Gastrointestinal Tolerance in very preterm babies

Editor(s): Virgilio Carnielli.

Gastrointestinal Tolerance in very preterm babies

Virgilio Carnielli

Feeding tolerance in preterm babies < 32 weeks is getting more and more a topic because the survival rates are getting better. However, many of these babies have growth failure and end up being too small to be 36 weeks old, the time they are usually discharged.

Even perhaps more importantly, over the last 20 years these babies did not get any neurological advantage. In addition these babies develop a nutrition deficit. Nick Embleton et al. concluded that preterm babies accumulate a significant nutrient deficit in the first few weeks of life that will not be replaced when current recommended dietary intakes are fed (Peds, 2001). This deficit can be directly related to subsequent postnatal growth retardation. Poor growth velocity and neurodevelopmental
outcomes had been brought forward by Richard Ehrenkranz et al. (Pediatrics, 2006). They studied a cohort of 600 babies (500 g–1.000 g) from the US and showed marked correlation between slow in-hospital growth and poor neurological outcome. Same findings were also reported in the UK. These are cohort studies, but only few randomized clinical trials are available. You have to be really careful when interpreting this type of data. So we became interested in seeing what happened in our unit, what happened in different places.

Today we have wonderful guidelines, however
are we able to apply these guidelines to our pa- tients? Some of these babies are less critically ill, some are more critically ill. We developed a statistically valid software and we are able to collect what is being the target, the prescription and the administration  of nutrients for each parenteral and enteral nutrition. With that nutrition dash- board you get all the calculations automatically on the bedside. You can see what is the target value, how far are you from target, you can even see the metabolic complications and the association between medication and illnesses. Next to it we have the growth data. Every patient’s growth rate is plotted  and we can monitor weight gain, head circumference gains and so on.

One of our major interests is, if the outcome of
the baby is nutrition-dependent or disease-de- pendent,  and whether there is an interaction between these two issues. In our Ancona co- hort there are nearly 1,500 babies from January 2004 up to March 2019. 1,129 babies < 1,250 g are
available for evaluation. They get routine parenteral nutrition and slower enteral nutrition progression. If they are > 1,250 g they get no routine parenteral nutrition and a faster progression enteral nutrition. That is our feeding scheme (Fig. 1) according to the day of life and according to birth weight segments.

Feeding intolerance and low enteral nutrition
There are quite different ways to define feeding intolerance. According to our definition of low enteral nutrition (EN): EN volume < 85 % of the target scheme over a given period time.

We did a concordance study and for instance we investigated if those being less than 85 % of the target value had a good concordance with the days reaching 120 mL/kg/d.

We compared five different methods to measure intolerance on the very same population and there is no difference no matter what method you use. The next step was to define critical illness. Sometimes you have a simple score especially given in the first days of life, but we came up with something a little more complex, where we took in account medications, signs, symptoms, and diagnoses. Every one of our patients gets this type of scheme with a score number for each individual day from day 1 to day 28. The sum of the score numbers tells you an approximation for severity and you can come up with the grade of sickness during the first 28 days of life or daily
sickness of a given time period, or  you can even do a cumulative calculation of points. We arbitrarily choose 15 points to separate the less critically infants and the more critically ones (>15 points). Comparing infants tolerating less than 15% of the target EN with control infants who tolerate the prescribed target value, profound difference  in fluids and energy intakes was found. What happened to low enteral nutrition tolerance <85% of the target? The more critically ill infants occurred in 70% of the cases, less critically ill infants in 30% and vice versa in control babies (28% and 72%, respectively). If you are sick, tolerance is less of course.

There are some risk factors during pregnancy and the intolerant babies, in spite of the contribution of parenteral nutrition, at  36 weeks are smaller, grow less. However, when these babies reached 2 years of age body size was identical. But not only body size is identical, the Bayley scores are 99 ver- sus 97 (cognitive), 103 versus 103 (motor) in the Ancona cohort.
We did not find that postnatal growth retardati- on over milder degree was associated with long term deficit at 2 years, neither in terms of an- thropometry nor in terms of neurodevelopment.

Case control group
In a well-matched case-control study we could select 88 cases of low enteral intake compared to standard enteral intake. Those infants on  low EN were much smaller at 36 weeks, showing lower body size, lower weight gain, lower growth velocity, significantly lower enteral intakes, somewhat higher parenteral intakes and lower total energy and protein (parenteral plus enteral intakes). What happened to these case-control babies? At 2 years identical body size and neuro- development were eventually found.

Some papers showed that human milk is better tolerated, so we compared human milk versus formula and the difference was trivial. In our unit, to our surprise, the tolerance of the human milk fed babies was similar to formula fed babies.

More critically ill infants
Is there any difference in terms of tolerance and of nutrients if the baby is much sicker? We have 276 of these infants and so we could match 138 babies who got lower tolerance, <85% target EN, identical clinical characteristics, and quite identical severity. Again you get less enteral energy, you grow less and at 36 weeks all the data are in terms of SD score. They are all smaller, weight gain during hospital stay was slower and the energy was significantly less in the low EN, as expected. But even in this group of more critically infant, tolerance of human milk and infant formula was identical. We could not find a difference in intolerance. (Fig. 2) At 2 years even in this more critically infants, 57 per group – because of drop out and  our neuro- development assessment started later than 2004 – no difference was found.

Even in critically ill infant receiving < 85% of the EN target we could not find any difference in the cognitive and motor Bayley scores. That is different from what was published before.

We divided babies who tolerated < 85% EN in- takes in less critically and more critically ill and the less critically ill did better, as you would expect. So the effect of illnesses was found to be the same in the infants who were intolerant and in the control EN. Also the less critically ill babies had higher neurodevelopmental scores than the more critically ill.

Multiple regression analysis

In a cohort of 1129 babies we tested all possible variables, with all possible tests focused mainly on the prediction of low tolerance for the babies
who got < 85% of the EN target volume during their first 28 days of life. The risk factors were not related to the type of milk or the clinical severity. We also looked at the cognitive development at two years, testing all possible variables and what came out was that cognition at 2 years was associated with PVL, BPD, SGA and maternal schooling. Using our severity score instead of the selected condition, the severity score
was highly significant.

Although the Ancona Cohort has nearly the same population and we used exactly the same criteria, we did not find the inverse correlation between in-hospital growth and indices of poor development reported in the US.

Very important: HAVING “OPTIMIZED” NUTRITION of the low birth weight infants!

Feeding intolerance in preterm infants of <32 weeks gestation is chiefly associated with sickness.

Infants who received less than 15% of the ANCONA target enteral nutrition volume were smaller at 36 weeks PMA (discharge) however they caught up at 2 years (body size and neurodevelopment comparable to control infants).

The use of Parenteral Nutrition was larger in Low-EN infants. Enteral tolerance of infant milk formula was similar to own mother’s milk.

Neurodevelopment in Ancona cohort was mainly affected by disease sever- ity and parental education.