Abstract
OBJECTIVE: Studies have presented conflicting results regarding the accuracy of ultrasonography (US) for diagnosing portal hypertension (PH). We sought to identify evidence in the literature regarding the accuracy of US for assessing PH in patients with liver cirrhosis. MATERIALS AND METHODS: We conducted a systematic review by searching databases, including MEDLINE, EMBASE, and the Cochrane Library, for relevant studies. RESULTS: A total of 14 studies met our inclusion criteria. The US indices were obtained in the portal vein (n = 9), hepatic artery (n = 6), hepatic vein (HV) (n = 4) and other vessels. Using hepatic venous pressure gradient (HVPG) as the reference, the sensitivity (Se) and specificity (Sp) of the portal venous indices were 69-88% and 67-75%, respectively. The correlation coefficients between HVPG and the portal venous indices were approximately 0.296-0.8. No studies assess the Se and Sp of the hepatic arterial indices. The correlation between HVPG and the hepatic arterial indices ranged from 0.01 to 0.83. The Se and Sp of the hepatic venous indices were 75.9-77.8% and 81.8-100%, respectively. In particular, the Se and Sp of HV arrival time for clinically significant PH were 92.7% and 86.7%, respectively. A statistically significant correlation between HVPG and the hepatic venous indices was observed (0.545-0.649). CONCLUSION: Some US indices, such as HV, exhibited an increased accuracy for diagnosing PH. These indices may be useful in clinical practice for the detection of significant PH.