Antibiotics for prevention with IUDs

An intrauterine device (IUD) is a small device placed in the womb for long-term birth control. Many people worry about the woman getting pelvic inflammatory disease (PID) with an IUD. This infection can lead to problems in getting pregnant. If PID occurs, it is often within the first few weeks. Antibiotics are sometimes used before inserting an IUD to prevent an infection. This review looked at how well these preventive drugs reduced problems. Such problems include PID, extra health care visits, and stopping IUD use in three months.

In February 2012, we did a computer search for trials that compared an antibiotic to a placebo ('dummy'). We contacted researchers to get more information. We also wrote to researchers to find other trials.

Women who took antibiotics to prevent infection did not get PID as often as those who had the placebo or no treatment. However, the numbers with PID were low for all groups, so the treatment did not have a major effect. Women who use the drugs for prevention had fewer extra visits for health care. The small difference may not be enough to provide all women with the drugs. Using antibiotics to prevent infection did not change how many women had an IUD removed in three months.

Authors' conclusions: 

Use of either doxycycline 200 mg or azithromycin 500 mg by mouth before IUD insertion confers little benefit. While the reduction in unscheduled visits to the provider was marginally significant, the cost-effectiveness of routine prophylaxis remains questionable. A uniform finding in these trials was the low risk of IUD-associated infection, with or without use of antibiotic prophylaxis.

Read the full abstract...
Background: 

Concern about the risk of upper genital tract infection (pelvic inflammatory disease (PID)) often limits use of the intrauterine device (IUD), a highly effective contraceptive. Prophylactic antibiotic administration around the time of induced abortion significantly reduces the risk of postoperative endometritis. Since the risk of IUD-related infection is limited to the first few weeks to months after insertion, contamination of the endometrial cavity at the time of insertion appears to be the mechanism, rather than the IUD or string itself. Thus, antibiotic administration before IUD insertion might reduce the risk of upper genital tract infection from passive introduction of bacteria at insertion.

Objectives: 

To assess the effectiveness of prophylactic antibiotic administration before IUD insertion in reducing IUD-related complications (pelvic inflammatory disease; complaints leading to an unscheduled visit) and discontinuations within three months of insertion.

Search strategy: 

In February 2012, we searched MEDLINE, POPLINE, and CENTRAL. We also searched for current trials via ClinicalTrials.gov and ICTRP. Previous searches also included EMBASE. For the initial review, we examined reference lists and wrote to experts on several continents to identify unpublished trials.

Selection criteria: 

We included randomized controlled trials using any antibiotic compared with a placebo.

Data collection and analysis: 

Two independent reviewers abstracted data. We made telephone calls to investigators to obtain additional information. We assessed the validity of each study using methods suggested in the Cochrane Handbook. We generated 2x2 tables for the principal outcome measures. The Peto modified Mantel-Haenszel technique was used to calculate odds ratios. We assessed statistical heterogeneity between studies.

Main results: 

The odds ratio (OR) for pelvic inflammatory disease was 0.89 (95% Confidence Interval (CI) 0.53 to 1.51) for use of prophylactic doxycycline or azithromycin compared with placebo or no treatment. Use of prophylaxis was associated with a small reduction in unscheduled visits to the provider (OR 0.82; 95% CI 0.70 to 0.98). Use of doxycycline or azithromycin had little effect on the likelihood of removal of the IUD within 90 days of insertion (OR 1.05; 95% CI 0.68 to 1.63). No statistically significant heterogeneity between study results was detected.