My guest today is Dr. Eli Y. Adashi, Immediate Past Dean of Medicine and Biological Sciences at Brown University. Dr. Adashi's accomplishments and productivity in reproductive medicine are vast in scope and number. His leadership in obstetrics and gynecology and public health policy has been critical to the development of our specialty. Among many other responsibilities and recognitions he is a member of the Institute of Medicine and an advisor to the WHO and the Gates Foundation. It is an honor to have him as a guest today.
In this interview Dr. Adashi gives his perspectives on the paper by Diamond and associates entitled, "Endometrial shedding effect on conception and live birth in women with polycystic ovary syndrome," published in the May Green Journal. The authors report that live birth and conception rates were higher in clomiphene cycles if not preceded by spontaneous or induced bleeding.
blO+G: Dr. Adashi, I'm going to ask you to go back in time with me. Back to 1980. You just finished your fellowship in Reproductive Endocrinology at Hopkins. Speroff was still a skinny book at that time. Take some time to tell our listeners how clomiphene was administered back then. These were the days when there were no such things as transvaginal sonography or rapid, sensitive serum hCG, estradiol, or progesterone assays. How was Clomid started? How was the response to stimulation monitored? When was the dose increased.
EYA: In the absence of all the technology that you just mentioned, in essence we were still relying on basal body temperature charts. We may have measured the occasional well-timed circulating estradiol. We may have measured a well-timed circulating progesterone level. And, of course, we would go by the patient's history as to trying to ascertain whether or not we were in any way successful. This was clearly a very different era in terms of our ability to monitor. There was no capability at the time to meaningfully--in real time--measure, for example, circulating LH levels. We didn't have sonography, as you mentioned. And so, in many ways, it was more of a clinical art complemented by a modest battery of hormonal tests that facilitated the process, but clearly were a far cry from where we are today.
blO+G: There was a saying back then: no period, no Clomid. Why was that?
EYA: Mostly I would say because of concerns that pregnancy has not been ruled out and that you would want to be sure that you're not dealing with a conception upon which you inadvertently superimpose another course of clomiphene. I would say that another reason for that--assuming that you ruled out pregnancy with a pregnancy test--was that there was a conviction for a long time that somehow clomiphene cycles better be started against the backdrop of some form of menstrual flow: natural or induced. I would think that those two factors combined probably conspired for somebody to have coined the term (and I'm not sure who did), "no [period], no Clomid."
blO+G: Then, in 1996, you, John Rock, and Zev Rosenwaks published an enormous book, Reproductive Endocrinology, Surgery, and Technology. You wrote the section on clomiphene induction of ovulation. You described something called the Triple 7 Rule. [Editor's note: the Triple 7 Rule consists of a serum estradiol level 7 days after the last clomiphene tablet, a serum progesterone level 7 days later, and a pregnancy test after 7 more days.] Would you explain that to us?
EYA: That was a practical, or at least aspiring to be practical, regimen that we felt had an easy ring to it, was memorable, and could potentially well-serve individuals who engage in the induction of ovulation with clomiphene citrate but may not be reproductive endocrinologists, maybe not as familiar with the fine points of the hormonal swings that may be involved in the context of clomiphene-induced ovulation. So I suspect it was a mnemonic of some sort that we felt was useful for teaching students, teaching fellows, and perhaps of some utility to practitioners who engage in the use of clomid but perhaps not all that frequently.
blO+G: In that same section you stated that initiation of clomiphene did not require spontaneous or induced bleeding. So I ask you, if it were the patient's first cycle, how would the doc know if the patient spontaneously ovulated within the prior two weeks?
EYA: I think that's a valid criticism. Clearly, the main thing to rule out would have been conception, as you may have mentioned earlier. And I would go further and say that even though we made that statement we couldn't have supported it with evidence probably one way or the other. It probably was our practice at the time, mostly for practical reasons. I'm interested to hear that we did in fact recommend initiation in the absence of endometrial shedding of some kind. I'd like to say that we were perceptive and could foresee the future, but I don't think those decisions back then were backed up by a large body of evidence. In this case, I would say no evidence, just practical experience and a modicum of pragmatism.
blO+G: It's not time to turn to the Diamond paper which is published in the May issue of the Green Journal. This was a secondary analysis of the PPCOS I study. In it, the authors report that live birth and conception rates were higher in clomiphene cycles if not been preceded by spontaneous or induced bleeding. What does this study contribute to our understanding of ovulation initiation with clomiphene for women with polycystic ovary syndrome?
EYA: It's an interesting and significant observation. As I mentioned earlier, the decision to bring about endometrial shedding or not was an uninformed decision up until this paper came out, which is why I was saying earlier that whatever statement we may have made in the textbook regarding this issue almost certainly was not based on firm empirical evidence. This, to my knowledge, is the very first time that somebody actually took upon himself to look into this treatment decision (it's really, in a sense, a fork in the treatment road) and subjected it to rigorous analysis, even if, in this case, a secondary or post-hoc analysis. This is the way to do it. No clinical measure that we employ is too small or too insignificant to be left to a situation wherein evidence is lacking. Since 1967 when clomiphene was first approved by the FDA to this day, which is 45 years, we didn't really have any intelligent response to the question: does a clomiphene cycle need to be preceded by some form of endometrial shedding and if so, what would be the impact of such a maneuver on the subsequent outcome? Back to what you said earlier about "no period, no Clomid," that was based on some concepts, perhaps, but certainly not on evidence. This study would argue that in retrospect that statement was flawed and inaccurate, especially if this study is confirmed or repeated especially if it is carried out in a prospective fashion by this or other investigators. I do think that this is precisely what we would like to see: evidence-based medicine. I would be the first to admit that for the longest of time, mostly out of necessity and sometimes out of pragmatism, we were practicing in a manner which was somewhat detached from evidence-based medicine. That's not a criticism of the reproductive endocrine community--we just didn't have the information. And so I applaud the authors for tackling what to some may seem like a small element of the decision-making process but one that in retrospect, as the paper indicates, has some real life implication. Yes, I am impressed and I applaud the authors for having taken this on. It's a wise use of the resources already invested in a prior study. All that was required was a post-hoc secondary analysis of existing information. The limitations of the study notwithstanding, I think it's a step forward that we ought to embrace.
blO+G: The next time you're in the office and you have a patient sitting across from you who is about to start clomiphene, she reports that last week she had the onset of spontaneous vaginal bleeding that was self-limited, would you advise her, in light of this study, to wait a month or two before starting her medication?
EYA: I think that would be a reasonable recommendation that would have to be made in the context of that individual. A month or two delay probably is not of major consequence in most circumstances. We would still be making an empiric decision yet again in this case because it's not as if this study specifically addressed the question just like the one you posed. But extrapolating from the observations of Diamond and associates, I don't think that would be an unreasonable posture to assume. On the other hand, it's not as if starting clomiphene after some form of endometrial shedding is futile. It's just that in this study it is, in relative terms, less effective.
blO+G: How do these results square with previously published work which showed an advantage to adjunctive oral contraceptive pills prior to initiating clomiphene? Are they contradictory results or can they be reconciled as derived from different contexts?
EYA: I don't know how well those studies were conducted or how well-controlled they were. I never fully understood the rationale, although the use of combination oral contraceptives prior to initiation of clomiphene citrate depends a great deal on when you superimpose the clomiphene. I assume you are referring to a situation whereby a cycle or two or three of combination oral contraceptives is provided and the patient is allowed to withdraw and then clomiphene is initiated. To me, that's not all that different from using an isolated progestin in the sense that combination oral contraceptives are predominantly progestational in nature. It's different in the sense that, of course, the exposure is longer. It's not a five-day, seven-day, or ten-day course of an isolated progestin. It's perhaps a 28-day exposure to a combination of largely progestin supplemented by an estrogen, so the situations are perhaps not entirely comparable. But, in the final analysis, it's about bringing about withdrawal, providing a hypothetical (and I emphasize hypothetical) starting point to a cycle even though, when dealing with ovulation induction we're not really talking about a conventional cycle. This might be a good point to reflect back on the notion of "no period, no Clomid" and add a third potential reason for why this was popular. That had to do with the sense that a cycle had to begin with a bleed and that there needs to be this reference point--day number 1 of the bleed, for example--to call it a cycle. These were the concepts of the day. It's hard to fault them. They were transmitted from one generation to the other, but faced with new evidence such as provided by Diamond et al., it behooves us to reevaluate.
blO+G: I know it's too early to discuss mechanisms to explain the observations, but we're going to, anyway. Several have been proposed by the authors and by Dr. Casper in his editorial. Do any of these proposed mechanisms ring true for you?
EYA: I was not fully convinced by the potential mechanisms proposed either by Dr. Diamond or by Dr. Casper, although all of them undoubtedly were valid efforts to try to account for the observations. But by definition, they were not backed up by any direct experimental data. They were efforts to stitch together, as best one can, based on precedent, based on prior studies, based on insights, a potential mechanism to account for the observations. Fundamentally, the notion that we should seek the explanation in one of two major areas is sound. The first being some form of differential effect on the hypothalamic-pituitary-ovarian axis and the second major area being some form of salutary effect that the lack of withdrawal afforded these women at the level of the endometrium. But beyond that, absent specific experimental questions that are submitted to the scientific method and experimented appropriately--especially the molecular events suggested at the endometrial level--those would require experimental verification.
blO+G: So where do we go from here? Should we invest in follow-up studies on this particular question?
EYA: That's a very valid question. I reflected on it. What was in one way very attractive about this study was the fact that it did not require substantial additional resources in that those have already been invested. All that was accomplished here--I think cleverly so--is a post-hoc secondary analysis. Conducting a full-fledged study, for example, by the NICHD Cooperative Reproductive Medicine Network, would be a decision that would have to be made by their steering committee in light of other priorities that have been identified as in need of investigation.
blO+G: What's on the horizon for initiation of ovulation with clomiphene? What would you like to see investigated? Are you still interested in a monoisomeric preparation of clomiphene?
EYA: I would say modestly so. We never really had the opportunity to fully explore the potential of monoisomeric preparation. It is a general truism, I would like to think, that when we use pharmaceuticals we want to limit the administration to the active principal and, if possible, minimize what I would characterize as "impurities" or, at least, allegedly inactive principals. So yes, I would say that would still be a useful proposition. We have learned in the last 45 years that there is not enough of a commercial incentive for the pharmaceutical industry to embark on this effort on their own volition. You and I could speculate that if such studies were to be conducted and if such studies were to demonstrate clear advantages of a monoisomeric preparation that it's at least possible that pharmaceutical interest would follow suit with production of such a preparation for widespread use. Here again, though, I like to temper any of those statements by the recognition that we are in a different time. There are many more options for infertile patients: both medical options and assisted reproductive options. Clomiphene is really just one agent, whereas when you took me back to 1980 it was all we had at our disposal. So today, with metformin, with aromatase inhibitors, the urgency of studying the fine details of the clomiphene paradigms such as should there be a precedent bleed or should there not, should we be using monoisomeric preparations or not becomes somewhat of a lesser issue than perhaps it would have been in 1980. My guess is that if you had asked me this question back then, I would have answered without hesitation that this is information I could use because that was the central tool I had at my disposal. That's probably not quite as important today--not that the question is not important, it's just that in the scheme of priorities in the face of limited resources we probably have to pause before we answer these questions in a definite fashion.
blO+G: Dr. Adashi, do you have any final words for our listeners about this topic?
EYA: I do, and it goes back to a point I made earlier about pragmatism. It turns out (and that probably is behind the statement we made in our textbook) that it has been our practice for some time not to precede clomiphene ovulation induction with an induced endometrial shedding. The reasons that I recall dominated that practice were pragmatic. You know, of course, as do many of our listeners that if you precede the administration of clomiphene citrate with a progestin-induced bleed, there are fairly involved logistics in the process and a great deal of education that needs to be imparted to the patient. The first thing you say is, "We're going to have to go through a course of 5-7 days of progestin. A certain number of days after the last pill you will experience endometrial shedding--some sort of menstrual flow--and then, on a certain day in the context of that flow we would ask that you begin to take clomiphene citrate." And so you sit there in front of a patient and you draw the chart with the timeline and you try to explain what will transpire and why and all of that happened long before the meaty part of the process (the clomiphene citrate part of it) actually is coming to light. So we found it to be simply challenging for the patients. It meant two prescriptions, fairly lengthy explanations, and often misunderstandings of a variety of natures. At some point we concluded that we could probably manage without it. Of course, I would like to argue today that we had the foresight to anticipate the findings by Diamond et al., but of course that was not the case. I think pragmatism at the time dictated our practice. We may have served our patients well in retrospect, but we will never know.
blO+G: Dr. Adashi, thank you for your perspectives.