Related Experiment Videos
Development and validation of a nomogram for early nutritional risk stratification in elderly patients with pulmonary
Xiaoming Yang1, Meiyan He1, Zhongda Liu1
1Department of Respiratory and Critical Care Medicine, Lishui Hospital of Traditional Chinese Medicine Affiliated to Zhejiang University of Traditional Chinese Medicine, Lishui, China.
Background:
Malnutrition is highly prevalent among elderly patients with sepsis and is associated with adverse outcomes; however, early identification of nutritional risk remains challenging. This study aimed to develop and validate a practical prediction model for early nutritional risk in elderly patients with pulmonary sepsis.
Methods:
In this retrospective cohort study, 323 elderly patients with pulmonary sepsis were included. Nutritional risk was defined as a Nutritional Risk Screening 2002 (NRS-2002) score of ≥3. Candidate variables collected within 24 h of admission were screened using least absolute shrinkage and selection operator regression, followed by backward stepwise selection. A multivariable logistic regression model was constructed and visualized as a nomogram. Model performance was evaluated using discrimination, calibration, decision curve analysis, comparison with the modified Nutrition Risk in the Critically Ill (mNUTRIC) score, bootstrap internal validation, and sensitivity analysis using proxy Global Leadership Initiative on Malnutrition (GLIM) criteria.
Results:
Nutritional risk was identified in 115 patients (35.6%). Seven predictors, age, body mass index, Barthel Index, Sequential Organ Failure Assessment score, serum sodium, albumin, and hemoglobin-albumin-lymphocyte-platelet score, were retained in the final model. The nomogram demonstrated good discrimination (area under the curve [AUC] 0.803, 95% CI 0.755-0.852) and satisfactory calibration (Hosmer-Lemeshow P = 0.845). Decision curve analysis indicated a net clinical benefit across a wide range of threshold probabilities (0.15-0.70). Internal validation confirmed model stability (mean AUC, 0.797). The nomogram outperformed the mNUTRIC score in head-to-head comparison (AUC, 0.803 vs. 0.658; P < 0.001). Sensitivity analysis using proxy GLIM criteria yielded an AUC of 0.708 (95% CI, 0.650-0.766).
Conclusion:
A parsimonious model integrating routinely available clinical variables enables accurate early identification of nutritional risk in elderly patients with pulmonary sepsis. This nomogram may serve as a potential bedside risk-stratification tool, although its use in routine clinical practice requires prospective external validation.