Does vitamin E make preterm babies smarter?

In this study about the potential effects of α-tocopherol on IQ in ELBW infants, there are a number of problems: Starting with the title, Long-term alpha–Tocopherol supplements may improve mental development in extremely low birth weight infants. Acta Paediatrica. 2014. (It is by Kitajima and others, and published on-line in Acta.) A title should give some idea of the methodology that was used, this was not a prospective study which you need to argue causation, but a retrospective review. Also using the title to promote one interpretation of the results, which could easily be challenged, is not a good idea. The title states ‘may improve’, but the phrase ‘may be associated with’, would be much more appropriate.

The introduction refers to the study as a  retrospective cohort study, but in the sentence just before that they talk about the results of the study. Maybe I’m an old stick-in-the-mud, but I think the introduction is where you explain why you did the study, the methods is where you describe what you did, and the results are where you say what you found. (The discussion section is then where you make extravagant claims about the significance and implications of your findings, and end up with a section calling for further research).

If we try to figure this out, there were 579 ELBW survivors from an NICU over 12 years, ending in May 1998. 45% of them were followed up, the last follow up was at 8 years of age, which means that the study was finished 8 years ago.

Now what really confuses me are some of the other details of the report, the babies had vitamin E started at between 3 and 4 weeks of age; the authors talk about the infants being ‘divided into groups’ either getting no tocopherol, tocopherol for between 1 and 6 months, or for more than 6 months. It seems as I delve deeper that the babies who were ventilated for less than 4 weeks did not get it, those ventilated for more than four weeks ‘continued’ the tocopherol until discharge or until 4 months, and those who went home on oxygen had the tocopherol until the oxygen was discontinued. So the groups are very different, the long use babies had much more dexamethasone use, for example.

The results actually show that the IQ at school age was highest in the group who did not receive any tocopherol. Which isn’t exactly what they say in the introduction, abstract or title! Only by combining the two groups, those who received no tocopherol and the 1 to 6 month tocopherol, were the authors able to claim that the very prolonged tocopherol had higher IQ, but there are so many confounders that I really don’t think that any analysis of these data can be considered reliable; even if you do all the data manipulation that these authors do, leaving out high income families, infants who had very large amounts of dexamethasone etc. All you can say is that infants in the group who got no tocopherol had higher IQ than those who got 1 to 6 months, and those who got extremely prolonged supplementation had intermediate values. Whether this was due to the tocopherol or to other differences between the groups is open to question. In fact that pattern suggests to me a random effect.

Posted in Neonatal Research | Leave a comment

Let’s scrub those hubs!

‘Scrub the hub’ is a part of many catheter sepsis prevention bundles. The idea being that organisms can be introduced into the infusion solution from inadequate cleaning of connectors, so 15 seconds minimum of vigorous application of alcohol wipes should make the intravenous tubing hub cleaner, and reduce subsequent infection. Many of the individual components of the ‘bundles’ have not been well studied, including the hub-scrubbing.

In the NICU, with our very high rates of late-onset sepsis (Health-Care Associated infections HCAI) compared to most health care settings, this might be even more important, or perhaps hubs are not a source of our infections.

According to the introduction of this before and after study, this is the only study of scrubbing the hub. The before part of the study was when the nurses were trained to wipe the hub with an alcohol swab, the after was specifying the 15 seconds of good scrubbing.

There were relatively few very preterm babies among the 860 or so total infants in the study (around 580 before and 280 after). Only 14 before and 6 after were under 28 weeks, so it is rather a different NICU to mine, a very high proportion of their central lines being in term and late preterm infants. There were 14 positive cultures for Coagulase Negative Staphylococci (CoNS) in the before period, and 3 in the after period, when possible contaminants were eliminated, there were 9 before and 0 after. All but one of the babies thought to have true sepsis were preterm, (less than 32 weeks). There is no reporting of any other infections in the paper, so we don’t know if they had no other infections (Gram negatives, staph aureus or enterococci) or they just didn’t report them. I really think that ought to have been included in the report.

The authors state that there were no other changes in sepsis control procedures between the 2 periods, but, of course there is no way to know if all risk factors were balanced without randomization of a large group of babies, which will, in general, balance for all other risk factors, known and unknown.

So this looks positive and the authors of course recommend a large randomized trial, as you are always supposed to. But: Really? Would that be ethical? Randomizing babies to have a reduced duration of ‘hub-scrubbing’ in order to see if they can get more infections? Surely when an intervention makes sense, costs nothing, and has no reasonable potential for adverse effects, we should just do it. The ‘cost’ of 20 seconds extra nursing time per catheter intervention, and of incorporating into their teaching that they should scrub for 15 seconds instead of just wiping the hub, is trivial. How could this possibly be harmful? Sure a large RCT might show that it is ineffective, but then there is still no-one harmed, just a few minutes of wasted time. RCTs of infection control interventions are indeed sorely needed, but we should concentrate on things which are either costly or have a potential for adverse effects.

In the meantime, scrubbing the hub for 15 seconds should be the minimum standard, we do need to concentrate on methods to ensure that it actually gets done, for a full 15 seconds (with a further 15 second delay while the alcohol evaporates). If, like me, you watch people in the NICU when they complete their 15 second alcohol hand gel application in 4 seconds, it is easy to get cynical about how long it is actually being done for.

One thing that helps I think (though I don’t know any good data) is to get people to sing ‘Baa-Baa black sheep’ while gelling their hands, the rhyme takes around 15 seconds if you don’t rush. We just have to decide which French rhyme to use for my hospital!

Incidentally a recently-published multi-center program to reduce catheter associated sepsis in paediatric haematology/oncology patients, which was highly effective, included ‘scrub the hub’ either 15 seconds with alcohol or 30 seconds with chlorhexidine. The swabs used in the neonatal paper above were alcohol with 5% chlorhexidine.

Posted in Neonatal Research | Tagged | Leave a comment

Umbilical cord flow after birth, before clamping

Well obviously before clamping, after clamping there isn’t much flow! In this study with full term babies they were delivered and placed on the mother’s chest, and then an ultrasound probe was placed to measure umbilical cord flow.  (Boere I, et al. Umbilical blood flow patterns directly after birth before delayed cord clamping. Archives of disease in childhood Fetal and neonatal edition. 2014).

This is great study, really important in helping to understand what is actually happening during delayed clamping, and what probably happened in the past as we were evolving and didn’t have cord clamps, although exactly how the cord was managed in the distant past is unknown, was it bitten by the mother? left to dry?

The babies were all vaginal deliveries at term who breathed spontaneously.  The cord was clamped at a time determined by the midwife doing the delivery, the local standard was to wait at least one minute and after the cord stopped pulsating. Clamping occurred on average at 5 minutes after delivery, and the researcher checked to see if there was still pulsation palpable just before the cord was clamped.

Arterial flow often stopped before cord pulsation disappeared, and often continued longer than venous flow. Venous flow continued for an average of 4.5 minutes in babies in whom is stopped before cord clamping, and as long as over 8.5 minutes (that is the third quartile, so the maximum duration isn’t reported); some babies still had some pulsation when the cord was clamped, which was at 5.5 minutes and up to over 9.5 minutes.

The duration of flow was extremely variable, but the fact that venous flow often stopped first suggests that prolonged delays in cord clamping might sometimes decrease placental transfusion.

The infant’s breathing pattern affected flows, in particular vigorous crying reduced or even reversed venous flow, but large inspirations increased flow.

This technology doesn’t allow volume measurements, as you would need to be able to measure accurately the vessel diameters, which are probably themselves changing. A similar studies in preterms, and during resuscitation would be fascinating.

Posted in Neonatal Research | Leave a comment

A CPS statement thought experiement

Its been a while since I ranted about the CPS position statement regarding decision making around extremely preterm delivery. But I am currently in Europe where certain countries have specific thresholds for willingness to provide intensive interventions, thresholds which are somewhat more rigid than the guidelines of the CPS, which are often ignored. This made me think (dangerous I know) about what justification you could have for giving a specific threshold in completed weeks of gestational age below which stabilization and active intensive care for an extremely preterm baby would not be offered.

I think that for such an approach to be rational you would need 2 things, you would need to be able to know the gestational age with certainty, and you would also need to show that there was some substantial change in outcomes (survival or another outcome that you thought was important) at that precise moment. Then you could justify intervening at say 24 weeks and 0 days, but not at 23 weeks and 6 days. Or using the threshold of the CPS position statement, be sometimes willing to offer intervention at 23 weeks and 0 days, but never at 22 weeks 6 days.

We know that the second criterion is not satisfied, there is no step-wise change in outcomes at any particular date, and certainly not at the end of an arbitrary 7 day cyclical period known as a week. There is a gradual progressive change in survival with advancing gestational age, and major effects of birth weight, sex and other variables. In contrast there is little or no effect of gestational age on long term neurodevelopmental outcomes among survivors.

Which led me to think of the following story:

Suppose a mother pregnant with twins following IVF was in threatened labour. She has passed 22 weeks, and has been seen by the neonatologist. She wishes institution of intensive care, but is told by the neonatologist that it is not an option before 23 completed weeks; the consensus of the hospital where she is being looked after is that they should follow the CPS guidelines, and be prepared to intervene actively at 23 weeks, (but only after the mother has had the litany of possible negative outcomes related to her), whereas prior to 23 weeks, life-saving interventions are not offered.

Her contractions continue and the cervix dilates.  Just before midnight on the day she reached 22 weeks and 6 days she delivers a little girl, who is placed in her arms for comfort care. 30 minutes later, just after midnight, the little boy delivers, and is intubated and taken to the NICU.

Would you actually do that? If you were to follow the CPS statement, the answer should be yes. Even if you are prepared to be flexible with twins, and treat them as a ‘package deal’ what about 2 mothers in adjoining rooms who deliver before and after midnight?

And what if the next day the mother comes to you and tells you that she had the Egg retrieval for her IVF at 2 pm in the afternoon. Oh no! The boy is actually only 22 weeks 6 days and 14 hours! We should not have intervened, maybe we should switch to comfort care…

You then find out that mother is referring to the original egg retrieval a couple of years ago, the obstetrician had known that and calculated the dates from the transfer of her embryos, which had been done at midnight because of a power failure during the day. Phew! So she really was exactly the gestation that they thought. we did the right thing after all.

The next day the obstetrician phones you sheepishly, to tell you that the mother actually had transfer of 5-day blastocysts not 3-day embryos. Which means that she had been actually 2 days further advanced in her pregnancy than they thought, so the little girl had actually been 23 weeks and 1 day gestation. So she should have been offered intensive care… Too late now.

This may seem absurd, but arbitrary limits to intensive interventions based on completed weeks of gestational age lead to such absurdities. When you add in the uncertainties of gestational age, the absurdities are multiplied.

This is a plea to recognize messiness. There are too many unknowns, and too many variables to have simplistic rules. Some of the variables can be quantified with precision, such as birth weight, some can be quantified but are imprecise (gestational age), and some can not be quantified at all, such as the values of the parents. We have to do the best we can, for the families we care for, making the best decisions possible in the uncertainties of real life.

Posted in The CPS antenatal counselling statement | 1 Comment

Brain Food for Preterms

All food is brain food when your brain is making 250,000 new neurones every minute.

In a small two-center trial first published in 2013 50 very preterm babies were randomized to different parenteral nutritional intakes. The main differences between the groups were that the controls started at 2 g/kg/d of protein, compared to 3.5 in the intervention group. The controls started at 0.5 g/kg/d of lipid given as ClinOleic, the intervention group started at 2 of SMOFLipid; both gradually increased to a max of 3.4. One they were receiving 110 mL/kg/d of milk the fortification was different between the groups, with the enhanced group getting an extra 0.8 g/kg/d of protein, to achieve 4.4 g/kg/d, and getting extra omega 3 fatty acids to arrive at a total of 166 kcal/kg/d, compared to 146. (which is a whole lot of calories!)

Unfortunately the planned sample size was not achieved, the study was stopped early because there was more sepsis in the enhanced nutrition group (p<0.04). Stopping a trial for a barely significant difference between groups is really not a good idea, but it is often a difficult decision, especially when the difference is an important adverse outcome. The primary outcome was post-natal growth restriction. The authors were following a number of secondary outcomes as were the data monitoring committee. The risk that one of those outcomes will be ‘statistically significant’ at some point during the data accumulation is enormous. That is why stopping rules should be much more stringent than that. Of course if you continue the study and it turns out that there really is more sepsis at the end of the study, you may then be criticized, and you might feel bad.  You might indeed face totally ridiculous criticisms similar to those that were directed at the SUPPORT trial by Public Citizen, who don’t understand how interim analyses work.

Having said that the authors published a second paper with the growth data; because of the small sample size there was a difference in the birth growth variables between groups, which actually favoured the controls, more SGA babies in the enhanced nutrition group. Despite this there were significant advantages of the enhanced protocol. None of the non-growth restricted babies in the intervention group became growth restricted, compared to 1/3 of the babies in the control group. The z-scores for weight dropped by about 0.3 for the intervention group, between birth and 30 days of age, and by about 0.7 in the controls. Head growth was also better.

A subset of the infants had MRI performed at term, which has just been published, and which is why I am writing this post. Strømmen K, Blakstad EW, Moltu SJ, Almaas AN, Westerberg AC, Amlien IK, et al. Enhanced Nutrient Supply to Very Low Birth Weight Infants is Associated with Improved White Matter Maturation and Head Growth. Neonatology. 2015;107(1):68-75. (yes its the first article of 2015 on the blog). There were differences between the groups. Which is about as much as I can say about head MRI, the mean diffusivity was lower in the intervention group. Which is good, apparently. In just about every region of the brain there were significant differences in what I shall from now on refer to as MD, as if I knew what I was talking about.

So give more food. It makes the babies grow better, including their head, and it improves their MD, which is a good thing to do.

Posted in Neonatal Research | Tagged , | 2 Comments

It’s not enough to be good

You have to keep practicing as well.

The way our health systems are organized has impacts on outcomes. Multiple small centers doing complex surgery and looking after fragile patients only infrequently will likely have poorer outcomes overall. Large centers which gain experience in those procedures, which develop protocols and approaches, and practice frequently are likely to do better, by which I mean to have more babies who are alive at the end.

It seems so self-evident, that to keep saying it should be redundant, but it isn’t; despite all the evidence that is out there, we still see a deterioration in regionalization, some of which is driven by the idea that the medicine is just a business like any other, so we should let the market decide who is going to do critical procedures. The idea, I guess is that if a center kills enough babies, then eventually they will receive fewer referrals.

Or you could take a different approach and say that highly critical care should be restricted to a few centers, and other centers should not be allowed to do those procedures.

These two new publications address this issue.

Karamlou T, Jacobs ML, Pasquali S, He X, Hill K, O’Brien S, et al. Surgeon and Center Volume Influence on Outcomes After Arterial Switch Operation: Analysis of the STS Congenital Heart Surgery Database. The Annals of Thoracic Surgery. 2014;98(3):904-11.

The arterial switch procedure is highly demanding, both in terms of technical skills of the surgeon, and peri-operative support. This article confirms that both the volume of the surgical center and the number of cases performed per surgeon are important in survival rates.

Kastenberg ZJ, Lee HC, Profit J, Gould JB, Sylvester KG. Effect of deregionalized care on mortality in very low-birth-weight infants with necrotizing enterocolitis. JAMA Pediatrics. 2014.

The progressive de-regionalization of neonatal care in some parts of the USA has led to multiple small units with little experience looking after critically ill babies. These data show that high volume, high-level NICUs have significantly lower mortality for infants suffering from NEC.

Which is consistent with a recent article from the EpiCure2 study that I already blogged about, showing that overall, tertiary centers have better outcomes for tiny babies than centers without tertiary NICUs.

None of which should be surprising, but the implications for the organization of our health care are often ignored.

Posted in Neonatal Research | Leave a comment

Neonatal Updates

I haven’t done one of these for a while, getting too verbose I  guess, but here is a selection of interesting recent articles with just a few sentences about them.

Mola-Schenzle E, Staffler A, Klemme M, Pellegrini F, Molinaro G, Parhofer KG, et al. Clinically stable very low birthweight infants are at risk for recurrent tissue glucose fluctuations even after fully established enteral nutrition. Archives of Disease in Childhood – Fetal and Neonatal Edition. 2014. VLBW and ELBW babies continue to have unstable blood sugars, with low and high values even when you think it was safe to stop monitoring them so closely. What do we do about this? I sure don’t know.

Okumus N, Beken S, Aydin B, Erol S, Dursun A, Fettah N, et al. Effect of Therapeutic Hypothermia on C-Reactive Protein Levels in Patients with Perinatal Asphyxia. American journal of perinatology. 2014(EFirst). Asphyxiated babies under hypothermia have elevated CRP. Asphyxia, and/or cooling is one of the many things that makes your CRP rise.

Hauglann L, Handegaard BH, Ulvund SE, Nordhov M, Rønning JA, Kaaresen PI. Cognitive outcome of early intervention in preterms at 7 and 9 years of age: a randomised controlled trial. Archives of Disease in Childhood – Fetal and Neonatal Edition. 2014. At 7 to 9 years of age the authors could find no benefit of their early intervention program on cognitive outcomes. I do not think this means that we should ditch the programs. They will be harder to justify, but I think we should still be helping families to learn how to help out their premies at home, and we need to focus on subgroups that have ore benefit, which in other studies are those with lower socio-economic status.

Meyer MP, Hou D, Ishrar NN, Dito I, te Pas AB. Initial Respiratory Support with Cold, Dry Gas versus Heated Humidified Gas and Admission Temperature of Preterm Infants. The Journal of Pediatrics. 2014. We don’t use cold dry gases to ventilate our premies at any other time, so why in the delivery room? This is a multi-center RCT examining the question. Using Warmed humidified gases resulted in more infants being normotthermic on admission to the NICU, especially the very preterm babies. With 100 babies in each group there were no other significant differences in clinical outcomes, but most differences were in the direction of a beneficial effect of warming and humidifying the gases. A larger study is worth doing I think but will be hard to fund with these modest differences between groups.

Posted in Neonatal Research | Tagged , | 1 Comment