Relationship between enteral nutrition timing and 28-day mortality in critically ill stroke patients in the MIMIC-IV database.

BACKGROUND

The ideal timing for commencing enteral nutrition (EN) in critically ill stroke patients in the intensive care unit (ICU) remains a subject of debate, with ongoing controversy regarding the impact of early EN (EEN) initiation. In this study, we investigated the association between the timing of EN initiation and 28-day mortality using data from the MIMIC-IV database.

METHODS

This study employed a retrospective cohort design using the MIMIC-IV database to identify stroke patients who received EN during their hospital stay. The main focus of this investigation was to examine 28-day mortality among these patients following hospital admission. Various demographic, clinical, laboratory, and intervention variables were considered as covariates. The Cox regression analysis was employed to assess the correlation between the timing of EN initiation and 28-day mortality, and restricted cubic splines (RCS) analysis was used to test for non-linear correlation. Patients were then stratified into two cohorts depending on the timing of EN initiation: within 2 days ( = 564) and beyond 2 days ( = 433). A multivariate Cox regression analysis was used to investigate the difference in 28-day mortality between the groups.

RESULTS

A total of 997 participants were included in this study, with 318 (31.9%) dying within 28 days. We observed that the timing of EN initiation correlated with 28-day mortality, but this correlation was not significant after adjusting for covariates (crude HR: 0.94, 95% CI: 0.88-1, = 0.044; adjusted HR: 0.96, 95% CI: 0.9-1.02, = 0.178). The RCS analysis showed that the correlation was not non-linear. Notably, in the multivariate regression models, early EN initiation was associated with a higher mortality rate compared to late EN initiation [odds ratio (OR) = 1.34, 95% CI: 1.06-1.67, = 0.012]. After adjusting for various confounding factors in the multivariate Cox regression models, we identified that patients in the early EN group had a 28% higher risk of mortality than those in the reference group (OR = 1.27, 95% CI: 1-1.61, = 0.048). These associations remained consistent across various patient characteristics, as revealed through stratified analyses.

CONCLUSIONS

Early commencement of EN in critically ill stroke patients may be linked to a higher risk of 28-day mortality, highlighting the need for further investigation and a more nuanced consideration of the optimal timing for commencing EN in this patient population.