ABSTRACT: INTRODUCTION: The EPaNIC randomized controlled multicentre trial showed that postponing initiation of parenteral nutrition (PN) in ICU-patients to beyond the first week (Late-PN) enhanced recovery, as compared with Early-PN. This was mediated by fewer infections, accelerated recovery from organ failure and reduced duration of hospitalization. Now, the trial's preplanned cost analysis (N = 4640) from the Belgian healthcare payers' perspective is reported. METHODS: Cost data were retrieved from individual patient invoices. Undiscounted total healthcare costs were calculated for the index hospital stay. A cost tree based on acquisition of new infections and on prolonged length-of-stay was constructed. Contribution of 8 cost categories to total hospitalization costs was analyzed. The origin of drug costs was clarified in detail through the Anatomical Therapeutic Chemical (ATC) classification system. The potential impact of Early-PN on total hospitalization costs in other healthcare systems was explored in a sensitivity analysis. RESULTS: ICU-patients developing new infection (24.4%) were responsible for 42.7% of total costs, while ICU-patients staying beyond one week (24.3%) accounted for 43.3% of total costs. Pharmacy-related costs represented 30% of total hospitalization costs and were increased by Early-PN (+608.00 EUR/patient, p = 0.01). Notably, costs for ATC-J (anti-infective agents) (+227.00 EUR/patient, p = 0.02) and ATC-B (comprising PN) (+220.00 EUR/patient, p = 0.006) drugs were increased by Early-PN. Sensitivity analysis revealed a mean total cost increase of 1,210.00 EUR/patient (p = 0.02) by Early-PN, when incorporating the full PN costs. CONCLUSIONS: The increased costs by Early-PN were mainly pharmacy-related and explained by higher expenditures for PN and anti-infective agents. The use of Early-PN in critically ill patients can thus not be recommended for both clinical (no benefit) and cost-related reasons. TRIAL REGISTRATION: ClinicalTrials.gov NCT00512122.