Beyond Description: A Functional GIUS-Based Algorithm for Enteral Feeding Decisions in the ICU

Gastrointestinal dysfunction is common in critically ill patients and frequently compromises the delivery and tolerance of enteral nutrition. Traditional bedside markers such as gastric residual volume or nonspecific abdominal symptoms provide only limited diagnostic accuracy and often fail to capture dynamic alterations in gastrointestinal function. Gastrointestinal ultrasound (GIUS) has emerged as a noninvasive, bedside-applicable method that enables structural and functional assessment of the gastrointestinal tract and may support more individualized nutritional management in intensive care. This narrative review summarizes the physiology and pathophysiology of gastric emptying and intestinal transit in critically ill patients, reviews established GIUS protocols, including Gastrointestinal and Urinary Tract Sonography (GUTS), Acute Gastrointestinal Injury Ultrasound Scoring (AGIUS), the Lai protocol, and the Ultrasound Meal Accommodation Test (UMAT), and proposes pragmatic GIUS-based algorithms for enteral feeding decisions. Three clinical use cases are addressed: 8 h monitoring during ongoing enteral nutrition, preprandial assessment of feeding readiness, and once-daily screening of gastrointestinal function. Current evidence supports the clinical relevance of key sonographic parameters such as gastric antral cross-sectional area and small-bowel diameter, whereas other measures, including mucosal thickness, colonic wall thickness, and Doppler-derived resistive indices, require further validation. UMAT adds a dynamic component to static sonographic assessment and may improve the evaluation of gastric accommodation and emptying in selected patients. Structured GIUS protocols offer a promising, evidence-informed extension of bedside assessment for enteral nutrition management in the intensive care unit. However, the available literature remains heterogeneous and is largely based on physiological studies, observational cohorts, and expert consensus. Prospective multicenter studies are needed to validate cutoff values, training standards, and outcome effects before GIUS-based algorithms can be adopted as stand-alone decision tools.