Alterations in the lipid profile of critically ill children in relation to outcome

Abstract

Critically ill adults typically develop hypocholesterolemia, associated with poor outcome. Whether similar alterations occur in critically ill children is less clear. In secondary analyses of the PEPaNIC RCT (n = 1440), we first documented the time course of plasma cholesterol and triglyceride concentrations, and the effect of randomization to early-parenteral-nutrition (early-PN) or late-PN hereon, for 96 matched critically ill children staying ≥ 5 days in PICU. Second, for 1165 children with available admission plasma samples, lipid profiles were determined and their independent associations with outcome (time to live PICU discharge, new infection and 90-day mortality) were assessed with Multivariable Cox proportional hazard and logistic regression, adjusting for baseline risk factors. Plasma HDL-cholesterol, LDL-cholesterol, total-cholesterol and triglycerides were low throughout the 5 PICU days, with only HDL-cholesterol further decreasing over time (P < 0.0001) and without effect of randomization to early-PN or late-PN, and with admission values lower in infants than older children and in patients with infection (P < 0.05). Lower admission HDL- and total-cholesterol concentrations were independently associated with a lower likelihood of an earlier live PICU discharge (P < 0.001) and with a higher risk of 90-day mortality (P ≤ 0.01), whereas higher plasma triglycerides were independently associated with higher risk of 90-day mortality (P = 0.004). Low admission plasma HDL-cholesterol was independently associated with a higher risk of acquiring a new infection (P = 0.05). Critically ill children presented with low circulating levels of lipids. Low plasma cholesterol concentrations were associated with poor outcomes, most robustly for HDL-cholesterol. Whether these associations are causal or casual requires further investigation.