Association between baseline neutrophil-to-lymphocyte ratio and short-term hearing recovery after glucocorticoid therapy in idiopathic sudden sensorineural hearing loss

Objective This study aimed to evaluate the predictive value of the neutrophil-to-lymphocyte ratio (NLR) for day-7 hearing recovery following oral glucocorticoid therapy in patients with idiopathic sudden sensorineural hearing loss (ISSNHL). Methods We retrospectively reviewed the medical records of 214 patients with unilateral ISSNHL who were admitted to the Department of Otolaryngology–Head and Neck Surgery between 2022 and 2025. Multivariable logistic regression was used to assess the association between baseline NLR and day-7 recovery status (recovered vs. non-recovered), and E-values were calculated to evaluate the potential impact of unmeasured confounders. Stratified analyses and interaction tests were performed to examine potential effect modifications across predefined subgroups. Receiver operating characteristic (ROC) curve analysis was conducted to assess predictive performance, and the optimal NLR threshold was determined using the Youden index. To address potential overfitting and improve reproducibility, internal validation and calibration of the NLR-only model were performed using bootstrap resampling, including the optimism-corrected AUC, Brier score, calibration-in-the-large, and calibration slope. Results A higher NLR was significantly associated with a lower likelihood of day-7 hearing recovery. In the fully adjusted model, the odds ratio (OR) for NLR was 0.40 [95% confidence interval (CI), 0.30–0.53]. The E-value was 2.54 (upper limit of the 95% CI, 2.09), supporting the robustness to unmeasured confounding factors. This association was generally consistent across subgroup analyses. ROC analysis showed strong predictive accuracy, with an apparent AUC of 0.90 (95% CI, 0.85–0.95). The optimal NLR cut-off was 4.59, with a specificity of 97.66% and a sensitivity of 80.23%. Bootstrap internal validation demonstrated minimal overfitting (optimism-corrected AUC 0.90) and satisfactory calibration (optimism-corrected calibration-in-the-large = 0.0039; calibration slope = 1.0032). Conclusion Baseline NLR is a simple and readily available biomarker that predicts day-7 hearing recovery following oral glucocorticoid therapy in patients with ISSNHL. An elevated NLR is associated with a lower probability of recovery, supporting its potential utility for early risk stratification; nevertheless, external validation in independent cohorts is warranted.