Retinopathy screening is undoubtedly painful for preterm babies, formal evaluation with PIPP scores routinely exceed 10 during screening, and may exceed 14, meaning moderate to severe pain. As a planned procedure there is always opportunity for pre-emptive analgesia prior to and during a screening exam, but what? Many of the interventions studied have limited efficacy.
A new systematic review has examined the efficacy of “pharmacologic” methods of pain control, and therefore, using their definition of “pharmacologic”, excluded breast milk, sucrose, swaddling; I know there has been some debate about whether sucrose should be considered “pharmacologic” or not, but that was the definition they used.
The review, which is unfortunately lacking in cute Forest plots to copy into this post, confirms that topical anaesthesia has limited effectiveness, there are 4 controlled trials that they found, with a mean PIPP score after local anaesthetic of between 10 and 15, reduced by an overall 1.6 points, which although not likely to be due to chance, is a very small reduction.
The other studies they reviewed all used topical anaesthesia in both groups, and examined other additional analgesia. They found 3 trials with acetaminophen (paracetamol to the Europeans) compared to control or placebo or sugar or milk. Overall, 2 of the trials showed a reduction in PIPP scores during the procedure, the 3rd reported PIPP scores after the procedure, which ere also reduced. The 3 trials used 15 or 20 mg/kg of acetaminophen given either 30 or 60 minutes prior to the procedure. As for opiates there were 2 trials of oral morphine which show relatively little effect, and one of intranasal fentanyl which appeared effective, when given in combination with sucrose and topical anaesthesia compared to that combination without fentanyl (Sindhur M, et al. Intranasal fentanyl for pain management during screening for retinopathy of prematurity in preterm infants: a randomized controlled trial. J Perinatol. 2020;40(6):881-7). That was a nice masked study with 50 babies per group, and has the advantage over acetaminophen that it works quickly, I don’t think it has become widespread, but why not?
You can see from that figure, taken from the aforementioned publication, that intranasal fentanyl was rather effective. You can also see from the controls that despite topical anaesthesia, swaddling, and sucrose, retinopathy screening examinations still hurt.
I think that either acetaminophen or nasal fentanyl given before the procedure warrants either routine introduction or more study. I’d like to see confirmation of the efficacy and safety of routine intranasal fentanyl before introducing the practice to large numbers of stable preterm infants, but it does seem effective from that one study, and the dose of 2 microg/kg as a single administration in the nose appears from other data to be safe. The reduction of pain scores with acetaminophen is interesting but it isn’t apparently as effective as fentanyl, although a comparative trial would be nice.
My optimal protocol for the present would be, administration of 15 mg/kg of acetaminophen 30 to 60 minutes prior to the procedure, at the time of application of topical anaesthetic and mydriatic, swaddling of the baby and administration of sucrose 2 minutes prior to the procedure, a second dose of sucrose with a soother just prior to speculum insertion, then repeated sucrose if the procedure takes more than 2 minutes, and trying to avoid scleral depression if possible.
The dose of caffeine that we used for the CAP trial was the dose that was being widely used at the time. It seemed to be a safe dose, that did not require serum concentration surveillance, but was not based on a large number of high quality dose ranging trials, indeed it was identical to the doses used in the very first publication of caffeine use in the preterm infant, published in 1977 by Jack Aranda, from Montreal (the McGill side of the mountain). (Aranda JV, et al. Efficacy of caffeine in treatment of apnea in the low-birth-weight infant. J Pediatr 1977;90(3):467-72), they used 20 mg/kg of caffeine citrate as a load, and 5 to 10 mg/kg/dose once or twice a day, starting 48 to 72 hours after the load.
The maintenance dose gradually stabilised over the years, although there have been repeated questions about whether it is the optimal dose. There were some pharmacokinetic studies, but few quality pharmacodynamic studies, those that existed rarely used objective quantification of apnoea frequency, which is essential to be able to say much about the impacts on apnoea, nursing records of apnoea spells being notoriously unreliable.
The CAP trial remains the standard, which showed an improvement in medium term, and very long term, outcomes after caffeine use among infants of less than 1250 g birth weight, less than 10 days of age started on caffeine (or placebo) because the attending physician thought they needed caffeine. The dose in that trial was 20 mg/kg of caffeine citrate as a load, and 5 mg/kg daily, that could be increased to 10 mg/kg of caffeine citrate daily if the physicians thought the baby needed more. I submitted an abstract to a PAS meeting which showed that babies who had the dose increase had the same advantage of caffeine as those who remained on the initial dose, which I never fully published (sorry!) but I think was reliable information that the higher dose of caffeine was safe.
Clearly if some caffeine is good, then it is possible that much more caffeine might be even better. If we can keep babies extubated for longer periods of time, and knowing that several animal models show brain protective effects of caffeine, then what dose of caffeine should we give?
In this study in 3 day old rats, for example, the animals received 20 mg/kg/d of citrate for 5 days starting the day before a classical carotid-ligation-hypoxia model, and they had less white matter injury. In another study in newborn mice with an IVH model, caffeine at 20 mg/kg/day for 3 days started after the IVH reduced brain injury and brain atrophy. In another study Jack Aranda returns to caffeine 45 years after the first publication (!), comparing neuroprotection in newborn rats who received a dose similar to the usual human preterm dose (20 mg/kg load of Caffeine citrate followed by 5 mg/kg/d) or a larger dose of 80 mg/kg load followed by 20 mg/kg/d for 12 days. They showed similar neuroprotection with the 2 doses.
But before we ramp up the caffeine dose too far, remember the results of the pilot trial from St Louis, which randomized babies to get 80 mg/kg as the loading dose, starting in the first 24 hours of life, compared to 20 mg/kg in the standard dose group. The 80 mg/kg was actually given over 36 hours, as 4 doses, 40 mg/kg as the initial load, 20 mg/kg 12 hours later, then 10 mg/kg 24 and 36 hours after the first dose, the controls received 20 mg/kg then 10 mg/kg 24 hours later. Both groups in that small pilot (n=37 per group) received the same maintenance dose 10 mg/kg/day of caffeine citrate. The results of that trial showed a greater seizure burden with high dose compared to standard dose, and the high dose babies also had more cerebellar injury; fortunately the 2 year and 5 year outcomes were very similar between groups.
This was actually a slower load than in the previous trial by Steer et al who gave a 80 mg/kg bolus over 15 minutes, and then a maintenance of 20 mg/kg/d compared to their controls who got 20 mg/kg load then 5 mg/kg/d. In that study, the 240 babies of <30 weeks gestation were a little older (2 to 12 days of age, average 4 days) when enrolled and the study was designed to look at extubation failure. The primary outcome, extubation failure was less frequent in the high dose group, and there was a little less BPD in the high dose group 34 vs 48% RR 0.72 (95% CI 0.52-1.01). The medium term outcomes at 1 year of age tended to be better in the high dose group, and are described in more detail in this publication. The developmental quotient from the Griffiths scale is a slightly higher in the high dose group, but I don’t see any publication with later follow up.
There are a couple of smaller studies randomizing babies to higher doses of caffeine. In one of them, published in Chinese so I can only read the English abstract, 162 ventilated infants <32 weeks were randomized to different maintenance doses, they all received 20 mg/kg load before 6 hours of age (presumably of caffeine citrate) then either 5 or 10 mg/kg (presumably the daily dose of caffeine citrate); it appears that the higher dose group were less likely to fail extubation, which was mostly because of apnoea. An Egyptian trial (fortunately for me published in English) randomized 120 ventilated babies <32 weeks to receive either a load of 40 mg/kg of caffeine citrate and a maintenance of 20 mg/kg/d, or a load of 20 and maintenance of 10. They showed less extubation failure in the high dose group, but there is no longer term follow up. Another study from China randomized ventilated infants <30 weeks gestation who were over 48 hours of age and thought to be within 24 hours of an extubation attempt to a maintenance dose of either 5 or 10 mg/kg of caffeine citrate starting 24 hours after the load of 20 mg/kg, which was the same in each group. Extubation failure was decreased with the higher maintenance dose.
Putting this together to me this suggests that very early, very high loading doses of caffeine might be risky, that increasing maintenance doses to 10 mg/kg/d is probably safe and beneficial, with an increase in successful extubation and less apnoea, but limited long term outcome data. Increasing the loading doses after the first couple of days of life probably improves extubation success, but without more safety data I would be reluctant to use very high loading doses as a routine.
One of the particular features of caffeine kinetics is a dramatic increase in clearance as babies approach term. Caffeine is mostly filtered by the kidneys unchanged in early postnatal life of the preterm, and there is a gradual development of various pathways of demethylation with maturation, acetylation probably develops even later. The half life, as a result, is often over 100 hours in the early preterm period, falling to about 4 to 6 hours in an adult.
If we are planning to give caffeine to babies who are more mature, as they approach term, doses will therefore probably have to be adjusted. A trial which randomized 95 preterm babies who were stopping caffeine at >33 weeks, to either usual care without caffeine or to restart it 5 days later with a loading dose of 20 and maintenance of 6 mg/kg showed that the babies had fewer intermittent hypoxic spells in the caffeine group, until they reached 37 weeks PMA, by which time there was no longer much impact of the caffeine, I think this may have been partly because of a lack of power, as the controls were having fewer spells, but may also be because caffeine clearance was rapidly increasing and the infants needed more. The authors of this study therefore enrolled a second cohort of 27 similar babies who received higher doses, which they compared to the controls from their first study. The doses in the 2nd publication were started just 24 hours after stopping clinically required caffeine, and were 10 mg/kg/d of caffeine citrate, increased at 36 weeks PMA to either 14 or 20 mg/kg/d, decided by random allocation. The doses were chosen to try and maintain salivary caffeine concentrations at above 20 microg/mL, thought to be a reasonable therapeutic target for efficacy and safety. In this trial there was a reduction in intermittent hypoxic spells compared to the controls from the previous study.
In a small short term study like this it is not possible to say whether there was a clinical benefit to the babies of having fewer intermittent hypoxic spells. As far as I can see there is no long term follow up of these babies published, but it would have very little power anyway.
All of which is a preamble to a newly published study from Auckland examining the use of caffeine in late preterm infants. We don’t normally worry about these babies in terms of apnoea of prematurity, even though they clearly very commonly have apnoeic spells, most of the spells are brief and resolve spontaneously. However they are accompanied by episodic hypoxia, and repeated hypoxia and re-oxygenation leads to an oxidative stress, which might have adverse long term effects. We also know that the neurodevelopmental long term outcomes of late preterm babies are different to those at term, with more cerebral palsy, and more schooling difficulties.
They showed fewer events at 2 weeks of age with the caffeine dose of 10 or 20 mg/kg/d compared to placebo (but not with 15 mg/kg/d), heart rates and episodes of tachycardia were increased in the caffeine groups. Presumably the lack of effect of the 15 mg/kg dose is just a random effect, as the groups were relatively small and the frequency of intermittent hypoxia spells is very variable.
This study confirms that to have an impact on respiratory drive and hypoxic spells near to term, the doses required are substantially higher. What it does not tell us is whether we should be doing this!
If you are caring for an individual child with troublesome clinically important apnoeas who is approaching term, and you want to treat with caffeine, you will need a higher dose to reduce the number of spells, because of the changes in metabolism.
What we need to know is whether it is safe and effective to give high doses of caffeine to very preterm infants as they approach term to reduce such spells. Even if, as seems likely, we can reduce intermittent hypoxia with higher doses of caffeine with routine continuation towards term, does it actually improve longer term outcomes. Even more important in some ways, given the very much larger numbers of at-risk infants, is it safe and effective to routinely treat late preterm infants to reduce their frequent hypoxic spells?
I am sure Dr Alsweiler and colleagues in Auckland are hoping to follow up the Latte Dosage trial with a large multicentre RCT examining long term impacts of high dose caffeine in late preterm infants, lets hope they get the funding. We tried a couple of times to get NIH funding for a trial in very preterm babies as they approached term, but the trial, as it was designed, was hugely expensive and never got high enough priority. A simpler pragmatic trial would be very valuable, and could inform our community about what to do with caffeine for the preterm infant approaching term.
A new study from Finland this time, raises similar concerns to those from a Canadian study. (Raikkonen K, et al. Associations Between Maternal Antenatal Corticosteroid Treatment and Psychological Developmental and Neurosensory Disorders in Children. JAMA Netw Open. 2022;5(8):e2228518). It is an analysis of the same dataset that the authors previously published about; in that previous study they lumped together developmental and behavioural disorders, and did not analyse neurosensory difficulties. Using the same enormous linked databases with over half a million children, of whom just over 2% had been exposed to antenatal steroids (presumably all fluorinated steroids for lung maturation). The authors state that they have previously validated that these cases are all indeed betamethasone in threatened preterm delivery. This is important of course, because people take steroids for all sorts of reasons, even during pregnancy, and most are metabolized by the placenta and have no direct fetal or neonatal impact of note.
This time they also differentiated between those who delivered at term and those who were actually preterm. The graphic below shows that most of the outcomes were similar between preterm exposed and non-exposed to steroids, even though the actual gestational age at birth was lower in the preterm babies who had been exposed to steroids (mean 32.8 (SD3.0) vs 35.5 (SD1.7) weeks). Among the babies who delivered at term the exposed and unexposed had very similar GA (39.3 vs 40.1) and other characteristics, but had more adverse outcomes, in just about every domain, with cerebral palsy being the most striking.
The absolute risks are small, however, for CP, for example, the Hazard Ratio is over 2, but the absolute percentages are 0.4 compared to 0.1.
What should we do about these data? I think that we should be more circumspect about steroid use in the late preterm, especially after 35 weeks when the benefits are small. Also we need to find ways to better target steroids to those who are more likely to deliver preterm.
About 60% actually delivered preterm in the study, only 30% prior to 32 weeks gestation, and another 10% prior to 34 weeks. About 20% in each group required surfactant, and the 95% confidence intervals for the difference in surfactant requirements crossed the non-inferiority boundary, leading the authors to conclude that they were unable to show non-inferiority.
Although this is an important, well-designed study, I do have doubts as I said, about the primary outcome. We don’t really give steroids to avoid giving surfactant, but for all the other benefits on maturation, and most importantly a reduction in mortality in the most immature infants, but a study designed with a mortality outcome would have to be absolutely enormous, so overall, I think needing surfactant is a reasonable proxy outcome.
What about the other important neonatal outcomes?
As you can see there is no clear difference in any of the outcomes, with the exception of the combination outcome among infants who were born within a week of getting the steroids.
It looks like the lower dose is just about as good as the full dose, but, being stringent in our interpretation, they have not shown non-inferiority, the outcome was a little bit more frequent in the low dose dose group, and is consistent with a poorer efficacy. Of note there were only about 250 in each group that delivered within 7 days of receiving steroids.
I was encouraged to read that the study was funded for long term outcomes, with a 5 year assessment including WPPSI and NEPSY subsets and a neuro exam. Also that a Canadian and Australian trial is underway SNACStrial.com with short and long term outcomes being investigated.
It is going to take a while, but we will eventually have good quality data about whether halving the dose of betamethasone is as effective, and potentially safer, than the current dose. If you remember, the currently used dose was derived directly from the original Liggins studies in sheep, and there has never been, until these new trials, any dose response data. It is about time.
In this trial called “MINVI”, babies of 35 to 42 weeks GA were delivered vaginally or by cesarean, and evaluated during the first 15 seconds, prior to cord clamping. Those with pallor, poor tone, or apnoeic were then exposed to the intervention. If they were vigorous they had usual care (natural cord management).
Hospitals were randomized to one of two approaches, and, depending on the randomization, babies who qualified then had either early cord clamping (at the latest by 60 seconds) or had cord milking, which was performed with the cord intact, the cord being milked 4 times over a 2 second period each time, and milking about 20 cm of cord. The cord being clamped at a median of 29 seconds in the milking group (IQR 20s- 30s), compared to 20 seconds in the early clamping group (IQR 10s-20s). These durations are presumably estimates, as I don’t think there was someone assigned with a stopwatch to time the interventions (which is why the numbers are suspiciously round numbers!). Once half the study was completed, which took about a year, hospitals then were switched to the alternate approach (with a 2 month “washout” period).
The primary outcome of the trial was admission to the NICU for the following reasons: “respiratory distress (tachypnea, grunting, retractions), bradycardia or tachycardia, hypotonia, lethargy or difficult to arouse, hypertonia or irritability, poor feeding or emesis, hypoglycemia, oxygen desaturations or cyanosis, need for oxygen, apnea, seizures or seizure-like activity, hyperbilirubinemia, and/or temperature instability”. Admission just for observation or for antibiotics or because of low cord pH, for example, was not considered. There were numerous secondary outcomes, including principally HIE, and other outcomes that could reasonably be impacted by the intervention, such as jaundice and hemoglobin levels.
The cord milking group were less frequently admitted to the NICU for the above reasons, 23% vs 28%, however, after adjusting the analyses for centre, the confidence intervals of the adjusted Odds Ratio included no difference between groups, OR 0.69, 95% CI 0.41-1.14. Respiratory distress leading to NICU admission was less frequent with cord milking, other secondary outcomes are shown below
As you can see the cord milking babies were less likely to need respiratory support. There was also less moderate HIE, leading to less cooling, numbers of these neuro outcomes were small which is why there is no adjusted Odds Ratio.
In the strictest sense this is a null study, with a primary outcome in the two groups being within the usually accepted limits of a possibly random effect. It certainly shows no adverse impact of the procedure, apart from the minor increase in bilirubin.
Anup Katheria, the principal investigator of this study, has previously published a review article (in 2018), which is open access, available via PubMed Central. He reviews the then available data, and the rationale which formed the basis for this study. I must say there is much less data from animal models regarding cord milking, and what is available is not reassuring, one study from Stuart Hooper’s lab, that I recently referred to, studied preterm lambs, (Blank DA, et al. Haemodynamic effects of umbilical cord milking in premature sheep during the neonatal transition. Arch Dis Child Fetal Neonatal Ed. 2018;103(6):F539-F46) and showed that if you did the procedure in one of the two ways they examined, they did not find much evidence that you actually increased blood volume. In that lamb study the milking was done in 2 different ways, the first was to release the cord between milks, so that it could refill from either end of the cord. The second method was to milk the cord, then keep it occluded near the lamb, so that it refilled from the placental end, then it was milked again. With the second method there was a net transfusion of about 9 mL/kg of blood. They also showed major haemodynamic fluctuations during the procedure, with blood pressure shooting up and down during milking.
You can see some of those impacts in this figure, of note there also was no net placental transfusion with physiological based cord clamping.
Although these impacts are concerning, the limitations of this lamb model are demonstrated by the lack of net placental transfusion with clamping after 3 minutes of positive pressure ventilation in the “physiological based clamping” group. Preterm babies with delayed clamping do have higher hematocrits and evidence that they receive a transfusion.
It isn’t clear to me whether the technique used in the MINVI trial is more reflective of the with or without placental refill group in the lamb trial, if I was doing it I think I would tend to hold the baby end of the cord closed with my fingers while the cord refilled from the placental end, (which is what Hooper’s group called WITH placental refill). It may be that the fluctuations in blood pressure and so on are less important in full term babies than they might be in the preterm.
What should be the response to these trials in the clinical realm? It might be too much to ask for a large RCT comparing clamping after PPV to cord milking in babies at or close to term who are non-vigorous at birth. In order to get an adequate sample size, an approach like MINVI, with cluster randomization and deferred/waived consent will be necessary I think.
That is probably the only way to resolve the conundrum, for now, based on these recent trials it looks like either approach would be acceptable, and both seem to be at least as good as early clamping, with no disadvantages of consequence, and all the differences being in favour of the alternative, physiology based clamping or cord milking. Personally, the physiology based clamping is something I find a more pleasing idea, in terms of the physiology, but it looks from MINVI that you can get the majority of the advantages, at least among term and late preterm infants who are non-vigorous.
Delaying cord clamping until respiration is well established is a physiologically pleasing approach, and avoids the dramatic decrease in left ventricular preload, simultaneously with an increase in afterload that occurs with early clamping. But does delaying clamping during positive pressure ventilation have the same physiologic benefits? I had previously thought that the decrease in intra-thoracic pressure associated with an inspiratory effort might increase placental venous return, and that therefore positive pressure ventilation with the cord intact might not have the same benefits. That is probably wrong, although the differential effects of delayed cord clamping with spontaneous respiration and assisted ventilation are not clear to me.
What is clear is that positive pressure ventilation decreases pulmonary vascular resistance, although surprisingly, we don’t really understand the mechanism. Stuart Hooper’s group has done much of this work and in one fascinating study (Lang JA, et al. Increase in pulmonary blood flow at birth: role of oxygen and lung aeration. J Physiol. 2016;594(5):1389-98), they showed that positive pressure ventilation of one lung with nitrogen, causes improved lung perfusion, of BOTH lungs. This was a study in near term fetal rabbits who were instrumented during partial cesarean delivery with the cord intact, but the actual procedures and images were taken after cutting the cord. So it tell us about the physiology of PVR reduction during positive pressure ventilation, but not about other aspects of delayed clamping and ventilation.
The fetal rabbit kits were ventilated unilaterally in the right lung with nitrogen or air or oxygen, then unilaterally with air (1LV2) then the tube was pulled back to ventilate both lungs with air. This is one selected image from the publication, showing the number of vessels that were seen in each lung, and that ventilating the right lung increased perfusion of both lungs, ventilating the right lung with oxygen increased perfusion further, especially of the right lung.
When you are doing physiologic studies in animals it is difficult to ensure that the animals make respiratory efforts reliably at the right moment, so most studies are about positive pressure ventilation. I guess in some ways it is less important for the future of delayed cord clamping what happens during spontaneous respiration, as it has become the standard of care to clamp the cord after at least one minute, if the baby is breathing. The responses to clamping before or after initiating PPV are of more relevance for the decision that we are still considering, whether we should routinely initiate PPV prior to cord clamping in depressed babies. The recent studies have not suggested any reason to me why we should clamp before PPV, if that is technically, logistically possible. In cases of an increased risk of needing PPV, I think the recent trial from Melbourne shows that it is not too difficult to get organized to do this. See the comment on my previous post from Doug Blank. (I’ve never linked to a comment before, hope that works).
I was thinking, based on the physiology, and the BabyDUCC trial, that we should all prepare to perform the initial steps of resuscitation during “natural cord management”. But hang on, what about an alternative… cord milking? (see next post!)
The Auckland group has been studying the treatment and implications of neonatal hypoglycaemia for many years now, with unique high quality studies. Two of their recent publications have addressed the safety of glucose gel for hypoglycaemia, the first (St Clair SL, et al. Effect of prophylactic dextrose gel on the neonatal gut microbiome. Arch Dis Child Fetal Neonatal Ed. 2022;107(5):501-7) was a substudy of the hPOD trial of prophylactic glucose gel in at-risk babies. If you remember, the original trial enrolled late preterm babies and term babies who were Infants of Diabetic Mothers, or small or large for GA. The trial showed a reduction in the number of babies with hypoglycaemia (<2.6 mmol/L) from 42 to 37%, but no difference in NICU admission (which was the primary outcome). Long term follow up of the hPOD trial, which I already blogged about, showed no major difference in neurosensory outcomes, except for an increase in motor delay in the active treatment group. This substudy examined the impact of glucose gel on the development of the intestinal microbiome, and showed no real difference in bacterial diversity between gel-treated, placebo-treated, and untreated non-randomized controls over the first 4 weeks of life. They did show the expected differences between vaginally delivered and cesarean delivered babies, and despite the tiny numbers of mothers not breast feeding, they showed impacts of milk source on the microbiome also. This is reassuring data that the intervention does not impact gut colonization.
The primary outcome occurred in about 38% of each group. Which seems like a lot! There are in particular many infants with MABC scores below the 15th centile, 25% vs 34% in the gel and placebo groups respectively. There are also many actively treated babies with VMI scores below 85, which is a standard deviation beneath the standardized mean, 24% glucose vs 15% placebo. There were a very large number of comparisons, so to find one that was “statistically significant” is hardly a surprise, but it does suggest that there is a need for further studies and further evaluation to see if the difference between groups in VMI scores is reproducible. The high rate of “neurosensory impairment” warrants evaluation, is that similar to other New Zealand infants at this age? I would have expected about 15% below -1 SD for each test, with a lot of overlap, so it is hard to guess how many in the general population would satisfy that definition of “NSI”. It makes me think that infants eligible for the study, and at risk for neonatal hypoglycaemia, are also at increased risk for these outcomes, regardless of how they are treated; which is consistent with other data I have discussed before.
Being maximally critical you could say that there are some minor indications of a possible adverse effect of oral dextrose gel, with a hint of a possible adverse effect on visual motor integration. Could this possibly be related to the effect that the same group reported a few years ago, infants with hypoglycaemia who had glucose therapy and had a more rapid rise in their glucose levels appeared to have an increase in risk of neurosensory impairment (RR 1.8)? This certainly isn’t an argument against glucose therapy, nor against glucose gel therapy, which I think is an advance that can spare many babies from intravenous infusions, but perhaps the first approach should be feeding with breast milk rather than glucose, at least for the majority of babies who have lowish blood sugars, above the threshold that HypoEXIT suggests is safe (2.0 mmol/L). It looks possible that a too rapid increase in blood sugar might have some minor adverse effects. Trying to balance this against the adverse effects of severe hypoglycaemia, and the apparent association of hypoglycaemia risk factors with poorer outcomes is a hugely complex undertaking.
There remains some doubt about the impacts on cerebral development, and thus on developmental progress of anti-VEGF treated babies. It is possible that there are cerebral effects of VEGF inhibition (or interception) and it isn’t clear what the exposure of the brain really is. In this new trial (Stahl et al, that I just posted about), most babies had undetectable serum aflibercept concentrations throughout the study, which is reassuring. Data from other observational cohort studies has shown some association between anti-VEGF “-mab” administration, and poorer outcomes. In this study from the NICHD network, which Roger Soll refers to in his editorial accompanying the FIREFLEYE trial publication, their long term primary outcome, “death or severe NDI” was not much different between bevacizumab and surgery treated babies (adjusted Odds Ratio was 1.42; with 95% CI 0.94 to 2.14). However, mortality was somewhat higher (most of the deaths, and the difference in deaths, was prior to hospital discharge; 9% of the treated babies vs 3.5%), and Bayley III cognitive composite scores were shifted lower by a mean of 3 points, so somewhat more babies were under 85 (aOR 1.78 [95% CI 1.09 to 2.91]). The motor scores were slightly lower in the bevacizumab group, and the language scores were almost identical. The babies in the bevacizumab group also had longer assisted ventilation, oxygen therapy, and hospitalisation, so perhaps they had more severe BPD.
Published follow up of the CNN and CNFUN Canadian cohort also shows lower scores among the bevacizumab treated babies compared to laser surgery, but the pattern is different, cognitive scores were identical between groups, but language and motor scores were worse. There were very few deaths between treatment and discharge. The CNN bevacizumab babies also seem to have been a bit sicker in terms of their lung disease than the laser treated infants. The difference in severity of lung disease is understandable, as bevacizumab was emerging as a treatment during these periods, we tended to use it for babies who we were most worried about deteriorating during surgery, so the simplicity of bedside intravitreal injections made us prefer it among the babies with the most unstable pulmonary status.
The real impacts at long term will require follow up of randomized trials, at present the observational data are conflicting and confusing, in addition to the two large multicentre cohorts discussed above, there are several others, and a systematic review in 2020 that found a total of 8 studies, actually showed no overall difference in outcomes. Another systematic review, (Kaushal et al) also published in 2020 included data from 13 studies. There are however, errors in their meta-analysis which I have just noted as I was reviewing it now. They calculated the mean and SD of the data from the CNN, estimating them from the median and IQR which were published. But the SDs they calculated are tiny, and are, I think, incorrect, using the method they claim to have used (I checked with an online calculator which is supposed to be based on the method the authors used, but gives SDs which are larger and more reasonable). As a result the CNN data are given huge weight compared to the other studies, which means that their calculations for the continuous outcomes are in error, I think.
A caution about the introduction to the trial, as you can see from the conflicts of interest statement, there was heavy involvement of Novartis in the trial. Novartis markets both bevacizumab (Avastin) and ranibizumab (Lucentis), which are derived from the same antibody. Novartis, however, want everyone to use ranibizumab for eye injection because they charge many times as much for it, they prepare it in the small doses needed for eye injections, and, according to them, it is less likely to cause systemic effects. The molecules are very different in size, with ranibizumab being a much smaller molecule which is probably cleared from the circulation much faster, which might possibly make it safer. Comparative trials in adults, however, haven’t shown much difference in either efficacy or complications. Bevacizumab is in fact “non-inferior” to ranibizumab in adult wet macular degeneration efficacy. (Moreno TA, Kim SJ. Ranibizumab (Lucentis) versus Bevacizumab (Avastin) for the Treatment of Age-Related Macular Degeneration: An Economic Disparity of Eye Health. Semin Ophthalmol. 2016;31(4):378-84).
All that being said, the first author is Neil Marlow, and I am sure he wouldn’t have written this unless he was sure about the data. These long term outcomes show equivalent development in almost all domains, but ranibizumab led to much less high myopia, and less ocular structural abnormalities than laser. There is very little follow up of the bevacizumab, trials, the BEAT-ROP trial only reported follow up from 16 infants from one centre, and therefore no power to show anything. Kennedy KA, et al. Medical and developmental outcomes of bevacizumab versus laser for retinopathy of prematurity. J AAPOS. 2018;22(1):61-5 e1. Even though cognitive scores were 20 points higher among those who received bevacizumab rather than laser there was such a wide range, and small numbers, that this might well have been a chance difference.
The long term, therefore, must be classified as uncertain, with some concerns from observational studies, and some reassurance from the little data available from RCTs. Which is a very unsatisfactory state of affairs when trying to counsel parents.
A newer anti-VEGF drug has been invented, and evaluated in retinopathy therapy. This new drug aflibercept works differently to the “-mab” drugs we have been using. Those others are monoclonal antibodies (hence mab) directed against VEGF, whereas this new stuff is some sort of protein that mops up VEGF (intercepts it, I guess, to get its generic name). It is used for therapy of colon cancer (and approved for this indication a few years ago) and for wet macula degeneration, much like bevacizumab. I don’t know what the price is for the tiny doses used for newborn retinal disease, but I guess it won’t be cheap!
You can see that there was little difference between the 2 groups, the aflibercept was slightly better in terms of treatment failure, or at least it was not worse. But according to the twisted syntax of the non-inferiority trial, it wasn’t not-worse!
The explanation of that is that they planned the trial such that, if the lower limit of the 95% confidence interval for the difference between groups was above a -5% difference, they would conclude that aflibercept was not inferior. But the difference between groups was 3.4%, with 95% confidence intervals -8% to infinity. Which is bizarre. A 95% CI up to infinity means that there is a 5% chance that aflibercept is more than infinitely better than laser!
The control group was quite small in this trial, and the failure of laser therapy was lower than their sample size calculations expected. Of the 43 randomized to laser, only 38 actually had laser, and for some reason the primary analysis is not by Intention to Treat, but only includes the 38 actually receiving laser. I can’t see anything in the statistical analysis plan to justify not including the 5 who did not receive laser treatment, but, as far as I can see, they at least weren’t analysed as intravitreal injection subjects (I don’t know what treatment they actually received, if any). The largest numbers of patients were from Russia (18) and Japan (17) with smaller numbers from Turkey, Bulgaria and Romania, and then very small numbers from each of a variety of other countries around the world.
One difference between this trial and the other “-mab” trials such as RAINBOW, is that the laser therapy had fewer failures, they thought that around 72% of laser therapy would be successful, whereas they actually had 82% success. The study was, as a result, underpowered. Definitions of failure were not very different between groups, but in RAINBOW there were many more babies in the laser group (18/74) who failed and then received rescue rabinizumab.
The authors try to explain the difference in failures, but I don’t think it needs any, in such a small group of controls 72% and 82% are almost identical proportions, they had approximately the proportion of treatment failures that would be expected.
What that means is, that although aflibercept was not inferior to laser in the primary outcome expressed as a simple percentage, we cannot say with more than 95% confidence that it isn’t worse than laser, the results are compatible with the possibility of aflibercept having a greater failure rate than laser.
If you want to understand a bit more about non-inferiority trials, you could do worse than watch the NEJM youtube video (not words I thought I would ever write) . Or better still, read this article from the NEJM from a few years ago, which includes the following figure for interpretation of different possible results.
You can see the figure doesn’t include an example which matches this trial (called FIREFLEYE), it does have an example which it calls “noninferiority and inferiority” which is a confusing phrase, “the test treatment is worse but at the same time not worse”, which could be restated, ‘the test treatment was worse, but the confidence intervals did not exceed our prespecified margin, so it might be an acceptable alternative”.
The type of findings that this study had, could be represented by the open circle in the imaginary results below, which I guess could be labelled “superiority but not noninferiority”, (smiley ironic emoji), or you could say “the results were numerically better, but not better enough to be sure they are better and, in fact, they are statistically compatible with a chance that they are worse”, or maybe we should just say “inconclusive”.
The study was underpowered to be able to be confident about the relative efficacy of aflibercept. It looks to be probably about as effective as laser, when compared to a group of babies in whom laser worked fairly well. Long term ocular and visual outcomes will, I guess, likely be much better than laser, if they are similar to the longer term ocular and visual outcomes of rabinizumab or bevacizumab, but of course that remains to be proven with longer follow up. Long term outcomes other than ocular/visual outcomes also need to be studied, as we will see in the next post.
Delayed cord clamping is now standard of care for all deliveries, at term or preterm. In term deliveries it leads to improved iron stores in the baby which difference persists during the first year of life. In very preterm infants mortality is reduced, and there is some reduction in many complications of prematurity; somewhat fewer cases of IVH, NEC, RoP (each of which, individually, are small reductions and potentially chance findings) and fewer transfusions.
Delaying clamping until pulmonary ventilation is established allows the reduction in pulmonary vascular resistance and increase in pulmonary blood flow and left heart preload to occur. It makes much physiologic sense therefore to ensure that the cord is left intact until this occurs, which complicates resuscitation in babies who don’t start breathing during the usual 60 seconds of delay in cord clamping. It is therefore common, for babies who don’t start to breathe spontaneously early, to clamp and cut the cord and take the baby to a resuscitation table to provide positive pressure ventilation.
Even thinking about how to do such a study makes my head hurt, you need to enrol mothers at increased risk of needing neonatal resuscitation, and then randomize the babies who do actually need resuscitation to either have the usual approach (immediate clamping) or to have initial steps of resuscitation, at least attempts to establish pulmonary ventilation, with the cord intact. But very few babies who are at risk actually need positive pressure ventilation, so you have to put in place the resources required in order to screen nearly 1000 babies, enrol about 500, and finally randomize only 120 of them, as they did in this trial. Those enrolled but not randomized are babies who have delayed clamping and don’t need assisted ventilation. So the team had the great idea of collecting saturation and heart rate data from them in order to update the percentiles for normal transition, using data from infants with delayed clamping. The percentiles we all currently use are from babies with immediate clamping.
Those percentiles are here for the babies in the study of at least 35 weeks gestation who started breathing with just drying suction and stimulation who had at least 2 minutes delay before clamping the cord:
Those babies who did not start breathing when the physicians thought they needed positive pressure ventilation, in the first 60 seconds, were then randomized using a smart phone App. The randomization actually occurred at a mean time of 26 seconds after birth, and the cord was subsequently clamped in the Early cord clamping, ECC, group, at a mean of 37 seconds, and after establishing pulmonary ventilation in what they called the Physiologically based cord clamping, PBCC, group at a mean of 136 seconds.
The primary outcome for the trial was the mean heart rate between 60 and 120 seconds after birth. ECG was installed prior to 60 seconds, and the heart rate was recorded every 10 seconds between 60 to 120 s and then averaged. This is the one thing in this trial that I think is questionable, and I am sure was the subject of many discussions. I presume it was based on the findings from the animal studies, of the same group, that heart rate was lower among preterm lambs who had ECC compared to lambs with PBCC. The clinical significance of an average heart rate during this period is questionable, time to resolution of bradycardia might have been a better measure, but of course, not all the babies would be bradycardic, and further restricting the study to those with slow heart rates would have made the study unfeasible. The results show that after 30 s from birth, fewer than a third of the babies ever had a heart rate under 100. They also show that many of the randomized babies didn’t actually need to have respiratory support, 23 of the 63 PBCC babies and 32 of the 60 in the ECC group did not need any respiratory support.
The primary outcome was not different between groups, or between subgroups, and the numerous secondary outcomes were also all similar. A few of the ECC babies started breathing very quickly after randomization, either before the cord was actually clamped, or within a few seconds afterwards, and the actual heart rates were almost identical during the intervention period, and then were slightly higher in the ECC group than the PBCC.
Despite the good physiologic rationale for the practice, this study doesn’t give any support for resuscitation with an intact umbilical circulation.
The eligibility criteria were a bit different (>34 wks) as were procedures and outcomes; the teams were following the Helping Babies Breathe algorithms, which are slightly different in the initial steps to NRP. Babies were randomized prior to birth, but not included unless they didn’t respond to the initial steps of suctioning and stimulation and start breathing within 30 s after birth. At which time the intervention started, which was either immediate clamping, moving the baby to a resuscitation area in a room right next to the DR, or starting PPV “close to the mother in her bed” with clamping delayed at least 3 minutes. Unfortunately only half of the delayed clamping group followed the protocol, with variable durations of delayed clamping, median 105 s (IQR 30 -191). Despite this, there were differences in the primary outcome variable, which for this study was the saturation at 10 minutes of age. The mean saturation with ECC was 85%, compared to 90% for the intact cord group. There were also differences in heart rate at 1 and 5 minutes, but they were actually lower in the intact cord group than ECC, a difference of about 10 bpm at each time. The babies also cried sooner and established regular breathing more quickly. Earlier on in the resuscitation, at 1 and 5 minutes, the differences in saturation were greater between groups, 72% vs 62% at 1 minute, and 84% vs 77% at 5 minutes.
Despite the problems with following the protocol to the full 3 minutes duration of delayed clamping, it looks like most of the intact cord resuscitation group did have either positive pressure or spontaneous breathing before the cord was clamped. 66% of the delayed clamping group, and 20% of the ECC group had breathing efforts before the cord was clamped, and it looks like most of the protocol violations occurred after PPV was at least initiated. They did analyse the data only from those who had the full 3 minutes of delay in clamping, and all the heart rate and saturation differences were a bit larger between groups, but there was no real change in the spontaneous breathing outcomes compared to the ITT analysis.
The clinical importance of having saturations 5 points higher at 10 minutes is questionable, but a trial to show improved clinical outcomes would have to be enormous. Is it feasible to do such a trial? It would probably need cluster randomization, or deferred consent if individually randomized. Otherwise we would have to consent many thousands of mothers prior to delivery in order to have an adequate sample size of babies who are at risk of complications.
It is interesting that the physiologic rationale for this approach includes the demonstration in animals that clamping the cord before PPV leads to lower heart rates, whereas in this study the heart rates were higher with ECC. If you look at the animal data such as this figure from a review article
you can see that the delay in PPV after cord clamping among the lambs was over 2 minutes, and the bradycardia was not really evident for the first 30 seconds (even though RV output is very low). As we don’t wait for 2 minutes of apnoea after cord clamping in human babies, we might well not see the bradycardia that is seen in lambs, despite a probably much lower RV output.
The break in my posts was partly due to a vacation in Hawaii, where I was fortunate to have some time to go birding, and take some photographs of birds from a group that has a remarkable evolutionary history, but which is much less well known than Darwin’s Finches. The Hawaiian Honeycreepers are now known to have diverged from a common ancestor. They are all now endangered, critically endangered or extinct, but those that I was able to see are beautiful unique birds.
Above is one of the least endangered, the Apapani, shown here in an Ohia tree, with its curved beak adapted to seeking nectar from the flowers of this tree.
The I’iwi is an iconic bird, shown on Hawaiian publicities or posters whenever they want a photo of a bird
It has somehow learnt to puncture the nectaries of invasive plant species, like the poka banana that this individual is perched on.
These birds are threatened by the loss of habitat, especially the loss of the Ohia tree which is being attacked by ROD (Rapid Ohia Death) due to a fungus probably brought in by the nursery tree industry. They also lack immunity to two mosquito transmitted diseases, avian pox and avian malaria. Mosquitoes were absent on these islands until recently, and the birds have no immunity to the disease carried by them. Each year, with climate warming, the mosquitoes are found at higher and higher altitude, and the birds retreat to smaller and smaller regions of Hawaii.
The Akiapola’au is a bird with a remarkable beak, but only about 600 are left, it is dependent on the Koa tree.
The lower bill is used like a woodpecker to tap and make holes in tree bark, and the long curved upper bill then probes to remove beetle larvae. It is an exclusive insectivore, unlike the other Honeycreepers.
The even more endangered Palila is only found in dry Mamane forest on the western slopes of Mauna Kea above 2000 m elevation. The remaining range is about 5% of its recent extent. It has evolved a resistance to the toxic phenols in the seeds of the Mamane, which is a leguminous flowering tree under threat from invasive species. Most other small animals find the Mamane seeds and pods toxic.
Not the best photo, but there are so few of these birds left I find myself fortunate to have seen one perched among the flowers of the Mamane. The are also very smart birds, as they were able to take the Hawaiian Department of Land and Natural Resources to court and win, to force them to provide some protection of their habitat. : Palila v. Hawaii Department of Land and Natural Resources, 852 F.2d 1106 (9th Cir, 1988).
They appear to have arrived after Kauai was formed, then cooled, then was colonized by plants, about 5.8 million years ago, with their forefathers being Eurasian Rosefinches. Each successive island has been colonized and then the various ecological niches been filled by adaptive radiation. Visiting the Hakalau Forest Natural Wildlife Reserve on Hawaii, where I photographed the I’iwi and the Akiapola’au was a highlight, special access with an approved guide is required, and we were fortunate to be able to go with Jack Jeffrey, who took the photos included in that figure above.
I put a few more of my photos of birds from Hawaii on a new page of the blog.
Personally I find the birds more impressive than Darwin’s Finches (which are not actually finches but Tanagers), the Honeycreepers in contrast are actually Finches, but have dramatically varying beaks and coloration, despite their common ancestry. Perhaps if they were more famous, and someone wrote a successful book about them, we might be able to stop their decline and prevent their disappearance.