In patients with atrial fibrillation (AF) under oral anticoagulant therapy (OAT), over half of the hemorrhagic complications occur in the gastrointestinal (GI) tract, with an incidence of 1-4% per year. This complication mainly involves older patients, often very compromised from the clinical point of view; mortality rates are not negligible, varying between 4% and 15%. The purpose of the present review was to evaluate the utility of resuming OAT after a major GI hemorrhage in patients with AF. Four observational studies were found in the literature that specifically investigated this issue; three of them had a retrospective design. In these studies almost exclusively warfarin was utilized. OAT was discontinued in all patients at the beginning of GI hemorrhage; in about half of the patients anticoagulation was then restarted and in the other half it was definitively stopped. The results of these studies suggest a beneficial effect of OAT resumption, since it reduced the incidence of thromboembolic events and mortality with a not marked increase in hemorrhagic recurrences. However, these results should be interpreted with caution since, very likely, OAT was resumed in patients in good clinical condition - as suggested by the very low mortality rate during hemorrhagic recurrences (0.7%) - and not in those with very severe hemorrhage and/or very compromised from the clinical point of view. Because of this type of patient selection, we do not know the real hemorrhagic risk in patients resuming OAT after GI hemorrhage. This is a strong limitation in the decision making; in order to acquire this knowledge, randomized studies should be carried out. The evaluation whether or not to restart OAT should be made in the clinical context by a team including the gastroenterologist (or the physician taking care of the GI pathology) and the cardiologist. At present, clinical variables such as site and/or cause of GI bleeding, severity of the anemia and the degree of prolongation of the international normalized ratio, do not appear useful for decision making. The available data suggest that OAT should be resumed in "robust" elderly patients, if the source of bleeding has been identified and corrected, whereas in frail patients and/or with multiple comorbidities, the doubt often remains. The available literature does not offer clear data on the optimal duration of OAT discontinuation after an episode of major GI bleeding. The evaluation should be made in the clinical context; however, therapy discontinuation between 1 week and 1 month appears to be adequate in most cases. On the basis of indirect comparisons, which show many limitations, the most appropriate anticoagulants after GI hemorrhage appear to be warfarin, apixaban and low-dose edoxaban.