(Image by Michal Jarmoluk from Pixabay)

One of the features of the abortion landscape that is relatively recent is the idea of “abortion pill reversal”. Pro-life advocates hail APR as a life-saving option that should be made available to women, and some believe we should promote APR via public information campaigns. Pro-choice people say APR is unproven and dangerous medicine, puts women’s lives at risk, and simply does not exist. 

Who is right? In order to get a proper understanding, we need to first understand how medical abortion actually works. For a comprehensive treatment of the issue, I recommend this post by Petra Wallenmeyer. Here, I’ll give a shorter overview of how APR works, and then consider some of the issues it raises for the pro-life movement.

How does APR work?

The abortion “pill” is actually two pills, mifepristone and misoprostol. Mifepristone is taken first, misoprostol a few days later. Abortion reversal refers to taking high doses of progesterone after taking mifepristone, but before taking misoprostol.

Pregnancy is associated with high, and constantly increasing, levels of progesterone. Progesterone is the dominant hormone during the luteal phase of the menstrual cycle (from ovulation to menstruation). Progesterone sustains the lining of the uterus, preventing the menstrual bleed. Progesterone is therefore necessary for a healthy pregnancy: without sufficient progesterone, the uterine lining will not be capable of sustaining the embryo. In fact, when progesterone levels decline towards the end of a luteal phase in which implantation did not occur, menstruation is triggered. If implantation did occur, progesterone continues to be produced in the ovary for the first trimester, and by the placenta from the second trimester onwards.

Mifepristone, the first drug taken in a chemical abortion, blocks the production of progesterone. Taking mifepristone during pregnancy therefore compromises the quality of the uterine lining and blocks blood flow to the embryo or foetus. The theory behind abortion pill reversal is that taking a high dose of progesterone can reverse the consequence of the mifepristone, allowing the pregnancy to continue to develop as normal. 

There is nothing particularly new, or novel, (or magic!) about taking progesterone in pregnancy in an attempt to sustain the uterine lining. In fact, progesterone is regularly prescribed as a matter of course to women undergoing fertility treatment such as IVF, and is also prescribed for women who may be at risk of miscarriage. APR is not some new, fantastic, wonder-drug – it’s just progesterone, taken under different circumstances.

Misoprostol causes the uterus to contract and completes the abortion. It is not possible to reverse the effects of misoprostol. This is why abortion pill reversal, if it is possible at all, is only possible if high doses of progesterone are taken before, and without, taking misoprostol.

In order to determine the efficacy of abortion pill reversal, it is important to establish whether the high dose of progesterone actually counters the effects of mifepristone, or whether simply not taking misoprostol would have the same effect. For this reason, it is important to compare women who took mifepristone and then took high doses of progesterone with women who took mifepristone and then did nothing. Comparing women who took mifepristone and then high doses of progesterone with women who took both mifepristone and misoprostol is not an accurate comparison.

Ok, but does the progesterone actually work?

Any relevant studies on this issue are summarised at the end of this post. The most we can say from them is that there is some evidence that abortion pill reversal is possible, with no negative side effects for woman or baby. However, we do not have the kind of evidence that would generally be required in order to prescribe, much less promote, a new treatment or drug. It is particularly worrying that most of the evidence we have is based on one paper published in a low-quality journal that has questionable publication practices.

There are clear ethical barriers to performing a randomised controlled trial on abortion pill reversal (or indeed abortion itself) and so any studies in this area will be limited by sample selection bias: in other words, whether the sample of women studied is representative of the general population, and how to draw accurate conclusions given the lack of a proper control group. The fact that no information is given on how or why the 754 women in the Delgado study decided to pursue abortion pill reversal, or on the number of women who didn’t pursue abortion pill reversal, is a bit of a red flag here (although the fact that the study spans four years ameliorates this concern for me somewhat).

What should pro-life activists do?

It seems clear that claims by pro-choice advocates that APR is unproven and/or unsafe are a bit of a stretch. However, I think pro-life activists should probably not actively promote APR for two main reasons: moral hazard, and maximising our impact.

Moral hazard is the term used to describe the possibility that being protected against the costs or downsides of an action increases the likelihood that you will undertake that action. For example, if you have insurance, you might be more likely to undertake risky behaviour. It is at least possible that large scale, or even small scale, promotion of APR, will inadvertently cause a woman who is unsure about abortion to be more likely to have an abortion, because she mistakenly believes she can change her mind if she wants to.

We don’t know how likely this is, because no research has been performed on whether providing information on the possibility of abortion pill reversal makes the woman more or less likely to proceed with the abortion in the first place. There is also no evidence on the factors that make women consider abortion pill reversal in the first place, or how prevalent this is. In short, there is no research on the behavioural and informational aspects of APR. On the other hand, what limited studies we have indicate that once a woman takes mifepristone, her baby is more likely to die than not, even if she does subsequently take progesterone. Given this background, efforts to legally require doctors to inform women of the existence of APR, for instance (some US states have introduced or attempted to introduce these requirements), seem particularly misguided. 

You might think that it is highly unlikely that someone could so badly get the wrong end of the stick. However, medical information is quite difficult to communicate clearly, and it’s actually very easy for someone to get a garbled message, especially if the message is coming from a general informational campaign. At a minimum, publicising APR is a big risk to take.

When it comes to maximising our impact, it’s worth reflecting on the fact that APR is the last line of defence. Once a woman takes the first set of pills, there’s a good chance her baby will die. It’s great that we can save some babies in this situation, but it really is leaving it to the very, very, very last minute, and for a lot of babies it will just be too late even if the woman does change her mind and access APR. The pro-life movement has limited time and resources, and should reflect on whether focusing efforts on APR may deflect from other activities, such as 

  • reducing unplanned pregnancies in the first place
  • changing people’s minds about abortion so they would not opt for abortion even if they have an unplanned pregnancy
  • changing social structures to make Ireland more pregnancy- and parenting-friendly so that the demand for abortion even from people who don’t have an ethical problem with abortion reduces as a matter of course
  • changing the law so that even if there are people who want to get abortions, it’s harder to do so, and 
  • providing real practical supports for people who in crisis pregnancy.

Finally, one thing that would really help is a high-quality medical study! Given the difficulties with randomised controlled trials on this topic, such a trial is probably not going to happen, but a large scale study on the outcomes of women who have availed of APR, conducted under current academic standards and published in a high-quality outlet, would be a real addition. This would challenge the narrative that the pro-life movement is “anti-science”, or promoting “unproven, unregulated and harmful” medicine, and would give the us confidence in what medical facts we can provide around APR.


Appendix: relevant studies

Yamabe et al. (1989): Mifepristone was administered in pregnant rats, with progesterone subsequently administered to a treatment group of the rats. Progesterone levels were higher in the treatment group compared to a control group. The study therefore supports the theorised mechanism of action for progesterone. (Source)

Delgado and Davenport (2012): a case study of seven women who received intramuscular injections of progesterone after having taken mifepristone. One woman was lost to follow up; of the six remaining, four carried to term. This implies a 66% success rate ignoring the seventh woman, or a 57% success assuming she did not carry to term. Given the tiny sample size, the confidence interval around these numbers is massive and these figures should not be quoted. This study should be considered as exploration of concept rather than providing quantitative evidence. Of the four babies born, there were no neonatal complications or birth defects noted.

The Journal is well-regarded, Scimago Q1 in the field general medicine.

Grossman et al. (2015): carries out a systematic review of pregnancy continuation after mifepristone is taken. Only the above study was found to specifically examine abortion pill reversal. Other studies reviewed examined continuation of pregnancy following mifepristone alone. In the latter category, the proportion of women with a continuing pregnancy 1-2 weeks after taking mifepristone ranged from 8% – 46%. The CI for the Delgado and Davenport study spans this range and so this study concludes that abortion pill reversal is not supported by the evidence. (Source)

The Journal is well-regarded, Scimago Q1 in the field Obstetrics and Gynecology.

Raymond et al. (2016): This is a funny study: it seems to have accidentally indicated that APR is possible. The authors wished to study whether administering a contraceptive injection, known as Depo provera, at the same time as performing a medical abortion, would lead to lower rates of future pregnancy, rather than waiting for women to return for the contraceptive injection after the abortion was completed. Depo provera is a high dose of progestin, a different type of progesterone to that typically administered in APR. The study found that pregnancy continued in a higher proportion of women who received Depo provera at the same time as they received mifepristone compared to those who received Depo provera after their abortion was completed, although the difference was quite small (~5%). (Source)

Bhatti et al. (2018): reviews the evidence on abortion pill reversal. They include Delgao and Davenport in the review but dismiss its findings on similar grounds to Grossman et al. Bhatti et al. conclude that there is a paucity of evidence for abortion pill reversal and the requirement by multiple US states that physicians notify women presenting for abortions of the availability of abortion reversal is bad science. (Source)

The Journal is well-regarded, Scimago Q1 in the field Obstetrics and Gynecology.

Delgado et al. (2018): This is the most comprehensive treatment of the issue. They criticise Grossman et al. for too wide an inclusion of studies. For example, Grossman does not distinguish between incomplete abortion (foetal or embryonic demise but incomplete evacuation of the uterus) vs continuing pregnancy (foetal or embryonic survival). They provide a new refined review and find that the mean embryonic survival rate following mifepristone is 12.6%. The sample size is again limited and so the CI is wide.

Delgado goes on to document the pregnancy outcomes of 754 patients who were given high doses of progesterone having taken mifepristone. The patients “were pregnant women… who had taken mifepristone, but had not yet taken misoprostol, and were interested in reversing its effects.” The women were referred to a physician in their locality having contacted a hotline seeking information on abortion pill reversal. The women contacted the hotline over a period of four years. Information on how the women were made aware of the website, hotline or indeed abortion pill reversal itself are not given.

The study reduced to 547 eligible subjects who underwent abortion reversal therapy, of whom 257 had a live birth and a further four reached 20 weeks gestation and then were lost to follow up. This equates to a reversal rate of 48%. This figure compares well with the pregnancy continuation rates found in the reviews by Grossman and Delgado in the absence of taking progesterone. Pregnancy continuation was highest when (a) progesterone was administered via intramuscular injection or orally and (b) mifepristone was taken at a higher gestational age.

This journal is poorly rated internationally (Scimago Q4 – the lowest tier, in the field health policy). The journal Issues in Law & Medicine journal has been criticised by pro-choice groups for promoting conservative and anti vaccination views. (Source)

Creinin et al. (2020): This was the only attempt to conduct a randomised controlled trial to establish the efficacy of progesterone as abortion pill reversal. Twelve women were enrolled in a trial. Two dropped out, leaving ten women, five of whom were given progesterone and the other five were given a placebo. Three women had a significant hemorrhage, of whom two were in the placebo group and one in the progesterone group, and at this point the trial was stopped. The woman who was in the progesterone group suffered the least adverse effects of the three (requiring no aspiration, while both women in the placebo group required aspiration and one required a transfusion); however given the tiny sample size it is impossible to know whether this is indicative of progesterone countering the effects of mifepristone or whether this is coincidence. Two of the five embryos in the placebo group survived versus four of the five in the progesterone group, but again it is impossible to draw any conclusions from this fact based on the tiny sample size. 

What this trial does appear to indicate is that taking mifepristone alone without following up with misoprostol is dangerous, leading to the potential for hemorrhage. Whether progesterone subsequently reduces this danger cannot be established from this trial. It is unfortunately unlikely that another will therefore be conducted due to safety concerns.

The Journal is well-regarded with Scimago Q1 ranking in the field Obstetrics and Gynecology.