Series B, Number 7
Intrauterine Devices

CONTENTS

Home (Key Points)

The IUD: An Important Method with Potential
Table 1. Overview of IUDs
Figure 1. Most IUD Users Are in China
Web Table 1. Current Use of Contraceptive Methods Reported by Married Women 15-49, 1976-2005
Table 2. Estimated Worldwide Use of IUDs Among Married Women Ages 15–49, 2005

Providing High-Quality IUD Services

Spotlight: Kenya Commits To Renewing Interest in the IUD

Feature: Good Counseling Increases Client Satisfaction

Very Low Overall Risk of Infection with IUDs
Figure 2. Risk of Pelvic Inflammatory Disease (PID) Greatest in First Few Weeks After IUD Insertion

Box: Evidence Shows Many Women with HIV Can Use IUDs

Minimizing the Risk of Infection

Clinical Characteristics of IUDs
Web Box 1. Managing Problems with IUD Use

Bibliography

Credits

Go to the IUD Toolkit at http://www.iudtoolkit.org for full-text resources, including tools and best practices, on IUDs.

• Checklist: Program Plan for Providing High-Quality IUD Services
• Checklist for Screening Clients Who Want to Initiate Use of the Copper IUD, from Family Health International
• Counseling Aid for Communicating IUD Effectiveness

• IUD Use, STIs, and HIV-Related Conditions: 2004 WHO Medical Eligibility Criteria
• Do IUDs Increase the Risk of PID in Women with STIs?

Order this report
Download PDF or PowerPoint
E-mail this page
See more Population Reports

Very Low Overall Risk of Infection with IUDs

• WHO Guidance Allows More Women To Use IUDs
• No Significantly Increased Risk of Infertility with IUDs

One of the most persistent questions about the IUD is whether it increases the risk of pelvic inflammatory disease (PID). Infectious organisms, most often those causing gonorrhea or chlamydia, are the direct cause of PID (232, 272).² The majority of evidence indicates that a woman who does not already have an STI—in particular gonorrhea or chlamydia—cannot get PID just from having an IUD inserted. It remains unclear whether the IUD increases the risk of developing PID in a woman with gonorrhea or chlamydia, however, beyond the usual risk just from having these STIs. The ideal study that would answer this question definitively cannot be conducted because it would require randomly assigning women with current gonorrhea or chlamydia either to a group having an IUD inserted or a group receiving no contraception at all.

Assessing a variety of evidence from several sources leads to the following conclusions:

Overall levels of PID in IUD users are low. In large studies mostly in developing countries, rates of acute PID among IUD users have been between 0.6 and 1.6 per 1,000 woman-years³ of use (50, 96). Long-term WHO multicenter studies report 4 to 11 IUD removals for diagnosed PID per 1,000 women over a 10- to 12-year period of use (249, 267).⁴ Although it is difficult to determine PID rates in the general population, studies among women in developed countries suggest rates range from 10 to 17 per 1,000 woman-years (207, 258).

Greatest PID risk is in the first few weeks after IUD insertion. Analysis of data from 13 WHO clinical trials conducted in Africa, the Americas, Asia, and Europe found that the risk of developing PID was 6.3 times greater during the first 20 days after IUD insertion than at any later time (50). After the first 20 days from insertion, the number of new PID cases occurring each year remained at a fairly constant low level—around 1.4 per 1,000 woman-years—throughout eight years of use (see Figure 2). This low level is similar to or even lower than that among women in developed countries who do not use IUDs (207, 258).

PID risk probably is related to having an STI at IUD insertion. The same analysis of WHO clinical trials indicates that, not surprisingly, PID rates among IUD users appear to vary according to the prevalence of STIs in the population (50). There were no PID cases among the 4,300 IUD users in China. During the 1980s, when the studies were conducted, most Chinese couples were thought to be mutually monogamous (82), and the country was virtually without STIs (282). In contrast, in Africa, where STIs are much more common, eight PID cases were found among the 846 IUD users, a rate of about six cases per 1,000 woman-years (50). This suggests that the increased risk of PID is associated with the presence of gonorrhea or chlamydia at the time of IUD insertion. The insertion process probably pushes organisms from the lower genital tract into the upper genital tract, where PID develops. If the organisms are bacteria normally present in the genital tract, then it seems that some mechanism automatically eliminates this contamination from the uterus soon after the insertion process without infection occurring (143). If gonorrhea or chlamydia is present in the lower genital tract, however, PID may develop.

Except for the first few weeks after insertion, an STI may be no more likely to progress to PID in an IUD user than in another woman. Data from six small studies of women with gonorrhea or chlamydia who had had IUDs inserted (57, 147, 163, 210, 222, 255) found percentages who developed PID within one month to two years after IUD insertion (0% to 5%)⁵ that are comparable to levels of PID among women in the general population with inadequately treated or untreated gonorrhea and/or chlamydia (166, 176, 228). This indirect comparison, together with evidence showing a fairly constant low level of PID after the first 20 days following insertion (mentioned above), suggests that, after the first few weeks following insertion, there may be no greater risk that an STI will progress to PID for an IUD user than for other women with STIs (145). A mathematical model taking into account STI prevalence and PID risk assessment data from other studies suggests that IUD users do indeed face a low risk of PID (205). The model estimates a 0.3% risk of PID attributable to the IUD, in a setting with a 10% STI prevalence, if clients are not screened for STIs before insertion. That is, about 3 more cases of PID would occur in every 1,000 women who have IUDs than would have occurred if they did not have IUDs inserted (or, in other words, 997 in every 1,000 women who have IUDs inserted would not develop PID). If IUD use were restricted to clients judged to be at low risk of STIs on the basis of risk-assessment questions (see 'Minimizing the Risk of Infection'), the estimated attributable risk would drop to 0.15%—that is, fewer than 2 women in every 1,000 would develop PID that would be related to IUD insertion.

WHO Guidance Allows More Women To Use IUDs

Current guidance from WHO indicates that women with PID, gonorrhea, chlamydia, or purulent cervicitis should not have IUDs inserted (see ' IUD Use, STIs, and HIV-Related Conditions: 2004 WHO Medical Eligibility Criteria'). (Purulent cervicitis is an infection of the cervix with a pus-like discharge from the opening of the cervix. It may be a sign of gonorrhea or chlamydial infection.) In contrast, women who develop any of these conditions after their IUDs are already in place can continue using their IUDs while the infection is being treated. Women who are at very high individual risk of exposure to gonorrhea or chlamydia should not have an IUD inserted unless other, more appropriate contraceptive methods are not available or not acceptable (268).

Changes to the WHO guidance now allow women with STIs other than gonorrhea, chlamydia, or purulent cervicitis to have IUDs inserted. In addition, a woman who lives in an area where STIs are common generally can have an IUD inserted as long as she herself is not at very high risk of gonorrhea or chlamydia (268). The guidance also generally allows women at risk of or with HIV infection to use IUDs (see 'Box: Evidence Shows Many Women with HIV Can Use IUDs').

The guidance against inserting an IUD in a woman at very high individual risk of gonorrhea or chlamydia reflects a concern that this woman may currently have gonorrhea or chlamydia but without any immediate symptoms. Only laboratory testing would detect these silent infections, and in many places these tests are not available (see 'Minimizing the Risk of Infection').

No Significantly Increased Risk of Infertility with IUDs

Complications of PID can be severe. PID can permanently damage the lining of the fallopian tubes and may partially or totally block one or both tubes enough to cause infertility. Well-designed studies find no significant increase in infertility associated with IUD use, however (77, 79). In numerous cohort studies 72% to 96% of women conceived within a year after their IUDs were removed (244). These levels are comparable to those among women who have never used contraception (244). These observations are consistent with findings that the absolute risk of PID in IUD users is low.

The most recent study looked at women who had not yet had children to find out if those with tubal blockage were more likely to have used IUDs than pregnant women were. The childless women with tubal blockage were no more likely to have used IUDs than either childless women without tubal blockage or the pregnant women. Childless women with or without tubal blockage were more likely to have been exposed to chlamydial infection than pregnant women, however. This finding confirms that exposure to STIs—specifically to chlamydial infection—increases the risk of infertility (94).

Another study reported that, among women who had not yet had children, long-term IUD users took longer to have a child after they had their IUDs removed than short-term IUD users (39). Previous use of oral contraceptives or barrier methods did not affect the time to childbirth.

Numerous problems with this study have been pointed out. They include relevant differences between users of IUDs and other women in the study, little information about who might have had an STI at the time of insertion, and no information on whether delayed time to childbirth was due to tubal blockage or some other cause (78). Also, the findings may not apply to current IUDs. In the 1980s, when the data were collected, most of the IUDs used by the women in the study had to be replaced every two or three years, presumably with elevated risk of infection just after each insertion (81).

2 Some researchers think that bacterial vaginosis (BV) may cause PID, and additionally that IUD users may be more likely to develop BV (11, 70, 90, 104, 105, 139). While some studies have shown that BV is common among women with PID, evidence proving that BV actually causes PID is lacking (116). One recent study followed women at high risk of getting an STI for a median of three years and found no overall increased risk of developing PID among women with BV (152).
3 "Woman-years" is an estimate of the actual time at risk that all women contributed to a study. It is calculated, in this case, by adding together the time each individual IUD user was at risk of PID, from the start of the study until either she was diagnosed with PID or the study ended, whichever came first.
4 New guidance from the 2004 WHO Medical Eligibility Criteria for Contraceptive Use indicates that there usually is no need to remove the IUD if the client wishes to continue using it while being treated for PID (268). There is no difference in the clinical course of PID once it is being treated whether the IUD is removed or left in place (4, 292, 299, 301).
5 If a broader definition of PID were used that included women with mild pelvic pain but without fever, the highest level in these studies might be 10% (145).