Six controlled epidemiologic studies of the association between intrauterine device (IUD) use and development of acute pelvic inflammatory disease (PID) are reviewed. The strength and weaknesses of each study are discussed as well as the consistencies of the results. All six studies point in the same direction: IUD use is associated with approximately a three-to fivefold increased risk of PID. The consistency of the association in six different studies, in three different countries and by five different investigators is strong evidence that the association is one of cause and effect. Evidence is presented for the biologic plausability of the association. There also is evidence that copper IUDs do not appear to offer any protection from gonococcal PID. Teenagers appear as likely as older women to be similarly affected by the use of IUDs. There is the intriguing but uncomfirmed suggestion that nulliparous IUD users are more likely to suffer from acute PID than are multiparous users. A public health measure that would control this effect of IUD use is to decrease the use of IUDs in populations at high risk of PID.