recognizing this complication when it occurs remote from the original operation, and even after hospital discharge. Our approach is based on an experience with 2 patients that we reported in 2006. Both patients sustained massive postpancreatectomy hemorrhage that was initially treated by open conversion of Billroth I procedures to Billroth II. In both patients, there was recurrent bleeding and, after angiography, successful placement of a stent graft in the hepatic artery for a pseudoaneurysm in the first patient, and coil embolization in the second patient. In both patients, we have documented long-term survival extending to the present time. The point I wish to make is that considerable bleeding postpancreatectomy should be a trigger for angiography, with the goal to define and treat a major vascular lesion, if possible, by endovascular means. I would then agree with the authors that, in the absence of demonstrable vascular pathology by angiography, exploration might be essential to find and manage other causes for the bleeding.