人群调查中尿白蛋白检测方法的探讨

Evaluation measurement of urinary albumin in population survey

目的探讨人群调查中尿白蛋白标本留取及检测方法。方法选取659名北京市居民，收集其24h尿测定UAER；并留取次日随机尿和晨尿，分别采用尿ACR及半定量尿白蛋白试纸法测定尿白蛋白。以24hUAER作为标准，建立应用两种方法检测晨尿和随机尿白蛋白的ROC曲线，比较敏感度、特异度及ROC曲线下面积。结果晨尿和随机尿ACR分别为9．36（5．12～33．29）mg／g及11．29（6．34～41．29）mg／g，组间比较差异无统计学意义（t=-1．382，P〉0．05），相关分析显示两者高度相关（r=0．932，P〈0．01）。晨尿和随机尿ACR与24hUAER具有高度相关性，相关系数分别为0．853及0．874，P均〈0．01。随机尿ACR诊断白蛋白尿的敏感度为77．9％，特异度为91．0％，晨尿ACR诊断白蛋白尿的敏感度为78．4％，特异度为95．7％。随机尿白蛋白试纸法诊断白蛋白尿的敏感度为90．3％，特异度为41．1％，特异度明显低于ACR法。随机尿与晨尿ACR的ROC曲线下面积分别为0．918±0．012及0．929±0．015，组间比较差异无统计学意义（X^2=2．13，P〉0．05）。随机尿白蛋白试纸法ROC曲线下面积0．661±0．021，低于随机尿ACR（X^2=248．41，P〈0．01）。结论随机尿的ACR兼具简便及准确的特点，是人群调查中诊断白蛋白尿的良好指标。

Objective To evaluate the spot urine sample collection method and value of urinary albumin measurement in population survey. Methods Six hundred and fifty-nine Beijing residents were requested to collect 24 h urine for detection of UAER, as well as random spot urine samples and morning urine samples in the next day. Rapid semi-quantitative urinary albumin-specific dipstick and ACR were measured in each spot urine specimen. The 24 h UAER was measured as golden standard to generate ROC curves and evaluate the sensitivity, specificity and AUC of each method. Results The value of ACR in the morning spot urine samples and random spot urine samples were 9. 36 （5.12-33.29） mg/g and 11.29 （6. 34- 41.29） mg/g respectively and there was no significant difference between these two groups （t = - 1. 382, P 〉 0. 05）. The correlation was significant in the two groups （ r = 0. 932, P 〈 0. 01 ） . The correlation coefficient between ACR in the morning spot urine samples and UAER was 0. 853 （ P 〈 0. 01 ） . The correlation coefficient between ACR in the random spot urine samples and UAER was 0. 874 （P 〈0. 01 ）. The sensitivity and specificity of ACR for diagnosis of albuminuria in the random urine samples were 77. 9% and 91.0%. The sensitivity and specificity of ACR for diagnosis of albuminuria in the morning urine samples were 78.4% and 95.7%. Concerning the semi-quantitative urinary albumin-specific dipstick, sensitivity and specificity were 90. 3% and 41.1%, respectively. The specificity was much lower than that of ACR. The area under the ROC curves of ACR in the random urine specimens and the morning urine specimens was 0. 918 ＋ 0. 012, 0. 929 ＋ 0. 015, respectively. There was no statistical difference between these two groups （X^2 =2. 13, P 〉 0. 05 ）. The area under the ROC curves of semi-quantitative urinary albumin-specific dipstick in the random urine specimens was 0. 661± 0. 021, lower than that of ACR （X^2 = 248. 41, P 〈0. 01）. Conclusion Measurement of ACR in random urine samples is a reasonable method with simplicity and accuracy for the detection of albuminuria in general population screening program.