The wide-acting effects of Fasting-Mimicking Diets (FMDs) on metabolic, inflammatory and regenerative pathways leading to reduced pathology or risk factors for various diseases in mice and humans, has the potential to be effective against Alzheimer's disease (AD). It is proposed to conduct a randomized clinical trial of twelve monthly cycles of the ProlonADTM diet (by L-Nutra) vs. placebo diet in patients with aMCI or mild AD (MMSE 18-23). The primary endpoint of the study will be the feasibility and safety of the twelve cycles of ProlonADTM.

Cognitive impairment that exceeds the physiological decline associated with aging can take the form of mild cognitive impairment (MCI) or, in its most severe form, dementia. In turn, patients diagnosed with MCI evolve towards outright Alzheimer's disease (AD) in an estimated percentage of 15%/year. MCI can occur in the form of amnestic MCI (amnestic MCI, aMCI; by far the most common) or non-amnestic MCI (single or multiple domain), where amnestic is the form that most frequently evolves into AD. Therefore this clinical trial will focus on the latter (as well as on the AD). Diet cycles with low sugar and protein levels followed by diets with normal levels of these lead to temporary reductions in growth hormone and IGF-1 levels. Both are potential mediators of the neuroprotective and regenerative effects of these diets not only in mice, but also in monkeys and humans . However, heavily restricted diets in terms of calories are often difficult to endure over time and are frequently associated with side effects, even significant ones and with progressive weight loss, in particular of lean mass. In a mouse model of AD, it has been shown that periodic cycles of a "fasting-mimicking" diet (FMD) restricted in protein content (PR-FMD) but not in terms of calories are able to reduce levels plasma levels of IGF-1 with contrasting effects on the neurodegeneration process. In particular, such FMD has been shown to reduce by about 30-70%, the levels of hyperphosphorylated tau protein (one of the typical markers of AD) at the hippocampal level, reducing the age-related deficit of cognitive performance. A relevant neuroregenerative effect (associated with a clinical improvement in motor coordination and memory) has been demonstrated in mice subjected to a diet based on a similar DMD during their "average life" (months 16-30).

It is proposed to conduct a study of twelve monthly cycles of the ProlonADTM diet (by L-Nutra) in patients with aMCI or mild AD (MMSE 18-23) diagnosed according to the criteria defined by Peterson and McKahn, respectively (1, 3 ).

It is proposed to conduct a randomized clinical trial of twelve monthly cycles of the ProlonADTM diet (by L-Nutra) vs. placebo diet in patients with aMCI or mild AD (MMSE 18-23). Patients in the treatment group alone will also receive a range of supplements containing omega-3 fatty acids, caffeine, tree nuts, coconut oil, olive oil and cocoa - which will be supplied with the ProlonADTM diet kit. Patients assigned to both arms will also receive personalized dietary recommendations matched to instructions for light-moderate physical activity to be carried out also at home and aimed, especially in the case of patients assigned to receive ProlonADTM, to prevent the loss of lean mass. The primary endpoint of the study will be the feasibility and safety of the twelve cycles of ProlonADTM. Feasibility is defined as taking at least one course of DMD every two months with the option of admitting consumption of 50% of the planned diet and / or a maximum consumption of 10 Kcal / kg of unforeseen food in only one of days 1-5 of each cycle. Investigators speculate that ProlonADTM will not cause severe side effects and that it will have no detrimental effect on the patient's body composition, specifically in terms of impact on lean mass measured by dynamometry and bioimpedance analysis.

Secondary objectives will include:

• conversion rate to AD (for patients with aMCI);
• episodic memory evaluated with Free and Cued Selective Reminding Test (FCRST);
• cognitive status assessed by MMSE, CDR-Sum of the boxes;
• functional status assessed with Barthel Index, IADL;
• emotional state assessed using the CESD-R scale;
• nutritional status (MNA and body composition - bioimpedance, handgrip);
• caregiver stress assessed through Caregiver Burden Inventory and NPI;
• quality of life of patients (QLQ-AD);
• inflammatory markers, oxidative stress markers, neuronal damage markers (Neurofilament Light, NfL), quantification of circulating stem cells, cell aging markers (eg evaluation of the telomerase activity of lymphocytes).
• assessment of frailty (Rockwood frailty index, which also includes walking speed, handgrip, breathing capacity)

Inclusion Criteria:

• Age 55-80;
• Presence of aMCI or initial AD (MMSE 18-23);
• Normal organ function (liver and kidney);
• BMI not less than 20 kg/m2;
• Bioimpedance phase angle (PA) > 5 °;
• Adherence to informed consent

Exclusion Criteria:

• Age> 80 years
• Diabetes mellitus;
• Organ impairment (liver, kidney);
• Food allergies to the components of ProlonADTM;
• Patients on therapy with vitamin K antagonist anticoagulants;
• PA <5 °;
• Patients who live alone or are not adequately supported by the family context;
• Other experimental therapies in progress.