Appraisement of Serum Immunoglobulins for Prediction of Hepatitis B Associated Cirrhosis by Binary Logistic Regression Analysis

ZHANG Zhan-qing,LU Wei,SHI Lian-guo,FENG Yan-ling

DOI: https://doi.org/10.3870/j.issn.1005-541X.2009.01.011

2009-01-01

Abstract:Objective To explore the diagnostic efficacy of serum immunoglobulins(IgG,IgA,IgM)for prediction of hepatitis B associated cirrhosis.Methods 172 consecutive patients were retrospectively studied,in which 28 patients and 144 patients were pathologically diagnosed as cirrhosis and non-cirrhosis.Serum immunoglobulins were determined by immunotrbidimetric assay.SPSS 12.0 software was used for statistical analyses.Independent samples test was used for comparisons of serum IgG,IgA,IgM levels between cirrhosis and non-cirrhosis.Binary logistic regression analysis was used for appraisement of serum IgG,IgA,IgM for predicting cirrhosis.Results Serum IgG and IgA levels in patients with cirrhosis were significantly higher than those in patients with non-cirrhosis(t=4.702 and 3.034,P=0.000 and 0.003),and serum IgM level in patients with cirrhosis was similar to that in patients with non-cirrhosis(t=0.132,P=0.895).Of the serum immunoglobulins,only IgG accorded with the entry criterion(P<0.05)for logistic regression analysis.The sensitivity,specificity,positive predictive value,negative predictive value,accuracy of the logistic regression model based on serum IgG were 0.107,0.986,0.600,0.850,0.843.Conclusion Serum IgG is an important index for prediction of hepatitis B associated cirrhosis.