Positive pressure ventilation puts infants at risk of developing chronic lung disease (CLD). Chronic lung disease in turn has been linked many times over, as a risk for long term impacts on development. So if one could reduce the amount of positive pressure breaths administered to a neonate over the course of their hospital stay, that should reduce the risk of CLD and by extension developmental impairment. At least that is the theory. Around the start of my career in Neonatology one publication that carried a lot of weight in academic circles was the Randomized Trial of Permissive Hypercapnia in Preterm Infants which randomized 49 surfactant treated infants to either a low (35-45) or high (45-55) PCO2 target with the thought being that allowing for a higher pCO2 should mean that lower settings can be used. Lower settings on a ventilator would lead to less lung damage and therefore less CLD and in turn better outcomes. The study in question did show that the primary outcome was indeed different with almost a 75% reduction in days of ventilation and with that the era of permissive hypercapnia was born.
The Cochrane Weigh in
In 2001 a systematic review including this and another study concluded that there was insufficient evidence to support the strategy in terms of a benefit to death or chronic lung disease. Despite this lack of evidence and a recommendation from the Cochrane group that permissive hypercapnia be used only in the context of well designed trials the practice persisted and does so to this day in many places. A little lost in this discussion is that while the end point above was not different there may still be a benefit of shorter term ventilation.
A modern cohort
It would be unwise to ignore at this point that the babies of the late 90s are different that the ones in the current era. Surfactant and antenatal steroid use are much more prevalent now. Ventilation strategies have shifted to volume as opposed to pressure modes in many centres with a shift to early use of modalities such as high frequency ventilation to spare infants the potential harm of either baro or volutrauma. Back in 2015 the results of the PHELBI trial were reported Permissive hypercapnia in extremely low birthweight infants (PHELBI): a randomised controlled multicentre trial. This large trial of 359 patients randomized to a high or low target pCO2 again failed to show any difference in outcomes in terms of the big ones “death or BPD, mortality alone, ROP, or severe IVH”. What was interesting about this study was that they did not pick one unified target for pCO2 but rather set different targets as time went on reflecting that with time HCO3 rises so what matters more is maintaining a minimum pH rather than targeting a pCO2 alone which als0 reflects at least our own centre’s practice. There is a fly in the ointment here though and that is that the control group has a fault (at least in my eyes)
Day of life
In the original studies of permissive hypercapnia the comparison was of a persistent attempt to keep normal pCO2 vs allowing the pCO2 to drift higher. Although I may get some argument on this point, what was done in this study was to compare two permissive hypercapnia ranges to each other. If it is generally accepted that a normal pCO2 is 35-45 mmHg then none of these ranges in the low target were that at all.
How did these babies do in the long run?
The two year follow-up for this study was published in the last month; Neurodevelopmental outcomes of extremely low birthweight infants randomised to different PCO2 targets: the PHELBI follow-up study. At the risk of sounding repetitive the results of Bayley III developmental testing found no benefit to developmental outcome. So what can we say? There is no difference between two strategies of permissive hypercapnia with one using a higher and the other a lower threshold for pCO2. It doesn’t however address the issue well of whether targeting a normal pCO2 is better or worse although the authors conclude that it is the short term outcomes of shorter number of days on ventilation that may matter the most.
The Truth is Out There
I want to believe that permissive hypercapnia makes a difference. I have been using the strategy for 15 or so years already and I would like to think it wasn’t poor strategy. I continue to think it makes sense but have to admit that the impact for the average baby is likely not what it once was. Except for the smallest of infants many babies these days born at 27 or more weeks of gestation due to the benefits of antenatal steroids, surfactant and modern ventilation techniques spend few hours to days on the ventilator. Meanwhile the number of factors such chorioamniotitis, early and late onset sepsis and genetic predisposition affect the risks for CLD to a great degree in the modern era. Not that they weren’t at play before but their influence in a period of more gentle ventilation may have a greater impact now. That so many factors contribute to the development of CLD the actual effect of permissive hypercapnia may in fact not be what it once was.
What is not disputed though is that the amount of time on a ventilator when needed is less when the strategy is used. Let us not discount the impact of that benefit as ask any parent if that outcome is of importance to them and you will have your answer.
So has permissive hypercapnia failed to deliver? The answer in terms of the long term outcomes that hospitals use to benchmark against one and other may be yes. The answer from the perspective of the baby and family and at least this Neonatologist is no.
Producing milk for your newborn and perhaps even more so when you have had a very preterm infant with all the added stress is not easy. The benefits of human milk have been documented many times over for preterm infants. In a cochrane review from 2014 use of donor human milk instead of formula was associated with a reduction in necrotizing enterocolitis. More recently similar reductions have been seen in retinopathy of prematurity. Interestingly with respect to the latter it would appear that any amount of breast milk leads to a reduction in ROP. Knowing this finding we should celebrate every millilitre of milk that a mother brings to the bedside and support them when it does not flow as easily as they wish. While it would be wonderful for all mothers to supply enough for their infant and even more so that excess could be donated for those who can’t themselves we know this not to be the case. What we can do is minimize stress around the issue by informing parents that every drop counts and to celebrate it as such!
Why Is Breast Milk So Protective
Whether the outcome is necrotizing enterocolitis or ROP the common pathway is one of inflammation. Mother’s own milk contains many anti-inflammatory properties and has been demonstrated to be superior to formula in that regard by Friel and no difference exists between preterm and term versions. Aside from the anti-inflammatory protection there may be other factors at work such as constituents of milk like lactoferrin that may have a protective effect as well although a recent trial would not be supportive of this claim.
Could Mother’s Own Milk Have a Dose Response Effect in Reducing The Risk of BPD?
This is what is being proposed by a study published in early November entitled Influence of own mother’s milk on bronchopulmonary dysplasia and costs. What is special about this study and is the reason I chose to write this post is that the study is unusual in that it didn’t look at the effect of an exclusive human milk diet but rather attempted to isolate the role of mother’s own milk as it pertains to BPD. Patients in this trial were enrolled prospectively in a non randomized fashion with the key difference being the quantity of mothers own milk consumed in terms of a percentage of oral intake. Although donor breast milk existed in this unit, the patients included in this particular cohort only received mother’s own milk versus formula. All told, 254 infants were enrolled in the study. As with many studies looking at risks for BPD the usual culprits were found with male sex being a risk along with smaller and less mature babies and receipt of more fluid in the first 7 days of age. What also came up and turned out after adjusting for other risk factors to be significant as well in terms of contribution was the percentage of mother’s own milk received in the diet.
Every ↑ of 10% = reduction in risk of BPD at 36 weeks PMA by 9.5%
That is a really big effect! Now what about a reduction in costs due to milk? That was difficult to show an independent difference but consider this. Each case of BPD had an additional cost in the US health care system of $41929!
What Lesson Can be Learned Here?
Donor breast milk programs are a very important addition to the toolkit in the NICU. Minimizing the reliance on formula for our infants particularly those below 1500g has reaped many benefits as mentioned above. The availability of such sources though should not deter us from supporting the mothers of these infants in the NICU from striving to produce as much as they can for their infants. Every drop counts! A mother for example who produces only 20% of the needed volume of milk from birth to 36 weeks corrected age may reduce the risk of her baby developing BPD by almost 20%. That number is astounding in terms of effect size. What it also means is that every drop should be celebrated and every mother congratulated for producing what they can. We should encourage more production but rejoice in every 10% milestone.
What it also means in terms of cost is that the provision of lactation consultants in the NICU may be worth their weight in gold. I don’t know what someone performing such services earns in different institutions but if you could avoid two cases of BPD a year in the US I would suspect that nearly $84000 in cost savings would go a long way towards paying for such extra support.
Lastly, it is worth noting that with the NICU environment being as busy as it is sometimes the question “are you planning on breastfeeding?” may be missed. As teams we should not assume that the question was discussed on admission. We need to ask with intention whether a mother is planning on breastfeeding and take the time if the answer is “no” to discuss why it may be worth reconsidering. Results like these are worth the extra effort!
There may be nothing that is harder in medicine. We are trained to respond to changes in patients condition with a response that more often than not suggests a new treatment or change in management. Sometimes the best thing for the patient is in fact to do nothing or at least resist a dramatic response to the issue in front of you. This may be the most common issue facing the new trainee who is undoubtedly biased towards doing something. Take for instance the situation in which the trainee who is new to the service finding out that their 26 week infant has a PDA. Their mind races as they digest this information from morning signover. There is less than 2 hours until they come face to face with their attending who no doubt will ask them the dreaded question. “What are you going to do about it?”. When having to choose a path, if they state “I want to sit tight and watch” they fear the thought of the attending thinking they don’t know what to do. Conversely they could stick their neck out and choose to treat with a variety of approaches but then might they be seen as too aggressive?! The likely path is suggesting treatment but in fact the more I think about it the option of benign neglect may be the best approach or at least one in which if you treat and it doesn’t work the first time you just shrug your shoulders and say “I’ll deal with it till it closes on it’s own”.
This post really is a follow-up to a previous one entitled The Pesky PDA. A Puzzle After All These Years. What triggered this writing was another before and after comparison of two periods in which the management of PDAs for a unit took a 180 degree turn.
Know When to Hold Em And No When to Fold Em
This is the essence of the issue for one unit. Sung SI et al published a paper this month entitled Mandatory Closure Versus Nonintervention for Patent Ductus Arteriosus in Very Preterm Infants. They describe a before and after comparison of 81 infants from 2009-11 and 97 infants from 2012-14. All babies were born between 23-26 weeks gestational age. In the first time period their unit had a mandatory PDA closure policy. That is they gave one course of indomethacin and if possible a second course followed by surgical ligation. A significant PDA was defined as one that had a left to right shunt and was at least 2 mm in diameter and the patient had to be ventilated. Any patient who had been extubated regardless of need for CPAP did not have to have their PDA closed. In the second time period the group attempted to avoid indomethacin and ligation at all costs and in fact in this cohort none received either.
So What Happened?
In the first time period 52 (64%) received indomethacin but only 29% responded and a full 37/52 (71%) went on to receive surgical ligation. Of the 29 that did not receive indomethacin due to contraindications they underwent primary ligation for a total of 82% receiving surgical ligation. The average day of closure for period 1 was 12.9 days.
In period 2 a number of interesting findings occurred. The average day of closure was at 44.2 days. Five infants were discharged with a PDA with 3 experiencing spontaneous closure after discharge and the remaining infants undergoing transcatheter occlusion. In period 2 there were more diuretics and fluid restriction employed. Comparing the two periods for a number of other outcomes reveals some other intriguing findings.
Even with such differing approaches there is no difference in mortality, severe IVH, ROP, PVL, NEC or sepsis. What is different though is the diagnosis of BPD yet there is no difference in total ventilation. In period 2 there is a shift towards more of this ventilation being HFOV less CPAP use at the same time.
What Might It All Mean?
It is retrospective and therefore we cannot be certain that there are not other variables that are not affecting the results that would have had a better chance of being evened out in an RCT. Having said that it is intriguing that having a PDA has been associated with BPD in the past but in this study having a PDA for a longer time is associated with a reduction in BPD. We know that longer periods of invasive mechanical ventilation increase the risk of developing BPD so it is intriguing that that there is no difference in mechanical ventilation yet there is more BPD when you are aggressive with the PDA. You might postulate that the need for surgery leads to greater need for ventilatory support and therefore damages the lungs but the needs for HFOV was higher in the second phase which at least hints that in terms of aggressiveness, Period 2 infants had a tougher go.
The culprit may be the heart. In period 1 there was a significantly increased rate of myocardial dysfunction and need for inotropes following ligation. It could well be that left ventricular dysfunction led to pulmonary edema such that in the 24-28 hours after the surgery ventilator requirements were increased and damaged the lung. The lack of a difference in overall ventilation days supports this possibility. Looking at the other common risk factors for BPD such as chorioamnionitis and lack of antenatal steroids these are no different between groups. Although not statistically significant there are more male infants in period 2 which would usually tip the scales towards worse outcome as well. It does need to be stressed as well that the rate of surgical ligation is higher than any study I have come across so the contribution of the surgery itself to the disparate outcome needs to be seriously considered.
What would I do?
Despite this study and some others that have preceded it I am not at the point of saying we shouldn’t treat at all. Our own approach is to give prophylactic indomethacin to such babies and then for the most part if a PDA remains treat one more time but at all costs try and avoid ligation. An RCT sounds like it is in the works though comparing the two approaches so that will certainly be interesting to see. It is tough to say what the future holds but to any young trainees who are reading this, the next time you are asked what to do about a PDA you are well within your rights to suggest “Maybe we should do nothing”!
Breast milk has many benefits and seems to be in the health care news feeds almost daily. As the evidence mounts for long term effects of the infant microbiome, more and more centres are insisting on providing human milk to their smallest infants. Such provision significantly reduces the incidence of NEC, mortality and length of stay. There is a trade-off though in that donor milk after processing loses some of it’s benefits in terms of nutritional density. One such study demonstrated nutritional insufficiencies with 79% having a fat content < 4 g/dL, 56% having protein content< 1.5 g/dL, and 67% having an energy density < 67 kcal/dL (< 20 Kcal/oz). It is for this reason that at least in our unit many infants on donor milk ultimately receive a combination of high fluid volumes, added beneprotein or cow’s milk powders to achieve adequate caloric intake. Without such additions, growth failure ensues. Such growth failure is not without consequence and will be the topic of a future post. One significant concern however is that failure of our VLBW infants to grow will no doubt impact the timing of discharge as at least in our unit, babies less than 1700g are unlikely to be discharged. With the seemingly endless stream of babies banging on the doors of the NICU to occupy a bed, any practice that leads to increasing lengths of stay will no doubt slow discharge and cause a swelling daily patient census.
What if increasing volume was not an option?
Such might be the case with a baby diagnosed with BPD. Medical teams are often reluctant to increase volumes in these patients due to concerns of water retention increasing respiratory support and severity of the condition. While diuretics have not been shown to be of long term benefit to BPD they continue to be used at times perhaps due to old habits or anecdotal experiences by team members of a baby who seemed to benefit. Such use though is not without it’s complications as the need to monitor electrolytes means more needle sticks for these infants subjecting them to painful procedures that they truly don’t need. Alternatively, another approach is to restrict fluids but this may lead to hunger or create little room to add enough nutrition again potentially compromising the long term health of such infants.
This paper is essentially a study within a study. Infants taking part in an RCT of Prolacta cream (Prolacta being the subject of a previous post) were randomized as well to a cream supplement vs no cream. The cream had a caloric density of 2.5 Kcal/mL and was added to donor milk or mother’s own milk when the measured caloric density was less than 19 Kcal/oz. The study was small (75 patients; control 37, cream 38) which should be stated upfront and as it was a secondary analysis of the parent study was not powered to detect a difference in length of stay but that was what was reported here. The results for the groups overall were demonstrated an impact in length of stay and discharge with the results shown below.
PDA ligation %
PDA treated medically %
Length of stay, days
PMA at discharge, weeks
What about those with sensitivity to fluid?
Before we go into that let me state clearly that this group comparison is REALLY SMALL (control with BPD=12 vs cream with BPD=9). The results though are interesting.
BPD control (N=12)
BPD cream N=9
Length of stay, days
PMA at discharge, weeks
So they did not reach statistical significance yet one can’t help but wonder what would have happened if the study had been larger or better yet the study was a prospective RCT examining the use of cream as a main outcome. That of course is what no doubt will come with time. I can’t help but think though that the results have biologic plausibility. Providing better nutrition should lead to better growth, enhanced tissue repair and with it earlier readiness for discharge.
One interesting point here is that the method that was used to calculate the caloric density of milk was found to overestimate the density by an average of 1.2 Kcal/oz when the method was compared to a gold standard. Given that fortification with cream was only to be used if the caloric density of the milk fell below 19 Kcal/oz where average milk caloric density is 20 Kcal/oz there is the distinct possibility that the eligible infants for cream were underestimated. Could some of the BPD be attributable to infants being significantly undernourished in the control group as they actually were receiving <19 Kcal/oz but not fortified? Could the added fortification have led to faster recovery from BPD?
Interesting question’s in need of answers. I look forward to seeing where this goes. I suspect that donor milk is not enough, adding a little cream may be needed for some infants especially those who have trouble tolerating cow’s milk fortification.
The decision to extubate an extremely low birthweight infant is one of the most common sources of disagreement and anxiety in the NICU. As a resident, I recall an unwritten rule that “no baby under 750g will have a trial of extubation as they will most certainly fail”. As time went on however, studies suggested that not only is this a false statement but also that the duration of intubation was directly correlated with risk of BPD. As Danan described, for infants less than a 1000g, delaying extubation when they had reached minimal settings did not result in lower rates of BPD. Moreover 60 – 70% of infants who have such extubation attempts are not reintubated within a week, so it is possible to have success.
The detractors however express concern about the 30-40% who do require reintubation and worry that harm will be caused by such acts. Frequent concerns are voiced with respect to potential trauma to the airway leading to subglottic stenosis; potentially increasing the risk of BPD and finally the belief that attempts are futile so why put parents through the stress in the midst of perceived near certain failure. The problem with these perceptions however is that they are mostly the culmination of confirmation bias. Patients that have multiple failed extubations are remembered especially when they develop significant complications, while the ones who extubate successfully are quickly forgotten. This is not intentional in any way but simply a human tendency to follow the pattern of “the squeaky wheel getting the grease”. The preemies who do poorly are front and centre as they take up a great deal of our time & attention while the ones who do very well while celebrated, are not as easily remembered for the success they demonstrated when extubated.
Being someone who believes in extubating as soon as possible though I was surprised several months ago to find a paucity of data demonstrating that there was in fact support for the practice in terms of better outcomes being associated with early extubation. While small studies certainly pointed in that direction the real question was about to be answered in the study that follows.
Is it better to fail multiple times or be intubated once for the duration of ventilation.
This is the central question that I believe has been answered as good as it will ever be by EA Jensen et al in JAMA this past month “Effects of Multiple Ventilation Courses and Duration of Mechanical Ventilation on Respiratory Outcomes in Extremely Low-Birth-Weight Infants.” This retrospective review included 3343 infants below 1000g and examined the primary outcome of BPD and secondary outcomes of death, continued supplemental oxygen use at the time of discharge, and tracheostomy among survivors. As you might expect, regardless of whether one compared duration of ventilation or number of reintubations, variables such as need for surfactant, postnatal dexamethasone, PDA treatment and NEC became more common as exposure to either outcome increased. When a logistic regression model was applied though it became evident that the odds ratio for BPD increased to a much greater degree with increasing duration of ventilation than the number of attempts.
To quote the authors:
“After adjustment for the cumulative duration of mechanical ventilation, a greater number of ventilation courses was not associated with an increase in the risk-adjusted odds of supplemental oxygen use at discharge. The number of ventilation courses remained associated with increased risk of BPD only among infants exposed to 4 or more courses.”
The Needs of The Many Outweigh The Needs of The Few
In other words, for an infant to have an increased risk of BPD they would need to have failed extubation four or more times! It is not easy to predict with certainty who these children will be but if we take these findings as accurate (since I believe we will not see an RCT of this magnitude anytime soon) is it ethical to keep a child intubated due to the belief that they might fail? On the contrary, how can we not try when failure is not associated with adverse outcome? In fact none of the secondary outcomes were found to be increased either. Importantly, there was no difference in the need for a tracheostomy in this population which may serve as a proxy for lack significant airway injury from multiple intubations. The goal of this post is to provide all of us who care for these preemies with the reassurance that if they fail extubation, once, twice and even three times that is still a good thing. For the ones that fail more than that, if only we had a crystal ball we might have spared them that risk but given that we live in a world without this ability, we owe it to our patients to try. Moreover, we owe it to the parents of these children to educate them that failure can still be a good thing.
May I never have a near 100% success at extubating my patients on the first try. If I truly care for them I will embrace my failures and know that I have done the right thing.