For those who are unaware, the Silverman Lecture is an annual honour bestowed by the AAP. A neonatologist who has made. or is making, major impacts in neonatology is asked to deliver the lecture. Bill Silverman was one of the early pioneers of neonatology in the USA, and was very outspoken about the need for evidence-based practice (having performed some of the earliest RCTs in neonatal medicine himself).
This years lecturer was Annie Janvier, and, for the first time ever, a second speaker was invited to collaborate on the presentation, a mother that Annie (and I) have worked closely with, Rebecca Pearce. Rebecca and Annie have both had extremely preterm infants, Rebecca recounted, at the start of her part of her talk, that she had twins Lily and Maren at 25 weeks. Lily died within her first few days of life of early onset septicaemia, and Maren, who also had EOS, survived. Rebecca and her husband have written about their experiences with the pre-discharge MRI, and they have become valuable members of our parent resource team at Sainte Justine. One particularly moving part of her presentation was a slide which listed on one side all of the diagnoses that Maren has had from doctors (including various delays and executive function disability), in the middle was a picture of a lovely smiling 14 year old girl, and on the other side a list of her characteristics as seen by her family, “caring”, “warm”, “bubbly” “hard worker”.
Annie’s presentation recounted, a little, our course with Violette, including this briefly shown picture of me doing kangaroo care in the NICU!
It also included pictures of Violette in the PICU, intubated after her tracheal reconstruction at the age of 15. She noted that Violette did not have “BPD”, as she came out of oxygen a couple of days before she hit 36 weeks, but that she certainly had respiratory fragility, and had serious upper airway problems, receiving several courses of steroids for stridor with signs of obstruction during her first few years of life. An outcome which has rarely or never been described in neonatal follow up studies, but which is well known to paediatric ENT surgeons. (As you can see, the image is from a day or two after the surgery when “ça va mieux”, things are going better).
Annie described the work of the “Parents’ Voices” project, which was created to “Engage Parents to Co-create” definitions of what are important outcomes of neonatal care, on which Rebecca is a co-investigator .
One of Annie’s slides, which illustrates that our current measures of “NDI” don’t reflect what parents think about their children, was derived from data in one of the Parents’ Voices publications. It showed that, when we categorize infants as having No NDI, mild-moderate NDI, or severe NDI, parents often disagree with us, a disagreement that increases the more “impaired” we think their child is. Only 12% of parents whose child was categorized as sNDI agreed that their child was severely impaired, a third of parents of such children don’t find them impaired at all. On the other hand there were 25 parents who found their child severely disabled, 22 of them were classified sNDI by the CNN definition, but 3 had either no or mild NDI.
Annie listed some of the publications from the project to date (and there are more to come!):
As you can see I am co-author on some of them, being a collaborator, rather than a member of the core group. I am proud to be associated with this group who are making great strides in re-defining what outcomes are of importance to families. The CNFUN is incorporating many of the measures that parents find important, for which validated instruments exist, into the national follow up database.
Here are the 2 Silverman Lecturers on the big day:
The lecture was greeted with a standing ovation from the audience,
This was followed by a Q&A session, which was generally thoughtful and appreciative. Unfortunately it ended with a comment from a retired neonatologist who said two things: Firstly, that quality of life is subjective, so shouldn’t be used as a primary outcome. I’m not sure I agree, you can measure quality of life just as well as you can measure the incredible complexity of child development; the person posing the question led a few very high quality studies (2 of which I was involved in), in which, unfortunately, like so many others, we dichotomised the richness of children’s developmental progress into “pass” or “fail”, “NDI or “no NDI”. I would agree, however, that you shouldn’t divide quality of life into 2 categories, of good or bad QoL. But in a sense, unless an intervention improves some aspect of QoL then what value is it? Secondly, they also stated that it is important that there is link of scientific plausibility between the intervention and the outcome. That is self-evident, and indeed what the Parents’ Voice project is trying to do is to develop more appropriate plausible outcomes: Parent-Important respiratory outcomes for respiratory interventions, Parent-Important developmental and neurological concerns for Neuroprotection studies.
The outcomes that are important to parents are also, often, the outcomes that are important to society as a whole, and that often have an impact on medical resource use. To take an example from the previous post, it matters little to anybody if a baby needs a few more days of oxygen while they are hospitalised. What matters to parents, and has an impact on society and resource use, is if discharge is delayed by pulmonary concerns (such as oxygen dependence), or if the baby needs oxygen at home, or has recurrent admissions post-initial discharge, or has difficulty feeding because of tachypnoea, etc. Until we start to measure such things, we have little idea if our interventions are having an impact on outcomes that really matter.
There was a hiccup yesterday, I was writing the part 2 and it suddenly got posted as just the title, and the entire post disappeared. I had to start all over (which has never happened before, there are usually automatic and frequent manual backups, but they all vanished). So as I rewrite it, here is an interim, a link to a video about someone I appreciate enormously.
One of the true giants of neonatology, and one of its true gentlemen, Alan Jobe, was honoured at the meeting, being given the Howland Award. Alan has made a huge impact on the care of babies, his achievements are too many to fit in a short video, but this link will take you to a 7 minute video. A full description of what he has done to improve neonatal care and help babies around the world would take far longer than 7 minutes!
This trial was a modestly sized RCT of low dose diazoxide for treatment of persistent or severe hypoglycaemia in newborn infants of at least 35 weeks gestation. I say “modestly sized” as a descriptor, not a criticism, because I think this trial is potentially very important. It included 75 babies admitted to the NICU in the first week of life who had ≥3 episodes of blood sugar <2.6 mmol/L in a 48 h period, or a persisting episode <2.0 mmol/L despite gel on 2 occasions, or any episode with a blood sugar <1.2 mmol/L, AND ongoing treatment for hypoglycaemia, which meant IV dextrose, continuous or frequent feeds (≤2 hourly), inability to wean formula
Babies were randomized to either 5 mg/kg of diazoxide followed by 1.5 mg/kg q12h, or placebo.
After the first 2 doses further doses were adjusted according to the following schema, the doses are given in mL/kg as it was a masked trial, but the drug was concentrated so that 1.5 mg/kg was 0.15 mL/kg.
The primary outcome was a complicated one, evidence of stabilised glucose metabolism, as shown by: established enteral bolus feeds (either breast or bottle); no IV fluids for at least 24 hours, and all glucoses in range (2.6 to 5.4 mmol/L with at least 4 of them prior to a feed).
The primary outcome was shorter with diazoxide, but not statistically different between groups (2.8 days with diazoxide cf 3.7 days in controls). Nevertheless, there seemed to be some major advantages of being in the diazoxide group. Only 2 of the diazoxide group had a mild hypoglycaemia after the loading dose was given, whereas, in the control group, 53% had ongoing hypoglycaemia, 21% had recurrent or severe episodes, 16% had hypoglycaemia after initial stabilization, 29% had study drug dose increases, and 3 required open label diazoxide to achieve control.
They redefined the outcome variable after the study (always a dangerous thing to do) to be a bit simpler, “Time to resolution of hypoglycaemia AND full enteral bolus feeding AND no IV fluids for ≥24 hours”, and that was substantially shorter in the diazoxide group, who had no adverse effects (2 babies with blood sugar in the high normal range) and, importantly, had many fewer heelsticks, a median of 15 instead of 20. In comparison many control babies had repeated and/or severe hypoglycaemia after the initiation of the trial: 53% had ongoing hypoglycaemia, 21% had recurrent or severe episodes, 16% had hypoglycaemia after initial stabilization, 29% had study drug dose increases and 3 required open label diazoxide to achieve control.
They also noted that babies had an IV infusion for a shorter time, and there was no recurrence of hypoglycaemia when the diazoxide was stopped. I think this warrants further investigation, especially in the severe group, less than 1.2 or multiple samples <2.0. And always remember that bedside glucometers are very inaccurate in the hypoglycaemic range, often over-diagnosing hypoglycaemia, as yet another study showed, presented as a poster in this meeting. The Auckland group have always used reliable methods in their studies, with high quality lab analyses (usually with a blood gas analyser or the EPOC device), which of course, have to be available very rapidly for this problem.
The ESC-NOW trial was a multicentre investigation of using ESC (Eat Sleep Console) instead of the Finnegan scale for neonatal opiate withdrawal. This secondary analysis showed that maternal buprenorphine treatment was associated with much less neonatal morphine use, and shorter medication use and shorter hospitalisation than maternal methadone treatment.
The ESC approach was advantageous in both groups, shortening hospitalisation from 11 to 7 days in the buprenorphine group and from 17 to 10 days in the methadone group. Need for medications in the babies was decreased from 39% to 15% with ESC in the buprenorphine group, and from 63% to 31% in the methadone group. As far as the impacts on the baby, it is clear that using buprenorphine is preferable to methadone during pregnancy. For an individual mother, of course, there may be enormous numbers of other considerations.
This was a single group study of the use of High Frequency Oscillation with the Draeger VN500 in pressure and then volume guarantee mode. The study was all funded by Draeger, and seems to have been done largely to try and get approval as a medical device in the USA. Subjects were 24 to 30 weeks infants in the 1st 4 days of life who still needed more than 25% oxygen after surfactant with an MAP of 7 or more. Initially the MAP was increased by 2 when the baby was switched, and the initial pressure amplitude was 15-20 cmH2O, which was adjusted by eye, looking for a chest jiggle that appeared right! TcCO2 was used and early blood gases to avoid hypocapnia.
Although not randomized, there was a concurrent control cohort, and outcomes were all very similar to those who stayed on conventional ventilation. Except for pneumothorax which was 4% in the HFO group, and 11% in the conventional group. 11% seems very high to me, after surfactant there are usually far less than that, but not knowing how the control cohort was enrolled makes it difficult to comment.
CO2 fell after the change in ventilation mode, with a mean about 45 mmHg at 2 hours, and a SD of about 4, which means that a mean -2SD is about 37, and there will be occasional babies with more severe hypocapnia. After stabilisation the babies were switched to Volume Guarantee Mode, which, just as in conventional ventilation means that the ventilator adjusts the pressure to give the volume that you set. The mean volume used was close to 2 mL/kg.
This is similar to previous data, that, at 10 Hz, the average volume required is close to 2 mL/kg. Because of those prior findings, and the occasional occurrence of severe hypocapnia in our babies after switching to HFO, we now initiate HFO with Volume Guarantee and a volume of 2 mL/kg. Since starting that approach we have had no severe hypocapnia: I think that makes much more sense than trying to evaluate if the chest is vibrating the right amount. We also try to keep all the preterm babies on 10 Hz, people get confused by the changes in high frequency tidal volumes and DCO2 when you change frequency, so keeping them all at 10 Hz simplifies things. However, with the VN500 you often can’t achieve your desired tidal volumes in larger babies at 10 Hz, and often have to reduce the frequency to 9, 8, or even 7 Hz, to get adequate ventilation.
2 mL/kg is however, just a starting point, some babies will still be mildly hypocapnic, on the first blood gas, which should be done rapidly after switching to HFO, and will need volumes eventually as low as 1.2 mL/kg, others may need volumes as high as 3.5 or occasionally even more.
Many of you will remember the MINVI trial of cord milking for non-vigorous term infants. That trial had a null primary outcome, but re-analysis using the DOOR methodology showed that short-term outcomes were improved with cord milking compared to immediate clamping.
As you can see from this illustration by Satyan, several outcomes of interest to parents were improved by milking, and there were no adverse effects.
At 2 years of age the babies were evaluated with the Ages and Stages Questionnaire, and with the Modified Checklist for Autism in Toddlers (MCHAT). The proportion of babies with abnormal scores on each domain of the ASQ-3, was identical in the two groups. There were no evident outcome differences between groups. I think this means that cord milking in such babies should be routine, having fewer babies with moderate to severe HIE and/or needing Cooling are substantial benefits. Especially as there were no adverse impacts, and the babies also have a higher red cell mass, and therefore higher iron stores, it may be that other benefits will show up later, as has happened with the term delayed clamping trials, where higher haemoglobin led to better very long term outcomes at 4 years, despite no difference at 1 year.
The Darbe trial was a multicentre RCT of darbepoetin among over 640 very preterm infants with the primary outcome being based on an interest in determining whether there was neuroprotection, it hasn’t yet been published in full, but seems to have had no impact on the primary outcome. 23 to 29 weeks infants were included, and the primary outcome is neurological and developmental findings at 2 years corrected age.
P
•650 infants born 230/7 to 286/7weeks’ gestation •Enrolled ≤24hr of age between 2017-2019 (f/u through 2022)
I
•Darbepoetin 10μg/kg weekly (IV or SC) through 35 weeks’ PMA
C
•Placebo (IV) or sham injection (SC)
All
•Parenteral or enteral iron supplementation per study guidelines •Restrictive protocol for red blood cell transfusions
O
•Primary: Bayley III composite cognitive score at 22-26 months •Secondary: Multiple pre-specified outcomes
The presentation didn’t have a slide saying “not to be shared” so I can copy the above data from the pdf, some of the others had such a slide, so I haven’t cut and pasted any figures or slides for some presentations.
Darbepoetin reduced the number of RBC transfusions compared to controls, the mean transfusion number was 1.9 (2.8) vs. 3.3 (3.6), and the proportion who never had a single transfusion was 46% compared to 24%.
The short term respiratory outcomes showed a reduction in grade 2 or 3 BPD with darbepoetin compared to control, from 46% to 34% among survivors; mortality was similar between groups. They performed a mediation analysis to show that it was possible that the reduction in BPD was mediated by the substantial reduction in transfusions.
Although they have gone to great statistical effort to show the potential mediation of the reduction in BPD by a reduction in transfusion, we should keep in mind that the TOP study, in which there were far fewer transfusions in the low hemoglobin group, showed no difference in BPD, or any other index of lung injury.
Short term respiratory outcomes that are of more interest to parents were, however, not different between groups. The proportion of babies going home on oxygen was the same, 37% vs 38%, and the proportions on respiratory medications at discharge were also very similar. Among respiratory outcomes at 2 years of corrected age, 5% in each group were still on oxygen, and 2% on respiratory support, about 1/3 were getting respiratory medications, and 1/5 had been readmitted to hospital for respiratory complications. There was no difference between groups in any of these outcomes.
Darbepoetin therefore had some effect on respiratory adaptation in the NICU, with babies in the active treatment group needing 7 days less of oxygen supplementation, and 7 fewer days of positive pressure, perhaps having something to do with the transfusion need. But no important impact on any outcome that matters to parents, or, indeed, matters to society in terms of health care resource needs.
Thanks to the NRN and Dr Jensen, for providing yet more evidence that “BPD” is of little value as an outcome variable in clinical trials.
Although there was a major reduction in BPD in this trial, it was because control babies needed oxygen or positive pressure for a little while longer, and thus they were more likely to still be in oxygen at 36 weeks PMA, but, as I already noted, none of the outcomes that we have shown to be concerns of families were affected. I think an ordinal outcome variable including several clinically important signs of lung injury would be much more useful for the future. Such an outcome should be co-constructed with parents (and former preterms, perhaps) and would probably include: discharge delayed because of respiratory difficulties; home oxygen after 40 weeks PMA; respiratory medications at discharge; home gavage feeding after 40 weeks PMA. Somewhere in there should also, be the rarer outcomes of prolonged ventilation past 40 weeks, tracheostomy, and so on.
If we keep using “BPD” there is a real risk that interventions that have no major real impact on lung injury, such as darbepoetin, may be widely used to improve pulmonary outcomes, which, based on these data, would be a big mistake. In contrast, using darbepoetin to reduce transfusions may indeed be something that appeals to parents. I think that being able to avoid transfusion might be something that parents would value, if there are no adverse impacts of the medication, which seems to be the case.
As usual the PAS meeting was packed with neonatology. The App was even worse than last year, crashing frequently on many peoples’ phones. It was completely useless for the posters, as they were listed by number, and the different sessions all started with #100; so there might have been 3 posters with the same number and no indication of which day they were on.
The organisers also decided to not have any indication of what subjects were in various areas of the poster session, so you might accidentally happen upon a group of posters about breast milk and breast feeding, for example, but there wasn’t any way to know that such a group existed, or how to find them. As there were several hundred posters in each session, it was impossible to actually find the things that interest you. If I wanted to see the posters about neonatal respiratory care, I just had to wander around the huge room and try to find them, the App didn’t help at all.
Sunday morning was very interesting, with Neonatal Clinical Trials 1 followed by the Silverman lecture given by Dr Annie Janvier then Neonatal Clinical Trials 2. Unfortunately there were other interesting neonatal things going on at the same time, which had tiny participation, a session on BPD and another on caffeine would have been super interesting, but I would guess had almost no-one there, as the Clinical Trials sessions were packed.
As for those Clinical Trials, I will outline a few of them here and give a link to the protocol if one is published, and may well return to them with a longer post when they are actually published. One of them was the Video-Laryngoscopy trial that I have already posted about. In no particular order the others were :
The PLUSS trial
The PLUSS trial showed that budesonide is not a plus when added to curosurf. Many of you will know that previous work, including moderately large trials from Taiwan, have suggested a reduction in early signs of lung injury when initial surfactant dosing was mixed with budesonide, comapred to budesonide alone. This trial randomized infants <28 weeks receiving surfactant either by intubation or LISA/MIST, and followed the babies until discharge.
As you can see there was no impact on either BPD at 36 weeks, or survival to 36 weeks. I think it is very unlikely that any longer term follow up will show a benefit from the budesonide, although there may be a risk of neurodevelopmental adverse impacts form this potent steroid, a small proportion of which is absorbed. There was no subgroup that appeared to have any benefit. A recent observational study has suggested that the budesonide only seemed to help the larger babies, perhaps that is the difference to the prior positive trials, which included many larger infants.
IBUPAR
The IBUPAR trial was a multicentre Spanish trial including preterm infants <30 weeks with a “haemodynamically significant” PDA, to either Ibuprofen or Paracetamol (=acetaminophen), which was diagnosed on average about 3 to 4 days of age. I found it odd that this was being presented, as the sample size is 274, but only 134 have completed the trial and 111 completed the treatment, nevertheless the data from an interim analysis have been unmasked in order to be presented at PAS. The primary outcome is closure of the PDA, which occurred in 60% with ibuprofen, and 40% with paracetamol. By the time of discharge most had closed in both groups, but it was 85% in the ibuprofen group and 70% with paracetamol. Acute Kidney Injury was much more frequent with ibuprofen, 16% vs 5%, but I am not sure what the diagnostic criteria were for that diagnosis, whether it was a transient oliguria, or persistent increase in creatinine, or some combination. There were no major differences in NEC, BPD, IVH, ROP or mortality. This shows I think that it doesn’t matter much if you give a less effective therapy to close the PDA, the babies turn out just as well despite having an open PDA.
There were two trials of prolonged use of caffeine in the preterm.
The MOCHA trial
The first was the MOCHA trial, which unfortunately. despite the name, did not give the babies chocolate along with their caffeine. In this trial 827 moderately preterm babies (29 to <34 weeks GA) who had received caffeine for clinical reasons, and had arrived at 33 to <36 weeks PMA, with a plan to stop their caffeine, were randomized to continue caffeine citrate at a dose of 10 mg/kg/day or placebo, with the plan to continue until and beyond discharge, for a further 28 days at home. The primary outcome was the number of days of hospitalisation, the idea being, of course, that less apnea, and fewer hypoxic events will lead to earlier discharge. There was no real difference in the primary outcome, which was slightly shorter in the placebo group, and identical after adjustment. Days to “physiologic maturity” (which meant apnea free, full oral feeding and stable in a crib) was also identical. Caffeine was having an effect, however, the caffeine group babies had fewer apneas, it was 2 days faster for them to be 5 days apnea free, but in the end it wasn’t apnea that kept most babies in hospital but feeding issues. There were no adverse effects, so in terms of duration of hospitalisation, caffeine is likely to only have an impact in the subgroup who are having persistent apnea.
The ICAF trial
On the other hand, the ICAF trial enrolled more immature babies <30 weeks GA at birth, who had arrived at 32 to <37 weeks and were about to have their caffeine stopped. Within 3 days of caffeine being stopped they were randomized to caffeine or placebo at the standard dose, but when they reached 36 weeks the dose was doubled to 20 mg/kg/day of caffeine citrate. The babies were on recording pulse oximeters until 44 weeks, and stayed on caffeine or placebo until 43 weeks. 160 babies were in the study, and these are the post-randomization findings.
Characteristic
Placebo N= 78
Caffeine N= 82
P-value
Clinical caffeine
3 (3.8)
3 (3.7)
0.95
Supplemental oxygen
14 (17.9)
4(4.9)
0.009
CPAP, BIPAP or High Flow NC
7 (9.0)
2 (2.4)
0.07
Mean (SD) PMA (weeks) at discharge*
38.5(2.2)
37.7(2.2)
0.04
Mean (SD) weight gain (g/day)*
38.5(25.3)
29.2(12.7)
0.004
Median (min – max) days from randomization to discharge)**
27 (21 –32)
17 (15 –23)
0.001
In these more immature babies caffeine did shorten hospitalisation, and improved their respiratory course, not just in terms of apnea, but by keeping them off oxygen or respiratory support. The oximeters showed dramatic effects on intermittent hypoxia, which continued to be significantly different until 42 weeks.
I think the big difference between the two trials is the less mature group enrolled in ICAF, who have more IH for longer, and have more lung disease.
We already know that caffeine metabolism changes at near to term, and levels are very low on the standard dose, so I think the increased dose was a good idea in ICAF. But I don,t think it would have made much difference to MOCHA as it wasn’t apnea, or IH that were keeping the babies in hospital.
Long term follow up will be important for both trials, and ICAF2 with a larger sample size is certainly warranted based on this information.
I posted about the short term results previously, it was an RCT of early closure of the PDA with ibuprofen that had basically a null result. The infants were followed at 2 years of age, with parental report of their development using a scale called PARCA-R, unfortunately the outcome was dichotomised into those with moderate or severe “neurodevelopmental impairment” or without. There were slightly more overall deaths with ibuprofen, and slightly more so called “impairments”, most of which I presume was slower development, in the controls who survived.
Longer term respiratory outcomes were also presented, including duration of oxygen treatment from randomization to stopping oxygen, and a composite of respiratory morbidity, the definition of which is shown in the legend to the figure below. It shows not a hint of any difference.
Trial of Prolonged CPAP therapy
Finally, in the first session in the morning there was a trial of prolonged CPAP therapy, which disconcertingly had no zippy acronym. This should be illegal, how can a poor blogger think up catchy titles for their posts? Babies of 24 to 32 weeks GA at birth who were on CPAP, who were thought to be ready to stop their CPAP and who met stability criteria, had FRC measurements and were then randomized to have 2 extra weeks of CPAP at 5 cmH2O, using a bubble system.
The lungs grew more on CPAP. As you can see, after the extra 2 weeks the FRC was close to 20% greater in the prolonged CPAP group. At 6 months of corrected age lung volume was again measured, with a different technique which gives alveolar volume, and they also measured the diffusion capacity for CO,
At 6 months the prolonged CPAP treated lungs has a higher alveolar volume, and had better gas exchange.
I think this is potentially a major advance in respiratory care, it may be that babies who are at high risk of long term respiratory problems might be helped just by leaving them on CPAP for two weeks. Of course we don’t know if they had better or worse small airways function, or any difference in any clinical outcomes. But, many of us have major concerns about the very long term outcome of our patients, who may have deteriorating lung function later in life if they start out with fewer alveoli. This is potentially a big advance, if we can show that it is safe and has medium term advantages.
On my way to the PAS annual meeting in Toronto, I decided to spend a couple of days at Point Pelee national park. It is the southern-most point of Canada, a peninsula jutting out into lake Erie that is a landing spot for birds migrating north in the spring. It was, fairly recently, proven that most Warblers migrate at night, so after a long night flying across the great lakes many of them land at Point Pelee to “refuel” before heading onward to the Boreal Forest. Some very pretty birds can be found there, and some rarities.
When I parked my car I immediately saw a group of enthusiasts watching a little brown bird, which turned out to be a Harris’s sparrow; it was the first time this bird had ever been seen in the Park.
Red-headed woodpeckers are one of my favourite Canadian birds, this was only the second time I had seen them.
Green Herons are fairly uncommon in the Park
Loggerhead Shrike don’t usually come this far north, I think this was a first for the Park also
One of the prettiest birds you could see is the Prothonatory Warbler (apparently named after the vestments of an official in the Catholic church!)
If you are interested, there are other photos on my other blog at keithbarrington.com
I have been increasingly using video laryngoscopy in my practice, both when I myself perform the intubation, and when I am supervising a resident or other trainee. I usually ask them to use a VL when it is a nurse practitioner or RT that is about to intubate also.
It already seemed to me that the evidence was very supportive, with lower failure rates with VL than standard laryngoscopes, and I really appreciate the ability to see what someone else is doing when training them. A recent large RCT in adults with over 8000 procedures showed that the failure of the first intubation attempt was about 7.6% with direct laryngoscopy and only 1.7% with a video device. Intubation failure (needing more than 3 attempts or switching to a different device) was 4% with standard technique, and only 0.26% with the video. Ruetzler K, et al. Video Laryngoscopy vs Direct Laryngoscopy for Endotracheal Intubation in the Operating Room: A Cluster Randomized Clinical Trial. JAMA. 2024;331(15):1279-86.
I wish we could have success rates that high; multiple intubation attempts lead to local trauma, pain, increased physiologic deterioration including spikes in intracranial pressure and are associated with increased IVH. In marked contrast to the above study in adults, a recent single centre randomised study of video versus direct laryngoscopy for nasotracheal intubation in the newborn (n=89) had desperately poor rates of intubation on the first attempt: 49% success on the first attempt with the video, and 44% with direct laryngoscopy. Tippmann S, et al. Video versus direct laryngoscopy to improve the success rate of nasotracheal intubations in the neonatal intensive care setting: a randomised controlled trial. BMJ Paediatr Open. 2023;7(1). Infants received fentanyl and midazolam, some of them received vecuronium “if intubation conditions were considered inadequate after analgesia and sedation”, I have no idea how you can determine this prior to laryngoscopy. Most of the intubation first attempts were by trainees (60%), and the babies were intubated either in the delivery room or the NICU, most were preterm. Although I said first intubation success was “desperately poor”, such results are similar to many other studies which also have very poor success on 1st attempt.
What can we do to improve these poor success rates? Well, in addition to the RCTs, there is an ongoing multicentre quality evaluation initiative Near4Neos that has shown that you are more likely to be successful if you use muscle relaxant, and if you use a VL. But overall, in those database studies, success of the first attempt was still very low. With muscle relaxant 56% 1st attempt success, vs 33% with sedation alone, and 58% with the VL, compared to 47% with direct laryngoscopy.
Which brings me to the new trial, presented in Toronto at the 2024 Pediatric Academic Societies meeting, and published the same day in the NEJM. Geraghty LE, et al. Video versus Direct Laryngoscopy for Urgent Intubation of Newborn Infants. N Engl J Med. 2024. This single-centre RCT from Colm O’Donnell’s unit in Dublin randomised just over 200 babies, in the DR or in the NICU, who required urgent intubation. It therefore included preterm and term infants, and babies in the NICU routinely received premedication, which included a muscle relaxant.
The results showed a dramatic difference.
The paediatric residents success rate went from 40% to 70%, neonatal fellows from 53 to 77%. There were very few performed by staff neonatologists, even though the slide from the presentation in Toronto shows Dr O’Donnell’s ear as he is intubating with a VL:
As you can see from the next figure the advantages applied to small and preterm babies as well as the overall group, in the DR and in the NICU.
The successful first attempts took about 10 seconds longer with the VL (60 vs 50 seconds), but that did not increase the number of babies with major desaturation or bradycardia.
Of note, some of my colleagues performed a fascinating trial among paediatric trainees, Michael Assaad and Ahmed Moussa are 2 of my colleagues in Sainte Justine, and Ewa Gizicki was one of our fellows, the other authors of this trial are colleagues from across Quebec. They randomized residents about to intubate a baby to a 10 minute training session (if there was time depending on baby’s condition, of course), which consisted either of a 10 minute video, or “Just In Time” training, which was a 30 second video accompanied by practice on a mannequin, supervised by a staff who used scripted feedback depending on the difficulties experienced. Gizicki E, et al. Just-in-time Neonatal Endotracheal Intubation Simulation Training: A Randomized Controlled Trial. J Pediatr. 2023:113576. First attempt success rate was improved in the JIT group 54% compared to 41%. Rethinking exactly how we train and supervise residents for this important skill has been a focus of my 2 colleagues for a while, this approach requires a lot of the staff supervising, but it seems to work.
I think the accumulation of evidence and this new RCT makes it clear that Video-Laryngoscopy should be considered the standard of care of neonatal endotracheal intubation. It has universally been shown to reduce failure of the 1st intubation attempt, and even though the 1st attempt may be slightly longer, the overall duration of laryngoscopy is much shorter as you are more likely to only do it once!
Optimal Endotracheal Intubation Procedures:
Ensure that the appropriate person is doing the procedure, someone who has been well-trained, with the use of simulation before practising on babies.
High risk intubations should only be performed by an individual with proven competence. Intubating a 500 gram infant is not the time for a junior resident to “have a try”.
If at all possible, the baby should be premedicated, using an analgesic with a rapid onset (not morphine), and a muscle relaxant with rapid onset and short duration of action.
Oxygen should be administered during the procedure, preferably by high-flow nasal cannulae.
Video-laryngoscopy should be used for all neonatal intubations, both in the DR and in the NICU (note to self, we need to get them available for the transport team also).
Senior supervision of trainees is essential, and is also facilitated by Video-Laryngoscopy.
Endotracheal intubation is the most traumatic procedure that we perform frequently in the NICU, multiple intubation attempts harm our babies, and we should do everything possible to reduce their number.
The standard method of pasteurization of donor breast milk, and I believe the only method approved by HMBANA (the human milk banking association of north america), is similar to what Louis Pasteur himself came up with a couple of centuries ago (in about 1865). I’m not sure why its called Holder pasteurization, it may be because the temperature is “held” at 62.5 degrees for 30 minutes, or maybe it’s named after someone. In any case, it is effective in killing most bacteria, not including spore-bearing forms (like Bacillus cereus), but, unfortunately, inactivates white cells, kills probiotic organisms and denatures many large molecules, especially proteins.
Some of the advantages of human milk are therefore impacted by Holder pasteurization (HP), and there are a few review articles available (including this one from the European MBA) which discuss the available pasteurization methods. Other means of pasteurization that have been evaluated, include ultraviolet treatment, high hydrostatic pressure, and most particularly High Temperature Short Time (HTST) which uses temperatures of about 72 degrees for 15 seconds. All the alternative methods seem to be equally good at inactivating bacteria, and many viruses, but the alternatives have much less negative impact on the other beneficial components of milk (Oligosaccharides are relatively unaffected by HP). The major current problem with changing over to alternatives are the new equipment that will be required, and the lack of good evidence that it actually makes a clinical difference.
A recent RCT from 2 NICUs in Madrid compared supplementing mothers milk with pasteurized donor milk that used one of 2 different pasteurization methods, HP and HTST. They enrolled 213 and analyzed results from 160 ELBW infants, with the primary outcome being the occurrence of CLABSI, or a positive blood culture in an infant who had a central line in place for more than 48 hours, and in whom the culture was drawn during catheter use or within 48 hours after.
The most striking result is the enormously high frequency of CLABSI in those units, in both groups, 42% with the HTST and 46% in the HP group. There was no statistical difference between groups, but the study was designed to have 150 babies per group, and powered for a 33% reduction in the outcome. They stopped adding patients after 160 were actually enrolled, which is never really explained, the confidence intervals for the relative risk of catheter related sepsis were 0.68-1.26 with HTST compared to HP, so a large potential benefit is consistent with this null result.
The definition of late-onset sepsis is quite rigorous, and required a positive blood culture, after 72 hours of life, with a central catheter in place (or removed <48 h) and at least 2 of a sequence of clinical signs, or in the case of CoNS also other lab results (CBC results or CrP raised). When expressed per 1000 catheter days, the incidence in the two groups was about 18/1000 catheter days.
As a comparison, among babies <1kg birth weight in the CNN, all LOS combined (which is any positive blood culture, not requiring any other criteria, either with or without a central catheter in place) was 28%. In a publication from the German Neonatal Network, they showed “only 45% of primary BSI were CVC related in the <1000 g birthweight group” in other words, only about half of LOS was catheter related. That percentage is going to vary a great deal, depending on how central catheters are used, and the GNN also showed also a large variation in use. Nevertheless, the infections reported by this new Spanish study are only a subgroup of all the LOS in their infants; again in comparison with the Canadian Network, among infants <33 weeks the incidence of CLABSI was about 7.5/1000 catheter days.
As far as I can tell, the Spanish study did not enrol particularly high-risk babies, some were eliminated because of a “risk of early death”, they seem to be a group with relatively standard risk profile.
I think that different ways to try and preserve the advantages of human milk, by using different pasteurization methods is really important. However, by themselves, such improvements will not be likely to have a huge impact on LOS rates, especially when only a small proportion of the milk received is the donor milk. Enteral colonization with pathogenic, mostly Gram negative, organisms is a known common occurrence prior to blood stream infection. This recent publication for example, Schwartz DJ, et al. Gut pathogen colonization precedes bloodstream infection in the neonatal intensive care unit. Sci Transl Med. 2023;15(694):eadg5562, showed that GI colonization with the same organisms was very common, and that the pathogens were more abundant, in babies who later developed LOS. The anti-infective properties of milk will have an impact on gut colonization, and it is likely that different pasteurization methods will have an impact on those benefits.
There are many other things we do that can affect the incidence of LOS, I note that in the Spanish trial the use of “anti-acid” medications was very common at 20 to 30%, there are no details, and they may have been given after the episodes of sepsis, but I think there is very little indication for the use of such medications, which clearly can increase gut colonization by reducing one of the important barriers.
I just hope this new study doesn’t discourage the on-going development of better ways to pasteurize donor human milk which preserve more of the advantages. I think we need more, adequately powered studies to, firstly, show impacts on the intestinal microbiome of different pasteurization methods, and, if possible, following that, the clinical benefits. But we must take into account the large number of other risk factors for LOS, and reinforce quality control. Stringent quality improvement methods can reduce LOS in the very preterm, and have to be a priority wherever ELBW infants are cared for.
The article title immediately alerts to the slant of the argument, apparently active intensive care for the most immature infants is “aggressive”, the authors justify this word by noting that Helen Harrison used it. Speaking as someone who was publicly attacked by Helen Harrison after a presentation I gave, in which I had noted that the long term outcomes of preterm infants were largely positive, I don’t find that justification adequate!
The authors of the article make a number of points that it is difficult to argue with: that outcomes are variable and uncertain; that decisions should be individualized and parents should always be involved; and that we should advocate for long term support for survivors and their families.
My question is: what is so different about this subgroup of babies? Surely the same things can be said about high-risk diaphragmatic hernias, or babies with severe variants of hypoplastic left heart syndrome, or babies born at 28 weeks after rupture of membranes at 20 weeks and persistent anhydramnios. Indeed, for each of those 3 examples, survival may be lower, and long term complications just as uncertain, as the baby born at 22 weeks gestation.
Surely all high risk infants, whatever their gestational age, have outcomes that are uncertain and variable, require individualized decision-making with parental involvement, and deserve long term support as much as they deserve active intensive care.
There are some other concerns about this paper, it is stated, for example, “As the rates of BPD were high, and BPD correlates with NDI, the loss to follow up may under-estimate rates of NDI” I don’t understand the logic of that sentence, unless there is some unstated evidence that infants with BPD have lower follow up rates. In fact the reference just given in the article, which was written by the first author of this new opinion piece, shows the opposite; higher follow up rates are associated with lower rates of “NDI”.
In that systematic review, it was shown that loss to follow up, on average, over-estimates rates of NDI. Which I presume is because infants who are doing well are, in general, less likely to be brought back by their parents for neuro-developmental assessment.
The authors also talk about the high rate of acute morbidity among the most immature infants, and it is true that a very high proportion have acute morbidity, including extremely high rates of BPD. But, still needing oxygen at 36 weeks is a consequence of exposing fragile and extremely under-developed lungs to oxygen and positive pressure, and doesn’t necessarily have huge impacts on the baby’s future quality of life. 2 of the other 3 examples I gave of high risk babies also have a very high incidence of chronic respiratory problems, and extremely high proportions have acute morbidity.
I think it would have been much more useful to point out that survival and other outcomes among these babies are extremely variable, but that with a consistent approach, co-ordinated with our obstetric teams, good survival is possible, and we should all be striving to institute best practice. They could also have indicated where to find the best information about how to improve survival and outcomes of these babies.
The Tiny Baby Collaborative has been developed in an attempt to improve the outcomes of such infants, their website gives you access to several recent Webinars. One interesting aspect of which is that the approach to such infants differs in many important details from one successful centre to another, in terms of fluid management, ventilatory approach, etc. What is universal in such centres is a collaboration with obstetrics, consistent, protocol-driven, approaches, and a commitment to and belief that these babies are worth the effort. I don’t find it useful to define one category of babies that are not worth as much as others; it is not many years since people were saying exactly the same things about babies born at 24 weeks. With “aggressive” quality improvement initiatives, survival is now over 70% across Canada, it has become infrequent to not start intensive care for such infants, and huge numbers of families, including my own, have benefited from a refusal to suppose that we have arrived at a limit of viability, but rather to push the barriers and progressively improve our care processes.
A new large RCT from PICUs in the UK randomly compared 2 saturation target ranges, 88-92 and >94%. (Peters MJ, et al. Conservative versus liberal oxygenation targets in critically ill children (Oxy-PICU): a UK multicentre, open, parallel-group, randomised clinical trial. Lancet. 2024;403(10424):355-64). Children <16 years of age being ventilated with supplemental oxygen were randomized within 6 hours of admission. The intervention stopped when the child was extubated. Primary outcome was “the duration of organ support” up to 30 days after admission. This was “a rank-based endpoint with death either on or before day 30 as the worst outcome (a score equating to 31 days of organ support), with survivors assigned a score between 1 and 30 depending on the number of calendar days of organ support received. Organ support … included respiratory support… cardiovascular support… and renal support. Other components of organ support included analgesia or sedation, exchange transfusion, neurological support, and metabolic support”.
2040 patients were randomized and data included for 1872 of them (more about that in a moment), half of the children were <1 year of age.
Differences in outcomes were small, but there was a reduction in the adverse primary outcome,
The size of the effect is expressed in a way which I find difficult to understand, it is “a probabilistic index of 0·53 (95% CI 0·50–0·55; p=0·04 Wilcoxon rank-sum test) indicating a higher probability for a better outcome in the conservative oxygenation group”.
I found this graph made the results a little easier to comprehend, the cumulative proportion of infants in each group who required organ support in the first 30 days of hospitalisation, with the children who died scoring 31.
This relatively small advantage of lower SpO2 targets was achieved despite many infants being above the target range for significant periods of time. As in the preterm O2 studies, much of this is because they were in 21% oxygen, but some was due to periods of time in supplemental O2 with SpO2 above target.
These 3 panels, using the same colours as the figure above, show the proportions of time by SpO2 in the top panel, the proportions of time with each SpO2 reading while receiving more than 21% oxygen in the middle panel, and the proportion of ventilated time with each FiO2 in the bottom panel. This is the version from the supplemental materials, which includes data up to day 30, data from the first 7 days are in the main article and look very similar. It looks like the lower SpO2 target group were extubated about 3 hours earlier on average, and may have had a shorter duration of cardiovascular support.
Previous studies in newborn infants of differing saturation targets include the early life trials which were meta-analysed in the NeoProm collaboration, and led to an increase in saturation targets, as mortality was higher with SpO2 targets in the high 80’s compared to the low 90’s.
Studies in older preterm newborn infants include STOP-ROP, and the first BOOST trial. STOP-ROP was a large RCT, published in 2000, among about 600 preterm infants who had developed pre-threshold retinopathy, as then defined, and who had a median SpO2 of <94% in room air. Pre-threshold was any zone 1 RoP of stage 1 or 2 without plus, or zone 2 RoP of stage 2 with a limited region of plus disease, or stage 3 without plus. The infants were randomized to target SpO2 in the low 90’s or the high 90’s (89-94 vs 96-99%), the idea being that some animal models showed decreased progression of retinopathy, after it had first appeared, when saturations were maintained in a higher range. It was thought that maintaining retinal ateriolar vasoconstriction with mild hyperoxia could prevent neovascularization. There was no overall difference in ophthalmic outcomes between groups, on subgroup analysis there might have been an advantage of the high oxygen group in the eyes that did not have plus disease, but the most severely affected infants, with plus disease, had no difference between groups. The intervention continued for at least 2 weeks, and until the baby’s eyes had either reached “threshold” (and treatment indications) or had shown regression on 2 subsequent eye exams. Almost all the babies were out of the study before discharge, as they had satisfied eye end points. Babies were usually randomized at about 35 weeks PMA, and therefore most had a BPD diagnosis. All the pulmonary outcomes were worse in the group with higher SpO2 goals. There were more babies still hospitalised at 3 months corrected age, more babies on oxygen at 3 months, were more episodes of pneumonia and episodes of BPD exacerbations in the supplemental oxygen group. In contrast to the expectations of the investigators, babies with higher oxygen saturations did not grow any better, with identical weight gain between groups.
BOOST was a trial from Australia of supplemental oxygen among 360 preterm infants <30 weeks GA who had reached 32 weeks and still required oxygen. They were randomized to target SpO2 of 91-94% or 95-98% (using masked oximeters) and remained in their group until oxygen was stopped (including when at home). The hypothesis behind this trial was also that higher saturations would lead to better growth and development, which was a reasonable supposition, based on observational data that babies who had been maintained with higher saturations had better outcomes in those domains. The RCT, in contrast to the observational data, showed no advantage in growth or development. In order to maintain the higher saturation, the high SpO2 group stayed in oxygen much longer, and were more likely to go home in oxygen. They also had more respiratory related deaths (1 vs 6, p=NS), more treatment with steroids and/or with diuretics, and had more re-hospitalisations.
Which suggests that even small increments of inhaled oxygen concentrations can have pulmonary toxicity, and that the balance between ensuring adequate oxygen delivery while limiting oxygen-induced lung injury is a very narrow one. Maintaining tight oxygen saturation limits, and avoiding even mild hyperoxia seems to be important for ensuring the best clinical outcomes, in older children, as well as in newborn infants.
It was fairly recently that I deconstructed a truly terrible database analysis which claimed that neonatal mortality was dramatically increased among very preterm infants who received mother’s own milk (MoM) and donor human milk (DHM), without any formula or fortifier, compared to a group which only received MoM and artificial formula. Almost certainly, this was because many of the deaths in the MoM+DHM group occurred before the babies had survived long enough to receive fortification. A baby who survived for a few weeks and then received some fortification or formula being deleted from the group, even if the change of diet occurred after the primary outcome (surgical NEC) had been determined! The study also claimed that MoM+DHM was associated with more surgical NEC (3.5%) than MoM+formula (1.7%), but with a difference much smaller than the enormous difference in mortality. This could well be another example of confounding by indication, and again, much of the feeding data were derived from medical record abstraction which included many days, or weeks, after the occurrence of the NEC, up to and including the day of discharge.
The details of the protocol are now available, with the full publication. Eligibility included a GA <29 weeks or birth weight <1000g, 483 babies were randomized. If the mother never initiated breast milk expression, or stopped before 21 days of age, or produced “minimal” amounts of MoM, then the infant could be included, and randomized, which could therefore occur any time between birth and 21 days of age. The median age of randomization was 16 days, and the median age of starting oral feeds was 4 days. Which makes me wonder why, on average, these babies were not fed for 4 days? I think most babies, of any gestational age, can be fed on day 1. Our protocol is to only withhold feeds for babies in shock and/or on inotrope/vasopressors, which is a small minority, most extremely preterm babies receiving either MoM, or if not yet available, DHM, within 24 hours.
The eligibility up to 21 days dilutes the potential differences between groups, I suggest, as many babies will have had some MoM, or DHM, or formula, before they were enrolled. Even after being enrolled, many babies in both groups received some maternal milk, which is reported as the number of weeks of any MoM, and averaged about 1.7 weeks.
I’m not sure how many babies were in those 3 slightly different subgroups (no MoM, mother stopped expressing before 21 days, and mother not able to produce enough milk), but some analyses were done for the subgroup who had zero MoM. In the supplemental materials, it looks like there were 369 babies in total that completed follow up, 79 of whom had zero MoM; in other words it is a minority of mothers who never expressed at all.
The primary outcome variable was the Cognitive Score on the Bayley version 3, performed at 2 years corrected age. Another thing which I find a bit weird, is that the babies who died were assigned a cognitive score of 54 (the minimum possible); it is one way of integrating the competing outcome of mortality with developmental outcomes, but, depending on the risk of death, it could well dilute any difference in cognitive outcomes, especially if there were an imbalance in mortality. It also makes the scores look a lot lower than they really were, mean scores being about 5 points higher when the non-survivors assigned scores were removed. Scores on the language composite (44) and motor composite (49) were also assigned for babies who did not survive.
In any case, all of the developmental and neurological outcomes were very similar between groups. In the supplemental material the analysis restricted to the survivors is given, which shows that all of the mean and median scores are slightly better in the DHM group, but none of the differences are large enough to really have any clinical significance (and none are “statistically significant”). Interestingly, the differences are all greater in the “sole diet” subgroup, the median scores on each subscale being 5 to 6 points higher with DHM than formula, but the numbers are, of course, much smaller, and remain within 95% CIs.
Assigning the lowest possible score to the non-survivors leads to some slightly strange findings, for example, the Bayley motor composite scores among survivors were slightly higher in the DHM group, but identical when the non-survivors are added as scoring 49; and there was actually more cerebral palsy (all grades) in the formula group, 20% vs 15%.
For the short term outcomes, the most striking difference is in NEC, which was twice as frequent, 9%, in the formula group than in the DHM group, 4.2%. You may want to argue that NEC is a somewhat subjective diagnosis, but, as I haven’t yet mentioned, this was a masked trial. The study group went to some lengths to maintain masking, with feeds prepared daily by study staff in amber tinted syringes, which were continued for 120 days maximum, or until 1 week before anticipated discharge.
Another interesting outcome is that length and head circumference increased very similarly in the 2 groups, but the weight gain was a little greater in the formula fed babies. Both groups “lost” percentiles of length, and gained a little in head circumference, being born with weight and length z-scores of about -1, and head circumference z-scores of -1.4. At discharge the DHM group weighed about 140 g less. The Fenton standards were used to create the z-scores.
The study confirms the safety of donor human milk, and the reduction in NEC compared to artificial formula. This is despite the limitations in the study design, which made the study feasible (only enrolling babies whose mothers decided to not provide breast milk would have enormously prolonged the trial), but potentially diminished any difference between groups.
A recent editorial in Pediatric Research raised some reasonable concerns about the commercialisation of human milk for preterm infants, but suggested, with no real evidence, that there are also concerns about harm from DHM. It misquotes O’Connor’s previous RCT of MoM-receiving babies who either received supplemental feeds with DHM or artificial formula, and which also showed only tiny differences in mean cognitive scores, of 92.9 with DHM vs 94.5 with formula. The spread of scores was wider in the DHM group, so a somewhat larger proportion had scores between 70 and 85 in the DHM group, but this was a post hoc secondary analysis. The editorial also suggests that the mortality morbidity index was worse in the DHM group, in fact the tiny difference 43% vs 40% in this composite outcome was far from being significant, the only clinical outcome that was different in that trial was a much higher rate of NEC (stage 2 or 3, 6.6% vs 1.7%) with formula than DHM.
The MILK trial therefore confirms the benefits of DHM as a replacement for inadequate or absent MoM, compared to formula. It shows that NEC is lower, and there are no adverse effects. With reasonable standardized nutritional practices weight gain may be slightly less with fortified DHM than with formula, but length gain and head growth are very similar. Developmental outcomes are also unaffected, or are perhaps a little better with DHM, and maybe with less CP.
Although many of us are uncomfortable with the commercial side of human milk trafficking, most human milk banks are non-profit, and/or government supported, and most donors are altruistic women who wish to help others, and are fortunate enough to have a surplus of this precious resource. Our own local bank, like many others, does not accept donations from mothers who are within the first month of breast feeding their own baby, in order to not impact adversely on breast feeding their own infant. There are many challenges to ensuring adequate DHM supply, standardizing and optimizing nutritional intakes, adjusting fortification to alter nutrient supply by when growth is suboptimal. The MILK trial results confirm that responding to those challenges is worth the effort.
Neonatal Research 