METHODS: Using data from a multicenter, cluster-randomized controlled trial, this secondary analysis involved patients with baseline circulatory shock as defined by a cardiovascular Sequential Organ Failure Assessment score of two or more. Patients were dichotomized into transient or persistent circulatory shock depending on the duration, while transient circulatory shock was defined by the resolution of shock within the first day of enrollment. Early enteral nutrition was defined as the initiation of enteral nutrition within 48 h after intensive care unit admission. The association between early enteral nutrition and a composite outcome (presence of any organ failure on study day 10 or 28-day mortality) was investigated by multivariable and propensity-score-weighted multivariable logistic regression analyses.

RESULTS: Seven hundred and eighty-five patients were included in the analysis, and early enteral nutrition was administered to 385 patients (49.0 %) in the whole study cohort. In patients with transient circulatory shock (n = 527), 221 patients (41.9 %) received early enteral nutrition, and in those with persistent circulatory shock (n = 258), 164 patients (63.6 %) did so. For the overall cohort, there was no difference in the incidence of primary composite outcome between early enteral nutrition and 'no early enteral nutrition ' groups (adjusted odd ratio 0.84, 95 % confidence interval 0.60-1.18) after adjustment for potential confounders. In patients with transient circulatory shock, receipt of early enteral nutrition, compared to no early enteral nutrition, was significantly associated with reduced incidence of the composite outcome (adjusted odd ratio 0.63, 95 % confidence interval 0.41-0.95, p = 0.027). On the contrary, there is no association between early enteral nutrition and the incidence of the composite outcome in patients with persistent circulatory shock (adjusted odd ratio 1.28, 95 % confidence interval 0.64-2.58, p = 0.485). The results of propensity-weighted multivariable analysis conform to the primary analysis.