Human serum fetuin A/alpha2HS-glycoprotein level is associated with long-term survival in patients with alcoholic liver cirrhosis, comparison with the Child-Pugh and MELD scores

Abstract

Background: Serum concentration of fetuin A/alpha 2HS- glycoprotein (AHSG) is a good indicator of liver cell function and 1-month mortality in patients with alcoholic liver cirrhosis and liver cancer. We intended to determine whether decreased serum AHSG levels are associated with long-term mortality and whether the follow-up of serum AHSG levels can add to the predictive value of the Child-Pugh (CP) and MELD scores. Methods: We determined serum AHSG concentrations in 89 patients by radial immunodiffusion. Samples were taken at the time of enrolment and in the 1st, 3rd, 6th, and the 12th month thereafter. Results: Forty-one patients died during the 1-year follow-up period, 37 of them had liver failure. Data of these patients were analysed further. Deceased patients had lower baseline AHSG levels than the 52 patients who survived (293 +/- 77 vs. 490 +/- 106 mu g/ml, mean +/- SD, p < 0.001). Of all laboratory parameters serum AHSG level, CP and MELD scores showed the greatest difference between deceased and survived patients. The cutoff AHSG level 365 mu g/ml could differentiate between deceased and survived patients (AUC: 0.937 +/- 0.025, p < 0.001, sensitivity: 0.865, specificity: 0.942) better than the MELD score of 20 (AUC: 0.739 +/- 0.052, p < 0.001, sensitivity: 0.595, specificity: 0.729). Initial AHSG concentrations < 365 mu g/ml were associated with high mortality rate (91.4%, relative risk: 9.874, 95% C.I.: 4.258-22.898, p < 0.001) compared to those with = 365 mu g/ml (9.3%). Fourteen out of these 37 fatalities occurred during the first month of observation. During months 1-12 low AHSG concentration proved to be a strong indicator of mortality (relative risk: 9.257, 95% C.I.: 3.945- 21.724, p < 0.001). Multiple logistic regression analysis indicated that decrease of serum AHSG concentration was independent of all variables that differed between survived and deceased patients during univariate analysis. Multivariate analysis showed that correlation of low serum AHSG levels with mortality was stronger than that with CP and MELD scores. Patients with AHSG < 365 mu g/ml had significantly shortened survival both in groups with MELD < 20 and MELD = 20 (p < 0.0001 and p = 0.0014, respectively). Conclusion: Serum AHSG concentration is a reliable and sensitive indicator of 1-year mortality in patients with alcoholic liver cirrhosis that compares well to the predictive value of CP score and may further improve that of MELD score.