Hello friends 👋
In this episode of Journal Club, Ben and Daphna are back in the studio for a lively discussion on the latest neonatal research. They start by highlighting upcoming conferences, including the BPD Collaborative meeting, before diving into a mix of clinical and policy-driven studies shaping neonatal care. They explore a JAMA Pediatrics study on the risk of apnea following two-month vaccinations in preterm infants, discussing how these findings impact vaccine counseling in the NICU. Next, they examine research on the link between pregnancy-induced hypertension and neonatal thrombocytopenia, offering insights into how maternal conditions affect early lab values and bleeding risks.
The episode also features an EBNEO segment with Dr. Srishti Jayakumar and Dr. Sarah DeMauro, covering a Brazilian randomized trial on parent-guided developmental interventions for very low birth weight infants. The findings reveal significant improvements in motor and language outcomes, reinforcing the power of early parental involvement.
Finally, they tackle the debate on NICU volume and outcomes, analyzing data that suggests higher-volume centers may improve mortality and morbidity rates—but with important caveats.
Packed with critical analysis, debate, and practical insights, this Journal Club episode is one you won’t want to miss!
----
Check out the BPD Annual Conference
The BPD Collaborative Annual Symposium is coming soon! The Symposium will be held Wednesday afternoon, March 5 and all day Thursday, March 6, 2025 @ Hilton Philadelphia @ Penn’s Landing.
We are excited to share the agenda (2025BPD COLLABORATIVE AGENDA) which includes exciting presentations from all our Ad Hoc groups, breakout times for all ad hoc groups to meet individually, updates from the data committee, a presentation from our international members, a poster session which will include work from Collaborative sites and an optional dinner.
New this year!
We will be using Eventbrite to register for the symposium and tickets for dinner.
Here is the Eventbrite Registration/Dinner link:
This form can be opened on any computer, tablet or your mobile device:
----
The articles covered on today’s episode of the podcast can be found here 👇
Apnea After 2-Month Vaccinations in Hospitalized Preterm Infants: A Randomized Clinical Trial. Greenberg RG, Rountree W, Staat MA, Schlaudecker EP, Poindexter B, Trembath A, Laughon M, Poniewierski MS, Spreng RL, Broder KR, Wodi AP, Museru O, Anyalechi EG, Marquez PL, Randolph EA, Aleem S, Kilpatrick R, Walter EB.JAMA Pediatr. 2025 Jan 6. doi: 10.1001/jamapediatrics.2024.5311. Online ahead of print.PMID: 39761016
Effects of pregnancy-induced hypertension on early-onset neonatal thrombocytopenia. Ye M, Zhou C, Li L, Wang L, Zhang M.BMC Pregnancy Childbirth. 2025 Jan 24;25(1):67. doi: 10.1186/s12884-025-07193-z.PMID: 39856602 Free PMC article.
Outcomes of extremely preterm infants who participated in a randomised trial of dopamine for treatment of hypotension (the HIP trial) at 2 years corrected age. Marlow N, Barrington KJ, ODonnell CPF, Miletin J, Naulaers G, Cheung PY, Corcoran JD, Khuffash E, Boylan GB, Livingstone V, Pons G, Straňák Z, Van Laere D, Macko J, Wiedermannova H, Dempsey EM; HIP consortium.Arch Dis Child Fetal Neonatal Ed. 2025 Jan 19:fetalneonatal-2024-327894. doi: 10.1136/archdischild-2024-327894. Online ahead of print.PMID: 39832819
Parent-Guided Developmental Intervention for Infants With Very Low Birth Weight: A Randomized Clinical Trial. Silveira RC, Valentini NC, O'Shea TM, Mendes EW, Froes G, Cauduro L, Panceri C, Fuentefria RN, Procianoy RS.JAMA Netw Open. 2024 Jul 1;7(7):e2421896. doi: 10.1001/jamanetworkopen.2024.21896.PMID: 39018071 Free PMC article. Clinical Trial.
EBNEO Commentary: Parent-Guided Developmental Intervention for Infants With Very Low Birth Weight. Jayakumar S, Burris H, Duncan A, DeMauro SB.Acta Paediatr. 2025 Jan 30. doi: 10.1111/apa.17608. Online ahead of print.PMID: 39887768 No abstract available.
The Impact of Hospital Delivery Volumes of Newborns Born Very Preterm on Mortality and Morbidity. Phibbs CS, Passarella M, Schmitt SK, Martin A, Lorch SA.J Pediatr. 2025 Jan;276:114323. doi: 10.1016/j.jpeds.2024.114323. Epub 2024 Sep 18.PMID: 39304118
Balancing Volume and Outcomes: Insights and Cautions from a Multi-State Analysis of Care for Infants Born Very Preterm. Jiang S, Lee HC.J Pediatr. 2025 Jan 24:114481. doi: 10.1016/j.jpeds.2025.114481. Online ahead of print.PMID: 39864501 No abstract available.
Unbound bilirubin and risk of severe neurodevelopmental impairment in extremely low birthweight newborns. Arnold CC, Maric I, Wong RJ, Tyson JE, Stevenson DK.Pediatr Res. 2025 Jan 23. doi: 10.1038/s41390-025-03872-x. Online ahead of print.PMID: 39849115
----
The transcript of today's episode can be found below 👇
[00:00:00.680] - Ben Courchia
Hello, everybody. Welcome back to the Incubator podcast. It is Sunday. We are back with another episode of Journal Club. We're here in the studio with Daphna. Daphna, how's it going this morning?
[00:00:10.760] - Daphna Barbeau
It's good. I don't know, I can't look at all of my things, now that we're in the studio.
[00:00:15.340] - Ben Courchia
What do you mean? You have your computer in front of you.
[00:00:17.650] - Daphna Barbeau
I know, but you get to see all the details! It's nice to be in the studio. We don't get to use the studio together that often, but this is fun.
[00:00:28.580] - Ben Courchia
Yeah, this is actually a lot of fun. What do we have to talk about today? I wanted to bring some attention to our friends at the BPD Collaborative. I think that it's a great collaborative to be a part of, if you're interested in chronic lung disease. The annual BPD meeting is coming up and it's open to all. The symposium will be held on Wednesday afternoon, March 5th, and all day Thursday March 6th. It will be taking place at the Hilton Philadelphia at Penn's Lending. The agenda is available on the BPD Collaborative website. It includes exciting presentations from their ad hoc groups. There's breakout times for all ad hoc groups to meet individually. There's updates from the data committee, presentations from international members, poster sessions, which will include work from collaborative sites, and an optional dinner. I'm sure people will take them up on the optional dinner. They also have a way to register using Event Bright. We'll share the link in the Show Notes. If you are interested, go check it out. It's a great group. If you're a fellow, consider looking into the BPD Collaborative because you will make very meaningful connections.
[00:02:01.660] - Daphna Barbeau
They're a fun bunch. You learn a lot. They're really accessible. I think it's a great way for people to get interested, either from a training level in education or research, or just being better at managing BPD babies. I guess we can reiterate our great interview with Dr. Steinhorn last week, which was also highlighting the Cool Topics Conference, coming up March 5th-8th, which includes QI Palooza and a neonatal POCUS workshop.
[00:02:44.180] - Ben Courchia
We had a coupon code, so go check out the episode page. Her interview is featured on the front page of our website, and so you'll be able to get a nice discount on registration, courtesy of the California Association of the Neonatologists, who's organizing this conference.
Let's begin with Journal Club today, we have a lot of articles. We have an EBNEO segment that will be coming up in a few minutes, but I will get us started if that's okay. There's an interesting article that came out in JAMA Peds called “Apnea After 2-Month Vaccinations in Hospitalized Preterm Infants: A Randomized Clinical Trial.” I think this is a very interesting article. It feels like a news article; this is something that people will be talking about. I think that there's some interesting data in the background that to me was quite appalling. Most notably, half of infants discharged from NICUs after a length of stay of 60 days or longer are fully vaccinated. Half of the babies who stay in the NICU more than two months are actually vaccinated!
[00:04:01.480] - Daphna Barbeau
We're missing it, huh?
[00:04:03.470] - Ben Courchia
Yeah, and preterm infants have been shown to be under-immunized at older ages anyway. They're mentioning that there are several factors that impact under-immunization in this population of preterm infants, but the perceived risk of apnea and other complications after vaccination in preterm infants likely contributes. This risk is noted as a warning and precaution in vaccine packages for routinely administered infant vaccines. Doing a little bit of a review, they mention that to their knowledge, there's only one randomized clinical trial that was conducted in the early 2000’s that showed no difference in the apneic events between the immunized and the control group. What they're really trying to do in this prospective randomized open-label clinical trial is to compare the proportion of preterm infants born before 33 weeks of gestation with clinically important adverse events, including apnea in the 48 hours after a simultaneous administration of two month vaccines, versus 48 hours after randomization to no vaccines (simultaneous meaning that they're giving all vaccines in one day). In terms of the design, I think it was a pretty well-designed study. It was a CDC-sponsored study, and it was a multi-center, prospective, randomized, open-label clinical trial to assess the risk of apnea in preterm infants following routine vaccination between 2018 and 2021 in three NICUs around the United States.
The participants were infants of gestational age less than 33 weeks that were of postnatal age between 6 weeks and 12 weeks. You had to be older than two months for the vaccination, and must have been hospitalized from birth. They had to have no plans of being discharged within the next 60 hours after the vaccination; they wanted to be able to monitor these infants. They were randomized 1:1 to vaccinated or unvaccinated groups. The study intervention that eligible infants were randomized to the vaccination group received the 13-valent pneumococcal conjugate vaccine; the diphtheria, tetanus toxoid, and acellular pertussis (the DTAP); hepatitis B; inactivated polio; and HiB within 12 hours of randomization.
I want to make sure that people understand that they did not encourage having a control group of unvaccinated. They used the babies who were not going to receive vaccination as the control group. The infants in the unvaccinated groups, while they did not receive vaccine during the study period of these 60 hours, were encouraged to be vaccinated in a timely manner subsequent to that study participation.
The primary outcome for this study is the proportion of infants with one or more apneic events as determined by manual review of the monitoring strips during the 48 hours after vaccination. The secondary outcomes were, within the 48 hour monitoring period: the average number of apneic episodes, the average duration of these apneic episodes, the proportion of infants requiring any increase in respiratory support, the presence of one or more severe cardiorespiratory events, and receiving positive pressure ventilation. They had some pre-specified exploratory outcomes, including fever, hypothermia requiring blood culture or antibiotics during sepsis evaluation, and significant adverse events in the 48 hours after.
A total of 223 infants were randomized: 107 in the vaccinated group, 116 in the unvaccinated group. Overall, 113 infants out of 222 (51%) received respiratory support (83 out of 130 infants less than 37 weeks postmenstrual age, and 30 out of 92 who were 37 weeks or older). 26% received caffeine during the 48-hour monitoring period (22% in the vaccinated group, and 28% in the unvaccinated group).
The primary analysis of outcome in the 48-hour monitoring period revealed that the proportion of infants with one or more apneic events was 24% in the vaccinated group, compared to only 10% in the unvaccinated group, with an adjusted OR of 2.7, a confidence interval of 1.27-5.73, and a p-value of 0. 01. Now, let's look at the gestational age. Among infants who were 37 weeks postmenstrual age or older, the proportion of infants with one or more apneic events was 30% in the vaccinated group, compared to only 6% in the unvaccinated group. Among those who were less than 37 weeks PMA, the proportion of one or more apneic events was 20% in the vaccinated group, compared to only 13% in the unvaccinated group. There was no statistically significant difference in the mean number of the apneic episodes between the vaccinated and unvaccinated group. There was no statistical difference between the mean duration of the apneic episodes between the two groups. There were no severe adverse events reported in either group during the 48-hour monitoring period. The proportion of other secondary outcomes and exploratory outcomes were also not very different, including length of stay.
The other exploratory outcome that I thought was interesting was the number of apneas between 48 hours and 14 days after vaccination. Of the 105 vaccinated infants monitored between 48 hours and 14 days after vaccination (with outcomes available), 23% experienced apnea during the 48-hour monitoring period. And of these, eight (which is a very small number, but represents about 32% of that group) experienced apnea between 2 and 14 days after vaccination. This is compared with 18% (which is much lower, but again, we're talking about very small numbers here) who did not have apnea in the first 48-hour period. It’s almost as if having an apnea in those first two days post-vaccine sets you up. Again, very small numbers. Apnea during the first 48-hour was not associated with subsequent increase in respiratory support, receipt of positive pressure ventilation, or blood culture/antibiotics given for sepsis evaluation. Of the 25% that were discharged within two weeks of vaccination, six of these infants had apnea in the 48-hour monitoring period, and of these infants, two were readmitted.
It is difficult to make general conclusions out of this. The conclusions of the authors is that based on their primary outcome, their findings of increased apneic events following vaccination in preterm infants are consistent with previous observational studies and the warnings present in vaccine packages inserts. Apneic episodes were brief. No infants experienced a serious adverse event during the 48-hour period, and there were no statistically significant differences between the groups for other secondary exploratory measures. The study supports the current Advisory Committee on Immunization Practices (ACIP) recommendations to vaccinate hospitalized preterm infants at two months of chronological age in most situations. Obviously, there are some kids who have NEC; we've all been there, maybe you want to take a break for that. Using the same contraindication and precautions as for full-term infants, we encourage neonatal clinicians to continue to provide evidence-based anticipatory guidance about post-vaccination apnea risk. Thoughts?
[00:12:11.970] - Daphna Barbeau
I have three thoughts. The first is that I would have liked to know the baby's day of life. They told us the correct gestational ages and that they were at least two months old. There have been some studies that show if you even wait 10 days, that events are less between the 60 days and 70 days. That's something I would have liked to know.
My two and three thoughts are connected. There's a lot of vaccine hesitancy, and so I think it's important to have this information. This is why it would be a great news segment. Like you said, it’s good to be aware of this information because when parents say, “What about the study about apnea?” You say, “It's fine, it’s no big deal.” Parents really lose trust in us. We really have to know the data, and say that there's increased apnea. There's no doubt, the study is very clear about that. But that was the only real measure that was different. You might have apnea, but then you'll be vaccinated. It also may help us encourage families to get the vaccination while they're inpatient instead of outpatient.
[00:13:31.670] - Ben Courchia
Partial answer to your question, in Table 1, they do have the postnatal age of randomization, and then they get vaccinated within 12 hours of that. It was about 8.7 weeks. But again, the full question you had was, if you wait a bit longer, did the babies improve?
[00:13:52.310] - Daphna Barbeau
They didn't use a continuous variable.
[00:13:54.230] - Ben Courchia
It's not big enough of a study to do that, in my opinion. I think it's a very interesting study. Like you said, I think that if we are upfront with the families and we say, yes, we know about this, and this is the reason why we're monitoring for two days and so on, it might create less hesitancy. Also, maybe don't discharge the afternoon after the vaccine in case that's something you were thinking about.
[00:14:17.840] - Daphna Barbeau
If the baby has apnea in those first 48 hours, then they're definitely staying, right? For the next week, in my opinion.
I had the paper, “Effects of Pregnancy-Induced Hypertension on Early-Onset Neonatal Thrombocytopenia.” This is something we see all the time. We see thrombocytopenia and think it's related to PIH, but it's nice to have this additional data. This is in BMC Pregnancy and Childbirth and the study is coming out of China. Basically, they analyze the data of neonates born to mothers with gestational hypertension and preeclampsia. They wanted to differentiate what thrombocytopenia looked like between those two severities of illness. They also had a third group of healthy mothers. The neonates were admitted to hospital within 72 hours of birth and they compared the incidence of early onset thrombocytopenia among the three groups.
China has a very rigorous prenatal care system. They bring in these mothers early, as early as they can, into prenatal care. The routine visit frequency is much like it is here in the States. They have a very rigorous algorithm for following elevated blood pressures. That was listed in this study, but it's basically their routine parameters for monitoring these mothers. They did exclude a number of neonates whose mothers had other diseases: diabetes, thyroid disease, hepatitis B infected before the 20th week of pregnancy, history of thrombocytopenia, history of chronic hypertension, chronic renal disease, significant heart conditions, PE, and DIC. They also excluded neonates who had history of anal atresia, trisomy 21, Tetralogy of Fallot, hemopathy, and those who passed away. They basically included any neonate who had a venous blood collection of platelet count within 72 hours after birth to diagnose neonatal thrombocytopenia, and they used the cutoff of less than 150,000.
Total thrombocytopenia in the whole group was about 14%; that feels about right. The gestational age of mothers who delivered neonates with thrombocytopenia range from 29+1 to 40+4. This did not include a lot of very preterm infants. This value did not significantly differ between neonates born to mothers with PIH or those born to healthy mothers. There were some differences in the two groups. Which babies had thrombocytopenia? What were the other maternal pregnancy risk factors? For premature rupture of membranes, that was seen in 14.5% of infants with thrombocytopenia, compared to 25% of infants without thrombocytopenia. So premature rupture of membranes was more commonly seen, actually, in the babies who didn't have thrombocytopenia. The C-section rate was 74%, very high, in infants with thrombocytopenia, compared to only 62% in infants without thrombocytopenia. Intrauterine growth restriction was seen in 20% of infants with thrombocytopenia, only 5% of infants without thrombocytopenia. Perinatal infection was 6.6% with thrombocytopenia and 1.9% without. There was no major difference in thrombocytopenia rates due to maternal age, weight gain during pregnancy, placental hypoperfusion, placental abruption, fetal distress, or oligohydramnios.
Then they wanted to look at the groups specifically. They had 220 neonates with mothers who had PIH, and they had 318 neonates born to healthy mothers. Of the neonates born to mothers with pregnancy-induced hypertension, 51 had a platelet count less than 150,000 (23%), and of the neonates born to healthy mothers, 25 had a platelet count less than 150,000 (8%). But pregnancy-induced hypertension consisted of both gestational hypertension, preeclampsia, and eclampsia. They wanted to look at the groups more closely. The incidence of early onset thrombocytopenia in neonates born to mothers with preeclampsia was significantly higher, even to those mothers with gestational hypertension, and certainly to healthy moms. It was 30% in those moms with preeclampsia, 13% in those with gestational hypertension, and 8% in healthy mothers. However, the incidence of moderate versus severe thrombocytopenia did not differ significantly among the three groups.
It's obvious that maternal gestational hypertension and certainly preeclampsia increased the risk of early neonatal thrombocytopenia. They looked at some other parameters. Compared with neonates with a birth weight greater than 1500 grams, those with a birth weight less than 1500 grams were also at a high risk of early onset neonatal thrombocytopenia, an odds ratio of 9.97. SGA neonates were at an increased risk of early onset thrombocytopenia compared to non-SGA neonates, with an odds ratio of 2.64. Neonatal hypoglycemia also showed an association, with an odds ratio of 2.17 of increasing the risk of early onset neonatal thrombocytopenia. Finally, compared with neonates without perinatal infection, neonates with perinatal infection were at a higher risk of early onset neonatal thrombocytopenia, an odds ratio of 2.64.
Then they wanted to look at bleeding events. A total of 63 neonates, 11.7%, suffered bleeding. 10.4% were born to PIH mothers. But interestingly, the incidence of bleeding did not significantly differ between neonates born to mothers with PIH and those born to healthy mothers. Bleeding occurred in 32% of infants with thrombocytopenia, of which 23% were born to mothers with PIH. The main manifestations of hemorrhage in neonates with thrombocytopenia were mild intracranial bleeding and scalp hematoma, but no major bleeding. This is for babies with thrombocytopenia born to healthy mothers. In contrast, bleeding events in neonates with thrombocytopenia of mothers with PIH, mostly manifested as GI hemorrhage (25%), pulmonary hemorrhage (16%), and intracranial hemorrhage (91%). There were nine cases of minor sub-ependymal hemorrhage and two cases of severe IVH who required transfer. The incidence of intracranial hemorrhage was significantly different between the neonates with thrombocytopenia born to mothers with PIH and healthy mothers. There was also a difference in scalp hematomas.
In conclusion, PIH, specifically preeclampsia, predisposes a greater risk of neonatal thrombocytopenia. They did show some differences in bleeding-related events. I thought it was nice to see these other associations with thrombocytopenia tht might help explain some of the lab values we find.
[00:22:40.540] - Ben Courchia
Daphna, question for you then. Figure 1 of this article is quite interesting, showing this linear relationship between the percentage of babies who will have thrombocytopenia (less than 150), based on whether the mother is normotensive, has gestational hypertension, and/or preeclampsia. You see that step-by-step progression. What are you making of this data in terms of translating that into clinical practice? How are you taking that information to your next call?
[00:23:09.520] - Daphna Barbeau
I'm not sure if it challenges anything that we already thought. Especially in our population, we have a lot of mothers with gestational hypertension, but not preeclampsia. Sometimes that's not always well-communicated. I think we know the moms that have preeclampsia, we have a lot of moms who deliver early with preeclampsia. But sometimes gestational hypertension, that's not quite preeclampsia, is not always well-communicated, and it may still explain some of the thrombocytopenia we see.
[00:23:45.960] - Ben Courchia
I think the problem also is the fact that when babies are born, they have thrombocytopenia. The question then becomes, this was a baby that was delivered “for maternal reasons.” Will your threshold for accepting low platelets be different? If it's just plain “gestational hypertension,” technically there's only 10% of them who should have low platelets. We tend to chalk off these low platelets like, oh the mother had blood pressure issues. But do you now say, huh, this mother didn't have preeclampsia but had gestational hypertension, so now I'm less tolerant of a kid with a platelet count of 80,000.
[00:24:33.070] - Daphna Barbeau
I think I might still sit on it a while, in an otherwise well-appearing baby.
[00:24:36.880] - Ben Courchia
Fair enough. Great.
[00:24:38.290] - Daphna Barbeau
But it's nice to have the details.
[00:24:42.550] - Ben Courchia
Yeah, I was thinking, do I need to adjust my threshold? But now that you say no, then no.
[00:24:54.840] - Daphna Barbeau
That doesn't mean you don't have to adjust your threshold!
[00:24:58.230] - Ben Courchia
I'll follow your lead.
I'm going to try to do a very quick paper, not a small paper, but a very quick paper before we go to EBNEO. This was an important study. It's called the “Outcomes of Extremely Preterm Infants Who Participated in a Randomized Trial of Dopamine for Treatment of Hypotension (the HIP trial) at 2 years Corrected Age.” This is published in the Archives of Disease in Childhood. The first author is Neil Marlowe and the second author is Keith Barrington, who spoke about the HIP trial a lot during his interview in January. So go check this out.
The premise of the HIP trial is that there's uncertainty among clinicians as to when and how to treat very preterm infants with low blood pressure during the first few days after birth. That's what really the HIP trial was trying to address. It was an international multi-center randomized trial that compared dopamine to placebo for the treatment of hypotension in extremely preterm infants between 2015 and 2017. The primary outcome was survival to 36 weeks without severe brain injury on ultrasound. Now, we talked about this with Dr. Barrington during the interview, but the trial was closed early because of challenges in recruitment. It was very, very difficult to recruit patients for this trial, for a variety of reasons. I think from a methodological standpoint, it's a great example of why in neonatology, for these particular questions, there are some shortcomings for traditional trial design, and that may be more of a cluster-randomized trial, where various hospitals are assigned to various interventions, might have been more successful. This trial really posed and raised a lot of these questions and brought them to the forefront. So in the shortened trial, they found that the primary outcome of survival without brain injury was reached in a similar number of patients between the two groups: 62% in dopamine, 69% in the placebo group. Additional treatments, however, were used less frequently in the dopamine group: 66% of the kids in the placebo group had to get other things to address the issue of hypotension, while only 38% in the dopamine group needed that.
Now, the extremely preterm infants who participated in this trial were followed up until two years corrected age. They undertook the follow-up of trial participants to determine survival without neurologic impairment or developmental delay as a co-primary outcome. So the HIP trial recruited infants from 10 centers from Europe and Canada between 2015 and 2017. By the way, when the trial has issues recruiting, and you're talking about an international multi-center trial that involves 10 large centers, just let that sit for a minute…these are centers that are definitely not low volume. We'll talk about volume in another paper later this episode, but it's so interesting. After two years of enrollment, they had some issues with funding and enrollment rate. The surviving infants and their parents were invited to follow up as close as possible to 24 months of age, corrected for prematurity. They were assessed by a psychologist or pediatrician using the Bayley-III. They used a score of 100; babies who scored 85 or below were considered delayed, less than 70 were considered severely delayed. Parent reporting was also used to determine hearing impairment and vision problems, as previously defined. The outcome of neurologic impairment and developmental delay was defined as any cerebral palsy, Bayley-III composite cognitive language or motor score less than 85, or moderate or severe hearing or vision impairment. I think this is maybe the last bit that we're going to hear from this particular cohort. And so the status at median follow-up age of 28 months was determined in 27 survivors in the dopamine group, 28 survivors in the placebo group. You're looking at about 50 patients. I'm still floored by this huge trial, and you just get 50 babies. It's crazy.
[00:29:52.440] - Daphna Barbeau
It underscores how hard recruitment is.
[00:29:58.190] - Ben Courchia
So 48% (13 out of 27) in the dopamine group survived without neurologic impairment or developmental delay, compared to 25% in the placebo arm. There were also similar proportions in each group, including the babies that died. Of the surviving infants evaluated, eight in the dopamine group (38%) had neurologic impairment or developmental delay, compared with 13 in the placebo group. Four infants in the dopamine group and three in the placebo group had cerebral palsy, with severe cerebral palsy present in one and two infants, respectively. The motor outcomes, respiratory symptoms, need for medication, or hospital readmission after discharge were similar between the two groups.
The babies who were on dopamine did better. Dopamine has been on the chopping block for so many years now. The authors, by the way, are not shy about outlining the limitations of a trial that was stopped early. They say, “Our results, therefore, increase the possibility that dopamine therapy for hypotension is beneficial for longer term outcomes, but are not definitive.” It's true with this data, it gives dopamine a little bit more of a leg to stand on. They also mention that another possible interpretation is that the rescue treatment that they used in the placebo group, which was epinephrine, maybe the frequent use [of epinephrine] in the placebo group resulted in the adverse events.
The conclusion is that the limitations of this small study mean that it cannot definitively answer the question, does administration of inotrope improve long-term outcomes of extremely preterm infants with low blood pressure over the first few days after birth? This question requires further evaluation.
[00:32:07.360] - Daphna Barbeau
But how are we going to do it?
[00:32:08.840] - Ben Courchia
I think you have to probably use a more clever study design where you can actually circumvent the issues that they run into. Which, by the way, if you're a young investigator, it's like they outline the roadmap. You have to tweak this one thing, and if you can make the methodology work, then you have this beautiful trial just waiting to be done. Go for it.
[00:32:26.550] - Daphna Barbeau
But you got to recruit the patients.
[00:32:28.730] - Ben Courchia
In my opinion, the problem is the placebo. You're going to have a hard time convincing critical care physicians to tell the parents, “The blood pressure is technically low, but we're going to watch it.” But maybe we can tailor different approaches. I think that maybe, I'm just spitballing here, but you take Iowa as an example and say we're going to have one center that's going to do a hemodynamic approach. We're going to have a center that'll say, if the blood pressure is lower than your gestational age, we give you dopamine. I think the placebo one is difficult. Or you take some preclinical data and decide to define blood pressure a bit differently. Maybe we're going to tolerate the blood pressure to go below your gestational age, as long as your systolic does X and your diastolic is above Y, or something.
[00:33:25.950] - Daphna Barbeau
That's the underpinnings of all hypotension research - what is hypotension? Until we can really agree on that, it's going to be hard to prove any of these things.
[00:33:37.940] - Ben Courchia
Correct. All right, we're going to take a quick break. We'll have a little ad, and then we will come back with an EBNEO segment featuring Dr. Sara DeMauro. See you in a little bit.
[00:34:31.840] - Ben Courchia
We are back this Sunday for another episode of our EBNEO series, where basically a special manuscript is identified by the EBNEO community as one to most likely impact neonatal care. We are joined by two members of the EBNEO team. We are joined by Dr. Srishti Jayakumar. Srishti, thank you so much for joining us.
[00:34:52.660] - Srishti Jayakumar
Thank you so much for having me, Ben.
[00:34:54.440] - Ben Courchia
You are a third-year fellow at Johns Hopkins in Baltimore, so almost an attending. You were mentored through this process by Dr. Sara DeMauro, who is an associate Professor of Pediatrics at the University of Pennsylvania Perelman School of Medicine and the Children's Hospital of Philadelphia, Thank you, and welcome to the show.
[00:35:17.340] - Sara DeMauro
Thanks so much for having me, Ben.
[00:35:18.900] - Ben Courchia
The pleasure is all ours. You guys have picked a super interesting study that came out in JAMA Network Open. The study was published a couple of weeks back, and it's called “Parent-Guided Developmental Intervention for Infants With Very Low Birth Weight: A Randomized Clinical Trial.” First author is Rita Silveira. Srishti, I'm wondering if you can take us through the study.
[00:35:45.530] - Srishti Jayakumar
Yeah, I'd be happy to. This is a very interesting study. It was designed as a randomized control trial, and it was set in Brazil at the Clinica de Porto Alegre. The idea was that the authors wanted to look at infants that were born either at a gestational age of less than 32 weeks or less than 1,500 grams, and who resided within 40 kilometers of the hospital. The idea was to see whether training parents to administer an enhanced developmental intervention would improve neurodevelopmental outcomes and motor outcomes at 18 months of age. They recruited parents at about 48 hours of life for an intervention that began at day seven. While the infants were still in the neonatal ICU, they had their usual standard of care, which included support for lactation, kangaroo care, and routine developmental interventions. But the enhanced intervention consisted of infant massage, visual stimulation, and tactile auditory stimulation that was administered in a specific fashion and overseen by developmental therapists.
The a priori stated objective was to have at least 100 participants, which they did. They had participants and clinicians that were caring for babies in the NICU that were, of course, aware of this randomization. But then the ultimate study assessments were all blinded. They recruited infants and then followed them for an 18-month period, between 2017 and 2022. They ended up not including any babies that had things like infections, inborn errors of metabolism, or major congenital malformations.
What they really wanted to look at for their primary outcome was cognitive scores on the Bayley-III. The Bayley-III was administered at 18 months, and the assessment was blinded, which was excellent. They also used a number of different measures to look at secondary outcomes. They looked at language and motor skills. They looked at the Alberta Infant Motor skills. They looked at assessments of the home environments of parental practice and knowledge, and also looked at enrichment in the home environment as well. Ultimately, they were able to recruit 100 participants. They started off with 207 that they assessed, came down to 138, of whom they lost 38 to follow up and/or passed away.
Their results are fairly interesting. Their primary outcome, the cognitive score in the enhanced developmental intervention group, was in fact higher than the scores seen in their usual care group, and these were statistically significant results. But they also reported increased language and motor scores in their secondary outcomes, and interestingly saw a much greater increase in their motor scores across both categories. They also talked about the interaction rating scale. This was a measure of parent-infant interaction, and this was meant to reflect whether the intervention had had any effect on the quality of parent-infant interactions. They actually saw that it was significantly higher on the group that had received the intervention, compared to those that had not. Very importantly, they found that they didn't actually have differences in the enrichment opportunities at home, parental care practices in general, or even the parents' knowledge regarding infant development - that was pretty well distributed in both groups. Overall, the takeaway is that they did find that this parent-guided intervention for preterm and/or very low birth weight infants in this LMIC setting improved cognitive, language, and motor outcomes at 18 months, specifically a significant increase in motor outcomes compared to the others.
[00:39:54.370] - Ben Courchia
Yeah, I think it's so interesting because we are familiar with the idea of parent coaching and trying to really get parents involved in the care of their infants early on. But it was interesting to see this rigorously implemented so early on in the NICU stay, with a direct sight onto the long term outcomes. I think it was eye-opening for me to see parents being involved in that care on postnatal day seven with the idea of, let's really try to see what these outcomes are going to be like at 18 months. Even though everything we do in the NICU is targeted towards neurodevelopment, this was outright black-and-white, this is our goal. I thought that was very interesting. Sara, I'm curious to hear your thoughts on the study and the study design.
[00:40:39.010] - Sara DeMauro
I completely agree with you, Ben. Parents are a captive audience in our unit. The study bridges the inpatient to outpatient stay, and it capitalizes on the fact that the parents are here in front of us in the unit. It really empowers them to start intervening early and practicing the skills that were taught to them. This really is an interesting approach to supporting developmental care, on breaking down that barrier between inpatient and outpatient. It empowers parents to start early, to learn about this intervention, and to provide this developmental care to their infants in a way that now has been demonstrated in a randomized trial to improve outcomes through 18 months.
[00:41:24.260] - Ben Courchia
Yeah, I think this is very interesting. I wanted to spend a little bit of time diving into your commentary. For the people who are first listeners of our EBNEO segment, all the EBNEO commentaries are published in the Acta Pediatrica Journal, so you can find that commentary there when it is released. It's interesting to me because from your commentary, what I'm taking away is that this study is framed as a study performed in LMIC country. But in truth, they do a lot of things very, very well, sometimes even much better than other places where you wouldn't consider these places to be LMIC. I'm just curious if you could tell us a little bit more about that.
[00:42:11.160] - Srishti Jayakumar
One of the things that I really liked about the study was the setting that many of the previous studies that we've read are published in traditionally high-income or high-resource settings. As we know, LMICs are a large heterogeneous group, and there's huge variation in what the follow-up looks like and what the availability of resources look like. To have a study in a dedicated LMIC setting, I think, is very, very helpful. Even looking at what their usual care looks like for follow-up, they talk about monthly follow-up for the first six months, which is significantly higher than we would do here. We do follow-up every three months in the first six months of life, and then space that out depending on how the infants are doing. Even being able to have access to constant developmental therapists and guided interventions that only in the NICU, I think was commendable.
[00:43:09.600] - Ben Courchia
Can you clarify that? The control group gets monthly visits in the clinic, and then the intervention group gets the same plus additional visits that are being done at home, if I understand correctly.
[00:43:24.630] - Srishti Jayakumar
Correct. The big difference for the enhanced development intervention group was that they had 10 additional visits by this multidisciplinary team to review how the intervention was being done, but then also get this real-time feedback to make changes to what was being administered. The usual care group got six visits in six months, and then was spaced out over the next six months. Then the following year or two, up to 18 months of age, they were seen with regular frequency.
[00:43:54.360] - Sara DeMauro
It's an extraordinary amount of developmental follow-up care, even in the usual care group, which speaks to the quality of follow-up care in Brazil in general. The authors sell this as important because it's based in a low or middle income country, which is true. But from that perspective, they really have considerable resources to provide follow-up care more than you'll see in many high-income countries. The interesting thing is that, if anything, it probably biases them towards the null. The fact that both groups were receiving quite a bit of follow-up care after discharge and very close surveillance, but yet the addition of the parent-administered intervention was still able to yield benefits at 18 months, tells you that potentially in settings where routine follow-up care is not as rigorous, you might even see a bigger benefit of an intervention like this.
[00:44:51.930] - Ben Courchia
That's a great segue into some of the outcomes that are presented, because the first thing I wanted to ask you was, what are your thoughts on the developmental tool used to assess these babies? They made the decision to use the Bayley-III. Any thoughts on that?
[00:45:11.430] - Srishti Jayakumar
That's a great question. At the time that the study was designed, several institutes were using the Bayley-III, and it was one of the best tools available. But significant research has since then has shown that the Bayley-III, in a sense, overestimates development. There's a good chance that we are missing developmental delay by using that tool. I think the one thing that mitigates that effect in this study is that the same tool was administered to both groups, and so the degree of overestimation was likely the same in both groups.
[00:45:47.600] - Ben Courchia
Sara, any thoughts on the use of the Bayley, especially when we're talking about using it in different languages? Are there any implications for research purposes in how we look at data?
[00:46:00.130] - Sara DeMauro
The Bayley is a very interesting tool. I think there's a considerable amount of controversy around the Bayley. The strengths of this study are the assessments at multiple time points. We haven't discussed that yet, but they actually assess the infants at four time points over the first 18 months of life. One of the strengths of the Bayley is that it's quite good for establishing trajectories of development. The fact that they could repeat it four times in all of the infants, demonstrate stability of their findings over time and a stable difference between the groups over time, is very reassuring. In addition, there was a secondary outcome using the Ames, which is a different motor assessment, which, if anything, showed a bigger benefit of the intervention, even than the motor score on the Bayley. I think we can be reassured by the fact that they did use a different instrument to assess motor outcomes, which also showed that the intervention is quite beneficial. The Bayley has its pros and its cons, but I think in this context, it was the best thing available at time. I don't think we should look at the Bayley results of these children necessarily in isolation without understanding that the overall scores may be somewhat overestimated. But I think the message of the paper remains clear that the benefit of the intervention was quite strong.
[00:47:20.770] - Ben Courchia
What do you guys make of the breakdown of the Bayley scores in terms of the improvements that were noted by the authors in language and motor, but somehow not significantly different in cognitive. Do you think this has to do with the discussion we had earlier about the control group being so well taken care of?
[00:47:40.880] - Srishti Jayakumar
This was another thing that we actually talked about extensively, that some of these results are actually contrary to what we've seen in larger studies and in other systematic reviews and meta-analyses. A Cochrane review actually talked about this, that we actually expect great differences in cognitive outcomes compared to motor outcomes, contrary to what the study described. The authors talk about the mechanism for some of these changes. It may have been that the families found it easier to administer motor interventions. Perhaps they had a greater positive feedback from the infants as they noted these things developing sooner, and so there was more immediate feedback. It's not entirely clear whether that was the case, but it is certainly interesting to see that the motor improvements were greater than the cognitive improvements.
[00:48:39.140] - Sara DeMauro
It just goes to show that developmental interventions are extraordinarily nuanced. Some may provide early cognitive benefit, while others may provide early motor benefit. We don't know the mechanism why this particular intervention seems to provide more motor benefit, but they don't doubt that that's real. I think it just demonstrates that we need to continue to work to understand the best way to provide comprehensive developmental care to high-risk infants.
[00:49:09.950] - Ben Courchia
We had this discussion with Susan Hintz on the podcast some time ago, where the effects that we're seeing in motor improvement are probably tremendously impactful. Both her and Barbara Schmidt were mentioning how it's difficult to tease apart the cognitive improvements we see early on and what to make of these things. But the motor improvement is such an impactful outcome that it makes me very optimistic when I read this study and see these outcomes for sure.
[00:49:40.580] - Sara DeMauro
Based on some of Barbara's work, there's reason to believe that motor outcomes may be more sustained over time than cognitive outcomes, which tend to be more strongly influenced by socioeconomic context. But that said, at least through 18 months, there was a small but significant benefit with this intervention.
[00:50:00.340] - Ben Courchia
I was really happy to see the investigators use the interaction rating scale. How does the nature of the relationship between a parent and a child evolve as the parent takes on the role of being more than just a parent, but also a caregiver? Does that pollute the bonding between the parent and his or her infant? It was surprising to me to see that this actually had beneficial effects. I'm just curious about how you interpreted this data and what you make of it, in terms of moving forward in this space of research.
[00:50:39.710] - Srishti Jayakumar
I actually thought there were very, very happening results to see as well. I think one of the greatest experiences the families and the NICUs have is that it's difficult to establish that type of early bonding with infants, especially medically complex infants or those that are born very premature. I think being able to provide an intervention that facilitates that, not just in the short run, but over this 12-18 month period can be incredibly helpful. I interpreted this intervention as then being beneficial in more than one way and certainly improving neurodevelopmental outcomes, but really also improving that fundamental parent-child relationship.
[00:51:23.160] - Ben Courchia
My last question for you both is, looking at this study, there's definitely two phases to this investigation. There's the interventions that are taking place in the NICU, and then the interventions taking place outside the NICU. We know that for us, we are all three practicing in the United States. We have inherent limitations to outpatient follow-up that other countries who have more socialized medical healthcare systems do not have. But irrespective of these limitations, what are your takeaways for your own practice?
[00:51:54.480] - Srishti Jayakumar
I think my big takeaway is that there's a lot of power in parent-directed and in parent-administered interventions. I think certainly, regardless of the context, being able to bring them into the fold in a timely fashion can be very, very helpful. I think this study sets the stage for many important studies moving forward, but certainly for us to try and include some of these evidence-based developmental interventions into our follow-up process as well.
[00:52:21.740] - Ben Courchia
Yeah, I think you mentioned something in passing here that was so important. You said in a timely fashion, because my unit is guilty of that as well. Very often, we involve the parents only once we are ready to start disposition planning, and it's already way too late. I think what this study highlights is that you can start this after seven days of life. Sara, any thoughts on this particular question?
[00:52:53.390] - Sara DeMauro
Yes, I completely agree with you, Ben. The parents began their educational journey in this trial at seven days, when their infants were seven days postnatal age. I mean, many of these infants were quite young and quite small at that point. But parents in that situation often feel quite powerless. These data tell us and give us a reason to tell them that they, in fact, are quite powerful when it comes to determining their infant's future. We can begin to teach them, support them, and help them learn how to encourage development, even starting at that early time point, and that it will make a difference for their children.
[00:53:32.920] - Ben Courchia
We had some of the team from Brazil come to the Delphi conference, and we had them on the podcast. One of the other aspects that is hinted throughout the study, but is not really mentioned, because I don't think this is something they even think about, is parental leave. The opportunity to have parents available at the bedside to be involved in this intervention is directly tied to their ability to be out of work while their child is critically ill in the NICU. No matter what, despite our best effort, it's difficult for our therapy teams and other providers in the NICU to be there late at night, when parents are out of work. It happens very often in the US. I thought to myself, as I was reading the paper, that this is another tidbit that for us in the US, it resonates quite loudly.
[00:54:21.100] - Sara DeMauro
I completely agree. We need data like these to support advocacy efforts for paid parental leave and other policy measures that will ultimately support development and health in our high-risk infants, and all infants.
[00:54:34.060] - Ben Courchia
I agree. Srishti, congratulations on this beautiful commentary. Sara, thank you so much for your help, for your work, and mentoring Srishti through this EBNEO commentary. It was a pleasure talking to you both, and will look out for more of your work coming up soon.
[00:54:49.140] - Srishti Jayakumar
Thank you so much, Ben.
[00:54:52.170] - Ben Courchia
Okay, we are back. Thank you again for the EBNEO team for participating in Journal Club on a regular basis. If you want to engage with the EBNEO team, they're very active on social media, but you can also check them out at www.EBNEO.org. Daphna, you're up next. Why are you taking us?
[00:55:11.190] - Daphna Barbeau
I know, but behind the scenes, you told me that I only get to pick one more.
[00:55:16.010] - Ben Courchia
Well, if it makes you feel any better, I have four more.
[00:55:21.660] - Daphna Barbeau
I have four more also. I know you wanted to hear about this paper. You know I love bilirubin. This was “Unbound Bilirubin and Risk of Severe Neurodevelopmental Impairment in Extremely Low Birthweight Newborns.” Lead author was Cody Arnold and senior author was David Stevenson. This is from Pediatric Research. Basically, the primary objective of this study was to quantify the association between unbound bilirubin levels and neurodevelopmental impairment. This study group was part of the multi-center study called “Aggressive versus Conservative Phototherapy for Infants with Extremely Low Birth Weight, Randomized Control Trial,” the AVSCRCT. This was conducted by the NICHD Neonatal Research Network. It provided strong evidence that aggressive phototherapy was protective against neurodevelopmental impairments in ELBW newborns. Now, this comes on the heels of changing phototherapy thresholds for term infants, not necessarily for ELBWs. This cohort was ELBWs.
Let's talk a little bit about unbound bilirubin; sometimes we need a reminder. This is a fraction of bilirubin that's not bound to albumin, hence unbound. It is felt to remediate the toxicity of bilirubin because, since it's unbound, it's more easily able to cross the blood-brain barrier. Now, we don't routinely measure unbound bilirubin. We measure conjugated bilirubin and total bilirubin. Then you might recognize that there's a difference between the conjugated, the unconjugated, and the total. Some of that probably consists of this unbound bilirubin.
What was their primary outcome? They wanted to look at the unbound bilirubin for these babies, and then they did neurodevelopmental assessments at 18 to 22 months corrected age. They created a composite primary outcome, labeled as severe neurodevelopmental impairment. This was profound impairment or severe bilateral hearing loss. Profound neurodevelopmental impairment was a composite outcome as a score on the Bayley II: mental or psychomotor developmental index of 50 or less, or a level of five for gross motor function (a level five indicates that movement requires assistance by an adult). Severe bilateral hearing loss was defined as hearing loss for which bilateral hearing aids were prescribed.
They did a single unbound bilirubin level measured in plasma from blood obtained on day 5 +/- 1 in 1,101 ELBW newborns. Of the 1100, 193 died prior to follow-up, 25 post-discharge interestingly, 67 lost to follow-up, and 24 were seen and followed, but did not have a full developmental evaluation. Of the 825 survivors that completed neurodevelopmental evaluation at 18 to 22 months corrected age, 90% had enough data to be judged about the exposure to severe neurodevelopmental impairment. Of note, they have excellent follow-up, so that's good.
The first thing the group really wanted to look at was the total bilirubin of these babies. I'll tell you, the graph makes it look like the correlation is actually pretty good. But using a Pearson's correlation coefficient, the total bilirubin to unbound bilirubin, the coefficient was 0.48, so not a really good correlation. Then they wanted to look at the unbound bilirubin Z-scores. They have a really nice graph. It is Figure 2, and it looks at the incidence of severe neurodevelopmental impairment by percentiles of unbound bilirubin Z-score. It's just a line graph. You can see that from the 1st-80th percentile of unbound bilirubin Z-scores, it has this saw-tooth pattern. Then you see that the graph jumps pretty drastically after 80. If you're above the 81st percentile of the unbound Z-score, the risk of severe neurodevelopmental impairment rises sharply from a baseline of about 14% to 33% at approximately the 85th percentile of unbound Z-scores, an inflection point that could not really be discerned before you mapped them out that way. They had a relatively flat baseline and then this sharp inflection point. They basically created these two dichotomous exposure variables, so less than or greater than the 85th percentile for Z-score. The incidence of severe neurodevelopmental impairment for unbound Z-Score percentiles 1-85 was 14%, versus 33% for Z-scores 86-100.
Then they wanted to look at the categories of severe neurodevelopmental impairment. Of 37 infants exposed to this threshold of greater than the 85th percentile Z-score that developed sensory neurodevelopmental impairment, 73% met the motor criteria, 73% met cognitive criteria, and 30% met the hearing criteria (requirement of bilateral hearing aids). That's a lot. We know that bilirubin affects hearing, so that's why it was included as one of the outcome measures. The majority of infants that met criteria for any of the three categories also met criteria for at least one other category. 82% of the infants that met hearing criteria, 78% of the infants that met the motor criteria, and 74% of the infants that met cognitive criteria also met one of the other criteria.
They also looked at the unbound bilirubin exposure by these phototherapy regimens. So using an aggressive phototherapy arm or the conservative arm. The aggressive arm used a bilirubin level of five. Now, I think conservative was eight, but I can verify that. The proportion of newborns exposed to the threshold of unbound bilirubin greater than the 85th percentile in the conservative arm was 21%, but only 8.4% of babies met this threshold in the aggressive phototherapy arm.
Then they wanted to look at mortality risk as a competing outcome. They basically showed similar mortality risk as the original paper, but that an increase of one standard deviation in unbound bilirubin level was associated with a mortality odds of 1.35. Basically, they are highlighting that unbound bilirubin is potentially an independent marker for severe neurodevelopmental impairment. It's interesting because it's not something that we routinely measure. We do think about albumin, we think about binding, we know that babies with hypoalbuminemia are at an increased risk - but we're not really measuring the unbound bilirubin. We're just hoping that having enough albumin is protective against this unbound bilirubin crossing the blood-brain barrier. It's interesting.
[01:03:53.830] - Ben Courchia
The last article that I have for us today is an article that was published in the Journal of Pediatrics. The title of the paper is “The Impact of Hospital Delivery Volumes of Newborns Born Very Preterm on Mortality and Morbidity.” First author is Ciaran Phibbs. There's a lot of interesting data in the background section, I'm going to share with you some of the tidbits. They mention how infants born at less than 32 weeks or who are very low birth weight make up over 40% of all deaths in the first year of life. There are numerous cross-sectional studies that have shown that hospital characteristics, primarily the level of the intensive care unit at the delivery hospital, is associated with the risk of mortality. Now, there's been an increase in the number of US hospitals with NICUs, particularly among lower volume, high-level NICUs that do not transfer infants based on their respiratory needs or gestational age at birth - meaning they're small, but they're still a high level of care. Most of this expansion has occurred in hospitals that are reasonably close to another hospital with a NICU, suggesting that there's improved geographic access, and that has not been the driving force for the growth of NICUs. These NICUs are not propping up because there's a dearth of NICUs or that there's a need to get that service.
Now, the potential reason for what they call this deregionalization is the lack of clear evidence of the appropriate volume threshold for high-level hospitals, and that maybe if we had an idea of what the interaction between volume and outcomes might be, we might be recalcitrant in opening more smaller NICUs. Now, a few studies they mentioned have shown that a number of very low birth weight infants or very preterm infants cared for in a NICU is associated with the risk of mortality. There are two studies that suggest that NICU volume has a stronger association with mortality than the actual level of the NICU. If you are a level 3 NICU and you have 2,000 very preterm babies a year, the fact that you have 2,000 preterm babies a year matters more than the fact that you are a level 3, because there could be a level 3 with six kids a year that are being born there. The goal of this study is really to examine the impact of different delivery volumes of infants born very preterm within a given hospital on the risk of mortality and common morbidities of premature birth, and to assess the effects of within-hospital changes in very preterm delivery volume.
They used the link to vital records patient discharge data for all in-hospital deliveries in the states of California, Michigan, Missouri, Oregon, Pennsylvania, and South Carolina from the years 1996 to 2015. The deliveries in those states together represent about one quarter of all US deliveries. That's a staggering number. The study sample was restricted to all in-hospital very preterm deliveries or fetal deaths with linked data, a valid hospital ID. They collected birth weight, gestational age. For the cases of babies that were born below 400 grams or who were born before 24 weeks, these babies were excluded because, rightfully so, of the inter-hospital variation in the decision to resuscitate infants with these characteristics, especially in the earlier years of the cohort. They also excluded infants from the cohort who had congenital anomalies with a significant risk of mortality. In terms of the outcomes, there are two outcomes for this study that matter to them. Number one is neonatal-related mortality and mortality plus major morbidity, which they defined as IVH, necrotizing enterocolitis, retinopathy of prematurity, bronchopulmonary dysplasia, which were identified through this database that they used using ICD-9 codes. The variable of interest was the number of infants born very preterm at each hospital for a given year.
I think it's a paper where you would want to read the statistical analysis. They performed a Poisson regression models with hospital fixed effects and hospital level annual very preterm delivery volume as the variable of interest to directly estimate the risk ratio for mortality or mortality plus major morbidity associated with the annual very preterm delivery volume. Now, the hospital fixed effect assesses the impact of a change in the annual very preterm volume within a given hospital over time during the period of the study. This is opposed to comparing volumes across hospital, which is biased by unobserved hospital factors more likely to occur within hospital of a given volume that are associated with the risk of either outcome. I asked ChatGPT, if I have to explain this on the podcast, what is a good way of explaining fixed effects to the audience? ChatGPT gave me a great answer. It said, consider you're trying to assess in a classroom whether giving kids homework is going to lead to them having better grades. There may be students in the class that are just naturally very, very bright, and no matter how much homework you give them, they'll get a good grade. Comparing them to another student who might be struggling is not a good idea, because they have these inherent intrinsic qualities that make them better at this particular subject. What fixed effect does is that it suggests that instead of comparing students amongst each other, you would want to compare students with themselves and say, if I give little Jimmy X amount of homework, let's see what grades he or she gets. If I give that student less homework or more homework, what is the change in the grade? Then by comparing it to themselves, you're removing these inherent issues of the nature of the subject you're studying.
They report that 24% of infants born very preterm were delivered in hospitals that had over 100 very preterm births in that year, compared to 42% of infants who were born in hospitals with less than 50 very preterm deliveries per year. Now, they're mentioning that of these 42%, 16% were born in very low volume setting, which represents less than 20 very preterm births per year.
To give you the full breakdown, they had a total cohort of about 208,261 infants. They categorized the NICUs into four different types of volume. We had the very low volume, less than 20 (representing 16% of deliveries). We had a little bit higher volume, 20 to 49 very preterm deliveries per year (26% of the deliveries were recorded there). Then we have slightly higher, 50-99 very preterm deliveries (33.9% of deliveries). Then we had 100 or more (24% of deliveries). Now, they mentioned that over the study period, the changes in the share of very preterm deliveries by NICU level and very preterm volume group was less than 2.1%. That's very little change, with the exception that the share born in hospitals with level 2 NICUs decreased by about 10%, and this is probably attributable to the fact that hospitals are upgrading their NICU levels. The share of NICU deliveries occurring in the higher volume groups remained relatively constant over the time, with just 25% occurring actually in NICUs with more than 100 deliveries of preterm per year.
The unadjusted mortality rate was 9.9% with some decline as VPT (very preterm) volume increased. The rate of mortality or serious morbidity was 37.8% and increased as volume increased, similar to other unadjusted results from studies of NICU level of care.
They have a few nice figures that I suggest you look at. They show the regression estimates and 95% confidence interval with the association of annual hospital volume and mortality (Figure 2). This shows that the relative risk of mortality, when delivering at a given hospital with annual very preterm volume within a specific category relative to a delivery hospital that has 100 or more deliveries per year, the mortality really goes up. There was a statistically significant increase in the relative risk of mortality for all very preterm volume groups less than 60 per year with a risk ratio of 1.39 for an annual volume of 1-9 very preterm infants, and a risk ratio of 1.13 for serious morbidity. There was also significantly higher risk of the combined outcome in all annual volume categories for less than 100 infants, compared with an annual volume of 100 or more, with the high risk of four hospitals that have less than 10 very preterm infants per year.
The authors conclude that they found that the annual number of infants born very preterm within a given hospital was associated with risk-adjusted outcomes, which is consistent with the results reported in previous studies. In addition, they found an association with mortality between the very preterm volume and the risk of common morbidities of premature birth, which confer long-term risk of neurodevelopmental impairment and health care use. The results show that outcomes improve as volume increases within a specific hospital. In the context of regionalized perinatal care, this implies that the progressive shift of very preterm deliveries from high-volume tertiary centers to NICUs to community hospitals with lower-volume NICUs is associated with worse outcomes at both the hospital losing volume of infants born very preterm and within the smaller volume hospitals. Their results imply that shifting infants born very preterm to create higher volume NICUs would likely improve outcomes with specific thresholds to aspire to: more than 60 for mortality, if you're just looking at mortality, or more than 100 if you're looking at morbidity and mortality.
But again, they mentioned that with any policy that encourages such a shift of infants born very preterm to higher volume centers in the US, we need to consider geographic access, trade offs between access and to any perinatal care, versus the potential gains for patients receiving care at hospitals that would see a larger number of patients annually.
The conclusions are that volume-outcome relationship within a given delivery hospital in both the risk of mortality and the combined risk of mortality and major morbidity for infants born very preterm in six US states, that relationship is well-established by this paper. The results were robust to a wide range of sensitivity analyses present in each state when examined individually. The development of perinatal health care systems should weigh these results with providing risk-appropriate access for all infants born in a given geographic region. They suggest that policies to optimize the annual number of infants born very preterm at a hospital in areas where geographic access is not an issue could reduce mortality and morbidity for very preterm infants. So the data never really sat well with me, and so that's why I wanted to bring up this editorial that was written by Dr. Lee and his colleague in the journal.
[01:16:10.960] - Daphna Barbeau
My only argument was here because not all tertiary centers are high volume centers, and not all community hospitals are low volume centers. So I take issue with that one sentence. But the rest makes sense. We know this about surgeons and surgical procedures. We know this about all procedures and the frequency with which you do them. This does make sense.
[01:16:37.500] - Ben Courchia
Let me go through a bit of that editorial that was published in the Journal of Pediatrics called “Balancing Volume and Outcomes: Insights and Cautions from a Multi-state Analysis of Care for Infants Born Very Preterm.” The first author is Siyuan Jiang, and co-author is Henry Lee. The beginning of the editorial mentions a lot of the study results that we just discussed, so I'm going to skip that. I want to pick up where the authors mentioned that in this current study, by incorporating hospital fixed effect, this study isolates the effects of volume changes within the same hospital, providing a clearer estimate of the causal relationship between delivery volume and outcomes. Second, it attempts to identify volume thresholds and offers actionable data to inform regionalization policies and designation of perinatal centers. It's something that we did highlight in our review.
However, the authors of the editorial mention that this approach introduces complexity and several aspects of the methodology and interpretation warrant cautious consideration. First, although the study identifies 60 and 100 as volume thresholds for better outcomes, these numbers require further validation. In both the study and real-world settings, large volume shifts from very low to high within a hospital are rare. The fixed effect model relies on extrapolations from small volume changes, and assumes that these trends apply to extreme contrasts. As a result, the reported risk ratios may not derive from direct calculation, and the reliance on sparse data could overestimate the mortality risk for the lowest volume categories, potentially affecting the accuracy of the estimated thresholds. The second point they're making is that the analysis assumes a monotonic effect of volume changes, regardless of whether the change is an increase or a decrease. However, the effects of these two directions may differ significantly. Outcomes may improve with increasing annual very preterm volume due to the accumulation of skills and organizational improvements. In contrast, the effect of a moderate decline in patient volume is less predictable. As existing skills and resources may mitigate adverse effects, the study does not differentiate between these opposing directions, which may bias estimated effects. The third point they're making is that significant variation exists in hospital-level characteristics and infant-level factors. The fourth point is that while fixed effect analysis provides more explicit information on the effects of volume changes within individual hospitals, implications of this approach may be limited to the hospital level. In the complex context of system-wide care regionalization, factors beyond volume, such as access to care and transport logistics, must also be considered. System-level regionalization requires a broader evaluation of these factors to ensure equitable and effective care. In the context of long-standing disparities in preterm birth and outcomes by race, ethnicity, and social factors leading to those at higher risk often receiving care at safety net hospitals, actions to shift hospitals or care may also have implications for health equity. The last point they're making is that another notable finding of this study is the fact that 42% of infants born very preterm were delivered in hospitals with annual volume of less than 50, and that 16% in settings of annual deliveries of less than 20 very preterm babies. Patterns that persisted over two decades, despite long-standing evidence supporting regionalized neonatal care, these rates of deregionalization remain high, understanding the underlying barriers to regionalization is critical for guiding feasible policy adjustments. For example, some low volume, very preterm deliveries may result from unavoidable emergency deliveries rather than hospital choice. In such cases, simply centralizing care by directing deliveries to higher volume hospitals may not benefit and could even worsen outcomes if timely access to care is compromised. In addition, given the variation of outcomes across hospitals, some low or moderate volume hospitals may achieve good outcomes, highlighting opportunities to learn and implement best practices in similar facilities. I think these points are crucial.
The authors of the editorial do give credit to the authors of the study, saying that they provided novel insight into the volume-outcome relationship for infants born very preterm and introduced a new methodological approach to this important topic. While the findings are valuable for informing policy, future research should explore barriers to regionalization, assess the applicability of volume thresholds in diverse settings, and evaluate how system-wide policies can balance accessibility, equity, and outcomes. This study marks an important step towards evidence-based policy, but highlights the complexity of translating volume outcome research into actionable strategies. After reading this editorial, I felt much better.
[01:21:39.140] - Daphna Barbeau
Yeah, that's what I was going to say. Preterm delivery is usually unexpected. I think the risk of regionalizing care was going to leave a lot of these hospitals then unprepared for a preterm delivery. That won't be good for the number of babies that end up at a hospital that has not had a preterm delivery in a very long time. That was the one point. The second point is that you and I have been in big hospitals, in little hospitals, and in medium-sized hospitals. I wonder if there's a dichotomous variable that if you reach a certain size, that you no longer can individualize care to babies and to families. I think there's probably a sweet spot that you need a certain threshold of deliveries. I hope people don't get the impression that then you should have 400 beds, because I don't think that's the right answer. No matter how many Neos you put in that unit, I think it loses the personalization of care. I think that's important to the babies we care for.
[01:23:06.890] - Ben Courchia
I wanted to do this paper last because I feel like it's going to get people talking.
[01:23:10.880] - Daphna Barbeau
You think about it. Let us know what you think.
[01:23:14.920] - Ben Courchia
All right, everybody. Thank you for joining Journal Club today. We'll see you next time.
Comments