Ninth Annual Maternal and Child Health Epidemiology Conference / December 10-12, 2003
AMY CASE: Prospective parents today can find themselves facing decisions that would have made Solomon’s head spin. Given the ubiquity of prenatal diagnosis, technology of perinataly diagnosing birth defects, parents may find themselves deciding whether to continue a pregnancy or indeed wondering if there’s any reason at all to prepare for a live-born child. While those of us who work in birth defects surveillance spend a great deal of effort trying to find an answer to these parents’ first question, which is usually, “How did this happen,” the data I am presenting today is an attempt to help health care providers answer the often subsequent questions, parents asking, “Will my baby survive,” and “What have other families done in this situation?” I’d like to point out that Mary Ethen is the primary author and analyst of this data. She also works at the Texas Birth Defects Monitoring Division. And although we worked closely on this information, I will refer you to her for some of the finer analytical points if you have questions on those. I’m going to move swiftly through the introductory and methods slides in order to give us more time to look at the results. We can always go back and revisit the methods at the end as time and interest allows.
The data used in this study came from the Texas Birth Defects Registry, which is an active surveillance system. We have about 50 field staff who visit medical facilities to collect information directly from the medical record. And the system is statewide, covering more than 365,000 live births annually. These records end up in a computerized database of all infants and pregnancies affected by the birth defects, and these birth defects are usually those--our emphasis is really on structural malformations. All of our cases are matched to a vital record for additional information on both outcomes and demographics. In order to be in the registry, the cases must meet these criteria. Of course, a mother must reside in Texas at delivery, although that’s not as straightforward a question as you might think, especially in certain areas of our state. We do include--and that’s important for this presentation--live births, fetal deaths at any gestational age, and elective terminations at any gestational age. So I’ll be showing you information on 20 weeks, pre-20 weeks, and post-20 weeks, basically. The infant and fetus must have one or more of the structural chromosomal birth defects or Fetal Alcohol Syndrome, and the diagnosis must be made within the first year after delivery or within the first six years for Fetal Alcohol Syndrome.
There were 23,455 individual--that is, non-isolated--birth defects among the deliveries in the Texas Birth Defects Registry during the years 1999 to 2000. For this analysis today, we used the defects that fell into 49 standardized major defect groupings, and that encompassed about two-thirds of all of the defects in the registry. Gestational age in this study was calculated from the medical record first using last menstrual period, and if that was unavailable or inconsistent, by the outcome of the neonatal examination, the findings. It’s important to note that as I present this information, that sometimes we’re counting deliveries, that is number of infants or fetuses in the registry with any monitored defect; and sometimes, we’re looking at the number of defects that occurred. About half of the deliveries, or fetuses, children, whatever you want to--whatever terminology you use in the registry have more than one defect. So when we’re looking at infants and fetuses with any monitored defect, we’re counting cases of, you know, actual children. When we’re looking at defects, we’re counting all of the defects that occurred, and some of those children may be represented many times. Pregnancy outcome for our purposes today is based on information abstracted from the medical records. The registry records, as I mentioned, we match to the appropriate vital records and found consistency with the appropriate vital records in 97 percent of the cases.
These next three slides will show the criteria that we use in our registry for what is a live birth, what is a fetal death, what is an elective termination? Determining those outcomes is not a very straightforward process from the medical records, as you might imagine; but we do have very specific criteria. And I’m not going to go and read those to you, but just on each of these slides, you’ll see the definition on the left-hand side, and on your right-hand side, you’ll see a listing of the defects that were overwhelmingly represented in that outcome. For example, in this list of defects, 100 percent of those defects ended up in a live birth: cataract, anaridia, *Hershsprung disease, *baileria *tresia, bladder *axtrophy, congenital hip dislocation, and FAS. I’d like to just let you know right now that I know many of you may not be familiar with all of the birth defect terms I’ll be throwing around here. So if there’s a particular defect you really want to know quickly what the definition of that is, just flag me down and I’ll stop and give you a quick idea of what type of defect that is. Here’s the definition in our registry of what a fetal death is, and here are examples of defects. And I’ll be looking at these in great detail a little bit later in the presentation, but these particular defects or defect groups ended up in in utero death quite a bit more often than some of the other defects or defect groups. No surprise there, anencephaly and some of the *trisemis that tend to be more lethal, even after delivery of a live birth. Likewise, induced termination. And this was the outcome in a little over two percent of the cases that we’re describing here today.
As I mentioned, we do collect information on pre-20 weeks and 20-weeks plus, and there is a very small percentage that we called “unspecified” because you really could not tell from the medical record whether it was a fetal death or a medical termination of the pregnancy. When I first looked over the results of this analysis, the way my mind works is “Well, why? Why some defects have this outcome and others do not? What are the external or subjective factors for the mother that determine whether this particular type of defect ends up as a fetal death more often than a live birth?” and so on. Some of the factors that influence pregnancy outcome that I found through a *literature search are the mother’s health factors. Did she experience any type of infection during pregnancy? Was she being treated for a chronic condition like lupus?
Prenatal diagnosis is a whole complex area, how likely the particular defect is to be diagnosed *prenatally will obviously have an effect on whether or not the parents choose to terminate based on the type of defect found. Also, whether prenatal diagnosis service are available to a particular woman or group of women. And then, of course, a consideration especially in the termination of a pregnancy is the prognosis for the child. How lethal is the defect? Will my baby live? Even if I continue this pregnancy, will it have a life? How impaired will my child be? Will they have life-long health problems? And also, are there co-occurring defects? A seemingly minor defect might often occur with something that’s quite serious, and so that seemingly minor defect will have a fairly high rate of, for example, fetal death. As I present this information, I hope that you will jot down for yourself other hypotheses that you might have, factors that would influence pregnancy outcome. Okay. The charts.
This chart is one that looks only at deliveries rather than defects or defect categories. In other words, these are all the infants and fetuses in the registry for 1999 and 2000 that had any monitored defect, one or more. Overwhelmingly, about 96 percent of our cases ended in a live birth. And the percentages of the others are almost too small to really compare here, so later on, I kind of tease those out and leave the live births off. Looking at these in specific defect groups, here we have a group, three different neural tube defects with very distinct patterns of outcomes. As has been mentioned earlier in this session, anencephaly is a defect in which the skull fails to form and the brain is not completely formed, and it is always fatal, even if the pregnancy does result in a live birth, which as you can see, only happens a little less than a third of the time in these cases in Texas. And you can also see that there is a relatively high rate of induced termination, both early and late in pregnancy. This is probably not real surprising given if you’ve had any experience with neural tube defects, with parents that have had an anencephalic pregnancy. On the other hand, spina bifida has a very different pattern, many of these surviving to birth. About 83 percent of them in this group resulted in a live birth. And again, this is not too surprising. Surviving children have a near-normal life span. And though they will have some level, rather--either slight or grave level of physical disability and probably or possibly some mental retardation, the outlook is not as grim for them and there are lower levels of induced termination, but also a very low level of fetal death compared to the other neural tube defects.
And finally, *enfaloseal, which is a very rare NTD that involves herniation of the brain tissue through the back of the skull. It exhibits another completely different pattern. Here, we’re comparing three different *tricimede, or chromosomal, defects. *Tricimede 21 many of you probably know most as Down Syndrome. And I think most of the people that have seen this chart, their first comment is that they’re surprised that that many babies with Down Syndrome survive to a live birth. That’s right at 90 percent. By contract, *tricimede 13, which is often referred to as *Pata Syndrome; and *tricimede 18, which is known as *Edward Syndrome, have much less optimistic outcomes. These children have many other associated birth defects often and very often do not survive past the first year. SO these results would be reflective of that, probably. Here are two abdominal wall defects that might seem similar in that they both involve development of the intestine and some of the intestinal organs--or the abdominal organs outside of the abdominal cavity during fetal development. However, *enfaloseal is one where the organs develop through the umbilical area. And again, this is often seen in combination with many other defects and tends to be a very serious outcome for the child. By contract, *excroskeesus has a better outlook and can usually be repaired soon after birth. And we see that they have a much higher rate of live birth.
We chose these two groups of defects to look at together because they’re often found together. There’s about a 25 percent overlap where babies with a renal agenesis or *disgenesis also have lung development problems and vice versa, and not too surprisingly, their outcome patterns in pregnancy are also similar, although the similarity of those outcomes probably can’t be entirely explained by the fact that they often co-occur. One thing that makes these particular categories--and I guess I’ll leap to the limitations at the end of this presentation--is that the groupings that we use for standard publication of birth defect rates have lumped together a large spectrum of outcomes of birth defects. For example, the category renal agenesis and disgenesis will include in there a baby who has a slightly malformed left kidney that was just discovered because they were screening for something else and will never have any *sequi from that malformed kidney all the way to the fetus who does not really form any kidneys at all, and that is always lethal. So there’s a blending of different types of defects for these purposes there.
This is the one that intrigued me because although I would not want to be born without any portion of my limbs, and I certainly would have been upset if any of my pregnancies had been diagnosed with a limb defect, I didn’t view it as being either something that was uniformly terminated or something that would have a high fetal death rate. But you can see that about 10 percent of these pregnancies died in utero. And if you remember back to the first slide I showed you, when you look at all congenital defects, that’s like one or two percent died in utero. And so how to explain that? Perhaps in the discussion afterward, someone will have some ideas. I did discover that reduction defects of the lower limbs are sometimes associated with monozygotic twinning, and so maybe it’s something in carrying a twin pregnancy that causes early fetal death, and also so then this would show up with the lower limbs. And likewise, there’s a--what’s that--about a little over seven percent induced termination. That seemed a little higher than maybe some of the other defect groups. Well, then we wanted to look at whether there were differences in pregnancy outcome by maternal race/ethnicity. So here, I’ve pulled out the live births so that we could see the small percentages of other outcomes and compare them more easily. And although we did not run tests for statistical significance for this data, these data for this presentation, there doesn’t appear to be an overwhelming pattern of different outcomes other than live birth between these three mutually exclusive groups. But when we looked at them by defect group, we did find some interesting patterns, although I’m not sure I can offer too many explanations or plausible explanations.
Here, we have the cases resulting in induced termination, and then within each of the defects, we’ve stratified by maternal race/ethnicity. And you can see one of the things that comes out fairly strongly is that Hispanic mothers tended not to select induced termination for neural tube defects, for anencephaly, spinal bifida, and *ensefaloseal. Is that a cultural influence? Is it because they tended to lose those pregnancies early? Is it because those services are not as available in some of the parts where more of the Hispanic moms live, parts of the state? These analyses really raise so many questions for us. We’re, like, “Well, next we’re going to have to look at this and this.” So you may have some hypotheses of your own. Interestingly, back to the reduction defects of the lower limbs, there’s a very distinctive pattern where the Black mothers were more likely to have their--if they had a child with a reduction defect of the lower limb, that it would end in fetal death. However, now we are getting into a problem of small numbers because that only represents--through the two years, I think that’s four--yeah, that’s only four births right there, that nearly 20 percent fetal death rate. Other than that, I don’t see anything that’s especially different. And I think a lot of us would have thought up front that there’d be more different patterns of outcomes for race/ethnicity. I’ve alluded to some of these limitations earlier, so I think we’re at a time--and I probably shouldn’t go into them in great detail. Three minutes? Okay.
But I wanted to point out the second point. It’s difficult to determine the intended outcome from medical records, and that is to say although perhaps we record something as a fetal death, we don’t know the entire pregnancy history and the history with the physician, whether that was intended to be a live birth and it was not or whether it was intended really to terminate the pregnancy and record it as such. So there are some subjective problems in the way that this is recorded in the first place. I would like to acknowledge the Texas Birth Defects Monitoring Division Surveillance staff, who go out every day into the hospitals and write all this information down; the Bureau of Vital Statistics that provides us with their data files so that we can match these and make sure that our records our consistent and valid; and Dr. *Canfield, our Director, who was invaluable in his advice and guidance in this project.