Article Text


P36 Morbidity and mortality associated with varices in patients with primary biliary cirrhosis
  1. I M Patanwala1,
  2. R Walter1,
  3. J Newton1,
  4. M Hudson2,
  5. D E Jones1
  1. 1Institute of Cellular Medicine, Newcastle University, UK
  2. 2Liver Unit Freeman Hospital Newcastle upon Tyne UK


Introduction Several studies have shown that gastro-oesophageal varices (GOV) are relatively common in patients with PBC, and can occur in pre-cirrhotic and asymptomatic patients as well as in patients with advanced disease. An important practical problem faced by clinicians managing PBC patients with GOV is the approach to their long-term management especially, defining the appropriate timing for referral for liver transplantation.

Aim To address the clinical impact of varices on morbidity and mortality in PBC in a large, long-term follow-up cohort of patients.

Method A retrospective study was designed to identify all PBC patients had had endoscopy (OGD) at the Freeman Hospital, Newcastle for any clinical indication. PBC patients with and without GOV at OGD were characterised by extensive review of their clinical records. Data obtained included survival and transplantation history. The log rank test was used to compare transplant free survival between groups.

Results 330 PBC patients (91.5% female, median age 64 yrs) were identified as ever having had an OGD at the Freeman Hospital. 159 [48% (95% CI 43% to 54%)] were found to have GOV. Subgroups with and without GOV were equivalent in terms of age, sex and time to endoscopy (Abstract P36 table 1). 39 (25%, 95% CI 18% to 32%) patients had GOV diagnosed at OGD performed at the time of their index bleed. In total, 83 (52%, 95% CI 44% to 60%) patients suffered 245 episodes of variceal bleeding during a median follow-up of 11 yrs (IQR 8). Of the 120 that did not present with a bleed 44 (37%, 95% CI 28% to 46%) bled a median of 1.5 yrs (IQR 3.75) after varices had been diagnosed In patients with varices that bled, there was no significant difference in the proportion that were on non-selective β blockers as compared with those that did not receive/tolerate these agents (48% vs 52%, p=0.75 Fisher's exact test). Unfortunately, data on the physiological adequacy of β blockade was not available. Importantly, 21 (13%, 95% CI 8% to 19%) PBC patients with varices had early stage (Scheuer Stage I, I-II, II) disease, and of these 3 (14%, 95% CI 3% to 36%) presented with a variceal bleed as the first presentation of their varices and a further 5 (24%, 95% CI 8% to 47%) bled during follow-up. Transplant free survival after diagnosis of PBC was significantly better in those without varices when compared to those with varices (p<0.001). There was no significant difference in survival in patients with varices that bled and those that did not (p=0.1) (Abstract P36 figure 1).

Abstract P36 Table 1

Baselines demographics of the PBC patients with and without varices

Abstract P36 Figure 1

Kaplan–Meier curves comparing transplant free survival of PBC patients (A). with and without varices and (B). varices that bled vs varices that did not.

Conclusion The development of GOV heralds a poor prognostic outlook for patients with PBC. Bleeding from these GOV does not further worsen survival in these patients and therefore prophylaxis against bleeding may not offer a survival advantage in these patients, who should be considered early for transplantation. The apparent importance of GOV in terms of prognosis, and decision making to optimise outcome means that we should re-look at strategies to screen for GOV in PBC.

Statistics from

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.