血清免疫球蛋白对慢性肝炎病理分级和分期的判别价值

Practical value of serum immunoglobulins for differentiating pathological grades and stages of chronic hepatitis

目的构建诊断慢性肝炎病理分级(G)和分期(S)的数学模型,评价血清免疫球蛋白(IgG、IgA和IgM)判别慢性肝炎病理分级和分期的价值。方法慢性乙型肝炎172例,肝穿刺活检标本进行病理分级和分期;免疫透射比浊法测定血清免疫球蛋白。以病理分级和分期作为因变量,血清IgG、IgA和IgM作为自变量,用Bayes逐步判别分析构建判别函数。结果血清IgG和IgA水平与病理分级呈显著正相关(r=0.324,P=0.000和r=0.468,P=0.000),与病理分期呈显著正相关(r=0.201,P=0.008和r=0.254,P=0.001);血清IgM水平与病理分级和分期均无显著相关性(r=0.046,P=0.547和r=0.104,P=0.176)。符合模型纳入变量、进入判别模型的指标只有血清IgG,该模型判别病理分级G1、G2、G3、G4的正确率分别为66.67%、14.44%、30.61%、66.67%,总正确率29.07%;判别病理分期S0、S1、S2、S3、S4的正确率分别为50.00%、2.94%、13.95%、27.12%、50.00%,总正确率60.71%。结论血清IgG符合该数学模型纳入变量,对判别肝脏病理分级和分期有一定价值。

Objective To build mathematical models for diagnosing pathological grades （G） and stages （S） of chronic hepatitis, and to evaluate the practical value of serum immunoglobulins （IgG, IgA and IgM） for differentiating pathological grading and staging of chronic hepatitis. Methods 172 patients with chronic hepatitis B were enrolled into present study. ＂Fhe specimens of liver biopsy were diagnosed for pathological grading and staging. Serum immunoglobulins were determined by inmmnoturbidimetric assay. Taking pathological grading and staging as the grouping variables, and serum IgG, IgA and IgM as the independents, the Fisher＇s linear discriminant functions were built by Bayes＇s stepwise discriminant analysis. Results Serum IgG and IgA levels were both correlated significantly with pathological grading （rs=0.324. P=0.001） and rs =0.468,P =0.000） and pathological staging （r = 0.201, P =0.008 and r =0.254,P = 0.001 ） ; serum IgM levels was not correlated significantly with not only pathological grading but also pathological staging （r= 0. 046, P= 0. 547, and r= 0. 104,P = 0. 176）. Only serum IgG levels were entered into the discriminant functions according to the entry criteria of the independents. The correctly classified rates for pathological grading of the original grouped cases by the discriminant models were 66.67% for G1, 14. 44% for G2, 30. 61 % for G3, 66. 67% for G4, respectively; and the total correctly classified total rate of the original grouped cases was 29.07%; The correctly classified rates for pathological staging of the original grouped cases by the discriminant models were 50.00% for S0, 2.94% for S1, 13.95% forS2. 27.12% for S3, 50.00% for S4, respectively; and the total correctly classified total rate of the original grouped cases was 60.71%. Conclusion The mathematical models have practical value to some extent, and only serum IgG of immunoglobulins has practical value to some extent for differentiating pathological gradings and stagings of chronic hepatitis.