Role of age of critically ill children at time of exposure to early or late parenteral nutrition in determining the impact hereof on long-term neurocognitive development: A secondary analysis of the PEPaNIC-RCT
Ines Verlinden, Karolijn Dulfer, Ilse Vanhorebeek, Fabián Güiza, José Hordijk, Pieter Wouters, Gonzalo Garcia Guerra, Koen Joosten, Sascha Verbruggen, Greet Van den Berghe
Abstract
BACKGROUND & AIMS: Early use of parenteral nutrition (early-PN), as compared with withholding it for one week (late-PN), in the PICU, has shown to slow down recovery from critical illness and impair long-term development of 6 neurocognitive/behavioural/emotional functions assessed 2 years later. Given that key steps in brain maturation occur at different times during childhood, we hypothesised that age at time of exposure determines long-term developmental impact of early-PN. METHODS: The 786 children who were neurocognitively tested 2 years after participation in the PEPaNIC-RCT were included in this study. First, for each studied long-term outcome, interaction between randomisation to early-PN versus late-PN and age was assessed with multivariable linear regression analysis. Subsequently, for outcomes with an interaction p ≤ 0.15, the impact of early-PN versus late-PN was analysed, after adjustment for risk factors, for 4 subgroups defined based on developmentally-relevant age at time of exposure [≤28 days (n = 121), 29 days to 11 months (n = 239), 11 months to <5 years (n = 223) and ≥5 years (n = 203)]. RESULTS: Interaction between randomisation and age was present for weight, and parent-reported inhibitory control, cognitive flexibility, working memory, planning/organisation, metacognition, total executive functioning, and internalising and total behavioural/emotional problems. Subgroup analyses revealed that none of the age-groups revealed benefit, whereas children aged 29 days to <11 months were most vulnerable to harm by early-PN for development of inhibitory control (p = 0.008), working memory (p = 0.009), planning/organisation (p = 0.004), metacognition (p = 0.008), and total executive functioning (p = 0.004), and for internalising (p = 0.005) and total behavioural/emotional problems (p = 0.01). Children aged 11 months to <5 years revealed harm by early-PN for development of inhibitory control (p = 0.003). In contrast, children aged ≥5 years and neonates aged ≤28 days appeared less vulnerable. CONCLUSIONS: Critically ill children aged 29 days to 11 months at time of exposure were identified as most vulnerable to developmental harm evoked by early-PN. CLINICAL TRIALS.GOV: NCT01536275.