Academic OB/GYN

by Sharon Phillips, MD and Nicholas Fogelson, MD

Lots of people have been talking about Ulipristal acetate (Ellaone), a new emergency contraception option now available in the UK.  The buzz is that it is effective for 5 days instead of 3.

Currently in the US we have only one option for EC (Emergency Contraception): levonogestrel (Plan B), which is FDA labeled for use up to 72 hours after unprotected intercourse.  However, we now know that it is effective for up to 5 days after unprotected sex(1), though the efficacy seems to decrease the further out from the episode of intercourse you are.

It’s great to have a new EC option.  Having 2 options might help bring the price down (Plan B can range from $30 to $60 for one dose).  However, it may not be any better than what we already have.

A previous study published by Creinin et al demonstrated that Ulipristal was as effective as levonorgestrel (Plan B) at preventing pregnancy when used up to 72 hours after unprotected intercourse.  A subsequent study in the Green Journal suggested that Ulipristal had effectiveness up to 5 days, but no comparison to levonorgestrel was made(2).  This month, a study of Ulipristal was published in the Lancet  that compared Ulipristal to levonorgestrel in a randomized trial, showing and that it may be more effective than levonorgestrel in the 4-5 day time period(3).  However, the study design of this paper precludes us from making this conclusion definitively. And this is why you can’t just read the one-liner you see in the paper.

Statistics can be complicated, but basically this study was not designed to show that ulipristal was better than levonorgestrel, only to show that it was not worse.  In order to show conclusively that one drug is better than another, you need lots and lots of people to participate in the study.  This study had around 1900 women for the final analysis.  This seems like a lot, but only a small percentage of women in the study ended up being pregnant despite using either of the emergency contraceptive options.  In the group that took the new drug, ulipristal, there were a total of 15 pregnancies.  In the group that took levonorgestrel, there were a total of 25 pregnancies.  Those that took the newer drug were 40% less likely to become pregnant than those who took the older one.  It might seem that this shows that ulipristal is better.

Unfortunately, statistics is not that simple.  In life there are always random variations.  For instance, if you toss a coin 3 times, you might get 3 heads, 3 tails, or some other combination.  Whenever we do statistical calculations, we try to take into account that random variation by giving ourselves a range that we’re pretty sure the true answer lies within.  When the authors of this study did just that, they calculated that women who took ulipristal were anywhere from 70% less likely to get pregnant to 10% more likely to get pregnant.

The reason the researchers found themselves in this situation was that they never intended to prove the new drug was better, just that it wasn’t worse.  They would have had to recruit lots more women in order to prove that it was better (because the more people you have in your study, the less random variation there is).  But that costs time and money, and the pharmaceutical company didn’t want to spend extra money for this, especially if it turned out to be no better than the drug we already have.

Of note, they did perform a meta-analysis, which involves adding up the data from several different studies with the hopes of having more people to do the analysis on, therefore less random variation.  This method is fraught with difficulties, for reasons we won’t go into here.  They did find that it just barely was better than levonorgestrel at 72 hours and was significantly more effective at 120 hours.  However, as we said, this statistical method is very problematic.

Also problematic is that this research was paid for by the company that stands to profit from the results.  Such an arrangement can often lead to the company only allowing results that are favorable to be published and burying any results that are not favorable (we saw this with many anti-depressants in the ’90s).  Though we cannot avoid the need for industry funding at times, there are a number of mechanisms for limiting this type of bias.  This study was registered, as required, with ClinicalTrials.gov, a government trial tracking system that exists to limit non-publication of negative trials.  Lacking this registration requirement, it would be possible to bury a negative trial such that no one would ever knew it existed.  With this registration, it can be verified that a trial was performed but never published, disincentivizing non-publication of negative work.  Though this registration is important, also important is a study statement that says that the sponsoring entity (in this case HRA Pharma, the company that will market ulipristal) had no access to study data until the study was submitted to publication.  This precaution prevents any possibility of publication bias from the funding source.  This study unfortunately lacks this statement, and as such we do not know if this was the case.  As this study is authored by three physicians directly employed by HRA, it is fairly certain that the study data was known to HRA prior to publication.

The bottom line is that this is another option now available, it works at least as well as levonorgestrel, and might work better than levonorgestrel in the 3-5 day range, though that is unclear.  New contraceptive agents offer new choices to women for preventing unwanted pregnancy, and that’s a good thing all around.

The other bottom line is that the news media routinely misinterprets scientific studies and it’s important to have some knowledge of how statistics works in order to interpret such studies.

Sharon Phillips, MD

Nicholas Fogelson, MD

Dr Phillips is a Family Planning Fellow at Albert Einstein College of Medicine  / Montefiore Medical Center in Bronx, NY.

1.            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 Dec 7;360(9348):1803-10.

2.            Creinin MD, Schlaff W, Archer DF, Wan L, Frezieres R, Thomas M, et al. Progesterone receptor modulator for emergency contraception: a randomized controlled trial. Obstet Gynecol 2006 Nov;108(5):1089-97.

3.            Glasier AF, Cameron ST, Fine PM, Logan SJ, Casale W, Van Horn J, et al. Ulipristal acetate versus levonorgestrel for emergency contraception: a randomised non-inferiority trial and meta-analysis. Lancet  Jan 28.