Clinical outcomes of variceal and non-variceal upper gastrointestinal bleeding in cirrhosis patients in Singapore

Dear Sir, Upper gastrointestinal bleeding (UGIB) in cirrhosis patients is associated with substantial morbidity and mortality. While the standard of care has substantially improved over the years, the outcomes of variceal bleeding (VB) and its mortality rate have remained high, ranging between 10% and 20% at 6 weeks.1 The updated data on the outcome of cirrhosis patients with UGIB in Singapore are limited, the last study with only 22 patients having been conducted more than a decade ago.2 While the role of prophylactic antibiotics is debatable in the setting of non-variceal bleeding (NVB) and early cirrhosis,3,4,5 it is clear that unnecessary antibiotics could promote the emergence of multidrug-resistant organisms and Clostridioides difficile infection, leading to poorer outcomes among cirrhosis patients.6 In this study, we sought to understand the comparative outcomes between VB and NVB, as it may help guide prognostication7,8 and rationalise antibiotic usage in cirrhosis patients with NVB. We retrospectively reviewed all cirrhosis patients with UGIB who were consecutively admitted to our institution between 2017 and 2021. The baseline characteristics of 218 cirrhosis patients with UGIB are summarised in Table S1 see Supplemental Digital Appendix at https://links.lww.com/SGMJ/A90. Patients with NVB were older, had higher baseline creatinine and more advanced cirrhosis, and were more likely to receive anticoagulation. The clinical outcomes of cirrhosis patients with VB and VCB are summarised in Table 1. Non-variceal bleeding was associated with a higher risk of bacterial infection (29.5% vs. 13.8%, P = 0.013), particularly pneumonia (18.2% vs. 7.5%, P = 0.031), which resulted in a longer median (interquartile range) length of stay (6 4–10 vs. 12.5 5–32, P < 0.001). Meanwhile, patients with VB had a significantly higher adjusted risk of developing new ascites at 1 year (odds ratio OR 9.4, 95% confidence interval CI 1.9–46.8, P = 0.006) despite a lower proportion of patients with Child–Turcotte–Pugh (CTP) class A (23.9% vs. 51.7%) and more liberal use of non-selective beta-blockers. These findings highlight the difference in pathophysiology between VB (which is portal hypertension driven) and NVB; thus, other portal hypertension–related complications were more likely in patients with VB.Table 1: Clinical outcomes between cirrhosis patients with variceal bleeding (VB) and non-variceal bleeding (NVB).Overall, the death-adjusted 5-day rebleeding rate and 6-week mortality were 3.3% and 18.3%, respectively. The most common cause of death was cancer (30%) followed by acute on chronic liver failure (20%), with bleeding and infection attributing to about 15% and 7.5% of deaths, respectively. Eight patients had early mortality within 5 days (25% was bleeding-related mortality). After adjusting for confounders (age, creatinine, CTP score and the use of anticoagulation), there was no significant difference between VB and NVB in terms of 6-week mortality (17.8% vs. 20.5%) or 1-year mortality (30.5% vs. 43.2%). Although the early rebleeding rate showed a decline compared to an earlier study (3.7% vs. 9%), the mortality risk of VB remained similar (18%).2 Compared to the study by Yew et al.,2 we found that hepatitis C and non-alcoholic steatohepatitis were the dominant aetiologies for cirrhosis. While this finding could be attributed to the increased screening for hepatitis C virus and growing prevalence of obesity, findings from this single-centre study should be validated in a more representative nationwide study. To conclude, NVB was associated with a higher morbidity comparable to VB. Higher risk of developing ascites among cirrhosis patients with VB highlighted the need to optimise strategies for portal pressure reduction using both non-selective beta-blockers and removal of underlying aetiology, whenever possible. Financial support and sponsorship Wong YJ is supported by the Nurturing Clinician Scientist Scheme (NCCS) award by SingHealth Duke-NUS Academic Medical Centre. Conflicts of interest Ang TL is a member of the SMJ Editorial Board, and was thus not involved in the peer review and publication decisions of this article.