Prediction of gastric fluid volume by ultrasonography in infants undergoing general anaesthesia

Background: Point-of-care ultrasonography can estimate gastric contents and volume to assess the risk of pulmonary aspiration; however, its use in infants has not been well validated. We aimed to develop a predictive model for estimating gastric fluid volume using ultrasonography in infants. Methods: This prospective observational study enrolled 200 infants (<= 12 months) undergoing general anaesthesia. After anaesthetic induction, while preserving spontaneous respiration, we measured gastric antral cross-sectional area using ultrasonography in both the supine and right lateral decubitus positions. We then suctioned the gastric content and measured its volume. The primary outcome was development of a gastric fluid volume prediction model with multiple regression analysis. Agreement between the predicted volume and the suctioned volume was evaluated using a Bland-Altman plot. Results: Overall, 192 infants were included in the final analysis. Pearson correlation analysis showed that the gastric antral cross-sectional area in the supine (P<0.001; correlation coefficient: 0.667) and right lateral decubitus (P<0.001; correlation coefficient: 0.845) positions and qualitative antral grade (P<0.001; correlation coefficient: 0.581) correlated with suctioned volume. We developed a predictive model: predicted volume (ml)=-3.7+6.5 x (right lateral decubitus cross-sectional area [cm(2)])-3.9 (supine cross-sectional area [cm(2)])+1.7 x grade (P<0.01). When comparing the predicted volume and suctioned volume, the mean bias was 0.01 ml kg(-1) and the limit of agreement was -0.58 to 0.62 ml kg(-1). Conclusions: Gastric fluid volume can be estimated using a predictive model based on ultrasonography data in infants.