应用受试者特性曲线分析血清胆红素 尿酸和同型半胱氨酸对冠心病诊断的临床价值

目的定量分析冠心病(CHD)患者血尿酸(UA)、胆红素(TBIL)及同型半胱氨酸(Hcy)的临床应用价值。方法收集285名对照者和721例CHD患者,将721例CHD患者分为稳定型心绞痛(SAP)、不稳定型心绞痛(UAP)、Q波心肌梗死(QMI)和非Q波心梗(NQMI)4组,测定285名对照者和721例CHD患者的UA、TBIL及Hcy,应用方差分析和受试者特性(ROC)曲线对不同类型CHD组内、CHD组与对照组之间进行比较与分析。结果①不同类型CHD与对照组相比:各组UA均显著高于对照组(P<0.01);除SAP外,TBIL均显著低于对照组(P<0.01);QMI和NQMI组Hcy显著高于对照组(P<0.01);CHD组Hcy、UA均显著高于对照组(P<0.01);CHD组TBIL显著低于对照组(P<0.01)。不同类型CHD组内比较:TBIL各组间差异无统计学意义(P>0.05)。与SAP组比较:UAP组、QMI和NQMI组,Hcy显著高于SAP组(P<0.01)。与NQMI组相比:QMI组UA和UAP组Hcy差异有统计学意义(P<0.05);SAP组Hcy差异有统计学意义(P<0.01)。②冠心病各组TBIL、UA、Hcy各指标诊断灵敏度偏低,特异性相对较高。TBIL对冠心病诊断的灵敏度31.2%,特异性80%,阳性和阴性似然比分别为1.56、0.86。Hcy对SAP的诊断灵敏度50%,特异性82%,阳性似然比2.78,阴性似然比0.61。UA对CHD的诊断评价指标都较低。③不同类型CHD及CHD组的TBIL面积均大于0.5;Hcy面积除SAP组为0.641外,其他组均小于0.5;UA面积均小于0.5。结论TBIL、UA、Hcy与CHD的发生、发展密切相关,对CHD的临床诊断价值均较低,可能会成为实验室诊断CHD的筛查指标。

Clinical value of ROC curve analysis for serum bilirubin, uric acid and homocysteine in coronary heart disease

Objective To quantitatively analyze the clinical value of analysis for uric acid, bilirubin and homocysteine in patients with coronary heart disease （CHD）. Methods Two hundred and eighty-five healthy controls and 721 CHD patients were recruited. The CHD patients were assigned to four groups as stable angina pectoris （SAP）, unstable angina pectoris （UAP）, Q wave myocardial infarction （QMI） and non-Q wave myocardial infarction （NQMI）. The serum levels of bilirubin （TBIL）, uric acid （UA） and homocysteine （HCY） were determined in all participants, compared among CHD subgroups and between CHD and control groups using analysis of variance and ROC curves. Results （1） Compared with controls, significantly higher level of UA was shown in all CHD subgroups （P〈0.01）, lower level of TBIL in all but the SAP group （P〈0.01）, and significantly higher level of HCY in QMI and NQMI groups （P〈0.01）. Overall, CHD patients showed higher levels of HCY and UA （P〈0.01）, and lower level of TBIL （P〈0.01） than the controls. Among groups of different CHD types： TBIL did not show statistical difference （P〉0.05）, the level of HCY appeared much higher in QMI, NQMI and UAP groups than in the SAP group （P〈0.01）. In comparison with NQMI group, the UA of QMI group （P〈0.05） and the HCY of both UAP （P〈0.05） and SAP （P〈0.01） groups showed statistical differences as compared with NQMI group. （2） In all CHD subgroups, tests for TBIL, UA and HCY appeared to be with lower sensitivity and relatively higher specificity for diagnosis. TBIL yielded a sensitivity of 31.2%, a specificity of 80%, positive and negative likelihood rates of 1.56 and 0.86, respectively. These figures for HCY was 50% in sensitivity, 82% in specificity, 2.78 in positive and 0.61 in negative likelihood rates. All diagnostic parameters with UA appeared unfavorable. （3） The area under curve for TBIL was over 0.5 for all CHD subgroups or the overall CHD cohort. The area under curve for HCY was less than 0.5 except for the SAP group （0.641）. UA showed an area less than 0.5 in all CHD subgroups. Conclusion TBIL, UA and HCY were demonstrated to have a close correlation with the incidence and development of CHD. With considerable clinical value for diagnostic purpose, these measures may be used for laboratory screening of CHD.