Inhaled steroids to prevent BPD? Think again…

The NEUROSIS trial was a high quality trial of inhaled budesonide started before 12 hours in extremely preterm infants (23 to <28 weeks gestation) receiving positive pressure respiratory support. The primary outcome of the trial was survival without needing oxygen at 36 weeks post-menstrual age. “The dose of budesonide was two puffs (200 μg per puff) administered every 12 hours in the first 14 days of life and one puff administered every 12 hours from day 15 until the infants no longer required supplemental oxygen and positive-pressure support or until they reached a postmenstrual age of 32 weeks”.

The initial publication of the results showed a reduction in oxygen needs at 36 weeks, but a possible increase in mortality. The combined outcome was less frequent with budesonide, a difference which had a p-value of exactly 0.05, i.e. a probability that the finding was due to chance alone of exactly 1/20. The initial report was of O2 requirement at 36 weeks, and mortality at 36 weeks, so later deaths were not included.

Since that original report the authors have been following the babies and have just published the neurological and developmental follow up at 2 years corrected age. (Bassler D et al. Long-Term Effects of Inhaled Budesonide for Bronchopulmonary Dysplasia. N Engl J Med. 2018;378(2):148-57). The cutoff they used for developmental delay was a BSID (version 2) score on the MDI of <85. There was no difference between groups, cerebral palsy was marginally higher in the budesonide group, and blindness, deafness, and developmental delay were marginally less frequent. The composite outcome was about 50% in each group, almost all due to the (very liberal definition of) developmental delay. More marked developmental delay and the average Bayley scores were very similar between groups.

They also note that there were more deaths in the budesonide group, an effect which was not likely due to chance (19.9% vs. 14.5%; relative risk, 1.37; 95% CI, 1.01 to 1.86; P=0.04). They note that there were 9 deaths after the 36 weeks limit set for the initial primary outcome, 8 of which were in the budesonide group, and 1 of which was in the controls. So eventually they had 82 deaths of the 413 in the budesonide group, and 58 deaths of 400 in the controls.

There was also the same proportion of babies in each group who had a hospital readmission for medical reasons (46% budesonide vs 48% control), the same number of long term inhaled steroids, bronchodilators, and leukotriene agonists (needing them for more than 2 months after discharge) the same number needing home oxygen for more than 2 months1 week after discharge.

In other words, there is no sign of long term pulmonary advantage, and no indication of long term developmental impact. The only possible benefit of prophylactic inhaled budesonide in extremely preterm babies is that more babies came out of oxygen before 36 weeks post-menstrual age, as shown by the results from the first publication, but they did not get off CPAP any earlier (median post-menstrual age of stopping positive pressure was 33.1 vs 33.4 weeks), nor get home any earlier (duration of hospitalisation 91 vs 93 days).

This is another good indication of why we should re-think our definition of BPD. Early, prophylactic use of potent anti-inflammatory steroids given by inhalation reduced the number of babies needing oxygen at 36 weeks, but does not seem to have had any other measurable benefit for the babies. Indeed a substantial increase in mortality, relatively unlikely to have been due to chance, is the only clinically important finding from this study.

I would love to see some more detailed pulmonary outcomes from this study, to determine if there is some other benefit, but at present this study has only shown that the babies were more likely to stop oxygen after 36 weeks if they were control babies compared to stopping before 36 weeks if they were budesonide receivers, but that doesn’t seem to have had any other impact on their lives. And the control babies mostly stopped their oxygen between 36 weeks and discharge, to have the same number going home with O2 tanks.

If we put this in the context of other studies, where are we now? There are 2 recent systematic reviews that are relevant. Unfortunately they both mix studies of very early prophylaxis using inhaled steroids, with early treatment of babies still ventilated after a few days, and one of them includes treatment of established BPD also.

The most recent is Shah VS, et al. Early administration of inhaled corticosteroids for preventing chronic lung disease in very low birth weight preterm neonates. Cochrane database of systematic reviews (Online). 2017;1:Cd001969. In that review of inhaled steroids, most of the studies were early ‘prophylactic’ type studies, albeit with differing entry criteria. But all, other than NEUROSIS, were tiny with less than 31 babies per group, except for one trial which enrolled babies between 3 days and 14 days of age who were being ventilated, babies were less than 30 weeks and less than 1250 g. That study had a reasonable sample size of 253. The other tiny studies enrolled kids within 12 hours, ❤ days, 3 to 4 days, and one study was 6 to 10 hours after the second surfactant dose. I don’t think that there is enough comparable between the Cole study (the one with 253 babies) and the truly prophylactic use of steroids in very early life. If we delete that study from this consideration we are left, basically with the NEUROSIS trial with about 420 per group, and about 90 steroid and 90 control babies from the other trials.

I think there is  a huge problem with the first report of the NEUROSIS trial, which reported a different number of deaths to this follow-up study. Clearly the study was designed with the primary outcome being death or BPD at 36 weeks, as such that should normally remain the reported primary outcome. But by the time of writing the first report they must already have known there were 9 extra deaths in the budesonide group (they reported average age to discharge and state in the manuscript that the babies were followed until discharge). Why didn’t they report in the initial publication that the number of deaths were significantly higher in the budesonide group than the controls? There was a 37% increase in mortality by discharge (RR 1.37, 95% CI 1.01-1.86), but I can see no mention of that in the publication or the supplemental data.

There really is no good reason for censoring deaths at 36 weeks, just because the diagnosis of BPD is made at that time. “Death or NEC” is another (questionable) composite outcome used in neonatal trials, and, of course NEC can occur at any time after birth, I don’t think that there are reports of the combined outcome of NEC at any time or death before 36 weeks. Death before discharge is a much more relevant outcome for babies and their families than death before 36 weeks.

Neonatal trials should always report death before discharge, even if death at some other point is also reported.

Earlier this year the Canadian Neonatal Network published an analysis showing that diagnosing BPD at 36 weeks is not very predictive of respiratory morbidity during the first year of life. Isayama T, et al. Revisiting the Definition of Bronchopulmonary Dysplasia: Effect of Changing Panoply of Respiratory Support for Preterm Neonates. JAMA Pediatr. 2017;171(3):271-9. It was noted that oxygen use at 36 weeks was not that good at discriminating between those infants with and without serious respiratory morbidity. The need for either oxygen or respiratory support (or both) at 40 weeks was substantially better. This is particularly important in the setting of a trial, where an intervention might reduce inflammation and reduce O2 needs at 36 weeks, but not improve long term respiratory problems.

We could imagine, for example, a randomized trial of intravenous dexamethasone as a single day’s treatment at 35 weeks and 5 days. Such an intervention would probably enormously reduce “BPD” but is unlikely to have much effect on clinically important pulmonary function!

Given the apparent increase in mortality, and the lack of clinically important impact on respiratory outcomes, prophylactic very early inhaled steroids should be avoided.

Posted in Neonatal Research | Tagged , , , | Leave a comment

Pre-discharge MRI for the very preterm infant?

For a pre-christmas present, one of the most important recent studies in neonatology.

The study is related to the following question: if I wanted to know how best to counsel my very preterm baby parents about potential long-term outcomes, when they were ready to go home, how should I do this?

Should I have an MRI for all of them, then talk about the results? Is a head ultrasound just as good? Perhaps imaging is not very accurate as a predictor and I should just examine the baby and summarize his clinical course?

The team from “the centre for the developing brain” in London, led by David Edwards, performed this landmark randomized controlled trial. (Edwards AD, et al. Effect of MRI on preterm infants and their families: a randomised trial with nested diagnostic and economic evaluation. Archives of disease in childhood Fetal and neonatal edition. 2017).
They enrolled over 500 babies who were born before 33 weeks gestation in several London hospitals. All the babies had a head MRI and an ultrasound at around term equivalent (38 to 44 weeks) at a single expert centre.

They were then randomized to receive the results of either one or the other test. The results were explained to them by an experienced physician who showed them the images, gave them a copy of some pictures and discussed the developmental prognosis. For the ultrasounds they discussed the prognosis using data from their own systematic review of the prognostic accuracy of head ultrasounds for prediction of cerebral palsy, and only when there was a lesion present with a positive predictive value of more than 25%. They don’t say exactly the same thing for the MRI, giving instead the reference to Woodward’s study from the NEJM in 2006; which was a study that did not calculate the PPV of the abnormal findings. I presume they were presenting the MRI results as being adverse, and predictive of an abnormal outcome, when they showed moderate to severe white-matter abnormalities. In that study there were 35 babies (of 167 total) who had moderate (29) or severe (6) white matter abnormalities. Of whom 7 and 4 developed cerebral palsy of any degree (the GMFCS system was not used). If we use those CP numbers to calculate the Positive Predictive Value of the term-equivalent MRI it works out to 31% (11 of 35) in the cohort of Woodward et al for moderate or severe findings.

In this new publication, there were 72 babies who had MRI findings considered to predict a poorer outcome, and 22 with ultrasound findings that were considered predictive.

The study wasn’t designed this way but I would be very interested to see what happened to the families who had ultrasound findings not considered predictive, but MRI findings which were…

What the authors did actually do was to present the results of the un-masked test to the families.  This was done as follows “An experienced physician unaware of the unallocated result discussed the allocated images and neurodevelopmental prognosis in a structured interview with parents, providing permanent examples of the images and a written summary of the prognostic information”. The “structured interview” by an experienced member of this group, among the world’s experts in evaluation and interpretation of imaging findings in such patients, would have been very interesting to see more details about.

What should we say to parents when we have the results of a test with very poor positive predictive value for adverse outcomes, but relatively good negative predictive value? When the prevalence of that adverse outcome is relatively low? By which I mean that 92% of the infants in this study did not have CP with a GMFCS of 2 or more, similar to data in many other studies; if you have a normal pre-discharge MRI that increases to about 96%, (based on previous publications) a modest improvement in NPV over not doing an MRI, but still not to 100%.

I guess you would have to say something like “your baby has an imaging study which is normal/abnormal, which doesn’t have much impact on our prediction of their long-term outcome. He/She will be followed-up with frequent evaluations, and we expect that things will go well. There is a small chance of developing problems with control of movement, and your baby will be examined so that if something develops they can get any therapy that they need. There is about a 1 in 3 chance of a delay in language development, and a 1 in 5 chance of overall slower than average development, we will be doing testing to see if those things happen and can offer support and intervention if they do turn out to occur.”

I think you could say almost exactly the same thing to parents whose baby had an “abnormal” or “predictive” result on MRI or on ultrasound as you would say when the imaging does not show such abnormalities.

I don’t know what was said to the parents in this study, but they did not have an increase in their anxiety scores after the interview, indeed the scores went down a little in both groups. Anxiety scores were in fact the primary outcome variable of this study. The score used has a scale for “state”, that is how anxious you are feeling right now, and “trait”, that is how anxious you are in general over a period of time. Higher scores reveal more anxiety,  and various cutoffs have been suggested as indicating clinically important anxiety, such as over 39 for young adults to over 55 for geriatric patients, the minimum score possible is 20. A previous study of parents with a baby in the NICU showed trait scores averaging in the low 50’s.

In this study the mothers’ “state” anxiety scores were slightly higher in the ultrasound group before the imaging visit (by about 0.8 points), and stayed slightly higher (by about 1.5 points) throughout the follow-up to 18-24 months. A difference which was not likely due to chance (p=0.02). The “trait” anxiety scores were more similar between groups, and the fathers’ scores were not different between groups.

In other words, with a structured explanation of the results from an experienced individual who understands the limitations of the imaging tests, there is no adverse impact on anxiety. Both MRI and US were followed by a small decrease in anxiety, a decrease which was slightly greater in the MRI group.

But were the tests actually useful? Were they actually predictive of long-term problems, based on Bayley version3 testing and neurological examination at 18 to 24 months corrected age? Here is table 4 from the publication showing the sensitivity, specificity and predictive values of MRI and ultrasound findings.

What they show basically is that the MRI finds more abnormalities, many of which are not associated with long-term problems, but some are. So the sensitivity of MRI is higher for all of the outcomes (but still relatively lousy, ranging from 18% to 60%) compared to head ultrasound (where they range from 5% to 16%), but the specificity is lower, around 90% for each outcome compared to around 97% for head ultrasound.

I think for an individual family the most important metrics are the predictive values, and, because of the increased details and increased detection of white matter abnormalities, the positive predictive values of an abnormal MRI are truly pathetic: from 28% to 48% for cerebral palsy and various components of the Bayley scores. The PARCAR-R test is a parent report of children’s abilities (revised), a parental questionnaire assessment which does not have the same standardization as the Bayley scales (i.e. it doesn’t have a mean of 100 and an SD of 15 in a standardized population), but a cut-off of 49 has been associated with a reasonably good distinction between infants with and without developmental delay, MRI had a very low sensitivity for this outcome, but was fairly specific, leading to a higher PPV but a low NPV. The M-CHAT is a screening test for autism features, which is affected by language delay; the MRI and the ultrasound prediction of failure on 2 or more items of the M-CHAT were quite similar to the results for prediction of language delay.

The authors of this article calculate the area under the  receiver operator characteristics curves, which were modestly higher for MRI than for US for prediction of moderate or severe motor dysfunction (0.74 compared to 0.64) and for motor dysfunction, but were very similar for cognitive and language dysfunction.

Finally they calculated costs, and the MRI cost about 300 pounds stirling more than the US.

I think this study confirms my prior evaluation of the usefulness of brain imaging at discharge of the very preterm baby. Sensitivity is poor for any modality of imaging. Specificity is higher for ultrasound, but is problematic for MRI.

Dr Edwards group and the developing human brain project (and other projects around the world) are incredibly important programs aimed at understanding how the human brain develops, and what can go wrong, so that eventually we will be able to figure out how to normalize brain development in more infants, term and preterm.

See these amazing images, for example, from the King’s College London human connectome project:

These are 3-D reconstructions of the cortical surface of newborn brains, with the images showing the surface of the white matter at the top, of the grey matter next, followed by the inflated surface (whatever that means) different structures, sulcal depth maps, brain curvature, cortical thickness and T1/T2 myelin maps. You can see the progression of complexity and maturation of the brain in the last 8 weeks of pregnancy.

Their detailed connectome images from developing brains are stunningly beautiful, showing the incredible profusion of connections within a developing brain.

Left: Multi-shell high angular resolution diffusion data decomposed into a free water component (greyscale background image) and a directionally resolved brain tissue component shown as rendered surfaces. Middle and right: Visualisation of anatomical connections in the developing brain derived from the brain tissue component.

MRI (or whatever you call these kinds of data maps, MRI seems much too simple a term) is an extremely important research tool, which is not the same as saying everyone should have an MRI!

Both term equivalent MRI and Ultrasound have very limited predictive ability for the former very preterm infant. This study shows, much more clearly than any before, that for an individual baby an MRI adds very little to an ultrasound, and that all former very preterm require health and developmental surveillance regardless of their imaging findings.

Posted in Neonatal Research | 4 Comments

Time to abandon the Papile classification? (part 3)

What should we do now?

I think we should stop using the Papile classification.

We should clarify that an intraventricular hemorrhage with acute hemorrhagic dilatation of the ventricle(s) is not the same pathophysiology as a hemorrhage followed by dilatation. (stage 2 followed by PHVD compared to a stage 3)

We should differentiate between an acute intra-parenchymal echodensity (which may be hemorrhagic or edematous) and PVHI, associated with intraventricular blood on the ispilateral side.

We should analyze and record the brain regions affected by the PVHI or IPE.

We should follow, and report the outcomes, of several hundreds of individuals, in order to have reliable information.

In other words, we should abandon the Papile classification, we should use descriptive terms to interpret head ultrasounds, we should use published percentiles to determine if a ventricle is dilated or not, we should record the brain regions affected on ultrasound, unilateral and bilateral, and we should work at correlating these findings with long-term outcomes.

Posted in Neonatal Research | 2 Comments

Time to abandon the Papile classification? (part 2)

As I mentioned in the last post, the initial report of outcomes by Bassan and others showed that the infants with severity score 0, (unilateral haemorrhage without midline shift affecting one zone) had reasonably good outcomes, 7 of 8 had cognitive scores within the normal range, 4 of 8 were without gross motor dysfunction. They also showed that having a VP shunt did not change significantly the neurological or developmental outcomes. As should therefore be obvious ‘grade 4 hemorrhages’ with these characteristics should be considered relatively benign, warranting close follow-up, and surveillance for complications such as ventricular dilatation or development of multiple cysts. Many PVHI resolve with a single porencephalic cyst in their wake, some resolve and leave findings similar to cystic PVL, that second group may well have worse motor outcomes than single porencephalic cysts, I will come back to this point later.

The Bassan study reports is obviously very small numbers, there were only 30 infants with PVHI in all (which isn’t bad for such a relatively uncommon lesion), so I asked myself if there are other similar data with larger numbers.

The largest I have found so far is a study from North Carolina (Maitre NL, et al. Neurodevelopmental Outcome of Infants With Unilateral or Bilateral Periventricular Hemorrhagic Infarction. Pediatrics. 2009;124(6):e1153-60) which has outcome data from 69 surviving infants with PVHI, 52 unilateral and 17 bilateral. The infants with bilateral PVHI were more immature than the unilateral babies, otherwise there was little difference in the early clinical course of the infants, but their outcomes at up to 36 months were very different.

Infants with unilateral PVHI had a Bayley II MDI at 18 to 36 months which averaged 82, compared to 49 with bilateral PVHI (almost all had an assigned score of 49). Cerebral palsy was much less common and much less severe with unilateral than bilateral PVHI, 15 of the 17 with bilateral PVHI had moderate or severe CP, compared to 19 of the 52 with unilateral PVHI. In concert with these findings the proportion with BayleyII scores above -2 SD on the MDI was significantly better with unilateral PVHI. In this study there is no description or analysis of the extent of the lesion on each side, or the worse side. They did report that those infants who had PVL on late ultrasounds had worse MDI, worse PDI, and more severe motor dysfunction than those who had porencephalic cysts.

Van Buuren et al (van Buuren LM, et al. Cognitive outcome in childhood after unilateral perinatal brain injury. Developmental Medicine & Child Neurology. 2013;55(10):934-40)
reported the outcomes of 21 babies who had unilateral PVHI, it is a somewhat unusual cohort in that the mean gestational age was 30 weeks, and ranged from 26 up to 41 weeks. One strength of the study is the later age IQ testing (performed, however, between 6 and 20 years of age) which showed that the children who had PVHI had full-scale IQ results which were lower than the standardizing population, but mostly within one standard deviation of that population, with a mean of 86. Verbal IQ scores were not significantly different from the theoretical norm at 96. Infants with post-hemorrhagic ventricular dilatation did worse, with a mean IQ of 80 compared to a mean of 96 for those without. The extent of the lesions is not clearly described they are noted as being frontal, fronto-parietal, or parietal and that the location had no impact on outcome, they don’t mention laterality (uni- vs bilateral); 12 of the 21 had a hemiplegia, but the severity is not described.

A study from Groningen (Roze E, et al. Functional Outcome at School Age of Preterm Infants With Periventricular Hemorrhagic Infarction. Pediatrics. 2009;123(6):1493-500)
followed 21 babies who had PVHI, there were 38 babies with this diagnosis in the cohort, 15 died and 2 were not included in the follow-up. They showed that only a minority had significant functional impairments, even though 9 of them had bilateral PVHI, and 5 had extensive (described as fronto-parietal-occipital) lesions. Although 16 children had Cerebral Palsy, 13 were GMFCS 1 or 2, only 3 had more severe motor problems. Most of the cognitive outcomes were normal or mildly affected with only about 10% more than 2 SD below the standardized mean at 4.4 to 12 years of age. They did not show an association between bilateral or more extensive PVHI and worse outcomes, but clearly the power is relatively low for these comparisons. Post-hemorrhagic ventricular dilatation was present in 8 infants, who had poorer scores, again the numbers are very small, but they seemed to have more cognitive difficulties.

The ELGAN cohort study included 54 babies with a diagnosis of PVHI. (O’Shea TM, et al. Neonatal cranial ultrasound lesions and developmental delays at 2 years of age among extremely low gestational age children. Pediatrics. 2008;122(3):e662-9). 44% of them had a Bayley II MDI less than 70 at 2 years of age, and 59% had a PDI under 70, there is no analysis of whether unilateral or bilateral PVHI were different, to unilateral or localised lesions.

A more complex system called the Abdi score is calculated as the square of the highest traditional IVH grade (so scoring 1, 4, 9 or 16), plus the IVH grade on the contralateral side, plus 5 for each hemisphere when more than two of its territories are involved, and plus 5 when there is a midline shift of the brain. The “territories” refers to the brain regions from the Bassan system. The score was first published in the Saudi Medical Journal and I can’t access the paper to see how it was derived, but as far as I can tell it was rather arbitrary. A recent outcome study using the Abdi scores (Al-Mouqdad M, et al. A New IVH Scoring System Based on Laterality Enhances Prediction of Neurodevelopmental Outcomes at 3 Years Age in Premature Infants. American journal of perinatology. 2016(EFirst) followed 183 very preterm babies at 3 years of age. They included 55 babies with intraparenchymal bleeds, whose Abdi scores could range from 16 (i.e. 4-squared) to 35 (4-squared plus 4 plus 5 plus 5 plus 5). The authors showed that  this score was better at predicting outcomes than simply using the Papile grade, but they did not compare to the Bassan system for PVHI. Although this score seems better for an individual patient prediction, it would be difficult I think to use it in daily practice, and the small numbers with each individual score mean that a very large database would be required to validate differences between each score

You may be asking yourself why is Keith Barrington/ going over all this older data? I was stimulated to do so by recent cases in my NICU and also by this publication from Texas (Sheehan JW, et al. Severe intraventricular hemorrhage and withdrawal of support in preterm infants. J Perinatol. 2017;37(4):441-7). The authors note that the long-term prognosis of very preterm infants with parenchymal hemorrhage has been shown to be much better than previously thought.

As you can hopefully see from this (non-exhaustive) review, the majority of infants with parenchymal bleeds (“grade 4 IVH” or PVHI) have outcomes which are within the range of healthy full term infants, only a minority have serious impairments (although, as most of the follow-up is only to 2 years of age, permanent functional impairments are impossible to quantify).  As you might expect, but with little data to support it, it seems likely that more extensive hemorrhagic lesions, and bilateral lesions have worse outcomes than limited or unilateral lesions, but there are very few patients who are informative in the published cohorts.

Sheehan’s study examined discussions and decisions regarding redirection of care among very preterm infants with a diagnosis on ultrasound of serious brain injury (“grade 3 IVH” or PVHI). They showed that the improved knowledge about the benign outcome of many of these lesions has not led to fewer decisions for comfort care. They also examined the outcomes of survivors with PVHI: 13/35 or 37% had no or mild NDI, 10 or 29% had moderate-to-severe impairment and 12 or 34% had profound impairment. Survivors with profound impairment had a median of 4 (IQR 3, 4) territories vs 2 (IQR 2, 3) territories in survivors without profound impairment.

Their data are therefore quite consistent with others, that parenchymal injury is often compatible with little or no impairment, and that the more severe the parenchymal injury the more severe the disability (although with wide confidence intervals). But this has not led necessarily to changes in decision making.

Why is this? In any other area of medicine, I suggest, if outcomes were shown to be much better (or at least, much less bad) than previous beliefs about long-term outcomes, decision-making would surely have changed, no?

I think we should reconsider our outcome data, and how good we are at predicting profoundly adverse outcomes.

What I suggest is the follows:

  1. The construction of robust large databases which relate head ultrasound findings to long-term, functional, outcomes. With enough detail to be able to relate a particular constellation of findings to a range of probable outcomes. Such databases will need to take into account mortality, but that must be predicated on whether care was re-directed or not. Withdrawal of life-sustaining interventions as a result of head ultrasound (or other considerations of long-term outcomes) must be factored into such databases.
  2. Development of tools to teach neonatal and perinatal trainees (and established caregivers) how to counsel parents regarding the results which would be routinely available from 1.

We must move on from a discussion focussed on “your baby has a grade 4 hemorrhage” and “life therefore is no longer worth living” (often implying, or overtly saying, that the long-term outcome will be “terrible”). To “your baby has these findings on imaging” which means that she (or he) will most likely be within the following range of outcomes, the majority of which are associated with an entirely acceptable quality of life.


Posted in Neonatal Research | Tagged , , , | 2 Comments

The end of the Papile classification? (part 1)

In 1978 Lu-Ann Papile published a landmark article detailing the incidence of intraventricular hemorrhage (IVH) in 46 VLBW preterm infants using CT scanning, describing the differing extent of hemorrhage that may be seen, and developing a grading system (there are over 3,000 citations to this article!). To recap, for those who don’t know, a grade 1 ‘IVH’ actually does not have blood in the ventricles, but rather in the germinal matrix region, also known as a GMH or an SEH (sub-ependymal hemorrhage). A grade 2 hemorrhage is bleeding into the ventricles without dilatation, a grade 3 is hemorrhage distending the ventricles, and a grade 4 is bleeding into the substance of the brain (with or without blood in the ventricles).

With the advent of head ultrasound more frequent and less invasive imaging became possible, and most centers adopted the Papile classification for grading hemorrhages seen on ultrasound.

There are several problems with this system, one is the definition of a grade 3 hemorrhage, as many people seem to think that a grade 2 hemorrhage which is followed by early post-hemorrhagic dilatation becomes a grade 3 hemorrhage. It is clear to me from the original description that a grade 3 hemorrhage is one where the ventricles are acutely dilated by the hemorrhage, even though in the original article this is called “intraventricular hemorrhage with dilatation”, the image shown is one of acute hemorrhagic dilatation of the ventricle. The early studies by Joseph Volpe’s group made this clearer by referring to grade 3 hemorrhages as those where “IVH completely filled and distended at least one lateral ventricle” (their articles don’t use a grade 4, more below).

This makes it very difficult to interpret and compare studies which include ‘grade 3’ bleeds, as it is sometimes unclear whether the hemorrhages are grade 3 as I understand them, or whether they are grade 2 hemorrhages followed by ventricular dilatation, which is often transient.

A bigger problem is the use of the term ‘grade 4 IVH’. Parenchymal echodensities without intraventricular blood are often not actually bleeds, and frequently resolve completely, there being a high degree of inter-rater variability in their diagnosis and interpretation. Many intraparenchymal bleeds in the preterm are associated with an ispilateral intraventricular hemorrhage, and have a somewhat heterogeneous appearance on ultrasound.  Such hemorrhages are often called Peri-Ventricular Hemorrhagic Infarction (PVHI) and they are not all equal, they may be small or enormous, unilateral or bilateral, and should not be lumped together in outcome studies. A classification scheme was developed in 2006 (also from Volpe’s Harvard group) to address the severity of the lesions. Bassan H, et al. Ultrasonographic features and severity scoring of periventricular hemorrhagic infarction in relation to risk factors and outcome. Pediatrics. 2006;117(6):2111-8. This classifies PVHI into 4 grades, depending on whether it is bilateral or unilateral, the extent on the worst side, and whether there is a midline shift of the brain. Extent is graded by dividing the brain into 5 zones by drawing lines around the thalamus.

This figure is from the article referred to above and shows the percentage of PVHI affecting each zone, the numbers add up to more than 100% because a PVHI may, of course, affect more than one zone.

In this scoring scheme, if the PVHI involves just one zone it scores 0, if more than one zone it scores 1 (if bilateral it scores 0 or 1 according to the extent on the worst side). The same goes for the other features, scoring 0 or 1 depending on whether the feature is present or not (bilateral, and midline shift, earning a 1 each). The severity score can therefore range from 0 to 3. It is a slightly confusing system as a unilateral PVHI without midline shift affecting one zone is referred to as a PVHI severity 0. In most scoring systems for other diseases a grade 0 means that the disease is absent. In this system a severity score of 0 does not mean that, so you have to remember this feature of the score. In the original report about 3/4 of the PVHI were unilateral, and about half of the unilateral PVHI affected one region only. In the original report the long-term outcomes reported were death, Ventriculo-peritoneal shunt, seizures (during the NICU stay or in the first 5 days after the PVHI), microcephaly, and neuromotor signs at follow-up at more than 12 months of age. Early seizures, death, and neuromotor abnormalities were all more common as the score increased, but those with more severe scores were more likely to have a redirection of care, and most of the deaths were associated with provision of comfort care. 8 of the 12 infants with a PVHI which scored 0 had a normal motor examination at 12 months.

The same authors provided more detailed developmental outcome data in what appear to be the 30 survivors of the same babies, in a 2007 article. The babies were examined between 12 and 66 months of age. There was a progressively increasing proportion of the survivors with abnormal scores  (<-2SD) on the various subscales of the Mullen scales of early learning with increasing scores on early ultrasounds. Most of the babies with a severity score of 0 did not have abnormal results on any of the scales, apart from gross motor where the proportion was 50%.

I think this should have led us to abandon the Papile classification, which was an important advance when it was described, but was a descriptive scale  and was not related to any long-term data. It is clear that there is a huge range of different severities of imaging abnormalities which are hidden by the use of the term “grade 4 hemorrhage”. We should be referring to intraparenchymal echodensities for those cases where they are not clearly PVHI, in particular those echodensities without intraventricular or at least subependymal blood. When a PVHI is present a scoring scheme such as the Bassan scheme should be used. Much larger studies could then give us a better idea of what the true range of outcomes is likely to be for each early appearance of the brain, whether having 3 zones affected is worse than 2, whether the volume of brain affected might be better than just dividing into zones (as a small hemorrhage which crosses zone lines might be less problematic than a large hemorrhage confined to one zone) and so on.

In the next post I will try and review the literature to see how much data of that sort is available.

Posted in Neonatal Research | Tagged , , | 3 Comments

Ethical conflicts and moral distress in the NICU, are they all bad?

The term “moral distress” was coined by nursing researchers who were, generally, referring to situations in which nurses were constrained to act in opposition to their ethical beliefs, usually by having to follow medical orders that they disagreed with. More recently the term has been applied to a wider range of issues that incorporate ethical conflicts.

A new publication from Melbourne (including my wife, Annie Janvier, as a senior author) questioned staff in two NICUs, both physicians and nurses, about their understanding of moral distress, whether they experienced it, and how often, and what could be done about it. (Prentice TM, et al. Always a burden? Healthcare providers’ perspectives on moral distress. Archives of disease in childhood Fetal and neonatal edition. 2017).

Nurses and Physicians experienced moral distress equally frequently, and those who looked after the sickest babies experienced it more frequently. However, they did not generally  consider it to be avoidable, nor to be always an adverse phenomenon. Indeed the majority didn’t even want to try to eliminate it completely.

The same 4 authors have followed this up with a review article.  They argue that since the term “moral distress” was developed, the decision-making landscape as changed. The nursing investigators who first created the term studied nurses who sometimes felt that they were forced into doing things that were against their moral beliefs about what was in the best interests of infants. In the newly emphasized “shared decision making” where decisions are made in a much more open way, and where parents are encouraged to participate in all the major decisions regarding the care of their infants, we often find a range of opinions about a treatment course, and a decision that therefore does not satisfy everybody.

Sometimes I myself feel that a patient that I am treating is not receiving care which is in their best interests, but the parents do feel that. We always try to avoid over-riding the parents wishes, working on compromises that are acceptable to the team, leading to days, or sometimes weeks, where many members of the team feel morally distressed. Unlike the caricature of the paternalistic over-invasive doctor imposing harmful treatments on a poor baby, by far the most common disagreements involve parents wanting to proceed with intensive care options that we feel are unlikely to succeed, or sometimes it is because of a different definition of “success”.

The authors of this review (Prentice TM, et al. The use and misuse of moral distress in neonatology. Seminars in fetal & neonatal medicine. 2017) encourage a process of self-reflection which can be used to build a type of moral resilience, to allow an appreciation of the complexities we deal with, prevent burnout, and encourage ongoing involvement, with a sort of moral courage.

After doing this for almost 40 years, I do not think that ethical conflicts are avoidable in neonatology, or in other critical care areas. I don’t actually they are any more or less frequent than in the past; we can do more, we are more able to support babies at more extreme degrees of illness, we also know much more about the long-term futures of our babies, when studied as groups of infants, but at the same time I am much less definite about my ability to predict what the future of an individual child will hold.

When I feel constrained to provide care that no longer seems in the child’s best interest (usually because of parental desires) it can create distress. Fortunately I work in an NICU where we frequently have open and wide-ranging multidisciplinary discussions about such cases. These really help the team to realize the diversity of opinions about the best course of action, that no-one is blind to the multiple potential courses of action, but are thoughtfully trying to seek what is optimal for the baby and family. We also often acknowledge the difficulty it may pose to the bedside nurse who spends many hours a day with a baby who is not being treated according to their own opinion of the right thing to do. I think this really helps to avoid statements such as “she’s always pro-life” or “he’s often ready to stop before anyone else”, but to, rather, appreciate the struggle we all go through to make the best decisions for our babies.

I think this approach in our unit (we are unique, I think in having three neonatologists who are also PhDs in clinical ethics!) reflects many of the suggestions of this review article as ways to reduce the impacts of moral distress; and, in fact, reflects much of what I have read about the so-called “Schwartz rounds”, an article about which has just appeared, promoting such rounds as a way of encouraging compassionate care in children’s hospital. That proposal irritates me a bit, as I have never noted a lack of compassion in children’s hospitals, or in general hospitals either! On the other hand, I do think that transparent, inclusive, multidisciplinary discussions about our most problematic cases can make a big difference to the emotional state of the caregivers.

Moral distress, and the ethical conflicts underlying such distress, help us to push past our preconceptions, to realize that caregivers and parents may have varying opinions while all still wishing to find the best decision for the baby under our care. I think that not having such distress, such emotionally charged conflicts, would mean that we thought we knew all the answers. I think this is why many people in the article that stimulated this post did not think that we should even try to completely eliminate moral distress. We should, instead seek ways to reduce and control and use it to advance our care for the newborn.

Posted in Neonatal Research | Tagged , | Leave a comment

Recent publications

Even though the blog has been quiet recently, other academic output has continued…

In the NICU we are often faced with babies with late- or early- onset sepsis. The worst cases develop shock, which carries a significant mortality; they may survive if they receive intensive treatment. We have no idea really which treatments are helpful, and which are harmful. We try to follow physiologic principles but without much confidence that we are doing the right thing; we have been analyzing our treatments to try and figure out which interventions are helpful, and which might not work.

What often happens is that babies have standard interventions, (antibiotics, fluid boluses, dopamine or epinephrine infusions…) and when things are not going well we introduce other therapies, such as steroids or norepinephrine.

The first of these two studies looked at steroids: Altit G, et al. Corticosteroid Therapy in Neonatal Septic Shock-Do We Prevent Death? American journal of perinatology. 2017. We examined the responses to steroid, almost always hydrocortisone, therapy for babies in septic shock (i.e. hypotensive and not responding to first line inotropes). As this was not a research protocol, but just desperately trying to save babies who were in dire straits, the use of steroids was variable. But usually was after their response to either dopamine or epinephrine was inadequate. After starting hydrocortisone we were able, almost always, to reduce the doses of, or stop, other catecholamines. As time has progressed we have tended to use steroids earlier, sometimes before getting to high doses of catecholamines…

We compared episodes of septic shock among preterm babies who received steroids in addition to their inotrope, to those who did not receive hydrocortisone. We would expect the babies who received steroids to be more sick, and indeed their inotrope use index was higher, and they received inotropes for longer than the comparison group. Babies in both groups almost always had cultures that were positive, apart from a few cases of severe NEC with negative cultures. The babies receiving steroids were more immature. Mortality was significant 22% of those who did not get steroids, and 40% among those who did, reflecting their increased severity of illness (the difference may have been a chance difference by statistical testing). There were further deaths in both groups prior to discharge, and among survivors severe BPD was frequent. Because of the severe BPD we examined survival to follow-up at one year corrected age, in fact there were no more deaths after discharge. When we statistically corrected for gestational age and duration of inotrope use, there were more deaths among the babies who received hydrocortisone than those that did not.

We looked at the hemodynamic responses to the hydrocortisone and found a rapid improvement in blood pressure, starting within 6 hours, this was followed by a progressive reduction in inotrope requirements, mostly after the first 6 hours, and an improvement in urine output, starting after the blood pressure had increased. As you can see in table 4, these babies, who had a GA at birth on average of 26 weeks, and were about 2 weeks old, on average had mean blood pressures when we started the steroids of only 28 mmHg.

The second study examined our use of norepinephrine in another group of babies with septic shock. Before moving to Sainte Justine I don’t think I had ever used norepinephrine in a newborn infant, but the experience in adults, which shows a better hemodynamic profile in sepsis with norepinephrine compared to other catechols, led us to use norepinephrine in septic shock, mostly late-onset septic shock or NEC with shock.

Rizk MY, et al. Norepinephrine infusion improves haemodynamics in the preterm infants during septic shock. Acta Paediatr. 2017. We looked though our pharmacy database to find preterm babies who had received norepinephrine, all of whom were considered to be in septic shock. The 30 babies were already receiving either epinephrine or dopamine before we started norepinephrine, and nevertheless were very hypotensive, they were on average 26 weeks gestation, and 18 days old, but had a mean blood pressure of just over 20. We started the norepinephrine almost always at 0.1 microg/kg/min, and increased progressively, most babies not needing more than 0.2 microg/kg/min, but occasionally we went as high as 0.6. On average it took about 6 hours to achieve blood pressure and urine output within the normal range for each baby; they were oliguric when norepinephrine was started (mean urine output less than 2 mL/kg/h). We were able to commence reducing the other inotropes at that point, with all but 2 babies having reversal of their shock. Those 2 died, and one other for whom palliative care was started, for a mortality of 10% during the shock episode.  There were 7 other deaths before discharge for an overall mortality of 33%. Long term outcomes were poor, with frequent disabling cerebral palsy and low Bayley scores.

Both of these studies have serious limitations of course, we don’t know if these babies would have done as well, or better, without these interventions, but there are very few studies examining therapies in septic shock, so we thought it worthwhile to examine our practice, to see how we can study this phenomenon for the future. A previous study on norepinephrine  in preterm babies, who were mostly septic, showed a more rapid reversal of shock than our study, but started at a much higher dose (0.4 microg/kg/min) and took 24 hours to see improvement of oliguria. Their infants were also a little more mature (27 weeks), and had younger postnatal age (mean 1.5 days of age) and were sometimes being treated for PPHN, rather than sepsis.

What does this mean overall?

We desperately need prospectively controlled trials in infants with sepsis and hemodynamic compromise. In the meantime, use of steroids and/or use of norepinephrine are usually followed by hemodynamic improvement in babies who remain in hypotensive shock despite a single inotrope. That, I think is about all we can say, from these studies and others. Are such interventions much better than waiting or increasing the doses of the primary inotropes? Do they improve survival or other important outcomes?

As mortality is very high in these babies, RCTs of interventions would not need to be huge in order to have power to detect a clinically important difference in mortality; long-term adverse outcomes are also common, and moderately sized trials could also be informative for those outcomes.

Posted in Neonatal Research | Tagged , , , | 2 Comments