Relationship between pure-tone audiogram findings and speech perception among older Japanese persons

Abstract

Objective: To clarify how the pure-tone threshold (PTT) on the PTA predicts speech perception (SP) in elderly Japanese persons.

Methods: Data on PTT and SP were cross-sectionally analyzed in Japanese persons (656 ears in 353 patients, aged ≥65 years). Correlations of SP and average PTT in all tested frequencies were evaluated by Pearson’s correlation coefficient and simple linear regression. After adjusting for sex, laterality of ears, and age, the relationship of average and frequency-specific PTT with impaired SP ≤50% was estimated by logistic regression models.

Results: SP correlated well (r = −0.699) with the average PTT of all tested frequencies. On the other hand, the correlation between patient age and SP was weak, especially among ≤85-year-old persons (r = −0.092). Linear regression showed that the average PTT corresponding to SP of 50% was 76.4 dB nHL. Odds ratios for impaired SP were highest for PTT at 2000 Hz. Odds ratios were higher for middle (500, 1000, 2000 Hz) and high frequencies (4000, 8000 Hz) than low frequencies (125, 250 Hz).

Conclusion: The PTT on the pure-tone audiogram (PTA) is a good predictor of SP by speech audiometry among older persons, which could provide clinically important information for hearing aid fitting and cochlear implantation.

Chinese abstract

目的：澄清PTA纯音阈值（PTT）如何预测日本老年人的言语感知（SP）。

方法：日本人的PTT和SP数据得以横向分析（353患者的656个耳, 年龄≥65岁）。通过Pearson相关系数和简单线性回归分析, 评估了所有测试频率中SP和平均PTT的相关性。经过对性别、耳朵的倾斜度和年龄进行调整后, 通过逻辑回归模型估测了平均的和频率特定性的PTT与减弱的语言感知能力（ SP ≤50％）的关系。

结果：SP与所有测试频率的平均PTT的相关性良好（r = -0.699）。另一方面, 患者年龄和SP之间的相关性较弱, 特别是年龄为85岁以下的患者（r = -0.092）。线性回归分析显示, 相对于SP值为50％的平均PTT为76.4 dB nHL。减弱的SP的优势比在PTT为2000Hz时最高。对于中频（500、1000、2000Hz）和高频（4000、8000Hz）的优势比比低频（125、250Hz）的优势比高。

结论：纯音听力图（PTA）的PTT对老年人的语音听力测定是言语感知的良好预测指标, 可为安装助听器和耳蜗植入提供重要的临床信息。

Keywords:

• Presbycusis
• audiogram
• hearing threshold
• speech perception
• hearing aid
• cochlear implant
• Japanese

Disclosure statement

The authors have no conflicts of interest to declare.