Jamil Zubia, Malik Maryam, Durrani Asghar Aurangzeb
Department of Medicine, Foundation University Medical College, Foundation University, Islamabad, Pakistan.
Turk J Gastroenterol. 2017 Sep;28(5):347-352. doi: 10.5152/tjg.2017.17090. Epub 2017 Aug 4.
BACKGROUND/AIMS: Endoscopy as a screening modality for esophageal varices is becoming difficult because of its invasiveness, cost, and increased burden of liver cirrhosis. This study aims to determine the diagnostic accuracy of simple and noninvasive markers in detecting esophageal varices.
Four variables (platelet count, portal vein diameter, splenic diameter, and ratio of platelet count to splenic diameter PC/SD] ratio) were studied in 150 cirrhotic patients. Endoscopy was performed to detect esophageal varices in the patients. The diagnostic accuracy of these variables was determined by obtaining area under the receiver operating characteristic (ROC) curve (AUC). The cutoff value of each variable and its sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios were obtained using the Youden index. Pairwise comparison of these variables was performed using the Hanley and McNeil method to determine the most reliable screening tool among them.
The PC/SD ratio was the most reliable indicator for the presence of varices: AUC=0.9 (p<0.0001; cutoff value, ≤1077.42; sensitivity, 88.75%; specificity, 81.43%). The AUC for platelets and splenic diameter was 0.85 (p<0.0001) and 0.77 (p<0.0001), respectively, showing they were also good indicators. The portal vein diameter was not a good predictor for esophageal varices (AUC=0.59). Pairwise comparison of these variables showed that the PC/SD ratio is statistically significant for predicting esophageal varices among these markers (p<0.05).
The PC/SD ratio is found to be the most reliable marker to prognosticate esophageal varices. It is easy to obtain and can be used with other markers to identify the high-risk patients for developing esophageal varices. It will definitely reduce the need for endoscopy as screening purposes and lower the medical expenditures.
背景/目的：由于内镜检查具有侵入性、成本高以及肝硬化负担加重等问题，将其作为食管静脉曲张的筛查方式正变得困难。本研究旨在确定简单且无创的标志物检测食管静脉曲张的诊断准确性。
对150例肝硬化患者研究了四个变量（血小板计数、门静脉直径、脾直径以及血小板计数与脾直径之比[PC/SD比值]）。对患者进行内镜检查以检测食管静脉曲张。通过获取受试者操作特征（ROC）曲线下面积（AUC）来确定这些变量的诊断准确性。使用尤登指数获得每个变量的截断值及其敏感性、特异性、阳性和阴性预测值以及阳性和阴性似然比。使用汉利和麦克尼尔方法对这些变量进行两两比较，以确定其中最可靠的筛查工具。
PC/SD比值是静脉曲张存在的最可靠指标：AUC = 0.9（p < 0.0001；截断值，≤1077.42；敏感性，88.75%；特异性，81.43%）。血小板和脾直径的AUC分别为0.85（p < 0.0001）和0.77（p < 0.0001），表明它们也是良好指标。门静脉直径不是食管静脉曲张的良好预测指标（AUC = 0.59）。这些变量的两两比较表明，PC/SD比值在预测这些标志物中的食管静脉曲张方面具有统计学意义（p < 0.05）。
发现PC/SD比值是预测食管静脉曲张最可靠的标志物。它易于获取，可与其他标志物一起用于识别发生食管静脉曲张的高危患者。这肯定会减少作为筛查目的的内镜检查需求并降低医疗费用。
