Effects of early enteral nutrition on persistent inflammation, immunosuppression, and catabolism syndrome in critically ill patients: A claims database study using a propensity score analysis
Shizuka Kashiwagi,Naoki Kanda,Minoru Yoshida,Yuji Wakimoto,Hiroyuki Ohbe,Kensuke Nakamura
DOI: https://doi.org/10.1016/j.clnu.2024.06.033
Abstract:Background & aims: Early enteral nutrition (EEN) potentially improves immune-related outcomes via the maintenance of intestinal immunity; however, the effects of EEN on clinical outcomes, including infectious complications, are controversial. Therefore, we herein investigated whether EEN affected persistent inflammation, immunosuppression, and catabolism syndrome (PICS), which represents the immunocompromised state after critical illness. Methods: This retrospective cohort study utilized the administrative claims database of inpatients and laboratory findings. Patients admitted to and treated in the intensive care unit (ICU) for more than 3 consecutive days were included. The primary outcome, a composite of PICS or mortality on day 14 after admission, was compared between the EEN group, which received enteral nutrition (EN) on the first 3 days (day 0, 1, or 2), and the late enteral nutrition (LEN) group, which did not receive EN on the first 3 days, but then received EN on days 3 through 7, using a propensity score-matched analysis. Secondary outcomes included the composite outcome on day 28, in-hospital mortality, the Barthel index, and laboratory data. Patients who met at least two of the following conditions were diagnosed with PICS: CRP >2.0 mg/dL, albumin <3.0 g/dL, and a lymphocyte count <800/μL. Results: A total of 7530 matched pairs were generated after propensity score matching. The primary outcome was significantly lower in the EEN group (risk difference -3.0%, 95% confidence interval (CI) -4.5 to -1.4%), whereas mortality did not significantly differ. The 28-day composite outcome was similar in the 2 groups (risk difference -1.5%, 95% CI -2.8% to -0.2%, no significant difference in mortality). There was no significant difference in in-hospital mortality between the EEN and LEN groups; however, the Barthel index at discharge was higher in the EEN group (the medians, 50 vs 45, P = 0.001). Laboratory data showed lower Albumin and CRP on day 14 in the EEN group, but no other significant differences. Conclusions: In patients admitted to the ICU, EEN was associated with a lower incidence of PICS on days 14 and 28, but was not associated with mortality. This positive association was not observed in sepsis, cardiac diseases, or gastrointestinal diseases.