| Autor: |
Stephen Rahul Atkinson, Mahmoud Aly, Katharina Remih, Luke David Tyson, Nurdan Guldiken, Robert Goldin, Alberto Quaglia, Mark Thursz, Pavel Strnad |
| Rok vydání: |
2021 |
| Předmět: |
|
| Zdroj: |
Liver international : official journal of the International Association for the Study of the LiverREFERENCES. 42(5) |
| ISSN: |
1478-3231 |
| Popis: |
Up to 30% of patients with severe alcoholic hepatitis (sAH) die within 3 months of presentation. The degree of ductular reaction, characterized by accumulation of biliary and liver progenitor cells, confers a poor prognosis. Keratin fragments are established serological surrogates of liver injury and keratin 19 (K19) is a histological marker of the ductular reaction. We assessed the relationship between serum K19 levels (viz. CYFRA21-1), histology and outcome in patients with sAH.Serum CYFRA21-1 was quantified in pre-treatment serum samples from 824 patients enrolled in the STOPAH trial. The cohort was randomly divided into two groups to test mortality associations; histological analyses were performed using the 87 cases with suitable samples.CYFRA21-1 levels were elevated in sAH and strongly predicted alcoholic steatohepatitis (ASH) on biopsy (area under the receiver operated characteristic [AUROC] 0.785 [95% Confidence Interval 0.602-0.967]) and 90-day survival (AUROC 0.684/0.693). The predictive ability of CYFRA21-1 was comparable with the model of end-stage liver disease (MELD) score and was independently associated with survival in multivariable analysis. CYFRA21-1 moderately correlated with hepatocellular injury markers M30/M65 but displayed a more robust predictive ability. A combination of MELD and CYFRA21-1 conferred a modest improvement in the AUROC value (0.731/0.743).In sAH serum, CYFRA21-1 levels associate with the presence of ASH on biopsy and independently predict 90-day survival. As a single marker performance is comparable to established scoring systems. Therefore, CYFRA21-1, which is available in many clinical laboratories, may become a useful component of prognostic models. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
Plný text