Abstract

Analyses of randomized trials of copper intrauterine devices (IUDs) substantiate three hypotheses: 1) Ratios of ectopic to total pregnancies conceived during copper IUD use did not differ among devices; 2) pregnancy rates per 1000 woman-years varied inversely with copper surface area; and 3) ectopic pregnancy rates consequently varied inversely by copper surface area. Randomized trials also demonstrated that progestin-only IUDs had a markedly higher ratio of ectopic to total pregnancies, 171 per 1000, than did copper IUDs, at 39 per 1000 (P less than .001). Ectopic pregnancy rates per 1000 woman-years with progestin-only IUDs varied inversely with dose, notably for levonorgestrel-releasing devices. In a large United States IUD cohort study, ectopic pregnancy ratios of copper IUDs varied significantly (P less than .05) with age, as they do in the general population. Copper and non-medicated IUDs in the United States and the United Kingdom had age-specific ectopic pregnancy ratios six times those of non-contraceptors in the United States population during 1970-1978. The estimated rate of ectopic pregnancy for users of IUDs having 200 mm2 of copper was four-tenths that of non-contraceptors in the United States during 1970-1978. For users of devices with 350 mm2 or more of copper, ectopic pregnancy rates were one-tenth those of non-contraceptors. Rates for users of devices releasing 65 micrograms/day of progesterone are estimated to have been 50-80% above those of non-contraceptors. Devices releasing 20 micrograms/day of levonorgestrel have had ectopic pregnancy rates estimated to be 0.1 of those of non-contraceptors. Most IUDs protect against ectopic pregnancy.