#456 - [Journal Club] - 🫀 From The Heart - Should an Abnormal Brain MRI Change the Decision to Operate in Critical CHD?

In this systematic review and meta-analysis, Nim and Adrianne dig into just how common brain injury really is in neonates with critical congenital heart disease. Pooling 31 studies spanning two decades, the authors found that nearly 70% of these babies show some form of ischemic brain injury, split roughly between pre-operative and post-operative timing. Counter to what most clinicians would predict, kids who went to surgery earlier (days 4-6) had higher rates of white matter injury than those who waited longer. MRI remained the most sensitive tool for picking up these lesions, well ahead of ultrasound or CT. But the data stops well short of proving these findings predict long-term outcomes, and Nim pushes back on the idea that an abnormal scan alone should steer decisions about whether to operate.

----

Prevalence of Ischemic Brain Injury in Neonates With Congenital Heart Disease: A Systematic Review and Meta-Analysis.

Kim C, Chetan D, Kazazian V, Alzamil J, Chau V, Seed M, Miller SP, Selvanathan T.Neurology. 2026 Feb 10;106(3):e214569. doi: 10.1212/WNL.0000000000214569. Epub 2026 Jan 9.PMID: 41512205

Transcript

Nim Goldshtrom (00:01.954)

All right, Adrian, how are you? Long time no seeing.

Adrianne Rahde Bischoff

Hi Nim. I'm good. I'm happy that it's spring when we're recording this. I don't know what time people are gonna listen, but I'm glad that spring, at least on the calendar, is finally here.

Nim Goldshtrom

I couldn't agree with you more. Snow is wonderful in waves. I think for you guys in Iowa it's much longer waves, but yeah, spring is a great time for us at least.

Adrianne Rahde Bischoff

Yeah. And it's a good time to catch up again and to get started on this year's series. I think we're gonna try to do something slightly different — we'll continue to do some cases, we'll continue to do journal club, and hopefully we'll even get some interviews in the next few episodes.

Nim Goldshtrom

Yeah, whole new world for our series, for our audience, to learn from not just us, but hopefully invited speakers and other experts on the topics of cardiovascular physiology, congenital heart disease, targeted echo. It should be a great year ahead for all of us. And so we have a nice smattering of both sides of the spectrum, right? A little bit of congenital heart disease and a little bit of targeted echo and echo evidence for the neonates today, which should be great.

Nim Goldshtrom (01:12.14)

So I'm going to kick things off with two topics to kind of start the new year off right, for whatever our audience gets to hear this on: congenital heart disease. We're gonna hear about the landscape of congenital heart disease and ischemic brain injury, and then the landscape of congenital heart disease trends globally — or actually in the United States specifically — over the last 25 years. And so to kick us off, let's start with this new, recent systematic review and meta-analysis, the title of which is "Prevalence of Ischemic Brain Injury in Neonates with Congenital Heart Disease: A Systematic Review and Meta-Analysis," published in Neurology just at the beginning of this year. First author is Dr. C. Kim, and last author is Dr. Thiviya Selvanathan. Apologies for the pronunciation. For those uninitiated to the topic, again, in critical congenital heart disease, especially in neonates, there is very high risk for neurological injury and later development of neurodevelopmental impairments.

A lot of prior work has reported widely variable rates for ischemic brain injury, limiting clinicians' ability to counsel families appropriately, design studies, and really gives us a not-great understanding of what the real burden is to this population. We'll talk in a little bit about the baseline physiology that exists for these kids, but the authors sought to determine the overall prevalence — looking at a wide variety of studies — of ischemic brain injury in neonates with CHD (congenital heart disease).

Identify potential risk factors and give us better evidence of how to counsel families, which I think is a great idea. So they performed a systematic review and meta-analysis. From several thousand potential studies, they were able to narrow it down to 31 eligible studies, most of which were cohort studies. Risk of bias was assessed using validated tools. They did pooled prevalence estimates using random-effects models, and meta-regression to explore associations with things like study year, age at surgery, and sex distribution, for example.

And ultimately, across the 31 studies, ranging all the way from the year 2000 to 2019 — so a nearly 20-year review of fairly recent data — their overall finding was that there was a pooled prevalence of any ischemic brain injury in the neonatal population of almost 70%, right? 68.5%, with low heterogeneity. So pretty good confluence of the grouped studies in that case. So almost three quarters of the population has some kind of ischemic brain injury.

Nim Goldshtrom (03:37.678)

In the neonatal period, in critical congenital heart disease — meaning they're going through surgery. So now let's break it down further. For pre-operative injury, the prevalence was about 35%. And here the heterogeneity starts to get more complicated, so not as clean a pooling of the data, but there's still no significant bias. For new post-operative injury, the prevalence increases up to 46.5%. And now when we look at the categories — the types of injury we have —

the highest-degree injury shown is white matter injury, the most common lesion, at almost 40%, but here there's a high degree of heterogeneity, which we'll talk about in a second. Followed by arterial ischemic stroke, which occurred in about 10 to 20% of patients. And lastly, hypoxic-ischemic injury, which is only about 3.5%, with a moderate amount of heterogeneity. Ultimately, they also found that age at surgery showed some association with white matter injury prevalence.

And while sex distribution was a little skewed for some prevalence of injury, this was all washed away basically when also controlling for CHD lesion type, which we'll explain a little in the discussion. So we have a study here focusing on ischemic brain injury, showing that it is extremely common, often clinically silent in our population. And lastly, out of all the methods evaluated by the study — which again I highly recommend the audience listen to — MRI (magnetic resonance imaging) still remains the most superior method, over ultrasound and definitely over CT (computed tomography), at detecting subtle lesions. With a high recommendation — not that this is what the study looked at — for considering other tools like EEG (electroencephalogram) to look for non-structural but functional parameters in our population, which we know CHD can have a significant amount of clinical or subclinical seizures as well. So I think this is wildly interesting to see how vast

and how prevalent this is in our population. And yet we don't have the answer to the question that this evidence is begging: so you have an injury, but does that mean your child is going to have a problem? Adrian, what did you think about this?

Adrianne Rahde Bischoff

I mean, these numbers are really high, especially compared to just a preterm population in general. But it is really fascinating to see that almost all these kids — more or less — are gonna come out with some kind of structural injury on a scan.

Nim Goldshtrom (06:02.488)

But what does this mean for them going long term?

Adrianne Rahde Bischoff

Yeah, I was gonna ask you that specific question because this is not a population that I have a lot of exposure to, and I definitely have no experience in follow-up of these kids. But I was wondering what your perspective is when you do see these kids — I don't know if you see them yourselves, or what your colleagues who do follow-up say about how these kids do when they come to clinic. Because I think it's important to put it in perspective, because the numbers look very doom-and-gloom.

But does it actually matter? And what impact does that really have on the quality of life and the true outcomes that families really care about — which is something we've been talking about here, and in many different forums, the last couple of years: what are outcomes that are clinically meaningful, not just for us? Yeah, it's important to know that there's, you know, like a "ditzel" there — abnormal on the MRI — or maybe a couple more than just a few ditzels.

But what do the families really care about, and how does that impact their child and their happiness going on? So what's your perspective? What do you guys see when these kids come back?

Nim Goldshtrom

So this is probably the most important question that you could possibly ask, right? And the reality is follow-up for this population can be challenging. But the data on neurodevelopmental outcomes — at least let's start with the first level, which is the two-year Bayley scores — most of these kids, at around the two-year mark, are going to show about one standard deviation, on average, lower Bayley scores than the general population.

Nim Goldshtrom (07:42.008)

So most of them, in the three domains — cognitive, behavioral, motor — you're gonna see somewhere in the high eighties or low nineties for most of these kids, which isn't bad, but it's also not great. It's also skewed a lot by the most severe lesions — the single ventricles are gonna have the lower scores, the kids who land up on ECMO (extracorporeal membrane oxygenation), the more premature, the SGA (small for gestational age) and growth-restricted. And one standard deviation is not great, but it's also not the most terrible thing.

But we still have not had strong associations in studies that show these kinds of findings — that these anatomical or structural findings on things like MR imaging are correlated with the incidence of these developmental follow-up outcomes. And this is where we have to be very careful. So some of the limitations of this specific study that we have to really hedge about — regarding the findings this paper points to, which is a lot of structural findings of ischemic injury —

which is, first of all, there was not a lot of preterm population in this group. So think of this as a term study, and they excluded genetic conditions. So you really have to caveat this to the non-genetically-diagnosed, full-term infant who is going to have these findings. We know from a lot of other great work that these kids are already at risk for this at baseline — that a full-term, 40-week CHD neonate has a brain the size of a 36-weeker,

compared to non-congenital-heart kids. Their brains are growing smaller, specifically the left side of the brain in certain kids — there's papers on this as well. Again, this study did not look at the association between NDO (neurodevelopmental outcomes) and ischemic brain injury. And I want to point out to readers who go through this article that they suggest these findings — because of MRI findings — may change parents' desires about how to pursue surgery. And I take strong exception to that statement in the discussion.

These findings, in my opinion, should not be the controlling factor of whether to proceed to surgery. The global rate of survival from neonatal critical congenital heart disease has been staggeringly improving — meaning things are getting better, we're able to help these kids survive longer — and all the NDO data shows, you know, one standard deviation. And there's not yet a very strong correlation that these injuries are the things definitively associated, depending on their grade,

Nim Goldshtrom (09:56.738)

with these kinds of developmental outcomes. So I'd be very careful to take this data and apply it to mean this impacts and means your child will definitely have some kind of these problems. And we also know from the five- and eight-year studies that the early problems you see in motor, cognitive, and behavioral aspects at two years differ and translate, as a child with CHD grows to five, eight, and school age, into maybe more problems of executive functioning, emotional regulation. And the problems you see at two are not the problems you see at five and eight. So we really need better studies that can

cover this arc of time and do a better job, because these are really interesting and fascinating findings, but the link isn't perfectly there yet. The one very fascinating thing they did find in this study is the association with age. And if readers go to specifically Figure 3 in the paper, they'll see that age at surgery — in the first two weeks of life — was associated with the findings of

white matter injury, but not in the way you'd expect. Usually you'd think that the longer you're sitting and waiting, the more you're at risk of, you know, ductal physiology — PDAs (patent ductus arteriosus) kept open, more hemodynamic instability. But actually, these kids were found to have a higher incidence of white matter injury the earlier they went to surgery. So day four, five, and six have a higher prevalence rate of injuries, and day seven through eleven have a lower prevalence rate.

And so this might be a little bit due to bias — why are kids getting surgery at day four and five? Are they sicker? It's not completely uniform who's getting it. But it was really fascinating to see that this link is a little different than what you'd expect from other studies. And the sex difference ultimately got washed away by lesion type, which is not unexpected. We know that males are slightly more prone to certain types of lesions, like left-sided lesions — aortic stenosis, coarctation, and TGA (transposition of the great arteries). So it's not completely unexpected that the sex difference gets washed out by these differences. So really,

interesting work. I think this clearly points to us doing more longitudinal connections — centers doing MRI studies that have great follow-up programs. It would be great to see these kinds of connections — how do those MRI findings, operative course, post-operative course, and then outpatient follow-up link together? What kind of services are they getting, what kind of engagement with families are they getting? It would be great to link these two things together. But an interesting observation —

Nim Goldshtrom (12:12.546)

be very hesitant to make strong recommendations about how this should impact care right now, but it can and should drive more research.

Adrianne Rahde Bischoff

So let me ask you a different question then. Do you think routine MRI — whether pre, post, or both — should become the standard of care? I bet we'd get a very wide range of answers if you put one of my colleagues from neurocritical care here versus some of the more skeptical folks, saying, well, if this doesn't really correlate with what we see in the kids later on, is it actually beneficial to have this information? Does it just

add stress to the system, stress to the families about knowing this is there, when it's not something they can change or that might actually cause anything? What's your take on that?

Nim Goldshtrom

So I have two opinions that are hopefully driven by the data. You're right — the logistics of getting an MRI in most places is not easy. Moving a child who is critically ill, or who is pre- or post-operative, is not a simple thing — on PGE (prostaglandin E1, used to keep the ductus arteriosus open), or having just recently had surgery. So what is the gain? The gain would potentially be there if, let's say, as a program, you're noticing that certain lesion types, certain surgery exposures, certain procedures are ending up with new post-operative white matter injury.

That's an actionable item — getting the pre-op scan, especially, let's say, with TGA patients, who are very high risk for pre-operative brain injury from balloon atrial septostomies, shows you whether your program is somehow one that is either not able to mitigate, or not meeting standards, or somehow experiencing children with new injuries — that's a quality-improvement angle. So you can argue there's value in doing it and learning about your program — is it in certain surgeries, certain conditions, certain bypass or perfusion techniques — that's one thing. And the second is,

Nim Goldshtrom (13:58.954)

if this is what it takes to argue, or to defend, or to help families get data to say, "my child is at risk," and as an outpatient, if they may be getting pushback or difficulty getting insurance to support early intervention services — these could be the added tools that help us show it. Here is the objective evidence — MRI data that can get these kids plugged in and give parents more

tools, and more of a sense of why we want them to get into early intervention, to come back to our clinics — which can sometimes be very difficult to keep up in follow-up, because "my child seems to be doing fine, they're full term, we had the heart surgery, so everything could be fixed." That's sometimes a lot of the message they hear, but it's not really reflecting the baseline of the child — when we bring them in for a developmental evaluation and look closely, we see they're kind of slightly behind in certain domains. And to that end, if systematically you can create those kinds of

support systems, it's probably worthwhile to consider. But, you know, getting an MRI on every neonate or every small child is not an easy thing. You have to really weigh the benefits and the cost to your system. But it can be a tool that supports both improvement at your center and long-term follow-up opportunities for your families.

Adrianne Rahde Bischoff

Is there a role for a pre-operative MRI to preclude a baby from getting surgery?

Nim Goldshtrom

I think very significant neurological injury would probably preclude it, because if you have something that's already really significant — either a bleed or a hypoxic-ischemic injury — first of all, if that were to happen, you'd probably have a clinical picture that matches it. Subtle brain abnormalities, or subtle ischemic or white matter changes, are not going to manifest with a situation that severe. I'll give you one example. Several years ago we had an emergency rescue

Nim Goldshtrom (15:52.972)

for a total anomalous pulmonary venous return (TAPVR) child born at another hospital. We brought them in at eight hours old, and they were extremis — we could not get the sats up. We got an echo, and we ended up putting them on ECMO first to rescue the child physiologically. We did that, we stabilized the child, we got an EEG — there was completely attenuated activity and seizures — and ultimately got head imaging and continued EEG that basically showed there was near brain death. And that dissuaded us

from wanting to do surgery on the child, because it was not going to help. But the clinical picture correlated with that consequence. These mild findings, again, in my opinion, should not be the things that are discouraging, because the clinical outcomes we're currently seeing do not match this prevalence of 70%, knowing that neurodevelopmental outcomes are, at best, maybe one standard deviation off. And most parents would take that in a minute, if they knew their child could have a potentially long life, or a long enough life that is meaningful to them and their family.

Adrianne Rahde Bischoff

Thanks. That was actually quite helpful. I know we can't extrapolate this to preterm neonates, but we do, every once in a while, have preemies here in our NICU with congenital heart disease. So it's helpful to get that perspective from term infants — which obviously is not the same — but I appreciate that.

Nim Goldshtrom

No, that's why I found it fascinating to give a landscape review of where the data is. I think this is a great systematic review to give us a cross-section — here's where we are, but we certainly have a ways to go to connect the dots. This is just one slice of the picture, and the clinical and functional aspects we see in these children at two, five, eight, and preteen years doesn't completely line up with what we see early on. But I guess that's a good thing, right — to have services and to be able to connect the dots over time.