Abstract
Background/Objectives: Tonal thresholds, typically assessed through pure-tone audiometry (PTA), are central to the audiological evaluation of Ménière's disease (MD). However, they fail to capture the complexity of real-life speech perception. This study aimed to characterize the relationship between PTA and speech recognition performance in unilateral MD and to determine whether a measurable dissociation exists between hearing sensitivity and verbal processing. We also evaluated frequency-specific audiometric patterns and potential threshold cut-off values associated with speech recognition decline. Methods: A total of 254 ears from 127 patients were included in the study across three groups: the Ménière group (affected and unaffected ears, n = 64 each) and the control group (n = 63). The pure-tone thresholds, speech recognition threshold (SRT), and the maximum word recognition scores (Rmax) were assessed in all participants. Results: Although the Ménière affected and control groups showed comparable pure-tone average (PTA) values (46.33 vs. 38.13 dB, p = 0.439), the affected group demonstrated significantly poorer speech performance (Rmax: 76.25% vs. 87.49%, p < 0.001; SRT: 50.64 vs. 38.45 dB, p = 0.009). The Ménière unaffected group exhibited near-ceiling performance (Rmax: 99.38%, SRT: 18.33 dB) and a mean PTA of 19.59 dB. A strong correlation between PTA and SRT was observed only in the Ménière affected group (r = 0.942, p < 0.001), whereas correlations were moderate in the unaffected (r = 0.671, p < 0.001) and control (r = 0.728, p < 0.001) groups. The ROC analysis revealed that PTA predicted impaired speech recognition with high accuracy in unaffected (AUC = 0.956, p < 0.001) and control (AUC = 0.829, p < 0.001) ears but far less so in affected ears (AUC = 0.784; all p < 0.001), confirming a functional tonal-verbal dissociation in MD. Conclusions: This study demonstrates a functional audiometric dissociation in unilateral Ménière's disease. Affected ears show poorer speech recognition and require higher intensities despite similar PTA values. The predictive link between thresholds and verbal performance is disrupted. These findings support the need for combined tonal and speech-based assessment in clinical practice.