Baveno VI and VII Criteria Are Not Suitable for Screening for Large Varices or Clinically Significant Portal Hypertension in Patients With Hepatocellular Carcinoma

Manon Allaire; Bertille Campion; Alix Demory; Edouard Larrey; Mathilde Wagner; Marika Rudler; Charles Roux; Lorraine Blaise; Nathalie G. Carrie; Dominique Thabut


Aliment Pharmacol Ther. 2023;58(3):346-356. 

In This Article

Abstract and Introduction


Background: Baveno VI and VII criteria are used in patients with cirrhosis to rule out large size oesophageal varices (EV) and rule in/out clinically significant portal hypertension (CSPH).

Aim: To evaluate their diagnostic performance in these patients.

Methods: We retrospectively included all patients with Child-Pugh A cirrhosis and HCC who had endoscopy, liver stiffness measurement (LSM) and platelet count within 6 months. They were classified according to the BCLC stage. Favourable Baveno VI criteria were defined by LSM < 20 kPa and platelets > 150 G/L (to rule out large EV), favourable Baveno VII criteria if LSM ≤ 15 kPa and platelets ≥ 150 G/L (to rule out CSPH, which was defined by a HVPG ≥ 10 mm Hg.

Results: We included 185 patients; 46% were BCLC-0/A, 28% BCLC-B and 26% BCLC-C. EV were present in 44% (23% large), and HVPG ≥ 10 mm Hg in 42% (mean 8 mm Hg). In patients with favourable Baveno VI criteria, 8% of the whole cohort (Se 93%, NPV 92%), 11% of BCLC-0-A (Se 89%, NPV 89%) and 10.0% of BCLC-C patients (Se 91%, NPV 90%) had large EV. Among patients with HVPG < 10 mm Hg, 6% had large EV and 17% small. CSPH was present in 23% of patients with favourable Baveno VII criteria among the whole cohort, and in 25% of those with BCLC-0/A. The specificity of LSM ≥ 25 kPa to rule in CSPH was 48%.

Conclusions: Favourable Baveno VI criteria are not appropriate to rule out the presence of high-risk EV, or Baveno VII criteria to rule CSPH in/out in patients with HCC.


Portal hypertension (PHT) and hepatocellular carcinoma (HCC) are major complications of cirrhosis and are closely related. Appropriate screening and management of PHT in HCC patients is a major issue as PHT directly impacts the choice of HCC treatments and patients' prognosis. To evaluate the presence of PHT, the hepatic venous pressure gradient (HVPG) measurement remains the gold standard.[1,2] A HVPG of ≥10 mm Hg indicates the presence of clinically significant portal hypertension (CSPH) in patients with viral and alcohol-related cirrhosis. CSPH exposes patients to liver decompensation, including ascites, pleural effusion, hepatic encephalopathy and acute variceal bleeding (AVB). Furthermore, CSPH is associated with a lower overall survival (OS) and a higher rate of complications in case of locoregional treatment for HCC such as liver surgery, ablation and transarterial chemoembolisation (TACE).[3–5] Prevention of PHT-related complications in non-decompensated patients relies on non-selective beta-blockers (NSBBs) in patients with large size oesophageal varices (EV) in order to prevent bleeding.[6] More recently, at Baveno VII consensus, NSBBs were recommended to be used in patients with CSPH to prevent liver decompensation and the occurrence of ascites.[1,7] The clinical utility of these criteria is to appropriately rule in or rule out patients with large EV or CSPH, without missing an opportunity to indicate NSBBs. In non-HCC patients, an algorithm including platelet count and liver stiffness measurement (LSM) value by transient elastometry (TE) allows to rule out large EV (favourable Baveno VI criteria: LSM < 20 kPa and platelet count >150 × 109/L), without endoscopy, restricting endoscopy to patients without favourable Baveno VI criteria. Similarly, the non-invasive assessment of CSPH is currently investigated. From a simple point, in patients without HCC, all patients with EV, whatever their size, ascites and collateral abdominal circulation exhibit CSPH,[1] whereas the absence of CSPH, per definition, precludes the presence of EV. Moreover, algorithms including again LSM and platelets count allow to rule in/out CSPH with a good reliability. A LSM ≥ 25 kPa is sufficient to rule in CSPH in patients with virus and/or alcohol-compensated advanced chronic liver disease (cALCD) and non-obese NASH (and a treatment with NSBBs is consequently recommended in those patients). On the other hand, the combination of a LSM by TE ≤ 15 kPa and a platelet count ≥150 × 109/L (Favourable Baveno VII criteria) allows to rule out CSPH (and the need for NSBBs therapy). In summary, the Baveno VI (to rule out large EV) and VII (to rule in/out CSPH) criteria provide clinicians with a simple and actionable risk assessment for patients with liver cirrhosis.

However, one of the most common complications of cirrhosis is the development of HCC. Whether all those assertions apply to patients with HCC remains to be determined. If a HVPG < 10 mm Hg allows to rule out CSPH in patients with HCC remains unknown. The interpretation of LSM and platelet count may also vary with HCC: indeed, in a series of Child-Pugh A patients with potentially resectable HCC, LSM was feasible in 88% and the correlation of LSM with HPVG was lower than in patient with cirrhosis, suggesting that factors related to tumour location might interfere with LSM, that is, the size and the localisation.[8] Platelets level may be increased by HCC presence,[9,10] thus decreasing the sensitivity of those criteria to rule out large EV. To date, no studies combining LSM and platelet count as proposed by the Baveno VI and VII consensus are available to evaluate the risk of large size EV and CSPH in patients with early HCC but also in patients with intermediate and advanced HCC.

Hence, the objectives of the present study were to assess, in patients with HCC: (1) if Baveno VI criteria are accurate to rule out large EV, (2) if a HVPG < 10 mm Hg allows to rule out CSPH and (3) if Baveno VII criteria allow to rule in/out CSPH as defined in Baveno VII consensus.