Abstract

Background:

Acute variceal bleeding (AVB) is a cause of serious morbidity and mortality in cirrhosis. While more severe disease (e.g. higher Child-Turcotte-Pugh [CTP] score) has been shown to correlate with worse outcomes after AVB, it is unclear if this correlation remains universal across all cirrhosis etiologies. Using a large multinational cohort of AVB patients, this study aims to compare the 6-week mortality after AVB between alcohol-related (AR) and non-alcohol-related (NAR) cirrhosis patients, adjusting for CTP score.

Methods:

Patients with cirrhosis hospitalized for AVB from 2012-2020 were retrospectively identified from 7 hospitals in the US and Singapore. Those receiving pre-emptive TIPS or liver transplant were excluded. AR cirrhosis was defined based on significant alcohol history during medical chart review. All patients were managed with endoscopy, antibiotics, and vasoactive therapy. The primary outcome was 6-week mortality (Baveno 6). A quasi-experimental method, using propensity score matching with the nearest 3 neighbors and maximum caliper distance of 0.1, was performed in a logit model, adjusting for age, gender, and CTP score. This was iteratively performed to evaluate the average treatment effect on the treated (ATET: the estimated causal effect of a variable [AR cirrhosis] on the observed difference in outcomes) in all patients, those with advanced disease (CTP>7), and those with CTP C 10-13 or CTP B with active hemorrhage on endoscopy, where a negative ATET value represents 6-week mortality benefit.

Results:

A total of 469 patients (204 [44%] with AR disease) were included. The baseline characteristics were similar across sites with the exception that the Singapore cohort was older, had less alcohol use, and had more HBV-related cirrhosis. The AR cirrhosis cohort had more men, lower age, and higher admission MELD/CTP/MELD 3.0 scores. Overall 6-week mortality was 15% with no differences between AR and NAR cirrhosis (p=0.77, [Fig A]). Adjusting for CTP score, AR cirrhosis was associated with lower 6-week mortality risk (OR 0.49, 95% CI 0.27-0.90, p=0.02) [Fig B]. After propensity matching with age, gender and CTP score, there was a significant mortality reduction from AR cirrhosis etiology among those with CTP B and active bleeding or CTP C (n=199, ATET -0.23, 95% CI -0.35, -0.10, p<0.01) as well as those with CTP>7 (n=272, ATET -0.16, 95% CI -0.31, -0.01, p=0.04). The overall beneficial effect from having AR cirrhosis compared to other etiologies neared significance (n=469, ATET -0.09, 95% CI -0.20, +0.02, p=0.10).

Conclusion:

In a large multinational cohort of patients with AVB who did not undergo TIPS, we found that patients with AR cirrhosis have lower 6-week mortality than those with other etiologies, particularly at higher CTP scores. This highlights the importance of underlying disease etiology on AVB outcomes and warrants further investigation to elucidate the impact of therapeutics on these subgroups.