“Flush sign” is a possible predictor of adhesions in EUS-guided transmural drainage of pancreatic fluid collections

Background and aims: Endoscopic ultrasound (EUS)-guided transmural drainage is an alternative to surgical or percutaneous drainage of inflammatory pancreatic fluid collections (PFCs). It is important, but sometimes difficult, to predict adhesions between PFCs and the digestive tract wall in EUS-guided transmural drainage. The aims of this study were to compare perforation and non-perforation cases in EUS-guided transmural drainage, and to establish methods for predicting such adhesions.

Patients and Methods: From May 2007 to March 2010, a total of 20 consecutive patients underwent EUS-guided transmural drainage of PFCs in our hospital. The best site for drainage was identified by EUS, and a 19-gauge needle was used to puncture the PFC cavity under EUS guidance. A guidewire was then advanced through the needle. An opening for cystogastrostomy was created using a wire-guided triple-lumen needle knife with diathermy. The opening was dilated with a biliary dilatation balloon catheter. A double pigtail stent and/or a pigtail type nasobiliary drainage tube were placed over the wire. CT was performed immediately after the procedure. We reviewed endoscopic movies or images recorded during the EUS-guided transmural drainage. One patient was excluded for lack of sufficient movies and images. We focused on whether the cyst fluid flushed into the stomach after creation of the opening for cystogastrostomy using the wire-guided triple-lumen needle knife. We named this the “flush sign”.

Results: Perforation occurred in two patients (10%). No tight adhesions between the gastric and cyst walls were revealed by surgical findings in these two patients. The flush sign was positive in 16 patients and negative in three patients. Among the three flush sign-negative patients, two were perforation cases and one had a pancreatico-colonic fistula and the PFC collapsed before EUS-guided transmural drainage. The occurrence of perforation was significantly higher in the flush sign-negative patients than in positive patients (p<0.05).

Conclusions: The flush sign may be useful for predicting adhesions between the gastric and cyst walls. When flush sign-negative patients are encountered, we must stop using a biliary dilatation balloon and place a single nasobiliary drain. Consequently, we may prevent unnecessary surgery.