目的探讨真菌(1,3)-β-D-葡聚糖检测(G试验)在侵袭性真菌感染患者中临床应用情况。方法回顾性分析湖南省人民医院、中南大学湘雅二医院和中南大学湘雅三医院2014年7月-2015年12月间进行G试验检测的2 588例患者血清标本的阳性率、阴性率、阳性似然比及阴性似然比,分析其临床相关因素。结果 2 588例患者中具有侵袭性真菌感染高危因素者1 678例,占64.84%,无高危因素者910例,占35.16%。在有高危因素患者中G试验检测真阳性169例,假阳性420例,真阴性904例,假阴性185例。阳性似然比为1.51,阴性似然比为0.76。G试验应用于存在血液系统恶性疾病、移植状态、重症感染、自身免疫病、慢性肾功能不全、恶性肿瘤、糖尿病、肝硬化、慢性呼吸系统疾病等9种具有侵袭性真菌感染高危因素的患者检测时阳性似然比分别为1.23、1.77、1.21、1.55、1.31、0.68、1.8、1.97、2.2;阴性似然比分别为0.91、0.64、0.86、0.66、0.75、1.18、0.72、0.64、0.76。结论 G试验临床应用价值有待考证,尤其是在患有血液系统恶性疾病、重症感染、恶性肿瘤的患者中,G试验应用价值较低,需改良其方法。 Objective To investigate the clinical application of fungal (1,3) -β-D-glucan test (G test) in patients with invasive fungal infection. Methods The positive rates and negative rates of 2 588 serum samples from Hunan Provincial People’s Hospital, Second Xiangya Hospital of Central South University and Third Xiangya Hospital of Central South University from July 2014 to December 2015 were retrospectively analyzed. Positive likelihood ratio and negative likelihood ratio, analysis of its clinically relevant factors. Results There were 1 678 cases (64.84%) in 2 588 patients with high risk of invasive fungal infection, and 910 cases (35.16%) without high risk factors. In patients with risk factors in the G test to detect true positive 169 cases, 420 cases of false positives, true negative 904 cases, false negative 185 cases. The positive likelihood ratio was 1.51 and the negative likelihood ratio was 0.76. G test is used in 9 patients with risk factors for invasive fungal infections such as hematologic malignancies, transplant status, severe infections, autoimmune diseases, chronic renal insufficiency, malignant tumors, diabetes, cirrhosis, and chronic respiratory diseases When the positive likelihood ratio were 1.23,1.77,1.21,1.55,1.31,0.68,1.8,1.97,2.2; negative likelihood ratios were 0.91,0.64,0.86,0.66,0.75,1.18,0.72,0.64,0.76. Conclusion The clinical value of G test remains to be verified, especially in patients with malignant hematological malignancies, severe infections and malignant tumors. The G test is of low value and need to be improved.