Cost-effective Imaging Approach to the Nonbilious Vomiting Infant

Abstract

Objective. To develop a cost- and time-effective algorithm for differentiating hypertrophic pyloric stenosis (HPS) from other medical causes of emesis in infants referred from community-based pediatricians and family practitioners to the imaging department of a tertiary children's care facility. Methods. Eighty-nine vomiting infants (22 females, 67 males) between the ages of 11 and 120 days (mean, 43.5 days) had received nothing by mouth for at least 1 hour before the study. Each child was assessed for duration of vomiting, status of body weight, time and volume of last ingestion, and time of last emesis. A #8 French (Sherwood Medical, St Louis, MO) nasogastric feeding tube was placed in the child's stomach. The contents were aspirated and measured to determine likelihood of HPS. An aspirated volume ≥5 mL implicated gastric outlet obstruction, and ultrasonography (US) was performed. If this study was positive for HPS, the patient was referred for surgery. If US was negative, an upper gastrointestinal series (UGI) was performed. An aspirated stomach contents volume <5 mL suggested a medical cause for the emesis, and UGI was performed. Pediatric surgeons with no knowledge of the volume results palpated the abdomens of 73 of 89 infants (82%). Results. Twenty-three of 89 patients (25%) had HPS. The aspirate criteria for HPS had a sensitivity of 91%, a specificity of 88%, and an accuracy of 89%. Of the false-positive studies (total = 8), six were related to recent significant ingestion (within 2 hours of the study), and two were attributable to antral dysmotility. The surgeons palpated the mass in 10 of 19 patients (53%). Sensitivity and specificity were 53% and 93%, respectively. Only 6 of 89 infants (7%) required both US and UGI to determine the etiology of the nonbilious vomiting. By performing the UGI in 66 patients, it was also found that 14% had slow gastric emptying and 79% had gastroesophageal reflux. Eighty-one percent of the gastroesophageal reflux was significant. Conclusion. The volumetric method of determining the proper imaging study is cost- and time-effective in the evaluation of the nonbilious vomiting infant for pyloric stenosis. If US was performed initially in all patients referred for imaging, two studies would have been performed in 68 of 89 patients (76%) to define the etiology of the emesis. Because we used the volumetric method, 62 fewer imaging studies were performed, representing a savings of $4464 and 30 hours of physician time. If children are given nothing by mouth for 3 to 4 hours before gastric aspiration, the specificity of the volumetric method improves to 94%, and the accuracy improves to 96%.