补体组分1q和胱抑素C对狼疮性肾炎的实验室诊断

Laboratory diagnosis of lupus nephritis by complement component 1q and cystatin C

摘要:

目的：探讨血清尿素（Urea）、肌酐（creatinine，Creat）、胱抑素C（cystatin C，CysC）、估算肾小球滤过率（estimated glomeru－lar filtration rate，eGFRCysC与c-aGFR）和补体组分1q（component 1q，C1q）检测在狼疮性肾炎（lupus nephritis，LN）诊断中的应用价值。方法：选取2017年3月至2017年12月在绵阳市中心医院就诊的系统性红斑狼疮（systemic lupus erythematous，SLE）患者共560例，其中未累及肾脏患者339例（SLE组），LN活动期患者150例（LNA组），LN非活动期患者71例 （LNI组），另有360例健康个体作为对照（HC组）。测定血清Urea、Creat、Urea/Creat、CysC和C1q水平，并计算eGFRCysC与c-aGFR，以此评估这些指标对LN的诊断性能。结果：各观察指标在各组之间均有统计学差异（均P=0.000）。Spearman相关性分析显示，C1q与Creat（r=0.046，P=0.160）、Urea（r=0.011，P=0.748）、Urea/Creat（r=-0.011，P=0.743）无统计相关，与CysC（r=-0.183，P=0.000）、eGFRCysC（r=-0.183，P=0.000）和c-aGFR（r=-0.075，P=0.023）成负相关；CysC与Urea（r=0.309，P=0.000）和Creat（r=0.382，P=0.000）成正相关，与c-aGFR（r=-0.430，P=0.000）成负相关，与Urea/Creat（r=0.003，P=0.927）不相关。受试者工作特征曲线（receiver operating characteristic curve，ROC）分析显示，各观察指标单独对LN的诊断性能（area under curve，AUC）以eGFRCysC（0.891）CysC（0.890）为最大，其后依次为：C1q（0.804）、Urea（0.660）、Urea/Creat（0.630）、c-aGFR（0.547）和Creat（0.501）。C1q+CysC+c-aGFR联合检测即可达到最大诊断性能（AUC=0.962），其敏感度为89.3%，特异度为93.1%，YI＝0.824。增加检测指标，不能改善诊断性能。结论：单独检测时，以CysC或eGFRCysC对LN的诊断性能最高，其次为C1q。联合检测以C1q +CysC+c-aGFR三项为最优，若出于经济考虑，C1q+CysC联合检测也可达到相近于最优的诊断性能。

Abstract:

Objective：To investigate the value of serum urea（Urea），creatinine（Creat），cystatin C（CysC），estimated glomerular filtra－tion rate（eGFRCysC and c-aGFR），and complement component 1q（C1q） in the diagnosis of lupus nephritis（LN）. Methods：A total of 560 patients with systemic lupus erythematous（SLE） who were admitted to our hospital from March to December，2017 were enrolled as subjects. In those patients，339 had the kidney uninvolved（SLE group），150 active LN（LNA group），and 71 inactive LN （LNI group）. Additionally，360 healthy individuals were used as controls（HC group）. Serum levels of Urea，Creat，Urea/Creat，CysC，and C1q were measured. eGFRCysC and c-aGFR were calculated to evaluate the diagnostic performance of these indices for LN. Results：There was a significant difference in each index between these groups（all P=0.000）. Spearman correlation analysis showed that C1q was not correlated with Creat，Urea，or Urea/Creat（r=0.046，P=0.160；r=0.011，P=0.748；r=-0.011，P=0.743），but negatively correlated with CysC，eGFRCysC，and c-aGFR（r=-0.183，P=0.000；r=-0.183，P=0.000；r=-0.075，P=0.023）. CysC was positively correlated with Urea and Creat（r=0.309，P=0.000；r=0.382，P=0.000），negatively correlated with c-aGFR（r=-0.430，P=0.000），but not correlated with Urea/Creat（r=0.003，P=0.927）. The analysis of the diagnostic performance of individual index for LN by re－ceiver operating characteristic curve revealed that eGFRCysC and CysC had the largest area under the curve（AUC=0.891，0.890），followed by C1q，Urea，Urea/Creat，c-aGFR，and Creat（AUC=0.804，0.660，0.630，0.547，0.501）. The maximum diagnostic performance was achieved by combined measurement of C1q，CysC，and c-aGFR，with a sensitivity of 89.3% and a specificity of 93.1%（AUC=0.962，YI=0.824）. The diagnostic performance was not further im－proved by increasing the number of indices. Conclusion：For the diagnosis of LN by a single index，CysC or eGFRCysC has the best diagnostic performance，followed by C1q. For the diagnosis by multiple indices，C1q，CysC，and c-aGFR are the optimal combination. Combined measurement of C1q and CysC is a more economic approach without comprising the optimal diagnostic performance.