Interoperator reproducibility of ultrasound assessment of gastric residual volume in critically Ill patients

Abstract

Background Gastric residual volume (GRV) is a key metric in assessing feeding tolerance among critically ill patients receiving enteral nutrition (EN). Ultrasound (US) has emerged as a noninvasive, bedside tool for GRV estimation. However, its interoperate reproducibility remains insufficiently validated. Objectives This study aimed to evaluate the reproducibility of GRV measurement using US among intensivists and to compare these measurements with traditional nasogastric aspiration. Patients and methods A prospective observational study was conducted at Alexandria Main University Hospitals on critically ill patients receiving EN. GRV was assessed by US in the right lateral decubitus position using a standardized protocol, with two attending intensivists performing measurements after specific training. The reproducibility was analyzed using Pearson’s correlation, Lin’s concordance correlation coefficient, and Bland-Altman plots. GRV measured by US was compared with nasogastric aspirate values to assess agreement and proportional bias. Results The study enrolled 60 patients (57% male, median age 52 years). The Pearson correlation coefficient (r) for US measurements among intensivists was 0.992, with a concordance correlation coefficient (ρc) of 0.992. Bland-Altman analysis revealed minimal bias (−1.5 ml; 95% CI: −3.7–0.7) and narrow limits of agreement, indicating excellent reproducibility. Comparison between US and nasogastric aspirate GRV measurements showed a high correlation (r=0.973, ρc = 0.948) with a mean positive bias of 21 ml (95% CI: 15–27). US demonstrated proportional bias at higher GRVs, with values diverging from nasogastric aspirate. Conclusion US is a highly reproducible tool for GRV estimation among trained intensivists, with results closely aligning with nasogastric aspiration. Despite minor proportional biases at higher GRVs, US offers a noninvasive and efficient alternative for assessing feeding tolerance in critically ill patients. Further studies may explore its role in guiding clinical decision-making and optimizing EN strategies. Trial registration: 0107755