Association between age and outcomes of evidence-based feeding protocol versus usual care in critically ill patients: secondary analysis of a multicenter cluster-randomised controlled trial

Why clinicians should know about it

• Picked for Critical Care and Intensive Care Medicine (paper of the day, 23 September 2026): Age‑based effect modification of feeding protocol in ICU patients

Abstract

BACKGROUND: Nutrition therapy is essential in critically ill patients, yet large randomised trials have shown inconsistent results regarding the effect of an evidence-based feeding protocol. The NEED trial showed a feeding protocol could effectively promote delivery of enteral nutrition (EN) but did not reduce mortality. Age-related physiological differences may modify responses to nutritional therapy. This post hoc analysis of the NEED trial evaluated whether age modifies the association between implementation of an evidence-based feeding protocol and clinical outcomes among critically ill patients. METHODS: This secondary analysis used data from the multicentre cluster-randomised NEED trial conducted across intensive care units (ICUs) in China. Adult patients newly admitted to participating ICUs were assigned to active implementation of a feeding protocol or usual care. Age was analyzed as both a continuous variable and predefined categories (≤65, 66-75, and ≥76 years). The primary effect-modification analysis evaluated the treatment-age interaction for 28-day all-cause mortality. Secondary outcome analyses, including organ failure (OF), persistent multiple OF, ICU length of stay, ICU-free days, infection within 7 days, and feeding intolerance, were exploratory. Mixed-effects Cox and logistic regression models were used to assess treatment effects and age-treatment interactions. RESULTS: Among 2672 patients included in the analysis (median age 62 years [IQR 49-74]; 1798 [67.3%] male), the effect of evidence-based feeding protocol on 28-day mortality varied with age. Treatment benefit attenuated progressively with increasing age, with each 10-year increase associated with a higher adjusted hazard ratio (HR) for mortality (1.28; 95% CI 1.13-1.46; continuous interaction P < 0.001). In age-stratified analyses, the intervention reduced mortality in patients aged ≤65 years (adjusted HR 0.64; 95% CI 0.44-0.92) but not in those aged 66-75 years (adjusted HR 1.01; 95% CI 0.64-1.60) or ≥76 years (adjusted HR 1.29; 95% CI 0.89-1.88; categorical interaction P = 0.010). An exploratory analysis showed a similar age-dependent pattern for persistent multiple OF (interaction P = 0.017). Results were consistent in per-protocol and sensitivity analyses. CONCLUSIONS: In this secondary analysis, associations between an evidence-based feeding protocol and clinical outcomes differed across age groups. These exploratory and hypothesis-generating findings suggest potential age-related heterogeneity in treatment response, which requires confirmation in future prospective studies. Research registration unique identifying number (UIN) of the original NEED trial: ISRCTN Registry, Registry number: ISRCTN 12233792, registered on November 24, 2017.

Abstract as published, via PubMed.