Prognostic significance of mild ascites in patients with cirrhosis

Abstract

Background: Overt (clinically detectable) ascites is the most common decompensating event in cirrhosis and is associated with a high mortality. The impact of mild ascites (only detectable by imaging) remains unclear. Methods: Retrospective cohort study in patients with cirrhosis using the Veterans Outcomes and Costs Associated with Liver Disease (VOCAL) cohort. Imaging reports at time of diagnosis of cirrhosis were analyzed using natural language processing and classified in three groups: no ascites, grade 1 (G1) or mild ascites, and grade 2/3 (G2/G3) or overt ascites. Mortality was compared across the groups using the Kaplan-Meier method and multivariable Cox proportional hazards analysis. Results: 41,669 patients were included in the study: 28,628 (68.7%) without ascites, 4,046 (9.7%) with G1, and 6,268 (15.0%) with G2/G3 ascites. Median follow-up time was 7.7 years (IQR 5.1-10.6). In patients with G1 (mild) ascites, median survival was 3.9 years (IQR 1.8-8.3), significantly lower than in those without ascites (6.5 years; IQR 3.1-12.2, p<0.0001) but significantly higher than in those with G2/G3 ascites (3.5 years; IQR 1.5-7.9, p<0.0001). After adjusting for potential confounders, the presence of any grade of ascites was associated with a higher hazard of mortality: G1 ascites (HR 1.50; 95% CI 1.44-1.57, p<0.001) and G2/G3 ascites (HR 1.63; 95% CI 1.57-1.69, p<0.001). These results were consistent across multiple sensitivity analyses. Conclusion: Grade 1 (mild or subclinical) ascites is associated with poorer survival compared to no ascites and represents an intermediate prognostic stage between compensated and decompensated cirrhosis.