Hard to know what to make of this report, given that the denominator is “estimated” by the manufacturer. The database to which the fatal bleeds were reported is the EudraVigilance, managed by the European Medicines Agency (the EC’s regulatory body). Unlike the FDA database, EudraVigilance cannot be queried by the public. Their report states that out of the 256 fatal bleeds, 21 occurred in Europe. I doubt that much of the others were from the US. Presumably a bulk comes from Japan, which was the first country to issue a modification to the package insert.
There has always been (to my knowledge), a higher rate of hemorrhagic strokes in Japan. The pharmacodynamics of the drug may also be different and less favorable in this population. These bleeding concerns may not apply to the US. However, the absence of a reversing agent for Pradaxa remains a handicap.