http://www.skepdic.com/acupunc.html
http://archinte.ama-assn.org/cgi/content/abstract/167/17/1892?maxtoshow=...
http://www.theness.com/neurologicablog/?p=362

Dr. Stephen Novella says:

Taken as a whole, the pattern of the acupuncture literature follows one with which scientists are very familiar: the more tightly controlled the study the smaller the effect, and the best controlled trials are negative. This pattern is highly predictive of a null-effect - that there is no actual effect from acupuncture.