Recent studies suggest the employment of 'conscious' sedation (1) for critically high - risk patients (2), showing more efficacy then deep sedation (3). The investigators want to compare intravenous injection versus enteral sedative drugs administration, purposing to maintain a 'conscious' sedation level compatibly with the needed cares, invasive procedures, and medical and nursing surveillance.

Recent studies suggest the employment of 'conscious' sedation for critically high - risk patients (2), showing more efficacy then deep sedation (3). From this point of view it's important to consider both the choice of sedative drugs and the way to administer them to maintain the constant 'conscious' sedation level that the investigators wanted.

This research compares two different procedures for sedation therapy: through a randomized and blind controlled trial the interventional group will collect patients receiving sedatives from enteral way (4, 5, 6). In the control arm, sedation administration forecasts intravenous injection (7), recognised like the best practice in the most of intensive care units of the world.

To consider two different clinical approach for administration the investigators have to use different molecules: in the control arm propofol and midazolam will be continuously administered through intravenous injection with daily suspension (7); in the intervention arm the investigators will administer melatonin like a physiological hypnotic inductor (6), the continuous 'conscious' sedation will be maintained through hydroxyzine and eventually lorazepam (4), and all the active principles will be given by enteral administration.

The goal of the study is to compare intravenous injection versus enteral sedative drugs administration, analysing their efficacy and the feasibility to maintain constant the appropriate sedation level (RASS measured = RASS wished ±1) (8) in high - risk patients admitted in Intensive Care Unit (2), purposing to conserve a 'conscious' sedation level compatibly with the needed cares, invasive procedures, and medical and nursing surveillance.

MATERIALS AND METHODS Prospective, randomized and controlled multicentric, single blind trial. Participant ICU centres: San Paolo MI (Iapichino), Policlinico MI (Gattinoni) , Fatebenefratelli MI (Cigada), Niguarda MI (De Gasperi), Desio (Ronzoni), Legnano (Radrizzani), Monza (Pesenti), Belluno (Mazzon), Lugano (Malacrida), Modena (Rambaldi), Torino (Livigni), Asti (Cardellino).

PROCEDURE Bare minimum sedation drug titration will be done to maintain and to achieve prematurely a 'conscious' sedation level (RASS=0). During every shift it will be discussed the appropriate sedation therapy: if physicians choose a deep sedation goal (RASS <-3) this decision has to be explained and registered; then enteral sedation has to be maximized in randomized enteral patients group. If necessary intravenous sedation has to be added , and this procedure dose not represent a violation of research protocol.

If there is pain (VNR>3 or BPS>6) it will be administrated analgesic therapy according to hospital guidelines bare minimum duration. The treatment of procedural pain will be applied both groups trough Fentanest/Morphine + propofol/midazolam bolus intravenous, this procedure dose not represent a violation of research protocol.

If acute cerebral malfunction appears (CAM-ICU positive), it will be administrated haloperidol (1mg per os, max 10 mg/die) or other antipsychotic therapy according to hospital guidelines.

Enteral artificial nutrition with prokinetics will be started as soon as possible both group, while parenteral nutrition will be administered only if strictly necessary. If gastric stagnation > 200 ml/4 hours exceeds 2 days duration, it could be positioned nasogastric tube or digiunostomy.

All patients will be sit in a semiortopnoic position (bed back rest inclination between 30 and 45°).

STUDY POWER AND STATISTICAL ANALYSIS The investigators suppose to obtain a difference of 15% per cent between two arms for patients' sedation adequacy benefit (RASS = desired RASS + 1). Knowing that the enteral approach reached 83% adequacy from a observational monocentric research (5), it should be necessary to enrol 141 patients for each arm (power 80%). In consideration of the missing data, the investigators expect to enroll 300 patients. An "ad interim" statistical analysis is planned after the enrollment of 70 patients in each group.

To provide a statistical analysis as "Intention To Treat" for possible arms change risk.

It's planned to obtain a selective analysis involving each hospital included in the study.

It's planned to obtain a selective analysis about septic patients (greater delirium prevalence) (9).

It's planned to obtain a selective analysis divided for age (greater delirium prevalence if age >70) (10).

The randomization will be achieved trough out a particular Internet Website with a specific program expressly built. It will be used minimization technique to maintain groups balanced in the patients' sample of each centre.

Inclusion Criteria:

• High Risk Patients (Ventilation days assessment >3, SAPS II >32).
• Until 24 h after ICU admission
• Age > 18 years

Exclusion Criteria:

• Neurosurgical patients
• Allergy to medications used in the study
• CNS diseases (epilepsy, ictus, dementia, anoxic coma…)
• Liver encephalopathy (Child C)
• Previous psychiatric or cognitive pathology
• Absolute contraindications to use enteral route (acceptable NGT, digiunostomy, ileostomy)
• Pregnant patients or in breast-feeding
• DNR patients