Progesterone to follicle index is better correlated with in vitro fertilization cycle outcome than blood progesterone level

Capsule:
The progesterone-to-follicle index has a linear and inverse relationship to cycle outcome, unlike blood progesterone, which is detrimental only at high levels.

Authors:
Yoel Shufaro, M.D., Ph.D., Onit Sapir, Ph.D., Galia Oron, M.D., Avi Ben Haroush, M.D., Roni Garor, M.Sc., Haim Pinkas, M.D., Tzippy Shochat, M.Sc., Benjamin Fisch, M.D., Ph.D.

Volume 103, Issue 3, Pages 669-674

Abstract:

Objective:
To investigate the impact of late follicular phase progesterone (P) elevation in relation to ovarian response on cycle outcome.

Design:
Cohort study. The progesterone-to-follicle index (PFI) was calculated by dividing the blood P by the number of follicles ≥14 mm. The clinical pregnancy rate was calculated against the range of PFI values and blood P levels.

Setting:
In vitro fertilization unit.

Patient(s):
A heterogenous population undergoing IVF with pituitary suppression and gonadotropin stimulation resulting in 3–15 follicles ≥14 mm and blood P≤10 nmol/L on hCG day and resulting in fresh embryo transfer.

Intervention(s):
None.

Main Outcome Measure(s):
Association of blood P and PFI with clinical pregnancy rate.

Result(s):
Data were retrieved for 8,649 IVF cycles in normal responders. The (reverse) odd ratios for pregnancy were 1.112 (95% confidence interval [CI], 1.077–1.165) for blood P and 4.104 (95% CI, 3.188–5.284) for the PFI. Elevated P levels were associated with a lower pregnancy rate only when they reached the >93rd percentile. The PFI was inversely and linearly related to the pregnancy rate for the whole range of values.

Conclusion(s):
A late increase in P level is detrimental if it is a consequence of increased P production per follicle (high PFI) but not if it is a consequence of additional follicular recruitment. The PFI enables clinicians to differentiate these conditions.

  • Pradeep Bhosale

    Thank you for the article,

    By monitoring the progesterone level and hence PFI index can we decide the optimum trigger time.

  • Micah Hill

    Thank you for the reply. As I understand it then, your hypothesis is
    that progesterone elevation is not detrimental in patients with many follicles
    but is detrimental in patients with fewer follicles. This is the hypothesis put forward recently
    from Griesginer et al., who
    demonstrated a negative effect of P4 <1.5 in normal responders, but not in
    high responders (http://dx.doi.org/10.1016/j.fertnstert.2013.08.045). This hypothesis was subsequently challenged
    by Dr. Bosch in a letter to the editor and an analysis of 4,000 cycles from his
    other paper (http://dx.doi.org/10.1016/j.fertnstert.2013.10.036). I think this concept that good response might
    protect from elevated progesterone is very interesting and controversial. While your paper looks at is from the
    progesterone : follicle ratio, they instead stratified patients by oocyte yield
    and looked at progesterone effect. But these are essentially looking to answer
    the same question.

    It is expected to me that the P:F index would be a
    more linear effect on pregnancy than serum P.
    Live birth rises almost linearly with oocyte yield up to 15 oocytes, the
    exact range in which your data was restricted (doi: 10.1093/humrep/der106. Epub
    2011 May 10.)
    Conversely, the large studies from Bosch et al. and Xu et al. show
    a sharp, but non-linear drop in pregnancy with rising progesterone. So the P:F ratio is linking two factors known
    to effect pregnancy in the opposite direction.
    In otherwords, someone with a high P:F ratio either has a very high
    progesterone (known to be bad) or a very low follicle count (known to be bad)
    or some combination of those factors.
    Conversely, a patient with a low P:F ratio eithers has a very low
    progesterone (good) or a high follicle yield (good). My question is what is the advantage of
    making these things a ratio of 2 variables instead of just using them each as
    independent clinic factors in our decision making?

    My final thought is I don’t believe this data
    demonstrates that patients with a high progesterone from many follicles do not
    suffer a detriment to their pregnancy.
    It simply demonstrates that they do better than a patient with the same
    high progesterone from few follicles.
    But that to me means that a good responder will have a higher pregnancy
    rate than a bad responder at a given P level.
    But I believe that good responder is still worse off than she would have
    been if her progesterone was not elevated. We have a paper in press that
    demonstrates this by statistically testing the interaction of progesterone and
    oocyte yield on live birth. I would be interested in comments when that gets published.

    In summary, I
    think the P:F is primarily a reflection of the benefit of being a good
    responder. But it is my belief that good responders still have a decline in
    pregnancy with progesterone that is significantly elevated. Appreciate your thoughts and or disagreement
    on these points.

  • Micah Hill

    Thank you for the interesting article. I have a couple of questions.

    1. In my mind, the serum progesterone value is a marker of something very different
    from progesterone : follicle ratio. The absolute serum progesterone value is a risk factor for endometrial advancement leading to asynchrony of the endometrium and embryo. The ratio on the other hand seems more likely to be a marker of follicular dynamics. You report an association with the ratio and oocytes aspirated. Did you find associations with and oocyte quality, oocyte maturity, fertilization, or embryo quality?

    2. You have over 8000 cycles you report on. Were patients excluded when they had multiple cycles? If not, did you control for repeated cycles in your analysis?

    3. Why did you choose just to index the ratio to follicles 14mm and over? Im
    curious if follicles 10-14 mm are also contributing to progesterone production?

    4. If you could, can you further expand on the concluding paragraph for me. It seems you are making the case that increasing progesterone from a follicle is a negative marker of oocyte quality. Biologically, why do you think this would be true? Similarly, the prevailing hypothesis about the negative impact of progesterone is related to
    endometrial advancement. Since the endometrium can be advanced over a certain absolute threshold of progesterone, why would it matter whether that threshold amount of progesterone came from 5 follicles or 20 follicles? Either way the endometrium is being exposed to that adverse amount.

    Thank you for your thoughts on these issues.

    • Yoel Shufaro

      Dear Dr. Hill,
      Thank you for your interest and comments. My answers to your questions are;
      1. There was no significant association between oocyte maturity, fertilization rate or embryo morphological grading and the PFI. I presume this is because of the limited usefulness of these parameters in predicting IVF outcome. In the multivariate analysis there was only a modest effect of the number of aspirated oocyte and transferred embryos on the odds ratio.
      2. Repeated cycles were not excluded because each cycle in this study stands on its own and has different unique parameters.
      3. The selection of this size range was for practical reasons; we were not sure that smaller follicles were indeed counted accurately. For the same reason we did not include cycles in which more than 15 follicles were counted. We believe that the accuracy of routine folliculometries (not prospectively performed for reasearch) is limited for counting small or very many follicles.
      4. This is exactly the novelty in this paper. The endometrium does respond to an absolute progesterone threshold, but the curve of progesterone level effect on pregnancy rate is bimodal (figure 1.). The pregnancy rate is not effected up to a certain progesterone level, than there is one step of decline and negative plateau. The effect of the progesterone level on the pregnancy rate is very limited.
      This is totally different for the PFI. The lower the average follicular progesterone secretion is, the higher the chance to conceive, even within the safe non-detrimental progesterone range. In the case that the progesterone level is elevated, the chance to conceive is substantially higher if it is secreted by many follicles because in this case the oocyte quality is better (though we cannot see with our our eyes). The effect on the endometrium has only a modest impact on the pregnancy rate.
      Thank you again for sharing your thoughts. I hope I answered your questions.
      Sincerely,
      Yoel

      • Micah Hill

        Thank you for the reply. I have some followup thoughts on this very interesting topic. As I understand it then, your hypothesis is
        that progesterone elevation is not detrimental in patients with many follicles but is detrimental in patients with fewer follicles. This is the hypothesis put forward recently from Griesginer et al., who
        demonstrated a negative effect of P4 <1.5 in normal responders, but not in high responders (http://dx.doi.org/10.1016/j.fertnstert.2013.08.045). This hypothesis was subsequently challenged by Dr. Bosch in a letter to the editor and an analysis of 4,000 cycles from his other paper (http://dx.doi.org/10.1016/j.fertnstert.2013.10.036). I think this concept that good response might protect from elevated progesterone is very interesting and controversial. While your paper looks at is from the
        progesterone : follicle ratio, they instead stratified patients by oocyte yield and looked at progesterone effect. But these are essentially looking to answer the same question.

        It is expected to me that the P:F index would be a more linear effect on pregnancy than serum P. Live birth rises almost linearly with oocyte yield up to 15 oocytes, the exact range in which your data was restricted (doi: 10.1093/humrep/der106. Epub 2011 May 10.)
        Conversely, the large studies from Bosch et al. and Xu et al. show
        a sharp, but non-linear drop in pregnancy with rising progesterone. So the P:F ratio is linking two factors known to effect pregnancy in the opposite direction. In otherwords, someone with a high P:F ratio either has a very high progesterone (known to be bad) or a very low follicle count (known to be bad) or some combination of those factors.
        Conversely, a patient with a low P:F ratio eithers has a very low
        progesterone (good) or a high follicle yield (good). My question is what is the advantage of making these things a ratio of 2 variables instead of just using them each as independent clinic factors in our decision making?

        My final thought is I don’t believe this data demonstrates that patients with a high progesterone from many follicles do not suffer a detriment to their pregnancy. It does demonstrates that they do better than a patient with the same high progesterone from few follicles. In otherwords, a high responder has a better prognosis than a poor responder for a given progesterone value. But I think that a good responder is still worse off than she would have been if her progesterone was not elevated. We have a paper in press that demonstrates this by statistically testing the interaction of progesterone and oocyte yield on live birth. I would be interested in comments when that gets published.
        In summary, I think the P:F is primarily a reflection of the benefit of being a good responder. But it is my belief that good responders still have a decline in pregnancy with progesterone that is signiifcantly elevated.

        • Yoel Shufaro

          Dear Dr. Hill,
          Than you for these enlighting comments. I certainly agree with your explanation of our observations. I will be more than interested in your paper offcourse.
          My answer to your question is that the PFI is most useful when a premature progesterone elvation occurs, and one has to decide upon its clinical significance to the outcome of the cycle. Here you need to use the second variable to determine prognosis. In the case that only a few follicles were recruited, such an elevation is bad news. If many follicles participate in progesterone production and secretion – the cycle’s prognosis is unaffected.
          We wanted to establish a prospective (i.e. pre- OPU) prognostic tool applicable to such cases, therefor we chose to divide the total blood progesterone in the number of follicles, and not in the number of oocyte eventually aspirated.
          Sincerely,
          Yoel

          • Micah Hill

            I really appreciate all the time you spent on this publication and answering my questions in detail!

            Thank you,
            Micah

Translate »