If the testing is sensitive enough to find trace amounts in the low parts per billion, then there's no "alibi"; it sounds like this is far, far below the performance enhancing level.
Is it possible that the low amounts could be that it's just that level because of the drug being metabolized? That being said, if the test was done 2 days previous, would he have been caught with higher levels? Does it work that way? I have no experience with something like this, obviously.