Refining the Baveno VII criteria for clinically significant portal hypertension: an individual-patient data meta-analysis.

Background And Aims
Clinically significant portal hypertension (CSPH) in compensated advanced chronic liver disease (cACLD) has therapeutic consequences. CSPH may be assessed with the Baveno VII criteria with lower performance in patients with obesity (body mass index ≥30 kg/m2) and metabolic dysfunction-associated steatotic liver disease (MASLD). The ANTICIPATE±NASH models predict the risk of CSPH. We aimed to validate Baveno VII criteria and refine them with the ANTICIPATE±NASH models.Methods
Systematic review of studies validating Baveno VII criteria of CSPH (hepatic venous pressure gradient as reference) with search strategy of "CSPH" (AND) "Baveno VII", from Baveno VII consensus until June 2024. A meta-analysis of Baveno VII criteria (ruling in: LSM (liver stiffness measurement) ≥25 kPa and ruling out: LSM ≤15 kPa + platelets ≥150x109/L) was performed. Using a risk threshold of CSPH by the ANTICIPATE±NASH for a positive predictive value (PPV) of ≥90% was explored. Individual patient data was used to assess model performance by center.Results
Five studies with 1433 cACLD patients (CSPH 34% to 62%) of different etiologies were identified. LSM ≥25 kPa had an excellent PPV (92%) pooled by studies and etiologies, except MASLD with obesity. A ≥75% risk of CSPH by the ANTICIPATE±NASH models improved PPV to 95%, including MASLD with obesity (PPV 0.67 to 0.83; p<0.001). The pooled NPV for ruling out was 99% for all etiologies. ANTICIPATE±NASH showed an excellent performance across centers.Conclusion
Baveno VII criteria for CSPH adequately classify patients across etiologies, except MASLD with obesity. Using a ≥75% risk threshold by ANTICIPATE models to detect CSPH improves global performance, including MASLD with obesity, supporting it can be a simpler way of predicting CSPH in clinical practice.Impact And Implications
This systematic review and meta-analysis confirm the validity for ruling out and ruling in CSPH in cACLD patients with the Baveno VII criteria. Using a threshold of ≥75% of the ANTICIPATE±NASH models, the global performance for detecting CSPH improves regardless of etiology. This represents a very practical approach for general hepatologists to select patients for prophylactic β-blocker therapy as its calculation relies on BMI, liver stiffness and platelet count with an online calculator.