDR=Very mild dementia (score: 0.5), q CDR=Mild dementia (score: 1), q CDR=Moderate dementia (score: 2) and q CDR=Severe dementia (score: 3)

ITT population. N=Number of subjects with assessment.

The Geriatric Depression Scale is a self-administered depression scale, which was developed as a basic screening measure for depression in older adults. By "yes" or "no" answers, scores permit to classify patients into groups of "severely depressed" (score of 21 to 30), "moderately depressed" (score of 11 to 20) and "normal" (score of 0 to10). It takes10 to 15 minutes to administer.

ITT population. N=Number of subjects with assessment.

Subjects completed a verbal fluency test. Higher scores represent higher levels of verbal fluency. Minimum score=0 and Maximum score= N/A.

Letter fluency: this task consists of enouncing as many words as possible that begin with a given letter of the alphabet. Participants are not allowed to use proper names.

Categorical fluency: in this task participants are asked to list as many words as possible that belong to a given semantic category (e.g. animals, fruits, towns) Each condition foresees 60 sec of word generation time. The score corresponds to the number of words correctly given. The verbal fluency task measures semantic storage and executive retrieval functions.

ITT population. N=Number of subjects with assessment.

MMSE is a brief screening instrument used to assess cognitive function in elderly participants. It assesses orientation, memory, attention, ability to name objects, follow verbal and written commands, write a sentence, and copy figures. Total score ranges from 0 to 30, with a lower score indicating greater disease severity.

ITT population. N=Number of subjects with assessment.

Clock drawing test is a visuo-constructive task, where subjects are asked to draw the face of a clock in a pre-drawn circle and then to draw in the arms to denote 16:45 (a quarter to five). The drawing can then be evaluated by a quantitative scoring method, which is based on the degree of completion of the drawing. The scoring system ranges from 0 to 6 with higher scores reflecting a greater number of errors and more impairment.

ITT population. N=Number of subjects with assessment.

Cube drawing: Subjects are asked to draw a cube by heart. In case of failure, a model of a cube is given to the subjects to copy.

The score system ranges from 0 (worse score) to 6 (best score). Score calculation is following: 1 point by face with 4 sides, 2 points for each face where each angle should be respected.

ITT population. N=Number of subjects with assessment.

Free and Cued Selective Reminding Test (FCSRT) Assessment of verbal episodic memory. By this test performances in free recalls, cued recalls and in a recognition task can be analysed, because the process of encoding is controlled.

Subjects are asked to remember a list of 16 words. Three tasks of free and cued recalls, as well as one recognition task and one delayed recall give the scores.

Total recall is obtained by the addition of cued recalls to free recalls. Maximum score is 48 for immediate: 16 words X 3 corresponding to immediate free recall + immediate cued recall + immediate recognition test.

Maximum score is 64 (better score) when delayed recall : 16 words X 4 . The minimum score is 0 (worse).

ITT population. N=Number of subjects with assessment.

Subjects are asked to memorize two short stories, one consisting of 24, the other one of 26 information units. After the stories have been read aloud by the investigator, subjects are asked to enounce all items of information they can remember (free recall). Correctly reported items are added for each story. This test applies analytical capacities, as well as auditive and verbal synthesis, working memory and episodic memory.

Score range from 0(worst) to 75 (better).

ITT population. N=Number of subjects with assessment.

Wechsler Adult Intelligence Scale (WAIS) In this test subjects are asked to say how two seemingly dissimilar items might in fact be similar (14 item couples). This test involves especially abstract thinking and concept capacities.

Cognitive Tests-Age-Adjusted Wechsler Adult Intelligence Scale (WAIS): 19 tests on similarities. The items 1 to 5 are graded from 1 (good) to 0 (bad), and the items 6 to 19 are graded from 2 (good) to 0 (bad). Item 6 and 7 can be repeated. At the end, the worst total score is 0, the best total score is 33.

ITT population. N=Number of subjects with assessment.

TMT is a neuropsychological test of visual attention and task switching. The task requires a subject to 'connect-the-dots' of 25 consecutive numbers (1,2,3,etc.) on a sheet of paper or computer screen. The goal of the subject is to finish the test as quickly as possible and the time taken to complete the test is used as the primary performance metric (in seconds). The maximum time allowed is 300 seconds. A negative change score indicates improvement.

First, in the TMT A, the subject has to connect numbers increasingly as fast as possible.The TMT B requires the subject to connect and letters in an alternating pattern (1-A-2-B-3-C, etc.) in as little time as possible.

ITT population. N=Number of subjects with assessment.

The Diagnostic and Statistical Manual of Mental Disorders: 4th Edition of the American Psychiatric Association (DSM-IV, 1994) also outlines diagnostic criteria for dementia of the Alzheimer's type that are generally consistent with the NINCDS-ADRDA criteria.

ITT population. N=Number of subjects with assessment.

The most widely accepted diagnostic criteria for probable AD are those offered by the National Institute of Neurological and Communicative Disorders and Stroke and by the Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA; McKhann et al., 1984). These criteria include the presence of dementia established by clinical examination and confirmed by neuropsychological testing. The dementia is described as involving multiple, progressive cognitive deficits in older persons in the absence of disturbances of consciousness, presence of psychoactive substances, or any other medical, neurological, or psychiatric conditions that might in and of themselves account for these progressive deficits.

ITT population. N=Number of subjects with assessment.

The relationship of an adverse event to the study medication will be classified according to the following criteria:

Related : reports including good reasons and sufficient information (e.g. temporal relationship, dose-response relationship, pharmacology, positive de-challenge and/or re-challenge) to assume a causal relationship with the study drug in the sense that it is plausible, conceivable, or likely

Not related: reports including good reasons and sufficient information (e.g. no temporal relationship and/or attributable to concurrent disease or other drugs) to rule out a causal relationship with the study drug.

Safety population. CNE patients withdrew during the double-blind phase and were also not included in the safety population for open phase (17 months).

Evolution of brain morphology (degree of cortical atrophy) after 18 months with EGb761, using MRI voxel based morphometry

ITT population.

Placebo 1 tablet twice a day for 4 weeks (double-blind phase) for CNE, MC, and AD patients

EGb761 120 mg tablet twice a day for 17 months (open phase) for MC and CNE patients