#469 - 📑 Journal Club - The Complete Episode from October 3rd 2026

Hello friends 👋

This week on The Incubator Podcast, Ben and Daphna dig into four studies testing what the evidence supports at the bedside. They open with the 24-month BeNeDuctus follow-up, where expectant management and early ibuprofen for PDA yield identical survival without neurodevelopmental impairment. Daphna then presents the Nourish trial, finding individually targeted milk fortification did not improve growth, body composition, or brain MRI outcomes. Next, a Dutch study links probiotics to halved NEC rates, but rising non-NEC mortality and a fatal probiotic sepsis case leave Ben ambivalent. Daphna closes Journal Club with an Australian cohort suggesting nearly a third of NICU deaths met organ donor criteria. The week wraps with Ben and Eli on Neo News, unpacking a sibling-matched study finding no link between prenatal Tylenol and autism or ADHD.

----

The articles covered on today’s episode of the podcast can be found here 👇

Long-Term Outcomes of Expectant Management versus Early Ibuprofen for Patent Ductus Arteriosus: 24-Month Follow-Up of the BeNeDuctus Trial. Apers WMJ, Hundscheid T, Kooi EMW, Onland W, Vijlbrief D, Steiner K, Donders R, Mulder AL, de Boode WP; BeNeDuctus Trial Investigators.J Pediatr. 2026 Sep 17;299:115292. doi: 10.1016/j.jpeds.2026.115292. Online ahead of print.PMID: 42637114

Individually Targeted Human Milk Fortification: A Randomized Clinical Trial. Belfort MB, Ahtam B, Bell KA, Berger P, Cherkerzian S, Ellard D, Foster L, Fusch C, Gagoski B, Grant PE, Kuncham M, Pepin H, Solti M, Steele T, Woodward L, Inder T.JAMA Pediatr. 2026 Sep 14:e264220. doi: 10.1001/jamapediatrics.2026.4220. Online ahead of print.PMID: 42734908

Probiotic Implementation and Necrotizing Enterocolitis Risk in Preterm Infants. Imren C, Jongejans V, Onland W, Keyzer-Dekker CMG, Been JV, Derikx JPM, van Kaam AH, Simons SHP, Twisk JWR, van den Akker CHP, Vermeulen MJ.JAMA Netw Open. 2026 Aug 3;9(8):e2631601. doi: 10.1001/jamanetworkopen.2026.31601.PMID: 42671836 Free PMC article.

Potential Neonatal Donors in Australia: A National Cohort Study. Barzegar R, Cavazzoni E, Popat H, Wong G.Pediatrics. 2026 Sep 3:e2025075110. doi: 10.1542/peds.2025-075110. Online ahead of print.PMID: 42686207

The Untapped Potential of Organ Donation from the NICU. Nguyen BV, Coloma M, Crouch EE.Pediatrics. 2026 Sep 3:e2026076954. doi: 10.1542/peds.2026-076954. Online ahead of print.PMID: 42686210 No abstract available.

Prenatal Acetaminophen (Paracetamol) Use and the Risk of Autism and/or Attention-Deficit/Hyperactivity Disorder Among Sibling-Matched Cohorts. Luo S, Gong Q, Ai Y, Zhang J, Chan L, Wong WCW, Ip P, Chan EWY, Tanuseputro P, Wong ICK, Wan EYF.JAMA Intern Med. 2026 Sep 1;186(9):1102-1111. doi: 10.1001/jamainternmed.2026.2215.PMID: 42371637

----

Watch this week's Journal Club on YouTube 👇

----

The transcript of today's episode can be found below 👇

Ben Courchia, MD (00:00)

Hello, everybody. Welcome back to The Incubator Podcast Journal Club. Definitely good morning.

Daphna Yasova Barbeau, MD (00:06)

Time is just flying. I don't remember the last time we did Journal Club, but it wasn't that long ago.

Ben Courchia, MD (00:13)

Yeah. We schedule them and think, "Okay, we have some time," and next thing you know, it's right around the corner.

Daphna Yasova Barbeau, MD (00:18)

And here they are. Well, we finally had a little breeze of cool air in South Florida this morning, so that was nice. We'll see.

Ben Courchia, MD (00:26)

Finally, huh? No more rain. The funniest thing is happening. I'm going to the AAP (American Academy of Pediatrics) meeting in San Diego, and I'm very much looking forward to this vacation slash conference, because I'm going a few days early. And there's a Category 5 hurricane in the Pacific. I was like, "Are you kidding me? I'm dodging hurricanes in Florida." I'm finally excited to go to California, and suddenly there's a Category 5. So yesterday I was on a call with members of the section, talking about the meeting, and I asked them, "I'm flying in. Should I be worried about this hurricane?" And they're like, "What hurricane? You guys in Florida... I'm sure you'll make it." They didn't know. It's on their side, and they had no clue.

Daphna Yasova Barbeau, MD (01:14)

It's your hurricane. Oh no.

Ben Courchia, MD (01:21)

I guess it shouldn't affect California too much. And I really need that vacation, so I'm going no matter what.

Daphna Yasova Barbeau, MD (01:30)

Well, I hope everybody will be very safe.

Ben Courchia, MD (01:33)

I think so. The conference is toward the end of the week, and I think the storm is mostly happening this weekend. Still, I hope everybody is safe. It looks like it's going to be mostly in Mexico, but Mexico and Southern California are very much connected.

Daphna Yasova Barbeau, MD (01:47)

Natural disasters.

Ben Courchia, MD (01:49)

Yeah, the biggest one they've ever recorded, they said. Anyway.

Daphna Yasova Barbeau, MD (01:51)

Oh my goodness. Maybe it's you. You're taking the hurricanes along with you.

Ben Courchia, MD (02:00)

I didn't think about that.

Daphna Yasova Barbeau, MD (02:05)

All right, we'll hope nobody has any hurricanes. But you have a very exciting paper to present.

Ben Courchia, MD (02:08)

That's right. Everybody stay safe. I have two papers this week. The first one is on the PDA (patent ductus arteriosus). I thought it was a good article: "Long-Term Outcomes of Expectant Management Versus Early Ibuprofen for Patent Ductus Arteriosus: 24-Month Follow-Up of the BeNeDuctus Trial."

Daphna Yasova Barbeau, MD (02:19)

Long awaited.

Ben Courchia, MD (02:37)

First author is Hundscheid and colleagues, so apologies if I've pronounced that incorrectly. We'll take the opportunity to review the BeNeDuctus trial a little, and some of the other PDA trials. I'm doing hemodynamics training, and even I sometimes get lost in all the trials coming out. They all seem to be saying the same thing, that pharmacological treatment doesn't work, but we should still understand those different trials. So we'll use this opportunity for a quick review.

In the introduction, the authors remind us of the usual stuff: PDA has lots of complications, it's debated, expectant versus pharmacological management.

Daphna Yasova Barbeau, MD (03:15)

Is that the background? Yeah, I think that about sums it up.

Ben Courchia, MD (03:22)

Almost. You read these introductions on the PDA, and it's always the same thing. But they do mention the three recent large trials that looked at pharmacological treatment: BeNeDuctus, Baby-OSCAR, and the PDA trial. They all show similar short-term outcomes for survival, BPD (bronchopulmonary dysplasia), and NEC (necrotizing enterocolitis), with no benefit of early pharmacologic treatment compared with expectant management or placebo. In fact, BeNeDuctus found a slightly lower rate of moderate to severe BPD, and the PDA trial found slightly lower mortality in the expectant management group. Now we're gathering the long-term follow-up of these trials, and that information is crucial. We know PDA is associated with changes in brain perfusion, cerebellar development, and motor function by age 5. So these childhood neurodevelopmental outcomes need to be assessed.

In this study, we're looking at survival and neurodevelopmental outcomes of the extremely preterm infants enrolled in BeNeDuctus at 24 months corrected age. The hypothesis is that expectant management, compared to early ibuprofen, would not lead to meaningful differences in long-term survival.

Before we begin, I want to go over some of the trials that have come out. We've reviewed most of them, I believe. I'll talk about BeNeDuctus in a second, but the other important one is Baby-OSCAR, published in the New England Journal of Medicine in 2024. It's a double-blind, placebo-controlled randomized trial in babies under 29 weeks with a PDA of 1.5 millimeters or larger in the first three days of life, comparing ibuprofen to placebo. They found no difference in the primary outcome of death or moderate to severe BPD at 36 weeks, and no difference in survival without moderate to severe neurodevelopmental impairment at 24 months. That was a big trial, 653 babies.

The other one is the PDA trial by Matt Laughon and colleagues. We interviewed Dr. Laughon at Hot Topics, so you can go back to that episode, where we discussed the trial shortly after it came out. It was published in 2026 in JAMA (Journal of the American Medical Association). It has just shy of 500 babies, and if you remember, it was stopped early. These are babies between 22 and 28 6/7 weeks with symptomatic PDA, comparing active treatment with acetaminophen, ibuprofen, or indomethacin versus expectant management. They had to stop the trial early because of increased mortality in the treatment arm. In the data they had collected, death or BPD at 36 weeks showed no difference, and there was that mortality signal: 9.6% in the treatment group versus 4% in the expectant group.

Then some other trials came out. The most recent is TRIOCAPI, in JAMA Pediatrics in 2026, by our French colleague Rozé and colleagues. It's another RCT (randomized controlled trial), almost 800 infants, same population, 23 to 28 weeks, with prophylactic acetaminophen versus placebo regardless of PDA status. There was no difference in survival without severe morbidity at 36 weeks, increased ductal closure by day seven, and no reduction in severe cerebral lesions or mortality.

There's also the SMART-PDA trial, which we reviewed recently, the Souvik Mitra paper in Archives of Disease in Childhood. It has a small number of babies, about a hundred, but remember it's a pilot feasibility study, so I don't know if it counts as a major PDA trial. They were really testing a very targeted SMART algorithm to decide which babies need treatment. So I'll stop here, because TRIOCAPI, the PDA trial, and Baby-OSCAR are the big ones, and we'll talk about BeNeDuctus in a second. Any questions so far, Daphna?

Daphna Yasova Barbeau, MD (07:49)

So far so good.

Ben Courchia, MD (07:50)

The BeNeDuctus trial was originally published in 2023 in the New England Journal of Medicine. It had about 300 babies, enrolled between 2016 and 2020 in the Netherlands, Belgium, and Denmark. Today we're looking at those babies' outcomes at 24 months.

Infants were included if they were born before 28 weeks of gestation and had a PDA confirmed on echocardiogram, larger than 1.5 millimeters at its narrowest point, with a left-to-right transductal shunt. People sometimes forget that a PDA can be measured in different frames and configurations, and you can get quite large measurements depending on where you put your cursor.

Between day one and day three of life, babies were randomized to early ibuprofen or expectant management. The intervention group received early ibuprofen according to the local protocol, with an echo at least 12 hours after each course. A second course was allowed, and then a third course or ligation. The expectant management group received no treatment. Unlike other PDA trials, this study was applauded because there was very little crossover, meaning almost no open-label treatment in the expectant arm. I believe only one baby crossed over, which for PDA trials is unique.

Daphna Yasova Barbeau, MD (09:25)

Remarkable. Very remarkable.

Ben Courchia, MD (09:27)

The primary long-term endpoint was survival without neurodevelopmental impairment at 24 months corrected, defined as a Bayley-III (Bayley Scales of Infant and Toddler Development, Third Edition) cognitive and motor composite score of 85 or higher. Other outcomes included survival without severe neurodevelopmental impairment, defined as a Bayley-III score of 70 or higher. They looked at individual Bayley-III scores, behavioral assessment, biometrics, general health, mild or major developmental disorders, and the need for paramedical support like rehabilitation services. General health was categorized as good with no readmissions in the first two years after discharge, moderate with one to three, and poor with more than three. The individual outcome components are fairly standard, so I won't spend too much time on them. The analysis was a modified intention-to-treat analysis of the available follow-up sample, based on the original randomization group.

So what did it show? The original BeNeDuctus trial looked at a composite of NEC, moderate to severe BPD, or death at 36 weeks. It met the threshold for noninferiority, with lower rates of moderate to severe BPD with expectant management. Of the 273 infants enrolled, 136 were randomized to expectant management and 137 to early ibuprofen. At 24 months, data were available for [117 and 116] infants, which is 86% of the expectant group and 85% of the treatment group. The primary endpoint could be assessed in about 76% of the expectant group and 70% of the intervention arm. Baseline characteristics were pretty much the same between groups. Interestingly, the children lost to follow-up had higher maternal age, higher gestational age, and lower rates of tocolysis and surfactant administration. So if anything, they were probably a population that did better, not a sicker one.

The primary endpoint, survival without neurodevelopmental impairment, occurred in 49% of the expectant management group and the exact same 49% of the early ibuprofen group, a relative risk of 1. Identical. There were no deaths after NICU (neonatal intensive care unit) discharge, so survival in the follow-up sample was 84% with expectant management and 80% with early ibuprofen, a relative risk of 1.04. Secondary outcomes showed no real differences in biometrics, general health, neurological exam, vision, or hearing impairment. On neurological exam, three children in the expectant group were considered abnormal, with GMFCS (Gross Motor Function Classification System) levels 2, 3, and 4, compared with two children in the ibuprofen group. The Child Behavior Checklist was completed for 66% of the 192 infants who completed follow-up, with no significant differences. And there was no difference in the main follow-up endpoint in the predefined subgroup analyses: gestational age (26 weeks and above versus under 26), birth weight (under or over 1,000 grams), sex, multiple pregnancy, and antenatal steroids.

So the authors conclude these results support expectant management as a valid approach to the PDA compared with early pharmacologic treatment. Now, that does not mean no PDA should ever be treated. I want to pause here, because understanding this statement is critical. We're currently recording a five-episode series with Dr. Joe Kaempf called Thinking About Thinking. In it, we try to understand the basics of biostatistics, how they relate to our interpretation of the data, and how we translate that into counseling at the bedside. We talk about this in episode two: just because we don't reach significance doesn't mean we throw the hypothesis out the window. So, as the authors say, it does not mean that no PDA should ever be treated.

These findings apply to early ibuprofen for a PDA larger than 1.5 millimeters with a left-to-right shunt and shouldn't be directly generalized to other treatment approaches. It's also important to know that infants with contraindications to ibuprofen, like active bleeding, pulmonary hypertension, or renal dysfunction, were excluded from the original BeNeDuctus trial. So infants who might have benefited from eliminating the ductal shunt weren't eligible. That raises the question: if we had closed their PDA, say with transcatheter closure, would we have seen something different? That question remains open, and the data are being collected. The other open question is whether a selected population would benefit from treatment, which brings us back to SMART-PDA. The authors call for developing and validating reliable clinical and echo markers to identify babies at higher risk of PDA-related complications, kind of what Souvik is doing with SMART-PDA.

And because pharmacologic treatment comes with both treatment failure and adverse effects, they suggest exploring definitive primary PDA closure with transcatheter intervention. That would allow a comparison of prolonged PDA exposure versus definitive closure in high-risk preterm infants. We reviewed data from the Canadian network not long ago showing closure rates of about 40% with pharmacologic treatment, so it mostly doesn't work. If we know we can definitively close the PDA, it would be very interesting to see the effect of closing it versus leaving it open. Finally, follow-up into school age and beyond is essential to capture the full trajectory of cognitive, motor, and behavioral development and quality of life, and to understand the long-term implications of different PDA management strategies.

These papers are always a nice opportunity to review the PDA literature, because there are so many publications and it can get muddy and confusing. This paper doesn't change where most of the data is pointing: for now, expectant management is probably the way to go, until we have more information about targeted approaches like SMART-PDA or data on definitive closure like transcatheter closure. Dr. Benitz from Stanford recently gave a series of talks at McGill, where I'm doing my hemodynamics training. He's very adamant that we shouldn't even attempt pharmacologic treatment. He's a staunch advocate of that approach, which resonated with the McGill folks, who also don't believe in treating pharmacologically. I think that's where the field is right now, and we'll see whether upcoming data moves us in another direction.

Daphna Yasova Barbeau, MD (17:05)

Well, let me ask you this. Most of us in the community are either studying whatever we study or at the bedside, and we're not looking at PDAs all the time. But you're in this interesting space where, in real time, all you see is PDAs. You're learning hemodynamics at a time in neonatal medicine when we're asking, "We've been doing this thing; should we be doing it?" How has that changed your experience at the bedside in your training?

Ben Courchia, MD (17:55)

I think the opportunity to use targeted neonatal echocardiography and hemodynamics really helps you understand what you're looking at. My old view of the PDA, how big it is and which direction the shunt goes, is a very primitive way of looking at a very complex system. These days, my assessment of a PDA and what intervention is needed is guided by its size, the degree of shunting, et cetera. But it's also very interesting to look at the PDA's downstream effects. What does cerebral perfusion look like? Are we compromising it, specifically in diastole? What kind of perfusion do we have in the descending aorta, again in diastole? And then look at the lungs. Sometimes a baby has some clinical instability that we're keen to attribute to the PDA. But then you pair your echo with lung ultrasound and see that these lungs are absolutely awful. Maybe the lungs should be your primary focus, and maybe a course of steroids would relieve a lot of what you're seeing. So a more comprehensive approach makes a lot more sense to me, and it lets you look at each baby more individually.

Which hints at another series we're about to record. We've been discussing some hemodynamics teaching, and I'll be hosting a series with Dr. Altit from the NeoCardioLab, where we'll review some of the basics of hemodynamics. The goal is not to make people hemodynamics experts. It's to ask: what parameters can we use from these studies, how do we interpret them, and how reliable are they? Like we were just discussing, six people could measure a PDA and get six different measurements, and that's a problem. So how do we agree on certain things? If the interventricular septum is flattened, how do we interpret that? Does it mean pulmonary hypertension, or something else? I think this will be quite interesting.

Daphna Yasova Barbeau, MD (20:16)

All right. I didn't even mean to set you up for that pitch, but I think we'll all be looking forward to it very much.

Ben Courchia, MD (20:22)

We've been teasing it on the air, asking if people think it's needed, and we received a lot of messages saying yes, we want more hemodynamics coverage.

Daphna Yasova Barbeau, MD (20:27)

A lot of good feedback.

Ben Courchia, MD (20:50)

My goal is to do it in a way that's helpful and compelling. Not just the basics, but going beneath the surface into the nitty-gritty. I think the first topics will probably be PDA and pulmonary hypertension. These are big ones. As a neonatologist, I've reviewed echoes for pulmonary hypertension assessments without understanding half of the measurements. Take pulmonary artery acceleration time: what is that? What does it measure? Where is it measured? What does it tell you? Or right ventricular pressures, whether they're half systemic, two-thirds systemic, or less than half systemic. I know half systemic or more is no good, but what does that mean, and what are the implications? All of this will be discussed. I'm very happy Dr. Altit is coming on the show, because he's a world-renowned expert, and I'll get to chime in. No pressure on me; all the pressure's on him.

Daphna Yasova Barbeau, MD (21:38)

The best-case scenario.

Ben Courchia, MD (21:40)

I'll have to be well behaved. He's my program director at the moment, so...

Daphna Yasova Barbeau, MD (21:42)

That's right. He'll be the boss this time. Looking forward to it very much.

Daphna Yasova Barbeau, MD (22:17)

I've got a paper. You had already planted the seed that I would like this paper, and I did in fact select it. It's in JAMA Pediatrics, titled "Individually Targeted Human Milk Fortification." The lead author is Mandy Belfort and the senior author is Terrie Inder, so quite a few big names on this paper. The question was: if we individually target fortification using a point-of-care human milk analyzer, compared with standard-of-care fortification, can we improve growth outcomes? So I was very hopeful about this intervention.

Ben Courchia, MD (23:05)

Don't spoil it.

Daphna Yasova Barbeau, MD (23:06)

That's all I'm saying. I don't have to be unbiased, because I didn't do the research. I just have to report the findings. So it was a randomized clinical trial of individually targeted human milk fortification. What does that look like? In the intervention arm, they took whatever human milk the infant was being fed, donor milk or mom's own milk, and analyzed it for several nutritional components, which I'll tell you about. If the base milk wasn't as good as it could be, they added nutrients on top of the fortification to bring the baby to a predetermined target. The control group got standard fortification. It was a single Level III NICU, and the trial was called the Nourish study, which I love.

Ben Courchia, MD (24:11)

Great name.

Daphna Yasova Barbeau, MD (24:12)

I know. They enrolled singletons or twins born at 24 to 30 weeks' gestational age whose mothers intended to provide expressed milk, and who would be younger than 21 days of chronological age when the study diet was expected to start. There were a number of exclusions: a major congenital anomaly; birth weight below the third percentile; existing NEC, intestinal perforation, or other significant GI (gastrointestinal) pathology; significant kidney or liver dysfunction; any inborn error of metabolism; anticipated fluid restriction below 140 mL per kilo per day for more than three days; an existing diagnosis of grade 3 or 4 IVH (intraventricular hemorrhage); and expected transfer before 36 weeks. They also excluded babies whose parents declined consent for donor milk in case it was needed, babies in DCF (Department of Children and Families) custody, and babies whose death was anticipated.

They enrolled infants from February 2020 to March 2025, and had to pause enrollment from March to June 2020. How frustrating: you start a project in February 2020...

Ben Courchia, MD (25:31)

If only they knew.

Daphna Yasova Barbeau, MD (25:33)

If only they knew that COVID would really mess up recruitment that first summer.

They used block randomization in a one-to-one ratio, and notably, twins were randomized individually. Parents and the entire clinical and research team were blinded to group assignment, except, obviously, the research assistants who prepared the milk.

Let's talk about the diets and the unit's standard of care. They shared some of their local guidelines. Briefly, enteral feeds were advanced, preferably with maternal milk supplemented with donor milk as needed, at volumes determined by birth weight. Once babies reached about 60 mL per kilo of unfortified human milk, the milk was routinely fortified with human milk fortifier; they used Similac HMF during this period. Volumes were advanced to a goal of 150 to 160 mL per kilo. For infants with birth weights under 1,500 grams, their standard of care was to add liquid protein once the baby reached goal volume. And for any baby with growth faltering, meaning weight gain under 15 to 18 grams per kilo per day and/or length gain under 0.8 centimeters per week, they increased milk volume or added protein or MCT (medium-chain triglyceride) oil.

Babies started the study diet once they were receiving and tolerating full enteral feeds, at least 140 mL per kilo. Each day, they made a 24-hour pool of unfortified milk for each baby and analyzed it every weekday with a point-of-care human milk analyzer, the Miris HMA. They measured macronutrients and calculated metabolizable energy content. For those studying for the boards: 9 kcal per gram of fat, 4 per gram of protein, and 4 per gram of carbohydrate.

In the control group, research assistants fortified the milk as ordered by the clinical team, without using the analysis. The milk was analyzed, but the clinical team didn't know the baby's group or the results, so they just ordered routine fortifier. In the intervention group, milk was fortified per the clinical order, plus additional protein and/or fat as needed based on the analysis. The goal was to reach minimum levels, within plus or minus 5%, of 1 gram of true protein per 100 mL and at least 67 kilocalories per 100 mL. What if the base milk exceeded those targets? It was still fortified per the clinical team's orders, because that's what the baby would have received without the analysis. They didn't target carbohydrates, basically because of product limitations. On weekends and holidays, bedside nurses prepared the milk routinely per the clinical team's order, I'd imagine because the study assistants weren't available. The study diet stopped at 36 weeks' postmenstrual age, after which babies received routine fortification per unit guidelines.

For each participant, they calculated the mean macronutrient and energy content across all analyzed days and looked at daily intakes. They also measured weight, length, and head circumference at the start and end of the study and again at NICU discharge. They looked at body composition using air displacement plethysmography with the PEA POD, which is really cool. You put the baby in a little container, and it measures how much air is displaced to estimate body composition. They derived Z-scores for postmenstrual age and assessed participants as close as possible to the end of the study diet.

They also did brain MRIs (magnetic resonance imaging), which I thought was super cool. All babies had a brain MRI at term-equivalent age, between 38 and 41 weeks, during sleep and without sedation. They measured total and regional tissue-specific brain volumes using segmentation models, and specifically looked at weighted average measurements of the PLIC, the posterior limb of the internal capsule. They collected covariates and clinical diagnoses from the EMR (electronic medical record) and maternal questionnaires. And notably, they're also following two-year outcomes, with two prespecified primary outcomes. That follow-up is ongoing and will be reported separately.

So let's get into the results. Of the 130 randomized participants who started the study diet, 87% completed it through 36 weeks' postmenstrual age. Mean gestational age was 28.5 weeks in the intervention group and 28.4 in the control group. Mean birth weight was 1,165... sorry, I don't have it for the control group. Hold on.

Ben Courchia, MD (32:23)

What?

Daphna Yasova Barbeau, MD (32:26)

Birth weight. Hold on. No, 1,165 grams is the overall figure for the whole group. The mean birth weight Z-score was 0.2 in the intervention group and 0.1 in the control group.

Then they looked at the milk's macronutrients. Median true protein content of the base milk was 1.18 in the intervention group and 1.24 in the control group. True protein intake from base milk alone, without fortifiers, was 1.55 versus 1.64. Mean protein intake from base milk plus all fortifiers was 4.47 versus 4.58. Content and intake of all other macronutrients and energy were similar. The proportion of enteral volume that was mom's own milk rather than donor milk was high and similar in both groups. All participants in both groups received at least the minimum recommended intakes for all macronutrients and energy.

And they did not observe any substantial differences between groups in growth, body composition, or brain structure. That was pretty much it: no differences across any of the measures.

In their discussion, they note something unexpected. Even though point-of-care milk analysis guided protein and fat adjustments up to five days a week, the targeted fortification group actually got slightly less protein. That was likely due to a chance difference in base milk, since the control group's base milk had slightly more protein. Fat and energy delivery were similar. Their overall conclusion is that this trial, together with other studies using direct milk macronutrient analysis, shows it's possible to meet recommended intakes for virtually all very preterm infants with standard-of-care fortification. The findings do not support widespread use of point-of-care milk analysis to individually target fortification. So I thought that was very interesting.

Ben Courchia, MD (35:16)

It is very interesting. Maybe a lot more interesting than people realize, actually, because...

Daphna Yasova Barbeau, MD (35:22)

Because it feels like it should work. I would have been interested in the babies whose base milk was very nutrient-dense. Is there a group of babies where we wouldn't need to add fortifier? I don't think so; I think we'll still need to add something for every baby. But I wonder. And looking at the outliers, is there base milk that's really not optimal that we could improve? This study says no, but that's at the group level. I wonder if there are some real outliers this could help. We'll see.

Ben Courchia, MD (36:11)

Yeah. But I do find the conclusion quite interesting, where they say it's possible to meet recommended intakes for virtually all very preterm infants using currently available commercial fortifiers. And if the analysis shows that the mother is producing exceptional milk, then maybe you don't need to fortify everything, just add protein. That's another avenue to explore with families.

Daphna Yasova Barbeau, MD (36:45)

You were getting too excited about the future. That's fine.

Ben Courchia, MD (36:50)

Yeah. A world where you analyze the mother's milk and fortify it with her own milk: that's the future.

Daphna Yasova Barbeau, MD (36:54)

I think so too.

Ben Courchia, MD (37:27)

There's a very interesting article in JAMA Network Open, "Probiotic Implementation and Necrotizing Enterocolitis Risk in Preterm Infants." The first author is Dr. Imren. The article starts by reminding us that probiotics may help prevent necrotizing enterocolitis by restoring microbial balance and suppressing pathogenic organisms. So far, everybody should be up to speed.

The point here is that individual strains may have distinct functional properties: stimulating gut motility, strengthening the mucosal barrier, modulating the immune response, and producing anti-inflammatory peptides, cytokines, short-chain fatty acids, and so on. However, the wide variety of probiotic products has led to inconsistent findings about effectiveness. That's why strain specificity matters so much when we evaluate effectiveness and safety.

Safety concerns, including rare cases of probiotic-associated sepsis, have led to cautious implementation around the world. Those concerns are what prompted the slowdown in probiotic use in the US after the FDA (Food and Drug Administration) memorandum. ESPGHAN (European Society for Paediatric Gastroenterology, Hepatology and Nutrition) and the American Gastroenterological Association endorse specific strains. Others, like the American Academy of Pediatrics, advise caution, specifically for babies under 1,000 grams. Since 2019, a combination of Bifidobacterium infantis, Bifidobacterium lactis, and Streptococcus thermophilus has been available in Europe, with documented effectiveness and safety in preterm infants. This study evaluates whether NEC incidence changed after it was implemented in Dutch NICUs. I'm going to pull up the paper, because this will be a long one and you'll have things to say.

This was a multicenter retrospective study. We don't usually do retrospective studies, but this one is quite telling, well done, and involves a lot of babies.

Daphna Yasova Barbeau, MD (40:02)

And kind of necessary at this point.

Ben Courchia, MD (40:17)

Absolutely. It was conducted at the Level IV NICUs of Erasmus in Rotterdam and in Amsterdam, from 2018 to 2024. Infants under 30 weeks' gestation, under 1,000 grams, or both, were eligible if they were born in or transferred to these NICUs within 24 hours of birth. GI anomalies requiring surgery and death within the first 48 hours were exclusion criteria. The primary analysis is intention-to-treat, based on whether babies were born before or after implementation. So this study evaluates the effect of a NICU-level policy change, not individual probiotic exposure, but I think it's an important way to look at this.

The intervention was a multistrain probiotic called ProPrems, made by NeoBiomics and manufactured under good manufacturing practice standards in accordance with European food law. Each batch came with certificates of strain identity and purity, viable counts at the end of shelf life, and antibiotic susceptibility and resistance profiles, in line with ESPGHAN recommendations.

Interestingly, for our US listeners, this three-strain combination of B. infantis, Streptococcus thermophilus, and Bifidobacterium lactis is what we used to have in the Tri-Blend probiotic. That product was available to US hospitals before the whole probiotics debacle. Daphna, correct me if I'm wrong: the probiotics mostly used were Evivo and the Tri-Blend, and I think both are off the market in the US now.

Daphna Yasova Barbeau, MD (42:11)

Well, everything is off the market in the US.

Ben Courchia, MD (42:13)

True. You can still get probiotics at Whole Foods, but these were hospital grade, for the NICU. I think the Tri-Blend was made by Abbott, but I forget, and that doesn't really matter. My point is that the product in this paper has the exact same strains as that one. So it's interesting for us, because it came through US NICUs at some point.

Each 0.5-gram daily dose contains a combined total of at least 10 to the 9th colony-forming units of those three bacteria. Once a day, the suspension was prepared by reconstituting the powder in 1.2 mL of sterile water in a dedicated room. It was drawn into an enteral syringe and given through the gastric tube within two hours of preparation, immediately before a feed. Supplementation started when enteral feeds reached 2 mL per bolus, typically on day two to four of life. It continued until 35 weeks' postmenstrual age or NICU discharge. During temporary feeding pauses to evaluate feeding intolerance, the probiotic was stopped. All infants in both cohorts received mother's own milk or donor human milk, fortified with a bovine-based fortifier once feeds reached 80 mL per kilo per day. They don't get into the feeding protocols.

The primary outcome is NEC incidence. We recently reviewed a US trial of a specific live probiotic agent that looked at mortality; here, the primary outcome is NEC. Secondary outcomes include all-cause mortality, non-NEC-associated mortality, focal intestinal perforation, late-onset sepsis, and time to full enteral feeds, with probiotic-associated sepsis tracked as a safety endpoint. NEC was defined as modified Bell stage 2 or greater. It was categorized as medical, surgical if surgery was needed, or fatal if the baby died of NEC.

Then they did something interesting: they created a reference cohort to account for confounders that change over time, which matter in a before-and-after comparison. They took babies born between 30 and 32 weeks weighing over 1,000 grams, who, per their policy, did not receive probiotics. If practice evolved over the years, this reference group should show it, so not every change would be attributed to the probiotic. They say clearly that it's not a great reference cohort, because these babies are more mature, with a lower risk of NEC and mortality. But it's the best they could do.

So let's get into the results. Over the study period, 1,413 infants met eligibility criteria, 47.5% of them female: 598 before implementation and 815 after. Baseline characteristics were generally similar. Median gestational age was 27.7 weeks before and 27.9 after, and median birth weight was 950 grams before and 975 after. One difference: prolonged rupture of membranes and cesarean delivery were more frequent after implementation. Adherence to the probiotic protocol was high, 96.3%.

What about NEC rates? They dropped from 11.9% to 5.3% after implementation. Medical, surgical, and fatal NEC were all lower, and both Bell stage 3A and 3B decreased. They give an adjusted risk ratio, and since we're doing some biostatistics this year, here's what that means. It's the risk after implementation divided by the risk before, once differences between the groups are accounted for. Below 1 means lower risk, above 1 means higher risk, and if the confidence interval crosses 1, it's not significant. After adjusting for gestational age, sex, birth weight, five-minute Apgar, prenatal steroids, and PPROM (preterm prelabor rupture of membranes), probiotic implementation was still associated with a 51% lower risk of NEC. That's an adjusted risk ratio of 0.49, with a number needed to treat of 15. The association held across categories: 0.5 for medical NEC, 0.44 for surgical NEC, 0.47 for fatal NEC, and 0.49 in extremely preterm infants alone. NEC rates were lower after implementation at every week of gestational age.

Let me see if I can show you this. We're now looking at Figure 1 together.

Daphna Yasova Barbeau, MD (48:03)

Wow. Yeah, it's their run chart. Impressive.

Ben Courchia, MD (48:09)

Exactly. You can see the NEC rates before and after the probiotic, and there's a nice drop. Some might say a 12% NEC rate is higher than usual, and this is where local statistics matter. But I wanted your take on something interesting in this graph: the timing of NEC. They seem to have spikes every year between July and October. It makes you wonder. There's been discussion in the literature about whether we see more NEC when viral infections are more common. They don't talk about this in the paper, but it jumped out at me.

Daphna Yasova Barbeau, MD (48:55)

Yeah, it's very cyclical. I think that's kind of a big deal.

Ben Courchia, MD (49:08)

Well, we'll let them know, I guess.

Daphna Yasova Barbeau, MD (49:08)

I told you once, "I feel like we get a lot of NEC in 'blank' month," and you said you'd look, but I don't know if you ever did.

Ben Courchia, MD (49:15)

We're fortunate that our NEC rates are quite low. I believe that without probiotics, we're at about 3%, so we're doing quite well, thankfully. It's hard to establish patterns when the base rate is low. It would be interesting to look at it regionally, like in South Florida, where we practice.

In the reference cohort, baseline characteristics were comparable before and after implementation. NEC incidence was lower after implementation, going from 5% to 2.7%, but that wasn't statistically significant. So first, they had a 5% NEC rate in a 30- to 32-week population. Is that generalizable to other units? I don't know. And they also saw a decrease in that cohort, even though those babies weren't exposed to probiotics. I'll let you decide what to make of that.

All-cause mortality was similar before and after implementation, 11.7% versus 11.5%, and mortality among extremely preterm infants was also unchanged. Here's what I think deserves attention: non-NEC-associated mortality was higher after implementation, 7.2% before versus 9.6% after. This was significant after adjustment, with an adjusted risk ratio of 1.42, a confidence interval of 1.01 to 1.98, and a p-value of 0.04. Median postnatal age at death went from about seven days to nine days. The same pattern appeared in extremely preterm infants, 11.6% versus 15.2%. The extra non-NEC deaths were spread across multiple causes that the authors say appeared largely unrelated to the probiotic: PPHN (persistent pulmonary hypertension of the newborn), perinatal asphyxia, and cardiac or neurological causes. There were also four additional deaths from focal intestinal perforation after implementation, although the overall rate of spontaneous intestinal perforation didn't change.

Late-onset sepsis rates were lower after implementation, from 40% to 27.6%, an adjusted risk ratio of 0.72, mostly because of fewer culture-negative episodes. Median time to full enteral feeds was shorter, eight days versus nine. And this one was a gut punch: there was one case of probiotic sepsis. The infant suddenly deteriorated, x-rays showed extensive pneumatosis, and the blood culture grew B. infantis. Within hours, the infant died of severe pulmonary hemorrhage despite full intensive support.

The authors conclude that routine supplementation with a well-characterized probiotic formulation was associated with lower NEC rates among very and extremely preterm infants. They see this as clinical evidence supporting this specific formulation. However, the increase in non-NEC-associated mortality calls for cautious interpretation and further investigation. They call for research to harmonize probiotic formulations, clarify strain-specific effects, and identify the subpopulations most likely to benefit.

You read the results section, and it starts with the primary outcome, and you think, "Boom, slam dunk, it went down." Then you keep reading, and by the end you're like, "I don't know how I feel about this." One case of probiotic-associated sepsis and death is concerning. Their NEC rates were relatively high at baseline. They also had a fairly high NEC rate in more mature babies, and it went down in the population that wasn't exposed to probiotics, too. I left the paper more ambivalent than when I started reading.

Daphna Yasova Barbeau, MD (53:47)

I agree with you. I think we're going to rely on these retrospective studies now to get enough data on probiotics. It's complicated, because every medication we give our patients has side effects, right? But when the side effect is an infection from the thing you gave, it's tough. Their finding of an increase in non-NEC mortality is interesting, because many other probiotic studies have found the opposite. We're going to need everybody to publish their data so we can move forward on this topic, one way or the other.

Ben Courchia, MD (54:39)

Interestingly, and maybe we'll review this in an upcoming Journal Club, Ravi Patel shared another article on LinkedIn about probiotic use in Sweden with the same composition. It would probably do this paper justice. They found that NEC and culture-proven late-onset sepsis really went down. They compared the periods before and after the probiotic guideline was adopted in 2020. The overall rate of death and/or NEC went from 3.4% to 1.5%, which is probably more in line with NEC rates in a lot of other units. So where you start matters tremendously. If your NEC rate is 12%, anything you can do to cut it in half is quite important.

Daphna Yasova Barbeau, MD (56:02)

A hundred percent.

Ben Courchia, MD (56:06)

And I wonder, are we going to get to the point, like with blood transfusions, where we tell parents, "We're going to use probiotics, and there's a very small risk your baby may get an infection from the bacteria"?

Daphna Yasova Barbeau, MD (56:28)

Yeah, I think it goes back to our number needed to treat and number needed to harm.

Ben Courchia, MD (56:32)

Yeah. Fifteen in this particular case. We'll see.

Ben Courchia, MD (56:39)

Hello, everybody. Welcome back to The Incubator Podcast Journal Club. We're back today for another episode. Daphna, good morning. How are you?

Daphna Yasova Barbeau, MD (56:45)

Good morning. I'm doing very well. Thank you for reviewing the probiotic literature. I was just saying that I'm feeling disappointed, that's all. I told you I was ready, and I was mentally ready, but apparently I was not logistically ready. Okay, here we are. Welcome.

Ben Courchia, MD (57:05)

That was quick.

Daphna Yasova Barbeau, MD (57:10)

I don't know about my new system. It's not working great for me today.

Ben Courchia, MD (57:15)

You're finally back in your recording booth at home.

Daphna Yasova Barbeau, MD (57:21)

That's right, so I'm just settling back in. We're going to highlight two articles, both in Pediatrics. One looks at an Australian cohort and organ donation from their NICUs, and the other is an editorial on organ donation from NICUs in the US. We'll start with the paper from Australia. The lead author is Rebecca Barzegar and the senior author is Germaine Wong. They wanted to look back at neonatal deaths and identify how many babies could potentially have been organ donors.

They used the ANZNN, the Australian and New Zealand Neonatal Network data registry, a clinical registry of every NICU in Australia and New Zealand. They looked at all registrants born between 2012 and 2022 who were admitted to one of these NICUs in the first 28 days of life and died before going home, at a corrected gestational age of 35 weeks or more.

There were a lot of exclusion criteria, which I'll get into. The exclusion criteria for donation were adapted from those used for routine pediatric and adult donors, following the Transplantation Society of Australia and New Zealand's clinical guidelines for organ transplantation from deceased donors, in affiliation with their organ donation program, Donate Life. They first sorted patients into two groups by cause of death: a neurological insult or a non-neurological cause. Then, as part of their audit criteria, patients were placed into four subgroups based on brain death outcomes:

Category A: confirmed death by neurologic criteria

Category B: probable death by neurologic criteria, but not tested

Category C: imminent death by neurologic criteria within 72 hours

Category D: not likely to die by neurologic criteria

Since they had registry data but not individual medical records, they couldn't identify categories A and C. So any baby who died of severe HIE (hypoxic-ischemic encephalopathy) was assigned to category B, and all other babies without exclusion criteria went into category D.

HIE becomes a common theme throughout this paper. So what were the exclusion criteria for organ donation eligibility? I think lots of people will want even just this part of the study. They excluded:

a weight, estimated or actual, under 3 kilos at the time of death

no mechanical ventilation, or unknown ventilation status, at the time of death

an unknown or unavailable cause of death, or unplanned death after cardiorespiratory arrest

a chromosomal or genetic condition diagnosed before or after birth

underlying kidney or liver disease, or multiorgan dysfunction at the time of death, which includes necrotizing enterocolitis

active bacterial, fungal, or viral infection at the time of death, or active infection with specific organisms, including Candida, at any point during life

inborn errors of metabolism, or underlying blood, lymphatic, or malignant disease

any major, multiple, or complex multiorgan congenital anomalies, such as abdominal wall defects

Neonates with relative or unclear contraindications were discussed individually by the clinical research team, and eligibility required consensus. So again, the first split was between donation after circulatory determination of death (DCDD) and donation after neurologic determination of death (DNDD).

During the study period, they identified 1,760 infants who met those criteria. The cohort was fairly evenly split by sex, with boys making up 56%. Median corrected gestational age at death was 40 weeks, and median age at death was 11 days. Median weight at death was about 3,230 grams, and 71% had a confirmed or estimated weight of 3 kilos or more. Gestational age at birth ranged from 23 to 43 weeks, which was interesting given those medians. However, almost 58% were born at term, 37 weeks or more.

The most frequent causes of death were HIE, pulmonary hypoplasia or pulmonary hypertension, and congenital heart disease. As a side note, the congenital heart disease group excluded hypoplastic left heart syndrome specifically, which made up 2.3% of the entire cohort. Causes also included chronic lung disease and intestinal conditions, including NEC. HIE was the predominant cause of death in term babies, and respiratory failure in preterm babies.

When they applied all the exclusion criteria, including weight, ventilation status, and planned end-of-life care, 750 of the 1,760 infants were further sorted into neurologic versus non-neurologic causes of death. Of this group, 25% were categorized as non-neurological, with some further exclusions based on weight and ventilation. Seven babies died of sudden or unexpected cardiorespiratory arrest before end-of-life care could be planned. Of the remaining patients in that group, 15% had an underlying diagnosis considered eligible for neonatal organ donation via the circulatory death pathway, approximately 250 infants.

Neonates with major congenital anomalies were excluded. That removed babies with abdominal wall defects, coloboma, heart defects, atresias, growth restriction, GU (genitourinary) anomalies, and vertebral anomalies. GI conditions like NEC were excluded because of the risk of systemic inflammation and infection.

Of the cohort, 17.4%, around 307 infants, were considered to have a neurologic cause of death. Severe HIE was diagnosed in 4% of patients who met all the other criteria. So overall, fewer neonates were eligible through the neurologic pathway than the circulatory pathway. In total, 28% of the cohort, nearly 500 babies, were eligible through the circulatory pathway. Altogether, 32% of the 1,760 neonates who died met criteria as potential organ donors. Among potential donors, the most common causes of death were HIE, respiratory failure, and congenital heart disease, and most were term infants who died within the first 14 days of life. Donation after circulatory death accounted for most potential donors, 28% of the cohort, compared with 4% through neurologic determination.

By contrast, in Australia and New Zealand over a 20-year period, only 5 of 374 consented donors under 16 who went on to donate weighed less than 5 kilograms. The authors say this highlights how little neonatal and infant donors contribute to the overall pediatric donor pool.

They acknowledge that the study couldn't capture important clinical details that may affect donor suitability, like hemodynamic stability or kidney and liver function at the time of death. They also couldn't talk with families individually, so family perspectives on neonatal organ donation were beyond the scope of this study. But it's an important area for future research. There appear to be babies who could qualify, and chances to refer neonates receiving end-of-life care that may be going unused.

I also wanted to highlight the editorial in the same issue of Pediatrics, "The Untapped Potential of Organ Donation from the NICU," by lead author Brandon Nguyen and senior author Elizabeth Crouch. I'll read some highlights, because I think they've already distilled their editorial as much as possible.

Less than 1% of all organ or tissue donations come from donors younger than one year, yet children in this age group have the highest waitlist mortality. Organ donation in the US follows the dead donor rule, and because brain death declarations are rare in neonates, most NICU donations happen through donation after circulatory death (DCD). The share of pediatric organs from DCD donors has increased significantly. That shift is being accelerated by new technologies like ex vivo perfusion, which circulates warm, oxygenated blood or specialized fluids through a donated organ. It lets clinicians resuscitate the organ and assess how well it works before transplantation, which improves outcomes.

Many critically ill neonates die in the NICU after life-sustaining interventions are withdrawn, but organ donation is often not considered beforehand, so opportunities are missed. There are obviously logistical considerations: ventilator management, timing of withdrawal, neonatal physiology, and NICU workflow. And this is different from the adult population. But they stress that the most common reason for missed donation is that the medical team doesn't refer, or refers too late. Studies indicate that up to 44% of NICU deaths in the US meet the minimum criteria for donation after circulatory death. Yet fewer than 10% are referred to organ procurement organizations, and fewer than 3% result in successful donation. Additionally, in a survey of NICU nurses and neonatologists, none of the respondents had any prior donation experience, and only 22% considered neonatal donation "very important."

Then they move on to what we can do about it, and it really comes down to standardizing neonatal protocols. To close these gaps, early referral to organ procurement organizations must become standard practice, and referral should happen before life-sustaining interventions are withdrawn. In successful neonatal cases, referral has happened as little as 90 minutes before organ recovery, so a lot can be done in a very short time. A central reason donation is so rare is the lack of a universally accepted protocol tailored to neonatal physiology and NICU workflows, and data from multiple countries strongly support standardization. For example, a Spanish NICU that implemented a formal organ donation protocol increased its referral rate from 2.5% to 16.7%. They go on to say that expanding this donor pool is clinically justified, because even these tiny organs can do quite well.

Of course, there are concerns about how parents will see this and how we fit it into end-of-life conversations. But focus group data show that families want donation presented as part of holistic end-of-life care, alongside decisions about baptism, family visits, or other religious practices, rather than as a separate, rushed conversation at the end. Empathic communication and sensitive timing seem to limit decision fatigue. And importantly, few families appear to suffer psychological harm from being offered the option of donation. Families of all potentially eligible infants should be offered the opportunity. Huge.

Ben Courchia, MD (1:12:40)

That's an important point, because sometimes we're reluctant to even bring up the subject. We think, my God, we're going to pile onto a family that's already distressed and cause more distress.

Daphna Yasova Barbeau, MD (1:12:46)

Yeah, we say, "We'll make it worse." Will we make it worse, or play into misconceptions about organ donation? There's a lot of mistrust in the community about organ donation: "Are they not going to do as much for my baby if we say my baby is an organ donor?" So I think it's important to address those things specifically.

Just as the Australian team challenged NICUs worldwide to move beyond anecdotal experience toward systematic, protocol-driven approaches, the editorial authors recommend that US NICUs work toward standardizing the offer of organ donation. They want it done in a way that honors both the potential of these tiny donors and the wishes of their families.

This reminded me so much of our last Delphi conference, when Nick Embleton talked about breast milk donation after death. He said that most families who were offered breast milk donation after their baby's death said yes. They had wanted the opportunity; they just didn't know how to ask. They wanted the chance to share their baby, to honor their baby. What I took away from him is that every family should be presented with those options. Will every family take them? No, but they should at least know they have them.

Ben Courchia, MD (1:14:28)

Yeah, agreed. And what do you think the impact of doing that could be?

Daphna Yasova Barbeau, MD (1:14:38)

I think it's twofold. I've been part of the care of some families who requested organ donation themselves, and they were actually quite disappointed to find out their babies weren't suitable donors. That happened in a few cases. And there were families who were thrilled to find out their baby could donate even parts of the eye. Their clinical picture ruled out organ donation, but they were able to help another child see, and that was incredibly therapeutic for those families. It was super powerful to be part of. So one, there are people who need organs. But our babies should never be seen just as organ donors. I think it's also a potentially powerful way to support families who are bereaved or about to be.

Ben Courchia, MD (1:15:43)

And by the way, there's also organ donation not just for transplant, but for medical research.

Daphna Yasova Barbeau, MD (1:15:52)

Sure, there are lots of opportunities. And we have to set aside our own biases and say, it's not about me, it's about what this family wants, and they should know all the options.

Ben Courchia, MD (1:16:04)

Yeah. Legacy is a very important part of neonatal death. And to be honest with you, it's a bit scary as well. You present these articles, and I'm nodding along, "Yep, yep," but if we had to do this tomorrow, I wouldn't know where to start.

Daphna Yasova Barbeau, MD (1:16:27)

I agree with you, and I think that's what they're trying to say. We can't just say, as individual neonatologists, "Let's wing it." We should sit down as national and international groups and say, "Let's standardize this," just like we're trying to standardize other things, so every family has the same opportunity. And we have some good data on this. Thankfully, so many people in our community have been brave enough to sit with bereaved families and ask what they would have wanted, or what was helpful. We have that information. Every family is different, but they should have the options. It's going to be an uncomfortable discussion, and standardizing it makes it less uncomfortable. Not because that's easier for us, but so that we actually do it.

Ben Courchia, MD (1:17:21)

Yeah. I don't mind being uncomfortable. I'll seek discomfort, as a famous YouTube channel says. But even then, I've never worked in a place where, if a family said yes, we'd know what to do. Who does the retrieval? Where does the organ go?

Daphna Yasova Barbeau, MD (1:17:38)

Yeah, we would scramble.

Ben Courchia, MD (1:17:45)

It will take a lot of legwork. You probably need a champion to get it started. And I agree with you, it should be done at the network level, identifying the points of contact in each area.

Daphna Yasova Barbeau, MD (1:17:58)

Yeah. I've been part of some of these cases, and most of these discussions happen through the organ procurement organizations, not with individual clinicians. I think that helps protect families from feeling any pressure from the clinical team, which is a big point to highlight. The way it worked was that we had a standardized protocol for bringing in the organ procurement organization to talk to families.

Ben Courchia, MD (1:18:38)

Yeah. You mentioned circulatory death versus brain death, and I feel like that's a big issue in the NICU. You said HIE is an important condition here, and I think that's because the concept of brain death may be better discussed and accepted in that setting. But in general, it's tough to diagnose brain death in the NICU. Do you think end-of-life discussions are sometimes held back by how hard it is to accept the reality? I'm curious about your thoughts. And if I'm bothering you, you can tell me.

Daphna Yasova Barbeau, MD (1:19:17)

No, I think it's an important conversation. This is what we do; we're at the water cooler, as they say. We've got to talk about these things. I think you're right. We struggle, internationally, universally, with neurologic declarations of death. We do have some guidelines, but they're sometimes easier said than followed. Whether families accept those guidelines is a different story. Whether institutions rely on family acceptance or make the declaration regardless is something we can discuss.

But what this Australian paper highlighted is that it didn't really matter whether there was a declaration of death per se. A lot of these babies were undergoing withdrawal of care, so they didn't necessarily have to be declared brain dead. If the team and the family have decided that withdrawing care is an option for this child, and how teams get there differs by institution, then that's a family we should offer organ donation as part of end-of-life care. I don't know if that answers your question. Did I skirt it entirely?

Ben Courchia, MD (1:20:56)

No, that makes sense. You're saying the conversation really comes when you're discussing withdrawal of life support, so you already have acceptance, and we can focus on that.

Daphna Yasova Barbeau, MD (1:21:08)

Correct. I don't know if that's the right answer; it's what I think. If, as a team, you've decided you're willing to withdraw life support, why would the exact definition of brain death rule a baby in or out for organ donation?

Eli (1:21:20)

We're moving on to our next article, which touches on a topic we really haven't addressed at any point in the history of this show. Just kidding. We talk about it all the time, and we'll keep talking about it because it is so stinking important. Sigh is how I feel about this.

Ben Courchia, MD (1:21:56)

This is exciting. I think we're going to talk about a very polarizing topic, somehow, some way. But the data we're going to review is very interesting.

Eli (1:22:05)

Yeah, let's talk about it. This belongs to a thread on the show that maybe we should call "Adventures with HHS." This is the latest edition of our Adventures with HHS (Department of Health and Human Services) thread. It touches on the alleged relationship between autism spectrum disorder and prenatal use of Tylenol. The study was published in JAMA Internal Medicine, and it was really interesting and very robust. It drew on a cohort of over 700,000 mother-child pairs in Hong Kong, about 40% of whom had prenatal Tylenol exposure. From that cohort, it built a sibling-matched cohort of about 120,000 children who were assessed for autism spectrum disorder and about 100,000 who were assessed for ADHD (attention-deficit/hyperactivity disorder).

Overall, they found that prenatal Tylenol exposure was not associated with an increased risk of either autism spectrum disorder or ADHD. Those findings held across every measure of Tylenol use they looked at: the timing of exposure across trimesters and gestation, the pattern of use (how consistently and in what manner the medication was taken), and the dosing. So across all three of those dimensions, there was no association between prenatal Tylenol exposure and autism spectrum disorder or ADHD. Ben, what did you think of this study: the way they conducted it, the data, and the findings?

Ben Courchia, MD (1:24:00)

I think this is a very important paper. It was published in JAMA Internal Medicine, which isn't a place we always look for articles in our specialty. But the way the data was collected and the way the analysis was built are really what let it answer the question that's been raised: is acetaminophen (or paracetamol, whatever we want to call it) taken during pregnancy linked to a higher rate of autism spectrum disorder or ADHD?

A lot of the data we've seen quoted relies too often on surveys, or on patients' recall of their acetaminophen use. If you've done any research at any level, you know that's not an ideal way to collect data, especially when you're trying to prove an association, or even causation, between a medication and a potential morbidity. So they were able to gather data on a massive cohort of more than half a million people over 20 years, using prescription data and good tracking of prescriptions, diagnoses, and so on. I think that already makes us feel a lot better about what this data shows.

Number two, it's important to emphasize that no link could be established between prenatal use of acetaminophen or paracetamol and an increased risk of autism or ADHD.

What I also find interesting is how they ran their analysis, because they did a lot of different things. It would take quite a while to go through the entire paper now. But they ran the analyses in several different ways and, almost like a teacher, they reproduced the association other papers have found and then showed you: "You see? If you don't handle the data properly, you can be misled." In other words, it's mostly an issue of confounding. One of the best examples they use is the supposed link between prenatal Tylenol and autism spectrum disorder and ADHD. Throughout the paper, they show how important it is to look at the data correctly, because if you do it the wrong way, you can end up drawing associations that are just false.

As you mentioned, and I want to say it again, this is a sibling-matched analysis. That's the most important part of the study. For me, this is one of the papers that should settle the conversation. I'm sure it won't, but it's done extremely well, for the reasons I mentioned.

Eli (1:27:48)

Yeah. I guess the question is how you take this study into your clinical practice. How do you explain this data and share it with your patients? Here's what I've seen, both while trying to follow how these conversations are evolving and while talking with some of the families I've had the opportunity to care for. Explaining evidence hierarchies and stronger methods isn't necessarily an accessible way to find common ground, or to show that some evidence is stronger than other evidence. So how do you describe the statistical techniques, the data handling, and everything behind the curtain that makes a study like this more reliable? How do you bring that to patients in a way that's accessible and builds trust between them and the healthcare system, rather than saying, "We understand this better than you do," or "Your studies are wrong"?

Ben Courchia, MD (1:29:12)

Yeah, for me, the way to explain this isn't to put things in opposition. It's to have a conversation with patients and families about the scientific process as a continuum, and about how the evidence evolves. You can't always dismiss the arguments people bring you, because they're rooted in things that were published, or, as I said, in things people in the government have said. They're not saying it out of nowhere. So just saying, "No, no, ignore that," is probably not the right approach.

In my opinion, presenting this as a story is always the best way. First, you can say the possible link between acetaminophen and ADHD or autism is a very interesting question, because some early studies asked, "Could there be a link?" That question led to a lot of activity in our field to find out whether there was one. You then present this new data as what came after the question was asked: an effort to answer it as robustly as possible and determine whether it's safe to take Tylenol during pregnancy. I think that paints a different picture.

When it comes to explaining the data and how it was handled, there are certain things that, like you said, we won't be able to explain. We're in a room trying to make a decision, and it's not easy to give a course in biostatistics. Most of us aren't biostatisticians. But what we mentioned earlier in this episode is something many people can understand: the early studies relied on whatever people could remember about what kind of Tylenol they took, how much, and so on. Anyone with common sense can see that this way of measuring acetaminophen intake is quite questionable. So you can explain that this problem was tackled head-on, and that researchers started looking at prescriptions and actual diagnoses and comparing siblings. People may not grasp every implication, but at least you show the work is robust.

And like I always say, some people are willing to have a genuine conversation with you, and they may be open to being convinced or not. Other people have made up their minds, and that's just the way it is. No amount of data you present will change how they feel or what they want to do. But the key is the idea that handling the data properly can answer the question, and can also show how we may have been misled before. The fact that the findings are different does create a bit of opposition, but that's okay. That's probably how I would handle it. What would you do?

Eli (1:32:35)

Boy, listen, what do I know? I think one of the privileges we have in the NICU, which many other parts of the healthcare system don't have, is continuity. These conversations can unfold over time. So much of the healthcare system runs on brief, one-time encounters, but we can do more than that.

Take families who are hesitant about vaccines, for example. My pattern is to start with a first conversation, and usually it doesn't go very far. I say, "Do you want to talk about vaccines?" They say, "No, I don't really want to talk about it." And I say, "Okay, I'm going to ask you about it. I'm on service for the next two weeks." I'm a fellow, so I'm always on service. "I'm going to ask you about this every single day for the next two weeks." I tell them in advance. Then I come back the next day and ask about it, and that conversation goes nowhere too. And I say, "I'm going to ask about this every single day for the next 13 days."

The third day I show up, they'll say, "Okay, okay, let's talk about it." And I say, "Can you help me understand where you're coming from? What are your concerns, and where is the hesitancy coming from?" Maybe they're worried about pain. Maybe they think the baby is too small, or that the baby is still recovering from an acute illness. Maybe they feel it's not necessary. Based on what they say, I tailor my approach to address those concerns both directly and indirectly.

Directly, I try to affirm that we're on the same team. I share my own concerns and put them side by side with the family's. If there's an opportunity to educate around their concerns, for example about adverse event rates, I put those rates in perspective against all the bad things that could happen if the baby doesn't get vaccinated.

Indirectly, I bring evidence. I say, "Don't take my word for it." Especially in this era of ChatGPT, with everybody doing their own research, I say, "You can take my word for it, and that's great. But let me also show you, transparently, all the evidence I'm looking at." Then I come back with a single study, or a manila folder of studies, and drop it off. Especially for families who are really curious about health information, I'll say, "If you get the chance to read this stuff, I'd love to talk about it." The next day on rounds, I ask, "Did you read the stuff?" They say, "No, it was a busy night," or whatever.

By the end of those two weeks on service, I want to have had a real conversation. I understand their needs. I've shared what I'm looking at transparently so they can review it. Hopefully, we go through it line by line. I think that's when we can get into methodological nuance: on day 12 of my service block, after we've been talking about this for 12 days. That's how I think about it. And who knows whether it works systematically, but I try to use that continuity and that time to my advantage. I also try to understand their concerns in a very deep way, communicate very transparently what I'm looking at and what worries me, and then give them access to all those resources so we can talk.

Ben Courchia, MD (1:36:36)

Yeah. So persistence is a big part of it.

Eli (1:36:40)

Hey, listen, that's life. It happens in patient conversations, and it happens with arterial lines. Sometimes the line doesn't go in, and you just have to find a new vessel.

Ben Courchia, MD (1:36:48)

Yeah, I agree with you. It's a very mature approach for someone at your level of training, and what you're saying is absolutely true. We have to stay available to families. They get their information from somewhere. Like you said, it might be social media or something else. But we have a duty to stay available as another source of information, one that could offer better data. One of our advantages, like you said when you mentioned the risks and the things you're assessing, is that this tailored approach is key. It lets us say, "I know your baby better than this influencer on TikTok, and we can think about this together in a much more sophisticated way than what you might see online." I think families might respond to that, and it gives them a lot of valuable information. Giving them space is also very important: "If you don't want to talk about it, we won't, until you're ready." It shows that we're not activists. We're really there for the wellness and welfare of the patient and the family. I think that's a big deal.

Eli (1:38:06)

Absolutely. Listen, if there's one lesson I've learned from journalism, it's that so much of the battle is just showing up: being present, sitting with people, and actually listening. I hope that part of my life adds real value to my clinical practice. When I have difficult conversations with politicians about issues, or difficult conversations about politics with sources, the way you get through them and come out stronger together is to sit there and listen, show that you really care, and respond in real time. I think we can take a very similar approach with the families we care for, on issues we both care about but may be coming at from different places.

Ben Courchia, MD (1:38:59)

It's interesting that you draw this parallel with journalism. With this paper, where we are today on acetaminophen (Tylenol, paracetamol, whatever we want to call it) is a bit like journalism. The data is pouring in, but the news was printed months ago, right? Now we're at the point of asking whether they'll issue a correction. I don't know that we'll see one in this case. I'm worried that the narrative has been set and that there won't be any further discussion at the political level. I think that's very unfortunate. First, as with newspaper corrections, it's always very hard to undo the impression that's already been given, even when a correction is issued. Journalists talk about this, and it's true: you can issue a correction, but it's sometimes hard to undo something that was done wrong. And if no correction is even issued, I don't know where we'll be. Again, we talk about this here, to an audience of committed, caring people. But the question is how we'll spread this particular body of evidence, because it looks like the people who spread the incorrect information won't even make an effort to correct themselves.

Eli (1:40:32)

Yeah. And the longer we stay silent, the more room there is for information from other sources to reach people. When people have questions, they'll look for answers. They won't wait for us.

Ben Courchia, MD (1:40:40)

Absolutely. Exactly. All right, this was fun. I'll see you next time. Take care.

Eli (1:40:51)

Thanks so much, Ben. See you next time.