Development and Internal Evaluation of a Nomogram for Clinically Significant Depressive Symptoms in Obstructive Sleep Apnea-Hypopnea Syndrome

Abstract

Purpose To develop and internally evaluate a nomogram for identifying clinically significant depressive symptoms in hospitalized patients with obstructive sleep apnea-hypopnea syndrome (OSAHS). Methods A single-center retrospective study included adults hospitalized at the First Hospital of Hebei Medical University between January and December 2025. OSAHS was confirmed by overnight polysomnography (PSG) and defined as an apnea-hypopnea index (AHI) ≥5 events/h. Depressive symptoms were assessed using the Zung Self-Rating Depression Scale (SDS), with a standardized score ≥53 defining clinically significant depressive symptoms. Thirty-six candidate predictors were extracted from electronic medical records and laboratory information systems. The cohort was randomly divided into a training cohort (n=574) and an internal validation cohort (n=246). Least absolute shrinkage and selection operator (LASSO) logistic regression with 10-fold cross-validation, followed by post-LASSO multivariable logistic regression, was used to select predictors and construct the nomogram. Discrimination, calibration, and clinical utility were assessed using the area under the curve (AUC), calibration analyses, and decision curve analysis. Results Among 820 patients, 88 (10.7%) had clinically significant depressive symptoms, including 62 events in the training cohort and 26 in the internal validation cohort. LASSO initially retained seven candidate predictors. Five predictors remained statistically significant after post-LASSO multivariable logistic regression and were included in the final nomogram: oxygen desaturation index (ODI), Pittsburgh Sleep Quality Index (PSQI), Epworth Sleepiness Scale (ESS), thyroid-stimulating hormone (TSH), and free triiodothyronine (FT3). The model showed favorable discrimination in the training cohort (AUC=0.910, 95% confidence interval [CI] 0.869–0.950) and internal validation cohort (AUC=0.915, 95% CI 0.853–0.977). Calibration was acceptable, with calibration intercepts of approximately 0.000 and −0.438 and calibration slopes of 1.000 and 0.814, respectively. Conclusion This internally evaluated nomogram integrating ODI, PSQI, ESS, TSH, and FT3 showed favorable performance for identifying SDS-defined clinically significant depressive symptoms in hospitalized patients with OSAHS. Prospective multicenter and temporal external validation is required to assess its generalizability and clinical applicability.