The development of EUS-GJ has largely replaced GPGJ, the traditional surgical approach, for patients with MGOO. However, in patients with higher survival, there are doubts as to whether this approach offers better results, due to the lower recurrence rate it has classically been associated with.

The main objective of the study was to compare the recurrence rate of MGOO symptoms between EUS-GJ and GPGJ in patients with longer survival. The secondary objectives were to compare the baseline characteristics, survival rates, technical aspects and adverse events between EUS-GJ and GPGJ in patients with MGOO and longer survival.

We designed a retrospective, single-centre, observational study. Patients with MGOO and >5 months post-procedure survival were considered. A search in the clinical history of these patients was conducted, including suitable patients until October 2025. We defined clinical success as the recovery of oral tolerance (GOOS ≥ 2) after the procedure and recurrence as a loss of oral tolerance (GOOSS = 0) at any point after the procedure.

In total, 36 patients were included (26 EUS-GJ vs. 10 GPGJ). The only statistically significant differences we found in their baseline characteristics were in tumoral etiology (the GPGJ group only included gastric and duodenal tumors, while the EUS-GJ group included other etiologies, p = 0.043) and stricture location (the GPGJ group mostly included pyloric stricture, while the EUS-GJ group included strictures at other locations, p = 0.081).

There were no statistically significant differences in clinical success (100% for both groups) or immediate adverse events (11.54% EUS-GJ vs. 0% GPGJ, p = 0.262). All the adverse events were managed during the index endoscopic procedure itself (AGREE IIIa). There were also no significant differences in the days of admission between both groups (13.3 EUS-GJ vs. 19.1 GPGJ, p = 0.1211).

After a median survival time of  286 days, we observed 3 recurrences in the GPGJ group, compared to 1 in the EUS-GJ group. The cumulative incidence of recurrence at 6 and 12 months was 20% and 30% (GPGJ) vs. 3.84% and 3.84% (EUS-GJ). The log-rank test showed a significant difference in recurrence-free survival between both groups (p = 0.016). The recurrences in the GPGJ group were all managed conservatively, while the recurrence in the EUS-GJ group could be treated with the placement of a new stent.

No statistically significant differences were found between the survival rates of both groups (p = 0.8682).

We found EUS-GJ is an effective approach for the treatment of MGOO in patients with longer survival. In comparison with GPJG, EUS-GJ is less invasive, and in our study, it even presents a lower rate of recurrence. Therefore, EUS-GJ could be considered as a first-line therapy for these patients. However, these results need to be confirmed with prospective intervention studies.