To assess feasibility and safety of integrating endoscopic ultrasound (EUS)-guided direct portal venous pressure (PVP) measurement into a one-stop endoscopy and to report endoscopy-relevant diagnostic performance. We prospectively compared PVP with the endoscopic portal pressure gradient (PPG = PVP − hepatic/inferior vena cava [IVC]) for two prespecified endpoints: oesophageal varices and clinically significant portal hypertension.

Prospective single-centre outpatient cohort (15 Mar 2024–15 May 2025). A standardized protocol used conscious sedation with midazolam and fentanyl, left-lateral decubitus, spontaneous breathing, and zeroing at the mid-axillary line. At least three acceptable end-expiratory plateaus were acquired in the portal vein and in a hepatic vein/retrohepatic IVC; technical success required ≥3 accepted measurements per site to compute PPG. Adverse events (AEs) were prospectively attributed to the pressure-measurement step vs EUS-guided liver biopsy (EUS-LB) and graded by AGREE classification. Clinically significant portal hypertension (CSPH) was defined a priori as a composite of ≥3 portal-hypertension signs (unweighted sum) from a literature-adapted list. Primary analyses estimated ROC/AUCs with paired DeLong tests; .632-bootstrap internal validation and decision-curve analysis quantified robustness and net benefit.

Fifty-two patients were analysed (63.5% male; median age 56.5 years). Technical success was 52/52 (100%); median EUS time 35 min [IQR 29–39]. EUS-LB was performed in 31/52 (59.6%). AEs occurred in 5/52 (9.6%): one Grade I epistaxis at insertion and four Grade IIIa post-biopsy bleeds (three clipped; one managed by interventional radiology); 4/5 (80%) were attributable to EUS-LB; none to the pressure-measurement step. For varices, AUCs were 0.981 for PVP vs 0.857 for PPG (paired DeLong p=0.031); Youden cut-offs respectively 23.5 and 8.5 mmHg. For CSPH (composite ≥3 signs), AUCs were 0.840 (PVP) vs 0.671 (PPG) (p=0.0049); cut-offs 27.5 and 10.5 mmHg. Internal .632-bootstrap and decision-curve analyses supported robustness and net benefit, consistently higher for PVP.

EUS-guided direct portal pressure can be embedded in a one-stop endoscopy with 100% technical success and a favourable safety profile, with events almost exclusively related to liver biopsy rather than the haemodynamic step. Absolute PVP provided stronger discrimination than PPG for both varices and CSPH, supporting its pragmatic value to anchor intra-procedural decisions and triage. These findings warrant multicentre confirmation under harmonized protocols (sedation, breathing mode, zero-leveling, replicate rules) and an evaluation of implementation pathways alongside traditional routes.