Determinants of effectiveness in orthopedic ERAS: a cross-setting perspective based on the integration of nutritional...

This study examined factors related to the effectiveness of Enhanced Recovery After Surgery (ERAS) pathways in orthopedic patients by using data from both the National Health and Nutrition Examination Survey (NHANES) and the electronic Intensive Care Unit (eICU) database. By comparing a community-based population with critically ill hospitalized patients, we aimed to identify potential prognostic markers and setting-specific factors associated with recovery outcomes.

A retrospective, multi-database comparative analysis was conducted. Participants with a history of orthopedic surgery (n = 2,528) were selected from NHANES cycles (2011-2018) and categorized into an "ERAS-inclined group" and a "control group" based on post-operative management characteristics. Orthopedic critically ill patients (n = 54,678) were selected from the eICU database (v2.0) and stratified into "survivor" and "non-survivor" groups based on hospital outcome. Data analyses included propensity score matching (PSM), multivariable logistic regression, survival analysis, machine learning models (SHAP), and decision curve analysis (DCA).

In the eICU cohort, the non-survivor group exhibited higher age, greater illness severity scores (APACHE and SOFA), and lower serum albumin levels (2.49 vs. 2.92 g/dL). Multivariable analysis confirmed that mechanical ventilation was the strongest factor associated with in-hospital mortality, while higher serum albumin was associated with lower mortality risk. In the matched NHANES cohort, male sex (OR = 0.45) and higher BMI (OR = 0.97) were identified as protective factors for improved long-term rehabilitation outcomes, while widowed status was a risk factor. Cross-database comparison showed that serum albumin levels were highest in the NHANES community cohort, generally around 4.0 g/dL, and lower in the eICU critically ill cohort, generally around 2.5-3.0 g/dL. This pattern suggests that albumin may be a useful marker for reflecting nutritional status and illness severity across different clinical settings. The NHANES model showed better discrimination, whereas the eICU model demonstrated modest discrimination, suggesting that the identified predictors may be useful for hypothesis generation and risk stratification, but require further validation before clinical application.

The findings of this study indicate that the factors associated with rehabilitation in orthopedic patients may differ depending on the stage of recovery. In the acute critical phase, the need for mechanical ventilation may reflect more severe physiological stress, while hypoalbuminemia may suggest underlying nutritional depletion; both were associated with poorer clinical outcomes. In contrast, during long-term recovery, baseline BMI and marital status appeared to have predictive relevance, possibly reflecting metabolic reserve and social support. These results suggest that nutritional risk assessment and evaluation of social vulnerability should receive greater attention in orthopedic ERAS practice. However, as this study is mainly exploratory, the findings should be confirmed in future prospective research.