Association between enteral nutrition timing and postoperative anion gap trajectory in cardiac surgery ICU patients: a retrospective study based on the MIMIC-IV database.
Source: PubMed, NCBI / U.S. National Library of Medicine
To investigate the association between enteral nutrition (EN) timing and postoperative anion gap (AG) trajectory in ICU patients undergoing cardiac surgery. Data were extracted from the MIMIC-IV database. Cardiac surgery patients receiving EN were classified into preoperative and postoperative groups according to whether EN was first recorded before or after surgery. Baseline comparability was assessed. AG levels at admission, within 24 h after surgery, and at discharge were compared, and multivariable linear regression and subgroup analyses were performed as exploratory analyses. Among 195 cardiac surgery patients, 83 had preoperative EN and 112 had postoperative EN records. Baseline illness severity was broadly similar between groups, although age, sex, alcohol use history, and diabetes prevalence differed. No significant between-group difference in admission AG was observed, whereas the preoperative EN group had lower AG at postoperative 24 h (15 [12, 17] vs. 17 [14, 19], P = 0.035) and at discharge (15 [13, 17] vs. 17 [13, 21], P = 0.027). EN timing remained associated with postoperative 24-hour and discharge AG in multivariable models. EN timing was not clearly associated with postoperative delirium, in-hospital mortality, or length of hospital stay. It was discharge-AG that positively correlated with in-hospital mortality. Preoperative EN was associated with lower postoperative and discharge AG levels in this retrospective cohort. H
Abstract
To investigate the association between enteral nutrition (EN) timing and postoperative anion gap (AG) trajectory in ICU patients undergoing cardiac surgery. Data were extracted from the MIMIC-IV database. Cardiac surgery patients receiving EN were classified into preoperative and postoperative groups according to whether EN was first recorded before or after surgery. Baseline comparability was assessed. AG levels at admission, within 24 h after surgery, and at discharge were compared, and multivariable linear regression and subgroup analyses were performed as exploratory analyses. Among 195 cardiac surgery patients, 83 had preoperative EN and 112 had postoperative EN records. Baseline illness severity was broadly similar between groups, although age, sex, alcohol use history, and diabetes prevalence differed. No significant between-group difference in admission AG was observed, whereas the preoperative EN group had lower AG at postoperative 24 h (15 [12, 17] vs. 17 [14, 19], P = 0.035) and at discharge (15 [13, 17] vs. 17 [13, 21], P = 0.027). EN timing remained associated with postoperative 24-hour and discharge AG in multivariable models. EN timing was not clearly associated with postoperative delirium, in-hospital mortality, or length of hospital stay. It was discharge-AG that positively correlated with in-hospital mortality. Preoperative EN was associated with lower postoperative and discharge AG levels in this retrospective cohort. However, this metabolic signal did not translate into clear clinical outcome benefit, and the findings should be interpreted as hypothesis-generating because residual confounding by indication remains possible.