NUTRITIONAL STATUS, ENERGY REQUIREMENTS AND METABOLIC MONITORING IN CRITICALLY ILL CHILDREN: THE NEW PERSPECTIVE OF ARTIFICIAL NEURAL NETWORKS

Nutrition plays a pivotal role in all humans. As early as during foetal life, a correct nutrition has the power to influence the development of organs and tissues, ultimately setting the basis for a healthy life. In critically ill children the risk of malnutrition is of particular importance. Accordingly, an appropriate monitoring of nutritional status and metabolic response, along with the correct assessment of energy requirements and energy balance, is gaining growing clinical relevance as a fundamental prognostic factor and should be considered a specific target in the management of critically ill children. The first step for a tailored nutritional support is the knowledge of patients’ resting energy expenditure (REE). Indirect calorimetry (IC) is the gold standard for REE measurement, however, its clinical use is limited across the world for both logistic and technical limitations. Alternatively, REE can be estimated using predictive equations, but this method has been found to be highly inaccurate in pediatric patients. Recent data pointed out that artificial neural networks (ANN) might represent a precise and accurate method to estimate REE in healthy and obese children. However, specific data regarding the applicability of the methodology on critically ill subjects are still missing. This thesis aimed to investigate the potential role of ANN on REE prediction for critically ill children by applying ANN to a dataset containing data on IC performed in our pediatric intensive care unit (PICU). We prospect that data derived from our observations could lead to a more accurate estimation of REE and to a better understanding of the energy requirements of critically ill children.