Abstract Book 2020 Contents Page Calne Williams presentations 3 Medawar Medal presentations 7 Oral presentations 15 Posters 51 2 CW1 The relationship between Cardiopulmonary Exercise Testing (CPET) and muscle volume in liver transplant candidates Mr Ben Fox1,2, Mrs Janice Miller2, Dr Helen Usher2, Mr Richard Skipworth2, Professor Stephen Wigmore2 1University of Edinburgh, Edinburgh, United Kingdom. 2Royal Infirmary of Edinburgh, Edinburgh, United Kingdom Introduction: Sarcopenia is described as the progressive and generalised loss of skeletal muscle mass and strength, which often accompanies liver cirrhosis. Cardiopulmonary Exercise Testing (CPET) is a key clinical tool used to assess exercise capacity and is key in predicting surgical outcomes. This study sought to investigate the relationship between CT measures of muscle volume and CPET performance in patients undergoing liver transplant assessment. The secondary aim was to investigate the relationship between sarcopenia and post-operative morbidity and mortality. Methods: A single-centre retrospective cohort study of 400 patients who underwent liver transplant assessment between 1st July 2016 and 1st July 2018 was performed. Routine pre-operative CT scans were used to analyse muscle and adipose tissue which were indexed for height to generate a Skeletal Muscle Index (Cut-off values shown in Table 1). These were compared with CPET variables collected as part of liver transplant assessment work-up. Groups were evaluated using the independent T-test, Chi-squared test and Kaplan-Meier distribution. Results: Of the 400 patients, 54 met the exclusion criteria, leaving 346 for analysis (121 sarcopenic and 225 non- sarcopenic). Clinical significance was found between sarcopenia and Anaerobic Threshold (AT) (p=0.033), Subcostal Girth (p=0.024), predicted Lung Transfer Factor (TCO) (p=0.010), peak Heart Rate (p=0.030), UKELD score (p=0.003) and transplant suitability (p=0.020) (See Figure 1). There were no significant differences between sarcopenia, Clavien-Dindo complications and mortality. Discussion: Sarcopenia is shown to be an accurate predictor of poorer CPET performance. Sarcopenia also has value in predicting worse pulmonary function tests (PFT’s) and higher UKELD scores, and thus a lower likelihood of being listed for liver transplantation. However, the presence of sarcopenia is unable to predict surgical outcomes following transplantation including long-term survival. Table 1 3 Figure 1 CW2 - withdrawn Long-term incidence of de novo malignancies in liver transplant recipients and weaning off immunosuppression in tor vergata l: results from an observational study 2 CW3 Time-trends in patient mortality and graft survival of patients receiving a DCD or DBD liver transplantation in the UK and Ireland between 2008 and 2016 Mr David Wallace1,2, Dr Thomas Cowling2, Dr Abid Suddle1, Dr Alex Gimson3, Dr Ian Rowe4, Mr Chris Callaghan5, Dr Gonzalo Sapisochin6,7, Dr Neil Mehta8, Professor Jan van der Meulen2, Dr Kate Walker2, Professor Nigel Heaton1 1Institute of Liver Studies, King's College Hospital NHS Foundation Trust, London, United Kingdom. 2Department of Health Services Research and Policy, London School of Hygiene and Tropical Medicine, London, United Kingdom. 3The Liver Unit, Addenbrooke's Hospital, Cambridge University Hospitals NHS Foundation Trust, Cambridge, United Kingdom. 4Liver Unit, St James' Hospital and University of Leeds, London, United Kingdom. 5Department of Nephrology and Transplantation, Renal Unit, Guy's Hospital, London, United Kingdom. 6Multi-Organ Transplant, Toronto General Surgery, Canada, Toronto, Canada. 7Department of General Surgery, University of Toronto, Toronto, Canada. 8Division of Gastroenterology, Department of Medicine, University of California, San Francisco, USA Introduction: Internationally, the UK is the primary exponent in the transplantation of livers from controlled donation after circulatory death (DCD). However, concerns exist in the efficacy of these grafts. We evaluated the mortality, graft failure and post-operative complications of DCD and DBD donor liver transplantation in successive eras between 2008 and 2016. Methods: All first-time elective adult liver transplant recipients in the UK were identified and hazard ratios comparing the impact of era (2008-2011 and 2012-2016) on post-transplant mortality, graft failure and complications were estimated. Results: 1 176 DCD recipients and 3 749 DBD recipients were included. The use of livers from DCD donors increased from 19.3% in era 1 to 26.7% in era 2. 3-year patient mortality decreased markedly from 19.6% in era 1 to 10.4% in era 2 (aHR:0.43, 95%CI: 0.30-0.62) for DCD recipients but only decreased from 12.8% to 11.3% (aHR:0.96, 0.78-1.19) in DBD recipients. Between eras no improvements in overall 3-year graft failure were observed for DCD (aHR:0.80, 0.61-1.05, p=0.11) or DBD recipients (aHR:0.95, 0.79-0.60, p=0.60) but the rate of re-transplantation increased from 7.2% to 10.1% in DCD recipients (p=0.14) and decreased from 4.8% to 3.7% in DBD recipients (p=0.01). In era 2, there was no difference in mortality between those receiving a DCD or DBD liver (aHR: 0.78, 0.56-1.09, p=0.14) however the incidence of biliary tract strictures increased for both cohorts (5.0% to 6.9% and 3.8% to 4.8%, respectively). Conclusion: Between 2008 and 2016, mortality more than halved in those who received a DCD donor liver. In the UK, mortality following liver transplantation now appears to be comparable for patients receiving DCD and DBD donor livers. Figure 1: 3-year patient survival across different eras of transplantation (2008-2011 and 2012-2016) in recipients receiving DCD or DBD livers (n=4 925). a. DCD (n=1 176) b. DBD (n=3 749) 3 CW4 Early anastomotic biliary complications after liver transplantation Dr Samuel Tingle, Miss Emily Thompson, Mr Shoaib Ali, Mr Ibrahim Ibrahim, Miss Ellie Irwin, Mr Gourab Sen, Prof Steve White, Prof Derek Manas, Mr Colin Wilson Institute of Transplantation, Freeman Hospital, Newcastle, United Kingdom Introduction: Biliary leaks and anastomotic strictures are common early biliary complications (EBC) following liver transplantation. However, their impact on outcomes remains controversial and poorly described. Methods: National UK data on adult liver transplantation between 2006 and 2017 collected by NHSBT was reviewed (n=8304). Multiple imputations were performed to account for missing data. Adjusted regression models were used to assess predictors of EBC, and their impact on graft and patient survival. 35 potential confounders were included, and backwards stepwise selection enabled unbiased selection of variables for inclusion in final models. Results: EBC occurred in 9.6% of patients and was associated with significantly worse graft and patient survival (Figure 1). Adjusted cox regression revealed that EBCs have a significant and independent impact on graft survival (Leak HR=1.325; P=0.021, Stricture HR=1.514; P=0.002, Leak plus stricture HR=1.533, P=0.034) and patient survival (Leak HR=1.218; P=0.131, Stricture HR=1.578; P<0.001, Leak plus stricture HR=1.507; P=0.044). Patients with EBC had longer median hospital stay (23 versus 15 days; P<0.001) and increased chance for readmission within the first year (56% versus 32%; P<0.001). On adjusted logistic regression the following were identified as independent risk factors for development of EBC: donation following circulatory death (OR=1.280; P=0.009), accessory hepatic artery (OR=1.324; P=0.005), vascular anastomosis time in minutes (OR=1.005; P=0.032) and ethnicity ‘other’ (OR=1.838; P=0.011). In addition, biliary stricture was significantly less likely with Roux-en-Y anastomosis (OR=0.558; P=0.001), and biliary leak was significantly more likely with T-tube anastomosis (OR=2.055; P=0.006) and in recipients with higher MELD scores (OR=1.015; P=0.023). Discussion: EBCs prolong hospital stay, increase readmission rates and are independent risk factors for diminished graft survival and increased mortality in liver transplantation. We have identified factors that increase the likelihood of EBC occurrence; further research into interventions to prevent EBCs in these at-risk groups is vital to improve liver transplantation outcomes. Figure 1: 4 CW5 Artificial Intelligence more accurately predicts individual graft survival than traditional modelling: Artificial Intelligence and Liver Transplant (AI4T) Dr Laura Wingfield1, Mr Simon Thorogood2, Mr Carlo Ceresa1, Mr Jacques Fleuriot2, Mr Simon Knight1 1University of Oxford, Oxford, United Kingdom. 2University of Edinburgh, Edinburgh, United Kingdom Introduction: Scoring systems predict outcomes following liver transplantation are useful for shared decision making, informed consent and organ allocation. Most scoring (Model for end-stage liver disease (MELD), Survival Outcomes after Liver Transplantation (SOFT)) utilise simple methodology like logistic regression with limited predictive performance. We investigated whether Artificial Intelligence (AI) and Machine Learning (ML) improve outcome prediction in liver transplantation. The aim is to evaluate ML methodology including random survival forests (RSF), artificial neural networks (ANN) and multi-task logistic regression (N-MTLR), compared to liver scoring indices to predict graft survival at 1- and 5- year periods. Methods: All UK liver transplants in patients 16 years performed January 2000 to December 2016 with outcomes recorded in the National Health Service Blood and Transplant (NHSBT) registry were included (n=10,388 donor-recipient pairs). Several ML models including RFS, ANN and N-MTLR were generated to predict individual graft survival at 1- and 5- years post transplant. Models were compared to traditional regression scores. Results: A total of 9,560 matched D-R transplants were included. The median donor age was 48 years old (n= 51% male, n=49% female). Using MELD score for survival prediction resulted in an area under the receiver operating characteristic curve (AUC-ROC) of 0.514 at 1 year and 0.471 at five years. The variables used in MELD using ANN resulted in AUC-ROC of 0.649 at 1 year and combined with RSF an AUC-ROC 0.605 at 5 years.