A Prognostic Model to Predict Hearing Recovery in Patients With Idiopathic Sudden Onset Sensorineural Hearing Loss

Abstract

ImportanceThe prognosis of idiopathic sudden onset sensorineural hearing loss (iSSNHL) is uncertain, which creates challenges in clinical decision-making for ear, nose, and throat (ENT) physicians and adds to the burden of the condition experienced by patients.ObjectiveTo develop and internally validate a prognostic model for hearing recovery among patients with iSSNHL to support ENT surgeons in making informed and individualized treatment decisions.Design, Settings, and ParticipantsThis prognostic study and model used cohort data from the Sudden Onset Sensorineural Hearing Loss study, which included 812 patients (age ≥16 years) diagnosed with iSSNHL at 76 National Health Service ENT departments in the UK from December 2019 to May 2022. Nine variables previously reported as independent prognostic factors for complete recovery of patients with iSSNHL were selected for inclusion. The final model was internally validated using bootstrapping with 500 repetitions, then coefficients were adjusted for the degree of optimism in the model. The model intercept was reassessed after adjustment of model coefficients. Impact of individual predictors was evaluated by estimating odds ratios with corresponding 95% CIs. Model performance was re-evaluated after internal validation and expressed by discrimination, calibration, and clinical utility. Data analyses were performed from March 2022 to April 2024.InterventionRoutine treatment (per National Health Service standards), including oral steroids and intratympanic steroid injections.Main Outcome and MeasuresComplete hearing recovery defined as a return to within 10 dB of the patient’s before iSSNHL hearing levels at all frequencies in the affected ear at 6 to 16 weeks after iSSNHL symptom onset.ResultsThe study sample included 498 patients (mean [SD] age, 58.7 [16.0] years; 215 [46.9%] females and 243 [53.1%] males) who met the criteria for inclusion in the model. Of those, 210 (46%) were classified as having experienced complete hearing recovery. Five variables were found to be independent predictors for complete hearing recovery: steroid treatment within 7 days from symptom onset (OR, 5.23 vs no treatment ), lower severity of hearing loss at presentation (OR, 0.19 if loss is mild), absence of vertigo (OR, 0.56 vs no vertigo), younger patient age (OR, 0.64 per year), and a history of cardiovascular disease (OR, 1.84 vs no cardiovascular disease). The model showed good performance after internal validation with a c-index of 0.77 (95% CI, 0.7-0.81). Predictions for complete recovery aligned well with observed complete recovery rates, and greater clinical utility than treat all or treat none strategies was shown.Conclusion and RelevanceThis prognostic model evaluated in this study may be able to assist ENT surgeons in making informed treatment decisions for individual patients with iSSNHL. It is available online at no cost.