Relationship between updated MELD and prognosis in alcohol-associated hepatitis : Opportunities for more efficient trial design