Hum. Reprod. Advance Access originally published online on February 16, 2006
Human Reproduction 2006 21(6):1467-1472; doi:10.1093/humrep/del029
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Preventing copper intrauterine device removals due to side effects among first-time users: randomized trial to study the effect of prophylactic ibuprofen
1 Family Health International, Research Triangle Park, NC, USA and 2 Instituto Chileno de Medicina Reproductiva (ICMER), Santiago, Chile
3 To whom correspondence should be addressed at: Family Health International, P.O. Box 13950, Research Triangle Park, NC 27709, USA. E-mail: dhubacher{at}fhi.org
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and methodsResultsDiscussionAcknowledgementsReferences |
|---|
BACKGROUND: Increased menstrual bleeding and pain are the primary side effects that lead to early removal of the copper intrauterine device (IUD). Ibuprofen and other non-steroidal anti-inflammatory drugs (NSAIDs) are proven treatments for such IUD-induced problems, but their effect on early IUD removal is unknown. METHODS: A total of 2019 first-time IUD users were recruited in Chile for this double-blind, randomized, placebo-controlled trial. Half of the participants were given ibuprofen and instructions to take 1200 mg daily during menses (for up to 5 days each cycle) for the first 6 months of IUD use. The other half were asked to take an identical appearing placebo in the same manner. The primary outcome was IUD removal within 12 months of insertion. RESULTS: A total of 1011 and 1008 women were randomly assigned to ibuprofen and placebo, respectively. During 12 months of observation, 190 had the device removed because of dysmenorrhoea and/or increased menstrual bleeding: 85 in the placebo group and 105 in the ibuprofen group. For ibuprofen users, the hazard ratio for removal for these IUD-induced side effects was 1.0 and 1.2 at 6 and 12 months, respectively (both not significant). CONCLUSION: Although increased menstrual bleeding and pain are common reasons for early IUD removal, prophylactic use of ibuprofen, at the dosage used here, does not reduce removal rates.
Key words: ibuprofen/IUD/prevention/removal/side effects
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionAcknowledgementsReferences |
|---|
The intrauterine device (IUD) is the most commonly used form of reversible contraception worldwide; approximately 145 µ
o
married women currently use it (United Nations, Department of Economic and Social Affairs, 2004), and, every year, approximately 41 µo have a device inserted. The predominant type of product, the copper IUD (a plastic T-shaped device with copper affixed to it) was developed over 30 years ago and has become the standard bearer for intrauterine contraception, due to its long-term effectiveness, safety and affordability (World Health Organization, 1987; Trussell et al., 1995
).
For some women, however, copper IUDs can cause increased menstrual blood loss and pain (Liedholm et al., 1975; World Health Organization, 1987; Andrade et al., 1988; Larsson et al., 1993; Petta et al., 1994; Reinprayoon et al., 1998). These side effects often lead to early removal (Sivin and Stern, 1979; Sivin et al., 1992; Trieman et al., 1995; Cox and Blacksell, 2000), which in turn has important consequences. First, women affected by these problems will often adopt less effective (or no) forms of reversible contraception, thus increasing the risk of unintended pregnancy. Second, women who have a device removed for side effects will probably never consider using intrauterine contraception again and may discourage friends from using this highly effective birth control method. Millions of women worldwide experience onset of these side effects, yet tolerate them, and retain use of the IUD; this may affect quality of life.
Ibuprofen is a non-steroidal anti-inflammatory drug (NSAID) that reduces IUD-induced menstrual blood loss and pain (Roy and Shaw, 1981; Makarainen and Ylikorkala, 1986); other NSAIDs do the same (Buttram et al., 1979; Davies et al., 1981) and this class of drug is a recommended treatment for these side effects (Grimes, 2004). Though increased bleeding and pain are common complaints of copper IUD use, NSAIDs are rarely provided as a take-home medication for new users. Moreover, the limited research on this topic suggests that NSAIDs may improve overall comfort and continuation rates of the IUD. Thus, for the approximately 40 x 106 women who receive a copper IUD each year and for the millions who have a device removed prematurely because of side effects, a simple, inexpensive, over-the-counter medication may provide substantial benefits. Prophylactic use of an NSAID may reduce side effects before they begin and help prolong use of the IUD.
|
Subjects and methods |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Participants were recruited at 42 Ministry of Health facilities and one private clinic (Instituto Chileno de Medicina Reproductiva, ICMER) in Santiago, Chile. The ethics committees of Family Health International (FHI), the Chilean Ministry of Health and Instituto Chileno de Medicina Reproductiva (ICMER) approved the protocol and data collection. Research staff obtained written consent from all participants before enrolment. Recruitment began in June 2002 and ended in August 2003; follow-up efforts ceased in December 2004. We complied fully with CONSORT guidelines for the conduct and reporting of this trial (Moher et al., 2001
).
Participants
Eligible women were aged 18–49 years, were literate, had menstruated in the last 6 weeks, had never used an IUD (repeat IUD users would bias results towards high continuation rates), were >6 weeks postpartum if recently pregnant and had no medical contraindications to IUDs or ibuprofen. On the basis of a feasibility study preceding the actual work, we estimated that only 11% of women with desire for intrauterine contraception would be eligible for our study.
Design and procedures
Study clinicians at all facilities were trained on how to administer the protocol, recruit participants, execute the informed consent process and complete forms. Once participants signed the informed consent document, they were given an identification (ID) number and corresponding plastic bottle containing either ibuprofen or placebo tablets; after this random assignment, clinicians inserted the copper IUD (marketed in the United States as ParaGard® T380A). Each active tablet contained 400 mg of ibuprofen, and participants were instructed to take three tablets per day (approximately every 8 h) as prophylaxis during menses, but for a maximum of five consecutive days. Participants were instructed to take the tablets for the first six menstrual periods of IUD use, though follow-up continued for a total of 1 year after insertion or until IUD removal, whichever came first.
A Chilean pharmaceutical company (Laboratorios Silesia S.A.) manufactured the identical appearing placebo and ibuprofen tablets. In separate task orders, the tablets were produced, bottled (90 tablets each) and temporarily stored. FHI generated a list of 2200 sequentially ordered numbers that were randomly assigned to group A or B using random permuted block sizes of 20, 10, 4 and 2; FHI wrote a computer program in SAS® to accomplish this task. Because of the large number of clinics and the desire to maintain balanced allocation at the end of the study, a block size of 2 was used for the 1601th number onwards; these latter numbers were reserved for the end of the study as recruitment was coming to a close. Labels containing the ID numbers for groups A and B were affixed to the bottles, and the bottles were interleaved back into sequential, consecutive order in smaller boxes for eventual distribution to the clinics.
Participants, clinicians and the research team were blinded to the true identity of the tablets. Before and after labelling, random bottles from both stores were pulled and tested by an independent laboratory to confirm placebo and ibuprofen status. The test results were concealed from everyone on the research team who could unblind the study prematurely, and the documents were stored in a locked cabinet with other blinding information at FHI. Blinding was maintained throughout data analysis.
Data collection
On the day of admission, clinicians recorded basic sociodemographic, obstetric and menstrual histories, as reported by participants. Clinicians recorded insertion-related results. Follow-up visits were scheduled for 6, 13, 26 and 52 weeks after insertion. When participants returned for scheduled or unscheduled visits, clinicians filled out follow-up forms to record current IUD use, participants’ stated reasons for the visit and reasons for IUD removal. Many follow-up interviews were conducted by telephone, and in some cases by home visits, when scheduled visits were missed.
Power calculations and statistical analyses
By using published estimates of IUD discontinuation rates (Pastene et al., 1977; Sivin et al., 1992; Farr and Amatya, 1994; Farr et al., 1995; Trieman et al., 1995), we assumed that 16% of women in the placebo group would have the IUD removed within the first 12 months. Further, we assumed that ibuprofen could reduce the rate to 11%. Given the one-sided hypothesis (ibuprofen can only improve continuation rates), our desire to achieve 85% power at a 0.025 significance level, an estimated 15% loss to follow-up rate and the use of a log rank test, we needed 909 participants in each arm of the study (Number Cruncher Statistical System, 1996). We did not perform any unblinded interim analyses.
The primary outcome for analysis was IUD removal, including timing and reason (e.g. bleeding and/or pain); thus, the treated population was used (women having successful IUD insertions). For these analyses, the product limit method (Kaplan and Meier, 1958) was used to estimate the 6- and 12-month cumulative probabilities of IUD discontinuation. As a supporting analysis, proportional hazards regression (Cox, 1972) was used to assess differences in the risk of discontinuation between study groups, controlling for known covariates. Competing reasons for removal (i.e. desire for pregnancy, expulsion and others) were considered independent and censored at the time of event. Only removals for the event of interest were coded as a discontinuation event. If multiple reasons for discontinuation included bleeding/pain, the event was recorded as a bleeding and pain event (except in cases of IUD expulsion, since these side effects would be considered a direct manifestation of the primary cause). The assumption of proportional hazards was assessed graphically and by testing for the significance of interaction terms between time (or functions of time) and other variables in the model.
Two physicians and one statistician in Chile (independent from the research organizations) made up the Data and Safety Monitoring Board (DSMB). The DSMB reviewed FHI-supplied data on three occasions to assess the conduct of the study in terms of IUD and ibuprofen-related adverse events. FHI biostatisticians provided the analysis tables with columns labelled as treatment A and B; a sealed envelope revealing which treatment was ibuprofen/placebo was in the possession of the DSMB, however, it was never opened since the committee had no concerns.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionAcknowledgementsReferences |
|---|
With analyses from a sample of screening forms, we estimated that approximately 20 000 women were assessed for eligibility; most were not eligible because they had used an IUD previously and/or because they were experiencing postpartum amenorrhoea.
Of the 2019 women enrolled, 1011 and 1008 were randomly assigned to ibuprofen and placebo, respectively (Figure 1). A total of 57 participants did not have the IUD inserted. Thirty-four minor protocol violations (15 in the placebo group and 19 in the ibuprofen group) were discovered after enrolment and included young age (4), previous IUD use (18), lactational amenorrhoea (8), within 6 weeks postpartum (3) and alcoholism (1); data from these participants were not excluded from subsequent analyses. Of the 1962 women who had an IUD inserted, 99.2% were contacted at least once during follow-up and 94.5% provided final (12 month) IUD status information; these percentages were similar across treatment group.
|
More than half of the participants were under 25 years of age, 72% were married/cohabitating and 92% had completed basic education (Table I). In terms of reproductive history, 5% of participants were nulliparous, 80% had a last pregnancy that ended more than 6 months ago and over two-thirds stated a desire for a future pregnancy. Ibuprofen and placebo participants did not differ on these characteristics.
|
During the follow-up period, a total of 503 women stopped using their IUD (voluntarily or involuntarily) within the first year; this included 246 events among the placebo group and 257 in the ibuprofen group (Table II). Expulsion was the most common reason for early termination, followed by side effects of menstrual bleeding and/or pain. Dysmenorrhoea and/or increased menstrual bleeding, specific a priori reasons that could be affected by taking ibuprofen during menses, accounted for the vast majority of side effect reasons. Neither overall probabilities for discontinuation nor reason-specific probabilities differed significantly by treatment, even in the first 6 months when participants were instructed to take the tablets.
|
Curves showing cumulative probability of removal reveal little differences between ibuprofen and placebo participants (Figure 2); the probabilities were calculated separately for overall discontinuation and reasons of dysmenorrhoea and/or increased menstrual blood loss (those hypothesized to be directly affected by use of prophylactic ibuprofen during menses).
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Prophylactic ibuprofen had no effect on IUD discontinuation rates. We applied rigorous methods to study this issue, had adequate power and had excellent participant retention (95% over 1 year). Despite evidence that ibuprofen and other NSAIDs can treat IUD-induced side effects, our study found no effect on removal rates with prophylactic use. This counterintuitive finding may have numerous explanations. First, the 1200 mg dose we chose, though adequate for treatment, may have been inadequate to impact the decision to remove the IUD. Second, participants may not have taken the tablets as instructed; for a trial of this size and scope, directly observed therapy would have been impossible. We rejected any notions that counting remaining tablets would have provided useful information on compliance; the validity of such information is questionable. Regardless of compliance concerns, however, the trial more accurately tests whether the provision of the medication can have a beneficial impact. We suspect that even if women had initially chosen not to take the tablets, they would have changed their minds if side effects became bothersome. We gathered information on other medications that participants took; some of these such as analgesics and more NSAIDs may have affected our ability to isolate the impact of the study medication. We did planned ancillary analyses censoring such participants and found that prophylactic ibuprofen still had no impact on removal rates.
Many women (12%) experienced an IUD expulsion (complete or partial); usually 2–10% of copper IUD users spontaneously expel their device in the first year of use (Grimes, 2004). Because of the high expulsion rate, we interviewed 170 study clinicians to understand how such a diagnosis is made. We found that patient complaints of side effects initiate an evaluation and that 85% of clinicians rely on a gynaecological exam for diagnosis and 15% use ultrasound to aid in the diagnosis. Because of the high number of expulsions and the fact that bleeding and pain are often symptoms of expulsion, we did ancillary analyses (by treatment) assuming that clinicians falsely made a diagnosis of expulsion. We found that even if clinicians had erred, ibuprofen had no impact on removals.
On the basis of our results, we cannot recommend prophylactic use of ibuprofen as a means of preventing early removals of the copper IUD. The external validity of our findings may be limited to settings where the IUD is well accepted/understood and where users have convenient access to pharmacies. To recruit enough participants in a reasonable amount of time, we needed a site such as Santiago, Chile, where the IUD is very popular. Regardless of our findings about the lack of impact on IUD continuation rates, clinicians should continue to provide NSAIDs (including ibuprofen) as a treatment for IUD-induced side effects, since the therapeutic evidence for this practice appears solid (Buttram et al., 1979; Davies et al., 1981; Roy and Shaw, 1981; Makarainen and Ylikorkala, 1986).
|
Acknowledgements |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionAcknowledgementsReferences |
|---|
This study was supported by a grant from the National Institutes of Health (HD39717 to D. Hubacher); the views expressed in this article do not necessarily reflect those of the authors’ institutions or the funding entity. We thank the 170 Ministry of Health clinicians for recruiting/enrolling participants and for completing our forms. We also thank Laboratorios Silesia S.A. of Chile for manufacturing and bottling the placebo and ibuprofen tablets. Finally, we thank Sandra Cameron and Monica Salvatierra for their careful attention to detail.
|
References |
|---|
|
TopAbstractIntroductionSubjects and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Andrade AT, Pizarro E, Shaw ST Jr, Souza JP, Belsey EM and Rowe PJ (1988) Consequences of uterine blood loss caused by various intrauterine contraceptive devices in South American women. World Health Organization Special Programme of Research, Development and Research Training in Human Reproduction. Contraception 38,1–18.[CrossRef][ISI][Medline]
Buttram V, Izu A and Henzl MR (1979) Naproxen sodium in uterine pain following intrauterine contraceptive device insertion. Am J Obstet Gynecol 134,575–578.[ISI][Medline]
Cox DR (1972) Regression models and life tables. J R Stat Soc Ser B 34,187–220.
Cox M and Blacksell SE (2000) Clinical performance of the Nova-T380 IUD in routine use by the UK Family Planning and Reproductive Health Research Network: 12-month report. Br J Fam Plann 26,148–151.[ISI][Medline]
Davies AJ, Anderson AB and Turnbull AC (1981) Reduction by naproxen of excessive menstrual bleeding in women using intrauterine devices. Obstet Gynecol 57,74–78.
Farr G and Amatya R (1994) Contraceptive efficacy of the Copper T380A and the Multiload Cu250 IUD in three developing countries. Adv Contracept 10,137–149.[CrossRef][ISI][Medline]
Farr G, Amatya R, Acosta M, Ekwempu C and Kisninci H (1995) Clinical performance of the TCu 380A and Lippes Loop IUDs in three developing countries. Contraception 52,17–22.[CrossRef][ISI][Medline]
Grimes D (2004) Intrauterine devices (IUDs) Contraceptive Technology. Ardent Medic, New York, pp. 495–530.
Kaplan EL and Meier P (1958) Nonparametrics estimation from incomplete observations. J Am Stat Assoc 53,457–481.[CrossRef][ISI]
Larsson G, Milsom I, Jonasson K, Lindstedt G and Rybo G (1993) The long-term effects of copper surface area on menstrual blood loss and iron status in women fitted with an IUD. Contraception 48,471–480.[CrossRef][ISI][Medline]
Liedholm P, Rybo G, Sjoberg NO and Solvell L (1975) Copper IUD – influence on menstrual blood loss and iron deficiency. Contraception 12,317–325.[CrossRef][ISI][Medline]
Makarainen L and Ylikorkala O (1986) Ibuprofen prevents IUCD-induced increases in menstrual blood loss. Br J Obstet Gynaecol 93,285–288.[ISI][Medline]
Moher D, Schulz KF and Altman DG (2001) The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomised trials. Lancet 357,1191–1194.[CrossRef][ISI][Medline]
Number Cruncher Statistical System (NCSS) (1996) PASS 6.0 User’s Guide: Power Analysis and Sample Size (PASS). Number Cruncher Statistical System (NCSS), Kaysville, UT.
Pastene L, Rivera M, Zipper J, Medel M and Thomas M (1977) IUD insertions by midwives: five years’ experience in Santiago, Chile. Int J Gynaecol Obstet 15,84–87.[Medline]
Petta CA, Amatya R and Farr G (1994) Clinical evaluation of the TCu 380A IUD at six Latin American centers. Contraception 50,17–25.[CrossRef][ISI][Medline]
Reinprayoon D, Gilmore C, Farr G and Amatya R (1998) Twelve-month comparative multicenter study of the TCu 380A and ML 250 intrauterine devices in Bangkok, Thailand. Contraception 58,201–206.[CrossRef][ISI][Medline]
Roy S and Shaw ST Jr (1981) Role of prostaglandins in IUD-associated uterine bleeding – effect of a prostaglandin synthetase inhibitor (ibuprofen). Obstet Gynecol 58,101–106.
Sivin I and Stern J (1979) Long-acting, more effective copper T IUDs: a summary of U.S. experience, 1970–75. Stud Fam Plann 10 (10),263–281.[CrossRef][ISI][Medline]
Sivin I, Greenslade F, Schmidt F and Waldman SN (1992) The Copper T380 Intrauterine Device: A Summary of Scientific Data. The Population Council, New York.
Trieman K, Liskin L, Kols A and Rinehart W (1995) IUDs an update. Population Reports, Series B, No. 6. Johns Hopkins School of Public Health, Population Information Program, Baltimore, MD.
Trussell J, Leveque JA, Koenig JD, London R, Borden S, Henneberry J, LaGuardia KD, Stewart F, Wilson TG, Wysocki S et al. (1995) The economic value of contraception: a comparison of 15 methods. Am J Public Health 85,494–503.
United Nations, Department of Economic and Social Affairs (2004) World Contraceptive Use 2003 (http://www.un.org/esa/population/publications/contraceptive2003/WallChart_CP2003_web.xls) (Accessed May 28, 2004).
World Health Organization (1987) Mechanism of action, safety and efficacy of intrauterine devices. Report of a WHO Scientific Group. World Health Organ Tech Rep Ser 753,1–91.[Medline]
Submitted on November 8, 2005; resubmitted on January 10, 2006; accepted on January 13, 2006.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  The ESHRE Capri Workshop Group Intrauterine devices and intrauterine systems Hum. Reprod. Update, May 1, 2008; 14(3): 197 - 208. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||