This article appears in the February 2002 edition of the Catholic Medical Quarterly

Return to February 2002 CMQ

Morning After Pills: How Do They Prevent Pregnancy?

P J Howard

Talk given at a conference organised by the Centre for Bioethics and Public Policy at The Royal Society of Medicine on Wednesday 28th November 2001 with subsequent discussion

Modes of Action of Morning After Pills (MAP)

There are several stages at which MAP may prevent pregnancy. MAP may have a true contraceptive effect by inhibiting ovulation and by enhancing the 'barrier� effect of cervical mucus on the passage of sperm through the cervical canal. There is a theoretical possibility that MAP may alter tubal motility and hence alter the passage of sperm up the Fallopian tube or alter the transit of the egg and conceptus moving in the opposite direction towards the uterine cavity. MAP may also disrupt implantation or cause the loss of a newly implanted embryo. There are ethical constraints on the study of the effects of MAP on implantation and for studying its effects on early pregnancy. Effects of MAP on implantation are clearly abortifacient as they cause the expulsion of the early human embryo from the womb with its inevitable death.

Anti-ovulatory Effect

The antiovulatory effect of MAP will depend upon its ability to inhibit the LH surge and the timing of administration in relation to the menstrual cycle. There would still remain a risk of pregnancy if MAP did not prevent, but merely delayed, ovulation.

Several studies have shown that Levonorgestrel has only a limited effect on preventing ovulation (<15%) even when deliberately administered just before the LH surge 2

In one study, Levonorgestrel 0.75 mgs was taken after intercourse during the peri-ovulatory period in 361 women. It was shown that ovulation was inhibited in 14.4% of cycles as shown by analysis of basal body temperature charts3. In another study designed to examine the effects of levonorgestrel as a post coital contraceptive, 77 women received an oral dose of 0.4 mg per coitus for 1011 cycles and 27 women were administered 0.75 mg per coitus for 226 cycles. In the first dose group seven women became pregnant (Pearl�s index 8.3), while two pregnancies resulted in the second group, one of the latter because of faulty drug intake (uncorrected Pearl�s index 10.6; corrected Pearl�s index 5.3).

Effects on Cervical Mucus

An effect on cervical mucus is unlikely to be important if given after intercourse has taken place. MAP administration is advised up to 72 hours after intercourse: the effects on cervical mucus arise within 24 hours of administration4. Moreover, sperm could be recovered from the uterus as soon as 30 minutes after intercourse in one study5 and within 4 hours of coitus in 8 of 10 women in another6.

Disruption of Implantation

Implantation is necessary for embryonic development beyond the blastocyst stage. However, prior to implantation the embryo is no less dependent on the maternal environment.

Successful implantation depends upon signalling between the embryo and mother, embryonic maturation, endometrial receptivity, synchrony between embryonic maturation and the endometrial 'window of implantation�.

Embryonic Maturation

As the embryo matures, various genes are activated, leading to the production of various factors involved in feto-maternal 'signalling� and the synthesis and expression of various cell surface receptors that are essential to adhesion to the endometrium and implantation. Prior to implantation the blastocyst must 'hatch� from the zona pellucida. The outer cell mass differentiates into specialised syncytiotrophoblast cells to form the placenta whilst the inner cell mass gives rise to the embryo. Different surface receptors are expressed as the process of implantation progresses. Secretion of HCG 'rescues� the corpus luteum, which is responsible for secreting progesterone and oestrogen, essential to the establishment and maintenance of pregnancy.

Endometrial Factors in Implantation

The endometnium undergoes cyclic growth and development. Endometrial cells are regulated directly by ovarian steroids and indirectly by various growth factors and cytokines. Maximal uterine re ceptivity is on cycle days 20-24 "window of im plantation". Cell surface projections or endometrial pinopodes are thought to be crucial to the process of adhesion and implantation.

The immunological relationship between the mammalian fetus and its mother during pregnancy has been considered similar to that between a transplanted allograft and its recipient. Aberrant implantation can cause a variety of clinical problems including miscarriage, intrauterine growth, retardation and pre-eclampsia.

The human placenta arises primarily through proliferation, migration and invasion of the endometrium and its vasculature by the embryonic trophoblast.


Studies in mice have shown that Levonorgestrel, administered in the form of sub-dermal implants following mating and ovulation, can not only prevent implantation but also cause resorption of those embryos that did implant. These effects depend upon the timing and dose of Levonorgestrel7.

A very recent study has shown that the Yuzpe method is effective between 72 and 120 hours after unprotected intercourse8. The Yuzpe method (using a combination of ethinyloestradiol 0.1mgs with Levonorgestrel 0.5 mgs given twice at 12 hour intervals) is thought to have a similar mode of action as Levonelle, though it is perhaps less effective. Whilst the fact that Levonorgestrel and Yuzpe may be more effective in preventing pregnancy the sooner they are given, it has been argued that, if they have, a delayed action beyond 72 hours, this pro vides evidence of a post-implantation effect.

Yuzpe and Levonorgestrel compared. (Lancet 1998;352:428-22).

1998 women were recruited into the study and given either Levonorgestrel 750 mcg (repeated 12 h later) or the Yuzpe regimen (100mcg ethinyloestradiol and 500mcg of Levonorgestrel repeated after 12 h) starting within 72 h of a single episode of unprotected intercourse. The primary outcomes were the crude pregnancy rates compared with expected pregnancy rates and side effects.

Women were included if they were otherwise healthy, had regular menses and no recent pregnancy. They were excluded if they had recently been pregnant or recently taken the oral contraceptive pill, or if they were breast-feeding. Samples of urine on blood were taken at entry for pregnancy tests, although the absence of pregnancy at entry was clearly not an exclusion criterion - since 4 of the 42 subsequently found to be pregnant during the study had been pregnant at entry.

Overall Levonorgestrel caused an 85% (74-93%) reduction in expected pregnancy rates and Yuzpe a 57% reduction (39-71%). However, in both groups approximately one third of women had further acts of intercourse: in these women the pregnancy rates were higher, but not significantly so for Levonorgestrel: Levonorgestrel...  6/372 [1.6%] v. 5/602 [0.8%] Yuzpe...............19/360 [5.3%] v. 12/619 [1.2%]

Hence, from the paper, the reduction in expected pregnancy rate in those who did not admit to further acts of intercourse were for Levonorgestrel 89% and Yuzpe 73%. For those who had further acts of intercourse and assuming that the expected pregnancy rate was between one and two times that of a single episode of intercourse, the pregnancy rates following levonorgestrel were between 90% - 79% (n=372) and for Yuzpe 64% - 28% (n=360). In other words, pregnancy rates when additional acts of intercourse occurred after taking levonorgestrel were not greatly different when compared to pregnancy rates after a single episode of pre-treatment intercourse. However, the reduction in expected pregnancy rates was less with Yuzpe. This suggests that Levonorgestrel acts mainly to disrupt implantation rather than having a significant contraceptive effect on cervical mucus or ovulation.

The issue of exactly how MAP works is of importance to women who will distinguish between a true contraceptive effect and an abortifacient effect. Indeed, in America, it is feared that some women may sue their doctors for failing to give adequate information for consent prior to taking MAP. "With out accurate information presented before prescribing, patients may experience emotional distress from an unanticipated result, an unforeseen side effect or the later discovery of a mechanism of action that is in conflict with their value system" McGaugham A. Am Fam Phys Nov 2000.


From the available literature, the main effect of Levonorgestrel is on the endometrium acting to disrupt implantation rather than acting at the level of the cervical mucus or by inhibiting ovulation.


  1. Pharmaco-kinetic study of Levonorgestrel used as a post-coital contraceptive. Shi YE, Zheng SH, Shu Yhm He C H, Yu PP, Fortherby K. Contraception. 1988:37(4):359-369.
  2. The effects of peri-ovulatory administration of Levonorgestrel on the menstrual cycle. Hapangama D, Glasier A F, Baird D T. Contra ception 2001 63:123-129
  3. A multicenter clinical study on two types of levonorgestrel tablets administered for post coital contraception. He-CH; Shi-YE, Xu-JQ; Van-Look-PF. Int-J-Gynaecol-Obstet. 1991 Sep; 36(1): 43-8
  4. Cyclic changes of cervical mucus in normal and progestin-treated women. Fertil Steril. 1966;17:63.
  5. Sperm survival in women. Rubenstein B B, et al. Fertil Steril 1951;2:51-19.
  6. Uterine flushing: a method to recover sperma tozoa and leucocytes. Williams Metal. Human Reproduction 993;8:2019-2026.
  7. An early pregnancy factor detected in human serum by the rosette inhibition test. Morton H, Rolfe B, Clunie G J. Lancet 1977;8008:394-7.
  8. A study of early pregnancy factor in activity in pre-implantation. Fan XG, Zheng ZQ. Am J Reprod Immunol 1997;37(5):359-64.
  9. Early pregnancy factor as a marker for the earliest stages of pregnancy in infertile women. Mesrogli M, Shneider J, Maas D H. Human Reproduction 1988;3: 113-5.
  10. Isolation, purification and partial characterisa tion of early pregnancy factor (EPE) from sera of pregnant women. Haq A, Mothi B A, Al Hussein K, Al-TufalilM, Hollanders J, Jaroudi K, Aiwaili N, Shabani M. Eur J Med Res 2001 29;6(5):209-2 14.
  11. Significance of the detection of early preg nancy factor for monitoring normal and disor dered pregnancy. Straube W, Loh M, Leipe S. Geburtshilfe Frauenheilkd 1988;48(1 2):854-8.
  12. Levonorgestrel as a postcoital contraceptive. Shriley B, Bundren J C, McKinney S. Contraception 1995 ;52(2):277-8 1.
  13. Effectiveness of emergency contraceptive pills between 72 and 120 hours after unprotected sexual intercourse. Rodrigues 1, Grou F, Jolly J. Am Obstet Gynecol 2001;184:531-7.
Dr P J Howard MA MD FRCP is Consultant Physician and Senior Lecturer in Medicine, Epsom and St Helier NHS Trust and St George�s Hospital Medical School.


Dr Peter Saunders (Chairman), Christian Medical Fellowship; Dr Graham Barker, Associate Medical Director of Schering Healthcare Ltd; Dr Philip Howard.

Dr Peter Saunders.
Now, perhaps we should start here before taking questions from the floor by giving Graham Barker an opportunity to respond to the two questions that Philip Howard has posed during his talk.

The first question you posed was what is the positive evidence that Levonorgestrel does not have an effect post-implantation? Is that correct?

The second question that you raised was how was it ethical to give the pill before establishing pregnancy or doing a pregnancy test? Is that correct?

Dr Graham Barker
Thank you. I think that the second one is easy and I can dispense of it in a few words. It wasn�t ethical and I am quite sure that it would not get past an ethical committee in this country now. I think it was wrong. No doubt about it.

With regards to the first one - the evidence of the post-implantation effect or lack of it from progestogens. I think I have two points to make. Firstly, I would say that I am always cautious of citing case histories from specific cases as evidence but in my position as the person responsible for pharmacovigilance at Schering Healthcare I am often asked what effects it will have if the woman has taken Levonorgestrel after she has become pregnant and it hasn t worked. There is no evidence of any harm. But that question is so often asked that I feel in itself it answers this one. There are no published studies that I know of, there are no ethical controlled trials that I know of, but there are an awful lot of case reports where established pregnancy has been exposed to Levonelle with no effect.

I would also ask Dr Howard a question myself - Could he explain for me, please, at which point in the development of the zygote to the embryo to an established pregnancy does the progestogen effect becomes a maintaining one, rather than, as he feels, a disrupting one?

Dr Philip Howard
I was quite interested in your comment that progesterone is a pro-pregnancy hormone, which of course it is. But that begs the question as to why on earth should it work to prevent pregnancy at all. It will have major effects on the endometrium. It could affect it at very many stages. I do not want to go into great detail in terms of the molecular biology and biochemistry of implantation. Clearly, the implantation process occurs within hours of fertilization in relation to all the preparatory stages that are necessary. It is a bit like the docking of a lunar module with the mother spacecraft around the moon. You have got to set off at the correct time, you have got to have the right trajectory, the right speed, you have got to have the right receptors and you have got to have good communication between the module and the space craft. If you get it wrong, you just go straight off into space. The analogy breaks down, however, because the implantation process of the human embryo is far more complicated than whatever they could have devised for lunar modules in Houston. The effect could very well be on many stages of implantation itself but could also operate through a luteolytic effect. You have quite rightly pointed out that the corpus luteum is essential for the production of progesterone and oestrogen to sustain a pregnancy. Now, if it has an anti-luteal effect then it will, of course, disrupt the pregnancy as pregnancy cannot be maintained without the corpus luteum. So an anti-luteal effect would be very important. What the early embryo has to do is to signal to its mother its presence and 'rescue� the corpus luteum and stop it regressing. So, if the corpus luteum goes, you are not going to sustain a pregnancy.

I would like now to come back and ask you a question if I may. The absence of evidence does not mean the evidence of absence. Even if that study [comparing Yuzpe and Levonorgestrel] was un ethical, and it was done in 1998, not long ago, by the World Health Organization and presumably put through its ethical committees, I would still like to know how many of those 1955 women that did not become pregnant were pregnant at entry. What is the figure? I would like that information. Of course, you have admitted that it would be unethical now. I would say that it was unethical then, to do that sort of study and to deliberately give Levonorgestrel or Yuzpe to women who are actually pregnant. But, even though this type of study is unethical, I am suggesting to you that the study has actually been done. It involved nearly 2000 women in a world wide study, supported by the World Health Organization which, you are now telling me as a representative of Schering, was unethical. But where is the data? Is it not equally unethical to obtain that data by taking samples of urine or blood for pregnancy tests and not to publish it?

Dr Peter Saunders
So the question is, if the research was done, is the data available and if so is it accessible and why has it not been published? Graham Barker?

Dr Graham Barker
I am sorry, I have no idea. It is not a company study. It is a WHO study. We had no input or sponsorship of it, and I would just point out, though, that the ethical problem that we have discussed is only one very small part of that study, I would not like to think that I am telling you that the whole study was unethical. But that small part of it has seemed to have slipped through. I would agree that that may very well have been wrong. Dr Howard did bring up with that study his interesting figures where he has shown that repeated intercourse after treatment appeared to have some effect on the efficacy; and it has just crossed my mind, and I would be interested in his thoughts, that that is showing us another possible mechanism for Levonorgestrel for providing a barrier to spermatozoa which we know from the POP (progesterone only pill) comes within hours and is remarkably effective within a very short space of time. I believe that repeated intercourse after taking Levonelle may well be blocked by that [effect].

Dr Philip Howard
Well if you see that sort of effect, it very much suggests an effect on the 'final common pathway� which I would suggest would be on the endometrium and not on ovulation or cervical mucus. If you were to see a grossly different effect, then you would expect the effects to be additive. But the fact that the effects are statistically very much the same would suggest that you are affecting the endometrium. In other words, if you give Levonorgestrel post-coitally, I have argued, and there is experimental evidence to suggest, that the effect on ovulation is going to be minor and usually only effective in this way if given before intercourse. One study of Levonorgestrel in which it was given just before the LH surge, showed that it doesn t usually prevent ovulation. It is really the prevention of ovulation that is important in populations of women and not just that ovulation is delayed, because if ovulation does occur but at a later stage, the woman is still at risk. But there have been various studies that have actually shown that its effect is minor pharmacologically. But of course it does matter, as Dr Barker has said, when actually you give it. In diverse populations of women it will be given at different times. But the fact that there does not seem to be a major additional effect or reduction in the effect on pregnancy if you like, by having further episodes of intercourse after the pill is given would suggest that whether you have inter course before or after you give levonorgestrel, it is affecting the endometrium - the final common pathway. That is how I would explain it.

Dr Peter Saunders
I wonder if our speakers could tell us about the effect on the biochemical markers for establishing very early implantation in the first two or three days because that seems to be a hugely important area in establishing the facts here.

Dr Graham Barker
I am afraid I cannot. My understanding is that it is when the human chorionic gonadotrophin (HCG) is sufficient to make a measurement to establish a pregnancy.

Dr Philip Howard
It has become important now in relation to infertility medicine and reproductive technology to be able to define exactly what the problem is in the infertile woman: whether it is a problem of ovulation, whether it is a problem of fertilization or whether it is a problem of implantation. A great deal of research work has been done in the clinical field to try and define those points. To some extent you can define ovulation by changes in body temperature, by changes in urinary hormones and of course, definitively, by ultrasound. A lot of these factors which I did not get to talk about are operative very early on in pregnancy. Early pregnancy factor (EPF) does arise long before human chronic gonadotrophin becomes apparent in the maternal serum A lot of these processes are experimental, but if you look hard enough you can detect pregnancy at a very early stage long before HCG.

Do you mean pregnancy, though, or do you mean morula or zygotes?

Dr Peter Saunders
A key question here is when pregnancy actually begins. Now we know that the Attorney General in 1983 said that the Abortion Act did not apply to the morning after pill: there was an assumption there. And the assumption generally in the medical profession now is that pregnancy does not begin until implantation. Could I ask each of the speakers when they think Pregnancy begins?

Dr Philip Howard
You can define pregnancy in two senses. You can either define it as the physiological state of the woman or you can define it in relation to the carriage of a conceptus (fertilized ovum) within the woman�s genital tract. Now ethically of course the important thing is, when does a genetically distinct human individual begin. Now there is nothing magical about implantation. In a biological sense the conditions which are required for the early growth and development of that early conceptus as it is passing down the Fallopian tube are no less critical over the first 4 or 5 days than they are subsequently in terms of that embryo�s early development and its survival. So the important ethical question is when does human life begin? When you look at the process of implantation, the more you see interactions at a very early stage between that conceptus signalling to its mother and her response. And even if you were to define pregnancy as the physiological state of the woman, you would find that from a biological point of view that woman�s physiology is being altered within hours of fertilization.

Dr Barker
I would not disagree with any of that. But a lot of those early changes do not progress to pregnancies. And to me a pregnancy is where it is self-sustaining and when the HCG is sufficient to switch-off the pituitary to suppress the production of any further ova.

Return to February 2002 CMQ