PTH-035 Coffee ground vomit:does it justify an urgent endoscopy?

Abstract

Introduction Coffee ground vomit is vomit that looks subjectively like coffee grounds. It is thought to occur due to the presence of coagulated blood in the vomit and hence is a common indication for inpatient admission and thereafter endoscopy. In an increasingly stretched inpatient endoscopy service it is important not to over burden it with endoscopies that could be performed safely as an outpatient. Therefore, our aim is to evaluate the need for inpatient gastroscopy in patients who are deemed to have coffee ground vomiting. We hypothesise that patients with coffee ground vomiting do not have significant upper gastro intestinal bleeding requiring endoscopic intervention. Methods A single centre, retrospective analysis was performed on patients endoscopied for the primary indication of coffee ground vomiting. Data was collected and scrutinised from the Electronic Patient Records (EPR) and Unisoft endoscopy-reporting tool at Barnet and Chase Farm Hospitals, Royal Free London for 12 months of 2017. Gastroscopy reports were studied to see whether endoscopic therapy was required (defined as use of adrenaline injection, banding, clips, haemospray or gold probe). EPR was subsequently used to assess whether these patients had a significant drop in their haemoglobin (Hb) defined as a Hb drop ≤20 g/dl. Two independent researchers carried this out. Results There were 2618 gastroscopies during the study period. Of these, 37 were indicated due to coffee ground vomiting with 29 being performed as an inpatient. Of these 29 patients, 27 (93%) had a significant drop in their Hb level prior to gastroscopy. One (3%) patient required endoscopic therapy. This patient had significant co-morbidty of ischaemic heart disease, hypertension, aortic valve replacement as well as a drop in Hb. In total, 12 patients had a diagnosis of oesophagitis, 3 had erosive gastritis, 3 non erosive gastritis, 1 oesophageal ulcer, 3 duodenal ulcers, 3 non erosive duodenitis, 1 pyloric ulcer and 1 hiatus hernia. 4 gastroscopies were completely normal. There were no patients with cancer diagnosis. Each diagnosis was reported separately if the report contained more than one diagnosis. Conclusions From this study we conclude that in the majority of patients endoscoped for coffee ground vomit do not require intervention during endoscopy. This study confirms our hypothesis and adds weight to the notion that patients with coffee ground vomiting do not necessarily require inpatient gastroscopy despite a significant Hb drop. If findings from this study were to be repeated in other centres we may be able to discharge stable patients with coffee ground vomiting to early OPD endoscopy thus reducing length of stay and pressure on already stretched inpatient emergency workload.