Nevin Manimala Statistics

Model of disease severity in alcoholic hepatitis and novel prognostic insights

Rom J Intern Med. 2024 Jan 5. doi: 10.2478/rjim-2024-0001. Online ahead of print.


INTRODUCTION: Harmful alcohol consumption is one of the leading risk factors for global disease burden and injury condition, causing death and disability early in life, with over 3 million deaths worldwide every year. Alcoholic hepatitis (AH) is a clinical syndrome characterized by hepatic failure with recent onset of jaundice, consequence of a heavy chronic alcohol drinking. The disease severity ranges from mild to severe cases, with high short-term mortality. Individual variety regarding disease outcome and therapeutic response complicates the prognosis stratification. Thus, novel parameters and continuously sought for a better disease outcome assessment.

AIMS AND OBJECTIVES: To highlight new parameters that accurately assess 30-day mortality (short-term) in patients with AH and to develop a new severity score that uses readily available parameters accessible to any clinician.

MATERIALS AND METHODS: This is a prospective study on patients diagnosed with AH between 2022-2023. We identified 70 patients with AH who met the National Institute on Alcohol Abuse and Alcoholism (NIAAA) criteria for diagnosis after exclusion of patients with severe comorbidities that could influence disease outcome. Clinical and paraclinical parameters were assessed at least on admission and day 7. Mortality at 30-day was considered the endpoint. The database was composed using Microsoft Excel (Microsoft Corporation) and the data was analyzed using SPSS Statistics version 26 (IBM Corporation).

RESULTS: A total of 70 patients were included in the study with a mortality at 30-days of 22.9% (n=16). The independent variables associated with increased short-term mortality identified using the univariate analysis were: fever, infection, esophageal varices, prothrombin time PT, INR, total bilirubin, CRP, LDH and CHI (creatinine height index). Using multivariate regression we determined a novel prognostic score, with criterion for retaining variable being p<0.05. Total bilirubin day 7, CRP, PT, fever and CHI resulted after the analysis and were included into a new mortality score. Our Prognostic Model Score obtained an area under the ROC of 0.950 (95% CI: 0.890-0.980, p<0.001), with a cut-off value of 13.75 (Sn=87.5%, Sp=91%). Regarding the consecrated prognostic scores, MDF and Lille score obtained good AUROCs=0.839 and 0.881, respectively (p<0.000), with cut-off values comparable with literature (MDF=34.35 vs 32) and (Lille=0.475 vs 0.450). The discriminatory power for ABIC (p=0.58), GAHS (p=0.16), MELD-Na (p=0.61) was not significant.

CONCLUSION: We obtained a new prognostic score for the assessment of 30-day mortality in AH that includes markers of inflammation (CRP, fever) and markers of sarcopenia (CHI) along parameters of hepatic disfunction (total bilirubin and PT). Amongst consecrated prognostic models, MDF and Lille scores were representative for our study, while ABIC, GAHS and MELD-Na did not attain statistical significance. Our score is unique by the addition of CRP and this could prove to be a useful tool in AH severity stratification.

PMID:38180800 | DOI:10.2478/rjim-2024-0001

By Nevin Manimala

Portfolio Website for Nevin Manimala