Insertion. A woman is most likely to develop PID just after insertion (33, 41, 78, 209, 247, 307, 326, 356). Analysis of data from 13 WHO clinical trials found that for women using IUDs the risk of developing PID was 6.3 times greater during the 20 days after insertion than at any later time. After the first 20 days, the incidence of PID remained at a constant low level—1.4 per 1,000 woman-years—throughout eight years of use (507) (see Figure 1). Similarly, the Women's Health Study found that women using IUDs (excluding the Dalkon Shield) faced the greatest risk of developing PID during the first month after insertion—an adjusted relative risk of 3.8. By 5 to 12 months after insertion, the relative risk of developing PID was 1.1—not significantly higher for IUD users than for women using no contraceptive method (209).

Providers can minimize the risk of infection just after IUD insertion by carefully following infection-prevention procedures during IUD insertion (see sidebar, Infection Prevention for IUD Insertion and Removal). In a recent study in which cervical infections were treated before IUD insertion and infection-prevention measures were followed, there was no increase in PID associated with insertion of the Nova T or LNG-20 (573). At the same time, providers should describe PID symptoms to women receiving IUDs, encourage special attention to any of these symptoms in the first month after insertion, and urge them to obtain medical attention promptly if symptoms appear (see sidebar, Important Information About the TCu-380A IUD).

Type of IUD. The Dalkon Shield—which is no longer on the market—was as much as five times more likely to be linked to PID and two times more likely to lead to tubal infertility than other IUDs (78, 81, 177, 209, 415, 495, 627). Moreover, with the Dalkon Shield, the higher risk of developing PID persisted in long-term users (209).

It is not clear whether the risk of PID varies among other IUDs. Several studies, including a recent analysis of WHO clinical trial data, show no difference in infection rates among unmedicated, copper, and hormonal IUDs (177, 209, 335, 343, 422, 507). In the Oxford University/Family Planning Association cohort study, however, the relative risk of developing PID was greater for women using unmedicated than copper IUDs (3.3 versus 1.8) (495). Likewise, two studies on infertility, rather than PID, reported lower risks of tubal blockage with copper IUDs than with unmedicated IUDs (78, 81, 502). A European multicenter study, although not designed especially to detect PID, found significantly lower rates of PID associated with the LNG-20 IUD than with the Nova T (530, 556, 573), but two other large international studies did not find that the LNG-20 protects against PID (334, 517, 546).

Exposure to sexually transmitted diseases. Much of the risk of PID in IUD users, apart from the first weeks after insertion, may be due to sexually transmitted diseases (208, 253, 437). Having multiple sexual partners—and, as a result, greater exposure to STDs—is thus a major PID risk for IUD users, as it is for all women regardless of contraceptive use. Also, as for all other women, if an IUD user's partner has more than one sexual partner, this increases her risk of PID (208, 356, 528). IUD users in mutually faithful sexual relationships face minimal risks, presumably because of less exposure to STDs (94, 208, 359, 360, 366, 423). Indeed, the low long-term relative risks of PID seen in the most careful studies raise doubt whether, beyond the first few weeks after insertion, PID risk is really any greater among users of currently available IUDs than among women with the same STD risks who do not use contraception.

Age. As in other women, PID in IUD users occurs more often among women under age 25 (24, 33, 41, 227, 360, 418, 421, 446, 528). In international clinical trials sponsored by WHO, for example, younger women suffered from PID at 2.5 times the rate of older women (507). This may be at least in part because these younger women are less likely to be married or to have mutually faithful sexual relationships (24, 33, 418, 446).

Duration of IUD use. In most studies the chances of ever developing PID remain unchanged or even decrease with duration of IUD use (41, 209, 332, 343, 385, 415, 463, 507, 550). Two studies concluded that the risk of very severe PID increased with time (356, 463). One involved few cases, and long-term users may have been more likely to have used the Dalkon Shield than copper IUDs (545). The other study involved predominantly inert IUDs (463). Of course, if, as time passes, women change sexual partners, their exposure to STDs and thus their risk of PID may change.