This is a prospective multicenter cohort study with a one-year follow-up. Tests were carried out at two types of medical organizations: HEC and MC. HEC are structures belonging to the French health insurance system which offers check-ups and screening for the prevention of health conditions, health education and actions, and chronic disease support. This is a free check-up and priority is given to frail or vulnerable people. MC (as defined by the circular issued by the Direction Générale de l’Offre de Soins (DHOS) (general directorate for healthcare provision) on 16^th April 2002) offer consultations and day hospitalizations which are dedicated to medical diagnostics, follow-up and nursing (treatments both with and without the administration of drugs). MC are responsible for research on AD and related disorders, participate in academic training and coordinate an interregional plan for local memory consultations.

The three HEC which participated in the EVATEM study are located in three French cities: Lyon Angers and Saint-Etienne. The diagnostics were performed through three MC in the academic hospitals of these three cities. The study is composed of several steps (Figure 1). Patients with a memory impairment were identified at HEC using questionnaires and cognitive tests during periodic health examination at T0. Memory consultations to diagnose MCI and AD and related disorders were carried out at MC according to a standardized neuropsychological evaluation and complementary exam at T0. A one-year follow-up exam, at T1, was scheduled for each of the patients.

All EVATEM subjects gave their informed consent to participate in this study. Their participation or refusal had no effect on the usual nursing performed at the HEC. The study protocol has been reviewed and approved by an ethics committee CPP Sud-Est III (comité de protection des personnes Sud-Est III (Persons protection comity of South-East III)) on the 7^th of July 2009. All procedures are in compliance with the Helsinki declaration. Data processing is approved by the CNIL (commission nationale de l’informatique et des libertés (National Commission for Informatics and Liberties)) under n°1362516.

EVATEM study has been registered in the clinical trials (Current Controlled Trials NCT01316562 NCT01316562).

Subjects were selected during free periodic health examinations in one of the three health examinations centers that participated in this study. These HEC were located to the cities of Lyon, Angers and Saint-Etienne.

Patients could be included if they were 65 years of age or older, lived at home or in a non-medical institution, volunteered to participate to a periodic health examination and if a memory impairment concern was identified (either by the patient themselves or by a member of their family).

Patients were excluded from the study if they had dementia, an inadequate level in French language ability, to a sensory deficiency which would compromise their evaluation.

We calculated the required sample size need to detect an expected sensitivity and specificity of 0.9 with the cognitive tests used in the HEC.

We set the minimum acceptable lower 95% confidence limit at 0.75, with a 0.95 probability.

Using the table of Flahault et al. 70 confirmed cases were required 19 . The number of controls required was calculated to be 513 with the equation: N controls = N cases [(1-P)/p] 19 ; where p was the expected prevalence of MCI or AD and related disorders (we supposed this to be 12% 13 )_. The total target sample size was therefore 583.

We predicted the loss of subjects between examinations in the HEC and the MC to be 30% and so aimed to recruit 833 subjects. Actual recruitment was consistent with these estimates.

1. Senior health periodic examination at HEC

a. Socio-demographic data

Age, sex, marital status, address, economic affluence and vulnerability level (evaluated with the EPICES score, the individual index of deprivation used at the HEC 20 ) were recorded.

a. Clinical data

We recorded weight, height, the number of different daily therapeutic groups attended, psychotropic dose (established by asking the patient or reading their last prescription), autonomy (evaluated by testing four activities of daily life 21 ), and mood (evaluated using the Mini-Geriatric Depression Scale (mini-GDS) 22 ).

a. Para-clinical data

Data recorded included long sighted visual acuteness (evaluated using the Monoyer scale), an evaluation of balance and walking (using tests such as the Timed Up and Go (TUG) 23 and the imagined TUG), patient proprioception sensitivity (measured using the diapason test), and the results of the handgrip test, unipodal position evaluation, chair rising, temporal analysis of walking and posture analysis (at the HEC of Lyon and Angers).

a. Memory complaint

A self-questionnaire composed of binary questions (Q1) was used to assess memory at the HEC Angers and Saint-Etienne. A questionnaire (Q2) was carried out by the doctor at the HEC of Lyon (Table 1). At least one positive answer to the questionnaire was required for the subject to be included in the study.

a. Cognitive testing performed

The EVATEM study assessed Dubois’ five-word test 24 , the cognitive disorders examination test (CODEX) 25 26 and the verbal fluency test 27 . These tests were chosen as they assess verbal memory, temporal and spatial orientation and executive function but are straightforward and rapid.

Dubois’ five-word test is an episodic memory test where subjects are asked to learn five words belonging to different semantic groups and lacking phonetic similarity (for example: beverage, cooking tool, vehicular, building and insect) with control of the encoding. The test is built in such a way that it enables differentiation between symptoms of hippocampus damage (associated with AD) and common memory impairment; it includes immediate free recall exercises and postponed free recall activities. In the case of recall difficulty, a cue is suggested.

The CODEX test is designed to detect dementia in elderly subjects. It is a very simple and quick test (taking less than three minutes) and can be performed by doctors who are not specialized in the evaluation of memory or other cognitive functions. This test consists of two stages: the three-word test and simplified clock test 28 are carried out before the MMSE (five questions of spatial orientation). In the EVATEM study, the three-word test, the simplified clock test and the five spatial orientation questions were all performed.

The test of verbal fluency measures the capacity of a subject to generate a list of words belonging to the same semantic category (animals), and a list of words starting with the same letter (P), within a time limit of two minutes. This test assesses semantic memory and executive memory functions. In the case of a neurodegenerative disease, particularly in AD, semantic fluency seems to be impaired and this may be associated with possible damage to the semantic memory stock 29 .

2. Neuropsychological evaluation performed at MC

The cognitive tests described below were performed routinely in MC. Episodic memory was assessed using the following tests: a free and cued recall test with 16 items (test adapted from Grober and Buschke 30 ), the DMS48 31 , the Signoret battery for mnesic efficiency (BEM 144) 32 and the Birmingham Object Recognition Battery (BORB) 33 . Executive abilities were assessed using the Frontal Assessment Battery (FAB) 34 , Baddeley’s dual task test 35 , the Wechsler Adult Intelligence Scale III (WAIS III) 36 37 , the Trail Making Test (TMT) 38 , the STROOP test (GREFEX version) 39 and semantic verbal and literal fluency tests (using ‘P’ and ‘Animals’; a two minute test) 27 ^. Instrumental efficiency was evaluated by gestural praxis assessment (symbolic gestures, pantomimes and abstract gestures), the Rey complex figure 40 , the picture naming test (DO 80) 41 and the Visual Object and Space Perception Battery (VOSP) 42 .

These tests were conducted by a neuropsychologist and lasted approximately 1.25 h.

The chosen definition of AD was based on the criteria of the National Institute of Neurological and Communicative Diseases and Stroke and the AD and Related Disorders Association (NINCDS-ADRDA) 43 ; this can diagnose AD as possible, probable or certain. The severity of dementia was based on the MMSE score 8 . Subjects were categorized as being in the mild stage if their score was between 25 and 21, in moderate stage if their score was between 20 and 10 and in the severe stage if their score was below 10.

The diagnostic criteria for MA at a stage of dementia were the clinical criteria for probable AD 43 44 .

At the stage of pre-dementia (corresponding to the amnestic MCI diagnosis 15 16 17 45 ), diagnostic criteria were the clinical criteria of the Pre-Al study and the paraclinical criteria recently proposed for research 46 47 . For other disorders frontotemporal lobar degeneration 48 , lewy body dementia 49 and vascular dementia 50 criteria were those referenced.

1. Baseline assessment: cognitive tests at HEC

The cognitive assessments were performed by trained physician and nurses. A guide, detailing how to perform the cognitive tests, specified the order in which the tests should be performed. Immediate recall and cued free test of five words was performed by the doctor, then an intercurrent test of attention for about 10 minutes and finally a free and cued delayed recall test of five words. The attention of the subject was intentionally diverted using several tests. For example, drawing pentagons as part of the MMSE and a visual-construction task in the Cognitive Assessment Battery (CAB 96) 51 . Time orientation questions (‘What day of the week is it?’; ‘What month is it?’; ‘What year is it?’), questions about drug intake (the number of therapeutic classes attended, the use of a psychotropic drug and treatment for osteoporosis), the TUG 23 and the imagined TUG were also used. The Instrumental Activities of Daily Living (IADL) scale (derived from the IADL by Lawton and Brody 21 ), was used to evaluate functional capacity. This test explores the subject’s degree of autonomy or dependence in relation to four practical activities of daily life (telephone use, transportation, responsible administration of medication and budget management). The three-word test, the simplified clock test, the five questions about spatial orientation and verbal fluency tests were then performed by the nurse.

2. Baseline assessment: memory consultation at MC

An appointment for consultation at the MC was made at the end of the periodic health examination. If this had not been possible, the patient was contacted by a secretary of the HEC to schedule the appointment. The standardized biological check-up performed at HEC included a neuropsychological assessment, screening for syphilis, serum electrolytes determination and serum assays of vitamin B12, vitamin B9, albumin, pre-albumin and Thyroid Stimulating Hormone (TSH). Depression screening was performed using the Geriatric Depression Scale (GDS); a 15-item scale of depressive symptomatology suitable for elderly patients 52 .

Check-ups at the University Hospital of Lyon were supplemented by the Hamilton Anxiety Scale, a questionnaire on therapeutic compliance and nutritional assessment. The established diagnosis, according to the criteria specified above, was sent to the doctor working at the HEC so that they could classify patients as “normal” subjects, subjects with MCI or subjects with dementia. A MRI scan was systematically performed at the University Hospital of Angers. At the University Hospitals of Lyon and Saint-Etienne, MRI scans were reserved for subjects where anomalies were detected during the examination.

3. Follow-up at MC

Subjects of the study will be re-called one year later (T1) at MC for a second neuropsychological check-up using the same protocol as at T0.

Collection and data entry are managed by the HEC of Lyon; collected data is entered by two secretaries of the HEC who have been trained in data management (double input). The collection tool was developed by the Regional Medical Service Officer (RMSO) of the Rhône-Alpes region.

Patients will be recorded as belonging to one of three groups: “normal”, MCI or dementia subjects. The main outcome measure will be the diagnosis of MCI or AD and related disorders at T0 and T1. The secondary outcome measure will be falls at T0 and T1. A fall was defined as an event which results in a person coming to rest unintentionally on the ground, floor or at a lower level.

Baseline characteristics of the subjects (for example, age and sex) will be described as above. Continuous variables will be summarized by the mean and standard deviation or the median and interquartile range (as appropriate). Categorical variables will be summarized as proportions of each category.

Socio-demographic and clinical comparisons will be performed between three groups (normal, MCI and dementia subjects). Group proportions will be compared using Pearson’s chi-squared test. The means of these groups will be compared using an Analysis of Variance (ANOVA) with parametric and nonparametric tests, depending on the distribution of the variables. We will study the relationship between the result of each cognitive test performed at HEC and the diagnosis confirmed by the memory consultation. A Receiver Operator Characteristic (ROC) curve will be generated for each cognitive test, using MCI and AD and related disorders diagnosis as references. The optimum threshold, providing the best compromise between sensitivity and specificity, will therefore be determined. Predictive and negative values, as well as positive and negative likelihood ratios, will be calculated. We will use a binary or multinomial logistic regression to determine which test, or combination of cognitive tests, are the most predictive for MCI and AD and related disorders diagnosis. Models will be adjusted for potential confounding variables.

A p value less than 0.05 will be considered as statistically significant.

Statistical analyses will be performed using SPSS (Statistical Package for the Social Sciences) version 17.0 for Windows (SPSS Inc., Chicago, Illinois, USA).

Inclusions at HEC took place from 15th November 2009 to 31st December 2011 (Figure 2). The one-year follow-up will end during the first semester of 2013.

A total of 4571 elderly subjects aged 65 years or over were eligible for research into the diagnosis of memory complaints. Amongst those subjects, 57.5% were recruited by the HEC of Lyon, 34.2% by the HEC of Angers and 8.3% by the HEC of Saint-Etienne.

The percentage of subjects with memory complaints, detected using the self-administered questionnaire (Q1), was 70% for the HEC of Angers and 59.4% for the HEC of Saint-Etienne. At the HEC of Lyon, where memory complaints were screened using a questionnaire administered by doctors (Q2), the percentage of complaints was 27.5%. Cognitive tests were refused by 33.6% of people with memory complaints. Amongst the 832 subjects for whom an appointment at MC had been made, 74.5% attended a consultation. Thus, the final study population included 620 subjects.