AIMTo evaluate the safety and efficacy of a modified cyanoacrylate [N-butyl-2-cyanoacrylate associated with methacryloxysulfolane (NBCA + MS)] to treat non-variceal upper gastrointestinal bleeding (NV-UGIB).METHODSIn our retrospective study we took into account 579 out of 1177 patients receiving endoscopic treatment for NV-UGIB admitted to our institution from 2008 to 2015; the remaining 598 patients were treated with other treatments. Initial hemostasis was not achieved in 45 of 579 patients; early rebleeding occurred in 12 of 579 patients. Thirty-three patients were treated with modified cyanoacrylate: 27 patients had duodenal, gastric or anastomotic ulcers, 3 had post-mucosectomy bleeding, 2 had Dieulafoy’s lesions, and 1 had duodenal diverticular bleeding.RESULTSOf the 45 patients treated endoscopically without initial hemostasis or with early rebleeding, 33 (76.7%) were treated with modified cyanoacrylate glue, 16 (37.2%) underwent surgery, and 3 (7.0%) were treated with selective transarterial embolization. The mean age of patients treated with NBCA + MS (23 males and 10 females) was 74.5 years. Modified cyanoacrylate was used in 24 patients during the first endoscopy and in 9 patients experiencing rebleeding. Overall, hemostasis was achieved in 26 of 33 patients (78.8%): 19 out of 24 (79.2%) during the first endoscopy and in 7 out of 9 (77.8%) among early rebleeders. Two patients (22.2%) not responding to cyanoacrylate treatment were treated with surgery or transarterial embolization. One patient had early rebleeding after treatment with cyanoacrylate. No late rebleeding during the follow-up or complications related to the glue injection were recorded.CONCLUSIONModified cyanoacrylate solved definitively NV-UGIB after failure of conventional treatment. Some reported life-threatening adverse events with other formulations, advise to use it as last option.