Fertility and Sterility On Air - TOC: Feb 2025

Take a sneak peek at this month's Fertility & Sterility! Articles discussed this month are:  

01:03 (Not) My body, my choice? - Should physicians be facilitating gestational carrier arrangements in the absence of medical indication?

05:22 Morphological changes of endometriomas during pregnancy and after delivery detected using ultrasound

12:17 Hormone-free vs. follicle-stimulating hormone–primed infertility treatment of women with polycystic ovary syndrome using biphasic in vitro maturation: a randomized controlled trial

24:37 Association of in vitro fertilization with severe maternal morbidity in low-risk patients without comorbidities

33:59 The association between primary ovarian insufficiency and increased multimorbidity in a large prospective cohort (Canadian Longitudinal Study on Aging)

41:28 Target trial emulation of preconception serum vitamin D status on fertility outcomes: a couples-based approach

52:59 Predicting a successful match among applicants to reproductive endocrinology and infertility fellowship

View Fertility and Sterility February 2025, Volume 123, Issue 2:
https://www.fertstert.org/issue/S0015-0282(24)X0015-1

View Fertility and Sterility at https://www.fertstert.org/

Welcome to Fertility and Sterility On Air, the podcast where you can stay current on the latest global research in the field of reproductive medicine. This podcast brings you an overview of this month's journal, in-depth discussion with authors and other special features. F&S On Air is brought to you by Fertility and Sterility family of journals in conjunction with the American Society for Reproductive Medicine, and is hosted by Dr. Kurt Barnhart, Editor-in-Chief, Dr. Eve Feinberg, Editorial Editor, Dr. Micah Hill, Media Editor, and Dr. Pietro Bortoletto, Interactive Associate-in-Chief.

Welcome back to another episode of Fertility and Sterility On Air. We are in February 2025, Volume 123, Number 2. I'm Micah Hill, the Media Editor for F&S, and we're joined today by Editorial Editor Eve Feinberg, and Editor-in-Chief Kurt Barnhart. Good morning, both.

Good morning, Micah. I'm glad to be here on the new year. Good morning to both of you.

Same. Eve, as always, you like to pull up controversial issues when you write stuff as an editorial editor. And this month, you're talking about, should we be using gestational carriers when there's not a medical indication for it? And your introduction says, not my body, my choice, question mark.

Tell me about this. Why did you think this was a good topic? And what are your thoughts? Yeah, well, I think that the lines between what's medically indicated for surrogacy and what's, let's just say, societally or socially indicated, have become really blurry. And in our practice, we have a weekly meeting where we bring forth any cases that are controversial, so to speak.

And I think that really got me thinking about, should there be society guidelines on what's medical? Should that be left to the individual physician? I've had cases where I have felt that this is very appropriate to use a surrogate, and others have disagreed. And so it really got me thinking about, should we have guidelines at all? Or should surrogacy just be available to anyone who feels that they cannot or should not be able to carry a pregnancy? And so I engaged Kate Schoyer and Brian Levine, and Kate had also, Kate and I had also had an offline conversation about this as a controversial topic, and it really sparked, I think it was Kate's email to me that really sparked the, aha, I'm not the only one who's struggling with this. Let's make this into a battle.

And so it was, I encourage everyone to read it. And I think to this day, I'm still not solidified in my opposition to it or my support for surrogacy for less than stellar medical indications. And I'm hesitant to call it social because I think that that trivializes it, and it really trivializes the very real trauma that pregnancy can bring to a person's body.

But I also, as I think our listeners know, I'm a fierce advocate for reproductive choice. But I think the really big question is, does choice for one person, is that all-encompassing or can that infringe on somebody else's bodily autonomy? And that's really where I struggle with this. It's a great debate because there's really good ethical arguments on both sides.

Just out of curiosity, do your practices allow gestational carriers if there's not a medical indication? So it's Illinois law that there has to be a medical indication. However, I will say it's up to the individual physician to determine whether or not something meets that criteria. And it's murky.

We've argued about this a lot also ethically, and we've had ethicists come and talk to us. And it really does drive emotions on either side. Personally, I don't think reproductive care should be commercialized and be a product.

I don't think that you just can come in and buy something. I can buy a carrier or I can buy a surrogate. So I really do think it's medicine.

There should be a medical reason. But ethically, we had some really strong discussions about it. And the only ethical tenet that came up with is that it's a scarce resource, and therefore it shouldn't be just for purchase.

There are a lot of women that need it for medical reasons, and no surrogate goes for those. It shouldn't go to the highest bidder, which tends to be the one that can afford it, those kind of issues. So it's a tough debate.

I'll let you know how much it's evolved. When I was looking for my first fellowship job, I was asked that very question, that the President's United States wife doesn't want to have a baby. She wants to come to you, and you provide all the services, and she'll bring a carrier.

Would you do it? And I think if I had answered yes, they would not have hired me. So that's how far it's come to now. If you have the cash, it really bothers me that we will do anything that you ask us to.

And that just gets under my skin a little bit. It's a great debate. I encourage everyone to read it.

Maybe someday we'll have artificial wombs, and science will offer us a new way to have this same debate from a different perspective. There's also a great inkling from editorial editor Laura Rienzi on the embryo ontogeny of parental genetics. I highly encourage you to read this.

There's a lot packed into just this two-page inkling from Laura. Kurt, we're going to jump right into the science now. You have the seminal contribution this month on morphologic changes of endometriomas during pregnancy.

What did we learn from this article? Yeah, I enjoyed this article because it's a little bit different from the daily practice of medicine, or REI, I should say here. And it's talking about kind of the natural history of endometriomas during pregnancy. So admittedly, I understand that we as REIs don't often follow our patients through pregnancy.

But we are consultants, and we should be given the best information, and we should learn about it. This study out of Sweden tackles the morphologic changes of endometriomas during pregnancy and after delivery detected in an ultrasound unit. Lund University and Malmo, Sweden, authors that did a very nice job here.

I have the impression these are not REIs, which I'm also happy that they published in Fertility and Sterility, because the article has a slightly different take in that they start with the idea that when we're scanning people during pregnancy, we come upon masses, and we have sometimes some difficult diagnostic dilemmas of what to do with these adnexal masses. They acknowledge that ovarian cancer is certainly a very rare disease, but it has real consequences in pregnancy. So the idea is that they could follow, and I guess it's very basic, they're just describing it, but they're following a relatively large number of women who were diagnosed in their first trimester ultrasound scan, and following them to see how those endometriomas changed.

We're talking about a sample size of around 57, which sounds low, but it really is the biggest series that we've ever seen. And it talks a little bit about the changes, both in terms of size, in terms of how you would characterize the cyst, and then the solid components, which gets everybody all angry and upset about when you see a solid component in a mass during pregnancy. And it does a nice job characterizing this.

Let me go through a little bit of it. So again, most of these people were found the mass during pregnancy. There were not patients that they knew had endometriosis and they followed, although they did make a comparison to about a third of them, which were known to have endometriosis before they went in.

And I'll just say that off the bat, that didn't seem to matter. And I'll also say off the bat, there really were only two that they were very concerned about borderline ovarian cancer and wanted to operate on, and these authors talked them nuts out of that, and they ended up to be endometriomas. So what happens is about three quarters of the endometriomas change in terms of the ultrasound characteristics during pregnancy.

The most important changes I'll list through is that about two thirds of them decidualize. Now decidualization, we understand, but again, as a radiologist or an obstetrician, you don't sometimes recognize that the decidualization of an endometrioma can look a lot more worrisome than just the typical ground glass or simple cyst characterizations of an endometrioma. It happens about the maximum around 12 weeks, and then it dissipates in some over the course of the time.

Other findings they found was that they decrease in size, so about three quarters of them decrease in size, mostly decrease in size after around 22 weeks, 81% of them were smaller after delivery than before they got pregnant or at least in their first trimester. And interestingly, in around a quarter of them, the endometrioma resolved. They didn't see it anymore, which I found interesting, and which I found even more interesting was at about half of those that resolved, they found it again in their non-pregnancy.

So it can make it a little bit more difficult. I've always told people that pregnancy can make your endometriosis better, but it doesn't make it go away. And then finally, the solid component was interesting, and they noticed that the solid component also increases in size over the course of the study, and that's what the radiologists are most worried about, and that seems to say the same after pregnancy.

So in summary, you could tell a patient your endometrioma is not going to stay stable, it's probably going to get smaller, it's probably going to decidualize, and that the solid component may actually get larger. Now, there's a neat discussion you guys might want to read about pure speculation of why that would happen. I think we understand the decidualization, but they're talking about the ratios of estrogen and progesterone as why the solid component might get larger, but I don't think we could call that the truth at the moment.

I think it's speculation. I thought just that information alone is something that we should learn and we should know about. Occasionally, I still get called by somebody that got pregnant in our practice that the radiologist wants to do X, Y, and Z, or even in de novo, somebody, you know, we are consultants for our OBGYN journalists, colleagues, so this is something we should know about and understand.

Did it surprise you guys? I just kind of like the way it characterized it. I'm not sure it really changed my perception, but it was a nice way of characterizing it. I think we think about decidualization of the endometrium, and to be honest, while, yes, we know all about decidualization, to be fair, this was the first time I'd actually heard about decidualization of an endometrioma and the characterization of that.

My only comment was I really wish they would have had a photo in this paper, but that part surprised me. And they talk about the decidualization of the endometrioma as if it's like a very common, well-known fact and finding, and this was the first time I'd really heard about it or had it well-described. And while it makes intellectual sense, it was a new concept to me.

I don't know if that's, Micah, I'm curious on your thoughts, too. Exactly the same. I had not thought about decidualization in an endometrioma.

I didn't really realize this was a thing. It all makes sense when you read it, and then just looking through, there's case reports of people being operated for suspected malignancy and pregnancy for this exact thing, and they turned out to be decidualized endometriomas. So having this relatively larger prospective case series is actually very helpful, I think, for these.

And I will take responsibility for this, generally, for the journal. There is a lovely picture, Supplemental Figure 1, that I know is hard to find in the journal, and I have to figure out how to make supplemented material more accessible to people. That's what I mean.

But if you look at Supplemental Figure 1, wow, that picture to me looks like borderline ovarian cancer. I mean, the decidualization is no longer ground glass, and the solid components are irregular and large compared to everything. I would be worried looking at that picture.

So at the very least, you should familiarize yourself by looking at this and say, yeah, I can see why they're worried. And I'm glad that this study at least calms us down a little bit. Fantastic.

That is great, Kurt. And I think the point of having non-REI submit some of these things that are more GYN or pregnancy-related that fit within our journal is a great example of that we're not just an ART journal. We're a reproductive medicine journal.

Right. And it goes again to my Venn diagram that this is not the most methodologic, sophisticated paper. It's just a well-characterized description, but of a very, very novel topic that teaches us something.

And that's exactly what should be in our literature. Eve, we're jumping now to assisted reproduction, and there's an RCT about in vitro maturation from this group in Vietnam who's just been doing a phenomenal amount of work with their clinical trials. So tell us what we learned from this trial.

Yeah, this also was very eye-opening to me as someone who does not do IVM on a regular basis. The title of this paper is Hormone-Free FSH-Primed Infertility Treatment of Women with Polycystic Ovarian Syndrome Using Biphasic In Vitro Maturation, a Randomized Control Trial. And this was by Lan Vuong and others from Vietnam.

So IVM is an alternative approach to controlled ovarian hyperstimulation. So for our learners, IVM protocols can vary a lot from center to center, and some centers use absolutely no gonadotropins prior to egg retrieval, and some centers use gonadotropins, and then these regimens of how they use gonadotropins can vary. IVM is not widely used in the U.S., but it is widely used in Japan, China, and Southeast Asia.

And previously, it was thought that gonadotropins were necessary, but this article really challenges that. So if you forego gonadotropin stimulation, you're eliminating the risk of OHSS, you're eliminating the discomfort associated with enlarged ovaries during stimulation, theoretically reducing your risk of things like ovarian torsion, and you're also dramatically reducing the cost of an IVF cycle. And so the purpose of this study was to compare two separate regimens of IVM using two days of FSH prior to egg retrieval, and then comparing that to a regimen where no FSH prior to egg retrieval was used, and they used a slightly different IVM system called a capacitation IVM biphasic system, and I'm going to describe that a little bit more.

So it was a randomized controlled assessor-blinded trial, and the objective was to compare the number of mature oocytes and the pregnancy outcomes in women ages 18 to 37 with PCOS, with and without FSH priming, and they did a one-to-one randomization. Interestingly, women were put on OCPs for seven days, and I suspect that was really just a timing issue. So women in the non-FSH group were scheduled for oocyte retrieval on 9 a.m. five days after the last birth control pill, and women in the FSH group received two days of 150 units of FSH, and then their egg retrieval was at 9 a.m. 42 hours after the last FSH injection.

No HCG trigger was administered in either group. Egg retrieval was performed, and then for the first 24 hours after egg retrieval, oocytes were placed in a supplemental media that maintained the meiotic arrest. This media has culture additives that are thought to enhance developmental competence, and that's the capacitation part.

After 24 hours, the cumulus oocyte complexes were washed and transferred to an IVM media and incubated under oil for 30 hours. ICSI was then performed for fertilization, and then embryos were cultured to day five and vitrified at the blastocyst stage. In the next cycle, these patients underwent FET.

They recruited 120 participants with 60 in each group. Participants were young, lean, and had high AMH levels on the whole. The number of mature eggs did not differ between the FSH and the non-FSH groups, and I think these numbers are actually really impressive.

There were 14 versus 13 M2 oocytes. The mean number of blasts was 2.6 and 2.8, and the live birth rate was 31.7 in the FSH group compared to 38.3 in the non-FSH group. So overall, not a statistically significant difference in either category of mature eggs or live birth rate.

So I thought this was incredibly interesting. I think it's ripe for discussion. Are the data convincing? Should we be doing more IVM in the U.S., and are we, how we do things here, are we thinking about things in the right way, or are we thinking about things in the wrong way? I'm fascinated by the number of cycles that are being done in Asia and the lack of doing these cycles in the U.S. So Kurt, Micah, I'm curious on your thoughts on this.

I thought it was a really well-done trial, and again, I learned a tremendous amount, not just about IVM but the different protocols and the idea of capacitation as well as FSH, no FSH, no HCG trigger, just retrieve the eggs and move forward. This is kind of a new frontier, if you will, or it's not a new frontier, it's a frontier we just haven't explored. We got so good at doing IVF the way we do it that we've just not been bothered by not understanding all the physiologic mechanisms that happened before folliculogenesis.

So I'm glad it's being done, because we really do need to understand these early sequences and this early biology. Time will tell how it fits into the clinical armamentarium and when it's used and when it's not, but I still think science like this is really important, and I applaud the authors for this well-done trial. Just two observations on it.

Their main outcome is M2, so they're looking at genetic competency after IVM, but there's also the cytoplasmic competency that happens within the oocyte. And so what you see in the clinical trials comparing IVM to regular IVF is that even if you get M2s, the development to BLAST seems to be impaired with IVM. And you see that here.

These patients on average were 29 years old and their AMH was 10. So these are patients you would expect a lot of BLAST. The mean number of good quality BLAST was one per patient.

So even though they're getting some M2s, it's fewer M2s than you would expect, and then the development after that to BLAST is very much lower than what you would expect from traditional IVF. So that sort of goes to Kurt's point about the, you know, clinically, who do we use these for? I guess if you want one baby, then maybe this is a reasonable route. If you want more than that, you're probably not going to get that from IVM, and that's what the clinical trials show.

You disagree? Yeah, but if you're, you could go through multiple retrieval cycles, right? So you're not incurring the cost of medication or the discomfort of stimulation. I could envision you go through two, three egg retrieval cycles for banking. The other question that I've always had about IVM that this does not answer is, as we are messing, so to speak, with meiosis by holding these oocytes in meiosis arrest, are we potentially altering chromosomal abnormality in the separation and segregation of the spindle apparatus? And are we seeing more aneuploidy in these blastocysts, which may account for the lower live birth rate? But I think that there's a role for it that we're perhaps not tapping into, and I don't think it can be like a one and done in the same way that you can have a 29-year-old with an AMH give you 8, 10, 12 blasts, and that person never has to go through another retrieval.

But is there a role potentially for IVM that we're not using? I don't think it's perfect by any stretch, but I think this article, to me, was very eye-opening to say like, hmm, should we or could we be doing things differently in some patients? Perhaps patients that can't tolerate high estrogen levels, breast cancer patients, where we might be exposing them to additional risk controversially. I know that's a loaded thought, but I don't know. I thought it was really interesting.

I'm glad it was published. And... I like the way you're pontificating where it can go, because I think my real point was that we need to learn this. And of course, you often learn the first clinical trial and the first biology and good prognosis patients.

But the more work that's done in this area, the more we're likely going to change the way we practice. Maybe not today. It might be 10 years or even longer.

But I mean, I love the idea that we're expanding the knowledge and expanding our understanding. Yeah. I mean, I have a patient right now who has severe antiphospholipid antibody syndrome.

She's had a clot in her cerebellum. She's had a thrombus in her heart. She's had multiple blood clots.

And she wants to use a gestational carrier for future pregnancy. But I'm very afraid of stimulating her given the high estrogen levels. Now, I'm not going to experiment on her in our center without adequate training and knowledge and expertise.

But it does start the wheels of my brain churning on thinking, like, I don't think that this is going to replace IVF in the way that we are doing IVF in our center in the US. But could we develop use of this in a selected patient population where traditional IVF may impose unacceptable risk? Just want to give a follow-up to your comment on two or three cycles of this to be equivalent to one cycle of IVF. That's exactly what these authors argue in their human reproduction randomized clinical trial where IVM was found to be both inferior and not non-inferior to regular IVF.

So they did worse. And they argue that a true trial should do three cycles of IVM and compare it to one cycle of IVF. I would argue that that's not patient-friendly.

You're going through three surgeries. You have three costs of incubation in the lab. It's not cost-effective.

And you're actually increasing costs to get to the same outcome to avoid 10 days of very low-dose gonadotropins that these patients would need. So I don't know that PCOS patients is the population. I think we're all probably agreeing with that.

We can avoid OHSS with a Lupron trigger, so you don't need to do this. So maybe it's some other cancer patients, other situations where we just don't have time. I know, Mike.

Again, I remain optimistic. Remember how inefficient IVF was when we first started? Oh, absolutely. Frequency rates of 10%.

And we were claiming victory with a success rate of 22% at one point. I mean, so again, we're all saying the same thing. I'm excited that we're exploring new technologies, and I'm glad we learned from it.

Yeah, I agree with that, Kurt. I just don't think the technology is to that point yet, but that's exactly why we need research like this to get the technology to that point. At some point, I imagine this will be there, which is not there yet clinically.

My only other point is more a statistical one. So they powered this as a non-inferiority trial to find that negative margin being less than three mature eggs, and that bottom margin was minus five. And so they say it's the same outcome, but actually they were not able to demonstrate non-inferiority using this other method.

So I think the interpretation was probably not accurate. It should have been that this is not non-inferior, so we can't rule out that it might be inferior, this new method. It's just that it's a nuance to how you interpret a non-inferiority margin, but I don't think it was quite accurate the way they interpreted it.

That's my opinion. Yeah, and I think that's a fair point. And I think even with that, to say we have an alternative treatment option that's available that might be a good fit in a certain population, it's a little bit worse.

But when you balance the risks and the benefits, again, in that patient, not so much that you don't want to expose them to the risk of OHSS, because I agree with you wholeheartedly, we can eliminate OHSS, but the patient who doesn't have time because they have a mediastinal mass and they're starting chemotherapy in two days, go for it. Try it if you are well-versed and you know how to do it. Or like my patient where high levels of estrogen may be very detrimental, we're planning on using letrozole, CoStim, low-dose kinatotropins, close help with her hematologist, but it scares me and rightfully so, I think.

Final comment is these technologies may actually help IVF. That surprising patient where their leads are 22 and all you get are GVs, you just weren't expecting that, these technologies may help us mature those eggs in regular IVF when you get that unexpected outcome. So I think it's great areas of research and I'm excited to see what else comes down from them.

All right, so we're staying within assisted reproduction and I have an article called Association of IVF with Severe Maternal Morbidity in Low-Risk Patients Without Comorbidities. So there is already a body of literature that shows that patients who conceive with IVF may be at increased risk of obstetric outcomes being adverse and a lot of these prior studies have tried to adjust for the confounders because there's certainly confounding that's probably happening. And so some of this may just be that IVF patients are inherently different than patients who conceive with unassisted, just naturally.

And so the way they approached it here was to sort of exclude those patients. So there's something called the obstetric comorbidity index score and they excluded anyone who had a score of one or higher. So you had to have a score of zero, which essentially means you should be a very safe pregnancy risk patient when you get pregnant.

And that's who they looked at, looking at IVF versus those who conceived without IVF. And they had almost 40,000 pregnancies within their network that delivered and about 1% of those delivered having conceived via IVF. Just of note, that's about half of the national average.

So the SART is currently at about 2% of pregnancies conceived by IVF. The study took place over 10 years, so that may just represent that 10 years of data and the increasing utilization of IVF in the country. So they're taking these patients who are at very low risk and then they are controlling or adjusting for potential confounders that still may exist, such as age, BMI, race, ethnicity, and nulliparity.

As you might expect, the IVF group was different than the group that didn't conceive by IVF. The IVF patients were older, they were more likely to be nulliparous, they were more likely to be white, and they were more likely to have private insurance as opposed to public insurance. And so that's why they adjusted for those things as their potential confounders.

So what did they find? Ultimately, having IVF was associated, getting pregnant by IVF was associated with an increased risk of severe maternal comorbidity in these patients that should have a very safe risk during their pregnancy. 6.4% severe maternal morbidity with IVF and 2.4%. That's about a risk ratio of 2.5. So it was different even after they did adjustments for those other comorbidities. Their secondary thing they looked at was cesarean section.

So c-section occurred in 37% of IVF pregnancies versus 19%. And again, that's a risk ratio of about 2. So they were more likely to conceive by a c-section. And so these authors conclude that IVF is associated with a higher risk of severe maternal morbidity.

When you go to Supplemental Table 3, that's where you see what these morbidities were. They were almost entirely just blood transfusion. So blood transfusion was more likely in IVF pregnancies.

And then c-section was higher. And so their ultimate conclusion is that recognizing this association may help on the obstetric side to monitor these pregnants and tailor postpartum care better, which I think is a very reasonable and interesting conclusion. I do think it's interesting when you look at Supplemental Table 3. So that's where these severe morbidities are, which is their primary outcome.

So I kind of wish it had been in the primary table so you could see it in the paper since it's their main outcomes. But you can find it in Supplemental Table 3. And again, really, the only one that was increased in the patients with IVF was blood transfusion. It was in 6% of IVF patients versus 2% of those without IVF.

Everything else was actually lower in the IVF patients. And some of these other risks, you might argue, are much more significant medically than a blood transfusion, which I think, in general, we would think blood transfusion in the United States is relatively safe. So when you look at these other outcomes that were lower in IVF patients, this is like having to go on a vent, having acute renal failure, amniotic fluid embolus, air embolus, thrombotic embolism, acute heart failure, DIC.

Those things are actually lower in the IVF patients. So I just thought this was an interesting way to interpret these data. So overall, this is an interesting study that certainly shows an increased risk of C-section and blood transfusion.

We know that C-section is probably the major risk factor for needing a blood transfusion in obstetric patients. And so it's possible that that's a mediator of that risk factor for these severe maternal morbidities. But certainly, I also find these data reassuring that some of these other severe risks were actually lower in the IVF population.

I don't think that I'd be clear when you say lower, you don't mean statistically lower. You mean just numerically lower. So I don't want our listeners to say that it's safer with those other outcomes with IVF.

No, it's numerically lower. And if you add those numbers up, it's 0.4% in the IVF patients versus 0.7% in the non-IVF patients. And so that's the numerical difference.

It's about half. They don't do statistics on that, and they don't give us an adjusted number for those. But there were no incidences of those.

And so I overall found that to be reassuring. But what I liked about this paper was what they did different from the prior literature in that they were bringing in these patients who had that risk score of zero to start with. So these should be very safe pregnancies.

And so they're sort of trying to protect from or adjust for those confounders up front by taking in patients who should have relatively good prognosis from an obstetric standpoint. What are your guys' thoughts? Yeah, I agree. I mean, I also want to highlight that the numbers in the study were low overall.

There were 166 patients in the C-section group and 288 patients in the non-C-section group. So while they had a very large cohort of deliveries within a single institution and network of institutions in New York, the numbers overall I thought were pretty low. But I agree.

And I think it's really hard because blood transfusion guidelines differ from hospital to hospital in our system. Like, we don't transfuse unless somebody has an extremely low hemoglobin and is symptomatic, where I have seen patients who've gotten transfusions who have hemoglobins of nine. And so I think that it's less concerning.

It's obviously concerning, but it's less concerning to me that it was blood transfusion and not acute renal failure. And I think the DIC part of it, you think that the most severe hemorrhages are in those patients that have DIC, and you really worry about that. And so I think that was reassuring that it wasn't associated with ICU admissions or widespread DIC.

So I have three different opinions. First of all, man, is it difficult to be pregnant? Pregnancy is not a disease. It's much more complicated than that.

So going back to the first piece is who's to say? It's why I don't call it social surrogacy, because pregnancy is risky for all. And, you know, it's... But my second thought, though, Eve, is it's clear to me that IVF pregnancies are somehow systematically different than unassisted pregnancies. But the question is, how much does it matter? So I think there's two questions there, which is their clinical hat looked pretty safe, not all that different.

But there clearly still is something different about conception and pregnancy with IVF than without. And I still think we need to figure that out and try to minimize that to the best of our ability. I agree.

I also would have liked to see another sub-analysis of patients who conceived with medical assistance but without ART. And so I think it doesn't answer the question of is IVF different or is there underlying risk factor in patients who have infertility that predisposes to obstetric complications? Now, I'll just push back a little bit. There's so many people that are getting pregnant now with IVF that really aren't infertile.

So I mean, we have to pay attention to the differences by this method of conception and overall. It's not just the person that's had five years of infertility that miraculously gets pregnant now. So it's a public health thing.

We just need to understand. I agree. And I think a lot of it is probably abnormal placentation and something with the relief.

And that's my challenge. All of you investigators listening to the podcast, somebody please define for me what abnormal placentation means. It's such a vague term.

I'm just saying it as a joke that it encompasses everything and nothing. So we really don't know what the specific molecular is or the physiological is or something like that. Right.

But I worry with assisted hatching. I worry with trophectoderm biopsy that we are disrupting that early syncytiotrophoblast and the implantation and the interdigitation of those cells with the maternal interface. And I think we see more subchorionic hemorrhages in IVF pregnancies.

We see more early pregnancy bleeding. And I think that, yes, I don't know exactly what it is, but that, quote, attachment is somehow a little bit different in the IVF patient than it is in the patient who conceived without medical assistance. But yeah, I think it's still reassuring, although this is a very small sample size.

I think it's very reassuring in this population that the differences were not more severe. We're going to stay with this theme of morbidity, but now we're going to talk about it with the primary ovarian insufficiency in the epidemiologic section. I'm excited about these next two discussions we're going to have.

Oh my goodness, I hope our podcast hasn't turned to morbid if all we're talking about is morbidity. But this is another study that I think we all need to know as clinicians. I'm not sure it's going to knock you off your seat, but again, it'll give you a very good feeling of what the differences are.

And this paper is getting a little out of our ART field, and it's looking about the association between primary ovarian insufficiency and what they call multimorbidity. So we're talking about women now menopausal, and we're trying to find out what morbidities they have for the rest of their lives and medical morbidities. So this study is out of McMaster as the primary author and the research institute at St. Joseph's Healthcare.

And basically, as I said, they assessed the association of multimorbidity, which I'll define in a second, in ovarian insufficiency compared to early menopause and compared to regular menopause. So it's a prospective cohort study using the Canadian Longitudinal Study on Aging, where they have a cross-sectional data from 50,000 community patients across Canada—well, not across Canada, but in this health system—between 45 and 85 years old, and the data is not too far out of date. It's 2010 to 2015 they're looking at.

So the primary exposure, again, is primary ovarian efficiency, which they define as menopause under the age of 40, and they have comparators, which is regular menopause, age 46 to 55. They didn't call it regular. They called it average.

And then they have early menopause, which encompasses the women that underwent menopause between 40 and 45. Multimorbidity is their primary outcome, which they define as two or more chronic conditions, and there is also a secondary outcome they call severe multimorbidity, which is defined as three or more chronic conditions. So they do a nice job.

I'll just gloss over it relatively quickly. As an epidemiologic study, the methods are quite sound. They control for many of these confounders, including such things as age, hormone replacement therapy, education, ethnicity.

Interestingly, some social constructs they control for as well, which I think is a benefit to the study, including self-reported loneliness and living alone, as well as the traditional ones like BMI and smoking habits and physical activity. So again, as far as you can go for an epidemiologic study, this seems well done. So just to get down to the business, we're talking about more than 12,000 postmenopausal subjects, some women, and about 3% of those have premature ovarian insufficiency, which I think is about right.

So the prevalence is reasonable. So to cut to the chase, and I'll go a little bit more into detail, the overall relative risk of having multimorbidity in that group is about 2.5 and statistically significant. When you adjust, it still remains above 2, and it's also doubled in the group that have the severe multimorbidity, again, with more than a doubling of the morbidity.

So if you dig down a little bit deeper, what's giving you this morbidity, and it's the usual factors that we would expect. There's an increase in ischemic heart rate, which is almost threefold, increase in ulcers, increase in osteoporosis. So when I got down to detail, I'm sure that's not surprising you.

But again, it's really well quantified and very convincing as far as an epidemiologic study can be that these women with premature ovarian insufficiency really do have more morbidity later in life. And they compare it again to early menopause, which is surprisingly, or not surprisingly, right in the middle in terms of risk. So if you have early menopause, you have more morbidity than average menopause and less morbidity than POI.

But really, when you compare the two extremes, the, quote, normal and the, quote, really early, there really is a big difference in some of these findings. Now, if you'll bear with me, when you look at the tables, because they're really interesting, as Micah did in the previous study, I gave you the highlights. But if you look down in table three, the prevalence of chronic conditions, there is every condition that you can see there that is elevated.

Diabetes, depression, stroke, anxiety, malignancy, hypertension, kidney disease, and I can go on and on. So this is very convincing to me that it's not a spurious finding. Many of these individual other morbidities might not be prevalent enough to show statistically significance, but it's not like they cherry-picked and found a couple things that were elevated and the others were normal.

So the message is, we do see patients with premature ovarian insufficiency. I see, you know, occasional patients. We're actually more than that.

And I think what has to be in our consulting is, of course, we'll take care of you. Of course, we can go through the fertility issues. But, you know, you have care that you need to pay attention to for the rest of your life, even if it's not with me, that this really is something different.

It's just not your ovary stopped working, to put it into perspective. The function of your ovaries clearly have implications for the rest of your life, even after childbearing. And that's something as all obstetricians and gynecologists, we should, I think, understand.

Yeah, I think it's interesting. And I agree. And I wasn't surprised by these findings.

But I was surprised that a very small percentage of these women, I think it was around half, were on HRT. And so I think the real question is, could we alter the course of these women's health trajectories by hormone replacement therapy at a younger age? And is it, again, is it the POI and the pathophysiologic processes that lead to POI that lead to increased morbidity in the future? Or is it hypoestrogenism at a young age? Obviously, we know that that impacts the bone. That may also impact some of these other factors.

And if we had HRT beginning at the time of ovarian failure, would we see less morbidity in these women in the future? And I think that that's the real question. Yeah, that's an age old question. And we're not studying as much anymore since the Women's Health Initiative and all of that.

We really thought simplistically estrogen could fix this. But I don't know. I think it's more than just one hormone.

Although I believe estrogen has lots of effects, don't get me wrong. And I think we've underutilized HRT since those data. But I still think it's more than just one hormone.

You summarized the main learning point for me on this was just thinking about the long term effects, because I haven't done that with my POI patients. With PCOS, I do that. I talk to them about endometrial hyperplasia and cancer risks.

I talk to them about cardiometabolic risks and what they need to be doing long term for their health while I'm helping them get pregnant. I haven't really approached POI with that mindset. And so I think this article definitely puts that high on my radar.

It's very good learning. We're teaching everybody quite a bit on this podcast. Look at that.

I actually was going to say, I think we learn a lot as we dive into these articles. I know I do. I think we're about to learn a lot more as we talk about this next article.

And Kurt, I'm hoping that you can correct me where I don't summarize this well. So this is another Epi article. And this is from a group that did the FAST trial.

So for those who don't remember the FAST trial, it was a randomized clinical trial looking at folic acid and zinc supplementation in men in couples trying to conceive and ultimately did not find a benefit to pregnancy for those couples by supplementing those two. But as anyone who knows Enrique Schisterman and Sunni Mumford who led that trial, they build in lots of secondary trials on purpose into that initial trial. And so they had serum samples and they were able to assess those for all sorts of different biologic markers for vitamin D status.

But what they did here with this secondary analysis was they emulated a target trial. So they're essentially emulating a randomized clinical trial using this observational data. And I'm going to do my best to sort of talk through how they did that.

And then hopefully, Kurt can help fill in where I'm maybe not getting the details right. So in all the studies we just looked at, these were retrospective cohort studies that we just talked about. And obviously, the main thing that we're worried about in those studies is confounding and how do we control for that confounding to make sure that we're getting to the right answer in the question and getting that answer as accurately as possible.

And so there's lots of different ways. We saw that the first trial that the first study we talked about where they're excluding patients that had any sort of risk factors. So they're trying to adjust for some of that confounding up front by just who they include in their trial.

We're probably all familiar with how we adjust in our regression models or other models for confounding to try to control for it. Some studies will stratify. You can use propensity score matching to try to make the two arms as equal as possible.

Or a newer method that's emerged is doing this target trial emulation where you're essentially trying to account for that confounding as best as possible. So how did they do that? If you go into their supplemental details, which is really great information. I would encourage fellows trying to learn about this to go in there.

They describe how they do this. They're essentially randomizing these patients to three different statuses of vitamin D, either deficient, insufficient, or sufficient based upon the normal accepted definitions. So how do you randomize patients to a vitamin D status? You can't really do that.

That's what they're coming in on. But that's what a target trial emulation is trying to do. So in Supplemental Figure 1, they do their directed acyclic graph where they're essentially showing you what they think is the causal pathway and the confounding pathways and the mediators as you go from someone's vitamin D status as they attempt to get pregnant by male and female, all the way to getting pregnant.

And then if they do get pregnant, whether they have a live birth or a pregnancy loss. And so they're able to identify what they think are those confounders. And then they use logistic regression to see how much those impact that outcome.

And by doing that, they're able to essentially randomize patients who have an equal risk of these confounders into each of these three statuses, deficient, insufficient, or sufficient. And then they're looking at this as a couple status. So they're combining the male and the female together into the different categories that you can get in the combination of those three statuses.

And so they get four different arms of vitamin D status. And then they further provide weights to those potential confounders based upon their association with pregnancy. So the general idea, as far as I understand it, is that as best they can, they're emulating a trial where you could randomize patients into different vitamin D statuses.

Now, they note that this isn't the same as randomizing patients to supplementation, because supplementation of vitamin D is just one way that you get your vitamin D status. Your race, your ethnicity, where you live, how much you're outside, your diet, your exercise, all of those things play into what your vitamin D status is. So this isn't trying to emulate vitamin D supplementation.

It's trying to emulate a trial of all the factors that would affect the patient's vitamin D status. So that's the basics of what they did. So what did they end up finding? Well, for the guys, vitamin D status was not associated with the likelihood of live birth or a pregnancy loss within that couple.

But when you combine the guys and the ladies, the couples where the lady was deficient, they did worse than those couples where they were either insufficient or sufficient. So in other words, lower vitamin D was associated with a lower likelihood of those couples being able to get a live birth. However, it was not associated with the risk of pregnancy loss.

So ultimately, they conclude that vitamin D is associated with the ability to get pregnant, but not pregnancy loss. But ultimately, this tells us that we need a clinical trial to see if supplementing vitamin D can actually change those outcomes. So my big picture take is this sort of confirmed what I've seen in this body of literature since I was a fellow.

If you look at all medicine, low vitamin D status is associated with all sorts of adverse health outcomes, including a lower likelihood of the ability to conceive. What I haven't yet seen is that clinical trial that tells us whether or not supplementation of vitamin D status can help us change that outcome. In other words, can we supplement it and change it? Or is this just a reflection of something else within overall health that's driving those lower outcomes? And as well as they try to adjust for those confoundings, I don't know that we have that answer yet.

Kurt, can you maybe explain a little bit better what a target trial emulation is doing? I tried to do it to the best of my ability, but I still am having a little bit of trouble wrapping my head around. I think there's two messages here. One, at the end of the day, let's talk about the data.

I agree with you that it's really hard to ferret out is vitamin D some sort of marker of overall health or really fixing it is the problem. And I'm not sure I know any better based on this sophisticated method. So that's the take-home message from the trial.

But an emulated trial, I think, is probably going to be the future. And we do need to learn what this means. And it basically means when you have a very large data set, or let's just say a large data set, instead of taking all of them, especially let's talk about like for the SART data set or something like that, you can actually first pare down the patients to patients that might have been eligible for a trial.

That's the first step. In other words, set up the population. And then you have still a large number of people that would have met the entry criteria of a trial.

Then now, admittedly, they're not randomized or different, but at least you've got your population to something that you want to study, not the entire population, which might skew you and confound you and other things like confounding by indication. We probably need to talk about the statistical methods another time, but now there's a variety of new statistical methods that say, once you've got your population down to something that's manageable, you really want to get as much confounding as possible. And we've gone beyond logistic regression to where people are using instrumental methods.

They're using propensity scoring. What they're trying to do is group, randomize, I'm sorry, we reduce the confounding of the population, not just the individual patient. And therefore, you can get closer to these are people that, if I had a trial, would have been eligible for my trial and would have been in each arm.

I hope our readers and our listeners will pay attention to this because I do think that we're going to be emulating trials more than doing them in the next decade or so because of the expense and the complication of doing a randomized trial and because we have such big data. So I don't think I went into the details here. Perhaps we can in other studies.

But the take-home message is the statistical methods to reduce confounding are much better than they were before. Do I still think all confounding is gone? Probably not. So you have to do it probably in multiple ways and sensitivity analysis to be confident in your answer.

But I do think the answers we can get from large data, like a good randomized trial that they present, will become closer to the truth than just a, I have a cohort and I just looked at the means in the two groups and all I did was adjust for age. I went down a rabbit hole and there are some really good papers, sorry, Kurt, not just in F&S but in JAMA, talking about the methodology of target trial emulation. And when you have big data, you can also look at those patients that did have an intervention.

So the classic example they were using were HRT and long-term outcomes. And you can look at those patients that did have an intervention and look at the outcomes. And if you so-called randomize them in the beginning, you can glean some inference about whether or not that intervention was effective.

And so in this paper, where I got really confused, I was trying to figure out like what was the intervention, but there wasn't an intervention in this paper. They were longitudinally following these patients to see what the ultimate outcome was and they compared that to their original groups. But you can do target trial emulations where you try to look at interventions as well.

Yeah, I've done this a couple of times even. It really is fascinating. And I'm lucky that I can work with people like Sunni and Enrique.

But when we emulated the Women's Health Initiative, since we just talked about HRT, based on large data sets before the WHI, we actually got the same answer as the WHI. In other words, when you take out some of the confounding and really focus on your population, now there was interesting ways of controlling for things called prior event ratios. We also emulated trials using the SART database about fresh versus frozen and came out, again, with the same answer that some of the meta-analysis and trials did.

So I'm not saying it's perfect, but it's not just take a data set and squeeze it. It really is trying to find out where you think the confounding is, eliminate that with the selection of patients, and then go three steps farther in how you control for confounding, not just the logistic regression model. And you can, many people think, get a more unbiased answer.

Now it's up for us to decide whether the unbiased answer is the truth or not, and that's the fun of science, but I do think it's going to be the wave of the future. This is going to be my final question for you, Kurt, but you already answered it. Is this how we are going to be doing our observational cohort data sets? Is this what our fellows need to be learning methodologically for? I do.

I think it took a long time in my lifetime to get people to understand logistic regression and confounding and that you actually need it. Now it's going to take even longer to say that it's like version 2.0 or 3.0, that there are far better statistical methodologies than just logistic regression, which is a start, don't get me wrong. And I think the sophistication of the papers in F&S are going to reflect this, just like they do in JAMA and New England Journal of Medicine.

It really is beyond just controlling for age and number of embryos. And I'm glad you talked about DAGs. We should talk about that in another podcast too.

I think that's going to be important in many observational trials. And I get a lot of authors pushing back at me about why do I have to DAG in my trial? And this is the answer. This is the reason you really need to control for these kinds of things.

Well, thank you, Kurt. That was very educational and I definitely encourage all the fellows out there to read this and dive deep into the supplemental material. It's great.

So we have one last paper we're going to talk about. Eve, you and I are both program directors and this is from a prior fellow of Kurt's and it's such an interesting study about matching into REI fellowship. I'm very interested in what you're going to teach us about this.

Yeah, I love this paper. I loved it so much that the second I read it, I texted Divya Shah, who is the senior author, to say, oh, this is fantastic. First author was Leigh Anne Humphries and I guess she did this when she was a fellow with the UPenn group.

And so I think it's going to be super helpful in counseling future REI applicants about the likelihood of matching into REI. The title of the paper is Predicting a Successful Match Among Applicants to Reproductive Endocrinology and Infertility Fellowships. The objective was to identify independent predictors for the REI match and to validate a prediction model for REI match results.

So they abstracted applicant-level data from ERAS, including age, race, sex, ethnicity, citizenship, allopathic versus osteopathic versus international medical school, region of school, USMLE and CREOG scores, region of residency, rank of residency program from Doximity, residency affiliation with REI Fellowship Program, number of research work and volunteer experiences, and number of first author publications in peer-reviewed journal as well as other publications, oral and poster presentations. So all of the things that we counsel our residents to get involved in on a daily basis if they are interested in pursuing an REI fellowship. They had USMLE scores that were available in 94% of applicants and CREOG scores in 65%.

But since USMLE went to pass-fail in 2022, they pre-specified that USMLE would not be included in the final model that was used in the creation of a web-based calculator. So they then confirmed match results by review of the match list published annually by SREI or they confirmed this by direct communication with program directors or review of fellowship websites. Application characteristics were stratified by match outcome and then analyzed using descriptive statistics and univariate logistic regression.

Factors that were significantly associated with the match in the univariate analysis were then included in a multivariable logistic regression to determine which of these were independently associated with the outcome. And the primary outcome was successful match into an REI fellowship program. So then they did forward selection with likelihood ratio tests that were used to identify the optimal set of predictors.

And the model was then refined using various strategies. So this dataset included 286 REI fellowship applicants received at Penn from 2019 to 2023, which represented nearly all applicants nationally. 2020 was excluded.

I'm assuming you guys had an internal candidate as they did not participate in the match that year. So 199 of the 286 applicants resulted in a successful match and the univariate analysis showed that successful match was associated with the following, younger age, attendance in allopathic U.S. medical school, higher USMLE and CREOG scores, and training at a residency program affiliated with an REI fellowship program or a residency program with a higher ranking on doximity. And then multivariable regression identified seven independent predictors of match success.

Six were positively associated, affiliation with an REI fellowship program, rank of the residency program, publication of one or more first author publications, number of unpublished articles, abstracts, number of poster presentations, and highest USMLE scores. Attendance in international medical school was negatively associated with match. Training at a residency program affiliated with an REI fellowship program had the strongest effect with an adjusted odds ratio of 5.43. And I think that really, we can talk a little bit more about that, but I think that really leads to a lot of the other outcomes like more research and more publications.

Having first author publications was also highly influential. Interestingly, they did a sub-analysis of candidates who had negative residency-related predictors. So 131 applicants from lower-ranked residency programs without REI fellowships.

And in this group, 46.6% of candidates matched and the probability of match was dependent on research productivity. Including publication of at least one first author manuscript and poster. And so the prediction model showed very good discrimination.

The AUC was 0.883 and the best fit model had a sensitivity of 88.5% and specificity of 65.8% and correctly classified over 80% of candidates. My favorite part of this is they integrated this equation from the refined model into a web-based interface that anybody can get. And the link is in the article.

Enabling users to estimate their own match probability based on application parameters. I spent a bit of time playing around with this and it's a fantastic tool. And I actually got an email this morning from one of our residents asking to meet as I meet with all the residents who are interested in applying to REI.

And I cannot wait to start using this calculator to plug patients in. So I would say as a program director, this was very much in line with how I review applications, how I counsel our residents. It was not surprising, but it was so beautiful to see the methodology that went into this paper and to have a tangible calculator tool was, I think is a big contribution.

And congratulations to this group for doing it. And I love this so much as a PD. I actually sent it to our residency program director too.

Went to the calculator and put what I had at that time when I applied and I had an 80% chance of matching. But my program was a military one in Hawaii, so we didn't have that fellowship. If I checked the fellowship box, it went up to 95%.

So it's just an interesting association that I believe just from what we see. It's interesting playing around with it. Yeah, I believe it.

It's very interesting. And I think I also had to look up where Northwestern's residency program was on Doximity. We're number one.

I don't know how much of those things change, but I was actually pleasantly surprised to see that. This group was great. And interestingly, Dr. Humphries was our internal candidate in 2020.

So it's a kind of self-fulfilling prophecy that we found a really, really good person at Penn and she had the time and energy to put this together. I wonder how this changes because we've kind of all talked to people forever about this is how you get into fellowship programs. And now that's mathematically demonstrated that it really does matter who you know.

And the academic programs are looking for people from academics. It's a little bit discouraging that if you don't go to an academic program, you're going to have trouble. But now you understand the bias against you.

But I do want to say one thing that's not in the article working this group. Congratulations to all the REI fellows and the future REI fellows. They did this for all matching and other subspecialties as well.

And the quality of REI fellowship candidates was so much higher than every other fellowship candidate in all of these metrics. It's astounding. So we really do have the most competitive fellowship program in OB-GYN.

And I guess that's nice for our field because we get such incredible people. But it really is telling people if you really want to do this, which is a fabulous job, it takes a long time to build your resume, not just your last couple of years in residency. Yeah, and I will say that the programs that are not affiliated with REI residencies, those candidates were able to match because they sort of tagged on to other programs.

They established connections. And I think we've all had those candidates that we've mentored who are maybe at a local community program or at a hospital system that doesn't have a fellowship. And the advice that I always give to those candidates is find somebody at an academic program who has an REI fellowship and tag on to some of the research like that will help you.

So I think it's reassuring. And even though those candidates did worse, about half of them still matched. And so I think that research is really important for so many different reasons.

That is absolutely how I counsel our residents get involved in research, but not just get involved, but actually write the papers and be the first author and be the presenting author. And that will help you. So obviously we're a bias group and that's why we're on this podcast.

And that's why we work for the journal. We're academic in nature, but I can't help but say the quality of people we get in REI is really important to our field. I don't think we should be upscaling people that don't have the same intellectual demonstrative prowess, the desire to be academic, the careful planning of their career, the productivity.

We wanna keep our field this strong and this academic. I really strongly disagree that we need more fellows and lower the criteria because there's a workload issue. I think that those are short-term economic issues and we shouldn't sacrifice the quality of people that are in our field.

Yeah, I agree. And I think one of my favorite papers was by Larry Lehman showing the academic productivity of REIs. And it didn't matter whether or not you were in private practice or in academics that within 10 years over a thousand papers were published.

I don't remember the exact metrics, but it just really shows that the research component of REI, which is what we use for applicant selection, I think is really important to drive the field forward in so many different ways. And learning how to do research early on in your career and continuing that research is I think the key to understanding. It's why we do this, right? Understanding the literature, the limitations of the literature, but also the innovation and the new application of technology and the ethical application of technology.

So I agree, Kurt, I always quote you. I say the solution to the pilot shortage is not to have less highly trained pilots. You really want to have people who are trained and experts in the field to drive it forward.

And I look at the topics we just talked about today. I mean, we're going to be seeing more sophisticated research epidemiologically. We need research in the basic, you know, IVM and basic biology.

I mean, there's so many more questions to answer. And that's what I think academic medicine is. What a fantastic discussion today.

Thank you, Eve. Thank you, Kurt. To all our listeners, please like and subscribe the podcast.

Listen to all of the F&S on air, including the sister journals. And just to let everyone know, we'll be starting a new podcast this year led by Emily Barnard that will be covering the editorial content. So the views and reviews and the fertile battles.

So please stay tuned and listen to those. Kurt, Eve, it was great seeing you. I can't wait till next month.

It was great seeing you both. It was a small but mighty discussion. It was it was terrific.

So I all you listeners, tell a friend like us this. I think this podcast can really be continue to be a real good resource for everybody, not just our trainees. This concludes our episode of Fertility and Sterility on Air, brought to you by Fertility and Sterility in conjunction with the American Society for Reproductive Medicine.

This podcast is produced by Dr. Molly Kornfield and Dr. Adriana Wong. This podcast was developed by Fertility and Sterility and the American Society for Reproductive Medicine as an educational resource and service to its members and other practicing clinicians. While the podcast reflects the views of the authors and the hosts, it is not intended to be the only approved standard of practice or to direct an exclusive course of treatment.

The opinions expressed are those of the discussants and do not reflect Fertility and Sterility or the American Society for Reproductive Medicine.