Extreme preterm survival and outcomes

There are frequently publications about the outcomes of extreme preterm infants; as a community I think we should be proud of our investment in outcome research. Indeed, neonatologists invented the entire field of outcomes research (Barrington KJ, Saigal S. Long-term caring for neonates. Paediatr Child Health. 2006;11(5):265-6). When very preterm babies started surviving, the obvious question was: how will they do in the long term? With the survival of extremely preterm (24-28 weeks) and now profoundly preterm (<24 weeks) infants the questions continue.

Multiple cohorts have been extremely informative, and give an opportunity to perform comparisons between those cohorts. A new publication comparing babies from France, Canada, and New Zealand has just been published. (Chevallier M, et al. Mortality and significant neurosensory impairment in preterm infants: an international comparison. Arch Dis Child Fetal Neonatal Ed. 2021). As a disclosure, the first author, Marie Chevallier, was one of the excellent fellows in our programme, and has gone on to be a neonatologist and researcher in Grenoble; the last author, Thuy Mai Luu, is a colleague and friend from my hospital.

They and an international groupd of researchers compared outcomes from 3 cohorts born in 2011. The Canadian cohort have their data collected in CNN, with standardized examinations and data collection in the Canadian Neonatal Follow-Up Network (CNFUN) at 18 – 21 months corrected age. The Australian and New Zealand Network is a similar prospective database with outcome data also being collected at 2 to 3 years, but without the same structured follow-up and data collection. The data from France are derived from EPIPAGE-2, with outcomes at 2 years corrected age derived from questionnaires.

The authors have focussed survival and on long term neurosensory impairments, which I think was wise, given the differences in ages and methodologies. Disabling cerebral palsy, blindness and deafness and relatively stable outcomes, and probably less affected by methods of data collection than, for example, developmental delay.

There are 3 findings of note, I think. One of which is not discussed in the article, that being the proportion of babies by completed week of gestational age, which was much lower at 24 weeks in France than the other 2 cohorts.

Birth and antenatal characteristicsANZNNCNNEPIPAGE-2P value
Gestational age, mean (SD), weeks25.7 (1.1)25.8 (1.1)25.9 (1.0)<0.01
 24 weeks, n (%)182 (19.0)159 (15.6)102 (9.5)<0.01
 25 weeks, n (%)218 (22.7)235 (23.1)258 (24.0)0.78
 26 weeks, n (%)254 (26.5)314 (30.8)361 (33.6)<0.01
 27 weeks, n (%)306 (31.8)311 (30.5)355 (33.0)0.48
Birth weight, mean (SD), g856 (201)864 (216)843 (172)0.05
Male sex, n (%)519 (54.1)537 (52.8)557 (51.8)0.58
Maternal age, mean (SD), years29.3 (6.5)30.7 (5.8)29.4 (5.9)<0.01
Complete course of antenatal steroids, n (%)601 (63.1)692 (70.0)622 (60.2)<0.01

I don’t think there is any biological reason why French women would have fewer deliveries at 24 weeks, this difference is probably because of a relatively lower willingness to provide active obstetrical and neonatal care to babies born at this gestation.

Keeping in mind that there are somewhat fewer of the highest risk babies in France, the outcomes, the primary and the various parts of the primary are here:

OutcomesANZNN, n/N (%)CNN/CNFUN, n/N (%)EPIPAGE-2, n/N (%)P value
Mortality or sNSI204/960 (21.3)210/1019 (20.6)305/1076 (28.4)<0.01
Mortality179/960 (18.7)177/1019 (17.4)283/1076 (26.3)<0.01
Any sNSI among survivors25/578 (4.3)33/621 (5.3)22/659 (3.3)0.22
Cerebral palsy with GMFCS >214/565 (2.5)14/610 (2.3)15/659 (2.3)0.97
Disabling hearing loss12/568 (2.1)14/607 (2.3)7/641 (1.1)0.23
Visual impairment4/570 (0.7)12/562 (2.1)2/623 (0.3)0.01

Mortality is substantially higher in France, but impairments are very similar; apart from more visual impairment in Canada. (But remember that the CNN/CNFUN have formal visual testing, which was not the case in France or in the ANZNN, so this may not be directly comparable).

One general implication of these results is that having a less “aggressive” intervention policy does not select babies who are more likely to have unimpaired outcomes. It just leads to fewer survivors.

These data are, of course, from babies born 10 years ago. Even though neonatal clinical science has not changed that much, attitudes can change much more quickly! In Canada in 2012 about 9% of babies delivering alive at 24weeks gestation had palliative care instituted at birth, in 2019 that was 6%. The CNN doesn’t detail why such babies did not receive active intensive care, but many would have been because of serious congential anomalies or severe growth restriction. It has become quite unusual in Canada for an infant born at 24 weeks gestation to not be admitted for active NICU care in the absence of such additional complications. (CNN annual reports available here, and the CNFUN annual reports here)

From what I have seen, the attitudes in France have also changed, and many more babies born at 24 weeks or profoundly preterm (using the Barrington classification above) now receive active intensive care. These data suggest that such “interventionism” should lead to more survivors, with a similar proportion of survivors having neurosensory impairments.

Posted in Neonatal Research | Tagged , | Leave a comment

Opioid infusions for ventilated preterm babies

Many practices in neonatology are variable between NICUs, and many vary from year to year; without any good scientific data practices become fashionable or routine or ingrained without necessarily having an evidence base to support them.

One such practice is that mentioned in the title, a few years ago it was rare for me to see babies with opioid infusions, but it seems to have become more common, despite the lack of a good rationale.

Infants who are unquestionably in pain, such as in the post-operative period, are excluded from this discussion; pain control after surgery is unquestionably necessary,(although there are serious questions about agents, route, and dosage, etc). I am referring to the use of opioid infusions (most commonly either morphine or fentanyl) for infants who are receiving routine care during assisted ventilation, which includes some potentially painful interventions, blood sampling and endotracheal suctining being the most common.

I think one of the first considerations is that the procedure of being intubated is very painful and should always be premedicated with analgesics, potent opioids with a rapid onset of action are needed (fentanyl or analogues; not morphine).

But I question whether being ventilated is painful. Earlier in my career I worked in the PICU, where it was not unusual to have chldren who were intubated (usually nasotracheally) and on assisted ventilation who needed no analgesia or sedation, I have played with intubated children and seen them reading and listening to music with no need for opiates!

Many babies can spend days or weeks ventilated, during which they can sleep for over 20 hours a day, without routine sedation or analgesia.

There are several considerations that should enter into our decision of whether to use opioids for a ventilated preterm baby. One of which is the lack of evidence from randomized trials of benefit. Indeed the latest version of the Cochrane review was recently published, Bellù R, et al. Opioids for newborn infants receiving mechanical ventilation. Cochrane Database Syst Rev. 2021 Mar 17;3(3):CD013732. doi: 10.1002/14651858.CD013732.pub2. As always you can access for free via the VON website, https://public.vtoxford.org/cochrane-at-von/. That review found 23 relevant trials, 15 of which compared routine opioid use to control (although, in all of them where I can see the information, opioids were permitted in the controls). They found no evidence of benefit, in the short or the long term. Of particular importance, there was little evidence of a reduction in pain scores, the best and easily the largest trial (NEOPAIN) showed, after the first few hours of infusion, a trivially lower PIPP score among the opioid group than the controls; 8 vs 8.7. For each of the long term outcome measures, only a single small study was available, therefore showing little difference, and with very little confidence.

We can learn, I think, from the recent evolution of practice in adult ICUs where light or no sedation has become routine. Trials have shown that routine heavy sedation prolongs assisted ventilation in adults. There are, as a result, a number of studies of patient experiences of being ventilated, many of which have addressed pain, as well as other parts of the patient experience of awake assisted ventilation. (Baumgarten M, Poulsen I. Patients’ experiences of being mechanically ventilated in an ICU: a qualitative metasynthesis. Scand J Caring Sci. 2015;29(2):205-14). In general, pain is not a major part of the reported experience of ventilated adults. They do, however, concur that being suctioned is very painful and unpleasant, although they often report feeling better after the suctioning, as the airway secretions are cleared.

Volume Controlled Ventilation in Adults works differently to volume “guarantee” ventilation in babies, with the presence of a background bias flow babies can take larger breaths if they wish, one of the complaints reported by adults during awake assisted ventilatoin was that they could not take a larger breath when they wanted, which should be less of a problem with our systems.

Also of note, many of these neonatal studies were performed in order to get babies to be better synchronised with the ventilator, at a time when non-synchronized assisted ventilation was the norm in the NICU. With the development of synchronous modes, which are now universal, that is no longer an issue, in general as long as the ventilator is set correctly.

With these considerations in mind, these are the major issues for deciding whether to give analgesia or sedation to a ventilated preterm baby, in particular morphine infusions:

  1. Being ventilated is not, by itself, usually painful. It is stressful for adults, but it doesn’t hurt. Non-pharmacological calming procedures help adults (in particular communication, less relevant for the newborn but calming sounds, voice, touch, swaddling, and gentle interventions are likely to be helpful).
  2. Opiod infusions are ineffective for preventing pain from blood sampling procedures;
    1. sucrose, skin to skin care, in combination and with the addition of non-nutritive sucking, can dramatically decrease pain from blood sampling.
    2. These interventions are still required even in a baby on an opioid infusion!
  3. More invasive skin-breaking procedures, such as lumbar puncture or chest drain insertion, benefit from local anaesthesia as well as sucrose etc. Lidocaine should always be administered for lumbar punctures and chest drain insertion.
  4. Endotracheal suctioning is unpleasant and painful but pain responses are not improved by any currently investigated modality. In particular opioids are ineffective.
  5. Retrospective and observational studies have shown more IVH, more death, and more NEC among preterm infants who receive opioid infusions.
  6. The prospective trials, in particular NEOPAIN, showed a very small increase in mortality and IVH (with 95% confidence intervals which included no difference), but did not report NEC.
  7. Morphine infusions often cause hypotension.
  8. Morphine infusions do not improve short or long term outcomes, additional doses are associated with worse pulmonary outcomes.
  9. Fentanyl is an unreliable sedative, but an excellent analgesic, as a relatively selective mu-opioid receptor agonist it is a poor choice when sedation is the goal.

There are some recent PK studies that I think are very relevant. The most immature babies risk major accumulation of morphine and its metabolites. There is a dramatic accumulation of morphine-3-glucuronide during morphine infusions in the very immature infant. It is a metabolite which is an opioid antagonist. In other words it prevents the analgesic effects of morphine, and interferes with the sedative effects also. This may be why morphine is not a very effective sedative in the most immature infants. Very immature babies accumulate M3G and have enormously high concentrations after a few hours of infusion.

There is a meta-model of neonatal morphine pharmacokinetics published a couple of years ago (Knosgaard KR, et al. Pharmacokinetic models of morphine and its metabolites in neonates:: Systematic comparisons of models from the literature, and development of a new meta-model. Eur J Pharm Sci. 2016;92:117-30) which paper has a link to this app https://unicph.shinyapps.io/MorphineNeonates/ . Play around with it for a while and you will see that the most immature babies risk having extremely high morphine and M-3-G concentrations after standard doses of morphine.

This is an example of predicted morphine and morphine metabolite concentrations for a 24 week 600g 1-day old baby who receives a 100 mcg/kg bolus and a 10 mcg/kg/h infusion, with the 95% confidence limits.

And here for clarity without the confidence limits

There are also huge variations in kinetics, one study showing a 40-fold variation in clearance! That variability can be reduced by taking into account gestational and postnatal age.

How about the long term? Well as mentioned there is very little information from randomized trials, so little confidence in the impact. However, there are observational studies, in particular a worrying series from Vancouver (Zwicker JG, et al. Smaller Cerebellar Growth and Poorer Neurodevelopmental Outcomes in Very Preterm Infants Exposed to Neonatal Morphine. J Pediatr. 2016;172:81-7 e2. Ranger M, et al. Internalizing behaviours in school-age children born very preterm are predicted by neonatal pain and morphine exposure. Eur J Pain. 2014;18(6):844-52) which show that the more morphine the babies received, the smaller were their cerebella (cerebellums?), and the more behaviour problems they had at long term.

They have recently followed this up to show that differences in morphine kinetics, or at least differences in the genes which metabolized morphine, were correlated with those behaviour problems. (Chau CMY, et al. Morphine biotransformation genes and neonatal clinical factors predicted behaviour problems in very preterm children at 18months. EBioMedicine. 2019;40:655-62), the exact meaning of which isn’t clear yet, but continues to make me anxious about the use of morphine for babies who are not clearly in pain.

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

Drug use and abuse in the NICU

The title does not refer to”drugs of abuse” but rather to how we use and choose medications for use in newborns, particularly the very immature. A new publication has just appeared on-line which focuses of medication use in the NICU, and the changes over an 8 year period. (Stark A, et al. Medication Use in the Neonatal Intensive Care Unit and Changes from 2010-2018. The Journal of Pediatrics. 2021).

The enormous database of the Pediatrix medical group has been trawled to find medication exposure data for nearly 800,000 newborns.

There are a number of striking findings, in particular the amazing relative growth of the use of dexmedetomidine. Practically speaking use was 0 in 2010, now 5 per 1000 NICU patients are exposed to it, and 23 per 1000 extremely low birth weight infants.

Dexmedetomidine is routinely touted as being “neuro-protective”, but that assertion is based on questionably relevant animal models, some of which show reduced neuronal apoptosis. I don’t believe there is any long term human outcome data with which to make the assertion that dexmedetomidine is neuro-protective in humans. But of course we don’t have much similar data for any of the sedative/analgesic medications that we use. Morphine probably being the only exception, but the data for morphine are not very robust or very reassuring.

One recent animal study showed that adding dexmedetomidine to a reduced concentration of sevoflurane reduced apoptosis, but if enough dexmedetomidine was given to achieve the same level of anaesthesia as the higher concentration of sevoflurane, then the neuronal apoptosis was identical. (Lee JR, et al. Effect of dexmedetomidine on sevoflurane-induced neurodegeneration in neonatal rats. Br J Anaesth. 2021;126(5):1009-21). So, in this model at least, dexmedetomidine was not neuro-protective. In contrast, this review article found several animal studies that did seem to show neuro-protection (van Hoorn CE, et al. A systematic review and narrative synthesis on the histological and neurobehavioral long-term effects of dexmedetomidine. Paediatr Anaesth. 2019;29(2):125-36) but it was not universal, the details of the animal models and experimental procedures vary greatly. How relevant each one is to the sick newborn is very uncertain.

Multiple use, prolonged infusions, and use in the most fragile babies are all things which need to be better investigated for dexmedetomidine, and for our other sedative/analgesia drugs.

In contrast the same article showed the reduction in use of other medications. Three of them because they are no longer available, (at least in the USA) THAM, chloral hydrate, and ranitidine. No great loss to neonatology, I think. I was pleased to see a dramatic reduction in metoclopramide use, for which I think there is no indication in neonatology. Also, and a little more surprising to me, a marked reduction in lansoprazole use. Again I don’t think that there is much role for the medication; treatment for babies with reflux by prescibing lansoproazole ignores the fact that 50% of reflux in the preterm is non-acid, and the clinical signs attributed to reflux are both non-specific for reflux, and not necessarily caused by acid. Also gastric acid is an important barrier to GI colonisation, helps to prevent respiratory infections, and is probably important for absorption of iron and calcium.

Although it hasn’t changed much over this period, there is still a lot of midazolam being used, being the 9th most frequently prescribed medication overall, and the 13th most frequent in the ELBW. I can’t remember the last time I prescibed midazolam, other than a case of status epilepticus unresponsive to 3 other anticonvulsants in a baby at term. 21% of ELBW babies were exposed to this drug, with a total of 3,700 days of use per 1,000 patients. My comments about sedative/analgeisc medications apply here. What little data we have for long term effects of midazolam are worrying.

The study points out how much we still need to know about the common medications that we use, the majority of which are not specifically licensed for the newborn.

Posted in Neonatal Research | Leave a comment

Pain studies with untreated control groups in babies are unethical.

If you are performing a study of pain control in the newborn and you assign babies to untreated controls, you are creating unnecessary avoidable pain in the control patients. That is true for any patient who is incompetent, obviously including newborns, but also older children and adults unable to give consent for themselves. I guess you could allow competent adults to consent to be randomised to control and have painful procedures without analgesia, but good luck enrolling subjects!

There is no way this is ethical, and it should never be done, and should never be allowed by ethics review boards. (Bellieni CV, et al. Should an IRB approve a placebo-controlled randomized trial of analgesia for procedural pain in neonates? Pediatrics. 2012;130(3):550-3). There is no benefit to anyone, especially not the babies, but not to medical knowledge either. We already know that sticking needles into babies causes pain, and that there are many ways to reduce that pain. Why on earth would you perform a study comparing different methods of reducing pain for blood sampling, and include an untreated control group?

Unfortunately, it still goes on. (Bellieni CV, Johnston CC. Analgesia, nil or placebo to babies, in trials that test new analgesic treatments for procedural pain. Acta Paediatr. 2016;105(2):129-36). I was stimulated to write this post after my weekly trawl through the literature found 2 such studies. I hesitate to give them any credence by referencing them, but maybe naming and shaming is the way to go.

The first is from Stanford, of all places, by a group of people that should know better (Chang J, Filoteo L, Nasr AS. Comparing the Analgesic Effects of 4 Nonpharmacologic Interventions on Term Newborns Undergoing Heel Lance: A Randomized Controlled Trial. J Perinat Neonatal Nurs. 2020 Oct/Dec;34(4):338-345). The study design is also poor, it seems that the control group was enrolled and assigned to have pain without being randomized, but the other 4 interventions were randomized. At first the authors (and apparently the IRB) didn’t think this was research, even though they were prospectively randomizing babies to different interventions and recording responses! The study clearly satisfies every possible definition of clinical research, and, if the researchers and the IRB can’t recognize that, it bodes poorly indeed. The study was therefore started without IRB approval, which means it should never have been performed and clearly should never have been published. They recruited newborns from the postpartum wards, who they say were identified by “medical record review”, which makes no sense, presumably there were individuals screening admitted babies at some point.

It was also retrospectively registered, which means that the authors don’t understand the basics of doing clinical research.

Not surprisingly, the group of 50 babies assigned to having more pain had more pain.

I think the researchers, and the IRB, and Stanford Lucile Packard Children’s Hospital owe an apology to these infants and to their parents.

The second study (Cakirli M, Acikgoz A. A Randomized Controlled Trial: The Effect of Own Mother’s Breast Milk Odor and Another Mother’s Breast Milk Odor on Pain Level of Newborn Infants. Breastfeed Med. 2021 Jan;16(1):75-81) randomized babies into 3 groups, one of which was an untreated control group. The study is behind a paywall, and I am certainly not going to pay to get it, so I don’t know the numerical results, but the abstract notes that the group randomized to having more pain had more pain.

As I performed a quick recent literature search, what is clear is that many of recent articles of painful procedures in the newborn with untreated control groups appear in specialist pain journals, where surely the reviewers should know how unethical it is to deny analgesia to babies having planned painful procedures. Other articles I have seen recently compared skin to skin contact with control (it is already clear that skin to skin contact is effective) another using a vibrating device compared to control (why not give both groups sucrose and see if the device has additional benefit?) and another with combined sucrose, music, non-nutritive sucking and massage compared to untreated controls, (completely useless as an addition to the literature, no idea whether any of the interventions was a useful addition to the others).

Other recent articles have shown how you can do such studies without assigning babies to have more pain, for example this one (Benoit B, et al. The influence of breastfeeding on cortical and bio-behavioural indicators of procedural pain in newborns: Findings of a randomized controlled trial. Early Hum Dev. 2021;154:105308) which randomized babies to either breastfeeding or oral sucrose prior to blood sampling, so all received an effective intervention, they showed no susbtantial difference in pain scores between the gorups. Or this one (Hoarau K, et al. “Holding-Cuddling” and Sucrose for Pain Relief During Venepuncture in Newborn Infants: A Randomized, Controlled Trial (CASA). Front Pediatr. 2020;8:607900.) in which all received sucrose and non-nutritive suckling, but one group also were cuddled during the procedure. Adding a new or additional intervention to previously proven analgesic intervention is ethically acceptable, if it is not already known that the combination of interventions is substantially superior; this study showed no difference in mean pain scores, but fewer babies exceeded a pain score threshold with the combined intervention compared to sucrose alone.

These two studies are a useful addition to the literature, showing that breast-feeding or sucrose are both reasonable alternatives for blood sampling analgesia, and that cuddling a baby in addition to sucrose and NNS has some benefit at the higher end of the pain scores.

Researchers want to perform studies showing differences between groups, it is easier to get them published, which helps to advance your career. But no study with an untreated control group has any scientific value, we already know effective ways of reducing pain. Subjecting babies to painful procdures to improve your CV is unconscionable.

My plea is that if you are asked to be a reviewer for an article for publication which includes babies prospectively assigned to be untreated during a painful procedure, you reject the article with a clear note to the editor that the study was unethical.

If you sit on an IRB, please reject any study which assigns newborn infants, or any incompetent participant, to have avoidable pain.

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

Plug the Lung Until it Grows: the FETO RCTs of antenatal diaphragmatic hernia intervention.

What was at one time called PLUG, and, with the change from open to endoscopic intervention, is now called FETO (fetal endoscopic tracheal occlusion) is a way to harness the normal physiology of the lung in congential diaphragmatic hernia (CDH) to improve lung growth and architecture in order to improve outcomes. Lung growth is partially dependent on the rhythmic increases in fetal intrapulmonary pressures which occur because of the active production of fetal lung liquid and intermittent breathing movements of the fetus associated with partial adduction of the vocal cords. Numerous fetal animal studies showed that interrupting this process (with phrenic nerve section or tracheotomy, for example) leads to pulmonary hypoplasia, while increasing the intrapulmonary pressures, by ligating the trachea, caused pulmonary hyperplasia. Finally animal models of diaphragmatic hernia followed by tracheal obstruction showed at least partial normalisation of lung growth. Interventional obstetricians and paediatric surgeons have attempted to temporarily obstruct the fetal trachea in some fetuses with CDH and very high predicted mortality. Initial attempts were apparently successful, but with a risk of preterm labour and delivery.

The improvements in technique and change to purely endoscopic approaches have led to lower complication rates (importantly prematurity), but it remained unclear whether, overall, survival was improved.

The recent publication of the results of 2 parallel trials in high-risk and moderate-risk patients has largely answered that question.

Deprest JA, et al. Randomized Trial of Fetal Surgery for Severe Left Diaphragmatic Hernia. N Engl J Med. 2021.
Deprest JA, et al. Randomized Trial of Fetal Surgery for Moderate Left Diaphragmatic Hernia. N Engl J Med. 2021.

This was a remarkable undertaking, a truly international collboration in a group of mothers carrying a fetus at extremely high risk of dying. And here, as an aside, I always feel a little uncomfortable with the way we talk about these interventions. Although it was indeed “fetal” surgery, there happens to be a woman in the way! Perhaps we should rather talk about “maternal-fetal” interventions. Even the pretty pictures below represent the mother just as an abdominal and uterine wall to be pierced…

As you can see from this summary, there was a dramatic improvement in survival. I do think it is important to emphasize that there is still a very high mortality of 60% in the treatment group; this is a terribly high risk group of babies. The eligibility criteria included a ratio of observed to expected lung-head ratio (O:E LHR) of less than 25%, predicting a mortality of about 82%, very close to the 85% actually seen in ther controls. The median age of delivery for the FETO babies was 34.6 weeks compared to 38.4 for the controls; this doesn’t sound like a huge difference, but 16 FETO babies delivered before 34 weeks compared to 0 controls, and 10 of them delivered before 32 weeks. Managing moderately preterm babies with CDH is very difficult. It seems likely that if we can find ways to prevent the rupture of membranes and preterm labour impacts of FETO, the benefits would be even greater.

The other trial was run in parallel, it enrolled mothers carrying a fetus with an observed lung to head ratio that was 25.0 to 34.9% of the expected, irrespective of liver position, or 35.0 to 44.9% with intrathoracic liver herniation. This was calculated to lead to a survival of 55% and sample size calculated for a 20% improvement.

The primary outcome was actually changed early on by the DSMC of the trial, which was initially designed with a primary outcome of BPD, or oxygen dependence at 28 days, with survival to discharge as a secondary outcome. I think the DSMC did exactly the right thing here, I don’t know who thought that 28 day oxygen requirement was important for infants with CDH, but to have given that priority over survival would have been a major problem. Perhaps that was initially chosen as the investigators are Obstetricians, not Neonatologists (he typed with tongue in cheek)?

Survival to discharge was higher in the FETO group, 62 of 98 treated compared to 49 of 98 controls, RR 1.27 (95%CI 0.99-1.63), as the confidence intervals for the RR just includes 1.0 the intervention was deemed not to show a significant benefit.

The impact on prematurity was very similar to the high-risk group trial, with the median gestational age at delivery being 2 weeks earlier with FETO than control, but both groups being about 2 weeks later than the other trial, that is 36 weeks for the FETO and 38 weeks for the controls. The relative risk and risk difference of being born before 37 weeks was very similar in the 2 studies.

Although this was a “moderate” risk trial, there was still a huge mortality of 50% in the controls (showing again that the O:E LHR is a reliable predictor of mortality), 20% of both groups required ECMO, and they had between 1 and 3 months of hospitalisation (average about 48 days in each group).

The division into “severe” and “moderate” risk groups (perhaps better termed “extreme” and “severe”) was completely arbitrary. A threshold of 25% O:E LHR could easily have been set to 30, or 31.5%. Such a threshold would have included more infants in the extreme trial, and I can guess would not have changed the result to a “non-significant” result.

In fact I don’t have to guess, because if these trials had been run as a single trial and included all the babies eligible for the two trials, with subgroups of extreme and severe risk, then the total survival to hospital discharge would have been 78 of 138 FETO babies and 55 of 138 controls; p=0.0081 (chi-square with Yates correction). In that case the conclusion would have been that FETO is beneficial for babies with an O:E LHR of < 35%, and for babies with intrathoracic liver with an O:E LHR of <45%, perhaps with a subgroup analysis showing a greater effect in the most severely affected fetuses.

It is evident that, at some point, the increased relative risk of prematurity (which was similar between the trials) will outweigh the benefit of the procedure, but we cannot from these data conclude that below a threshold of 25% O:E LHR, FETO is overall a beneficial intervention, but that above 25% it is not beneficial. Unfortunately, I think that is how the data will be interpreted. A secondary analysis of the overall survival benefits of FETO according to baseline O:E LHR needs to be performed, otherwise mothers carrying fetuses who would benefit will not be offered the intervention.

There are not many conditions in medicine for which a trial of 80 patients will show a dramatic and reliable difference in outcome between groups. The substantial benefits of this intervention for a profoundly serious condition mandate that all centres that have potential links to an experienced FETO centre should find a way to offer the intervention to mothers carrying an eligible fetus.

Currently, mothers who have had the procedue need to stay within reach of the FETO centre until the balloon is removed, for very good reasons. This creates limitations for many families, which may be insuperable, our FETO centre for example is 550 km away, for some mothers for other parts of the province it will be 2000 km away! Moving to live in Toronto for 8 weeks is not necessarily feasible for some mothers, especially as health care costs are all covered by our provinicial systems, but not daily living expenses. I wonder if we could create satellite centres who would develop the expertise required to remove or puncture the balloon in case of preterm labour, or routinely at 34 weeks, including the team, the equipment, and the protocols. Then a mother could travel to the FETO centre for the procedure, stay a couple of weeks in the region, and then return to be close to a satellite centre for the rest of the pregnancy. I have no relevant expertise in the interventions, but it seems to me that ultrasound guided balloon puncture is not that different to many of the interventions performed by our MFM specialists, at least compared to fetal bronchoscopy which seems to me to be a greater level of complexity.

Thank you to the mothers who were prepared to be randomised in this trial, which will have a major impact for many future babies. Your willingness to help others is enormously appreciated.

Posted in Neonatal Research | Tagged , , | 2 Comments

What is critical in a “critical sample”?

It is common practice in the evaluation of neonatal hypoglycaemic episodes, especially if unusual or prolonged, to perform a “critical sample”. This is performed to rule out underlying metabolic or endocrine disorders. But what should the critical sample consist of?

I have looked at several recommednations, and there is great variability in what is included in such samples; it is near universal that a lab glucose, a serum cortisol, insulin and growth hormone are included, but after that there is little consensus.

One exception to those inclusions is the American Pediatric Endocrine Society, who are quite minimalist, recommending that the critical sample includes a glucose, bicarbonate, Beta-hydroxy-butyrate and a lactate. They further recommend that plasma be kept in reserve for further tests (and give the following as examples: plasma insulin, FFA, and C-peptide : total and free carnitine and acyl-carnitine), they don’t even overtly mention cortisol and growth hormone as part of the critical sample. Their recommendation is meant for screening for neonates but is also meant for older children, and it therefore ignores the relative incidence of the various causes of hypoglycaemia in the neonatal period. With substrate deficiency (transitional hypoglycaemia) and hyperinsulinaemia being much more common than other causes, and endocrine causes being next and relatively speaking fairly uncommon.

Beta-hydroxy-butyrate is recommended by many to be included in initial screen, it is the most commonly measured of the ketone bodies, and is depressed in hyperinsulinaemic infants, but is low also in infants with transitional hypoglycaemia, so is really only helpful as a flag for endocrine deficiencies and for the rarer glycogen storage diseases, where it is raised during hypoglycaemia.

Ketones, such as beta-hydroxy-butyrate, are increased during hypoglycaemia due to growth hormone or cortisol deficiency. The diagnosis of those entities is sometimes tricky, growth hormone secretion being pulsatile and the range of normal serum cortisol being wide. Often, also, when an infant has a critical sample performed because the bedside glucose is low, the lab glucose will be taken a few minutes later and often ends up being low normal, either because of the delay, and/or because bedside glucose is inaccurate and usually lower than the lab glucose. So, a serum cortisol which is on the low side of normal on a critical sample with a lab glucose of 2.8 (for example) what does that mean? Knowing that the ketones were high would be a good clue that further endocrine evaluation is required.

Ketones are also elevated in ketotic hypoglycaemia, which is one justification for measuring them in some guidelines, but it is not on the differential of neonatal hypoglycaemia, becoming important in older children.

Free Fatty Acids are also in many recommendations, and mainly serve to distinguish fatty acid oxidation defects, where they are elevated during hypoglycaemia despite low ketone bodies, the total incidence of all those disorders is probably about 1:10,000 births, but they are treatable, and picking them up when an infant is hypoglycaemic in the first few days of life is probably beneficial, and may well improve outcomes. I think we should keep them in our critical sample, but I don’t know the proportion of babies with fatty acid oxidationn defects who present with hypoglycaemia in the first few days of life, so diagnosing them from a critical sample is likely to be very uncommon.

Some recommend including a C-peptide measurement, with the idea that a high insulin with a low C-peptide is evidence of exogenous insulin administration. As that is an extremely unlikely scenario in the neonatal ward or NICU I think we can drop the C-peptide, (especially as some varieties of commercial insulin do not even register with some insulin assays). It is mostly paediatric endocrinologists who seem to want a C-peptide, and as part of the work-up of an older child it might be more relevant.

Many recommendations suggest growth hormone and cortisol estimation in the critical sample, and these seem to be the next most likely to lead to a diagnosis, after high insulin concentrations. Inappropriately low concentrations have led to a diagnosis in several babies I have seen over the last few years.

The Canadian Paediatric Society recommends obtaining a critical sample, but does not mention what should be measured in the sample. Others have suggested IGF binding protein-1 levels, without a good explanation why, and then several suggest other tests in later work up depending on the initial findings, including ammonia, urine organic acids and serum amino acids, triglycerides, carnitine and acylcarnitine profiles.

The volume of blood required is an issue for most blood sampling in the newborn, so I think the tests required for specific diagnosis of other rare conditions that rarely present with neonatal hypoglycaemia can be left out of the initial “critical sample”.

The other consideration is that you can get an idea whether a hypoglycaemic infant is hyper-insulinaemic from the amount of glucose that they are requiring. Those that require very high glucose intakes to remain normoglycaemic (more than 8 to 10 mg/kg/min) are likely to be hyper-insulinaemic. That includes a proportion of those who are Small for Gestational Age, or with birth asphyxia as well as infants of diabetic mothers or LGA infants. Of course the uncommon babies with congenital hyperinsulinism syndromes will fit this picture.

Putting all this together, I think the most appropriate critical sample when it is required; that is, unexplained, prolonged, or recurrent hypoglycaemia, should measure the following, and if blood volume is a problem, start with the first 2 items on the list and work down:



Serum Cortisol

Growth Hormone

Ketones (either Beta-Hydroxy-Butyrate or ketone bodies depending on your lab)

Free Fatty Acids

If you can get enough blood, then do a blood gas with bicarbonate and lactate concentrations.

After that, if you still don’t have an answer and hypoglycaemia is recurrent, a call to your helpful local paediatric endocrinologist would be a good idea!

Posted in Neonatal Research | 1 Comment

Two amazing trials, at the opposite ends of the spectrum. What a weird world.

In the most recent NEJM two trials impacting newborn/paediatric care. One including 3,211 preterm infants, which shows that a very simple intervention could save, I estimate, tens of thousands of lives each year at almost no cost; the other with 50 infants, of one of the highest tech interventions possible, and which could make life immeasurably better for a tiny number of children, but which will probably have an extremely high price tag.

The prosaic first study was a randomized trial in “five tertiary-level hospitals in Ghana, India, Malawi, Nigeria, and Tanzania. All live-born infants in the participating hospitals whose birth weight was between 1.0 and 1.799 kg, regardless of gestational age, type of delivery, or singleton or twin status, were eligible for inclusion” WHO Immediate KMC Study Group. Immediate “Kangaroo Mother Care” and Survival of Infants with Low Birth Weight. N Engl J Med. 2021;384(21):2028-38. This trial randomized the low birth weight infants to either standard care, which involved separation of the mother and baby, with the baby placed in an incubator or with a radiant heater for at least 24 hours with no kangaroo care, then once they started to recover, defined as “CPAP not required, SpO2 90-94% on low concentration of additional oxygen (FiO2 <30% by nasal prongs), Tolerating partial enteral feeds (maybe on partial IV fluids)”, at which point the mother “will come to the SCNU to provide brief sessions of KMC a few times a day”.

Once a control group baby was more stable, defined as “when the following criteria are met for at least a continuous period of 24-hours:
(i) Breathing spontaneously without additional oxygen, and oxygen saturation on room air >90%
(ii) No need for CPAP
(iii) Respiratory rate 40 to <60 breathes per minute
(iv) No apnoea
(v) Heart rate 80 to <180 beats per minute
(vi) Axillary temperature 36.0 to 37.4°C
(vii) No need for intravenous fluids”

they were transferred to the Mother-NICU, units which were specially renovated spaces where Kangaroo Care was facilitated. These standards for the control babies followed current WHO guidelines.

The Mother–NICUs “included mothers’ beds and reclining chairs, were built or converted from existing NICUs. All equipment, staff, and care provision in the Mother–NICUs remained the same as in the control NICUs. At two sites, completely new Mother–NICUs were built in a nearby location and the existing NICUs were retained as the control NICUs. At the other three sites, modifications were made to convert half the existing NICUs to Mother–NICUs, and the other half served as the control NICUs. Infants receiving kangaroo mother care were secured firmly to the mother’s chest with a binder that ensured a patent airway.”

Babies randomized to the intervention were admitted as soon as possible after birth to the Mother-NICU, being randomized either before birth or as soon as possible afterwards, with a mean age of enrollment of about 30 minutes. The goal was to have the babies in KC for 20 hours a day, by the mother or a designated female relative (fathers are not allowed in the majority of the units!) Each had a KC support person who ensured they had access to food and toilets (as did the surrogate) and had routine obstetric postnatal care. The same approaches to neonatal care were employed as in the controls. The countries involved were, clearly very poor; the mothers enrolled had an average monthly family income of about 170 US dollars, and they were in countries where per capita annual health expenditures were as low as 35$ in Malawi, up to about 80$ in other involved countries.

The 28-day mortality in the controls was 15.7%, which was reduced to 12.0% with the immediate kangaroo care intervention. The study was, in fact, stopped early because of the mortality benefit in the intervention group, with no evidence of any harmful effect. The relative risk of death was 0.75 with KC, 95% CI 0.64, 0.89.

Previous studies of KC in low-income countries have also shown a benefit in reducing mortality, the Cochrane review notes a 40% reduction in mortality, but most studies were started after the baby was “stabilised” and therefore excluded a large number of deaths, those occurring soon after birth.

This intervention has the possibility of dramatically decreasing neonatal mortality in low and middle-income countries (NNT=27) at almost no cost; there may be some initial costs related to the logistics, which will not be huge, but may need specific budgets assigned in countries with such low health care expenditures. India and Nigeria, two of the countries involved in this trial, have the highest numbers of neonatal deaths in the world, with a combined 800,000 annual deaths of babies <28 days, much of which is related to low birth weight.

At the absolute opposite end of the health care spectrum, and in the same issue of NEJM, it now looks like gene therapy can “cure” children with ADA deficiency Severe Combined Immunodeficiency (ADA-SCID). Kohn DB, et al. Autologous Ex Vivo Lentiviral Gene Therapy for Adenosine Deaminase Deficiency. N Engl J Med. 2021. This is a disease with an incidence somewhere around 1 per million live births, with the Adenosine Deaminase enzyme being hypo- or non-functional, leading to serious life-threatening, life-shortening, immune deficiency. Some of the states in the USA perform universal screening for this disorder, not because it makes any sense to screen, but because of family advocacy. That is a side issue, but it explains why the average age of patients in the 3 studies, reported together here, is much lower in the USA than in the UK. It is easy to understand why parents wanted neonatal screening, diagnosis is usually delayed until after the first couple of serious infections; however, there was previously no treatment, and the extreme rarity of the condition made it a questionable part of the routine neonatal screen. In the last few years, enzyme replacement therapy has been available and has improved the quality of life of the affected infants, which makes screening a bit more reasonable. This new study shows that administering gene therapy can provide long term improvement in ADA levels, prolonged high-quality survival and cessation of enzyme replacement. The study was partially supported by a commercial enterprise, Orchard Therapeutics, but with major funding by the NIH and the English Medical Research Council and other foundations.

Forty-eight of the 50 children studied had a sustained increase in their ADA levels, and had clinical improvements in their immune function. It is an amazing advance for the tiny number of infants with this condition, with a good probability that their lives may be almost normal after the intervention.

But I worry about what this might cost, and if crowd-funding initiatives will be necessary to get treatment for affected children; the example of Werdnig-Hoffman, Spinal Muscular Atrophy, does not give one much hope.

I also fear that the annual budget expenditure for one case of ADA-SCID in a high-income country may be more than the entire costs needed to implement immediate kangaroo care across the whole of Malawi.

Unfortunately the world makes no sense.

Posted in Neonatal Research | Tagged , | 1 Comment

What do you think are meaningful long term outcomes for preterm infants? Give your opinion!

Annie Janvier is part of a group of collaborators who are trying to get multiple opinions about meaningful outcomes in preterm babies. It is part of a project being run by the Canadian Follow-up Network CNFUN, which is called Parent-EPIQ.

If you want to participate you can follow this link


Feel free to answer from wherever you are in the world, I just completed and it took less than 5 minutes.

Posted in Neonatal Research | Leave a comment

Which is worse; death or a low Bayley score? Comparing composite outcomes between groups, taking into account clinical priorities.

I keep harping on about this issue as I think we make a mistake in the design of our research studies when we include death and a much less important outcome in composite outcomes. For example in the STOP-BPD trial, the primary outcome “death or BPD at 36 weeks” was “not significantly different” between groups, but death before 36 weeks was somewhat lower in the steroid group than controls. There were more survivors to 36 weeks in the steroid group but exactly the same proportion of survivors with BPD in each group (65% vs 66%), therefore there were numerically more babies with BPD in the steroid group. Surely being alive with BPD is better than being dead! The distribution of moderate and severe BPD also favoured the steroid group, but that is not factored into the primary outcome analysis.

Methods exist for analyzing not only death, but the severity of the BPD, and giving a greater importance to death than to severe BPD, and lesser importance again to moderate BPD. Those methods can also take into account time to an event, so that, for example, a baby who no longer has oxygen at 6 months of age, is considered less severe than a baby who still needs oxygen at 12 months.

Death between 36 weeks and discharge is not better than death before 36 weeks, all deaths before discharge (there are very few after discharge in the first years of life) should be considered an equally adverse outcome, which is worse than surviving with BPD. Of course in adults, many of whom in cardiac research studies are even older than me, delaying death is a good thing to do, indeed it is the main aim of many interventions, preventing death is not an option! In newborn infants delaying death a few days or weeks is not necessarily a good outcome.

The problem of composite outcomes with differing clinical importance is not only a problem in neonatology, in cardiac research the outcomes often include death and hospitalisation for cardiac reasons, for example. In those studies the time to the adverse outcome may included in the analysis, and ways of comparing patients, so that a patient who dies is considered a worse outcome than one who survives but is hospitalised, and one who is hospitalised 2 years after intervention is considered a better outcome than one who is hospitalised after 2 weeks; also the time to a patient dying is clearly important.

As a result, ways of analysing composite outcomes that take into account the clinical importance of the outcome have been developed, in particular the Win Ratio. Patients are compared in pairs and if one has a longer survival than the other then the treatment they received is considered the winner, if the next pair has one patient dying and the other being hospitalised, then instead of a traditional analysis in which “death or hospitalisation” is the primary, and both patients are considered to have equivalently bad outcomes, in win ratio analysis the patient who survived despite being hospitalised is considered the winner, and their treament is given an extra point and so on.

When subjects are paired, by design, then the analysis is actually quite straightforward, if patients are not paired then you can compare every result from the treatment group to every result from the control group, and the analysis gets much more complicated, especially calculating the confidence intervals of the win ratio which seems to require heavy duty bootstrapping. One can also take into account stratification, and only compare within stratified groups.

I thought I would try this out on the recent data from the Inositol trial. The parts of the composite outcome at 2 years are dichotomous, death, yes or no, “NDI”, yes or no, so I did not need the primary data set to do this. In the Inositol group there were 60 deaths, who therefore were losers for all their comparisons to the control babies, except for comparisons with the 39 control babies who died, where there was no preference. Similarly the Inositol babies with “NDI” were winners when compared with the control deaths, but losers when compared to control baby survivors without “NDI”.

As there were 287 Inositol babies with known death or NDI outcomes, and 289 controls, there were 82,943 possible comparisons (which I evaluated one by one of course!!) The controls won 31,901 comparisons, Inositol won 29,171 comparisons and 21,871 comparisons were null. The win ratio therefore was 1.09. (to be honest Excel is very good at copying large ranges and if you know how to use relative and absolute cell addresses it is a bit laborious but not too difficult).

The calculation of the p-value is then not too complex, and using the formulae in the supplementary data of this article I calculated that z=2.45, which gives a p-value <0.01. (I hope I have calculated correctly), I tried to use an SAS program that was supplied to me, but never having used SAS before I have not yet been able to get it working, even though I downloaded SAS and followed some initial tutorials trying to learn it a bit (the things I do for my readers). The same source kindly sent me a program for calculating the 95% CI of the win ratio, but again I could not get it to work, so I used an approximate method from one of Pocock’s publications, which gave 1.02 and 1.17.

If you use the standard methods, as used in the publication, which give equal weight to death and “NDI” (which is mostly low Bayley scores), there is no significant difference between Inositol and control in the combined outcome. If you use the Win Ratio method, that takes account of the fact that being dead is worse than a low Bayley score, you find that the odds of any pair of patients having a better outcome if they were the one that got placebo was 1.09, with a 95% CI 1.02 to 1.17 (p<0.01).

That means you are significantly more likely to be a winner if you get placebo than if you got Inositol; using the same data that say there is no significant difference in “death or NDI”.

It has been said that, for the example of the SUPPORT trial, “does it really matter that the primary outcome variable was not statistically different? Everyone can read that the lower saturation group had more mortality”. My response is that, yes that is true for SUPPORT (even though, in fact, unadjusted comparisons of death using a chi-square, are not “significantly” different between groups), but in fact it is not true for the Inositol trial, that trial was suspended because of a manufacturing issue, and the unexpectedly higher mortality in the intervention group was not significantly greater at the time of evaluation of the primary outcome (50 in the inositol group and 33 in controls, chi-square with Yates’ correction=2.86, p=0.09). The trial also had more RoP in the treated group than the controls which made the combined outcome better in controls.

The great advantage of the win ratio method is that it can be used, as mentioned in the first paragraph, to give more importance to death than to severe BPD, and also more importance to severe BPD than to moderate BPD, etc. We could, preferably, include in a composite, outcomes that are more clinically important, such as numbers of rehospitalisations in the first 2 years, or duration of home oxygen therapy.

Sometimes parts of a composite may be difficult to prioritize, such as non-surgical NEC and severe BPD. Which is worse? They both have adverse long term impacts as well as short term morbidity, I guess that if we asked a group of parents the answers to that particular comparison would be mixed, so they could be given equal weight in the analysis

Trials are now being designed using methods such as this, and methods for sample size calculation have been published. (Redfors B, et al. The win ratio approach for composite endpoints: practical guidance based on previous experience. Eur Heart J. 2020;41(46):4391-9).

Other methods which can be used for composite outcomes which incorporate the clinical importance of the parts of the outcome are also available (Capodanno D, et al. Computing Methods for Composite Clinical Endpoints in Unprotected Left Main Coronary Artery Revascularization: A Post Hoc Analysis of the DELTA Registry. JACC Cardiovasc Interv. 2016;9(22):2280-8) such as the Weighted Composite Outcome method.

The time has surely come to design trials, especially those in which mortality is a potential outcome (the majority of neonatal ICU trials), using methods that take account of the clinical importance of the various outcomes that we are measuring. And to cease the methods that imply that death and BPD, or death and retinopathy, or death and low Bayley scores, are equivalent.

Posted in Neonatal Research | Tagged , | Leave a comment

Renal Function in the very immature preterm, what is a “normal” creatinine

The last time I blogged about this issue, there were a few comments from renal specialists which pointed out some limitations in my interpretation of the data. I defended myself admirably (it is my blog after all!) but I recognize that this is a complex issue.

Two new publications analyse serial serum creatinine in a cohort of preterm babies who all had multiple measures of creatinine concentrations. the first (Rios DR, et al. Creatinine filtration kinetics in critically Ill neonates. Pediatr Res. 2021;89(4):952-7) introduces the concept of creatinine filtration delay, with the idea that newborns have a delay before they start filtering creatinine of varying duration, being longest in the most immature infants. This is why creatinine concentrations take a while to fall and may indeed increase in the first few days of life.

I find this difficult to comprehend, if a baby is producing urine, then surely they must be filtering creatinine, even if GFR is very low (as it is immediately after birth) and tubular function is poor, I don’t think that there can be a selective delay in the filtration of creatinine. Perhaps they mean the term as a way of visualizing something else?

The idea I wrote about last time was that an increase in serum creatinine despite usual amounts of urine output may well be evidence of creatinine being reabsorbed in the tubules, one commenter noter that there is no reported mechanism for the reabsorption of creatinine, challenging that concept (that I must admit was not my idea, but the interpretation of a pediatric nephrologist of their data showing that creatinine clearance was lower than inulin clearance in the newborn rabbit, probably because of “back-leak of creatinine across leaky immature tubules“).

Around the time I posted that previous blog post, there was a publication from one of the commenters on the post (Askenazi D, et al. Acute changes in fluid status affect the incidence, associative clinical outcomes, and urine biomarker performance in premature infants with acute kidney injury. Pediatr Nephrol. 2016;31(5):843-51), which pointed out that infants lose extracellular fluid after birth, and the most immature babies lose the most weight after birth. This postnatal weight loss is due largely to a reduction in interstitial fluid, and thus to total body water. Creatinine is evenly distributed in all the body water compartments, therefore a reduction of total body water, without equivalent elimination of creatinine by the kidneys, will lead to an increase in serum creatinine concentrations.

I realize when re-examining the data from Rios et al, and their story of creatinine filtration delay, that they did not take into account the postnatal weight loss and, as the most immature babies have the highest extracellular fluid compartment (as much as 800 mL/kg, compared to about 400 at term), the impact of postnatal weight loss on serum creatinine concentrations is greatest among that group. I think it is most likely that the majority of the delay in the fall in serum creatinine is because of contraction of extracellular fluid and concentration of the creatinine while GFR is still low, rather than a delay in creatinine filtration.

I think this is an important part of the answer to the issues, Askenazi et al give a way of calculating an adjusted serum creatinine concentration accounting for the increased concentration due to weight loss. Which means, for example, that a small preterm baby who has a serum creatinine of 100 μmol/L at birth, the same as her mother, and then loses 13% of their body weight over the first 3 days and then has a re-measured serum creatinine of 120 μmol/L actually has a new adjusted serum creatinine of 100.5; If the baby weighs for example 1000g, and we assume a total body water, TBW, at birth of 800mL/kg then we can calculate the adjusted creatinine as 120 x (670/800) =100.5 (800 being the TBW in grams at birth and 670 being the TBW after loss of 130 g weight).

The total body creatinine, adjusted for weight loss, should be falling if renal function was “normal” to eventually arrive at a value which is appropriate for the size of the infant, but at least the adjustment for weight loss does give us a better indication of the real situation. Then we have to figure out why the adjusted serum creatinine is not falling, despite a urine output that has commenced.

This graph from Askenazi’s paper shows that this is a frequent finding in the very immature baby.

You can see that under 26 weeks, on average serum creatinine rises, but even the “fluid-adjusted” creatinine, taking into account weight loss, does not fall as you might expect; in more mature babies the “fluid adjusted” Serum Creatinine (FA-SCr) falls immediately after birth.

In this study there were 16 babies under 27 weeks of a total of 41 who had aute kidney injury according to the definition of an increase in serum creatinine of at least 27 micromol/L (0.3 mg/dl), 11 of them still had AKI by that definition when using adjusted creatinine values.

Of course, if there is glomerular filtration, creatinine is being filtered (I don’t think that creatinine filtration delay is a real thing) the Fluid-Adjusted Creatinine concentration will vary depending on the balance between GFR, and thus the creatinine clearance, and how much creatinine is being produced. Creatinine is produced by degradation of creatine, which is largely from muscle mass. So tiny preterm babies probably produce very little, but they also have very low GFR especially on day 1.

The other new study I mentioned is a study of 158 ELBW babies from Leuven, which modelled creatinine kinetics (van Donge T, et al. Characterizing dynamics of serum creatinine and creatinine clearance in extremely low birth weight neonates during the first 6 weeks of life. Pediatr Nephrol. 2021;36(3):649-59) and in their model they included a term for creatinine production which averaged 3.55 mg/day (SD 1.44) which is 31.4 μmol in modern units, the average weight of the babies was 820g, so a production of about 38 μmol/kg/d. (An adult produces somewhere around 160 μmol/kg/d).

These are the reference ranges they produced:


The model takes into account the fluid loss and reduction in total body water which leads to increased concentration in serum creatinine, the very low initial GFR, and the ongoing production of creatinine.

I think this is the best explanation of how all those factors interact, and gives reference ranges which can be used. It does use data from a group of babies of widely varying disease severity, some of whom were sick and at risk of kidney injury. However, their data did not seem to show an impact of inotropes, but did show a small impact of ibuprofen use.

How the use of these ranges helps in the definition of clinically important kidney injury will be important in the future. But it starts to explain to me the changes in serum creatinine in the very preterm.

For example: if creatinine clearance is as low as 0.2 mL/min on the first day of life, and an 820 g baby starts with a serum creatinine of 100 μmol/L, then they have a total body creatinine load of (0.820 x 0.8 (TBW) x 100)= 66 μmol. Filtering 0.2 mL/min means they filter 29 μmol of creatinine on day 1 while producing 31 μmol of creatinine, which would lead to an increase in total body creatinine load to 68 μmol, and if they lose 5% of their body weight (and 9% of their TBW) then their serum creatinine concentration will rise to 110 μmol/L.

A baby with a greater weight loss and a somewhat lower creatinine clearance will of course have a greater increase in their serum creatinine concentration..

I don’t know if you could use these data to try and back-calculate the creatinine clearance of an individual baby; then perhaps a certain lower limit of clearance could be used to define Acute Kidney Injury in the preterm. I do, however, think that these data clearly invalidate the definition of neonatal AKI as an increase in SCr of 27 μmol/L (0.3 mg/dL), and even using fluid adjusted SCr with that threshold. I think we need a new definition which is based around creatinine clearance which is lower than the gestational and postnatal age adjusted “normal”, or if that is too complex for routine use, then graphs such as those above with the addition of percentile lines may be enough for a working definition: a baby with a creatinine which rises too fast and crosses percentiles could be considered to have AKI.

I know I went a bit down the rabbit hole with this post, but I like understanding things, and when I don’t I get agitated and follow reference chains until I either get fed up or find a conclusion! I guess the take home message is that creatinine concentrations rise after birth, partially explained by postnatal weight loss, and also by low GFR, but we don’t have a good definition of AKI in the preterm infant.

Posted in Neonatal Research | Tagged | 3 Comments