Here is an Ashenden paper published a few years ago that was discussed here quite a bit at the time. He showed that the passport won't pick up a lot of EPO-induced changes, but he also discussed the odds that the rider would test positive for EPO, and concluded there was a good chance he wouldn't.
The link I have provided is through Google scholar. Though it's only the Abstract, some of the full paper comes up, and if you're lucky--I have been in the past, but not now, possibly a browser issue--you might get the full paper. The Discussion section, which I can't access now, goes over the issue of whether the rider would test positive for EPO. As I said, we discussed this extensively before, but I'm not going to try to dig that out now.
However, the gist of it is that if the doses are IV and low enough, the glow time is only 12-18 hours. This was Ferrari's discovery, of course, and in 2006 Ashenden published a paper showing that. In the article I have linked here, EPO was given IV to the volunteers twice weekly, so they might have tested positive, but obviously it would be a chance thing. If you assume they were glowing for 18 hours after each injection, they would still be negative about 80% of the time. Of course it depends on the dose, but no question you can take a significant amount of EPO and have a good chance of not testing positive. The question is whether a rider wants to take that risk.