2025 · LANCET GASTROENTEROLOGY & HEPATOLOGY
Pre-emptive TIPS for gastric variceal bleeding in patients with cirrhosis (GAVAPROSEC): an open-label randomised clinical trial
Cervoni, Jean-Paul, Weil, Delphine, Desmarets, Maxime, Lannes, Adrien, D'Alteroche, Louis, Bouzbib, Charlotte, Larrue, Helene, Lemaitre, Elise, Faure, Stephanie, Latournerie, Mariane, Jezequel, Caroline, Billioud, Claire, Carbonell, Nicolas, Saliba, Faouzi, Tanne, Florence, Hiriart, Jean-Baptiste, Olivier-Hourmand, Isabelle, Khac, Eric Nguyen, Archambeaud, Isabelle, Hilleret, Marie-Noelle, Oberti, Frederic, Elkrief, Laure, Meunier, Lucy, Gerster, Theophille, Rode, Agnes, Bardou-Jacquet, Edouard, Robic, Marie-Angele, Ozenne, Violaine, Sacleux, Sophie-Caroline, Reboux, Noemi, Chermak, Faiza, Calame, Paul, Borentain, Patrick, Thevenot, Thierry, Di Martino, Vincent, Thabut, Dominique, Louvet, Alexandre, Rudler, Marika, Bureau, Christophe, Club Francophone l'Etude l'Hypertension Portale, & GAVAPROSEC Study Grp
- Journal
- LANCET GASTROENTEROLOGY & HEPATOLOGY
- Année
- 2025
- Volume
- 10
- Numéro
- 8
- Pages
- 726-733
- Mois
- AUG
- DOI
- 10.1016/S2468-1253(25)00156-6
Abstract
Background In patients with cirrhosis, variceal glue obliteration is recommended for the treatment of acute fundal variceal bleeding, and for the prevention of rebleeding in combination with non-selective beta blockers (NSBB). We evaluated the efficacy of pre-emptive transjugular intrahepatic portosystemic shunt (p-TIPS) in this setting. Methods This open-label randomised trial was conducted at 17 tertiary centres in France. Patients with cirrhosis and acute fundal variceal bleeding (excluding type 1 gastro-oesophageal varices), who achieved initial haemostasis with endoscopic glue injection and vasoactive therapy, and remained stable for at least 12 h, were randomly assigned to receive either a covered p-TIPS within 72 h or to continue with on-demand glue obliteration sessions combined with NSBB. Randomisation was centralised, stratified by centre, and performed in blocks of four (1:1 ratio). The primary composite endpoint was all-cause mortality or clinically significant rebleeding within 1 year. Analyses were conducted in the modified intention-to-treat (mITT) population. This completed study was registered on ClinicalTrials.gov (NCT03705078). Findings Between Jan 3, 2019, and Feb 25, 2023, 292 patients were screened, of whom 105 were enrolled and randomly assigned. After excluding two patients who were randomly assigned by error and two who withdrew consent, 101 patients were included in the mITT population (mean age 582 years [SD 97], 81 [80%] male, 91 [90%] alcohol-related cirrhosis, mean MELD score 143 [SD 50]). Of these 101 patients, 47 were allocated to p-TIPS and 54 to glue obliteration and NSBB. The 1-year probability of being free from death or rebleeding was 77% (95% CI 62-87) in the p-TIPS group versus 37% (24-50) in the glue obliteration and NSBB group (hazard ratio 025 [95% CI 012-051]; p<00001). Rescue TIPS was required in 20 (37%) patients in the control group. Glue migration was reported in eight patients (three [6%] in the p-TIPS group and five [9%] in the control group). One case of cardiac decompensation occurred in the p-TIPS group. The 1-year cumulative incidence of hepatic encephalopathy was similar between groups (35% [95% CI 21-49] vs 32% [19-45]). Interpretation In patients with cirrhosis and acute bleeding from fundal varices, p-TIPS significantly reduced the risk of rebleeding or death at 1 year and should be considered as first-line therapy. Funding French Ministry of Health. Copyright (c) 2025 Elsevier Ltd. All rights reserved, including those for text and data mining, AI training, and similar technologies.
Lire l'article complet
