Bartlett R C, Zern D A, Ratkiewicz I, Tetreault J Z
Department of Pathology and Laboratory Medicine, Hartford Conn Hospital 06115-5037.
Arch Pathol Lab Med. 1994 Nov;118(11):1096-101.
We assessed the ability of reagent strip screening to predict the finding of blood cells and bacteria using an automated urinalysis workstation (The Yellow IRIS, International Remote Imaging Systems, Chatsworth, Calif) in 427 specimens submitted for urine culture. The sensitivities of leukocyte esterase, hemoglobin, and nitrite detection on reagent strips were 71.9%, 70.8%, and 56.7%, respectively, at 5 or more white blood cells per high-power field, at 3 or more red blood cells per high-power field and bacteria observed using the IRIS. Screening results for leukocyte esterase associated with negative results using the IRIS for white blood cells represented mostly false-positive screening test results based on chart review. Positive screening test results for hemoglobin associated with negative results obtained with the IRIS for red blood cells consisted of equal numbers of false-positive screening test results and IRIS test results based on chart review. A common screening algorithm using a combination of these three reagent strip variables exhibited a false-negative rate of 30.1%: review of medical records found clinical evidence of urinary tract infection in 14 patients and genitourinary or renal disease or hypertension in another 13 patients. Adding more variables to the algorithm increases the sensitivity and decreases the specificity. Both microscopic examination and reagent strip testing of urine are necessary for the detection of abnormalities associated with disease.
我们使用自动尿液分析工作站（黄色IRIS，国际远程成像系统公司，加利福尼亚州查茨沃思），对427份送检尿培养的标本评估了试带筛查预测血细胞和细菌检测结果的能力。在每高倍视野有5个或更多白细胞、3个或更多红细胞以及使用IRIS观察到细菌的情况下，试带上白细胞酯酶、血红蛋白和亚硝酸盐检测的敏感度分别为71.9%、70.8%和56.7%。根据图表回顾，与使用IRIS检测白细胞呈阴性结果相关的白细胞酯酶筛查结果大多为假阳性筛查试验结果。与使用IRIS检测红细胞呈阴性结果相关的血红蛋白阳性筛查试验结果，根据图表回顾，假阳性筛查试验结果和IRIS检测结果数量相等。使用这三个试带变量组合的常见筛查算法显示假阴性率为30.1%：病历审查发现14例患者有尿路感染的临床证据，另有13例患者有泌尿生殖系统或肾脏疾病或高血压。在算法中增加更多变量会提高敏感度并降低特异度。尿液的显微镜检查和试带检测对于发现与疾病相关的异常情况均很必要。
