Department of Gastroenterology, University of Medicine and Pharmacy TimiŞoara, 300482 Timişoara, Romania.
World J Gastroenterol. 2011 May 7;17(17):2206-10. doi: 10.3748/wjg.v17.i17.2206.
To determine if liver stiffness (LS) measurements by means of transient elastography (TE) correlate with the presence of significant esophageal varices (EV) and if they can predict the occurrence of variceal bleeding.
We studied 1000 cases of liver cirrhosis divided into 2 groups: patients without EV or with grade 1 varices (647 cases) and patients with significant varices (grade 2 and 3 EV) (353 cases). We divided the group of 540 cases with EV into another 2 subgroups: without variceal hemorrhage (375 patients) and patients with a history of variceal bleeding (165 cases). We compared the LS values between the groups using the unpaired t-test and we established cut-off LS values for the presence of significant EV and for the risk of bleeding by using the ROC curve.
The mean LS values in the 647 patients without or with grade 1 EV was statistically significantly lower than in the 353 patients with significant EV (26.29 ± 0.60 kPa vs 45.21 ± 1.07 kPa, P < 0.0001). Using the ROC curve we established a cut-off value of 31 kPa for the presence of EV, with 83% sensitivity (95% CI: 79.73%-85.93%) and 62% specificity (95% CI: 57.15%-66.81%), with 76.2% positive predictive value (PPV) (95% CI: 72.72%-79.43%) and 71.3% negative predictive value (NPV) (95% CI: 66.37%-76.05%) (AUROC 0.7807, P < 0.0001). The mean LS values in the group with a history of variceal bleeding (165 patients) was statistically significantly higher than in the group with no bleeding history (375 patients): 51.92 ± 1.56 kPa vs 35.20 ± 0.91 kPa, P < 0.0001). For a cut-off value of 50.7 kPa, LS had 53.33% sensitivity (95% CI: 45.42%-61.13%) and 82.67% specificity (95% CI: 78.45%-86.36%), with 82.71% PPV (95% CI: 78.5%-86.4%) and 53.66% NPV (95% CI: 45.72%-61.47%) (AUROC 0.7300, P < 0.0001) for the prediction of esophageal bleeding.
LS measurement by means of TE is a reliable noninvasive method for the detection of EV and for the prediction of variceal bleeding.
确定通过瞬时弹性成像(TE)测量的肝硬度(LS)是否与显著食管静脉曲张(EV)的存在相关,以及是否可以预测静脉曲张出血的发生。
我们研究了 1000 例肝硬化病例,分为 2 组:无 EV 或 1 级静脉曲张的患者(647 例)和有显著 EV(2 级和 3 级 EV)的患者(353 例)。我们将 540 例有 EV 的患者分为另 2 个亚组:无静脉曲张出血的患者(375 例)和有静脉曲张出血史的患者(165 例)。我们使用配对 t 检验比较各组的 LS 值,并使用 ROC 曲线确定存在显著 EV 和出血风险的 LS 值截断值。
无或 1 级 EV 的 647 例患者的平均 LS 值明显低于有显著 EV 的 353 例患者(26.29±0.60 kPa 比 45.21±1.07 kPa,P<0.0001)。使用 ROC 曲线,我们确定了存在 EV 的截断值为 31 kPa,其敏感性为 83%(95%CI:79.73%-85.93%),特异性为 62%(95%CI:57.15%-66.81%),阳性预测值(PPV)为 76.2%(95%CI:72.72%-79.43%),阴性预测值(NPV)为 71.3%(95%CI:66.37%-76.05%)(AUROC 0.7807,P<0.0001)。有静脉曲张出血史的患者(165 例)的平均 LS 值明显高于无出血史的患者(375 例):51.92±1.56 kPa 比 35.20±0.91 kPa,P<0.0001)。对于截断值为 50.7 kPa,LS 的敏感性为 53.33%(95%CI:45.42%-61.13%),特异性为 82.67%(95%CI:78.45%-86.36%),阳性预测值为 82.71%(95%CI:78.5%-86.4%),阴性预测值为 53.66%(95%CI:45.72%-61.47%)(AUROC 0.7300,P<0.0001),用于预测食管出血。
通过 TE 测量 LS 是一种可靠的非侵入性方法,可用于检测 EV 和预测静脉曲张出血。