A88 Gut June 2013 Vol 62(Suppl 1):A1–A306 BSG abstracts discharged on propranolol. Only 35% of patients received sucral- fate post banding. Only 13% of patients had a TIPS procedure. A further 48% of our sample could have been considered for TIPS where no contra-indica- tion was found (i.e hepatic encephalopathy not secondary to UGI bleeding or renal dysfunction). The average length of stay was 14 days and the 30 day mortality rate was 13%. Conclusion The pharmacological management was generally good and our mortality rate of 13% was better than the quoted fig- ures of 30% in the literature. However, we identified a possible 48% of the sample could have been considered for TIPS which is no lon- ger considered rescue therapy alone with good evidence for its early use, with subsequent prevention of readmission from a variceal bleed. We recommend early pharmacotherapy with terlipressin and antibiotics as soon as varices are suspected with early ITU involve- ment, airway protection at endoscopy and early TIPS in selected patients. Disclosure of Interest None Declared. SINGLE CENTRE MANAGEMENT OF PYOGENIC LIVER ABSCESSES: SURPRISINGLY POOR BUT MORTALITY STILL LOW doi:10.1136/gutjnl-2013-304907.194 1 A Puri, 1 M Bhuva, 1,* F Sampaziotis, 1 B Macfarlane. 1 Gastroenterology, West Herts NHS Trust, Watford, UK Introduction Untreated, pyogenic liver abscesses have a mortal- ity approaching 100%. Three admissions in a week sparked inter- est in the best management of this condition. Although common, no comprehensive management guidelines could be found, prompting further review into how well this condition was man- aged locally Methods Retrospective analysis of all patients admitted to Wat- ford Hospital between 2006 and 2011 with a diagnosis of pyogenic liver abscess. Data was collected to evaluate use of cultures, radio- logical intervention (aspiration or drain insertion), source of infec- tion, investigation for cause, follow up and outcome Results Fourty four admissions were identified: 39 patients with 5 re-admissions. Mean age was 62 yrs, 59% male, 41% female. Eleven patients were managed by the Gastro team. Assumption of source was made on CT imaging results: 46% pre- sumed portal translocation (most diverticular disease), 36% biliary, 18% unidentified. Blood cultures were taken in 24 patients (42% positive). Abscess aspirates were taken in 33 cases, sent for culture in 30 (50% positive). Presumed biliary or unidentified sources grew gramme negative organisms in 12/13 cases. Presumed portal sources grew gramme positives in 7/8 and anaerobes 1/8. Abscess size was < 3 cm in 5 cases (incl. 2 readmissions). Four received anitbiotics (Abc) alone:resolution in 3/4, 1/4 no follow up. One was managed to resolution with Abc and aspiration. Mean length of stay was 11 days. In 6 patients the abscess was 3–5 cm. In this group, 1 patient with malignancy died, 1 treated successfully with Abc alone. The remaining 4 were treated with Abc and aspiration: 1/4 resolution, 1/4 readmitted, 2/4 no follow up. Mean length of stay 15 days. Thirty three patients had abscesses > 5 cm (incl. 3 readmissions). Nineteen were treated with Abc and drainage: 2/19 had underlying malignancy and died, 6/19 resolution, 3/19 readmitted, 7/19 no follow up, 1/19 referred to surgery. Of the remaining fourteen, 3/14 had Abc alone (2 resolved, 1 patient with two readmissions no follow up), 1/14 a readmission referred for surgery and 10/14 Abc and aspiration. Out- come in these ten: 1/10 mutiple aspiration, 2/10 drain insertion, 1/10 surgical referral, 1/10 readmitted, 1/10 partial response, 3/10 no follow up, 1/10 resolution. Mean length of stay in > 5 cm group : 27 days. PTU-104 the ascitic fluid samples sent to our microbiology department. Case notes for these patients were reviewed and data were collected on patient demographics, aetiology of cirrhosis, use of blood products and human albumin solution (HAS) and volume of ascites drained. Results 56 LVP were performed on 28 patients. 24 were male, age range 30 – 84 years (median 59 years). Alcohol was either the only or a contributory cause of cirrhosis in 25 (89%) of patients. None had hepatitis B or C virus infection. 5 patients received fresh frozen plasma (14 units in total) and 1 received octaplex® prior to LVP. The total cost was £1024. 8 patients had less than 5L ascites drained and received a total of 19 units of 20% HAS. 16 patients received more than 8g albumin per litre of ascites drained (a total of 31 unnecessary units). The total cost of this was £1400. The potential cost saving per procedure was £49.47. However data on albumin administration was unavailable for 7 patients and this could be an underestimate. Conclusion Alcohol is the predominant cause of cirrhosis requir- ing LVP in our population and working age men constitute the largest proportion. Significant cost savings can be made by avoid- ing unnecessary blood products and by avoiding excessive use of albumin or administering other fluids when less than 5 litres of ascites are drained. Trusts should ensure relevant protocols are in place. Disclosure of Interest None Declared REFERENCE 1. EASL clinical practise guidelines on the management of ascites, spontaneous bac- terial peritonitis, and hepatorenal syndrome in cirrhosis. Journal of hepatology 2010; 53(3):397–417. AUDIT INTO THE MANAGEMENT OF ACUTE VARICEAL BLEEDS AND THE ROLE OF TIPS doi:10.1136/gutjnl-2013-304907.193 1 V Chhaya, 1 A Bevan, 1,* F Stanley, 1 M Pollard, 1 S Clark. 1 Gastroenterology, St Georges Hospital, London, UK Introduction The mortality associated with acute variceal bleed- ing is significant with a 70% risk of recurrent haemorrhage in survi- vors. Our aim was to assess the outcome from variceal bleeding at St George’s Hospital over a one year period, to determine whether current clinical guidelines in the management of variceal bleeding are being adhered to, and to assess whether we are utilising the role for early TIPS (transjugular intrahepatic portosystemic shunt) in patients with variceal bleeding. Methods A dataset of all adult patients admitted from 1/4/11 for a period of 12 months was obtained with a primary diagnosis code of K922 Gastrointestinal haemorrhage, unspecified (n = 378). Genu- ine cases were confirmed by reference to the Micromed endoscopy reporting tool, CEPOD emergency theatre lists, bereavement records and old inpatient lists for the Hepatology firm. Case notes were obtained for the final sample of 23 patients. Results The main cause of variceal bleeding (65%) was alcoholic liver disease (ALD). 78% were rebleeds of which 83% were within the last 6 months. 61% of patients had features of decompensation (ascites 86%, renal dysfunction 29%). Only 4% of cases were Childs-Pugh A, with 61% of cases being Childs-Pugh B and 35% Childs-Pugh C. The predicted 3 month mortality according to the MELD (model for end stage liver disease) score was 6–19.6%. An average of 2 to 3 units of blood was transfused to 78% of patients and 60% of patients required either FPP, platelets or both. All patient received an endoscopy during their admission, of which 74% were carried out within 12 hours. Only 52% were intubated for procedure and 39% were admitted to ITU post procedure. 96% received antibiotics, 87% received terlipressin and 79% were PTU-103 on June 10, 2020 by guest. Protected by copyright. http://gut.bmj.com/ Gut: first published as 10.1136/gutjnl-2013-304907.194 on 4 June 2013. Downloaded from