In this presentation, we will discuss different approaches to prediction of pregnancy complications, including discovery of exosomal biomarkers.
Dr. Louise Laurent, is a Professor and the Vice-Chair for Translational Research in the Department of Obstetrics, Gynecology, and Reproductive Sciences at the University of California, San Diego. She received her medical and doctoral degrees from the University of California, San Francisco, and completed her residency in Obstetrics and Gynecology and her fellowship in Maternal Fetal Medicine at the University of California, San Diego. Her research focuses on using stem cell and genomic approaches to understand early embryo development and normal and complicated pregnancy.
Dr. Carmichael has disclosed that she does not have any real or perceived conflicts of interest in making this presentation.
Dr. Sarah Holdt Somer: Dr. Suzan Carmichael is a perinatal and nutritional epidemiologist in the departments of Pediatrics and Obstetrics and Gynecology at the Stanford University School of Medicine. She holds a PhD in Epidemiology from the University of California, Berkeley. Her research focuses on finding ways to improve maternal and infant health. Exposure themes include nutrition, social context, care, environmental contaminants and genetics. Outcomes include severe maternal morbidity, stillbirth, birth defects, and preterm delivery. She is particularly interested in understanding the intersection of these varied types of exposures and outcomes and how they interact to impact health and health disparities, for the mother-baby dyad, in domestic as well as global health settings. Dr. Carmichael’s presentation is titled, Maternal Health Before and After Stillbirth. Thank you.
Dr. Suzan Carmichael: Hi. Thank you so much for the invitation to be part of this important conference and to share my work with you. I really appreciate being included here. The title of my talk is maternal health before and after stillbirth. Basically, first, I’m going to talk about stillbirth prevalence and trends, make some important points there, and then talk to you about maternal health factors that increase risk of stillbirth, and then talk about stillbirth and its impact on maternal health, and interspersed in there give you some idea of some of the challenges we face doing research on this topic.
For the purpose of this talk, the definition of stillbirth is an infant who was born at 20 or more weeks gestation, and who died in utero. This is partly dictated by the fact that we have fetal death certificates starting at this gestational age.
The first point I want to make is that stillbirth is now as common as infant death. This figure shows you the trends over time in both infant mortality and fetal death or stillbirth and how that has been going down, and it’s been going down more rapidly for infant mortality than stillbirth. Now those lines, if we project them out to current years, which I couldn’t find as good a graph of online or in the literature. They meet or actually infant mortality is a little bit below stillbirth. Currently about six per thousand births are stillborn.
I also want to make the point that the decline that has occurred for stillbirth is primarily among those that occur, as we say, “later”, which is defined as 28 or more weeks gestation. The earlier stillbirths, 20 to 27 weeks, have not really been declining.
Then this slide gives you a visual on the very different gestational age distribution of stillbirths and live births. Most live births are at term or 37 or more weeks gestation. For stillbirth, the curve looks very different, weighted much more heavily towards the earlier births, and a majority of stillbirths occur then.
Bringing those pieces together, they’re illustrated in this graph, which shows that of all births, when we’re looking at the earliest births, like the 20 to 25-week period, most births are stillborn. For example, at 20 to 25 weeks here, gestational age on the X axis, if live births and stillbirths are the denominator, we see that 70 to 80% of all births are stillborn at that time. The two different colors are just different for males and females.
I want to make these particular points about “periviable” births, which is considered the gestational span of 20 to 25 weeks because that’s when viability outside the womb is really just starting. There’s been a lot of interest in these births in particular because when they’re liveborn, they have very low survival and the highest risk of more morbidities, especially the most severe morbidities that have long-term consequences for health.
Just to bring this home, the importance of these, NIH has even made an announcement to encourage studies of understanding care and causes of these births at periviable gestation.
I just want to make the point that there’s a lot of interest there. However, given what I showed you in the prior slides, I was making the point that really to understand these births, we need to include stillbirths.
Unfortunately, they’re usually excluded and they aren’t even really discussed in this NIH announcement. That was really the point I wanted to make there that if we’re going to understand early delivery, we need to understand stillbirths. What we do instead is typically in the literature there’s silos, and there’s the stillbirth literature, and there’s the preterm birth literature, but I would argue that we really need to have more coherence across those inquiries, and we will learn a lot more.
Now I want to talk a little bit about maternal risk factors for stillbirth. Again, focusing more narrowly on our work, because I’m sure that, this audience, all of you have somewhat of a feel for this already. A study that I led really wanted to make this point. I was concerned about the points that I’ve just made to you. What we did was we did an analysis that compared risk factors for stillbirth and live birth at periviable gestational age, or that 20 to 25 weeks.
This slide shows you the risk factors that we examined. You see the risk factors in the left-hand column, and then you see what the risk was for among fetal deaths or the association with fetal death, and the comparison group for all of these is live births born at term. Then the second column is looking at the risk factors associated with having a live birth who died in the first week of life or a live birth who survived at least a week.
For example, in epidemiology, we talk in odds ratios and relative risks. To explain it briefly, you can think of it as 1.0 is null. Meaning, there’s no difference. It’s one-to-one if you compare two things or two groups. Increased means increased risk. Or above one means increased risk. Here, our 1.03 would mean there’s a 3% increased risk. Up here, for example, for maternal age, 1.03 means for every increase in age of one year, you have a 3% higher risk of having a stillbirth.
The point made in this paper is that these risks are very similar across these three groups. Really that was the point that we were making is to compare really the fetal death risks with those of having a live birth at those early gestational ages. As you can see, they’re much more similar than different. Here again, smoking, very similar. Pre-pregnancy hypertension, diabetes, a little bit more variability there with higher risk being seen for fetal death.
Basically, in conclusion, inclusion of stillborn and liveborn babies in studies of deliveries at periviable gestations is important. The paper brought that home, I hope. In order to understand preterm live birth, we need to understand stillbirth.
I’m going to show one more study, which is not for my team, but it’s very recent from the CDC. I thought it might be of interest to highlight. This is from the CDC’s maternal morbidity and mortality weekly report looking at racial and ethnic disparities. Just to bring that point home, this graphic shows that there are persistent racial/ethnic disparities in stillbirth. Basically, what it’s showing you top left is a heat map, so to speak, of fetal stillbirth rates across states, and then showing it by race/ethnicity. The higher the rate within that group, the darker and more solid the colorings is. You see that among Black babies. Black women are much more likely to have stillbirth.
The point to make here is really I think we need more research on stillbirth to understand social determinants, which could include anything from social disadvantage, to built environment, to inequities in the healthcare system, such as access and quality of care, and really need to understand how those are impacting and driving these disparities and the pathways by which that happens by which they lead to physiological vulnerability. I think that stillbirth is very understudied in this respect, and hopefully we’ll be seeing more of that research in the future.
Then one more graphic from this paper is showing that these disparities persist regardless of the cause of death of the stillbirth. Also wanted to point out that from fetal death certificates, we get some good information, but there’s still a lot that have an unspecified cause. So it’s not perfect, but it does give us some useful information. Then the final point I wanted to make from this slide, if you look at these different types of causes, most of them and where we see the greatest disparities are those that are related to maternal health. A cause here, fetal affected by maternal conditions unrelated to pregnancy, but still maternal or the mother. Maternal complications, diabetes, and so forth.
Now I want to move on to talking a little bit about one study we did that looks at the co-occurrence of the same timing of maternal health at delivery and stillbirth. First, I want to make a little detour and talk to you about the data that all these studies that I’m presenting to you from our team derive from. They’re focused on California. Some of the reasons we’re using the California data are that it’s about half a million births per year, so it’s a large state. A lot of data. About one in eight US births is what that amounts to. The population is diverse in many ways. Geographically, racial/ethnically, socioeconomically, rural, urban, and so forth. Fortunately, we have some really rich, unique data sources here.
The one that I am really focusing on for the work I’m highlighting here are what we call the California OSHPD or Office of Statewide Health Planning and Development linked birth cohort files. They’re these huge files where they’ve linked– They take vital records as the hub of the wheel, and you start there and that would include live birth and fetal death certificates at 20 or more weeks gestation. Then link that with hospital discharge records for the mom and the infant during pregnancy through nine months postpartum. This is really key to understanding maternal health because the vital record is great for some infant outcomes, but it’s a very poor source of data to understand maternal health. This linkage really opens up the opportunity to look at maternal health.
These data also link birth to the same mom over time, so we’re able to look at recurrence risks. We’re doing some of that right now, but I’m not going to talk about that today. Also importantly, we’ve gotten permission to get access to the maternal residential address from the vital record. Well, it’s actually just the live births. It hasn’t been available, but it will in current years moving forward for stillbirth records as well. That allows us to say, “Where do women live?” Then we can characterize social determinants and neighborhood factors and things like that to understand disparities better and environmental exposures.
Why California and then why maternal health? Well, the last five or six years, I’ve really been focusing my own work on maternal health. Before that, I was really focused more on maternal health as a risk factor for infant outcomes such as preterm birth and congenital anomalies. In the United States, we really do not perform well on maternal health outcomes. Just thinking about maternal mortality in the United States, more women die from pregnancy and childbirth related causes than any other developed country.
These deaths are on the rise. If you just look at this graph here that I’ve included, it makes the obvious point that all these other higher income countries, it’s been on the decline, and the USA it’s been on the up-rise. Even if we correct for some difficulties we’ve had with data, maybe it’s flatter, but it’s certainly not decreasing and it’s still higher. Most of these deaths are considered preventable through better care, and disparities, again, are a really important problem being highest for Black and Indigenous American women.
Where we focused in our group is actually to study severe maternal morbidity, and I’ll give you the reasons for doing this. Death is obviously a sentinel event and a tragic event, but there are only about 700 deaths per year. It’s 700 too many, but it is hard to study when something is that rare. Severe maternal morbidity is about 50 to 100 times more common than maternal death. SMM precedes most maternal deaths, probably close to 90%, especially of those that occur in hospitals.
The way we measure severe maternal morbidity is it’s a composite of conditions that put a woman most at risk of dying. We use something called the CDC Index, which is a number of procedures and diagnostic codes from those hospital discharge records to come up with the identification of women. It’s got a lot of similarities to maternal death. It’s doubled in recent decades. It’s got the high disparities. Unfortunately, a high preventability and many shared common causes, which makes sense with maternal death. Some of the most common being severe hemorrhage, cardiovascular events, hypertensive disorders, and sepsis.
Just one slide very brief thinking about the literature on maternal health at delivery and its co-occurrence with stillbirth. There have been some studies, not a lot, but here’s my quick summary and a couple of references. Prior literature showing there’s more hemorrhage, infection, and peripartum hysterectomy, those are looked at one at a time and associated with the co-occurrence of stillbirth. These are all things that in their severe forms are part of that CDC index of severe maternal morbidity.
I want to highlight a study that a postdoc led in my lab, Elizabeth Wall-Wieler, who’s now on faculty at University of Manitoba, and just had a baby herself a couple years ago, her first. What we did was look at the prevalence and risk of severe maternal morbidity among delivery hospitalizations for stillbirths compared with live births. This work was in collaboration with all the people shown here and, especially, I want to note my collaborator, Alex Butwick, who is the senior author on this paper.
Basically, we use California births from 1999 to 2011, which is around 7 million births. Out of those, we have records on about 26,000 stillbirths, and we define SMM based on the CDC algorithm. Here we see the prevalence of severe maternal morbidity and its specific component indicators and what their prevalence is among stillbirths versus live births.
This is where you get to see what indicators go into this SMM index or composite. For example, that first line, we see that among stillbirths, just divide by a thousand, 1.5% of the moms had severe maternal morbidity. Whereas it was much more rare among the live births. I’m sorry, I misspoke. I was using the case number. The 578 means that there was almost 6% prevalence of SMM among the moms who had a stillborn baby, and it was more like 1%, 0.99, among the moms who had a live birth. Then you can see, again, these prevalences, how they vary and how they vary between stillbirths and live births.
Then this shows you we basically calculated a relative risk to put a number on these sort of ratios and that’s what’s shown here. Even after adjusting for a bunch of different factors, meaning holding them constant, we see at least a four-fold increased risks, so people down here. See it’s four to fivefold increased risk of SMM among women who have had a stillborn baby versus a liveborn baby. Then you can see that many of these indicators, we see these even greater risks. These things are more likely among women who have a stillbirth versus a live birth.
Then this is breaking it down, instead of by SMM indicator, it’s breaking down by what we know about cause of fetal death from the fetal death certificate. This is basically showing the SMM prevalence per 10,000 stillbirths. You can see which ones are most common. Most common being hypertensive disorders and placental conditions.
Then one more point we made in this paper was to look at the prevalence of SMM and compare it between live births and stillbirths, but look along the gestational age continuum. What this slide shows is that stillbirths are the gray line, and live births are black line. At these earliest gestational ages, there’s not really much of a difference in the risk of SMM among stillbirths and live births. Where it diverges and where we’re really seeing this higher risk being driven is among the stillbirths that occur later.
Now I want to move on to thinking about maternal health after stillbirth. The most work that’s been done is really effects on mental health, but in general, beyond that, it’s understudied and underappreciated. What we’ve done is a few studies looking at risk of postpartum hospital readmission. Again, partly because it’s important, it has public health significance, and it’s also what we can get from these data files that we have access to.
Hospital readmission has been in the public eye. It’s an important quality metric. It’s an important cost that is often avoidable. As far as postpartum readmissions, this is understudied, but it’s a very vulnerable time. Probably 1% of women who give birth in the US are readmitted in the first month after giving birth. In California, 1 in 12 women visit the ED within three months post-discharge after having a baby, and 40% of women don’t attend a postpartum checkup. It’s a really important time, and we’d like to understand it better.
Hold on. I just need to reset my timer. There we go. Just want to recognize that we did get some funding from NIH for a small grant for this work. Alex Butwick has been my fabulous colleague on this work, and he’s an OB anesthesiologist.
Elizabeth Wall-Wieler, again, she led this study and she looked at maternal health after stillbirth and looked at risk of postpartum readmission. Again, using these California data, we had about 30,000 stillbirths and she looked at readmission within six weeks of giving birth and looked at diagnosis and procedure codes associated with those readmissions. We grouped them into some clusters of related diagnostic and procedure codes to simplify the picture.
Then this figure shows you the cumulative percent of women who had a readmission from zero to 42 days and shows you the contrast between live birth and stillbirth and shows that it was much higher risk among stillbirths. Women who had a stillborn baby.
There’s a lot on this slide, but I’ll just give you a few highlights. This is the bulk of our results. This shows you both the rate of readmission among women who had a stillbirth or a live birth, and then showing the relative risk of readmission among a woman who had a stillbirth versus a live birth. To walk you through this, the first row shows you the overall results. Basically, we saw that readmission was– and you can divide these by a hundred. The rate was 2% among women who had a stillbirth versus 1% among those who had a live birth. That relative risk was about twofold higher or a hundred percent higher among the women who had a stillborn baby.
Then model 1 and 2, model 1 adjusts for some sociodemographic factors, things like age and parity. Thinking about, are those driving the results? Not very much. Those are very similar. Then we adjusted additionally for more maternal clinical factors and comorbidities. Then it was a more modest increased risk, but still about 50% increased risk. Or 47%.
Then we looked at this by specific diagnostic and procedure clusters, and those are the diagnosis and procedures that were listed as reasons or things that happened during the readmission. The three little arrows point out the top three causes for stillbirths and live births. A point here is that they shared the number one cause of infection, but the next couple of causes were different for live births and stillbirths. What you can also see is how these increased risks occurred for most of these more specific reasons for readmission paralleling what we saw overall. Then this is just showing what we adjusted for which I explained.
Basically, after adjustment for confounders, women who had a stillborn baby were at 50% higher risk of postpartum readmission within six weeks. We saw a little bit different distribution of what the most common diagnoses and procedures were. That is, reasons for readmission. This really led us to conclude we need closer postpartum follow up for these women.
Then to follow this, Julia D. Tosto, who’s a master’s group in our lab, and she went the next step to think about, just among the stillbirths, what were the risk factors for readmission among women. She looked at readmission within six weeks, but she also looked at later readmission from six weeks to nine months. It says 42 weeks. It was nine months.
Then this is again using similar data and this shows you the cumulative percent of postpartum readmission within nine months. I showed you this earlier. It was just really for that first 42 days, and this shows you what it looks like when it’s extended out. By nine months, about 6% of women were readmitted. There’s a lot on this slide, but again, this is showing for a comparison of women who were– it’s all women who had a stillbirth and it’s comparing the characteristics of women who got readmitted with those who did not get readmitted.
Here, for example, race/ethnicity, walking you through it and showing the unadjusted and adjusted odds ratio. For example, Black women were 40% more likely to have a stillbirth. I’m sorry. 40% more likely to be readmitted than non-Hispanic white women. Another example, women with lower education were more likely to be readmitted. You can see the other more clinical factors over here. Women who had SMM, for example, at birth were two to threefold times more likely to be readmitted.
I’m not showing you the results for the sake of time here, but we saw a similar pattern of results for the later time period. Basically, we saw that a lot of factors were associated with increased risk of readmission just looking at stillbirths because they hadn’t really been looked at carefully before. Most of these readmission studies focus on live births. It gives us some idea of who’s most at risk and that more attention needs to be paid in postpartum care to these risks.
My final slide is just to– hopefully, I’ve given you an appreciation for what’s being done related to maternal health and stillbirth from this population based big data epidemiologic perspective. My take home, maternal health matters in so many ways, and I really appreciate your attention and for listening. I want to thank the funder, NIH, for supporting this work and also all of my fabulous collaborators who I appreciate so much. And end with a final quote, “We make the road by walking.” It is a journey. I appreciate you listening today, and I look forward to your comments. Thank you.
This presentation was part of the Stillbirth Summit 2021. This individual lecture will be awarded .75 hours of continuing education credit to include viewing the lecture and completing evaluation and post-test. Once received a certificate will be emailed to the address you provide in the post-test. If you did not register for the Summit WITH Continuing Education, you can purchase the continuing education by clicking here. This purchase will provide you access to all Stillbirth Summit 2021 lectures including continuing education credit. There is no charge for viewing the presentation.
Please feel free to ask questions of the presenter. We will obtain their answers/comments and provide them here as received.