Feeding and NEC

Necrotizing Enterocolitis remains a devastating disease. One of the major causes of mortality in Very Low Birth Weight Infants, it often strikes when babies are starting to do well. Because it is relatively unpredictable, observational studies are potentially useful, but can easily be misleading. In particular, observational studies which are performed as a result of a perceived change in incidence might easily be biased.

One recent study that was published has received some publicity, I myself received some links on Linked-In pointing to this study, which at first sight seems to show that advancing feeds slowly might decrease NEC. Lets examine the data a bit more carefully.

In 2009 a hospital changed their feeding policies. A paper published in 2014 reported their data from 2009 to 2012. According to the published article the paper was received on July 24th 2014, and accepted on August the 5th. Which immediately makes me wonder two things, why the authors decided to perform this study, and how do they get a paper accepted so quickly? If the authors performed the study because they saw a decrease in NEC, and then decided to do the analysis, that immediately creates a bias, many other hospitals could possibly have made a similar change and not noted any change in NEC, and therefore not looked at their data. Studies such as this are more likely to be reliable if the decision to prospectively collect and try to publish their data is made at the time that the change in therapy is introduced. But then a similar study showing no effect on NEC would be very difficult to publish, and probably not be accepted in under 2 weeks!

There are a couple of concerning things in their findings, one is that after the change in the protocol there was not actually statistically any less NEC overall than before the change in protocol. Only by a subgroup analysis, was the group of babies under 750g significant, and then they did not report the p-value of the interaction, which is really essential. They showed much less NEC in the babies under 750 g birth weight than those between 750 and 1000g. Which really strongly suggests to me that this is just the result of the random nature of NEC incidence, I don’t know any large study that has ever found more NEC in the larger babies.

More worryingly, the new protocol led to babies having later commencement of feeds, longer use of intravenous nutrition, longer use of picc lines, and really horrendous nutritional outcomes, with over half of the babies being under the 10th percentile at discharge (admittedly that was better than the 75% with their older protocol).  The feeding protocol that they introduced is entirely non-evidence based, as I have noted before here, there is no evidence from controlled trials that varying the advancement of feeds has any effect at all on the incidence of NEC.

A slightly older study (from 2009) provides interesting data from the German Neonatal Network, and is, I think, much more reliable. They compared the outcomes of babies from centers where it took on average less than 12.5 days to get their VLBW babies on full feeds, to centers where it took on average more than 12.5 days to get the babies to full feeds.


You can see here the clinical outcomes from the study. The only things that were different between the groups were the late-onset sepsis, which was much higher in the babies with slower feeding. Surgical NEC was slightly (non-significantly) more frequent with slower feeding, and severe retinopathy was also higher, which is entirely consistent with the recent information linking retinopathy with poorer nutrition.


Posted in Neonatal Research | Tagged , | 1 Comment

Frozen poop sounds much more palatable

Fecal transplants are effective in treatment of persistent clostridium difficile infections. The publication showing that, in the small RCT published in the PNEJM, was remarkable in the pre-screened donors who were ready and available to produce a “donation” on demand.

This new study (Youngster I, Russell GH, Pindar C, Ziv-Baran T, Sauk J, Hohmann EL. Oral, Capsulized, Frozen Fecal Microbiota Transplantation for Relapsing Clostridium difficile Infection. JAMA. 2014;312(17):1772-8.) takes poop from screened donors and freeze dries it and puts it in capsules: 70% of the small group of patients resolved their diarrhoea after a single treatment. The others almost all resolved after repeat treatment. This way of treating has the advantage of not needing a list of on-call volunteers, and not needing to pass a tube into the intestine. It has the disadvantage that if you forget, and bite down on the capsule….

Posted in Neonatal Research | Tagged | Leave a comment

Magnesium for brain protection and longer term outcomes.

Two new articles describing the longer term follow up (to school age) of antenatal magnesium, when used for brain protection.

1. Doyle LW, Anderson PJ, Haslam R, Lee KJ, Crowther C, Australasian Collaborative Trial of Magnesium Sulphate Study G. School-age outcomes of very preterm infants after antenatal treatment with magnesium sulfate vs placebo. JAMA. 2014;312(11):1105-13.

The ACTOMag trial presents data from 669 of 867 known survivors. There are slightly fewer deaths with Magnesium (14% compared to 17.6%) which could be due to chance (p=0.08), and no evidence of any other benefit, or indeed of harm.

2. Chollat C, Enser M, Houivet E, Provost D, Benichou J, Marpeau L, et al. School-age outcomes following a randomized controlled trial of magnesium sulfate for neuroprotection of preterm infants. The Journal of pediatrics. 2014;165(2):398-400 e3.

431 of 616 survivors from the PREMAG trial were seen, no substantial differences in outcomes were seen. Most of the minor differences were in the direction of benefit of Magnesium. Mortality was again a little lower with Mg than control.

Think we need a new systematic review; I nominate Lex Doyle. All those in favour…

Posted in Neonatal Research | Tagged , | 1 Comment

Murky Guidance

After the OHRP meetings and their evaluation of the SUPPORT controversy, the OHRP have now released what they call “Draft Guidance on Disclosing Reasonably Foreseeable Risks in Research Evaluating Standards of Care

The conclusion of the draft guidance starts with the following sentence, which is supposed to clarify their position.

In summary, if a research study examining standards of care includes as a purpose evaluating identified risks associated with those standards of care, the identified risks associated with the standards of care being evaluated that are different from the risks of standards of care at least some of the subjects would be exposed to outside of the research study are generally considered by OHRP to be reasonably foreseeable risks of research.

As we say, “clear as mud”.

Immediately before the conclusion they give an example which has the advantage of being somewhat clearer :

Consider this hypothetical example: It is known that treatment using surgery and radiation has a high likelihood of curing a particular form of childhood cancer, but that the radiation produces a significant risk of other cancers developing later in the child’s life. Consequently, some doctors treating children with this cancer use a smaller amount of radiation.  Both amounts of radiation are consistent with clinical care guidelines and considered to be within the standard of care. There is little evidence available comparing the outcomes of the two treatments in terms of their cure rates or the development of later cancers.  A randomized clinical trial is proposed with subjects to be assigned to treatment with the higher or lower amount of radiation to compare the effectiveness of the two treatments in curing the current cancer and how often later cancers occur.  Parents should receive and evaluate appropriate information before deciding whether to permit their child to participate in this greater-than-minimal-risk study. This decision is important and consequential. They should be told about the reasonably foreseeable risks of the study, that the treatment with the larger amount of radiation poses a possibly increased risk of cancer in later life, and that the treatment with the smaller amount of radiation poses a possibly increased risk of the original cancer not being cured.

Of course what that implies, but is not specifically stated in this draft guidance or in this example, is that you should then go on to say, the treatment with a larger amount of radiation poses a possibly increased risk of death, and treatment with a smaller amount of radiation poses a possibly increased risk of death.

In this example there is good reason to guess what the different risks might be, and to perform a study to try and find the balance between those risks might be justifiable, I might suggest though, that if those risks are already known, and can be quantified, based on prior data or experience, then the study might be questionable. Don’t you already have enough data to engage the child and family in a shared decision?

All standards of care do not pose identical risks, either in terms of the nature or the degree of the risks, and individuals often may have reasons to prefer the risks of one standard of care over the risks of another. Disclosing the reasonably foreseeable risks of research to prospective subjects recognizes the ethical obligation to give prospective subjects sufficient information to make a knowledgeable decision about whether or not to participate.

Its hard to disagree with those sentiments, but putting them into practice when you really have little idea what the balance of risks and benefits are is more of a problem. What is a reasonably foreseeable risk of the research into (to pick an example entirely at random) whether we should treat hypotension numbers with dopamine, in very preterm babies, or whether we can wait to see if a baby develops signs of end-organ hypoperfusion? Dopamine might increase the risk of IVH, or, under-perfusion of the brain might predispose to IVH, and may not be detectable clinically, and the same comment could be made about any of the common complications of prematurity. Do parents want 2 lists of all the common complications of prematurity, and a statement that they may be either increased or decreased by the therapy?

I do agree that parents must know that the whole idea of the research project is that there might be a difference in one or other of those outcomes, and, at the end of the trial, it may be that there is an advantage to being in one group or the other.

In one sense that is what we want; we don’t want to always do research with a null result. We want to have a real difference so we can advance care, and participants also, I think, want the same thing; they want to participate in research that improves future care, as long as we don’t expose them to additional risks over what they may be exposed to in any case.

If we try to apply the reasoning in the guidance to the SUPPORT trial, then I guess you would have to say, if today you were asking for consent for the trial, with the same understanding that we had when it was performed, that randomization to the low saturation group poses a possibly increased risk of adverse outcomes, including death, cerebral palsy, and developmental delay, and randomization to the high saturation group poses a possibly increased risk of adverse outcomes, including death, cerebral palsy, developmental delay and retinopathy.

(Before the trial was performed, the extra oxygen exposure in the high group was thought to maybe increase the risk of bronchopulmonary dysplasia, which is associated with more death and adverse long term outcomes. Indeed, there is more oxygen dependence at 36 weeks, by a little less than 10% in the high saturation target groups. For most of the trials it is not significant (and therefore could just be due to chance), but in the combined BOOST-2 data it is statistically significant. )

I think it must also be stated, and this is not clear I don’t think from this guidance, that when there is no known advantage of any of either arm of the study, then that should also be in the consent process, as an explanation of why the heck you would want to do the trial.

I think we should focus on the fact that the consent forms are only a small part (not necessarily the most important) of the consent process. They constitute the basic minimum, and the part which is documented.

Shouldn’t we be looking for other, adaptive, more complete ways of informing participants in trials? I think we should be developing websites for informed consent, sites where participants (or parents) can go and get whatever detail of information interests them, where the parents can find the complete protocol, all the references, or a selected group of key references, and all the justification for the trial; if they want it. If they don’t want to wade through all the detail (and they should and could decide themselves) they could give consent after a ‘bare-bones’ description of the trial. Also a parent/participant could then go back to the consent website and get more information later, whenever they wanted. If there were a web platform designed to do this, it could be relatively simple to fill the various parts with the information.

Posted in Neonatal Research | Tagged , | 1 Comment

What do parents of twins (and the twins) think about randomizing multiples

There has been some controversy recently about how we should randomize twins (and triplets or more) in neonatal trials.

The question being whether randomizing the babies independently or randomizing them as a group is preferable. Those who wish to randomize as a group think it will help enrollment, without compromising the trial, those who think we should randomize individually are worried about the loss of power, and that there should at least be an analysis of such babies as a cluster. For some studies, especially in the preterm infant, a large proportion of eligible babies may be twins, so it can become an important issue.

Now fortunately someone has asked parents their opinion! And twins!

Bernardo J, Nowacki A, Martin R, Fanaroff JM, Hibbs AM. Multiples and parents of multiples prefer same arm randomization of siblings in neonatal trials. Journal of perinatology. 2014.

The authors from Rainbow Babies Hospital in Cleveland contacted about 200 parents of multiples, and over 300 multiples (97% twins) and asked them their preference. They were contacted from numerous sources, and quite a large proportion had actually been in an NICU (about half the parents and 1/3 of the multiples) and/or participated in research (about 19% of parents, and 9% of the multiples).

Most of the correspondents had a preference as to randomization strategy, about 70 to 80%, and most of those, particularly among the parents, preferred randomization to the same group (70 to 80% among parents and 60% among the multiples themselves), very few preferred being randomized into the alternate group (a strategy I didn’t talk about above but which has been proposed as a way of balancing groups), and a few more preferred independent randomization, (between 12 and 20%).

I think this is pretty conclusive, and very important. If parents and babies (or rather former babies) largely prefer being randomized into the same group, then we should probably in the future plan to randomize multiples as clusters, and analyze as clusters.

I gave a presentation at the 2012 PAS meeting about this very issue. In fact you can look at the presentation if you want on this website/blog, under presentations from our group, the 2012 Boston PAS meeting. My conclusion was the the impact of randomization method will really depend on the intra-multiple correlation. In a disease with a strong genetic component, and often similar outcomes between twins (such as BPD) the implications are different than for a disease such as severe IVH where there is much lower correlation between siblings. One of my slides says this :

If infants are randomized to the same group, i.e. as a cluster, then the analysis should take account of that

If there is little correlation between twins (eg severe IVH) then a strict cluster approach, such as the multiple outputation technique will reduce power, and risk  type 2 error

If there is strong correlation between twins (eg BPD) then not correcting for clusters will, in general, make confidence intervals more narrow, and risk type 1 error

And even if you independently randomize twins, some of them will end up in the same group by chance ; you should probably, strictly, account for that in the analysis (particularly for diseases with a stronger genetic component), but even worst case scenarios show the total impact is minor.

One of my recommendations at the end of the presentation was that we should look at the impacts on families, I wasn’t exactly thinking of this kind of questionnaire, but I am very happy it was done.

The other question I have from a methodological point of view is this, now that we have this information, and, although clearly in favour of randomization as a cluster, it is far from unanimous:

What if we gave parents, at the time of randomization the option of having both their twins (or all the triplets) in the same group or of being randomized independently?

They could then be counted as either individuals or as a cluster in the randomization schedule, and analyzed as a cluster if randomized as a cluster. It would be a bit more complicated, but I can’t immediately think of any objection, and it would respect parental wishes, and show them how important their views are, at a critical time in the relationship of the researchers with the family.

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

Reassuring Prophylactic Indomethacin Data

One of Annie Janvier’s first research projects was a case control study of the influence of prophylactic indomethacin on intestinal perforations. Under my supervision she analyzed cases of spontaneous intestinal perforation (SIP), and we analyzed the influence of prophylactic indomethacin, which was highly significantly related.

As we knew at the time, there are numerous biases in this type of research. For one, we decided to do the study because we had just had a run of SIP; often such studies are stimulated by just this kind of phenomenon. Which immediately introduces bias: what you should probably do, if you have a cluster of adverse outcomes, is to take your suspicions and analyze data from a completely independent data set, of which you are not a part. But few of us have access to such data.

We had recently introduced indomethacin prophylaxis, and had this run of SIP, so the analysis did show an associations between the 2, which made us worry, but we certainly weren’t convinced that the association was causal, nor that our data were free of bias.

As an aside, Annie, a fellow at the time, was really pissed off when a well known senior researcher, who shall remain nameless, walked past her poster and sniggered. Don’t do that folks (most of my good friends and colleagues wouldn’t dream of it): even if there are flaws in the research, we should encourage junior people to be thoughtful, to try and analyze their experiences and find ways to ask and answer questions.

That is a long introduction to a new publication from the NICHD database.

And another beef, is it really necessary, when someone does an analysis of the NICHD database, that there are 19 authors?  Did all 19 have a substantial intellectual contribution to this work? Or is it more of a little present thrown to the waiting hordes, a sort of quid pro quo, you scratch my data, I’ll scratch your CV?

Just asking.

Kelleher J, et al, et al, et al. Prophylactic Indomethacin and Intestinal Perforation in Extremely Low Birth Weight Infants. Pediatrics. 2014. This prospective cohort comparison used the generic database to examine the incidence of SIP in babies who received prophylactic indomethacin or  not, and subdivided each group as to whether they had early feeding or not. Early feeding was defined as receiving any feeds in the first 3 days of life. My European readers might laugh at that as a definition of early feeding, but there were plenty of these babies, 400 to 1000 g birth weight, who were npo for at least 3 days, indeed there were 11,000 of the 15,000 who got no enteral feeding during the first 3 days (between 1999 and 2010).

As for the results, there was no evidence that prophylactic indomethacin increased SIP, and early feeding is associated with a lower rate (there may be some confounding here, as babies who have abdominal signs and maybe early signs of SIP might have feeds held as a result). Babies who had early feeding had less developmental delay at 20 months, (or, as the NICHD persists in calling it, ‘neurodevelopmental impairment': if I have to say it again, a low Bayley score is not an impairment!) and the association of early feeding with better Bayley scores, was seen whether or not the babies got prophylactic indomethacin.

The prophylactic indomethacin babies also did not have more NEC, and they had fewer PDA ligations. They also had much less frequent treatment with indomethacin or ibuprofen later in their hospital course, but the RR is reported as being over 1, and significant, which I don’t understand.

The early enteral feeds groups also had much quicker advancement to full feeds, and many fewer days of parenteral nutrition.

This confirms the RCT results, with a lot more babies but an observational study design, indomethacin doesn’t seem to increase SIP (despite our case-control findings) decreases PDA ligations, and adds very positive data about early feeds, at least starting feeds before 3 days of age. I think you should have a very good reason for not feeding a very preterm baby, such as shock requiring inotropes perhaps.

Posted in Neonatal Research | Tagged , | Leave a comment

Effects of (not very) NICE guidelines

Call me prescient, OK, you won’t, but I will. Two recent observational studies suggest that the recent NICE guidelines have had adverse effects on infants evaluated for potential early neonatal sepsis. ‘NICE’ of course is not an adjective for how good the guidelines are, but the acronym for the National Institute of Health and Care Excellence in the UK.

Any long time neonatalresearch watchers will remember vividly my perceptive analysis and critique of those guidelines. Which included the good (to stop antibiotics at 36 hours if no signs of sepsis) and the questionable; universal measurement of CRP at the start, with a repeat at 18 to 24 hours, and stopping the antibiotics if the trend is ‘reassuring’. My concern being that CRPs are relatively sensitive but with low specificity. Infants exposed to many different stressors, infectious and not, may have increased CRP. As early onset sepsis in term infants is relatively uncommon in most of the at-risk situations outlined, the proportion  of babies who have unnecessary prolongation of antibiotic therapy solely for a CRP which is not ‘reassuring’ might be substantial.

You have, of course, to balance that against the theoretical benefit of continuing antibiotics in an infant who is truly septic, in whom the antibiotics would have been stopped were it not for the non-reassuring CRP. A benefit which is likely to occur much less often.

Two recent studies have sought to quantify these impacts.

Mukherjee A, Davidson L, Anguvaa L, Duffy DA, Kennea N. NICE neonatal early onset sepsis guidance: greater consistency, but more investigations, and greater length of stay. Archives of Disease in Childhood – Fetal and Neonatal Edition. 2014.

This is a before-after study which investigated the impacts on babies evaluated for possible sepsis.

I think this is from a single hospital in London (it is a short report so some details are missing, also weirdly there is not a single reference, not even to the NICE guidelines), and it appears that about 8-9% of babies had sepsis work-ups, with around 70 babies affected in each of the 2 month cohorts, before and after the NICE guidelines.

They looked at how many babies stayed in the hospital less than 72 hours, this decreased from 38.1% to 18.4%. More babies stayed over 5 days, from 20.9% up to 27.7%. They found that 58% of the repeat CRPs were used to change management, including leading to more LPs (14% up to 23%). In all of their babies there were no positive cultures.

We envisaged shorter hospital stays with new NICE standards, particularly, with the aim of 36 h blood culture reporting. However, repeat CRP led to further investigations, increased LPs and longer durations of treatment and stay. This, in turn, impacted on workload and cost, and influenced parental experience in the first few days of life

Naydeva-Grigorova T, Manzoor A, Ahmed M. Management of early-onset neonatal infections. Archives of Disease in Childhood – Fetal and Neonatal Edition. 2014.

This experience, published as a letter in Archives, recounts what happened with about 40 babies, before and after the guidelines. 1 of 40 before the guidelines had an LP, compared to 8 afterward. Entirely because the CRP was raised. Hospital stay increased after the guidelines, and duration of antibiotic use increased. None of the babies had positive cultures.

Our results indicate that babies in group B had prolonged hospital stay requiring longer duration of intravenous antibiotics without much clinical benefit.

I think there should be a rapid re-evaluation of these guidelines, which seem to be only having negative effects, at least from these 2 experiences. I think the reason that the NICE guidelines are not very nice is that the data being used to support the recommendation are based largely upon a single study from 1998, a good study from Bill Benitz in California, unfortunately (for the NICE guidelines) that study was not among asymptomatic babies, but only babies who had symptoms, the list of which includes shock, new apnea, lethargy, respiratory distress and so on, and also includes preterm infants down to 550 grams birth weight. It only included babies treated for sepsis in one of 3 NICUs. (See evidence table 10.3 in the guideline that you can find here). So it is of no relevance to the healthy full term baby with risk factors alone, in whom the false positives are clearly going to be much more frequent.

In the more recent study by Thierry Lacaze and colleagues, which only included asymptomatic infants being evaluated because of signs of sepsis, a single CRP at 18 hours of age had a PPV of 14% for proven or probable sepsis. I discussed the article when it first came out, and I think the suggestion of those authors, that a low CRP at 18 hours of age, using a method which gave a result immediately to the medical team, could lead to earlier stopping of antibiotics and earlier discharge is reasonable. The implication being that a non-reassuring CRP would then mean waiting until 36 hours and making a decision independent of the CRP.

For asymptomatic babies CRPs are too sensitive, being elevated for all sorts of reasons unrelated to sepsis requiring treatment.



Posted in Neonatal Research | Tagged | Leave a comment