This study aimed to investigate risk factors of prolonged length of hospital stay (LOHS) in patients with acute exacerbations of chronic obstructive pulmonary disease (AECOPD) and to develop a predictive nomogram.

AECOPD patients admitted to the Department of Respiratory and Critical Care Medicine at West China Hospital between January 2010 and December 2019 were retrospectively analyzed and randomly divided into training and validation sets (7:3). Least absolute shrinkage and selection operator (LASSO) and multivariable logistic regression analyses were performed to screen independent predictors, which were incorporated into a nomogram. Model performance was evaluated through receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA).

A total of 3146 patients were included, with 2202 in training set and 944 in validation set. Twelve variables were independently associated with prolonged LOHS and incorporated into the nomogram, including onset seasons, hemoglobin (Hb), white blood cells (WBC), albumin (ALB), sputum cultures, respiratory failure and respiratory support modes, use of anticoagulants, antibiotics and systemic glucocorticoids, heart failure, and malignant tumors. The area under the ROC curve (AUC) was 0.760 (95% CI: 0.737-0.783) in training set and 0.746 (95% CI: 0.710-0.783) in validation set, demonstrating strong discrimination. Calibration curves closely aligned with the diagonal, indicating high consistency between predicted and observed values. DCA showed substantial net benefits, highlighting the model's clinical utility.

A nomogram prediction model was developed and validated, facilitating early identification of AECOPD patients at risk for extended LOHS and supporting informed clinical decision-making and optimal resource allocation.