Ectopic pregnancy and past IUD use. Two recent studies have concluded that IUD use somewhat increases the risk of ectopic pregnancy after the IUD is removed (576, 624). These studies, conducted by some of the same researchers in the US and in Indonesia, reported a similar increase in risk—1.6 and 1.7 times greater—associated with past IUD use. Both these studies were case-control designs, comparing women who had ectopic pregnancies with nonpregnant women of the same age and area of residence, and not currently using IUDs, and seeing whether one group was more likely to have used IUDs in the past.

Earlier studies—all using case-control methology—of ectopic pregnancy and IUD use before conception produced mixed findings. Two studies involving sexually active nonpregnant woman as controls found low or no risk linked with past IUD use (261, 442). In contrast, some (582, 602) but not all (442, 612) studies involving, as controls, pregnant women or women who had just given birth found a somewhat greater level of risk. Still other studies found risk only for former users of the Dalkon Shield (81, 580). A meta-analysis of all studies available through 1994 concluded that past IUD use might increase the risk of ectopic pregnancy by about 40% and that choice of pregnant or nonpregnant controls made no difference (651).

Still, these findings are difficult to interpret because neither pregnant nor nonpregnant women is an ideal control group (645, 651). Carolyn Westhoff has argued that the best control group for an analysis of whether past IUD use increases the risk that a pregnancy will be ectopic would be women who conceive, including those who have spontaneous or induced abortions (646). She points out that the duration-of-use effect, as seen in the recent US and Indonesian studies (576, 624), may appear because long-term IUD users are more likely to stop IUD use to become pregnant, whereas short-term users are more likely to stop IUD use because of side effects and then switch to other methods, thus protecting themselves from pregnancies, including ectopics (646).

At the same time, in studies to assess whether former IUD users face an increased risk of ectopic pregnancies, control groups that include ever-users of contraception will tend to increase the apparent relative risk. Thus, for example, when Irving Sivin used data from the Women's Health Study (261) to compare former IUD users with women not protected from ectopic pregnancy by use of other contraception, he calculated the relative risk for past IUD use at 0.7, suggesting a modest protective effect (327). Similarly, a WHO case-control study of ectopic pregnancy found that the relative risk of past IUD use was 0.7 whether past IUD users were compared with currently pregnant women or with nonpregnant women (442). A small study examining tissue from women operated on for ectopic pregnancies found that inflammation of the fallopian tubes, which might be related to PID, was not more common in current or past IUD users than in woman who had never used IUDs (607).

Infertility. Most women who discontinue IUD use to become pregnant conceive as rapidly as nonusers. As noted, however, IUD insertion can increase the risk of developing pelvic inflammatory disease (PID). The extent to which this leads to tubal infertility has been debated (437, 586, 595).

Two US case-control studies reported in 1985 that, overall, childless women with tubal infertility were two to three times more likely to have used IUDs than women having their first child (78, 81). The risk of tubal infertility varied markedly with the number of a woman's sexual partners. For example, in one study women who had had only one sexual partner in their lifetimes, regardless of the type of IUD used, had no increased risk of tubal infertility. Women who had had more than one partner had three to four times higher risk.

In contrast to these two case-control studies, most cohort studies that have followed women who stopped using IUDs have found no indication of impaired fertility. In over a dozen studies, from 72% to 96% of women conceived within a year after discontinuation (5, 9, 23, 84, 267, 291, 294, 332, 337, 431, 485, 514, 529, 549, 550, 553) and in one large study 51% gave birth within a year (and therefore a higher percentage presumably were pregnant) (414). These rates are in the same range as rates among women who have never used contraception (32, 323) and apply to the LNG-20 IUD as well as copper IUDs (480, 550). When studies have followed former IUD users for longer periods of time, on average for four years, they have found the prevalence of tubal infertility to be low (from 3 to 14 per 1,000 IUD removals) compared with rates in the general population (337, 431, 539, 551).

Of course, cohort studies cannot be expected to gauge whether IUD insertion leads to infertility in a very small fraction of users. In fact, most of these studies have involved women who had no complications with IUD use (5, 9, 23, 84, 267, 291, 294, 332, 337, 514, 529, 549, 553). In most studies all the women were married (9, 294, 337, 414, 514), and thus were not at high risk for STDs, or had had children (9, 23, 84, 267, 414, 514). Two studies did examine conception rates in women who had never been pregnant before using IUDs (337, 431). After these women stopped using IUDs, they conceived at a slightly slower rate than women with children, as would be expected among women in general. The difference was not statistically significant, however, and the gap diminished over time. Also, studies have found no clear difference in the return of fertility between women who had discontinued use of the IUD for medical reasons, which might have included PID or its symptoms, and women who had stopped using IUDs in order to become pregnant (414, 431).

Most cohort studies have found that women who use IUDs for long periods of time conceive about as rapidly as short-term users (9, 20, 23, 179, 291, 294, 337, 431, 486, 514, 550, 553). One of the US case-control studies, however, found that the risk of infertility was slightly higher with longer use, once women who had used their IUDs for less than three months were excluded (78).

A large case-control study found that current and former IUD users were no more likely than nonusers to develop tubal adhesions (fibrous bands of tissue), which can be caused by PID and are a major reason for tubal infertility. These results fail to support other reports that IUD use increases the risk of infertility (462).