Usually when I experienced that behavior from IE there would be a redirect involved. So link in PDF takes you to intermittent location which then redirects to the final location. This may or may not be the case here.
Additionally it is hard to rule out environmental factors, such as which pdf reader do you use, what version of ie, pdfreader, etc. Do you have proxies, do they strip or add headers to your requests. Do you have spyware that could be hijacking links, do you have kill bits set for activex objects or javascript in pdfs, etc. If you can ensure us that it is reproducible behavior it might be easier to check using some inspection tools (such as etheral or something similar).
Confucius says kill mosquito unless cannon