Many things that we put in the intestines of preterm infants increase the risk of necrotising enterocolitis, including xanthan gum, kayexalate, and now, it seems, gastrografin.

A group from Vienna performed a masked RCT in 96 very low birth weight, preterm infants who were administered either 3 mL/kg of a hypertonic contrast agent (gastrografin) diluted 1:3 via the NG tube in the first 24 hours of life which they say gave 9 mL/kg of liquid, or the controls who got 9mL/kg of water. (Now I am not a mathematician, but I would have thought that 3 mL of gastrografin diluted 1:3 should give you 12 mL of liquid, not 9; I guess the authors and the reviewers aren’t mathematicians either!) Haiden N, Norooz F, Klebermass-Schrehof K, Horak AS, Jilma B, Berger A, Repa A: The effect of an osmotic contrast agent on complete meconium evacuation in preterm infants. Pediatrics 2012, 130(6):e1600-1606.

The idea behind this is that delayed meconium passage is associated with bad outcomes (delayed feeding intolerance, and in at least 1 study with NEC) and that therefore if you can speed up meconium clearance, maybe you can reduce those bad outcomes. The primary outcome of the study was time to clearance of meconium, which is rather questionable as an outcome of clinical significance. They also recorded other secondary outcomes, feeding tolerance, duration of hospital stay and NEC. The first thing to note is that they enrolled 96 babies out of 789 eligible infants, this immediately makes you wonder how representative the babies in the study were. The second thing to note is that the ‘intention to treat’ analysis showed no difference in anything. The only differences were in the ‘per protocol’ analysis (when they re-analyzed the data after excluding 18 infants with protocol violations) which showed a much shorter NICU stay in the gastrografin group, and some evidence of achieving full feeds faster. Here again there are some weird findings, the duration of hospital stay was as low as 4 days. So some babies in Vienna who weigh less than 1500 g at birth and are less than 32 weeks can go home in 4 days!! I don’t know how they do that. They are also clearly very fixated on stools, many of the babies in both groups also got glycerine suppositories, and in addition the babies averaged more than 1 enema each before discharge.

But most importantly, there were 8 cases of NEC in the gastrografin group vs 3 in the controls, and there were 5 NEC deaths in the gastrografin group. I would suggest that this is not a good thing to do!

Also just because slow clearance of stools might be associated with other signs of poor intestinal function and maybe with NEC, does not mean that enhancing meconium passage will improve feeding tolerance or NEC. The same authors have already investigated giving dilute glycerine enemas to their babies (Haiden N, Jilma B, Gerhold B, Klebermass K, Prusa AR, Kuhle S, Rohrmeister K, Kohlhauser-Vollmuth C, Pollak A: Small volume enemas do not accelerate meconium evacuation in very low birth weight infants. Journal of Pediatric Gastroenterology & Nutrition 2007, 44(2):270-273.), they gave the enemas every day until all the meconium was gone, and in the controls they left them alone. They showed no benefit of the intervention, even when they analyzed as ‘per protocol’ rather than intention to treat.

Which brings me back to the new study, why were the authors allowed to emphasize the non-ITT analysis in the abstract and to present this as a study showing decreased length of stay? In a small pilot study with a lot of exclusions, I am not averse to presenting a ‘per protocol’ analysis as a secondary suggestive analysis that might warrant a future study with more strict enforcement of a protocol. But the deviations from the protocol might well be a result of the intervention (indeed in this trial the exclusion was commonly because of vomiting after the large volume of fluid) so it is essential that the primary analysis which is presented is the intention to treat, only that way can a true estimate of the use of an intervention in the real world be estimated.

Finally a word about NEC prevention. I was surprised recently by an editorial in the Journal of Perinatology (Swanson JR: Necrotizing enterocolitis: Is it time for zero tolerance? J Perinatol 2013, 33(1):1-2) which made the following statement : ‘Experts have suggested that we could cut the incidence of NEC in half if neonatologists would just do four things: (1) practice as a group (instead of individuals), (2) promote breast-milk feeding and start it as trophic feeds early in life, (3) have a standardized feeding advance that is gestational age-based and (4) minimize the feeding intolerance episodes with liberal use of glycerin.’

A number of things bug me about this, first of all the ‘call to authority’ is not what we should be doing as scientists, secondly even if experts suggest this it had better be based on some evidence, which it clearly isn’t, and finally the reference given doesn’t even say these things! It NEVER mentions glycerin. (or indeed mentions working as a group rather than individuals, nor does it mention trophic feeds) Indeed the review article that the editorial references does refer to the data regarding having a feeding protocol and using more breast milk. It also neglects to mention the most effective intervention that we have for prevention of NEC: Probiotics.