The Reliability Of 'Coffee-Ground' Vomiting As An Indicator Of Active Ugib: A Tertiary Centre Experience

Introduction Acute Upper GI Bleeds (UGIB) are a common medical emergency and account for significant morbidity and mortality. ‘Coffee-ground’ vomiting is considered indicative of UGIB however its reliability as an indicator of active bleeding is contentious. OGD is the current gold standard in the diagnosis and management of UGIB however UGI capsule endoscopy is an emerging alternative which can rapidly screen for bleeding and has benefits over OGD in terms of invasiveness, patient tolerance, rapid deployment anywhere and thus reduced hospital admission rates and ultimately cost-effectiveness. The aim of this study was to establish whether coffee-ground vomiting is a reliable indicator of active UGIB. Methods Electronic referrals for OGD, which contained the phrase ‘coffee-ground vomiting’, were analysed and endoscopy reports assessed to identify whether active UGIB was identified. Results During a 12-month period; 552 electronic endoscopy referrals for UGIB were submitted. 62 referrals contained the specific phrase ‘coffee-ground vomiting’; 9 patients were excluded. 33/53 referrals came from inpatient medical wards. 35 patients were male, mean age at time of referral was 67.5 years (22.0 – 90.0). Mean Glasgow Blatchford Score at referral was 7.4 (1.0 – 15.0). Mean time between referral and endoscopy was 1.3 days (0.0 – 12.0). The most common endoscopic diagnoses were oesophagitis (16/53 cases) and gastritis (9/53 cases). OGD findings were normal in 10 cases. Active bleeding was identified in 6 patients and in 2 patients a visible vessel was visualised without evidence of active bleeding at endoscopy. 4 patients required endotherapy namely injection and clipping, 1 patient required variceal banding. All 6 patients with confirmed active bleeding on endoscopy had melaena on presentation. Mean length of stay (LOS) was 8 days (1.0 – 203.0) and in 1 case a repeat OGD was required during the same admission. Conclusions The results of this study demonstrate that only a minority of OGDs performed for the indication of ‘coffee-ground’ vomiting demonstrated active UGIB. In addition, a significant proportion of the cohort had a normal study. No patients with coffee ground vomiting without melaena had active UGIB and none of these patients required endotherapy. This study could be used to support the proposal that patients with coffee-ground vomitus without other corroborating evidence of UGIB are very unlikely to have active bleeding and/or the need for endoscopic haemostasis. These patients could have an UGI capsule endoscopy in the emergency department as opposed to hospital admission and subsequent wait for an inpatient OGD. This could consequently have significant implications on admission rates, LOS and hospital associated morbidity.