A study conducted by the World Health Organization (WHO) Task Force showed that the progestogen, levonorgestrel, given alone in 2 separate doses of 750 μg to be more effective and associated with fewer adverse effects than the Yuzpe regimen.1 Levonorgestrel has essentially replaced the Yuzpe regimen as the standard method used for providing oral hormonal emergency contraception in many parts of the world, and over 80 countries have now approved the progestogen-only emergency contraception regimen.2,3
Studies have evaluated the use of different doses of mifepristone for emergency contraception and reported high efficacy.4 The Cochrane Database of Systematic Reviews combined the results of the studies comparing high doses of mifepristone (> 50 mg) with midrange doses (25–50 mg) and low doses (≤ 10 mg).4 The review concluded that the efficacy of the different doses of mifepristone seemed to be similar, but the frequency of the delay in onset of the subsequent menstrual cycle after treatment increased with increasing doses. In 2002, the WHO Task Force reported a multicenter international study comparing low-dose mifepristone (10 mg) with levonorgestrel 750 μg given 12 hours apart and levonorgestrel 1.5 mg given in a single dose and reported similar efficacy and adverse-effect profile for all 3 arms.2
We carried out this 2-center trial, which commenced before the completion of the WHO Task Force study, in our own environment to compare the efficacy, adverse effects, and acceptability of a single low-dose mifepristone regimen (10 mg) with the levonorgestrel regimen (750 μg given 12 hours apart) administered up to 120 hours after unprotected sexual intercourse.
MATERIALS AND METHODS
The study was carried out between July 2000 and June 2003 in the Family Planning Clinic in Aberdeen with a small number of women (n = 24) recruited in the Family Planning Clinic in Dundee during a 4-month period in the final year of the study. Ethical approval was obtained from the Grampian and Tayside Research and Ethics Committees. Eligible participants were women over the age of 16 years with regular menstrual cycles (21–35 days), who requested emergency contraception within 120 hours of unprotected sexual intercourse. Advice was given to women to avoid further episodes of unprotected sexual intercourse within that cycle and to attend a follow-up appointment 7–10 days after their next expected menstrual period. A urine pregnancy test was carried out if the patient's next menstrual period failed to occur within 7–10 days of the expected date or if menstruation following treatment was not normal.
Women presenting within 72 hours of unprotected intercourse were asked to take part in the study. Women presenting beyond 72 and up to 120 hours were offered a copper intrauterine device (IUD) insertion as the first treatment choice. Those declining IUD insertion were offered participation in the study and were randomized to receive mifepristone or levonorgestrel. The study protocol was based on evidence suggesting decreased efficacy of hormonal emergency contraception as the duration between sexual intercourse and treatment increased and high efficacy reported for the copper IUD in this group.1,5 Women with more than one episode of unprotected sexual intercourse within 120 hours of presentation were also included in the study, although this was taken into account in the analysis. Women using any form of hormonal contraception were excluded. Those who were certain they would continue with the pregnancy if the emergency contraception failed (based on the lack of safety data on the effect of mifepristone on a continuing pregnancy), women with known allergy to mifepristone or progestogens, and women using hepatic enzyme–inducing medications were also excluded. Women who previously used hormonal contraception were eligible for recruitment if they had experienced one normal period after discontinuing. Women were randomized to receive a single tablet of mifepristone 10 mg (provided by Exelgyn Laboratories, Oxon, UK) or 2 tablets of levonorgestrel, 750 μg given 12 hours apart, by opening sequentially numbered opaque sealed envelopes prepared using random number tables. The randomization envelopes were prepared in the Family Planning Clinic in Aberdeen by a health care assistant not involved in the recruitment or data collection. The study was not blinded, and both medical staff and patients were aware of the treatment assigned.
The primary outcome measure was unintended pregnancy, confirmed by a positive pregnancy test followed by a transvaginal ultrasound scan. We assessed the crude number of pregnancies, the number expected, and the estimated reduction in expected pregnancies. Prevented pregnancies were calculated using the following formula: 1 − (observed/expected pregnancies). Expected pregnancies were calculated by multiplying the number of women having unprotected intercourse on each day of the menstrual cycle by the probability of conception on that day of the cycle, using the pooled recognizable conception probabilities described by Trussell et al6 The estimated date of ovulation was assessed by subtracting 14 days from the expected date of onset of the next menstrual period after treatment. Secondary outcomes included adverse effects experienced by the women, acceptability of the method of emergency contraception used, and the timing of the first menstrual period after treatment. These were assessed using questionnaires given to women on the day of treatment and collected at the time of follow-up.
The outcomes were assessed through the patient questionnaires at follow-up, by telephone follow-up for women giving consent, or through subsequent attendance to the Family Planning Clinic. A follow-up appointment was made 7–10 days after the next expected menstrual period after treatment. A urine pregnancy test was carried out if no bleeding had occurred or if menstruation after treatment was not normal. If bleeding had not occurred but the pregnancy test was negative, a further appointment was arranged a week later. Those who conceived were counseled about their options regarding the pregnancy.
The Hospital Patient Administration System database, the Maternity Hospital, and Termination of Pregnancy Databases for the Aberdeen Royal Infirmary, as well as hospital and community (Family Planning) case notes, where necessary, were reviewed to exclude pregnancies for women recruited in Aberdeen who failed to attend follow-up and those who were not able to be contacted and did not return the study questionnaires. Of the 24 women recruited to the study from Dundee, the outcome was assessed through follow-up at the Family Planning Clinic in Dundee or telephone follow-up.
The sample size was calculated by using the equivalence criterion.7 The efficacy of the levonorgestrel regimen was based on a crude pregnancy rate of 1.1%, as reported by the WHO Task Force.1 It was estimated that a sample of 1,890 women in each group would demonstrate equivalence within 1.1% (2-sided alternative hypothesis), with 90% power at the 5% level of significance. To compensate for an anticipated loss to follow-up of 10%, a total sample size of 4,200 women was needed to demonstrate equivalence between the 2 methods.
Recruitment was slower than anticipated because of the stringent entry criteria and an observed change to more complex presentations for emergency contraception within the clinic. This resulted in a higher-than-expected exclusion rate, and the study had to be discontinued because the shelf life of the mifepristone tablets had expired.
Approximately 3,000 women request emergency contraception at the Family Planning Clinic in Aberdeen annually. Furthermore, approximately 1,500 women attend the Family Planning Clinic annually in Dundee in request of emergency contraception, and during a 4-month period in the final year of the study, 24 women were recruited.
The total number of women recruited was 2,065. Of these, 1,030 women were randomized to receive mifepristone and 1,035 to receive levonorgestrel. A total of 8 women who were randomized to receive mifepristone and 12 women randomized to receive levonorgestrel did not fulfill the study criteria and were excluded from data analysis.
Data were analyzed using the SPSS 11.5 (SPSS Inc, Chicago, IL). Variables that were normally distributed (assessed by superimposing a normal curve on the data chart created by SPSS) are presented as means and standard deviations and were analyzed using the independent and paired t test. The χ2 or Fisher exact were used for independent nominal data. Confidence intervals (CIs) were used where appropriate, and statistical significance was defined as P < .05. Multiple logistic regressions were carried out to adjust for age, body mass index, the method of contraception in use at the time of unprotected sexual intercourse, circumstances leading to emergency contraception request, the interval between unprotected intercourse and treatment, number of episodes of unprotected intercourse before treatment, and the day in the menstrual cycle on which unprotected intercourse occurred. Epi Info 6 (Centers for Disease Control and Prevention, Atlanta, GA) was used for the χ2 calculation of pregnancy trends in the 5 successive days from the time of unprotected intercourse for the 2 groups.8
Of the 2,043 women included in the data analysis, 1,022 were in the mifepristone group and 1,021 in the levonorgestrel group. Key characteristics of the women are shown in Table 1. There were no significant differences in age, weight, height, or past reproductive history for women in the 2 groups.
Of the total number of women (Aberdeen and Dundee), 761 (37%) attended follow-up at the Family Planning Clinic, 635 women (31%) had a telephone follow-up, 356 women (17%) attended the Family Planning Clinic or hospital subsequently for another reason, while 291 women (14%) did not attend follow-up.
Treatment outcome for women (Aberdeen and Dundee) was known for 860 women (84.2%) in the mifepristone group and 858 (84.1%) in the levonorgestrel group. Of women with unknown outcome, 15 (9.3%) in the mifepristone group and 19 (11.7%) in the levonorgestrel group subsequently attended the Family Planning Clinic or hospital in Aberdeen but had no documentation of the outcome of their treatment. No further pregnancies were identified from the review of the hospital databases and, where necessary, the patient case notes for women with unknown outcome in Aberdeen.
Of the 24 women recruited in Dundee, 8 women (33%) attended follow up at the Family Planning Clinic in Dundee and 7 (29%) had a telephone follow-up, while the outcome was unknown for 9 women (38%). None of the subgroup of women recruited in Dundee and where the outcome was known conceived.
Treatment outcomes in relation to the time interval since sexual intercourse are shown in Table 2. Excluding the cases with unknown outcome, we observed failure rates of 1.5% and 2.3% for the mifepristone and levonorgestrel regimens, respectively (difference of 0.7%, 95% CI 0.5–2.2%). The failure rates for mifepristone (1.5%) did not cross the stipulated 95% CI of 1.1% for the difference between the 2 treatments, whereas that for levonorgestrel (2.3%) did. This suggests that we cannot rule out the possibility that levonorgestrel could be less effective than mifepristone. That would also suggest that levonorgestrel could result in an absolute difference of 2.2% more pregnancies than mifepristone.
There was no significant trend in the pregnancy rates in the 5 successive days from the time of unprotected intercourse for the 2 groups (χ2 for linear trend:8 0.27, P = .61), although the number of women receiving emergency contraception beyond 72 and up to 120 hours after unprotected intercourse was small.
A total of 23 women (1.5%) in the mifepristone group and 13 women (0.9%) in the levonorgestrel group required repeat emergency contraception in the same menstrual cycle. Reasons for repeat emergency contraception included further episodes of unprotected intercourse, missing the second levonorgestrel dose, and vomiting after receiving emergency contraception. Of these, one woman in the mifepristone group conceived and chose to have an abortion. The outcome was unknown for two, while all the remaining women did not conceive.
Multiple logistic regressions showed no significant association between the efficacy of emergency contraception (defined using unintended pregnancy) and age, body mass index, the method of contraception used, circumstances leading to emergency contraception request, the interval between unprotected intercourse and treatment, number of episodes of unprotected intercourse before treatment, and day in the menstrual cycle on which unprotected intercourse occurred for the 2 emergency contraception methods used. These findings are shown in Table 3. Features of the subsequent menstrual cycle after emergency contraception use for women in both groups are shown in Table 4.
A total of 743 women (36%) completed the acceptability questionnaires. Of these, 373 (50%) were in the mifepristone group and 370 (50%) in the levonorgestrel group. This low return rate reflects the low attendance for follow-up at the Family Planning Clinic. Adverse effects experienced by women in both groups are shown in Table 5. Women were presented with a list of possible advantages and disadvantages for the emergency contraception method used. The views expressed by women are shown in Table 6.
A total of 339 (93.6%) women in the mifepristone group were satisfied with the method used, compared with 329 (90.9%) in the levonorgestrel group. Seventeen (4.7%) and 21 (5.8%) answered “don't know,” while 6 (1.7%) and 12 (3.3%), respectively, were dissatisfied with the method used (P = .28). A total of 345/363 (95%) women in the mifepristone group said they would use the same method again if they required emergency contraception in the future, compared with 332/360 (92%) in the levonorgestrel group (P = .28), whereas 349/364 (96%) and 341/361 (95%) of women, respectively, said they would recommend the same method to a friend.
This randomized study suggests that mifepristone is at least as effective as levonorgestrel for emergency contraception. However, because of the early conclusion of the study, we cannot rule out the possibility that levonorgestrel could be less effective than mifepristone. The study showed high acceptability for both methods used, although only one third of the women in each group returned the acceptability questionnaires. Some women felt the number of levonorgestrel tablets and the dose timing was a disadvantage to the levonorgestrel regimen. Adverse effects experienced by women in both groups were similar, although women in the mifepristone group were more likely to experience a delayed onset of the subsequent menstrual period after treatment, similar to findings reported in previous studies.9–11 Furthermore, women in the levonorgestrel group were more likely to experience an early onset of the subsequent menstrual period after treatment, compared with those in the mifepristone group.
Unfortunately, expiration of the shelf life for the mifepristone tablets necessitated an early conclusion of the study, and therefore, the initial estimated sample size was not achieved. However, data analysis using the equivalence criterion showed that the failure rates for mifepristone (1.5%) did not cross the stipulated 95% CI of 1.1% for the difference between the 2 treatments, whereas that for levonorgestrel (2.3%) did, suggesting that we cannot rule out the possibility that levonorgestrel could be less effective than mifepristone.
This study was not blinded, and both patients and medical staff were aware of the emergency contraception method used. Although the primary outcome for the study was unplanned pregnancy, women's acceptability of the method was an important secondary outcome for the study.
Only one third of the women returned the acceptability questionnaires, and we acknowledge that this may have introduced bias and should be taken into consideration when interpreting the study findings. Furthermore, only 37% of the women attended the Family Planning Clinic for follow-up, a problem reported in previous studies involving a similar study group.12 However, a further third of women had a telephone follow-up and, through this means or subsequent attendance at Family Planning or hospital, treatment outcome was known for a total of 84% of women taking part in the study. The Aberdeen Royal Infirmary, where the majority of women (2,019/2,043, 98.8%) who took part in the study were recruited, is the only referral hospital within a 50-mile radius for both gynecologic and maternity cases. The population base is largely stable, with little migration (outmigration including deaths of about 3%).13,14 It would have been likely that, if any pregnancies had resulted, these women would have attended antenatal care or the pregnancy advisory service (for termination of the pregnancy). We acknowledge that, for the small group in Dundee who had no follow up (9/2043, 0.4% of the total number of women), no similar criteria (hospital databases or notes) were available, and the possibility remains that some pregnancies may have been missed. However, it would have been expected that these women would have returned for a pregnancy test or for counseling about options if they had conceived, as seems to be the pattern experienced in Family Planning services.
Women who presented beyond 72 and up to 120 hours from unprotected sexual intercourse were offered a copper IUD, and those declining were offered entry to the study. We acknowledge that this may have introduced bias. However, in a previous report from this center, Ashok et al15 reported a study of late postcoital contraception over a 2-year period and showed that only 15/219 (6.8%) women wished to have the copper IUD fitted15 We, therefore, believe that the numbers declining entry to the study in favor of having a copper IUD fitted would have been very small.
Women in the levonorgestrel group were more likely to have a longer interval between sexual intercourse and treatment. Some women may have taken the first levonorgestrel tablet later in the evening on the day of presentation to allow the second tablet to be taken at a convenient time 12 hours later, a point highlighted in the acceptability questionnaires. The clinical relevance of this difference may not be highly important as such, but it highlights the potential implications for women's compliance with the 2-dose emergency contraception regimen. That difference, however, was not noted in those receiving emergency contraception beyond 72 and up to 120 hours, although a smaller number of women was available for comparison in that group.
Previous studies have excluded women with more than one episode of unprotected sexual intercourse at presentation.4 We included that group of women to allow the study to simulate standard nonstudy emergency contraception use. None of these women where the outcome was known had conceived.
Although the standard levonorgestrel regimen uses 2 doses of 750 μg given 12 hours apart (Schering. Summary of Product Characteristics 1999),16 studies have suggested similar serum concentrations for single- and two-dose administration,17 and the recent WHO study showed similar efficacy.2 Following on from this, the levonorgestrel product license was changed to single-dose administration in October 2003 (Schering. Summary of Product Characteristics 2003). Administering the levonorgestrel regimen as a single dose could improve compliance and allow the tablets to be taken at more convenient times for women.
Piaggio et al18 reported a meta-analysis of the randomized trials comparing the different doses of mifepristone (5–600 mg) in emergency contraception. This showed that there was no significant overall dose effect. However, there was evidence of a small dose effect in the lower-dose range (< 50 mg), with an increase in pregnancy rates by a factor of 1.6 when 10 mg was used instead of 25 mg. Nonetheless, in terms of the number of women needed to treat, using 10 mg in place of 25 mg implied having one extra pregnancy for every 146 women requesting emergency contraception. The authors concluded that, in balance, it might be more beneficial to use a lower dose to allow a greater number of women to have access to treatment.
In conclusion, this study shows that low-dose mifepristone (10 mg) is as effective as the levonorgestrel regimen (750 μg in 2 doses given 12 hours apart). Adverse effects were minimal and similar in both groups. Both regimens were reported to be highly acceptable to women. Further research is needed to assess the efficacy of the different doses of mifepristone and to address other needs of women requesting emergency contraception, including screening for sexually transmitted infections and contraceptive needs.19
1. WHO Task Force on Postovulatory Methods of Fertility Regulation. Randomised controlled trial of levonorgestrel versus the Yuzpe regimen of combined oral contraceptives for emergency contraception. Lancet 1998;352:428–33.
2. von Hertzen H, Piaggio G, Ding J, Chen J, Song S, Bartfai G, et al. Low dose mifepristone and two regimens of levonorgestrel for emergency contraception: a WHO multicentre randomised trial. Lancet 2002;360:1803–10.
3. Webb AM. Emergency contraception. BMJ 2003;326:775–6.
4. Cheng L, Gulmezoglu A, Oel C, Piaggio G, Ezcurra E, Look P. Interventions for emergency contraception (Cochrane Review). In: The Cochrane Library, Issue 3, 2004. Oxford: Update Software.
5. Fasoli M, Parazzini F, Cecchetti G, La Vecchia C. Post-coital contraception: an overview of published studies. Contraception 1989;39:459–68.
6. Trussell J, Rodriguez G, Ellertson C. New estimates of the effectiveness of the Yuzpe regimen of emergency contraception. Contraception 1998;57:363–9.
7. Piaggio G, Pinol AP. Use of the equivalence approach in reproductive health clinical trials. Stat Med 2001;20:3571–7.
8. Schlesselman JJ. Case-control studies: design, conduct, analysis. New York (NY): Oxford University Press; 1982. p. 203–6.
9. Webb AM, Russell J, Elstein M. Comparison of Yuzpe regimen, danazol, and mifepristone (RU486) in oral postcoital contraception. BMJ 1992;305:927–31.
10. Glasier A, Thong KJ, Dewar M, Mackie M, Baird DT. Mifepristone (RU 486) compared with high-dose estrogen and progestogen for emergency postcoital contraception. N Engl J Med 1992;327:1041–4.
11. Liu JH, Garzo G, Morris S, Stuenkel C, Ulmann A, Yen SS. Disruption of follicular maturation and delay of ovulation after administration of the antiprogesterone RU486. J Clin Endocrinol Metab 1987;65:1135–40.
12. Ashok PW, Stalder C, Wagaarachchi PT, Flett GM, Melvin L, Templeton A. A randomised study comparing a low dose of mifepristone and the Yuzpe regimen for emergency contraception. BJOG 2002;109:553–60.
13. Hall MH, Campbell DM, Fraser C, Lemon J. Mode of delivery and future fertility. Br J Obstet Gynaecol 1989;96:1297–303.
14. Pandian Z, Bhattacharya S, Templeton A. Review of unexplained infertility and obstetric outcome: a 10 year review. Hum Reprod 2001;16:2593–7.
15. Ashok PW, Wagaarachchi PT, Flett GM, Templeton A. Mifepristone as a late post-coital contraceptive. Hum Reprod 2001;16:72–5.
16. Tremblay D, Gainer E, Ulmann A. The pharmacokinetics of 750 microg levonorgestrel following administration of one single dose or two doses at 12- or 24-h interval. Contraception 2001;64:327–31.
17. Johansson E, Brache V, Alvarez F, Faundes A, Cochon L, Ranta S, et al. Pharmacokinetic study of different dosing regimens of levonorgestrel for emergency contraception in healthy women. Hum Reprod 2002;17:1472–6.
18. Piaggio G, von Hertzen H, Heng Z, Bilian X, Cheng L. Meta-analyses of randomized trials comparing different doses of mifepristone in emergency contraception. Contraception 2003;68:447–52.
© 2004 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
19. Kettle H, Cay S, Brown A, Glasier A. Screening for Chlamydia trachomatis infection is indicated for women under 30 using emergency contraception. Contraception 2002;66:251–3.