MCHB EPI Atlanta Conference
 
December 5 - 7, 2006

Genetic Epidemiology, Genetics, and MCH

SIOBHAN DOLAN: Well, thank you so much. It’s really lovely to be here this afternoon and join this distinguished panel and what I want to try to do here as I conclude this session is, kind of, marry a lot of the genetics and genomics principles really with the focus on genomics, with this very human face of preterm birth, and the statistics which I honestly think this is like bringing calls to Newcastle to show you all the slide because I know you really been very instrumental in producing these data, but this is really what we’re facing.

So the question is, how can we bring the innovations, the excitement about genetics and genomics? And the opportunity to improve health into maternal child health world, and we need to, sort of, do it right now because the need is certainly great. So--sort of, just run us quickly through risk factors but then try to move in to genomics principles, building on what Cindy Moore discussed, then talk about what we’re really thinking of preterm birth as a complex disorder means, and then really try to lay out an ambitious research agenda, and hopefully, we’ll have some time for questions.

So we look at risk factors, these really go—are very well established. We have a good understanding of, sort of, be able to describe, and from this, we have demographic and clinical standpoint who’s at risk for preterm birth. And the clinicians will take some of those risk factors and then think, sort of, path of physiologically. We have preterm labor, we have ruptured membranes, we have medical intervention. And so we can understand a lot about how you get from pregnancy to preterm birth, but really, the question remains why, why did these things occur? And most importantly, what can we do to try to interrupt some of these pathways? We take a lot of histories and clinical obstetrics practice. We sit down and we take a 10-page poppers form or whatever, clinical obstetrical intake form, and we note and we write all these risk factors down. And then, I actually still teach the residents like a couple of mornings a week and they say, “What are we going to do about this?” And so we still need a lot of clinical interventions. But with these risk factors, we do have a very good landscape of the condition of preterm birth, and we’ve really put a lot of effort on the ground level into issues like the prevention of tobacco use and alcohol use into folic acid deficiency; into understanding socio-economic factors and racial factors and working with them on a community level. But we still sometimes do have the quite question mark of what’s causing what and whom, and really then, how do we get out of this? Where do we go to bring down the rate of preterm birth?

Now since this is the genomic session, I’m going to bring us back and orient us to a risk factor that actually leads the list in terms of risk factors for preterm labor and delivery. So the best two predictors or the strongest two predictors are, the first, having a current multi-fetal gestations, so currently carrying twins, triplets or more. And there’s a lot of literature around that, and a lot of understanding of why that occurs and what we can do about it. But the other risk factor that’s very well established is a history of a preterm labor and delivery. A woman who’s had a previous preterm birth we know to be at risk in the current pregnancy. And I can actually remember a fellow telling me that when I was probably an intern and he was a fellow, and he told me that, “So these patients were coming to high risk clinic because they were at risk.” And I was like very naïve, “So what do we do for them?” And he’s like, “Well, we just follow them. We bring them in more often, right?” So this is established but what does this mean? How are we going to intervene? Well, I’m going to suggest that genetics and genomics can help us start to understand susceptibility and risks, and then hopefully help us target interventions. But to do that we really need to go back to the understanding of genetics as classically, the study of the patterns of inheritance of specific traits. Now that is, sort of, classically done in a clinical geneticist office with a formal pedigree taken, and we can look for patterns like autosomal recessive, autosomal dominant, et cetera. And I’m not suggesting that we’re going to find preterm birth following a single gene inheritance pattern. But the great example we have in this community of the efforts that have been made around neural tube defect, a multifactorial condition; a condition we understand to be very complex with both genetic and environmental influences is really the paradigm for how to approach preterm birth.

So what am I, kind of, trying to say with that? Well, I’m going to try to suggest that complex genetic traits are phenotypes or clinical presentations that don’t fit the classic Mendelian segregation or assortment patterns, so they’re not following single gene inheritance patterns. But they do, in fact, exhibit a preferential familial clustering and that cannot be claimed exclusively by cultural or environmental causes. So we know that preterm birth runs in families, we know that we have some high-risk families in our clinical settings. And I mean I would suggest, from a clinical standpoint, a lot of folks do think of that and frame that as shared environmental factors, and there's some of that I'm quite certain. But I think we can step back and think about now shared genetic susceptibility and how that interplayed with various environmental factors, really gives us the strong effect that we really be a place for targeted interventions. And I mean we try to actually demonstrate that with a couple of pieces of literature and I don’t want to belabor it but really try to just articulate that this is a very promising research direction.

So complex disorder--excuse me, how does--as I said, genetic contribution, environmental influences and gene-environment interactions. And I'm proposing to you that preterm birth really fits the bill here. And this is what I'll try to do in pointing out a few studies.

The first is the study on the risk of preterm birth across generations. In this study, a cohort in Utah was looked at and in intergeneration data set there was 405 preterm mothers, so the mother herself was born preterm 20 to 30 years ago when her birth occurred. And then there were 2,781 term mothers who they themselves were born term. And what the study did was went on to follow the gestational age of which they had their children. And if we look over here on the right, we can get the pointer, we can see here as a gestational weeks of the mothers own birth goes down, her odds ratio of having her child preterm goes up, and it's a really a doze response relationship. So we start to think, distinct from environmental factors, which probably don’t completely coincide over 30 years or so. There's something that's going from one generation to the next and most certainly that would make us think about genetics.

Another really classic way to look for genetic influence on condition is twin studies, because of course we look at monozygotic twins who share 100 percent of their genes and compare that to dizygotic twins who share about 50 percent of their genes. And if we see the condition more commonly occurring in both of the monozygotic twins than the dizygotic, we can start to infer that there's more of a genetic contribution to that condition.

So in this study that was done in Sweden, they looked at 868 monozygotic female twin pairs, and this is the mother’s and then they followed the mother’s birth outcomes. And then they compared that with 1,141 dizygotic female twin pairs, again, the mothers were twins and then their children were assessed. And using some statistical techniques on twin concordance, they suggested a heritability estimate in the range of 25 to 40 percent, preterm birth had a point estimate of 36 percent. Which suggest that the variation in preterm birth had about a 36 percent contribution from genetics, which is actually a really high number considering that lots of us think of preterm birth as an environmentally induced condition, and that a lot of our risk factor list don’t talk about family history or genetic susceptibility.

So I wish I had an hour's worth--well, I don’t actually, I wish I had an hour's worth of slides to show you because I couldn’t finish them, but I also want to point out that there's a limited literature to support this and I think we could do more and kind of adding to these population-based studies that continue to support the genetic contribution to preterm birth. But there's one interesting, one that just came out earlier this year looking at the paternal contribution to preterm birth and looking at a Norwegian study, a group of children, the gestational age of children at birth increased on average 0.58 days for each additional week in the father's gestational age, so there was a paternal contribution to gestational age of a child, and 1.22 days for each of additional week in the mother's gestational age. So you would rightly argue to me that if it's genetic, there's should be both the paternal and maternal contribution. And here's some early insights and studies demonstrating just that.

Now that's, as I said, limited but I think kind of convincing literature, or pieces of that literature to talk about a genetic contribution to this complex condition of preterm birth. Environmental influences, I'm going to hope that you'll trust me that they've been well studied. I know many of you in this audience have studied them, and that we know a lot about that smoking and infection and stress do in fact contribute to preterm birth, there's no, I think no doubt in that question. But now I want to try to take what we know about some of those environmental influences and sort of say, and I think it really answers the question. Remember when bacterial vaginosis came the--the literature hit the popular press that bacterial vaginosis was associated with preterm birth. And it was kind of exciting because people were like, “Uh, this is great, finally.” Like, it’s infection, and we know how to screen for that, we know how to treat it. And it seems really terrific except as the studies were replicated and as they went from high to lowest populations, things kind of fell apart. And so the question of what gene environment interactions, who are the women who are at risk for infection? And who are the women who when treated for bacterial vaginosis will experience decreased rates of preterm birth? That is a sort of more subtle question that doesn’t make it into the popular media but really still challenges us to think more deeply about this gene environment interactions.

So here’s a pretty classic one that we all know well: Cigarette smoking and birth weight and gestational age. And this is built on what Juan showed us very early in this set of lectures this afternoon, which is that we have the classic two by two table. Any table you set up with smoking and preterm, or smoking and birth weight is going to show an association. But in this study, the investigator looked at two enzymes CYP1A1 and GSTT1, which are in the liver and their site to compete for 50 enzymes. And they help our bodies break down the smoking by-products. So what we found or what the investigators found and then hypothesized upon was that if you break down the by-products of smoking more quickly, you could assume that your fetus would experience those by-products less--for less long period of time and might have less of an effect. And so when they went ahead and looked at singleton live births in smokers and non-smokers, they were able to set up as Juan pointed out not the two by two table but the two by eight table, which had smoking with a never or continuous status taking up eight rows because now we have the different variants of types of CYPY1A1 and GSTT1. So you could be the fast breaking down and slow breaking down, et cetera. And what we see was for all seven of these rows, which--the details don’t really matter at all, its just variants in these genes. For all seven we see that the gestational weeks of birth were non-significant. But here we see that if you were a continuous smoker and had this genotype in this enzyme and absent GSTT1, okay, so these are your variants. And there was 11 women in the group. But those women experienced the gestational weeks at birth that was five weeks shorter than the reference group. So here we have an example of a gene environment interaction. Is smoking associated with preterm birth? Yes. But is smoking have a stronger effect in the context of these two gene variants? Well, absolutely. And so this starts to let us target for whom is smoking cessation absolutely critical, for whom will we see the biggest benefit if we targeted smoking cessation? And I’m going to go move pass this because we don’t have time.

There’s a little bit of data to suggest similarly with bacterial vaginosis and TNF-alpha. That bacterial vaginosis affects a subgroup of women with a variant in TNF-alpha. And I’m happy to show anybody the slides that go through this, or point out the literature that demonstrates this. So this brings us into this idea of genomics. We can start to think about the structure and function of the entire genome and then interactions and we need human genome epidemiology as Dr. Moore pointed out in her first talk to help us decipher this. But I’m going to just put up this slide to suggest that we have the ability to do it now because we have a good sense of the path of physiology of preterm birth. We have stress and activation of the maternal fetal hypothalamic pituitary access. We have inflammation causing preterm birth. We have a lot of clotting abnormalities causing preterm birth. And then we have uterine distention causing preterm birth. But now we can take that another step and think about enzymatic variants that act along those pathways and think about studying not just the risk factor but who’s at risk. And now we have the technology to look at interlooping variants or to look at factor V Leiden variants and add that additional piece to our assessment of risk.

Sorry, I’m out of time so I’m just going to try to do one last point. Genomics is sort of the sense that we have a high tech approach in the future, like one day you’ll come in and you’ll have an Aphimetrics trip and we’ll put your blood on, we’ll tell you what you’re at risk for. We’re going to do that in general, right? We already know that that’s coming in a bigger picture. But we’ll probably do it for prenatal care, right? You come to first OB visit and then it’ll all be spit out. Until we get there, we have a much more low-tech trick or tool in front of us, which is family history. And I think that we have a good paradigm for the rule of family history in the prevention and interventions in the role of folic acid. So we already have a sense that if you’re at average risk, you would get the routine components of preconception and prenatal care but if you’re in increase risk for neural tube defect, through a bunch of criterion including having a previous affected child, we would give you routine care plus a targeted four milligram folic acid supplement. So we could take that paradigmatic approach. This is all I want to try to point out and say, “If your family history suggests you’re at average risk, then you get routine care. But if your family history puts you at a high risk of preterm birth, in other words, maybe you’ve--your sister had a preterm birth or you were born preterm yourself, we could already start to think about adding some targeted interventions like smoking cessation or weight reduction. And then furthermore, if you were found to be an extremely high risk like you’d previously had a preterm birth yourself, today we could consider progesterone supplementation for the prevention of recurrent preterm birth. So I think that this is just a way of thinking that genomics can add to and supplement but for now we do have some tools in our armamentarium.

I’m going to close there by just sort of a call to arms that we really can’t leave genomics just as the thing that’s going to help adult chronic disease and hypertension and diabetes and asthma. But really we do in MCH need to sort of get on the map and have the sense and I certainly do, that genomics is really going to help us address and bring in healthier babies into this world. So thank you for your attention.