#442 - [Journal Club] - 📌 Does NIRS guided treatment improve clinical outcomes for extremely preterm infants?

In this episode of Journal Club, Ben and Daphna dive into the results of the NIRTURE trial, recently published in JAMA Network Open. Building on the lessons of SafeBoosC 3 , the NIRTURE investigators aimed to reduce the burden of cerebral hypoxia and hyperoxia in extremely preterm infants using a standardized NIRS guided treatment protocol. While the study showed a dramatic improvement in maintaining cerebral normoxia, driven largely by a reduction in hyperoxia , the clinical outcomes before discharge remained neutral. Join us as we discuss whether regional oximetry is a must have bedside tool or just another data point in search of a clear clinical benefit.  

----

Cerebral Oximetry-Guided Treatment and Cerebral Oxygenation in Extremely Preterm Infants: A Randomized Clinical Trial. Jani PR, Goyen TA, Balegar KK, Maheshwari R, Saito-Benz M, Schindler T, Moore J, Merhi M, Cruz M, Song Y, McDonagh H, Luig M, Tracy M, D'Cruz D, Perdomo A, Morakeas S, Dasireddy V, Culcer M, Shingde V, Bennington K, Michalowski J, Fucek A, Querim J, Stevens S, Santanelli J, Elhindi J, Gloss B, Halliday R, Shah D, Popat H.JAMA Netw Open. 2026 Feb 2;9(2):e2557620. 

Transcript

Ben Courchia (00:00.57) Hello everybody, welcome back to the Incubator Podcast. We're back this week for another episode of Journal Club.

Daphna Yasova Barbeau (00:06.786) We're here. We're here. We made it to Journal Club. I love Journal Club.

Ben Courchia (00:15.522) I love Journal Club. I feel like in a dream world, I get paid to read papers and do Journal Club.

Daphna Yasova Barbeau (00:20.791) Hmm.

Ben Courchia (00:27.77) What do you think about that?

Daphna Yasova Barbeau (00:28.952) I think, yeah, in a dream world. This is why people write papers. When you go to the bedside, you're like, "Well, maybe I don't have the exact right answer for this exact right patient, but this is what I do know".

Ben Courchia (00:47.75) The only reason I'm saying that I would like to get paid to do this is because it takes a lot of time. The pile of papers accumulates. If we don't record for a week, I see the pile accumulating and then I'm like, "But now...". We still don't have a nice process when it comes to that, because I read a lot of papers and then halfway through I'm like....

Daphna Yasova Barbeau (00:53.751) It does take a lot of time.

Daphna Yasova Barbeau (01:05.282) Yeah, we're like, "That's good".

Ben Courchia (01:15.864) "Nah, I'm going to put this one in the bin just for my own knowledge". It's not going to make it onto Journal Club, and that is why it takes a long time.

Daphna Yasova Barbeau (01:27.586) I think that's what some people don't realize. We do read papers that we don't report on, and not because they're not useful papers, just for one reason or another in that timeframe. A paper beats out another paper.

Ben Courchia (01:41.198) For sure, because we only have so many slots. All right. Do we have to talk about anything else? Nope. Where am I taking you this particular week? Daphna, you're going to pick. We can either talk about NIRS or we can talk about Apgar scores and umbilical artery pH.

Daphna Yasova Barbeau (02:07.502) I think people want to hear about NIRS.

Ben Courchia (02:11.97) Okay, let's do NIRS. I found this paper in JAMA Network Open. It's called "Cerebral Oximetry Guided Treatment and Cerebral Oxygenation in Extremely Preterm Infants". This is the NIRTURE trial. It is the Near-Infrared Spectroscopy Targeted Use to Reduce Adverse Outcome in Extremely Preterm Infants.

Daphna Yasova Barbeau (02:15.97) It's a stretch, but they got there.

Ben Courchia (02:42.242) It's like in Formula One when they lose the tires and they just get the car back on the road. It's just enough to make it. Let's talk a little bit about the background. The stability of cerebral oxygenation by using NIRS, which is near-infrared spectroscopy, compared with systemic assessment of oxygenation using pulse oximetry.

Ben Courchia (03:09.402) It gives you another alternative to understand and potentially reduce brain injury. The appeal of this approach is that pulse oximetry reflects systemic arterial oxygen saturation. NIRS gives you regional cerebral oxygenation directly, giving you a window into what is actually happening at the brain level where it matters most. Before I get into the paper, we're working on a mini-series on NIRS specifically. Stay tuned; there will probably be a five-episode mini-series on that later this year.

Ben Courchia (03:42.549) The big topic of discussion is the SafeBoosC trial. You need to understand the SafeBoosC trial before you understand this. SafeBoosC-2 was published in 2015 in the New England Journal of Medicine.

Ben Courchia (04:03.002) It was a randomized trial of about 170 preterm infants randomized to either treatment guided by cerebral oximetry monitoring during the first three days of life, or usual newborn care. The primary outcome was the burden of cerebral hypoxia and hyperoxia, defined as the total duration and magnitude of cerebral oxygenation below 55% or above 85%. The SafeBoosC trial was not really powered to detect differences in clinical outcomes. It was a proof-of-concept study demonstrating that the intervention could move the physiological target. SafeBoosC-3 was published in 2023 with 1,600 infants in 17 countries.

Ben Courchia (04:03.002) The primary outcome was death or severe brain injury on cerebral ultrasound at 36 weeks. The results were unambiguously neutral: 35% versus 34% in each group, showing the intervention did not reduce death or severe brain injury. This is where NIRS has lived for some time: it is a great tool, but it doesn't seem to be making a lot of a clinical difference. The SafeBoosC-3 investigators identified some explanations for this neutral result in the preamble of this paper.

Ben Courchia (04:03.002) Many sites were using different NIRS platforms and sensors, introducing variability in absolute oxygen readings. Training was recommended but not mandatory at certain sites with little prior experience. The monitoring window of 72 hours may have been too short. Additionally, the hypoxic threshold of 55% was calibrated to a specific adult sensor, the INVOS, but had to be extrapolated to different devices, adding measurement uncertainty.

Ben Courchia (06:29.966) NIRTURE was designed to address some of these gaps. SafeBoosC showed NIRS-guided treatment could stabilize cerebral oxygenation, but what if we had more consistency with the same sensors and devices? We tested the hypothesis that the burden of cerebral hypoxia and hyperoxia could be reduced by combining cerebral oximetry with a dedicated treatment guideline using NIRS from a single device. They used the SenSmart Model X100 by Nonin Medical.

Ben Courchia (06:29.966) Unlike the SafeBoosC trial, they looked at this for a little bit longer and tried to look at more outcomes. This was a multi-site, single-blinded, two-arm randomized controlled trial with 1:1 allocation stratified by gestational age (less than 26 weeks or over 26 weeks) and study sites. The trial was conducted across five tertiary units in Australia, New Zealand, and the US. This included Westmead Hospital, Nepean Hospital, the Royal Hospital for Women, Wellington Hospital, and Connecticut Children's Hospital.

Ben Courchia (06:29.966) Investigators at each site had prior experience with cerebral oximetry, and staff without experience underwent structured training and internet-based certification. Eligibility criteria included infants born at less than 29 weeks who were younger than six hours old. Exclusion criteria included congenital anomalies requiring major surgery, genetic disorders associated with neurological impairment, or multiple births beyond twins.

Ben Courchia (08:00.546) Infants were randomly allocated to standard care or the intervention group. Cerebral oximetry in both groups was performed using a single neonatal NIRS sensor placed on the forehead. Data was recorded continuously for five days, or 120 hours. In the intervention group, real-time cerebral oxygenation readings were visible to the staff, who managed the patient to keep cerebral saturation between 65% and 90%. In the standard care group, the monitor was hidden from view.

Ben Courchia (08:23.354) They used a pragmatic consensus-based approach to define the range. The lower threshold of 65% was based on SafeBoosC-2 and adjusted for higher absolute values with neonatal sensors. The upper threshold of 90% was chosen based on local data from Westmead Hospital. Sensor sites were inspected every four hours. A treatment guideline for cerebral hypoxia was activated when oxygenation was less than 65%.

Ben Courchia (08:23.354) Steps included checking the sensor and assessing if there was a decrease in oxygen delivery or perfusion. They looked at saturations, hemoglobin, systemic perfusion, PDA, and $PCO_2$. Based on these, they might increase $FiO_2$, give a transfusion, treat the PDA, start inotropes, or change ventilator settings.

Ben Courchia (10:47.426) When cerebral hyperoxia was noted (above 90%), they would wean $FiO_2$, wean pressure, or adjust the ventilator. If it was due to hyperglycemia, they would adjust the GIR. The primary outcome was the burden of cerebral hypoxia and hyperoxia expressed as percentage hours. For example, an hour event with a mean cerebral oxygenation of 55% equals 10% hours of hypoxia.

Ben Courchia (10:47.426) They ensured deviations lasted at least one minute to count as clinically significant events. Secondary outcomes included mortality, brain injury on imaging (IVH, cerebellar hemorrhage, PVL), BPD, necrotizing enterocolitis (Bell classification), and ROP. 104 infants were randomized: 53 to the intervention group and 51 to standard care. Baseline characteristics were similar, though the intervention group had a higher proportion of chorioamnionitis exposure and higher birth weight percentiles.

Ben Courchia (12:16.044) In the intervention group, the median burden of cerebral hypoxia and hyperoxia was 5.7% hours, compared to 82.3% hours in the standard care group. The adjusted relative change was 42.8% with a p-value less than 0.001. This is striking, showing that if you have the information, you are better at keeping oxygenation within range.

Ben Courchia (14:38.106) The improvement was predominantly driven by a reduction in cerebral hyperoxia, which decreased from 23.7% hours to 3.5% hours. This contrasts with SafeBoosC, where the primary improvement was driven by a reduction in hypoxia. The effect was larger in the most immature infants.

Ben Courchia (14:38.106) No significant differences were observed for secondary outcomes before hospital discharge. The adjusted relative risk for mortality was 1.28. No skin injuries occurred from the sensors. NIRS-guided treatment significantly improved the stability of cerebral oxygenation during the first five days. Neurodevelopmental outcomes are still being collected.

Ben Courchia (17:06.362) Whether this translates into better neurological outcomes is uncertain. Resolving this will require well-conducted randomized trials and individual patient data meta-analyses.

Daphna Yasova Barbeau (17:39.733) So what do you think?

Ben Courchia (17:45.038) As a clinician, I believe maintaining normal NIRS at the bedside is good. I don't see how swings wouldn't have a negative impact. But as a researcher, if the clinical metrics don't improve, do we really need to favor this intervention? It is one more data point.

Ben Courchia (18:14.956) Do we need to put NIRS on everyone? Probably not. But for certain babies, like those with hypotension or extremely low birth weight, maintaining that level might be critical. I find it hard to believe a baby swinging up and down does as well as one maintained in a normal range.

Daphna Yasova Barbeau (19:14.511)

It is a reminder that what we see on the pulse ox is not necessarily what the brain is seeing.

Ben Courchia (19:26.786)

Exactly. Cerebral autoregulation is a peculiar system that doesn't adapt very well to the rest of the body. All right, buddy, I will see you tomorrow.

Daphna Yasova Barbeau (19:44.345)

Sounds good.