Article Text

PDF
PWE-050 United Kingdom Model for End-Stage Liver Disease and Model for End Stage Liver Disease scores predict outcome after oesophageal variceal haemorrhage
  1. N Mangrolia,
  2. W Gelson,
  3. A Zissimopoulos,
  4. K Lee,
  5. M Allison
  1. Department of Hepatology, Addenbrooke's Hospital, Cambridge, UK

Abstract

Introduction Variceal haemorrhage is a common, life-threatening complication of portal hypertension. In this study, we aim to assess mortality and predictors of outcome in a cohort of patients presenting acutely with a bleed.

Methods We performed a retrospective review of consecutive cases of oesophageal variceal haemorrhage presenting acutely to our centre (a regional Hepatology Unit) in 2008. 30 cases were identified. Median age was 52 (range 29–79), 14 (46.7%) patients were male and underlying liver disease aetiology was alcohol-related in 15 (50%). The Mann–Whitney U test was used to compare United Kingdom Model for End-Stage Liver Disease and Model for End Stage Liver Disease (UKELD and MELD, respectively) scores at presentation by mortality. Receiver operating characteristic (ROC) curves were plotted when significant associations with UKELD and MELD scores were found. Binary logistic regression analyses were used to assess for associations with re-bleeding and death.

Results 30-day mortality from admission to hospital was 13.3%, 90-day 30.0% and 180-day 40.0%. 7 (23.3%) patients had more than one bleed. Rescue TIPS shunt was required in 3 (10%). UKELD score at presentation was significantly higher in those who died within 30 days (mean 51.9 vs 60.6, p=0.02), but not thereafter. A similar finding for 30-day mortality was observed with MELD score (5.8 vs 20.5, p<0.0001). In ROC analyses for death prediction at 30 days, the area under the curves were 0.89 for UKELD and 0.97 for MELD. Using a UKELD cut-off of 53.5 to predict death within 30 days, sensitivity was 75% and specificity 73%. With a MELD cut-off of 11.4, these were 75% and 89%, respectively.

Univariate binary regression analyses identified positive associations of gender, creatinine and sodium with death, and ALD, bilirubin and creatinine with re-bleeding (p<0.10). However, none of these variables achieved significance with multivariate analyses.

Conclusion This study confirms a high mortality in patients presenting with variceal haemorrhage. UKELD and MELD scoring systems show promise at stratifying risk in this cohort.

Statistics from Altmetric.com

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.