Levonorgestrel is More Effective, Has Fewer Side-effects, than Yuzpe Regimen

The Programme organized and supported a large-scale clinical trial to compare levonorgestrel, given as two 0.75 mg doses 12 hours apart, with the Yuzpe regimen for emergency contraception. A total of 21 centres in 14 countries (Australia, Canada, China, Georgia, Hungary, India, Mongolia, New Zealand, Nigeria, Panama, Slovenia, Sweden, the United Kingdom and the USA) took part in the study which was completed in 1997. The results were published in The Lancet last year (1998, 352:428–433) as the work of the Programme's Task Force on Postovulatory Methods of Fertility Regulation.

Background

Programme-supported research by Ho and Kwan (Human reproduction, 1993, 8:389–392) in Hong Kong had earlier suggested that a two-dose regimen of 0.75 mg levonorgestrel was as effective as the Yuzpe regimen but was associated with fewer side-effects in treatment given within 48 hours of unprotected intercourse. The Programme's new study set out to examine whether the findings from the Hong Kong trial could be confirmed in research on a larger scale and including women from different populations. Thus, a double-blind, randomized, multinational study was carried out using the same treatment regimens but extending the maximum delay between intercourse and the start of treatment to 72 hours.

A total of 1998 women were enrolled in the study, of whom 997 were randomly assigned to the Yuzpe regimen and 1001 to the levonorgestrel regimen. In all, outcome was unknown for 43 women, most of whom were lost to follow-up despite efforts to reach them. Thus, a total of 1955 women (979 in the Yuzpe group and 976 in the levonorgestrel group) was included in the analysis. All were healthy women with regular menstrual cycles of 24–42 days' duration, who had had only one act of unprotected intercourse during the treatment cycle, and who were asked to avoid further acts of unprotected intercourse during that cycle.

Women who were breast-feeding or had used hormonal contraception within the current menstrual cycle were excluded, as were those who had contraindications to hormonal contraception or were uncertain about the date of their last menses.

Randomization ensured that the women in both treatment groups had similar baseline characteristics. The participants in the study had a mean age of 27 years and a minority (22%) of them had used emergency contraception before. Similar proportions of women in both groups cited lack of contraception (56%) and failure of a barrier method (41%) as the reason for requesting emergency contraception.

Treatment began within 24 hours of unprotected coitus in nearly 50% of the women in each group and within 48 hours in more than 80%. A total of 42 women were found to be pregnant after treatment. However, retrospective urine analysis showed that four had already been pregnant on enrolment and the pregnancy status at admission of a further five could not be determined. Five women continued their pregnancies with normal outcomes and the others opted for termination.

Efficacy

The pregnancy rate was 3.2% [95% confidence interval (CI) 2.2–4.5] among women assigned to the Yuzpe regimen and 1.1% (95% CI 0.6–2.0) among those assigned to the levonorgestrel treatment. The crude relative risk of pregnancy for levonorgestrel compared with the Yuzpe regimen was 0.36 (95% CI 0.18–0.70). In both treatment groups, women who had further acts of intercourse had higher pregnancy rates than women without further intercourse (Yuzpe regimen 5.3% versus 1.9%; levonorgestrel 1.6% versus 0.8%). When the number of pregnancies observed in the levonorgestrel group (11) was compared to that expected without treatment (75), it could be concluded that the regimen prevented 85% of pregnancies. One the other hand, the Yuzpe regimen in this study prevented 41 of the 72 pregnancies that would have been expected to occur (i.e. 57%). These figures of prevented pregnancies are underestimates, however, because they include the four women who were pregnant and the five women whose pregnancy status was unknown at the time when the treatment was started.

An important finding of the study was the "consistent linear relationship" between efficacy and the time from intercourse to treatment. The trend for pregnancy rates to rise as treatment was delayed was found to be significant (p<0.01). Postponement of the first dose by 12 hours raised the odds of pregnancy by almost 50%. The percentages of pregnancies prevented at different exposure-to-treatment time intervals decreased from 95% for the first 12 hours to 47% for the interval of 61-72 hours (Figure 1). These percentages were calculated for the Yuzpe and levonorgestrel regimens combined, excluding four women who were pregnant at enrolment (The Lancet, 1999, 353:721).1 The finding of decreasing effectiveness with time elapsed since the act of sexual intercourse has major implications for emergency contraception services and for counselling of women.

Side-effects

The study found that levonorgestrel was better tolerated by the women than was the Yuzpe regimen. Nausea, vomiting, dizziness, and fatigue were all significantly less common among women who received levonorgestrel. For instance, nausea was reported by 23% of the women in the levonorgestrel group versus 51% of those in the Yuzpe group, while vomiting was reported by 6% and 19%, respectively. Also, other side-effects were less common in the levonorgestrel group. The time to resumption of menses was similar for women in both groups. For both groups combined, menses returned within three days for most women (57%), had an early onset for 15% of the women, and was delayed by more than seven days for 13% of the women.

Figure 1: Effect of delay on pregnancies prevented

The study produced the following findings of public health importance:

• the levonorgestrel treatment was better tolerated than the Yuzpe regimen;

• the levonorgestrel treatment was more effective than the Yuzpe regimen, in terms of both crude and adjusted pregnancy rates and pregnancies prevented; and

• both methods were more effective the sooner they were used after unprotected intercourse.

The single most important message from this study is that women should receive treatment as soon as is practicable after unprotected sex. Extrapolating from the significant trend found in the trial, it is estimated that treatment after 72 h will have even lower efficacy.

The authors admit that little is known about why the Yuzpe regimen of levonorgestrel with ethinylestradiol is less effective than treatment with levonorgestrel alone. The lower efficacy could, they suggest, be due to an interaction between the estrogen and the progestogen as well as to the lower dose of levonorgestrel used in the Yuzpe regimen. However, they point out that a separate assessment of the effects of the estrogen, the dose of levonorgestrel, and the interaction between the hormones "would require randomized clinical trials of questionable ethical value" since levonorgestrel has now been shown to have fewer side-effects and estrogen alone in the dose used in the Yuzpe regimen is unlikely to be effective.

The authors suggest that replacement of the Yuzpe regimen with levonorgestrel should improve the acceptability of hormonal emergency contraception and that family-planning programmes providing emergency contraception should consider making this change. Indeed, the study has already had a major impact on emergency contraception services. In 1996, on the basis of interim results from the study, the International Consortium for Emergency Contraception selected levonorgestrel as the method of emergency contraception for pilot introduction programmes in Indonesia, Kenya, Mexico and Sri Lanka. A two-pill pack of 0.75 mg levonorgestrel was subsequently registered for emergency contraception in China, Hungary, Kenya, Nigeria and Sri Lanka. A few months ago, the levonorgestrel-only method was also registered in the USA, and drug regulatory authorities in several other countries, including the European Union, are currently considering to register the method and update guidelines for providers. Where 0.75 mg levonorgestrel tablets are not available, levonorgestrel-only minipills (30 µg) can be used, although this regimen is not very convenient as it involves taking 25 pills per dose.

Obviously, the Yuzpe regimen will continue to be used in many places for quite some time although a shift to the more effective levonorgestrel treatment with its fewer side-effects is expected over time. As the two methods continue to be used, however, the concluding words of the authors of the study are worth remembering: "With either regimen, the sooner treatment starts, the better it works."

Go to Progress