The etiology of Alzheimer's disease (AD), the most common type of dementia, remains largely undefined and the early diagnostic and effective treatments are still not available. In addition to the neuropathological hallmarks, cerebrovascular dysfunction has been identified as an important component of AD. Using the experimental models, we showed that cerebrovascular reactivity (CVR), the ability of cerebral vessels to dilate or constrict in response to stimuli, is impaired very early in AD. We designed this trial to compare CVR to carbon dioxide (CVR CO2) in AD patients and in persons with subjective cognitive impairment (SCI), the cognitively healthy individuals which began to worry about worsening their memory, and to correlate CVR CO2 with AD markers in cerebrospinal fluid and the blood markers of endothelial function. We hypothesize that CVR represents a potential diagnostic/prognostic marker and an attractive target for the development of new therapeutics in AD.

The primary objective of this study is to determine, whether the measurement of CVR CO2 makes it possible to differentiate AD patients from control SCI subjects (subjects with a memory complaint but without cognitive impairment as assessed with neuropsychological tests).

We also pursue the following secondary objectives:

• Establish correlations between cognitive level and CVR CO2 in the 2 groups: SCI and AD.
• Investigate a correlation between CVR CO2 and AD markers in cerebrospinal fluid
• Assess the correlation between cognitive function (Z score of neuropsychological tests) and CVR CO2.
• Look for a correlation between CVR CO2 and plasma levels of markers of endothelial function: endothelin-1 (ET-1), big-ET-1, asymmetric dimethylarginine (ADMA) as well as plasma renin activity.

We plan to enroll a total of 24 male and/or female patients (> 18 years) with a diagnosis of AD (group AD) and 24 male and/or female SCI subjects (> 18 years) which came to our hospital for consultation about poor memory function.

Each patient will be informed on the study protocol during his / her regular follow-up visit at Broca hospital's day clinic. Informed consent/assent will be obtained from each subject or his/her representative prior to conducting any study related procedures.

Each study participant will make a total of two visits. The first (baseline) visit will be made to the Geriatric Department of the Broca hospital. At baseline visit, a complete medical history, physical examination, neurological examination, as well as the MRI and lumbar puncture data, if available, will be collected. A blood sample of 10 ml will be drawn in addition to the blood sample taken as part of patient usual care. A medical appointment will be made at the Clinical Physiology Department of the Lariboisière Hospital for CVR CO2 measurement (visit 2).

The second and the final visit will take place approximately one month after the baseline visit. The CVR CO2 will be assessed using Doppler ultrasonography, as a change in blood flow in the internal and common carotid arteries during the 10th minute of inhalation of a gas mixture containing 5% CO2, 16% O2 and 79% N2, compared to the baseline blood flow value measured after 10 minutes rest in lying position. Ten minutes after the end of the inhalation of the hypercapnic mixture, patients will receive a sublingual administration of a nitroglycerin spray (NTG 300 μg, Procter and Gamble Pharmaceuticals) and the carotid blood flow will be recorded between the 3rd and the 5th minute after NTG administration. In this way, the maximal blood flow obtained after endothelial-independent vasodilatation will be evaluated.

Changes in carotid blood flow (before and during hypercapnia) will be compared by a covariance analysis (variables known to interfere with CVR will be included as covariates). The analyses will be conducted considering the CVR CO2 variable in absolute values and relative values.

A multiple regression model will be used to test the hypothesis of a relationship between CVR CO2 and cognitive level (Z score of neuropsychological tests), CVR CO2 and AD markers in cerebrospinal fluid, and CVR CO2 and the plasma concentration of ET-1, bigET-1, ADMA and plasma renin activity in the two groups: SCI and AD. Variables known to be cognitively related and known to be CVR related will be included in the model. The slopes of the regressions between CVR CO2 and the different parameters of interest will be compared between the two groups. Models taking into account each biomarker of interest separately and then a global model will be constructed.

Inclusion Criteria:

• Age 18 years and more
• Alzheimer's disease as defined by the established clinical consensus criteria (DSM IV-TR and NINCDS-ADRDA) or
• Subjective cognitive impairment: subjects who consulted for cognitive complaint without cognitive impairment on neuropsychological tests defined by normal cognitive scores located between - 1.5 and + 1.5 standard deviations (σ) compared to the average of subjects in their reference group (same age, same socio-educational level) and normal performances according to daily life activities scores.

Exclusion Criteria:

• Non-Alzheimer's dementia (vascular dementia, Lewy body dementia, fronto-temporal dementia, dementia linked to Parkinson's disease and other)
• Other diseases that may interfere with cognitive performance evaluation (severe depression, epilepsy, Parkinson's disease, psychosis, bipolar syndrome)
• Active smoking
• Diabetes
• Uncontrolled hypertension (PAS / PAD> 140/90 mmHg despite correctly dosed antihypertensive therapy)
• Heart failure
• Recent myocardial infarction
• Stroke
• Stenosis of the carotid commune or interne with plaque > 20%
• Cerebrovascular lesions on MRI (major white matter lesions Fazekas 3,> 2 deficiencies and sequelae of cerebral infarction)

1. Patients with Alzheimer's disease
2. Patients with subjective cognitive impairment

There is not a plan to make IPD available.