Malnutrition is very common in critically ill patients. It is quite important to evaluate nutritional status precisely. Heyland et al firstly reported NUTRIC score including age, APACHE II score, SOFA score, number of commorbidities, days from hospital to ICU admission and IL-6. Because the IL-6 is not routinely checked at ICU. A modified NUTRIC score without IL-6 is more practical. Previous studies showed lower in-hospital mortality in higher nutritional risk patients with higher caloric intake compared with lower caloric intake. However, there is still controversial regarding the in-hospital mortality between full caloric feeding and permissive underfeeding in critically ill patients. Herein the investigators conduct a study to investigate what kinds of nutritional supplements will decrease in-hospital mortality in different nutritional risk patients.

Inclusion Criteria:

• 1. Medical intensive care unit patients 2. age elder than 20 years old 3. Intubated with ventilator support patients 4. Predicted ICU stay longer than 72 hours

Exclusion Criteria:

• 1. NPO patients 2. Contraindication for enteral feeding 3. TPN use 4. Upper or lower GI bleeding patients 5. Feeding with gastrostomy or Jejunostomy patients 6. Metoclopramide related EPS or Torsades de pointes history