ObjectiveRex shunt (mesenteric-to-left portal vein bypass) is considered a more physiologically rational treatment for EHPVO than other portosystemic systemic shunts in children. However, about 13.6% of children with EHPVO do not have usable left portal veins and up to 28.1%. Rex operations in children are not successful. Hence, a Rex shunt in these children was impossible. This study reports a novel approach by portal-to-right portal vein bypass for treatment of children with failed Rex shunts. Material and methodsEight children (age 6.1years, range 3.5–8.9years) who underwent Rex shunts developed recurrent gastrointestinal bleeding and hypersplenism 13months (11–30months) postoperatively. After ultrasound confirmation of blocked shunt, they underwent exploration. Three patients were found to have right portal vein agenesis. Five patients (62.5%) were found to have the patent right portal vein, with the diameter of 3–6mm. Four patients underwent bypass between the main portal vein in the hepatoduodenal ligament and the right portal vein by interposing an inferior mesenteric vein autograft, whereas the remaining patient underwent a bypass using ileal mesenteric vein autograft. ResultsThe operations took 2.3h (1.9–3.5h). The estimated blood loss was 50ml (30–80ml), with no complication. The portal venous pressure dropped from 34.6cmH2O (28–45 cmH2O) before the bypass to 19.6cmH2O (14–24cmH2O) after the bypass. The 5 patients were followed up for 10.2months (4–17months) and the post-operative ultrasound and CT angiography confirmed the patency of all the grafts and disappearance of the portal venous cavernova in all five patients. ConclusionThe portal-to-right portal vein bypass technique is feasible and safe for treatment of children with EHPVO who have had failed Rex shunts. Our preliminary result indicates that this technique extends the success of Rex shunt from left portal vein to right portal vein and open a new indication of physiological shunt for some of the children who not only have had failed Rex shunts or but also are not suitable for the Rex shunts. Type of studyTreatment study. Level of evidenceLevel IV.