Feeding intolerance scoring system in very preterm and very low birth weight infants using clinical and ultrasound findings

Abstract

Very preterm infants are at a high risk of developing feeding intolerance; however, there are no widely accepted definitions of feeding intolerance. This study aimed to develop a scoring system for feeding intolerance in very preterm infants by combining clinical symptoms and ultrasonography (US) findings. This prospective cohort study included very preterm and/or very low birth weight infants. We defined feeding intolerance as the inability to achieve full feeding (150 ml/kg/day) by 14 days of life. The clinical findings included vomiting, abdominal distention, and gastric fluid color. US findings included intestinal peristaltic frequency, gastric residual volume, peak systolic velocity, and the resistive index of the superior mesenteric artery. We conducted multivariate analyses to evaluate the potential predictors and developed a scoring system to predict feeding intolerance. A total of 156 infants fulfilled the eligibility criteria; however, 16 dropped out due to death. The proportion of patients with feeding intolerance was 60 (42.8%). Based on the predictive ability, predictors of feeding intolerance were determined using data from the US at 5–7 days of age. According to multivariate analysis, the final model consisted of 5 predictors: abdominal distention (score 1), hemorrhagic gastric fluid (score 2), intestinal peristaltic movement ≤18x/2 min (score 2), gastric fluid residue >25% (score 2), and resistive index >0.785 (score 2). A score equal to or above 5 indicated an increased risk of feeding intolerance with a positive predictive value of 84.4% (95% confidence interval:73.9–95.0) and a negative predictive value of 76.8% (95% confidence interval:68.4–85.3). The scoring system had good discrimination (area under the receiver operating characteristic curve:0.90) and calibration (p = 0.530) abilities. This study developed an objective, accurate, easy, and safe scoring system for predicting feeding intolerance based on clinical findings, 2D US, and color Doppler US.