Phenotypic Profiling of People With Subjective Tinnitus and Without a Clinical Hearing Loss

Our objective was to study the characteristics of patients with subjective tinnitus and normal hearing and to investigate whether the features correlated to different shapes on audiograms. In this retrospective study, 313 patients with subjective tinnitus and clinically normal hearing were enrolled from the tinnitus outpatient department of the Eye and ENT Hospital of Fudan University. The following phenotypic variables were collected: age, dominant tinnitus pitch (TP), tinnitus loudness, tinnitus duration, tinnitus severity, sex, education, hearing thresholds, tinnitus position, and tinnitus condition. The dominant TPs of patients with normal hearing were mostly high-pitched, with a mean of 4866.8 ± 2579.6 Hz; thus, we speculated that the condition is related to high-frequency hearing threshold elevations. We further divided the patients into four subgroups based on the matched TP: (i) TP ≤ 500 Hz (n = 34), (ii) 500 Hz < TP ≤ 3,000 Hz (n = 15), (iii) 3,000 Hz < TP ≤ 8,000 Hz (n = 259), and (iv) TP > 8,000 Hz (n = 5). We studied the phenotypic profiling of different audiograms and found that the group with TP of ≤500 Hz had an average "inverted-U" shaped audiogram, and the group with TP between 500 and 3,000 Hz had a slowly ascending slope audiogram below 2,000 Hz, followed by a drastically descending slope audiogram ranging from 2,000 to 8,000 Hz; further, the high-frequency (3,000-8,000 Hz) and ultra-high-frequency (>8,000 Hz) groups had flat curves below 2,000 Hz and steeper slope audiograms over 2,000 Hz. Our findings confirmed a consistency ratio between the distributions of dominant TPs and the frequencies of maximum hearing thresholds in both ears. The dominant TP was positively correlated with the maximum hearing threshold elevation frequency (left ear: r = 0.277, p < 0.05; right ear: r = 0.367, p < 0.001). Hearing threshold elevations, especially in high frequency, might explain the appearance of dominant high-frequency TP in patients without clinically defined hearing loss. This is consistent with the causal role of high-frequency coding in the generation of tinnitus.