When “trace” is not trivial: Mild ascites as the missing prognostic middle ground in cirrhosis