Thinking About Ob/Gyn

Episode 10.3 Post-Cesarean Antibiotics, Hysteroscopy, and More!

Antonia Roberts and Howard Herrell

Recent evidence challenges the practice of prescribing oral antibiotics after Cesarean delivery in obese patients, finding no significant reduction in infection rates compared to standard preoperative antibiotics alone. Howard and Antonia analyze studies showing why this once-promising intervention may not be necessary.

• ACOG updates delayed cord clamping guidance to minimum 60 seconds for preterm infants
• Baby born at 21 weeks and zero days celebrates first birthday, highlighting advances in neonatal care
• Systematic reviews show no difference between chlorhexidine and iodine for vaginal prep before hysterectomy
• Conservative management of placenta accreta spectrum disorders shows improved outcomes over immediate cesarean hysterectomy
• Labor arrest Cesareans have highest blood loss among non-accreta cesarean indications
• New HPV testing terminology recommends "HPV detected" rather than "positive" to avoid relationship misunderstandings
• USPSTF preeclampsia prevention guidelines classify 89% of pregnant women as aspirin candidates despite limited evidence
• Endometrial sampling best practices include stepwise approach starting with ultrasound before considering hysteroscopy

In two weeks, Jacqueline Vidosch returns to discuss her son Noah who has trisomy 18, following a feature in the New York Times.

00:00:00 Episode Introduction

00:06:43 Post-Cesarean Antibiotics: Evidence Review

00:17:11 Delayed Cord Clamping Updates

00:22:13 Extreme Preterm Survival Case

00:26:40 Vaginal Prep and Placenta Accreta Management

00:30:11 Cesarean Blood Loss by Indication

00:34:21 HPV Testing Language Changes

00:37:45 Aspirin for Preeclampsia Prevention

00:51:33 Endometrial Sampling Question

Follow us on Instagram @thinkingaboutobgyn.

Speaker 1:

Welcome to Thinking About OB-GYN. Today's episode features Howard Harrell and Antonia Roberts discussing post-cesarean antibiotics and more.

Speaker 2:

Howard.

Speaker 3:

Antonia.

Speaker 2:

What are we thinking about on today's episode?

Speaker 3:

Well, we've got a few updates and lots of just short things from literature that we haven't been able to cover in the last few episodes, so a bunch of nice little things coming up for you. But first we should do a thing that we do without evidence.

Speaker 2:

All right, yeah, some quick hits yeah.

Speaker 3:

Well, what is the oral cephalexin and metronidazole or Flagyl to patients for 48 hours after a cesarean, if they happen to be obese, in order to reduce the risk of infection.

Speaker 2:

All right, this is a good one. I really like the ones that I also do it. To really examine that. So I think it was just towards the end of my training that there was a study that came out in support of doing this practice. It's not like it was taught to me from day one, but something came out that said, hey, we should do this. So some of my colleagues and I started doing it in those patients for that reason. So the idea is that undergoing a cesarean delivery increases the risk of infectious morbidity compared to vaginal birth, as does being obese, and then even more so if there's extra risk factors for infection, let's say prolonged membrane rupture or chorioamnionitis, of course. So with this study that had come out at that time, they had concluded that giving some oral Keflex and Flagyl for a couple of days after the cesarean would be beneficial in terms of reducing infectious morbidity for that at-risk population.

Speaker 3:

Okay, Well, we can cover for sure that original study you talked about. But, as it happens to be, the July 2025, Green Journal has a new randomized controlled trial that looked at this very practice. So they randomized patients from August of 2017, which was a bit after that paper was published up until August of 2023. And these patients were obese and they had ruptured membranes for at least four hours and then they had a cesarean. They had ruptured membranes for at least four hours and then they had a cesarean. And that's important because the subset of patients in that original study were people with ruptured membranes, not intact membranes.

Speaker 3:

So they were looking at the higher risk population for that original study group. So they enrolled 321 women and half of them received a combination of the two antibiotics she mentioned orally for 48 hours after delivery, and the other half received just the standard cefazolin and azithromycin that they might've received anyway as preoperative antibiotics.

Speaker 2:

All right. Well, what did they find?

Speaker 3:

They found that it makes no difference.

Speaker 2:

Okay, well, I was at first a bit surprised to see this study come out, but I really shouldn't have been, because, after all, there really was only that one other study out there before this one that looked at this at all. So, of course, the right thing to do is to try to replicate studies or expand upon them to determine should these really be standard of care. Just to be clear, it wasn't standard of care, it wasn't anything ACOG had said, but it just seemed, I think, a compelling enough study. But in this new one from 2025, they did both an intent to treat and a per protocol analysis, which was very good of them. We had just discussed a couple episodes ago about how good studies will do both types of analyses, a couple episodes ago about how good studies will do both types of analyses.

Speaker 2:

The intent to treat is important because there might be reasons in real life why people are planning to get the antibiotics but they can't finish the course. Let's say, one of them develops some kind of allergic reaction for some reason, or there's, of course, other reasons too or maybe in the real world, they were supposed to get some pre-op antibiotics but they didn't, and so now you're looking at they're already off of standard of care. So does it help to give them these extra antibiotics afterwards or not? So that's why you want to look at just the intent to treat and see what was the original plan and what was the outcome. And then the per protocol analysis is also important because that tells you like if everything was optimized and it was all done by the book, would this still be a worthwhile intervention or not? But by either analysis they found that it was not effective.

Speaker 3:

Right, and they did have some infections, of course. In this new study, their primary outcome was a composite of surgical site infection markers, but they looked at individual issues like fever or endometritis or wound cellulitis, but neither the primary outcome nor any of those individual outcomes were any different when you looked at the two groups. Another thing I thought was interesting was they looked at the type of skin closure and only about 10% of the patients in the study had Dermabond applied and about 20% of the patients had stapled incisions. We know that stapled incisions are associated with higher rates of wound complications and I think there's at least some compelling evidence that using Dermabond provides a lower risk of wound complications as well. Although that data is more mixed, you could do a subcuticular and then put stereostrips on it and a dressing, and so in both of these studies, the older one and the newer one, it's a little hard to know what percentages in which patients had what types of dressings. Particularly that's important if you're thinking about wound complications.

Speaker 2:

Yeah, that is an interestingly, I think, high amount of staple use. But we do tend to see, I think, in a lot of different types of studies, higher rates of complications. In general, but especially when we're looking at C-section infections, we're going to see higher rates of them in studies compared to in community settings because well, firstly, there probably is better chart review and better tracking in facilities that perform research like this, but also they almost always take place in teaching facilities where there's necessarily some learners involved in performing the cesareans. So we've talked about that. There's a lot of different reasons you see higher rates of complications when learners are involved. Some of that is common sense, but the point is there's a higher baseline rate of infection in residency programs than in private practice or community settings, even when you're adjusting for the same patient factors. So there are probably some other things that could be done to decrease the wound infection rates seen in certain patient populations and, of course, in certain training settings, but apparently the post-op oral antibiotics are not it.

Speaker 3:

Yeah, so the earlier study, as you mentioned, came out in 2017 and they randomized 403 obese women who had cesareans between 2010 and 2015 at the University of Cincinnati Medical Center, either to oral Keflex Flagyl Postop or placebo, in addition to the standard pre-op antibiotics. Now, that's important because we give azithromycin now to patients with ruptured membranes. That study, I believe by TADA, came out in 2016.

Speaker 2:

I'm riffing here, but I think that's right. Yeah, I think you're. Yeah, that sounds right.

Speaker 3:

And so most of the patients in this University of Cincinnati study would have predated the routine now routine use of azithromycin. And that could be a major difference in the studies too, because the new study those patients got azithromycin for ruptured membranes. They all had ruptured membranes. So that could be an important difference as well and maybe in a world where we don't do azithromycin maybe this has more utility. Is what I'm saying.

Speaker 3:

Now in the original study they stratified randomization according to membrane rupture so that it wouldn't skew the numbers, but they didn't have a power analysis for the subgroups, including the subgroups for ruptured membranes. It was just a mixture of patients, so they ended up being about 60 patients out of 200 in each group actually had ruptured membranes, for what it's worth. When they tried analyzing just the ruptured membrane patients, they also didn't see a difference in outcomes, but that again could be due to sample size. They also excluded anyone who developed chorionitis. The oral antibiotic group overall, though, had a post-op infection rate of 6% compared to 15% for the placebo group.

Speaker 3:

That included any superficial, deep or organ space infection within 30 days of surgery, and when they looked more closely at what type of infectious outcomes each patient had, only the cellulitis showed a significant difference. So for non-medical listeners, that's when you look at the incision and see that it's red, hard, swollen and tender. There's some subjectivity in that diagnosis. Other things like fever, wound separation and endometritis they weren't different between the two groups and I'll point out those are more objective Fever is present or it's not, for example. So here they also said that subcuticular stitches were standard, but staples were allowed at the surgeon's discretion and they didn't specify what dressings were put on the skin, whether that was glue or teri strips, things like that, and they only did an intent to treat analysis, not a per protocol analysis.

Speaker 2:

All right, yeah, so that's the study that me and my training colleagues were looking at and thinking maybe we should do this for certain patients Looks like it could reduce their risk of post-op cellulitis, and we were also starting to add the azithromycin at the same time and this was published in JAMA, so that, especially as a resident, carried some credibility for me. But that was a really good, just quick breakdown of that paper. So how do you look at that paper, especially now in comparison to this newer Green Journal study?

Speaker 3:

Well, I think there's several things and this could just be as simple as we give azithromycin now, things like that, it's okay. I also think we always need to be reminded that really we should have two RTCs that agree with something before we implement into practice. This is the McKenna mistake. This is now the mistake of Flagyl and Keflex the list goes on where we don't await confirmatory results, and maybe that's okay in some cases. But there weren't confirmatory results and we were making decisions based upon subset analyses, like the subset of the small number of women that had ruptured membranes that might have washed out in a more adequately powered study. But also it's another example, like the ARRIVE trial, of comparing unrealistically high numbers of infections, probably due to the effect of, as you said, it being in a training facility. So in their introduction they had cited previously reported post-Ossarian infection rates that were somewhere in the range of 3% to 12%. Usually 5% or 6% is what you see typically, and that's what they saw in this new Green Journal paper somewhere in the range of 5% to 6%. But they had 20%, for the obese patients in their institution were cited as having some sort of infection. So for that alone it's hard to say that this is broadly applicable to non-training settings or maybe even settings outside of their own institution. And so you have to wonder why is that rate so high? That's what happened with McKenna, where in the original study there was a very high rate of subsequent preterm birth, much higher than the rate that you would normally see. And you should ask the question why is that term birth much higher than the rate that you would normally see? And you should ask the question why is that? There was no statistically significant difference in outcomes that we also like, really care about. So they were all related to those subjective things like some redness around the incision.

Speaker 3:

And that's one of the difficulties here is that it could just be over-diagnosing cellulitis. One man's cellulitis is another man's rubor, just normal redness, that's healing. And in an academic institution you have the intern rounding on the patient and they're really cautious about it. Looks a little red and, oh gosh, we've got to put them on antibiotics. So that phenomenon too is one of the reasons why you see higher infection rates in residency programs.

Speaker 3:

It's just the over-diagnosis of it and the quick to use antibiotics for anything that looks red. But with time and experience you or I might look at that and say, oh, that's okay, it's day two and it's a little red. I doubt that's cellulitis, right. So lack of experience there. The other thing is that there's no evidence that giving these antibiotics prophylactically, even if it didn't decrease the rates of superficial wound issues, is any better than giving the antibiotics once you've diagnosed the superficial wound issue. In other words, why expose 100 women to antibiotics, with the inherent risks of antibiotics, things like resistance or even, for some women, perhaps developing C diff, in order to prevent seven women from eventually being diagnosed with cellulitis and being prescribed an antibiotic for that, if you could have just given the antibiotics at the time of diagnosis and the outcome was the same ultimately? So it's not like that they're saving hospital readmissions or reoperations or febrile morbidity for that matter, or anything right.

Speaker 3:

So this is another example of, again, the 2017 study, of a preliminary study that needed to be validated, and there are methodologic concerns in the sense that we've been talking about where objective markers weren't really different, but some subjective markers were different. But even if you're okay with all that, even if you think these findings are valid and, in fact, it really did reduce by seven percentage points the risk of cellulitis, well then ask yourself how does this patient population and these demographics mesh with your own population? 2017 study. The average BMI was around 40. And two-thirds of the patients in the study were scheduled cesareans with intact membranes. Almost the majority of patients had intact membranes, as you mentioned. And yet, in the 201 women in the placebo group, most of whom had intact membranes and were scheduled cesareans, 11 of them received transfusions and 31 of them were deemed to have a surgical site infection.

Speaker 3:

So this does not apply to my practice. I can promise you that, in fact, if it applied to any of the physicians that work in my hospital system, there would be a micro review for what's going wrong, right? So before you use a paper like that to change your practice, you should ask yourself if what you're seeing seems to match your own experience in your own patient population and do one out of every six of your patients have surgical site infections. Does that many require blood transfusions on a routine basis, that sort of thing. So JAMA is a good journal.

Speaker 3:

It wasn't in the Gray of the Green Journal. It probably got passed on by the gray or the green journal, you never know. It reminds me of a study a few years ago that looked at routinely giving antibiotics to every patient who had a laceration at the time of vaginal delivery and of course they found that there were lower infections, some statistically significant lower number. But it just doesn't make sense. Ask yourself how many times have you had to treat with antibiotics a infected vaginal delivery laceration? I don't think I ever have, honestly. And would the outcome be any different if you gave the antibiotics before versus after?

Speaker 3:

Yeah you'd cause a lot of women to be exposed to antibiotics unnecessarily so. In this new study from the Green Journal, the overall infection rates were 5% or 6%, like we would expect from all the other reasonable literature, and that would probably mesh with our own real-world experiences. Maybe you're a little above or below that, and so I would say that the devil is in the details here. Only 2.7% of the patients in the new study had cellulitis, and so that seems to match with what I've seen in my population, and maybe the difference is the azithromycin. It's perfectly fine, but the 2017 study just doesn't apply to our patient population or in a world where we're giving azithromycin.

Speaker 2:

All right, very good, now we know, stop doing that routinely. I think we have better evidence to not do that than to do it.

Speaker 3:

Okay, so let's, should we do some of these quick hits and catch up on some literature we've missed?

Speaker 2:

Yeah, let's do it so. But before we hit the quick hits, I just want to compliment the last episode that you and Scott Guthrie did. I learned a lot listening to him. That field of neonatology has changed so much, even in the few years since I've come out of training. We used to have a lot of lectures by neonatologists and the numbers are very different now, so it's very impressive, are very different now, so it's very impressive.

Speaker 3:

And I do have a couple of relevant updates from the literature that have even come out since he was on just the last time.

Speaker 2:

So the first one you guys were already ahead of the curve on this. You've had this quality improvement project in Tennessee. It was recently published in the Pediatrics Journal, focusing on optimal cord clamping for at least 60 seconds. In a prior episode we had discussed the reasons for 60 rather than 30 seconds. That was the previous recommendation from ACOG to do at least 30 seconds. So your new goal of 60 seconds or more is based on more recent scientific evidence.

Speaker 3:

Yeah, the 30-second recommendation, I think, was taking into account the babies might potentially need resuscitation, so it was always 30 to 60. You weren't going to wait the full 60 seconds to start chest compressions, for example, just to get 60 seconds of cord clamping. So I think it was partly due to that. Patients still come in routinely asking me all the time for delayed cord clamping, as if it's not standard of care that we've been doing forever. But it is. It's been an evidence based standard since probably 2010. It's been an ACOG recommendation since 2016,. Nearly a decade. And, yes, in the past it's been phrased 30 to 60 seconds to leave some leeway, I think, for some of these different clinical scenarios that come up.

Speaker 2:

Yeah, and even I think this morning I saw in my little reel of news headlines there was a study about giving supplemental oxygen while still doing delayed cord clamping and giving at least 30% up to 100%. So that's something I've not routinely done, but it seems like that's a way that you could still do resuscitation but not cut the cord too early.

Speaker 3:

That's right. You can do resuscitation on a mother's chest.

Speaker 2:

Yeah, but ACOG did just change their guidance about this issue this past week.

Speaker 3:

They work on these things for months. The literature moves as it moves and hopefully, if people are paying attention. These things all mesh and come to be so our guidelines and things are necessarily a slow moving thing. To make sure we're not. For example, there was never a guideline that says give Keflex and Flagyl because you wait for some confirmation, you wait for the data to mature and there you go.

Speaker 2:

Yeah yeah, so you were just on the same wavelength at the same time. But the new guidance recommends 60 seconds as a minimum time for delayed cord clamping in preterm infants. They specify who do not require immediate resuscitation.

Speaker 3:

Yeah, and this guideline change focuses right now on preterm because the magnitude of effect is definitely greater in these smaller babies, and so the literature is becoming very clear how important this is for the smaller infants.

Speaker 3:

I think that's why the guidance says increase it to at least 60 seconds. Ironically, these smaller babies are the ones that we most often want to hand off as quickly as possible to the pediatrician. We just assume automatically that they need immediate resuscitation because they're preterm, but, as Dr Garthry put so well, they're the ones who benefit the most from the delay, because you're literally giving them a blood transfusion right as they're born. I think over time we might see two minutes become the standard for term babies. I'm frankly doing that as much as I can already as that literature evolves. But the magnitude of benefit is so small in those term babies for whether it's 60 seconds or 90 seconds or whatever, that it just makes it hard for well people like ACOG to make a firm strike your fist against the table recommendation. But for these preterm babies 60 seconds is now the standard of care.

Speaker 2:

Okay, and the other thing that you two talked preterm babies 60 seconds is now the standard of care, okay, and the other thing that you two talked about a lot, of course, was how neonatal survivability in the preterm period has improved so much. It's no longer uncommon to see resuscitation efforts at 22 weeks with good outcomes, and I'm sure a lot of our listeners have seen the headline lately about baby having his one-year-old birthday, and he was born at 21 and zero. He was born vaginally in Iowa, so this appears to be the earliest, very well, reliably dated pregnancy to have survived to at least one year of age.

Speaker 3:

Yeah, they said that it was 10 ounces, which I would assume just doing the conversion is about 280 grams, so it was just amazing. This is still big enough to have a breathing tube place and to access the tiny blood vessels, but apparently they would not have offered resuscitation even a day earlier. That's their current cutoff 20 weeks and six days. There is no real reporting about the child's neurologic condition that I've been able to find or the degree of disability that he may have. They did note that he had a perforated bowel and survived a very risky surgery for that and spent six months in the NICU. These are absolutely extraordinary, rare outcomes and I do wonder or worry that one of the dangers is that patients come across these sorts of stories and then they expect the rarest of the rare outcomes for their own pregnancies, not realizing that this is one in a million, to put it mildly.

Speaker 2:

Yeah, it's important to remember. If this was common it wouldn't even be reported. But that's exactly why it was reported is no one would have expected this kid to survive this long. Every other pregnancy ever that's been born at 21 week zero days has not made it to one year of age and we still don't know what kind of long-term outcomes they're going to have to deal with. But it does show that a lot of the things you and Scott talked about on that prior episode have dramatically improved survivability even in the last few years and certainly the last 20 plus years. This means that even the 24, 23, 22 week babies are doing a lot better, with less disability, less mortality, and it's just like the rising waters lifting all the ships, if you will. But yeah, this 21 and zero week baby still represents the absolute cutting edge.

Speaker 2:

Clearly the institution was planning and comfortable offering resuscitation at that age, but I'm not sure that all facilities can do that and I'm not that far out from training. It's been less than 10 years and I remember during my time there was a very hard line at 24 weeks and even a day under 24 weeks no baby was considered viable. They only offered comfort care and just allowed the baby to pass away from not breathing. So this is really a huge difference from just a few years ago and I'm sure in this 21-week case they counseled the patient, I'm sure very severely, just very realistically, that there's no guarantees and it could be a very painful road, no matter which way they chose. But obviously this outcome of the baby turning one years old now certainly gives hope and I'm sure a lot of facilities have been taking note of this for a while and are probably working on shifting their timelines for earlier resuscitation.

Speaker 3:

Yeah, and they are going to be facility dependent.

Speaker 3:

So we discussed in that episode the outcomes of these pregnancies, and the outcomes are improving.

Speaker 3:

It's also tempting for us as obstetricians to take credit for some of the improved mortality and morbidity that we're seeing or the improved neurologic outcomes that are being achieved.

Speaker 3:

If you've been doing this for the last 20 years or so and you're seeing how much better many of these outcomes are you this for the last 20 years or so and you're seeing how much better many of these outcomes are you could be tempted, for example, to attribute that to the success of magnesium sulfate for neuroprotection, but there's simply no evidence that these that well, that intervention in particular is contributing to these improved neonatal outcomes.

Speaker 3:

For our younger colleagues, they may not realize that we've been using magnesium sulfate as a tocolytic long before we were using it for neuro protection specifically and at the same dose since the early 1980s, and so it might feel like it's this new intervention that we invented in 2009 that's led to these improved outcomes. But it hasn't, and the long-term data from randomized controlled trials that have followed these infants versus placebo later out have not found improvements in these outcomes, and nor has a long-term epidemiologic data. But when everything's getting better, it's easy to misattribute why it's getting better, and so a cautionary tale, not specifically about magnesium, but about a lot of things, and I think that episode we talked about some of the things that actually should get the credit.

Speaker 2:

Yeah, or like some kind of prenatal care intervention. I agree I don't think anyone out there is trying to argue that magnesium is why a baby has lived, born at 21 weeks, although I did listen to that episode. Like I said, and if Dr Scott Guthrie is doing my baby's resuscitation and he says, give magnesium, give steroids, delayed cord clamping, all of those things, and I'm just going to do all of those things. But you guys did talk about some of the actual neonatology improvements that have most contributed to these improved outcomes. You mentioned surfactant and better ventilation management as a couple examples there, and better ventilation management as a couple examples there. So we'll definitely have to try to get Scott back on in the future. It was just very delightful to listen to him and we always get a lot of positive feedback whenever he's on.

Speaker 3:

Yeah, he's great at what he does. Okay, well, onto the May issue of the Gray Journal. There was a systematic review that concluded that there's no difference in infection rates at the time of hysterectomy when you use a vaginal prep with chlorhexidine versus an iodine prep.

Speaker 2:

I saw that too. That was another pleasant surprise. I remember a time I think it was during my training where vaginal chlorhexidine was considered superior, and it was the standard prep in one of my facilities I trained in. But one is not superior to the other in this study, and I do believe betadine is cheaper. So it's good to know that you can stick with betadine as the standard. But then, let's say, someone has an iodine allergy. You can just as safely use chlorhexidine instead. If I remember correctly, there's a certain type of formulation that was used that's less irritating to the vaginal mucosa and that was preferable. So, moving on and sticking with the same journal, there's another systematic review looking at conservative management of placenta accreta spectrum disorders compared with doing an immediate cesarean hysterectomy, and this found that conservative management was associated with improved outcomes such as lower blood loss, less need for blood transfusion, less genitourinary injuries and fewer ICU admissions.

Speaker 3:

And this is great also and it confirms what we've known, I think, for a while now that for particularly focal placenta accretus disorders there are alternatives to hysterectomy. This would include focal resection and repair of the uterine wall or also just leaving the placenta in situ. Those patients have been managed medically and of course, it not only would result in better surgical outcomes, as you described, because you're not doing surgery, but in many cases, preservation of fertility because they still have their uterus. So lots of options that 20 years ago just weren't considered an option. And in that regard too, there's another article that I thought was interesting that looked at the impact of different indications for cesarean on the amount of blood loss.

Speaker 3:

This study looked at almost 5,000 cesareans and looked at the quantitative blood loss by indication for cesarean, and some of this was going to make common sense. The largest amount of blood loss was among cesareans performed for labor arrest, if you exclude the ones that were done for placenta accretum spectrum disorder. Now, those arrest patients likely have the most exposure to oxytocin during their labor, but you also might look at that as they're the patients who were least responsive to oxytocin during their labor. But you also might look at that as they're the patients who were least responsive to oxytocin and that's why their labor has arrested or not progressed as you would expect it. So labor arrest is a risk factor for hemorrhage, as is oxytocin exposure, which are perhaps surrogates for the same thing, but the common mechanism is likely related to poor responsiveness of the myometrium to oxytocin. So even when you give oxytocin postpartum, the uterus just doesn't contract well and so they bleed more.

Speaker 3:

The other thing that I thought was interesting about this was that planned cesarean hysterectomies for placenta accreta spectrum disorder. In those cases, the estimated blood loss was only about not even 400 mLs greater than the cesarean's performed for labor arrest, and I really appreciated that point, because cesarean hysterectomies that are scheduled, they usually go very well and they're rather straightforward and people shouldn't be afraid of them, not much more than they are afraid of a labor arrest, cesarean. But it's those cesarean hysterectomies that are unplanned and that develop over time. Those are the ones that can sometimes be nightmares and have unexpected massive hemorrhage.

Speaker 2:

Yeah, I would imagine, as long as they're not like, maybe a percreta would still have more blood loss if there's blood vessels all over.

Speaker 3:

But even then you've gotten them to Stuart Winkler and they've done the preoperative stuff and you've planned for it.

Speaker 2:

Yeah, exactly, and so that almost sounds like an argument to do the hist for all of the sena accretas. But obviously you're going to. If you're going to elect for conservative management, you're going to make sure it's hemostatic, and if it's not hemostatic then you just go do the hysterectomy. But this decision to do a hysterectomy in a cesarean that was not planned typically comes after prolonged bleeding and other problems, so it can be associated with massive amounts of blood loss just before you even get to the hysterectomy. So usually the blood loss is not because of the hysterectomy itself, it's because of all the decision, all the things that led to the decision. And that's an important point, because we know a hysterectomy is an extreme thing to do during a cesarean. It's not the routine thing by any means, and it's so much so that we almost always discourage a hysterectomy for any other reason than intractable bleeding, Like even for a cervical cancer patient. Or let's say they have a ginormous fibroid or they have a history of really bad periods, bad endometriosis.

Speaker 2:

I get some patients that will say I'm pregnant, I'm planning a cesarean, why can't we just take out the uterus at the same time and be done with it? And the answer is almost always no. It's safer to have a whole separate surgery later on, when you're further out from delivery, because those blood vessels are huge. There's so much more risk for all kinds of post-op complications. You're essentially throwing a major surgery on top of a major surgery on top of a pregnancy. So for us to do a hysterectomy with a cesarean, we tend to think of it as a very last resort, very morbid kind of thing to do that we're trying to avoid if at all possible. We're trying to avoid it in some accreta cases even. But in the setting of severe bleeding you also don't want to wait too late to perform the hysterectomy because that'll paradoxically increase the risk of massive blood loss and other complications, including death. So the hysterectomy is a life-saving procedure if you do it in time. So do it in time when it still can save their life.

Speaker 3:

Yeah, you could be getting into DIC, for example, before you even made the decision and then you're in the deep doo-doo, as they say. Yeah, the deep doo-doo as they say, yeah, okay. The other interesting quick hit from that same edition was an opinion piece that provides an updated understanding of the natural history of cervical human papillomavirus infection.

Speaker 2:

Yeah, this one was interesting.

Speaker 2:

They recommended using different language when we discuss HPV test results. So if the test is positive, then we should say HPV is detected, and then if it's negative, we should just say HPV is not detected. The concern is that when we just say you're HPV positive, you're implying firstly that there's an active infection going on that wasn't there before, and then when you say you're HPV negative, you imply that there's no virus at all, when really there could be and it just wasn't detected. So patients may come in, they may have an HPV negative pap smear and then the next time they come in it'll show HPV detected, and so their initial assumption is going to be that they've gotten HPV since their last pap through some kind of sexual contact during that interval, when in fact they may just simply have had a recurrence of a prior HPV infection and nothing has actually changed for them. The test really reflects immune function and immune control of the virus rather than total clearance. The HPV virus and its presence in the cervix probably works more like herpes as opposed to, say, chlamydia.

Speaker 3:

Right. I think we've known for the longest time that you don't truly clear HPV in most cases. But you can see the impact on the patient if she assumes that she has a new HPV infection when in fact she may simply have re-inoculation with a virus that's been dormant and she's had it for 20 years or more. Misunderstandings about this can destroy marriages and relationships, so the new terminology is thought to maybe avoid some of those unintended implications of saying positive or negative, but ultimately we have to educate our patients about the natural history of HPV infections.

Speaker 2:

All right. Moving on, We've covered a lot on here in the past about baby aspirin for preventing preeclampsia in pregnancy and there seems to be a very real possibility that we're not helping anyone with that lowest dose, the 81 milligrams in particular, and that we've also expanded the use of it to too broad of a population with the recent US Preventative Services Task Force recommendation on who qualifies for it. So there's a new JAMA Network article from July 2025 that tries to assess how good a job these USPSCF guidelines do at predicting risk. These USPSCF guidelines do at predicting risk and their conclusion is that the guidelines provide very little utility for estimating preeclampsia risk.

Speaker 3:

Right. This is a great example of the concept of therapeutic drift. So you have a reason to believe that something works for some population. Let's say you've got preliminary evidence that aspirin prevents preeclampsia in patients who've had preeclampsia in the past. Then, because aspirin is viewed as a relatively low-risk intervention, you implement this intervention without confirmatory evidence that it's actually effective or meaningfully so. And maybe that's okay.

Speaker 3:

But then you ask the question well, if it does seem to work with people who've had a history of preeclampsia, you ask the question well, if it does seem to work with people who've had a history of preeclampsia, why don't we use it for everyone who has the same risk of preeclampsia as people with a history? We can calculate the statistic right. They're at X percent above a certain threshold. Maybe we should use it. And that's what's happened. Or obesity, or eventually, every patient except for well, a thin, white, multiparous woman under the age of 35 with no prior preeclampsia or hypertension. Essentially, that's the new definition of low risk.

Speaker 2:

Yeah, yeah, because if she was, nulliparous.

Speaker 3:

That would be a risk factor as well. That would be a risk factor. Yeah, so a lot of things develop this way. Our antenatal testing guidelines have developed the same way, where we do antenatal testing on every patient who has a certain risk, percentage-wise, of fetal demise, even if the testing hasn't been shown to change that risk Think cholestasis and obesity, for example. And then we get this slow creep of indications for testing or, in this case, for who should take a baby aspirin.

Speaker 2:

Yeah, so the current guideline says you're at increased risk for preeclampsia this is going to be such a long list If you have one or more high risk conditions. So prior preeclampsia, having a multifetal gestation, twins or more, chronic hypertension, pre-gestational diabetes, kidney disease or autoimmune disease. So those are the high-risk ones. And then you're also considered at increased risk and should get aspirin if you have two or more moderate risk factors, and these include nulliparity, obesity, black race, age 35 or greater, having a mother or sister who had preeclampsia, having low income, having an IVF pregnancy, having a personal history of certain adverse pregnancy outcomes that includes low birth weight or small for gestational age, and then just other unspecified adverse outcomes or having more than a 10-year interval from your last pregnancy. So that list of the moderate factors are really what just explode the applicability to almost everyone. So even I fall into that because of my age and having IVF, for example.

Speaker 3:

Even though you've had two pregnancies without hypertensive disorders you're as low risk as they come in another worldview. So with those criteria, the authors looked at a population of over 5,600 patients and found that 89% of their patients would be considered at increased risk and deserving of aspirin by this rubric. It basically just does get everyone, except for a thin white woman with a previous healthy pregnancy, who's not poor and hasn't had IVF. I mean, it's crazy. And it's under age 35. At some point you have to decide what level of risk is sufficient to justify the intervention, and maybe the logic here is that aspirin is cheap and harmless, so why not give to everyone, which is something I hear people say.

Speaker 2:

Yeah, maybe that even was their goal here, because clearly they're covering almost 100% of people and the lower the perceived risk of any intervention, you know, the lower that threshold is going to be for just applying it. So, and for what it's worth, I did not take the aspirin because it gives me hives, so I'm allergic to it, and I've certainly had a few patients push back and say if I don't need it, I'm not going to take it, and so if they're in that moderate risk category, I don't really die on that hill over it.

Speaker 2:

Absolutely Okay. But overall, in general, I think most doctors and also most patients don't really overthink it. They'll just say oh, the guideline says aspirin, and the patient says, oh, the doctor says aspirin, it might prevent preeclampsia. It costs like maybe a dollar, so no harm, no foul, and they just go ahead and take it.

Speaker 3:

Right, and in the same way, antenatal testing is viewed as harmless, though there is more cost and inconvenience to the patient, that's for sure. I also think people could do a better job seeing the harms of false positive results that lead to unnecessary induction or atrogenic prematurity. With antenatal testing, at least, I think that they just accept that those risks are worth it, because somehow we're preventing a stillbirth In exchange for that. Somewhere we're preventing a stillbirth, and there's also economic incentives for the clinic to do more antenatal testing, whereas there isn't really any for prescribing aspirin or vaccines. Rfk.

Speaker 3:

Yeah, but in both cases we started widespread interventions in the absence of clinical trials that show specific benefit. Now the authors of this current study were talking about aspirin. They were simply pointing out that the current rubric that the U specific benefit. Now the authors of this current study were talking about aspirin. They were simply pointing out that the current rubric that the US Preventive Task Force Service has put forward doesn't do much to tell us who's actually at high risk for developing preeclampsia. Just everybody right. And we know that the rate of preeclampsia is, or that the risk of preeclampsia is not 89% of patients or whatever. Frankly, I bet most physicians aren't following it anyway, and they've shown that the uptake is low, at least not for those patients with moderate risk factors.

Speaker 2:

Yeah, I think if you take residents and med students and you give them a guideline or at least gave me a guideline as a resident, I was like the guideline police back then and I had that list of risk factors posted up next to my workstation and I was recommending it to 89% of all my patients because that's what the USPSTF said. But yeah, I've backed down on that since my training because some of those are vague. Like what exactly means low income? I think that's a relative term. What is a prior adverse pregnancy outcome?

Speaker 2:

Let's say they had a prior stillbirth for a genetic abnormality or an emergency cesarean for cord prolapse. Those are adverse. But how does that increase their preeclampsia risk? It just doesn't make sense. And then even on the race question, not everyone is just straightforwardly white or black. If they have a really diverse family of origin, then how do you draw the line on whether their race fits enough into the category to be a risk factor or not? So yeah, over the years I've shifted only to recommending it when the patients are very clear-cut high-risk candidates. So prior preeclampsia and now they have twins, that kind of thing.

Speaker 3:

or maybe if they have all of those moderate risk factors too, yeah, but again, the med students and residents may be doing a great job of this, but even when we tell patients we know this to take the aspirin, most of them probably aren't taking it?

Speaker 2:

Yeah, probably not.

Speaker 3:

Now in this study they also found in this population of patients that among the individual moderate risk factors they looked at this and they controlled for this Only noliparity. Of all those moderate risk factors that you read I think it took four minutes of all those moderate risk factors, only noliparity was consistently associated with a modest increased risk of preeclampsia. In other words, obesity, black race, advanced maternal age were not associated independently with an increased risk of preeclampsia. If anything. In their data, advanced maternal age was protective when you adjusted for the other confounders.

Speaker 2:

That's very interesting. That, I think, gets to an underlying problem with a lot of data analysis, because it can be true that if you just take a group of women over 35 and compare them to a similar group of women under 35, the older ones will have an increased risk of preeclampsia because of other things it's more likely that those older patients are going to have hypertension, diabetes or other comorbidities, that those are why they have the increased preeclampsia risk. So, yeah, this information you cited says a 35-year-old without other risk factors does not have a higher preeclampsia risk than a 25 year old. And yeah, that's surprisingly true even for obesity and black race. If they don't have the other risk factors, then they don't have the risk.

Speaker 3:

So so I guess this lets me off the hook at least for the aspirin, and that's a problem that that the authors even point out that all of these things in that list have come from independent studies that looked at one thing, not from a study of heterogeneous population of patients that adjusted for confounders. So among obese patients you're going to find more diabetics and hypertensives. But if you control for obese women who aren't hypertensive or diabetic, then their risk of preeclampsia is not different, and that's the point. Extend that to older patients who've had a longer time to be smokers or be obese or have hypertension or diabetes, or black patients, for whom obesity is more common things like that.

Speaker 3:

For the nulla parity, the relative risk of preeclampsia was about 1.5, which in and of itself would be below the threshold of risk that anyone would recommend prophylaxis for. It's too small of a percentage. Even if you believe that the prophylaxis with a baby aspirin is effective and we've talked before that there's still no clinical trial that shows that a baby aspirin is itself effective.

Speaker 2:

So if these studies findings are true and you just take away all those moderate risk factors, then we really are back to where we started with just that high risk list, and it's still reasonable to prescribe it or recommend it if they've had prior preeclampsia or they have active hypertension or diabetes.

Speaker 3:

Yeah, maybe. Now we're back to the issue of what dose of aspirin should we be using. And it should be noted that, despite the dramatic increase in the use of aspirin for this indication over the last 10 to 15 years, the rate of preeclampsia is only increasing. It's not decreased. It's surprising, right. It could be increasing despite aspirin having a positive benefit, because the headwinds, the negative headwinds, are greater or the attention that it gets is more intense, or something like that.

Speaker 3:

But we still don't have a trial that shows that, and we have trials for many specific cases, like among black patients. It doesn't help them to put them on aspirin when that's their risk factor. So to all the social media influencers who believe that aspirin should just be given to every single pregnant woman well, they might have drunk the Kool-Aid here a little bit, but this is what we do. It's a fad. It goes through the hype cycle, it gets overused, and then this article and other recent trials that show a lack of utility and that sort of thing. They're often the first salvo on pushing back against the overutilization of aspirin if you buy the hype cycle idea.

Speaker 2:

And this paper about the USPSCF still talks about the idea that why not just give the aspirin to everyone? I think they're concerned that the guideline waters down the recommendation and we're not getting the aspirin to the people who most benefit, and so one solution is just to tell every single patient to take it. If it really does reduce the risk of recurrent preeclampsia in patients with prior preeclampsia, then we need to make sure those patients are getting covered within the guidelines and that we're focusing on them. And we can get burned out by telling, looking at this little list and trying to be like, wait, what's her BMI? Okay, what's her race? Again, did she have a sister? Like we can get burned out, getting too far into the details, burned out, getting too far into the details, what's her income? Et cetera, and then that can take away from our focus on the truly high risk patients, the patients who actually benefit.

Speaker 2:

Yeah. So they're saying, instead of that, just why not just have a universal recommendation?

Speaker 3:

Yeah, Well, this can be our social media segment for this episode, because I do see it among OBGYNs on social media just feeling like they're evidence-based and they're doing this great thing by telling everybody to take an aspirin and a prenatal vitamin because they believe it does something that's not been established in the literature.

Speaker 3:

So I would say don't give any medicine unless you have a clear benefit proven by two randomized controlled trials right.

Speaker 3:

But when the medicine is perceived again as having no real downside, then we give it on the hope and prayer that it might be beneficial for somebody and it's not gonna hurt anybody. But in response to that idea the authors of this study that the idea that we should give it to every single patient they do cite a couple of studies as a cautionary tale. The first one is from Sweden, from 2021, published in the Gray Journal, that noted that patients who use aspirin have an increased risk of postpartum bleeding and postpartum hematoma and potentially an increased risk of neonatal intracranial hemorrhage. I think this study is one of the reasons why we've seen some recent trials encouraging earlier cessation of aspirin to avoid those sort of third trimester, antinatal or intrapartum risks and we reviewed the evidence here a while ago studies that looked at really no evidence that aspirin is beneficial when you continue it past 24 to 28 weeks, and not any different than if you continued all the way to 36, where you might be more worried about some of those particular negative outcomes.

Speaker 2:

Yeah, we talked about that one before and, just to refresh people, that one was published in JAMA in 2023. And those patients had that SFLT1 blood testing done and had a normal result. And those patients were the ones where they said if you stop it at 28 weeks or even 24 weeks, there's no difference. It's just as safe to do that. And yeah, that Swedish study about hemorrhage risk with aspirin is important. I was specifically taught and trained that there's no increased risk of hemorrhage even if you take it to the day of delivery. But clearly that's incorrect.

Speaker 3:

Yeah, that's what I mean by drinking the Kool-Aid. Once we identify these things as key to what we do and important, then we see everything through that lens. They also cited a Cochrane review that said that aspirin increases the risk of postpartum hemorrhage and perhaps the risk of placental abruption. So, again, no intervention is without risk and of course, we're at a point now where we have to really consider whether we use a higher dose of aspirin, which might increase some of those risks. If those are real risks for 81 milligrams, they should be more real for 162. So we need to make decisions based upon quality replicated interventional trials.

Speaker 2:

Yeah, that one's tricky, especially because preeclampsia also increases the risk of flucentylabruption. So anyway, we'll continue updating on this when more information comes out, but for now we're going to move on. We've got time to cover a really great listener question we got, so I'm going to read it out and then we'll discuss. Dear OB, I've been thinking a lot about endometrial sampling for both pre or perimenopausal women as well as postmenopausal women, specifically the issue of a blind endometrial biopsy with a pipel versus visually directed endometrial biopsies with hysteroscopy. The evolution of office hysteroscopy and the call in quotes to end the practice of blind sampling and blind curatage by leaders in the field are compelling. But do the purported benefits? Yeah, now that's their real moniker.

Speaker 3:

We like it when you guys provide the name for us, so we don't have to make it up provide the name for us so we don't have to make it up. So this is one of those great questions, because I think this question exists at the intersection of industry marketing, a potentially enhanced billing and reimbursement for the doctor who adopts the practice and at least the probability or possibility of an improved patient outcome. And when you have any two of those three things, this is a really people go crazy, a big uptake in clinical practice.

Speaker 2:

When you see any two of those three things present, yeah, I think if you go to any OBGYN meeting you're going to see reps like demonstrating their little office hysteroscopy setups. The industry certainly wants us to do more of these office hysteroscopies. The industry certainly wants us to do more of these office hysteroscopies. And you do see how some practices have made hysteroscopy at the point of care in the office a routine part of their workup. For abnormal bleeding there is a larger equipment cost that probably gets probably mostlyed to insurance, but to some degree I'm sure the patients absorb it as well and then also a larger physician fee associated with this. So the physicians will earn more for doing this. The question is what value does it offer? Our gardening gyne friend poses this question separately for the pre perimenopopausal bleeding patients as well as the post-menopausal patients.

Speaker 3:

Yeah, and that's a marketing term really, when you think about is it a cost to the patient or cost to the insurance, well, cost to the insurance is cost to the patient. It just gets funneled down and then it's a cost to the physician because you see lower reimbursement for your other stuff, because it cuts your reimbursement to pay for how much you're spending on in-office hysteroscopy.

Speaker 3:

It's a limited pool of money and the principle of medical justice dictates that we are thoughtful about this. Well, I think we have to start with our traditional workup, and how good is what we're doing and what are the limitations? So the traditional workup centers around ultrasound and or endometrial biopsy with the PIPEL, and we have to consider postmenopausal women separately, as our gardening friend suggested. We have to do that different from the pre and postmenopausal women. So the utility of a PIPEL endometrial biopsy, which was first introduced in 1984, is much more effective in detecting endometrial cancer in postmenopausal women with bleeding than it is in premenopausal or permenopausal women with bleeding than it is in premenopausal or perimenopausal women. Of course that makes sense because the pretest probability for cancer is higher in the postmenopausal patients. About 90% of women who have endometrial cancer will present with postmenopausal bleeding. But of all women presenting with abnormal bleeding in the postmenopausal period, but of all women presenting with abnormal bleeding in the postmenopausal period, only a small percentage of those women actually have endometrial cancer or hyperplasia. It's less than 5% for the two pathologies added together.

Speaker 3:

In the year 2000, there was a landmark meta-analysis published of 39 studies, over 7,900 women. Women that reported a 99.6% rate of detection for endometrial cancer among postmenopausal women, and that's the number that we frequently cite from that meta-analysis. So of course this is a blind biopsy. It's not perfect and it can be performed inadequately too. You may not do it right, you may not get into the cavity or you may not take sufficient passes. The number of passes, the rotation of the device while you're doing that around the cavity, that increases your sensitivity. So it's possible that in the real life numbers are less accurate than the numbers that you see from carefully playing clinical trials, where everybody is instructed to do it the right way and that sort of thing. And it's certainly the case that you can miss a cancer, particularly if that cancer is located in a polyp.

Speaker 2:

Okay, so what about premenopausal women?

Speaker 3:

Well, from that same meta-analysis the detection rate in premenopausal women was only 91% for endometrial cancer.

Speaker 3:

Now there are newer studies that show a lower detection rate for both of these populations than 91 and 99.6%. And this gets back to what you talked about earlier. It depends on whether or not you use an intent to treat statistical model or a per protocol model. So that 2000 meta-analysis it looked at studies that were per protocol. It assumed that we performed the test technically correct and that when we got the pathology back they were adequate, that we sampled the cavity, that we had some uterine or endometrial pathology on it, to know that we actually sampled correctly. So they in those data they might exclude, for example, someone who had inadequate pathology on it and obviously you didn't detect the cancer if he didn't detect proliferative endometrium or something. Newer studies have interpreted the same information but instead of using an intent per protocol, have used an intent to treat model and so now you see often reported lower sensitivity and lower detection rates and of course those are the studies that those product reps that you mentioned at the meetings are citing to people.

Speaker 2:

Well, it makes sense. But couldn't you also just say that if you had an inadequate biopsy or you know you couldn't perform it for some reason, you couldn't get in the cavity or they couldn't tolerate it then you just move on to hysteroscopy. But if your biopsy is adequate, the results show an endometrial pathology, then that 2000 intent to treat data does apply to your patient.

Speaker 3:

Or per protocol data.

Speaker 3:

Yeah right, I think that's exactly right, and this is why we're trying to be thoughtful about is this intent to treat or per protocol when we talk about these studies, and what does that mean for how we use it? If you're marketing in-office hysteroscopy products, you're going to use the newer studies without really explaining all of that. We just said to make endometrial biopsy with Papil seem pretty bad. But yes, we can and we do use an algorithm that would ask us to take action for insufficient sampling, as well as for persistent bleeding or persistent symptoms after the sampling has been performed, even if it was normal.

Speaker 2:

All right. Well, do these numbers also apply to the pre-cancers, the hyperplasia results, because that's going to be more common than the actual cancer?

Speaker 3:

Right? Well, the PIPEL doesn't work nearly as well for hyperplasia. So even in the 2000 study the sensitivity for atypical hyperplasia was only 81%. There are definitely studies out there that show better sensitivity than that number and that probably reflects the pretest probability of the patient population being studied in a given study. If you're only doing really high risk patients, you're going to see better detection and if you're doing a much broader population with a lower pretest probability you'll see a lower number. But it also reflects this important fact Endometrial cancers tend to be, at the time of diagnosis, more diffuse, whereas hyperplasias are often very focal or even limited to a polyp, for example. So we think that the pipel only samples about 4% of the cavity. If you could go inside and spray paint it and then look at it after you've done it, 96% of it would be unsampled by your pipel. So cancer cells, since they are typically more diffuse, they're going to be more likely detected with a pipel than these often very focal hyperplastic cells.

Speaker 2:

Yeah Well, and I think that fact that you're only sampling blindly 4% of the cavity is a big argument for doing these visually directed biopsies, because then you can see the lesion or the polyp and you can sample that directly. But if we're reasonably excluding cancer, then we're not as worried about also catching those focal precancers. The patient with some kind of hyperplasia is likely going to have persistent symptoms anyway. So even if we come back with an adequate negative biopsy, if they keep having abnormal bleeding, the plan already is that we're going to follow it with additional actions, and that might be going to hysteroscopy, that might even be medical treatment or doing a hysterectomy for them. Sooner or later we're either going to catch it or we're going to suppress it with progestins before it becomes cancer.

Speaker 3:

I think that's exactly right. So we've pretty much worked out the algorithm that I think most of us used to evaluate patients with risk factors and if their symptoms resolve or are managed with medical therapy in that premenopausal population, then we should feel pretty good. And if their symptoms are persistent or the biopsy is inadequate, then yeah, they need hysteroscopy or they need the next step that might be clinically appropriate for them. So that's what we already do and that's the evidence basis for that treatment algorithm. The question is are we better off skipping the blind biopsy and just instrumenting every single patient with hysteroscopy, even if it's less invasive, in the form of an office procedure versus a hospital procedure? I say, should we just do everybody? And this is like the Keflex and Flagyl thing Do we give 100 people Keflex and Flagyl to save seven women having to need it later, if the ultimate outcome is no different?

Speaker 2:

having to need it later if the ultimate outcome is no different.

Speaker 2:

Well, we can definitely say hysteroscopic biopsies are the gold standard for sampling and assessing for endometrial lesions and they are the reference to which other tests like ultrasound or blind biopsies are compared. But just because it's the gold standard doesn't mean we apply it as the first step to every single patient. It's actually rare that we would use the gold standard of any kind of test in medicine as the initial test. We almost always employ some other type of screening test and then, based on the aftermath of that screening, we would then move on to the gold standard in selected cases. So the companies that are selling in-office hysteroscopy are always going to be able to say that their test is better than the current standard. But it doesn't necessarily mean that we should use it universally. Really, I'm thinking a bigger utility for it is that we can use it as a next step instead of going to hysteroscopy under sedation in the OR, for the right patients at least. So maybe it could save OR time and equipment, but it shouldn't take away from using the PIPEL.

Speaker 3:

Yeah, this conversation reminds me of an editorial in the Green Journal that we talked about I think it was Green Journal in the last couple of years about the best test isn't always the right test or the best procedure, and I think that was in reference to using hysteroscopic resection for the products of conception after missed abortion. And, yeah, it does a better job, but that doesn't mean the patients that we employ that universally. So when you say that hysteroscopic biopsy is a gold standard, that traditionally has implied the operating room, the patient's asleep, you've got the full setup. You're actually taking these biopsies under direct visualization through an operative channel and then you're doing a thorough curatage of the uterus. And so one thing is these office scopes. They're a bit more limited and it's not always the same thing. Some office hysteroscopies are basically the full setup with the paracervical block. Many are not.

Speaker 3:

What's being marketed is a flexible hysteroscope without an operative channel, anything like that. So you have to be sure that we're making an apples to apples comparison and also that's not what people are necessarily doing. When you're saying taking the directly visualized biopsy, again, depends on what you're talking about. Not everybody's doing that. So I still say the burden is on the company that sells these products to show that somehow it ultimately leads to better patient outcomes, and there is no study I'm aware of currently that shows that it leads to better outcomes when implemented in this universal way. It just leads to higher cost.

Speaker 2:

That's fair, and people need to remember that none of these tests are perfect, not even the hysteroscopy and the OR, certainly not the office hysteroscopy. There's a 20-year retrospective study of over 2,600 patients who had office hysteroscopy and that only showed a sensitivity of 87% for endometrial cancer and 75% for hyperplasia. So those numbers really don't look much better than blind biopsy. So what are we gaining there? And on the flip side, we know that inpatient hysteroscopy still will not catch 100% of those lesions either. I think I've heard it quoted you sample 4% with the blind pipelle, but you sample maybe 50% at best, even when you think you're doing a global DNC. So we've got to know the numbers and realize nothing's 100%.

Speaker 3:

Yeah, and when you look at studies that look at, like you said, these gold standard numbers, remember these are being done by experts. Whose job is they do? You know hundreds of these a year and they are intimately familiar with what gross pathology looks like on hystereoscopy not by the average OBGYN, and that matters. That's not an insult to the average OBGYN or anything at all, but studies like that are going to be in the best hands. That's the numbers you get.

Speaker 2:

If the argument is that every single patient with abnormal bleeding, regardless of their menopausal status, should undergo in-office hysteroscopy, that's purely the industry talking and there's no science behind that claim at all.

Speaker 3:

Right, and we haven't so far considered their utility of ultrasound as a complementary tool to endometrial biopsy. We know that ultrasound, just like everything else, isn't perfect and it can't always discriminate between focal lesions, from just generalized thickening, or focal lesions in a background of generalized thickening. And obviously, if the lining is thin, well, it has no utility for detecting the rare thin cancerous lesions. Either right, they're going to be missed. Utility for detecting the rare thin cancerous lesions either right, they're going to be missed.

Speaker 3:

But nonetheless, ultrasound is probably the first-line diagnostic tool for abnormal uterine bleeding rather than biopsy.

Speaker 3:

Postmenopausal patients, in particular, with an endometrial thickness of less than or equal to four millimeters have a less than one percent risk of endometrial cancer. So that's comparable and maybe even a little bit better than an endometrial biopsy. Plus, the ultrasound may show you a focal thickening, a non-global thickening that could indicate a polyp or a cancer, or fibroids or some other anatomic abnormality that may be leading to the abnormal bleeding that you're investigating, which a blind biopsy wouldn't show you Even a cervical scar bleeding or something like that that you're going to find, or fibroids certainly. And then you might be cued by some of that information to skip your in-office workup or biopsy and go straight to the OR for appropriate DNC or hysteroscopy and things like that. So ultrasound gives you more information and it gives you information at least as good, if not better, than the pipe pill. And the same rule applies again where ultrasound, if it's reassuring but the abnormal pattern continues, well then you still go on and you do your biopsy, maybe you even do hysteroscopy because you haven't found the cause of it.

Speaker 2:

And we should also point out that there's some newer studies showing that some of these tests don't work as well in black women. There was an article in JAMA Oncology June 2024, looking at ultrasounds of a cohort of black women who ultimately underwent hysterectomy, and among those who were pathologically confirmed to have endometrial cancer, almost 10% of them had a less than four millimeter stripe on ultrasound, and it's unclear why this is, but it's possible. Maybe if they had intramural fibroids, that might have obscured the ability of the ultrasound to measure the thickness accurately. And it's also known that Black women tend to have higher risk of those more rare, high-grade subtypes of cancer, like papillary serous carcinomas, which don't present with the thickened endometrium, and so the ultrasounds miss those as well.

Speaker 3:

Right, and what I would say to that is if you're not confident in your ability to measure that stripe on the ultrasound, then you should perform a biopsy. You should not rely on an image you can't interpret. And even if your ultrasound and your biopsy are normal and the patient has persistence of symptoms, you should go on to the next step, meaning biopsy or hysteroscopic directed biopsies. Even the women in that study all eventually underwent a hysterectomy, I presume, because despite their normal ultrasounds, they had persistence of symptoms or they, for whatever reason, they ended up with a biopsy or hysteroscopy because of, presumably, persistence of symptoms. So it's a stepwise diagnostic approach.

Speaker 2:

And we need to always remember the pretest probability when we use these tests. 120 pound white woman who's 50 years old, with two days of spotting, no other symptoms, does not have the same pre-test probability for endometrial cancer as, let's say, an obese 76 year old who's been bleeding constantly for six weeks. And white women in general don't have the same risks for the, even for the same types of endometrial cancer as Black women. We all know the classic risk factors for classically, how I was taught was the type 1 endometrial cancer that does have a lot of environmental contributors. So we should consider these when we choose our tests and for that matter, we should also think about their risk factors when we do get an abnormal result, let's say an EIN or atypical hyperplasia. Because if they have a ton of risk factors for cancer, those would probably be more prudent to send straight to the oncologist rather than just go for a simple hysterectomy because they're more likely to have underlying cancer.

Speaker 2:

But in the initial workup it's very reasonable to triage patients with a transvaginal ultrasound. Look for an endometrial thickness under four millimeters. Then if their symptoms resolve, stop there, you're done. If it's more than four millimeters, or if they have persistent symptoms, then you do the blind office biopsy and then, if they still persist with their symptoms beyond that, even with a normal pathology, then go on to the next step, do hysteroscopy with directed biopsies or maybe discuss hysterectomy with them if that's what they're really interested in. Perhaps for Black women or other patients who you feel have that higher pretest probability of cancer based on their symptoms, you should start earlier with the biopsy and ultrasound as your first step and then move up that algorithm from there. So thank you to our friend for the question and just keep them coming.

Speaker 3:

Yeah, great question, and something that we all have to work through every day. Just keep them coming. Yeah, great question, and something that we all have to work through every day, and hopefully there are technological improvements over time, but we have to avoid falling into that industry trap where we can. So, okay, well, in two weeks we're going to have Jacqueline Vidosch back on to talk about her son, noah, who has trisomy 18. And so that's going to be a really good episode, and if you're listening to this as it comes out, then last week in the New York Times there's a very long form, well-written piece about her. So go ahead and read that if you haven't, and she'll be on the podcast in two weeks.

Speaker 2:

All right, looking forward to it, we'll see you then.

Speaker 1:

Thanks for listening. Be sure to check out thinkingaboutobgyncom for more information and be sure to follow us on Instagram. We'll be back in two weeks.