Recent data indicate that the number of ectopic pregnancies in the US has quadrupled since 1965, while the frequency has tripled, from 38/100,000 women aged 15-44 in 1965 to 113/100,000 in 1981. The frequency of ectopic pregnancies in Sweden has also increased significantly, while in England and Wales a less marked but still notable increase has also occurred. The finding in the 1970s of a correlation between ectopic pregnancy and IUD use raises questions about the causal factors involved which are addressed in this article through an examination of the relevant literature. It has been noted in previous work that rates of ectopic pregnancy associated with inert or copper IUD use in the US are not different from rates associated with use of barrier contraceptives or in the absence of contraception. Comparison of rates of ectopic pregnancy among women using IUDs with those of women using combined oral contraceptives (OCs) sterilization, or sexually inactive are methodologically questionable. In 1981, the Women's Health Study, a large case control study, demonstrated that current IUD use protected against ectopic pregnancy when the control group was composed of women who had never used an IUD. Current IUD users had a relative risk of .4 when compared with women using no contraception. Users of barrier methods had a relative risk very similar to that of IUD users, while women on OCs had a relative risk of .1. Differences in the rates of ectopic pregnancy among users of barrier methods in England and the US are largely explained by the much higher failure rates of barrier methods in the US. The relative risk of ectopic pregnancy among IUD users was also under 1 in a Swedish study. The single study with data on the relative risk of previous IUD users indicates a relative risk of 1.4 compared to women who never used the IUD, but a risk of about .7 compared to women currently at risk of conception but using no method. The 3 large studies analyzing risk of ectopic pregnancy by duration of IUD use yielded conflicting results and have been challenged on methodological grounds. Supplementary proof of a correlation from prospective studies is needed. Studies of the rates of ectopic pregnancy for different types of IUD are difficult to plan and conduct. In general they indicate that inert and copper devices are not associated with an increased risk, although progesterone-releasing devices elevate the risk significantly. It is concluded that there is no causal relationship between use of a nonmedicated or copper bearing IUD and ectopic pregnancy, although OCs and effective barriers afford better protection against ectopic pregnancy than do IUDs.