Do estimates of survival change decisions made?

Kidszun A, et al. Effect of Neonatal Outcome Estimates on Decision-Making Preferences of Mothers Facing Preterm Birth: A Randomized Clinical Trial. JAMA Pediatr. 2020.

This is a short report of an interesting idea, published as a research letter. The authors from Germany randomized pregnant women hospitalised for threatened preterm labour, but who had already reached at least 28 weeks gestation, to respond to a scenario of an extremely preterm birth, at 22 or 23 weeks gestation, with either a 60% or a 30% survival rate. They were then asked whether they thought active intensive care provision or comfort care were preferable.

The attitudes toward intensive care provision for the very preterm infant were not different between the 2 scenarios, 47% would have wanted active intervention for the baby with a 60% chance of survival, and 50% for the baby with a 30% chance of survival.

We are advised by professional societies to make shared decisions with parents, prior to extremely preterm delivery, after ensuring that parents are well informed. These recommendations are often accompanied by long lists of potential complications and percentage risks that we are supposed to ensure that the parents understand prior to delivery. This new small, limited, study suggests that all of our information giving doesn’t have much impact, at least between these 2 percentage survival figures; what matters to the decisions which are made is the underlying attitudes of the parents. The authors state, without explaining where their conclusion came from that “an attitude that mere survival is at least as important as quality of life was associated with a preference for life-sustaining treatments”.

I think that we should use the antenatal encounter with potential extremely preterm parents to investigate their values, as much as that is possible, rather than trying to transfer complex information.

Many studies have examined ways to ensure that parents are well informed, but how much that numerical information impacts the decisions that are made is not clear. Decision aids are promoted by various groups as ways of ensuring that information is transferred, and they may indeed increase the amount of information retained by parents in the short term, but do they change decisions? One study from last year examined whether Decisional Conflict was affected by the use of a decision aid. This was an RCT using a decision aid that the authors have previously published. (Guillen U, et al. Evaluating the Use of a Decision Aid for Parents Facing Extremely Premature Delivery: A Randomized Trial. J Pediatr. 2019;209:52-60 e1).

I have previously criticized this specific decision aid for the numbers that are used to describe long term outcomes; the range of blindness for example for babies between 22 weeks and 25 weeks 6 days is shown as being 1 to 15%, I have never seen long-term data from this century showing a 15% incidence of blindness, the majority of studies give an upper limit of serious visual impairment of about 2 to 3%. Similarly, the decision aid includes a section about the risk of “mental disability” which is apparently something that happens in 18 to 54% of former preterm babies of this gestational age. Even the use of this term shows up the limitations of such decision aids, the person presenting the decision aid will have to explain what that term is supposed to mean, and their own prejudices and beliefs and values will become part of the discussion, a decision aid such as this is not value-free! I don’t think you can actually create a value-free aid unless you stick to objective outcomes such as death, and even then, whether death is the worst outcome, or whether some survivors would have been “better off dead” (and which ones) is something that medical caregivers and our patents/parents often disagree about.

The primary outcome of the Guillen study was whether the mothers were definite about their decision or remained uncertain, measured on a decisional conflict scale. The trial showed no difference in that primary outcome. About 200 mothers were recruited between 2013 and 2017, but it was actually the counsellors who were randomized (92 0f them). 123 babies finally delivered before 26 weeks gestation.

A secondary outcome of this trial was: understanding of the complications of extreme prematurity, measured using a 47-question true/false knowledge test. I can’t find an example of the test anywhere, but I wonder how anyone could have a good understanding of the complications of extreme prematurity after being presented with this aid! If the true-false question was “the prevalence of blindness among survivors born before 26 weeks is 1 to 15%” then retaining the knowledge provided by the decision aid would give you the wrong answer!

Even though parents may be more able to recollect the information they are given if it is presented differently (with the decision aid), that may have little or no impact on the decisions that are made, or how definite parents are about their decisions.

A contrasting question was asked by Marlyse Haward a few years ago: whether framing the same data as either positive or negative affected potential decisions made. Their group compared wishes for intensive, compared to comfort, care after receiving identical scenarios, one group received a document which mentioned the chances of survival and being without disability, the other received a document which described death and handicap rates. The information was identical except for the following section: “25 out of 100 infants will survive if provided intensive care. Of those who survive, 15 out of the 25 infants will not have severe developmental disabilities.” The negative version was: “75 out of 100 infants will die even if provided intensive care. Of those who don’t die, 10 out of the 25 infants will have severe developmental disabilities.”

In that study, 3/4 of the respondents (volunteers who weren’t in that situation) overall preferred active intervention and the remaining 1/4 preferred comfort care. When the same data were posed as positive, more respondents preferred active intervention than when the data were presented as death and disability. The impact wasn’t huge, but it seemed to be there.

Putting all these results together, it seems that the actual percentages of good and bad outcomes presented have very little impact on decision-making. The 2 things that matter are the pre-existing underlying values of the parents and whether the person doing the counselling thinks that the outcomes are good or bad.

One illustration of the importance of that second factor is found in the studies from the NICHD neonatal network. In some centres, 100% of babies who deliver at 22 completed weeks gestation receive comfort care. In other centres, 100% of those born alive get active intervention. I am sure that physicians and other caregivers in each centre believe that they practice shared decision making. I wouldn’t be surprised to learn that the counselling professionals in centres with universal intervention emphasize survival and good quality of life, whereas in the other centres they emphasize mortality, suffering and “mental disability”. Even if both groups describe a possible 30% survival, for example. (Of course, some centres just say “we don’t resuscitate at 22 weeks”).

Efforts to improve information transfer probably only improve information transfer, but don’t change the decisions, nor even how certain parents are about the decisions that they make.

I think repeating the German study with more extreme values of survival would probably reveal different preferences, but then the relevance to the real world decisions that we make would become less. Presenting survival of over 90% at 25 weeks among females with good prognostic factors, compared to below 5% at 21 weeks and 0 days for a growth-restricted boy might well reveal that parents respond to those figures with different preferences, but, in the range of outcomes where counselling and decision-making usually occur, the pre-existing values and beliefs of the parents are probably much more important than outcome percentages or lists of complications. And even more important is how those outcomes are presented, as being either a chance of a good life or a probability of suffering, death or disability.


Posted in Neonatal Research | Tagged | Leave a comment

Glucose screening and treatment in the newborn; what now?

This is a field that stagnated in darkness for a while, “But, soft! what light through yonder window breaks? It is the east, and science is the sun” (with apologies to Romeo).

Recent research has shown that

1. In healthy term babies, plasma glucose concentrations frequently fall to levels which cause symptoms in adults

2. Infants who are considered at risk of low glucose concentrations are only slightly more likely to fall to those levels than those who are not so identified.

3. Most term and preterm babies do not have any clinical or metabolic signs that these low concentrations are causing them a problem.

4. Many term and preterm babies do not produce ketones in response to hypoglycemia, whether or not they are hyperinsulinaemic.

4. Screening for short-lived, asymptomatic low glucose concentrations requires blood sampling in about 50% of term and late preterm babies. But, as mentioned, the remaining 50% are at almost the same risk.

5. Low plasma glucose concentrations will often respond to oral glucose gel administration without having recourse to intravenous therapy. But infants who receive treatment also have their breastfeeding impacted and may be separated from their mothers, even if IV therapy is avoided.

6. Waiting to determine if the plasma glucose falls to below 2.0 mM/L before treating, compared to treating immediately at a plasma glucose of 2.6 mM/L, dramatically reduces the number of babies treated, and does not have an adverse impact on developmental scores at 34 months. While having less impact on breastfeeding.

What do do with all this new information? It seems to me that we should :

1. Redefine who needs screening and when. The data summarized above suggests that we should either screen every newborn infant or none of them. I am not comfortable with either idea! If we never screen then very severe prolonged hypoglycaemia without symptoms will never be diagnosed. If we screen everyone then we inflict pain on an awful lot of babies, we will have to ensure reliable methods are rapidly available and protocols in place to minimize harm, and a great proportion of those who receive intervention will not have needed it. Even if we use a low threshold, confirmed as safe by HypoEXIT, of 2.0 mM/L for plasma glucose or 1.7 mM/L for blood glucose, that will include somewhere between 10% and 30% of term and late preterm infants. I am sure that 10 to 30% of newborn infants do not have significant adverse impacts of low blood glucose.

Should we only screen the most severely growth-restricted infants? Only those who are truly growth restricted rather than the normal small baby? Only those from diabetic pregnancies, or those with poorly controlled diabetes, or those from diabetic mothers who are obese?

2. Redefine who needs treating and when. Even if we use a threshold of plasma glucose of 2.0 mM/L then between 10% and 32% of babies will be treated. I am sure that the large majority of those babies do not gain any benefit from their treatment, and may indeed be harmed.

3. Redefine the purpose of screening. There is very little indication that having 1, 2, or 3 plasma glucose concentrations between 1.5 and 2.6 mM/L (for example) has an adverse impact on newborn infants. So, in order to decide on #1 and #2, I think we should focus screening efforts on the detection of babies who are at risk of prolonged or repeated severe hypoglycaemia. The purpose of screening should be to detect hypoglycaemia which increases the chances of adverse outcomes, what are the characteristics of such infants? Can we identify them by history, or examination, or a single blood test with a full metabolic profile performed at 6 hours of age (for example)?

In the meantime, the developing evidence suggests to me :

Screening asymptomatic infants in the first 48 hours of life is of questionable value, regardless of risk factors.

During the first 48 hours plasma glucose < 1.5 mM/L is very unusual.

After 72 hours of age a plasma glucose <2.0 mM/L is very unusual.

Perhaps the best way to address the issues would be to perform a trial of universal screening before the second feed, to only intervene or retest if <1.5 mM/L (blood sugar equivalent of 1.3 mM/L) and compare the outcomes to a randomized group screened and treated according to current practice (choose which one).

Posted in Neonatal Research | Tagged | Leave a comment

Let it GLOW… Normal blood sugar profiles in newborn infants.

To stay with this recent “hot topic” of this blog; what low blood sugar threshold should be used for intervention in the neonatal period?

The usual risk factors (SGA, LGA, IDM, late preterm, maternal beta-blockers) select a group of babies whose blood sugar frequently falls below 2.6, 2.4, 2.0, or whatever threshold you want to choose. But what is not clear is how frequently infants who are not considered at-risk will fall below those thresholds. With modern early skin to skin contact, maintenance of normothermia, and breastfeeding starting early after delivery; what is the normal glucose concentration trajectory after birth? If we measure glucose continually, or intermittently, how often do babies fall below thresholds that are considered abnormal and meriting treatment?

Harris DL, et al. Glucose Profiles in Healthy Term Infants in the First 5 Days: The Glucose in Well Babies (GLOW) Study. The Journal of Pediatrics. 2020.

The important work which is continuing to appear from the Auckland group now includes the GLOW study which followed the continuous interstitial and intermittent capillary plasma glucose values of not-at-risk full-term infants. They enrolled 70 babies with birthweights between the 10th and the 90th percentiles, and no history of maternal diabetes, obesity or other risk factors. The sample was from births in Hamilton, a town a little way south of Auckland. A continuous monitor was placed as soon as possible after birth, capillary glucose was taken about 1 hour later, and then another 3 times in the first day, and then twice a day for 5 days, prior to feeds if possible. Highly reliable methods were used to analyse the blood samples, which are reported as plasma glucose concentration equivalents. They were able to include babies born in the hospital, birthing centre and at home, and 67 babies’ data were eventually included.

These are the ranges of results that they found.

large img

One third of these healthy babies had at least one plasma glucose below a value of 2.6 mM/L; one of the most widely used thresholds (47 mg/dl). From continuous monitoring (interstitial glucose concentration) well over half of the babies were “hypoglycaemic” at some point. As you can see from the figure, the lower limit of normal plasma glucose (or interstitial glucose) doesn’t really start to increase until about 48 hours of age.

If you compare these results to the recommended thresholds for treatment of low glucose results from various learned societies (as in the table below from the publication), you can see that some recommendations lead to a majority of healthy newborn infants being defined as abnormal, and would be treated if the recommendation was followed for them.

Numbers of infants with episodes of glucose concentrations below recommended thresholds for treatment

Postnatal ages (h) 0-4 4-24 24-48 48-72 72-120
American Academy of Pediatrics
 Plasma 0/64 (0) 2/67 (3) 1/67 (1) 1/67 (1) 0/67 (0)
 Interstitial 0/60 (0) 6/55 (11) 3/57 (5) 1/56 (2) 0/47 (0)
British Association of Perinatal Medicine
 Plasma 3/64 (5) 3/67 (4) 2/67 (3) 2/67 (3) 0/67 (0)
 Interstitial 4/60 (7) 9/55 (16) 5/57 (9) 1/56 (2) 0/47 (0)
World Health Organisation
 Plasma 12/64 (18) 16/67 (24) 9/67 (13) 7/67 (10) 1/67 (1)
 Interstitial 23/60 (38) 35/55 (63) 19/57 (33) 17/56 (30) 4/47 (9)
Pediatric Endocrine Society§
 Plasma 16/64 (25) 27/67 (40) 15/67 (22) 31/67 (46) 4/67 (6)
 Interstitial 30/60 (50) 40/55 (73) 33/57 (58) 41/56 (73) 26/47 (55)

Data are number (%).

<25 mg/dL [1.4 mmol/L] if 4 hours, <35 mg/dL [1.9 mmol/L] if 4-24 hours.
<36 mg/dL [2.0 mmol/L].
< 47 mg/dL (2.6 mmol/L).
§50 mg/dL (2.8 mmol/L) in the first 48 hours, ≤60 mg/dL (3.3 mmol/L)] after 48 hours. The American Academy of Pediatrics and Pediatric Endocrine Society guidelines refer to plasma glucose concentrations. The British Association of Perinatal Medicine and World Health Organisation guidelines refer to whole blood glucose concentrations.


The PES thresholds would, therefore, define 40 to 73% of healthy normally grown full-term infants as being hypoglycaemic, which is worrying; the WHO standards, and the new Canadian Pediatric Society guideline, define 24 to 63% of healthy babies as being abnormal and needing intervention.

To be fair the CPS guidelines do note that they only apply to at-risk babies, and that some healthy babies have low blood sugar below the thresholds; they accompany that by a statement that “outcome data support raising the interventional threshold” for at-risk infants. But is that true? I don’t know where the idea comes from that hypoglycaemia is more damaging to at-risk babies than to the remainder. The reason for defining at-risk groups was surely to try and define infants who are more likely to have low blood sugar, not because there was any data that they are more likely to have long-term adverse impacts. It used to be taught that hyperinsulinemic babies were more at risk than babies who were hypoglycemic because of low substrate (especially glycogen stores), and that was because hyperinsulinaemia inhibits the production of ketone bodies, but recent research shows that most “hypoglycaemic” newborns do not produce ketone bodies, even those with low insulin concentrations.

In fact, if you compare the incidence of low sugars in this new study to the previous results they obtained from “at-risk” infants (as they do in this publication) you find that at-risk infants are not at much more risk than not-at-risk infants.

Table V. Number of healthy and at-risk infants with episodes of low glucose concentrations

Thresholds, mg/dL [mmol/L] Plasma glucose Interstitial glucose
Healthy infants At-risk infants Healthy infants At-risk infants
<47 [2.6] 26/67 (39) 159/326 (49) 37/51 (68) 33/44 (75)
<36 [2.0] 7/67 (10) 48/326 (15) 12/51 (23) 14/44 (32)
<27 [1.5] 0/67 (0) 9/326 (3) 0/51 (2) 3/44 (3)

Data are number (%).

I took out the risk ratios and p-values, from the table. (Although I understand the desire to put them in, comparing 2 independent data sets and presenting a p-value of the result is questionable, it gives the probability that 2 samples are from the same population and we already know that is not true!) Just looking at the data shows that there is very little difference in the proportion of “at-risk” and “not-at-risk” babies who have low glucose concentrations.

In other words, “at-risk” infants have only a marginally increased chance of having a glucose that falls below each of these thresholds, compared to perfectly healthy, highly selected “not-at-risk” infants.

Posted in Neonatal Research | Tagged | 3 Comments

Not COVID-19. Who is at risk for hypoglycaemia?

In my previous post about hypoglycaemia, I made certain estimates about the proportion of well-appearing term and late preterm babies who would be considered at-risk. Karen Puopolo and her group at the Pennsylvania hospital have just published a study covering over 10,000 infants delivered at 36 weeks or later. This retrospective study reports the proportion of babies considered at-risk,  which included SGA, >4kg, preterm, post-term (over 41 weeks) infants of diabetic mothers or those who had beta-stimulants or beta-blockers. (Mukhopadhyay S, et al. Clinical impact of neonatal hypoglycemia screening in the well-baby care. J Perinatol. 2020.)

The first finding of importance is that nearly 50% of all deliveries (48.7% to be exact) were considered to be at risk and were therefore screened. The study includes babies who were screened because of non-specific signs that were perhaps due to hypoglycemia, that is, a baby who is a bit jittery or is found with a lowish temperature is often (and should be) screened for hypoglycemia; these infants were about 2000 of the 5,140 screened infants in the study.  These figures highlight the importance of the definition of who is at risk; subtracting the babies who were only screened for possible symptoms, over 30% of late preterm and full-term babies qualify for screening.

The clinical definition of hypoglycemia used for most of these infants was 2.78 mmol/L (50 mg/dl) at all time points, but after the first 18 months of the study the threshold for the first glucose at 2 hours of age or less, according to their protocol, was reduced to 2.5 mmol/L (45 mg/dl) for the remaining 14 months of this study period.

Using these definitions 43% of all the screened infant had at least one hypoglycaemic blood glucose (by bedside stick testing- Accu-check), and received an intervention (usually an extra feed of some sort) before 72 hours of age (when data analysis for this study finished).

Overall, 52% of the infants of diabetic mothers had at least one glucose <2.78, and 42% of the others. Of the babies screened for possible symptoms, 31% had a blood sugar between 2.0 and 2.78 mmol/L at some point, although very few needed extensive treatment and NICU admission.

In other words, using these definitions nearly one quarter of healthy newborn babies, born at term or late preterm, were considered abnormal. One of the adverse consequences of this is that extra feeds are often not breast milk, so among the 3/4 of mothers who planned to breastfeed, 3/4 of them received formula if they had a blood glucose under 2.78 mmol/L. Of those who never had a blood glucose below 2.78 then 56% were exclusively breastfed.

The authors also examine which criteria identify which infants for screening, and compare criteria from 3 different hospitals as well as using the Fenton charts.

The figure below shows how they overlap; it is interesting that the Fenton charts identify a total of only 13.2% of babies as being either under the 10th or over the 90th percentile. I guess that there is a high proportion of LGA babies who are infants of diabetic mothers and thus not included in this group.

A big limitation of this study is the reliance on the Accu-chek device for blood glucose monitoring, which means that there is likely an overdiagnosis of hypoglycemia, as bedside devices all have inaccuracies and tend to read lower than the true concentration.

I think it is still somewhat unclear whether LGA babies whose mothers are not diabetic really require screening. Some studies suggest that they do indeed, but the adequacy of antenatal screening and the criteria used for the diagnosis are not clarified in several studies I found, it does seem that mothers with impaired glucose tolerance, even if not satisfying standard criteria, have babies who are larger than average, and who may be at risk. More work is needed, I think.

Using more restricted, and physiologically relevant thresholds for making a diagnosis of hypoglycemia has several potential benefits, born out in the hypoEXIT trial: fewer babies with a diagnosis, fewer babies receiving unnecessary supplemental feeds, which are usually not maternal breast milk, less separation of mothers and babies.

It is important that we also find ways to reduce the proportion of babies who are screened, as well as having a more evidence-based threshold for treatment.Those babies are also adversely affected by the blood draws, repeated testing to confirm low point of care tests, overtreatment while awaiting confirmation, and all of the impacts which happen to screened babies, even those who are never hypoglycemic.

We need to go beyond the rather 20th-century method of putting babies on a scale at birth to determine their risk of having too little glycogen or too much insulin.

Perhaps intra-uterine growth restriction could be flagged by the obstetricians using deviations from their intrauterine growth curves on sequential ultrasound, many mothers already have a scan at 20 and at 36 weeks, and we could probably eliminate many small babies who are not growth restricted.

At the other end of the weight scale, babies of mothers with completely normal GTT and a large baby could probably avoid being poked by nurses or lab techs.

Posted in Neonatal Research | 1 Comment

Linezolid seems safe for preterms, probably

A few years ago we started having difficulty clearing Coagulase-Negative Staphylococcal (CoNS) sepsis from the blood cultures of some babies in our NICU, children with CoNS also seemed to be sicker, and to more often have thrombocytopenia. It was at that time that I learnt about heteroresistance. This is a phenomenon, as I understand it, where within a colony of vancomycin susceptible CoNS there are a few individuals which are much less sensitive, sub-cultures of the original strain continue to show this heterogenicity of vancomycin sensitivity. This makes it difficult to eradicate the condition with vancomycin.

As a result, we started using linezolid as an alternative for a few cases, and for a short time, it even became our standard anti-CoNS antibiotic, but especially for babies presenting with moderate-to-severe illness severity and thrombocytopenia.

There was very little to base dosing on in the preterm, and little or no knowledge about toxicity in the newborn. There is apparently a risk of peripheral neuropathy in adults after prolonged use, which is usually reversible, so we were concerned that there were potential neurological impacts. The long term safety of many treatments is not clear in the preterm infant, which is not an excuse for introducing more interventions of unknown safety, but does help to put this in context.

We had actually used linezolid once or twice before this occurrence, as it is well absorbed orally, so we had decided, for some babies with difficult IV access and in whom the only indication for IV access was for administration of antibiotics, to finish their antistaphylococcal treatment with oral linezolid.

There does seem to be an increased risk of neurologic impairment and developmental delay in infants who have had a CoNS sepsis, but it is probably less important for that outcome than gram-negative sepsis. We decided we had better investigate the impact of treating CoNS with linezolid, so we decided to compare our outcomes with those of 2 other large Canadian NICUs in the CNN, Mount Sinai in Toronto and BC Children’s in Vancouver.  Sicard M, et al. Neonatal and Neurodevelopmental Outcomes Following Linezolid for Coagulase-negative Staphylococcal Infection: Real World Evidence. Pediatr Infect Dis J. 2020. In the period covered, Sinai had not used linezolid at all, BCCH had used it a few times, and we treated 3/4 of our cases with the stuff.

The baseline data show that babies who received linezolid were indeed sicker, more getting vasopressors, and more receiving a transfusion of something (mostly platelets). We, therefore, corrected for these factors in the analysis of survival and of neurological impairment or developmental delay. In terms of survival, this was a bit lower in the infants who received linezolid in the first 30 days after the episode, but the difference was not great and disappeared after adjusting for severity of illness.

Long term neurological and developmental outcomes were very similar after adjusting for severity of illness. Even though there were more deaths in the very long term after linezolid use, as mentioned there was no difference in the first 30 days after using the medication, so it seems very unlikely to be causative.

Since this period our CoNS sepsis has again become less virulent, and we don’t seem to be having the heteroresistant strains any longer, so we now only use linezolid in rare cases. I wish I knew why changes like that happened!

Why not use linezolid more? It can be given orally, monitoring levels is fairly easy, and not clearly required. We now have about as much data about the safety of linezolid as about many other drugs that we use in the newborn; vancomycin, for example, is described as having about a 5% frequency of nephrotoxicity, and we have no idea about long term safety, but it seems no worse than linezolid from our study. A randomized comparison of vancomycin vs linezolid, examining long term outcomes in preterm infants with CoNS would be the best approach but seem unlikely to be done in the near future. Good quality registry RCTs could answer the question quickly and cheaply, but would still require some funding. In the meantime, observational studies like ours will help to allay some fears, but risk missing some adverse impacts, especially as we did not collect some data, for example acute creatinine changes.

I guess that the common response in such situations is that we have been using vancomycin for a long time, we think we know it, and its possible complications, and introducing a new agent, which does not have overwhelming advantages, is something we try and avoid.



Posted in Neonatal Research | Tagged , , | 1 Comment

What to do about early postnatal steroids?

Steroid metabolism in the very immature infant is… immature. Adrenal function is still developing in the fetus between 20 and 26 weeks, and a source of precursors from the placenta is important, but obviously disappears at delivery. Very preterm babies might have limited responses to stress, and therefore might benefit from administration of steroids. Some studies seem to show that extreme preterms who have lower cortisol levels, and/or lower responses to ACTH stimulation tests, have more mortality, or perhaps more BPD, but those data are confusing, and somewhat inconsistent. Babies exposed to chorioamnionitis have higher cortisol concentrations over the first week of life at least.

This micro-summary of the evidence underlying the rationale for replacement trials does suggest that some form of replacement is worth investigating, but for whom, and when?

There are a number of possible approaches to the very early use of hydrocortisone.

1) Immediate replacement therapy for all extreme preterm infants.

2) Immediate replacement therapy for selected infants

3) Slightly delayed (12 to 48H) therapy for selected infants with ongoing respiratory problems, or other diagnoses.

4) More delayed therapy (48h to 7 days) for selected infants with ongoing respiratory problems, perhaps with larger doses.

5) Therapy after 7 days of age, when lung inflammation has set in, with larger hydrocortisone doses.

A recent individual patient meta-analysis of studies which were referred to as “prophylactic” included 4 trials which, from this viewpoint, were in categories 1 and 3. (Shaffer ML, et al. Effect of Prophylaxis for Early Adrenal Insufficiency Using Low-Dose Hydrocortisone in Very Preterm Infants: An Individual Patient Data Meta-Analysis. The Journal of pediatrics. 2019;207:136-42 e5). I think that the implications of those 4 trials is quite different between group 1 (which is only the PREMILOC trial of 521 babies), and the other 3 trials which include 360 babies from Kristi Watterberg’s trial, and 100 total from 2 other trials, those three trials are in my category 3.

The PREMILOC trial (Baud O, et al. Effect of early low-dose hydrocortisone on survival without bronchopulmonary dysplasia in extremely preterm infants (PREMILOC): a double-blind, placebo-controlled, multicentre, randomised trial. The Lancet. 2016;387(10030):1827-36) did not have any postnatal illness criteria, but did exclude infants with severe growth restriction (<3%le), early rupture of membranes (<22wks). asphyxia (5-min Apgar <4) and congenital anomalies seen before birth. The growth restriction criteria would probably have eliminated more than 3% of babies from eligibility, depending on the definitions used. The other exclusions, probably fewer than that. The study did not include any babies of 23 or 22 weeks best-guess gestational age (BGGA).

Those other 3 trials included in this IPD meta-analysis are :

1. Watterberg, included infants of 500-999g if they were ventilated between 18h and 48h; they were randomized at an average age of 33 h. (Watterberg KL, et al. Prophylaxis of Early Adrenal Insufficiency to Prevent Bronchopulmonary Dysplasia: A Multicenter Trial. Pediatrics. 2004;114(6):1649-57). The 360 randomized infants received 1 mg/kg/d for 12 days, then 0.5 for 3 days, or placebo.

2. Peltoniemi 2005, who randomized 51 infants of 23 to 30 weeks and 500-1250 g birth weight, ventilated before 24 h (the larger babies were ventilated for >24h with O2). they received 2mg/kg/day for 2 days, then 1.5 for 2 days, then 0.75 for 6 days. (Peltoniemi O, et al. Pretreatment cortisol values may predict responses to hydrocortisone administration for the prevention of bronchopulmonary dysplasia in high-risk infants. The Journal of pediatrics. 2005;146(5):632-7). Hydrocortisone was started before 36 hours of age, but I can’t find the average postnatal age of administration.

3. Bonsante 2007, included 50 infants born at 24 to 30 weeks, and 500 to 1250 g, ventilated after surfactant, at less than 48h of life. They received 1 mg/kg/d for 9 d, then 0.5 for 3 days. (Bonsante F, et al. Early Low-Dose Hydrocortisone in Very Preterm Infants: A Randomized, Placebo-Controlled Trial. Neonatology. 2007;91:217-21). The actual age when the hydrocortisone was started is not clear.

Although I do think that including all of these trials in an IPD-SR is not unreasonable as a way of investigating the issues, it makes figuring out the implications for practice difficult;  Premiloc asks the question: is it safe and effective to give all infants between 24 and 28 weeks BGGA low dose hydrocortisone within the first 24 hours of life? Will it likely improve outcomes without increasing hazards? Or more than the increase in hazards? The 3 trials in category 3 inform a response to a somewhat different question; is it safe and effective to give steroids to all infants of less than 28 weeks who remain intubated for respiratory problems at around about 24 hours of age?

If we divide the trials in this way we end up with much less power, and with results with wider confidence intervals as a result. The results of the PREMILOC trial must also be evaluated with reference to the much higher mortality in that trial than in my practice, or indeed in Canada overall. In PREMILOC the mortality for the combined group of 24 to 25 weeks gestation was 43% (almost identical between groups, 42% hydrocortisone and 44% control). In 2016 in the CNN report the mortality at 24 weeks was 27%, and at 25 weeks was 19%; 23% for the 2 GA weeks combined. (As a side issue, survival continues to increase, in the 2018 report mortality at 24 weeks is down to 22%, and at 25 weeks remains stable at 18%, for an overall of 20% mortality for the two weeks combined).

It is hard to know what to do with the results of a good quality trial like PREMILOC when the intervention group, who had better results after early universal postnatal hydrocortisone than the controls, have much worse results than our babies without routine hydrocortisone; indeed, double the mortality.

When the mortality is twice as high in the better group in a trial than our mortality without the intervention, can we expect any beneficial impact of the intervention on our outcomes? PREMILOC also showed a dramatically higher risk of late-onset sepsis with hydrocortisone than controls among the more immature group, 40% compared to 23%. It is also worth mentioning that 30% of the 24 and 25 weeks babies received hydrocortisone after the study period (slightly fewer in the hydrocortisone group than controls) and 11% of the HC babies and 13% of controls received post-study betamethasone, it isn’t stated how many received both.

The results in the other trials also need to be taken in context, but as they are a selected subgroup of babies it is much harder to compare with local or national results. In Watterberg, for example, the mortality was 17% before discharge, and not different between groups. The original publication notes that the survival without BPD was basically identical between groups, 34% and 35%, or 43 and 42% using a physiologic definition. The IPD meta-analysis, in contrast, gives a slightly larger difference, and numbers which are different to either of the two reported numbers, 38% vs 41%. But in any case, if we put all the trials into a standard-type meta-analysis it looks like this, for the result of death or BPD:

If we take out the group 1 trial (Baud) and just look at those with early selective treatment of babies being ventilated after the first few hours of life, it looks like this ;

And if we take out the small trials, which are much more likely to show positive effects which are routinely exaggerated compared to larger trials, and take out the trial with co-intervention of hydrocortisone and T3 you are left with this:

There has been quite a lot of debate recently about systematic reviews and the fact that the Cochrane approach, of including all randomized babies, risks exaggerating the benefits of treatments, as smaller trials are much more likely to have positive results. Lower quality trials are also much more likely to have positive results.

I don’t like the idea of just eliminating smaller trials from the analyses, but sensitivity analyses focussed on larger trials, and on higher-quality trials, sometimes give different results to those obtained when all trials are included, and should be routine, I think, in a systematic review. Analyses could also be limited to trials which were registered prior to being performed, and further limited to trials where the primary outcome in the registration document is the same as the one in the final publication, to give an objective way of only including trials of high quality, even if they are small.

To return to the original question, what to do about postnatal steroids, I haven’t mentioned group 4 and 5 trials, but the Dutch multicenter trial of Wes Onland et al I discussed when it first came out, is in my group 5. It does seem to show a decrease in mortality with higher dose hydrocortisone after the end of the first week. There are some problems with extrapolatability for this trial also, I can’t find any published Dutch survival and BPD data more recent than 2011, but this article (de Kluiver E, et al. [Perinatal policy in cases of extreme prematurity; an investigation into the implementation of the guidelines]. Ned Tijdschr Geneeskd. 2013;157(38):A6362), after the change in Dutch guidelines to promote active treatment at 24 weeks gestation (thank you Google translate!), reported a 43% survival at 24 weeks, and 61% at 25 weeks, much lower than contemporary Canadian results, and 64% BPD incidence among survivors at 24 weeks, and 44% at 25 weeks. The Onland trial started in 2011, so these figures have some relevance. In the Flanders region of Belgium in 2011 (Draper ES, et al. Variability in Very Preterm Stillbirth and In-Hospital Mortality Across Europe. Pediatrics. 2017) survival at 24 and 25 weeks combined was only 40% and was 0 at 22 and 23 weeks.

For this trial, it is much harder to think about whether it may be relevant to my practice, the babies included were already receiving intensive care and had survived to a week of age, I don’t know whether the results of a similar group of babies in my current practice would be similar enough for me to be able to expect a similar impact on survival. But I am concerned that a mortality of 24% in the control group seems very high. If you use the online BPD calculator of the NICHD, using data from 2000 to 2004 (based on the publication by Matt Laughon), you can input data from a virtual baby similar to the average infant in the control group (male, 26 weeks, 710g, and 35% oxygen); the predicted mortality is 9.5%. substantially less than the mortality before discharge of the intervention group in Onland et al, which was 15.5%.

I know that comparing a point prediction from the online tool for a single baby similar to the mean baby in the Onland trial is not directly comparable to their findings, but it does give me pause to see so much higher mortality among the controls in the trial.

I think that I am not ready to give routine hydrocortisone to all babies between 24 and 28 weeks based on these data, nor to give them to 22 or 23 week BGGA babies. There are no clear data that giving them to babies at 24 hours of age who are still ventilated is beneficial either. I think, prior to changing practice and instituting a potentially hazardous treatment (increase in late-onset sepsis) I need a trial similar to Permiloc, which includes babies of under 24 weeks, in an environment with mortality more similar to current mortality in Canada, the USA, the UK, Germany, Australia and New Zealand, or Scandinavia (not an exclusive list!)

Perhaps a reasonable approach for the present would be to consider hydrocortisone therapy at 7 days of age if an individual baby’s predicted mortality is over say 24%. The Onland trial would suggest that this had a chance of reducing mortality, with little adverse consequence.

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

We are all treaty peoples

We are all treaty peoples if we live in North America!

We, at least the “we” who are concerned about such things, have a tendency to think that the treaty people are only the aboriginal descendants of those who had their lands expropriated by the European invaders. We think that term only relates to the people who then signed treaties to protect the small areas that were left. But a treaty is an agreement between 2 peoples, that creates a requirement on both sides to abide by that agreement. Just because the treaties were signed over the last 300 years doesn’t relieve me of my responsibility, as an immigrant to Canada, to be bound by those treaties.

Some years ago Annie and I wrote an article with Brett Schrewe, one of our brightest residents. It was about a case of near-death in an otherwise healthy full-term infant placed in skin to skin with his mother, who suffered an arrest. Since then I have been very interested in that particular phenomenon, and I try to keep up with the literature.

Brett, on the other hand, has moved on to other things and is now a paediatrician in British Columbia.

I was delighted to read his recent article (Schrewe B. Who matters? Paediatr Child Health. 2019) about the provision of health care to indigenous children, even though some of the terminology (“allyship”, really Brett?) is a bit beyond me, the basic message, and the emotion behind it, is clear and extremely important.

Non-Indigenous Canadians cannot know the intricacies of the hurt or ever claim to know what it has been like to live this history from that perspective. Yet these events nonetheless bind together those of non-Indigenous and Indigenous heritages, for in this story we have all lost: what happened and continues to happen has deprived every single Canadian of the benefits of an equitable country. This matters morally as human beings, but it also matters as citizens: we are, by Canadian law, treaty persons, compelled to acknowledge historical title to unceded lands that comprise the majority of British Columbia as we are to honour the 97 treaties and Land Claim Agreements made by the Crown since 1701. Being a treaty person means having lifelong obligations to those with whom we live in treaty as well as a duty to ask why their benefits continue to be asymmetrically distributed. To ignore these existential relationships is akin to assuming our bodies could exist without their hearts or that the book of our history is comprehensive despite chapters obviously torn out.

The descendants and heritors of those who signed these treaties continue to be bound by their conditions. The injury to our society which is evident from the profound deprivation of many aboriginal children, and their families, is an injustice which has impacts on all of us, whether you are a first nations person, or not.

Posted in Neonatal Research | Leave a comment