Baveno VIII rules

Non-invasive assessment of compensated advanced chronic liver disease (cACLD), clinically significant portal hypertension (CSPH), carvedilol or NSBB treatment and the need for screening endoscopy. It uses liver stiffness by FibroScan, platelet count, BMI and spleen stiffness, following the Baveno VIII consensus (2026).

Stage Compensated: no current or past decompensation Recompensated cirrhosis: past decompensation, cause removed or controlled

MeasurementsLiver stiffness (LSM) kPa, FibroScan (VCTE), fastingPlatelet count × 10⁹/LBMI kg/m², needed for MASLDSpleen stiffness (SSM) kPa, FibroScan 100 Hz, optionalPrevious LSM kPa, optional, to assess change

Cause of liver disease Hepatitis B Hepatitis C (untreated or cured) Alcohol MASLD Other (for example cholestatic or autoimmune)

Main cause of liver disease Still active Removed or suppressed (for example HCV cured, HBV DNA suppressed, sustained abstinence)

Carvedilol or non-selective beta blocker (NSBB) Taking carvedilol, propranolol or nadolol Cannot take one (contraindication or intolerance) Not yet decided

Hepatocellular carcinoma (HCC) No HCC BCLC 0 to A, no nodule over 5 cm in the right lobe Other BCLC 0 to C Macrovascular invasion

Enter liver stiffness, platelet count and the cause of liver disease.

Changes from Baveno VII

CSPH can now also be diagnosed by an ANTICIPATE or ANTICIPATE-NASH probability of 75% or more, by NICER, or by spleen stiffness over 55 kPa, alongside LSM ≥25 kPa. This replaces the ANTICIPATE 60% bands and the 21 and 50 kPa spleen stiffness cut-offs for viral hepatitis.
If carvedilol or an NSBB cannot be used, endoscopy can be avoided when LSM is under 20 kPa with platelets ≥150 × 10⁹/L, or spleen stiffness is under 40 kPa. If there are no varices, endoscopy is repeated every 2 or 3 years.
Non-invasive tests are repeated every 12 months in cACLD, including when results are indeterminate.
For change in LSM, a confirmed relative change of 30% or more (alcohol and MASLD) replaces the Baveno VII definition of a clinically significant decrease, and the most recent LSM guides decisions.
The Baveno VII rule for discharge after hepatitis C cure (LSM under 12 kPa and platelets over 150) is not part of Baveno VIII.
New statements cover recompensated cirrhosis, HCC, and LSM under 10 kPa with signs of portal hypertension (consider vascular liver disease).

Statements used

Summarised from the Baveno VIII consensus. Numbers are Baveno VIII statement numbers, with the level of evidence (LoE, Oxford 1 to 5) and strength of recommendation. Check the paper for the full wording.

cACLD and prognosis

1.3: LSM under 10 kPa rules out cACLD if there are no other clinical or imaging signs; 10 to 15 kPa suggests cACLD; over 15 kPa strongly suggests it. (LoE 3, strong)
1.4: LSM should be done fasting. An index LSM of 10 kPa or more should be repeated (LoE 2, strong) or complemented by another validated fibrosis test (LoE 3, weak).
1.5: LSM under 10 kPa carries a negligible 3-year risk (1% or less) of decompensation and liver-related death. (LoE 1)
1.6: LSM under 10 kPa with clinical or imaging signs of portal hypertension should raise suspicion of vascular liver disease. (LoE 3, strong)
1.7: Refer patients with cACLD to a liver specialist. (LoE 3, strong)
1.8: Rule of 5: LSM of 10, 15, 20 and 25 kPa marks progressively higher risks of decompensation and liver-related death, whatever the cause. (LoE 2)
1.9: In cACLD, repeat non-invasive tests for CSPH every 12 months. (LoE 3, strong)

Clinically significant portal hypertension

1.14: Estimate the probability of CSPH with ANTICIPATE, or ANTICIPATE-NASH in MASLD with obesity (BMI 30 or more) (LoE 1, strong), or NICER (LoE 2, strong).
1.15: LSM ≤15 kPa with platelets ≥150 × 10⁹/L rules out CSPH (negative predictive value over 90%). (LoE 1, strong)
1.16a: A CSPH probability of 75% or more by ANTICIPATE or ANTICIPATE-NASH (LoE 1, strong), or NICER (LoE 2, strong), establishes CSPH.
1.16b: In viral and/or alcohol-related cACLD, and MASLD with BMI under 30, LSM ≥25 kPa establishes CSPH. (LoE 1, strong)
1.16c: Spleen stiffness (100 Hz) over 55 kPa establishes CSPH. (LoE 2, strong)
1.13: Spleen stiffness may further improve the classification of CSPH. (LoE 2)
1.17: If results are indeterminate, re-evaluating at 12 months may be the preferred approach. (LoE 4, weak)

Carvedilol, NSBBs and screening endoscopy

3.15: With CSPH, carvedilol (preferred) or an NSBB is recommended to prevent decompensation and improve survival. (LoE 2, strong)
3.16: Without CSPH, carvedilol or NSBBs should not be used to prevent decompensation. (LoE 2, strong)
3.18: Patients on carvedilol or an NSBB do not need screening endoscopy, as it would not change management. (LoE 2)
1.19: Patients who cannot take an NSBB should have screening endoscopy, unless LSM is under 20 kPa with platelets ≥150 × 10⁹/L, or spleen stiffness is under 40 kPa. (LoE 1, strong)
3.19: If carvedilol or an NSBB cannot be used and large oesophageal varices or GOV1 are found, band ligation is recommended. (LoE 1, strong)
1.20: If there are no varices at the first endoscopy, repeat it every 2 years if the cause is still active, or every 3 years if it has been removed or suppressed, unless the criteria in 1.19 are met later. (LoE 4, strong)

Hepatocellular carcinoma

1.21: With macrovascular invasion, non-invasive tests for CSPH and high-risk varices may not apply. (LoE 3, weak)
1.22: In BCLC 0 to C, LSM ≤15 kPa with platelets ≥150 × 10⁹/L rules out CSPH, and endoscopy may be avoided. (LoE 3, weak)
1.23: In BCLC 0 to A without a nodule over 5 cm in the right lobe, an ANTICIPATE or ANTICIPATE-NASH probability of 75% or more may be highly suggestive of CSPH. (LoE 3, weak)

Change in liver stiffness

1.10: Changes in LSM carry prognostic information, but the most recent value should guide decisions. (LoE 2, strong)
2.3: In alcohol-related cACLD, a confirmed relative change in LSM of 30% or more goes with a matching change in the risk of liver-related events. (LoE 5)
2.9: In MASLD-related cACLD, a confirmed relative change of 30% or more is a clinically meaningful change in risk, particularly when it crosses 15 kPa. (LoE 2)
2.6: In MASLD-related cACLD without CSPH, a confirmed relative fall of 30% or more may indicate regression by at least one fibrosis stage. (LoE 3)

Recompensated cirrhosis

7.4: Recompensation needs all of the following for more than 6 months: the cause removed, controlled or suppressed; ascites resolved on imaging off diuretics; no hepatic encephalopathy off lactulose, rifaximin or LOLA; no recurrent variceal bleeding; and liver tests consistent with Child-Pugh A5 or A6.
7.19: LSM under 10 kPa (LoE 2, weak), or LSM under 15 kPa with spleen stiffness under 25 kPa (LoE 4, weak), may rule out CSPH; LSM over 25 kPa may rule it in (LoE 4, weak).
7.16: HVPG under 10 mmHg off carvedilol or NSBBs confirms resolution of CSPH. (LoE 2)
7.17 and 7.18: Varices and collaterals can persist despite resolution of CSPH (LoE 3). Carvedilol or an NSBB may be stopped only if resolution of CSPH is confirmed (LoE 3, weak).
7.13 and 7.14: Continue hepatology follow-up every 6 months (LoE 3, strong) and HCC surveillance (LoE 2, strong).

Models

ANTICIPATE: logit = −6.3320165 + 3.1001014 × ln(LSM) − 0.020481545 × platelets, with platelets above 150 entered as 150.
ANTICIPATE-NASH: logit = −3.9529402 + 2.2835809 × ln(LSM) − 0.033777725 × BMI − 0.014490895 × platelets.
Probability of CSPH = 1 ÷ (1 + e^−logit). LSM is in kPa and platelets in × 10⁹/L.
NICER uses spleen stiffness, LSM, platelets and BMI on a published points chart, and is not calculated here.

References

Berzigotti A, Mandorfer M, Pons M, et al; Baveno VIII Faculty. Baveno VIII – Advancing consensus in portal hypertension. J Hepatol. 2026. Accepted pre-proof, 31 July 2026.doi:10.1016/j.jhep.2026.07.030

Model formulas as published in: Rabiee A, Deng Y, Ciarleglio M, et al. Noninvasive predictors of clinically significant portal hypertension in NASH cirrhosis: validation of ANTICIPATE models and development of a lab-based model. Hepatol Commun. 2022;6(12):3324-3334.Free full text