Emergency physician ultrasound diagnosis of pneumoperitoneum in intraoperative patients with peritoneal insufflation

Abstract

Objectives

To determine the accuracy of emergency physicians (EPs) in diagnosing pneumoperitoneum with POCUS, and if the volume of pneumoperitoneum affects accuracy.

Methods

POCUS clips were obtained from patients undergoing intraperitoneal insufflation for an elective laparoscopic procedure. Video clips of the right upper quadrant and epigastric regions were obtained prior to insufflation and then after 500 ml, 1000 ml, and 1500 ml of insufflation. These clips were randomized and reviewed by three blinded ultrasound-trained EPs. For each clip they determined whether pneumoperitoneum was present or not.

Results

EPs chose correctly 71 % of the time. Overall sensitivity for detecting pneumoperitoneum was 66 % with a specificity of 85 %. Sensitivity for detecting small, medium, and large volumes of air was 53 %, 70 %, and 73 % respectively. The AUC for pneumoperitoneum overall was 0.753. The AUC for detecting small, medium, and large volumes of air was 0.688, 0.773, and 0.789 respectively. There was substantial agreement between EPs with a kappa of 0.658.

Conclusions

POCUS for pneumoperitoneum had moderate sensitivity, high specificity and moderate accuracy. The sensitivities increased with increasing volumes of air. This data suggests that POCUS has similar sensitivity to CXR for pneumoperitoneum.