- In the current era, hospital acquired infections are relatively uncommon early during a pediatric ICU admission. In the study patients (overall and in each treatment group), could you provide summary statistics (e.g., median (IQR)) of the time from ICU admission to development of new infection? Please comment on the temporal relationship between timing of exposure to parenteral nutrition (early or late) and development of new infections. Were there any differences in infection rates and time to infection among centers?
REPLY (Dr. Van Den Berghe): We are going to publish the time to new infections in a separate paper, so I cannot give you these data yet, so sorry. There were no center differences what so ever regarding the impact of late versus early parenteral nutrition, neither on infection rates nor on the time to infection. The findings were very robustly center-independent and thus generalizable.
- Approximately 50% of study patients were discharged from the ICU by day #4. In retrospect, do you feel that such patients had disease severity that may have benefited from parenteral nutrition?
REPLY: With hindsight, none benefited from parenteral nutrition. For short-stayers and long-stayers alike, early parenteral nutrition did not provide benefit and instead caused harm.
- Please expand on the types of neonates that were included in the study. For example, were pre-operative cardiac neonates receiving a prostaglandin infusion enrolled in the trial? If so, if clinicians felt that such patients could not be fed enterally, were they randomized or excluded from the study (i.e., were they part of the 73 patients excluded in the category “had short-bowel syndrome or other condition requiring parenteral nutrition”?)
REPLY: Again, it is our plan to analyze the neonatal subgroup in more detail and publish these data. As for diagnoses, I can inform you that the major subgroups in this neonatal subgroup were as follows: 35% were newborns admitted after cardiac surgery for major congenital anomalies, about 7% were “medical cardiac neonates” (so not after surgery, some could be those you refer to but I do not have that detailed information readily available to give you exact numbers) and 31% were newborns admitted after abdominal surgery for major gastrointestinal congenital anomalies such as congenital diaphragmatic hernia etc.. Whenever clinicians felt kids could not be fed enterally, they were still included in the trial. Only those specified in the consort diagram were excluded.
- Was parenteral nutrition continued past ICU discharge in a subset of study patients? If so, are there data that patients assigned to receive early parenteral nutrition had more new infections after ICU discharge but prior to hospital discharge?
REPLY: The continuation of parenteral nutrition beyond PICU discharge was at the discretion of the pediatricians on the regular wards. We did not assess the number of infections after PICU discharge. In an ongoing health economy analysis, however, we will assess the use of antibiotics in that specific time window separately. We are aiming to report these data in a separate publication.
- Lack of preservation of autophagy and overfeeding were mentioned in the manuscript and accompanying editorial as potential causal mechanisms for the association between early parenteral nutrition and worse clinical outcomes. Please expand on these potential causal factors and/or discuss alternative etiologies.
REPLY: We have good reason to believe from our more basic research that autophagy is essential for clearing damaged organelles such as mitochondria, misfolded or unfolded proteins and protein aggregates in the context of critical illness. In addition, autophagy is known to contribute to removing intracellular microorganisms as well as engulfed microorganisms within macrophages. We think that providing macronutrients early, with the amino-acids as major suspects given the results from the EPaNIC trial (Casaer M et al. Am J Resp Crit Care Med 2013; Hermans G. et al. Lancet Resp Med 2013) and those from our earlier animal studies (Derde S. et al. Endocrinology 2012), suppresses this evolutionary conserved quality control system in cells. We are further testing this hypothesis for the PEPaNIC trial. Of course one can speculate about other potential mechanisms that are activated by “lack of macronutrients” (thus other fasting-induced pathways, which we are currently further investigating) but the published data so far do point to autophagy.
- A footnote in Supplementary Table 7 states that “… receiving a higher amount of enterally administered kcal per kg per day were independent risk factors for infections and for delayed live discharge from PICU.” Please comment on this issue. Should clinicians infer that in addition to the adverse effects of early parenteral nutrition (the key finding of the trial), the receipt of more kcal/kg/day of enteral nutrition during the early portion of an ICU admission increases the risk for new infection and greater time to live discharge from PICU?
REPLY: We were asked to do this additional analysis by NEJM reviewers and editors. The data are the data… If this result is not biased (which we cannot be sure of as we did not randomize for enteral nutrition) then that is indeed the interpretation: even forceful enteral feeding could be harmful… Evolutionary it makes sense to couple effects of “fasting” (e.g. autophagy activation among others) during the early phase of illness (in prehistorical times, illness would preclude going after food …) to benefits for the host in the struggle for survival (bacterial killing and cell damage removal to recycle substrates).
- Have the study findings changed practice regarding the timing and use of parenteral nutrition in the ICUs at the participating centers? If so, how?
REPLY: Yes. We all changed practice to the “late parenteral nutrition” strategy, which means that we no longer start parenteral nutrition prior to day 8 in PICU. But we do continue to give micronutrients early to all patients, as this is how we did the study, in order to prevent refeeding syndrome upon initiation of parenteral nutrition whenever necessary beyond day 8. Also, although different blood glucose targets are used per center, the “late parenteral nutrition” is still being combined with frequent blood glucose monitoring and some form of blood glucose control (looser in the other centers than in Leuven) from admission to PICU onwards.