Percutaneous endoscopic colostomy (PEC) is increasingly proposed as an alternative to surgery to treat various disorders, including acute colonic pseudo-obstruction, chronic intestinal pseudo-obstruction and relapsing sigmoid volvulus. We report on a severe complication that occurred two months after PEC placement. A 74-year-old man with a history of chronic intestinal pseudo-obstruction evolving since 8 years was readmitted to our hospital and received PEC to provide long-standing relief. The procedure was uneventful and greatly improved the patient's quality of life. Two months later, the patient developed acute stercoral peritonitis. At laparotomy, the colostomy flange was embedded in the abdominal wall but no pressure necrosis was found at the level of the colonic wall. This complication was likely related to inadvertent traction of the colostomy tube. Subtotal colectomy with terminal ileostomy was performed. We review the major features of 60 cases of PEC reported to date, including indications and complications.

Background and study aims: Upper gastrointestinal (UGI) bleeding is a frequent cause of hospitalization. Its severity may be assessed before endoscopy using the Glasgow-Blatchford Bleeding Score (GBS), a score validated to identify patients requiring clinical intervention. The aim of this study was to assess whether the GBS was effective for shortening hospital stay and reducing costs in patients with an UGI bleeding predicted at low risk of requiring clinical intervention. Patients and methods: Consecutive outpatients presenting with UGI bleeding at our hospital were prospectively included. In the observational study phase, UGI endoscopy was performed in all patients according to routine clinical practice. In the interventional study phase, patients with a GBS of 0 were discharged with an appointment for an outpatient UGI endoscopy. All patients had follow-up at 7 and 30 days. Need for clinical intervention was defined as performance of endoscopic hemostasis, blood transfusion or surgery. Results Two-hundred and eight patients were included, 104 in each study phase; complete follow-up was obtained in 201 patients. GBS varied from 0 to 18, with 15 (14 %) and 11 (11 %) patients having a GBS of 0 in the observational and interventional study phase, respectively. For patients with a GBS of 0, hospital stay was shorter (6 versus 19 h, P < 0.01), and costs were lower (845 EUR versus 1272 EUR, P = 0.002) in the interventional versus the observational study phase. For patients with a GBS > 0, hospital stay duration did not significantly differ between study phases (189 versus 207 h, P = 0.726). No adverse event was observed in the patients sent home with a GBS of 0 during the interventional study phase. Conclusions Implementing the GBS as a tool for triage of hospital outpatients who present with UGI bleeding allowed us to identify those who could safely be discharged for ambulatory management. Implementing this change in the hospital strategy significantly shortened hospital stay and decreased management costs.