Background: Hyponatremia, a frequent yet potentially life-threatening electrolyte 
imbalance, poses heightened risks in intensive care contexts. This investigation sought to 
explore contributory factors linked to hyponatremia following linezolid administration in 
critically ill (CI) individuals and to formulate a robust predictive framework. 
Methods: A retrospective evaluation was conducted on clinical records and follow-up data 
from 200 CI patients who received linezolid therapy. To isolate key determinants, logistic 
regression modeling was utilized, followed by validation using Receiver Operating 
Characteristic (ROC) curve analysis. A nomogram-based risk assessment tool was then 
constructed, with calibration tested via the Hosmer-Lemeshow goodness-of-fit approach. 
Findings: Adverse reactions were recorded in 23.5% of the cohort. Statistically significant 
disparities (P < 0.05) emerged between CI and non-CI patients across several variables, 
including linezolid serum levels, therapy duration (DOM), baseline sodium values (BSS), 
estimated glomerular filtration rate (eGFR), white blood cell (WBC) count, total bilirubin 
(TBIL), albumin (ALB), and key biomarkers (NGAL, suPAR, Cystatin C), as well as concurrent 
spironolactone usage. The Z-score presented the highest diagnostic efficacy for 
hyponatremia, with a threshold of -3.24. The model demonstrated an 85.5% predictive 
accuracy, and the nomogram—based on multivariate regression and fit assessment—
 exhibited excellent alignment with actual outcomes. 
Interpretation: Independent predictors of hyponatremia included DOM, drug 
concentration, BSS, eGFR, and TBIL. Incorporation of novel biomarker profiles modestly 
improved model precision, suggesting added value in patient risk stratification. The 
developed tool offers promise for early detection and intervention in vulnerable ICU 
populations.