The liver is situated behind the lower right rib cage, under the diaphragm. Therefore, it is typically only palpable during inhalation. Its border feels sharp but flexible in the normal state. It has the shape of an elongated triangle across, from the right hypochondrium to the epigastric hollow. The liver is the largest organ of the human body, weighing between 1 and 2 kg, depending on body weight. The liver is reddish-brown in color due to its high concentration in ferric pigments. Its smooth surface consists of a very fine capsule (<1 mm on histological sections), called Glisson’s capsule.

Patients with cirrhosis are at risk of developing esophageal varices (EV). Recently, criteria based on elastography and platelet count (Baveno VI criteria) have been adopted and included in practice guidelines to circumvent screening endoscopy (EGD). However, elastography measurement is not widely available. The aim of the study was to determine predictive factors excluding liver stiffness in order to predict high-risk EV in patients with compensated cirrhosis. Retrospective chart review of all compensated cirrhotic adult patients who underwent screening EGD at Saint-Luc Hospital between 01/2014 and 12/2016. Patients with decompensated cirrhosis (ascites, hepatic encephalopathy, jaundice), past history of EV/TIPS or liver transplantation, acute upper gastointestinal bleeding, acute alcoholic hepatitis were excluded. High-risk EV were defined as medium or large EV and/or presence of red wale signs. Splenomegaly was defined as a spleen ≥ 13 cm on imaging. Thrombopenia was defined as a platelet count < 150 x 109/L. Baseline characteristics, laboratory values and EGD findings were analyzed. A total of 463 patients were included. The median (IQR) age was 60.2 (13.4) years and the majority were males (n=289; 62.4%). The most frequent causes of liver disease were chronic hepatitis C infection (n=117; 25.3%) and non-alcoholic steatohepatitis (n=112; 24.2%). At screening EGD, the median (IQR) MELD score was 7.5 (2.7), the median (IQR) platelet count was 154 (87) x 109/L, and 203 (43.8%) patients had splenomegaly. A total of 45 (9.7%) patients had high-risk EV at screening EGD. In multivariate analysis adjusting for age and gender, the following variables were predictive of high-risk EV: thrombopenia (OR 4.7; p= 0.001) and splenomegaly (OR 3.8; p=0.001). MELD > 6 score was not predictive of high-risk EV (OR 1.7; p=0.337). Among the 172 patients having normal spleen size and platelet count ≥ 150 x 109, only 2 (1.2%) patients had high-risk EV at screening EGD. The presence of splenomegaly and/or thrombopenia had a sensitivity of 95.6% and a negative predictive value of 98.8% for the presence of high-risk EV at screening endoscopy. In compensated cirrhosis, the combination of normal spleen size and normal platelet count translates into a 98.8% chance of absence of high-risk EV at screening EGD. Using these criteria, 172 (37.1%) screening EGD could have been avoided in our population. Therefore, the use of elastography is not mandatory in the decision of recommending screening EGD in patients with compensated cirrhosis. None

Two algorithms based on sequential measurements of liver and spleen stiffness using two-dimensional shearwave elastography (2D-SWE) have been recently proposed to estimate clinically significant portal hypertension (hepatic venous pressure gradient [HVPG] ≥10 mm Hg) in patients with cirrhosis, with excellent diagnostic accuracy.To validate externally these algorithms in a large cohort of patients with cirrhosis.One hundred and ninety-one patients with stable cirrhosis (Child-Pugh class A 39%, B 29% and C 31%) who underwent liver and spleen stiffness measurements using 2D-SWE at the time of HVPG measurement were included. Diagnostic accuracy of the 2 algorithms was assessed by calculating sensitivity, specificity, positive and negative predictive values.The first algorithm, using liver stiffness <16.0 kilopascals (kPa) and then spleen stiffness <26.6 kPa, was used to rule-out HVPG ≥10 mm Hg. In our population, its sensitivity and negative predictive value were 95% and 63% respectively. The second algorithm, using liver stiffness >38.0 kPa, or liver stiffness ≤38.0 kPa but spleen stiffness >27.9 kPa, was used to rule-in HVPG ≥10 mm Hg. In our population, its specificity and positive predictive value were 52% and 83% respectively. Restricting the analyses to the 74 patients without any history of decompensation of cirrhosis or to the 65 patients with highly reliable liver stiffness measurement did not improve the results.In our population, diagnostic accuracies of non-invasive algorithms based on sequential measurements of liver and spleen stiffness using 2D-SWE were acceptable, but not good enough to replace HVPG measurement or to base clinical decisions.