Well, it is what it is, right? I don't see a practical approach that would allow one to adjust for the "reserve ccy" issue without making all sorts of completely arbitrary assumptions. Moreover, there's all sorts of other well-known issues with national accounting, but that's life, innit? Nothing's perfect in this funny world of ours. I might add that my personal favorite at the moment is the measurement of the contribution of financial services to GDP. I really like the stuff that Andy Haldane at the Bank of England produces on the subject.
Again, as I mentioned earlier, there's no such assumption made when calculating GDP using the expenditure approach. The (im)balance of trade is captured by the last term in the famous national income identity.