1770只未筛选耳中颈性VEMP与后管vHIT之间的功能性分离

文献信息

中文摘要

目的： 颈性前庭诱发肌源性电位（cVEMP）和后管视频头脉冲试验（PC-vHIT）通常被视为同一前庭下单位功能的测量指标。我们量化了二者不一致的频率，以及这是否随年龄变化。
方法： 对单中心前庭登记数据库进行回顾性横断面分析。提取所有具有双耳500 Hz cVEMP振幅和后管vHIT增益的记录，不进行诊断或年龄筛选。就每只耳而言，若反应缺失，或耳间振幅绝对差≥40%且该耳振幅较小，则cVEMP为异常；双侧反应缺失被归类为双侧异常并单独分析。PC-vHIT异常定义为增益<0.7。采用患者聚簇自举法和广义估计方程，以Cohen's κ量化一致性。
结果： 共分析来自868例患者885次检查的1770只耳（平均年龄53.9岁）。在至少一项测试异常的1027只耳中，747只（72.7%）仅一项测试异常；280只（27.3%）为一致性异常。一致性极低（κ=0.097，95% CI 0.047-0.150）。两种异常率均随年龄升高，但κ在各年龄层间保持平稳（<40、40-59、60-69和≥70岁分别为0.073、0.056、0.040和-0.012），因此与年龄相关的原始不一致率下降反映的是患病率升高，而非二者关系改变。这种升高几乎完全由双侧cVEMP缺失所致。
结论： 在约四分之三的受累耳中，cVEMP与PC-vHIT不一致，且在各年龄段经机会校正后的一致性均接近零。二者更宜被理解为共享外周节段的两个反射环路，而非同一实体的可互换测量指标。

英文摘要

OBJECTIVE: Cervical vestibular evoked myogenic potentials (cVEMP) and posterior canal video head impulse testing (PC-vHIT) are conventionally treated as measures of one inferior vestibular unit. We quantified how often they disagree, and whether this varies with age.
METHODS: Retrospective cross-sectional analysis of a single-center vestibular registry. All records with 500 Hz cVEMP amplitudes and posterior canal vHIT gains for both ears were extracted without diagnostic or age filtering. Per ear, cVEMP was abnormal if the response was absent, or if the absolute interaural amplitude difference was ≥ 40% and that ear had the smaller amplitude; bilaterally absent responses were classified abnormal bilaterally and analyzed separately. PC-vHIT abnormality was gain < 0.7. Agreement was quantified with Cohen's κ, using patient-clustered bootstrapping and generalized estimating equations.
RESULTS: A total of 1770 ears from 885 examinations in 868 patients were analyzed (mean age 53.9 years). Of 1027 ears abnormal on at least one test, 747 (72.7%) were abnormal on only one; 280 (27.3%) were concordantly abnormal. Agreement was minimal (κ = 0.097, 95% CI 0.047-0.150). Both abnormality rates rose with age, yet κ was flat across strata (0.073, 0.056, 0.040, and -0.012 for < 40, 40-59, 60-69, and ≥ 70 years), so the age-related fall in raw discordance reflects rising prevalence, not a changed relationship. That rise was almost entirely bilateral cVEMP absence.
CONCLUSION: cVEMP and PC-vHIT disagree in roughly three-quarters of affected ears, with chance-corrected agreement near zero at every age. They are better read as two reflex loops sharing a peripheral segment, not interchangeable measures of one entity.