摘要
目的 产生激素敏感性肾病与难治性肾病的判别函数 ,了解引起肾病难治的危险因素。方法 使用SPSS软件包 ,对激素敏感与难治性肾病 2组患儿的临床资料进行了逐步判别分析和条件logistic逐步回归。结果 筛选出血尿 (blo)、水肿程度 (ede)、感染 (inf)、白蛋白输入 (per) 4个变量对 2类总体有显著贡献 ,其判别方程为 :X1=- 6 .16 18+5 .2 6 2 2blo+2 .5 86 0ede+0 .710 8inf+0 .5 6 86per,X2 =- 11.1970 +6 .72 18blo +3.40 42ede +2 2 873inf+2 .75 0 3per。此判别函数的Wilks统计量为 0 .6 947,P <0 .0 1,回代判别总的准确率为 77.88%。由这些自变量组成的logistic线性方程为 :Y =- 5 .5 789+1.15 38blo +0 .8711ede +1.3714inf+1.5 72 3per,这 4个变量的条件危险比分别为 3.170 3、2 .3894、3.940 9和 4.8178。结论 血尿、高度水肿、感染。
Objective To formulate a discriminant function between children with prednisone responsive nephrosis and those with refractory nephrosis and to determine the risk factors of refractory nephrosis.?Methods Using software SPSS for MS WINDOWS Release 8.0, discriminant analysis and conditional logistic regression were performed about the clinical data of children with prednisone responsive nephrosis and of those with refractory nephrosis.?Results Four variables, i.e. blood urine(blo),edema degree(ede),infection(inf) and albumin perfusion(per), were selected into Fisher′s linear discriminant function. The canonical discriminant functions were: X 1=-6.1618+5.2622blo+2 5860ede+0.7108inf+0.5686per,and X 2=- 11.1970 +6.7218blo+3.4042ede+2.2873inf+2.7503per;Their Wilks lamda was 0.6947, P <0.01. The percentage of “grouped” cases correctly classified was 77.88%. The logistic linear function including these variables was Y =-5.5789+ 1.1538 blo+0.8711ede+1.3714inf+1.5723per.The risk ratios of these variables were respectively 3.1703, 2.3894, 3.9409 and 4.8178.?Conclusion Blood urine,high degree edema,infection and albumin perfusion are dangerous factors for refractory nephrotic syndrome.
出处
《徐州医学院学报》
CAS
2000年第2期112-114,共3页
Acta Academiae Medicinae Xuzhou
