Thoughts on NSAIDs for PDA

Great editorial by Carl H. Backes and Jonathan L. Slaughter on the BeNeDuctus trial results. Very interesting read. Similar to the recent text posted on here (see below - news of January 2023).

The Beneductus trial was, to my opinion, one of the best trial published in 2022. Here is my little summary of the results.

Expectant Management or Early Ibuprofen for Patent Ductus Arteriosus by Hundscheid T et al.

Why this study? Meta-analyses of RCTs showed that pharmacologic treatment with cyclooxygenase inhibitors induced patent ductus arteriosus closure but had no beneficial effect on clinical outcomes. This trial in extremely preterm infants with echocardiographically confirmed PDA assesses whether expectant management would be noninferior to early ibuprofen treatment with respect to necrotizing enterocolitis, moderate-to-severe bronchopulmonary dysplasia, or death as assessed at a postmenstrual age of 36 weeks.

How: International, multicenter, RCT noninferiority conducted at 17 neonatal intensive care units in Europe. Included are those born <28 weeks with PDA confirmed by ECHO (>1.5 mm and left-to-right shunt between 24 and 72 hours). Randomly assigned to receive either expectant management or early ibuprofen treatment. In the expectant-management group, no treatment was initiated with the intention of closing the PDA. In the early-ibuprofen group, ibuprofen was administered according to the local protocol, preferably within 3 hours after randomization. If closure had not been achieved, a second course of ibuprofen was given. The primary outcome was a composite of necrotizing enterocolitis, moderate-to-severe bronchopulmonary dysplasia, or death.

Results of the RCT: Between December 2016 and December 2020, a total of 1600 infants assessed for eligibility. 273 underwent randomization (136 expectant and 137 early-ibuprofen group. The baseline characteristics of the infants and their mothers were similar (except more HELLP in expectant group). The primary adverse outcome occurred in 63/136 infants (46%) in expectant group and in 87/137 infants (64%) in early-ibuprofen (absolute risk difference, −17.2%; P<0.001). Moderate-to-severe BPD was diagnosed in 33 (expectant) vs 51% (ibu) - absolute risk difference, −18%; 95% CI, −30.2 to −5.0). The results of subgroup analyses (<26 weeks at birth) were consistent with the overall findings.

Interpretation: Expectant management of PDA in preterm infants was noninferior to early-ibuprofen treatment at a postnatal age of 24 to 72 hours with respect to necrotizing enterocolitis, moderate-to-severe bronchopulmonary dysplasia, or death. This trial had a true non-intervention control group (with almost no open-label treatment), which allowed for a clearer comparison between expectant management and ibuprofen treatment. Of concern, the primary-outcome results suggest harm associated with early ibuprofen exposure in this vulnerable population, especially with an effect on higher BPD following early ibuprofen exposure. This is supported by in vitro and in vivo studies suggesting that angiogenesis may be inhibited by ibuprofen.

The patent ductus arteriosus in prematurity - Is it unethical to administer NSAIDs?

Disclaimer – this is only my own opinion, and I truly believe that we are all scientists that try to do the best appraisal of the evidence in order to provide the most optimal care to our vulnerable patient populations. 

After carefully reviewing the literature (some references below - 1 to 7, but there are 100s), I believe there are numerous publications outlining that 1000s of premature infants in randomized control trials have failed to show any benefits from exposing these newborns to prophylactic, early or targeted medications to attempt acceleration of ductal closure. Actually, the preliminary results of the Beneductus trial in Europe (presented at the Pediatric Academic Societies 2021 in Denver by Pr Willem de Boode), has demonstrated that there may even be an increase in harm when exposing extreme premature infants to early ibuprofen (even with a targeted approach). The study was very well done, and there was almost no open-label treatment. 

Could these medications be toxic, and not even provide any sort of benefit to these vulnerable infants? Possibly... And likely according to what I am getting from these studies. Is there a subgroup (very limited number of infants) that may benefit from accelerated closure by other means? Some of the studies looking at catheter-based approaches or paracetamol/acetaminophen are ongoing (but will unlikely show any benefits in my biased opinion). 

For all these reasons, I strongly believe that NSAIDs, are not the way to go. Our local practice has been to abandon the use of these toxic medications. They are known to cause systemic vasoconstriction (undifferentiated, reference 8-9, but there are so many articles on this). They are associated with increased spontaneous intestinal perforation, end-organ ischemia and renal dysfunction ( in part due to the renal hypoxic vasoconstriction that is sustained). 

Of course, any care bundles have to be taken into context. We do not just “ignore” the ductus, we have implemented successfully multiple care guidelines for our preterm newborns. We have an extensive golden hour protocol. We have trained all our health care professionals to be expert at the most extreme premature newborns (nurses, nurse practitioners, nutritionists, respiratory therapists, medical trainees, etc.). We have guidelines standardizing the care of our extreme premature population for the first week of life and beyond. We have clinical pathways standardizing ventilation, investigations, caffeine use, nutrition (with a rapid exposure to breastmilk and aiming to get to full feed in the first 7 days), etc. We are aggressive at obtaining breast-milk and promoting early and prolonged skin-to-skin care. Our rate of intubation in the delivery room (DR) is very low, and most of our newborns <29 weeks are brought to the NICU (including those <26 weeks) on bubble CPAP. With bCPAP, it is quite easy to install the cPAP in the delivery room and transport the baby to the unit. It is also a reliable way to sustain adequate peep, avoiding multiple disconnections and loss of FRC. (There is also some data that the frequency of oscillation, which changes with underlying pulmonary compliance, provides some benefits to the infant and more articles showcasing its benefits in the premature population). I have seen a dramatic drop in the need to intubate these infants (especially those born at 26 weeks and more). We even have newborns 23 weeks arriving to the NICU on bCPAP! These infants seem to have a much smoother transition (than what I was used to in my previous practice), as well as optimal recruitment and oxygenation. The drop of PVR occurs possibly more smoothly, allowing for less torrential left to right shunt via the ductus that eventually spontaneously constricts to various degree (until it finally closes at various ages, but often before 36 weeks). I scan almost all the <29 weeks infants in the unit at 10 days and 36 weeks, and it is extremely rare that the ductus is still patent.

We only provide intubation and surfactant in those infants with the most severe respiratory distress syndrome (our protocol has been published, references 10-11). As such, with FiO2 above 50% consistently (or other signs of significant respiratory distress or apneas), these infants get intubated and get surfactant. With a more selected population of infants exposed to surfactant, there is less acute drop in the pulmonary vascular resistance, which makes our rate of pulmonary hemorrhage low (6%) in our extreme premature population (same rate in the <26 weeks newborns). These events of pulmonary hemorrhage often occurs in infants that are outborn, and rarely yield to catastrophic decompensation since they are often these babies that remain intubated and found to have traces of blood upon suction. We do not restrict the total fluid intake (aiming 150-160 mL/kg/day total) and we aim for good nutrition for adequate growth and caloric intake. Patient gets ventilated in volume-guarantee when mechanically-ventilated and extubation occurs after the 72 hours window (in the <27 weeks newborns), once they do get intubated (we are often more rapid at extubating those 27-28 weeks infants, but they so rarely do get intubated in the first place to be honest). Extubation is deemed ready on a pulmonary perspective if FiO2 is <35%, mean airway pressure <10 and drive is present (that is not a set practice guideline). Infants that remain dependent on mechanical ventilation will receive low dose dexamethasone at >14 days (median 21 days of initiation) to get off mechanical ventilation. 

Initially, I was quite skeptical about the absence of medical attempt to close the ductus. It my hemodynamics mind, it did not make sense. I compared the rates of death or BPD of my unit to the ones of another unit and found that there may have been an increase of this rate after the adoption of the policy compared to the status quo in another unit. The other unit, though, had excessively higher rate of death or BPD all throughout the 2 periods (so very hard to increase when your baseline is already 80-90% death or BPD!). The rates of those infants exposed to a no NSAIDs policy, was still dramatically low in terms of death or BPD. Also, the unit adopting the conservative policy, had seen a dramatic increase in the admission of outborn infants 23-24 weeks in the 2015, 2016, 2017 period (following the move to a much bigger single-bed unit, fusing 2 previous units). As we all know, these infants are at the highest risk for death or BPD. The previous EPOCH using NSAIDs had very low numbers of infants born at 23 weeks (most of them being inborn and selected because of their "good profile" to survive). We now care for infants <25 weeks, <600 grams at birth that are born with already very significant medical burden (severe maternal hypertensive or rheumatological conditions, extremes of obesity or diabetes, maternal oncological processes, maternal ECMO, extremely severe growth restriction, twins with TTTS, triplets, etc.). We are the only center in my province that can provide ICU care to the mother with all subspecialty (liver/renal/cardiac/oncology), as well as all array of medical/surgical care to these infants. As such, we do attract some of the most complex cases. Also, as you all know, the association between prematurity-related averse outcomes is logarithmically associated with the degree of immaturity... so I did remain with a doubt in this relative increase in the outcome of death or BPD in the EPOCH 2...

As such, humble to dig more into the question, I recently published the updated follow-up data of our approach (adding the numbers of 2018 and 2019) since we moved to this new unit (single room). Our BPD rates kept on declining and they were actually were very good despite the high admission rate of infants 23-24 weeks that was maintained through the years. We had a 32% death or BPD in those <29 weeks… Further, we only had 7% of infants <26 weeks that had Grade 3 BPD as per the new NICHD classification! Looking at national rates, we have one of the lowest rate of mortality in this high risk population (despite those factors that I have exposed earlier regarding the complexity of the population we care for). These are amongst the best reported outcomes of this particularly vulnerable population in Canada. Our NEC rates are also consistently very low in this population as per our annual report. Our IVH rates are within national average (disclaimer: many of the <26 weeks outborn infants in our catchment area are re-routed to our centre, and as such, close to 20% of the <26 weeks admitted are outborns). The majority (>90%) of these infants underwent spontaneous closure. Since 2015, we have had only 1 premature infant (born <29 weeks) requiring a tracheostomy (and we have admitted since that time: 400 babies <29 weeks, out of which there are 40 infants <24 weeks, 50 at 24 weeks, and 60 born at 25 weeks). Further, we recently reported on our population with pulmonary hypertension, and compared to previous reports, our mortality rate within this subgroup was extremely low. There were only 24 infants with >2/3 systemic pulmonary pressure estimates at the 36 weeks echocardiography (out of a source sample of 387 infants, 8%). Reported literature quotes that mortality can be up to 25% in this subgroup, while only 3 infants did not survive among these 24...

For some reason, the PDA debate is very emotional for many. My physiological mind tells me that the left to right shunt through the ductus is probably not great for these infants with such vulnerable status in the post-natal setting. However, the attempt at accelerating its closure (with invasive methods, or with potentially toxic medications) does not confer an advantage and seem to be at their detriment. I am now quite confident about the approach of not using NSAIDS in this population. With all the studies out there, it is time to accept the null hypothesis and stop loosing so much resources, time, energy, money on studying yet-again NSAIDS that have been shown in countless well designed randomized studies to not benefit these babies! No matter how much we would like these medications to benefit them. At this point, one may argue that it is unethical to expose a patient to such medication. It is also not a judicious use of scarce research resources to try to enrol infants in yet another study that was already done, to replicate negative results (or worst, increase harm)

Regarding cath-closure, I am worried that we will be intervening on very small infants that are transitioning, potentially exposing them to interventional complications for no benefits. I hope to be wrong... I have seen some infants post-embolization of the occluder (not good)... I have also seen some premature infants loosing a limb due to vascular injury when attempting ductal closure at early age. I have heard some centres with mortality and low output from valvular injury, atrial rupture and sustained arrhythmia... 

These are my humble opinions, that will certainly trigger many disagreements ;). I more than welcome more discussion - My Twitter: @CardioNeo

The Controversy Surrounding PDA in Prematurity

Despite countless trials conducted in various forms (and a lot of money invested in running these trials), reflecting real-world neonatal practice, none have shown a meaningful improvement in outcomes. In fact, they have highlighted the potential harm associated with these interventions. We cannot conduct research solely to "prove" what we wish to be true and then disregard the results when they do not align with our expectations. Science demands that we remain objective, follow the evidence wherever it leads, and adapt our practices accordingly. This is a recording of a talk I gave to our neonatal hemodynamics fellows. Perhaps I can persuade you to reconsider practices that do not appear to benefit some of the most vulnerable babies we care for and instead advocate for evidence-based approaches rooted in science. More importantly, this is also a call to pursue research and explore new strategies to continuously improve outcomes for these fragile newborns. Hope you enjoy it!

Recent Trials results

1. Gupta S, Subhedar NV, Bell JL, Field D, Bowler U, Hutchison E, Johnson S, Kelsall W, Pepperell J, Roberts T, Sinha S, Stanbury K, Wyllie J, Hardy P, Juszczak E; Baby-OSCAR Collaborative Group. Trial of Selective Early Treatment of Patent Ductus Arteriosus with Ibuprofen. N Engl J Med. 2024 Jan 25;390(4):314-325. doi: 10.1056/NEJMoa2305582. PMID: 38265644.

2. Hundscheid T, Onland W, Kooi EMW, Vijlbrief DC, de Vries WB, Dijkman KP, van Kaam AH, Villamor E, Kroon AA, Visser R, Mulder-de Tollenaer SM, De Bisschop B, Dijk PH, Avino D, Hocq C, Zecic A, Meeus M, de Baat T, Derriks F, Henriksen TB, Kyng KJ, Donders R, Nuytemans DHGM, Van Overmeire B, Mulder AL, de Boode WP; BeNeDuctus Trial Investigators. Expectant Management or Early Ibuprofen for Patent Ductus Arteriosus. N Engl J Med. 2023 Mar 16;388(11):980-990. doi: 10.1056/NEJMoa2207418. Epub 2022 Dec 6. PMID: 36477458.