Early progesterone cessation after in vitro fertilization intracytoplasmic sperm injection A randomized, controlled trial

Luteal phase supplementation with vaginal progesterone after in vitro fertilization/intracytoplasmic sperm injection can be safely withdrawn at 5 weeks’ gestation.

Graciela Kohls, M.D., Francisco Ruiz, M.D., María Martínez, M.D., Erik Hauzman, M.D., Gabriel de la Fuente, M.D., Antonio Pellicer, M.D., Juan A Garcia-Velasco, M.D.

Volume 98, Issue 4, Pages 858-862, October 2012


To investigate the effect of stopping progesterone (P) support at week 5 versus week 8 on ongoing pregnancy rate after in vitro fertilization (IVF)/intracytoplasmic sperm injection (ICSI).

Prospective, randomized, controlled trial.

University-affiliated infertility center.

A total of 220 patients with intrauterine pregnancy demonstrated by transvaginal ultrasound after IVF/ICSI.

Luteal phase support with micronized vaginal progesterone was suspended at week 5 or ar week 8.

Main Outcome Measures:
Ongoing pregnancy rate, miscarriage rate, and number of bleeding episodes.

Progesterone levels were similar on the day of the first pregnancy ultrasound exam (149 ± 108 vs. 167 ± 115 ng/mL). Significantly more bleeding episodes were observed in the first trimester in the group with early cessation of P supplementation (18.0 ± 2.6 vs. 7.2 ± 1.3 episodes). Miscarriage rates among singleton pregnancies were similar in the two groups (5/80 vs. 6/79).

Vaginal P supplementation after IVF/ICSI can be safely withdrawn at 5 weeks’ gestation, because cycle outcome was similar to conventional luteal phase support up to 8 weeks of pregnancy.

Clinical Trial Registration Number:

  • Micah Hill

    This study continues to be provocative for me and I had another discussion point. The authors are saying that the miscarriage rate between the two P4
    supplementation groups is similar. In such a study, you worry about the type II error, that is that the 2 groups are actually different when the study found that they were not different. With the power of .8, there is a 20% chance of such an error. Further, this study was powered to detect a 10% difference in miscarriage, but for me even 3-5% differences in miscarriage would be clinically significant. But to perform a study that minimizes the type II error and can detect smaller differences in miscarriage, you need a very large sample size, far greater than 220 patients. So while I admire this study, for me it is not powered adequately to determine that clinically significant differences in miscarriages do not exist with a reasonably low chance of error.

  • Dr Steven Ory commented about this article on his reflections paper about it–in the discussion on this forum when asked what he does in his practice he replied:

    Great question! Our group of 9 reacted much as the bloggers on the forum
    for Juan’s article. It prompted a spirited discussion. We are slow to
    let go of our dogma, yet much of our IVF practices are not evidence
    based and that is certainly true for progesterone supplementation. As a
    result of Juan’s submission and other RCTs, we have reduced our period
    of progesterone support from 10 weeks for fresh cycles and 12 weeks for
    frozen and recipient cycles to 8 weeks for all. Hardly a bold step. As
    our comfort level increases, I expect that we will go to 6 weeks but
    this process represents a consensus of 9 practitioners.

    his related article is at :

  • Micah Hill

    For programs that have implemented a shortened luteal support, what was patient reaction? For the reasons outlined by Paul and Isiah, I would imagine significant restistance from patients to shorten the course fo progesterone, even if the evidence supports it.

    • Micah– its important this study showed significantly more bleeding in the early stop group which without doubt will be concerning because i expect patients will think its cause and effect

      • Juan Garcia-Velasco, MD

        Absolutely right, but this is like bed rest after embryo transfer: we may advise patients to do regular life but, how many of them do??

  • Paul Brezina

    I think that this is a really valuable topic. Certainly progesterone support in very early pregnancy is valuable… I do not think this point is debated by most people. The results of this article are interesting. I agree with some of the comments below. Specifically, the low cost of continuing progesterone support until 8-10 weeks or so seems prudent. Until there are multiple such studies confirming this result, it would be difficult for me to want to adopt this strategy for my patients.

  • Isiah Harris

    While I agree with many of the positive comments made about this article, I would argue that in the era of vaginal progesterone supplementation that is at no cost to most patients after a positive pregnancy test, it seems premature to discontinue P supplementation before the physiologic production of P by the placenta at 7-9wks. It seems intellectually dishonest to state that it is needed for the first 3 weeks and not to continue it until the corpus lutei are no longer needed. There is not to my knowledge any identifiable harm in continuing therapy, and even if the number needed to treat to prevent one miscarriage was several hundred, it would be worth it to me and to that one patient.

    • Juan Garcia-Velasco, MD

      Thanks Isiah for your comments – while I fully understand your empiric approach (low cost medication and very low risk of harm), what it is true is that patients do not want to use any medication at early stages of pregnancy unless it is needed (there are a lot of unneeded medications being given without any benefit for the pregnancy). And on top of that, the trophoblast is inducing hCG secretion that is enough to support the pregnancy, just like in any spontaneous unsupplemented pregnancy. Supplementation is needed until trophoblast is capable to support the pregnancy as luteal phase after IVF is clearly defective.

  • laurenwroth

    It is our routine practice to stop progesterone supplementation at about 6 1/2 weeks (when fetal cardiac activity is confirmed on ultrasound) after a fresh embryo transfer. This article offers further support to that practice. Anectodotally, we do not see an increase in bleeding episodes after progesterone cessation.

    • Juan Garcia-Velasco, MD

      Thanks for your comments! In the Blockeel paper they did not see an increase in bleeding episodes either, but the nordic study did. We did see more bleeding episodes (minor bleeding) but miscarriage rate did not increase, which was very reassuring.

      • Juan– it seems to me there was an very LOW early loss rate between 5-9 weeks in your study can you comment on that as it might reflect on the population studies and #2— this was a non inferiority design so what would be the maximal difference that might exist in miscarriage rates and be missed with your sample size


        • Juan Garcia-Velasco, MD

          Thanks for your comments, Steve – 5 and 6 miscarriages respectively from 110 pregnant patienst per group is slightly lower than the average EPL rate, but we are including good prognosis women under 40, so this might explain it. We considered a 10% difference allowed to be missed, anything below would not have been detected.

  • Micah Hill

    This was a well designed RCT to address stopping luteal progesterone support at 5 versus 8 weeks. The study was powered to detect an absolute difference in miscarriage of 10% between the arms (5 versus 15%), which potentially could have been underpowered for clinically significant differences in miscarriage rates. But given that the actual miscarriage rates were only 1% different and the large increase in sample size that would have been needed, this seems less important.

    I am curious if the author’s institution has moved to a shortened luteal support? Or if any other readers would be willing to clinically make that change?

    • Juan Garcia-Velasco, MD

      Yes Micah, we have changed our practice and currently stop progesterone on week 5 in any IVF cycle that has been triggered with hCG; not in agonist triggered cycles, women over 40, endometriosis patients or women with early bleeding.

  • laurenwroth

    This is an interesting study asking an important question regarding necessary length of progesterone supplementation in fresh IVF/ICSI cycles. The study was excellent in that it had an appropriate sample size calculation and had biologic plausibility based on prior studies. It is comforting that no difference was found in any pregnancy outcome and certainly supports withdrawing P support early which may save patients money and may remove a potential teratogen. The only negative is that patients with early P withdraw had more bleeding episodes which may lead to more anxiety and patient visits.

Translate »