Abstract
BACKGROUND: Based on the clinical characteristics of patients, a nomogram predicting the prognosis of patients suffering from sudden sensorineural hearing loss (SSNHL) was constructed, which could aid in personalized treatment. METHODS: Data on the clinical characteristics of patients with SSNHL were collected and statistically analyzed. A nomogram for predicting the hearing prognosis of SSNHL patients were then constructed. RESULTS: A total of 356 patients were included in this study, including 227 and 129 in the recovery group (63.76%) and non-recovery group (36.24%), respectively. Univariable logistic regression demonstrated that age, gender, body mass index (BMI), marital, Audiogram curve, vertigo, hearing loss degree, and time to initial treatment were associated with hearing outcomes. Multivariate logistic models showed that age [odds ratio (OR): 0.479, 95% confidence interval (CI): 0.301-0.748, P<0.001], descending (OR: 0.116, 95% CI: 0.047-0.275, P<0.001) and flat audiogram curves (OR: 0.397, 95% CI: 0.159-0.979, P=0.045), profound hearing loss (OR: 0.047, 95% CI: 0.013-0.152, P<0.001), and treatment initiation after 1 week (8-14 days: OR: 0.047, 95% CI: 0.013-0.152, P<0.001; >14 days: OR: 0.131, 95% CI: 0.039-0.413) were risk factors for the hearing recovery. Logistic regression analyses were conducted to construct the prognostic nomogram. As estimated by the area under the receiver operating characteristic curve (ROC), the model had an accuracy of 0.867 (95% CI: 0.709-0.747). The validation analysis confirmed the high accuracy of the nomogram, and the decision curve showed that the model has potential clinical application value. CONCLUSIONS: This study demonstrated that age, descending and flat audiogram curves, profound hearing loss, and initiating treatment after 1 week of SSNHL onset were independent risk factors associated with a worse hearing recovery prognosis. Using these factors, a nomogram with a high prediction accuracy was developed to predict the hearing recovery rate of SSNHL patients.