PP1502 - cVEMP Age-Related Normative Data: Different Methods to Define Norms

Abstract:
In the clinic, cVEMP amplitude, interaural amplitude asymmetries, and peak latencies are measured to identify potential vestibular dysfunction. We collected cVEMP amplitude and threshold data from 444 otologically and neurologically normal individuals examined in the audiology department at Mayo Clinic Florida between 2018 and 2022. Distributions for each parameter were analyzed, and normal reference values from the second to the ninth decades were generated. Latency data follow standard parametric methods (means and standard deviations). Amplitude data were better described after log transformation. Age-specific normal reference values and differences between transformed and non-transformed descriptive methods will be reviewed.

Summary:

Rationale
In the clinic, cVEMP amplitude, interaural amplitude asymmetries, and peak latencies are measured to identify potential vestibular dysfunction. In particular, cVEMP amplitude can be classified as either small/absent, normal, or too large. Each classification implies different forms of vestibular dysfunction. Many studies have demonstrated systematic age-related changes in the cVEMP response, necessitating a need for age-specific normative data. We noticed that commonly applied methods for describing cVEMP amplitude (mean and standard deviation estimates) could result in misclassification of observed cVEMP amplitude, complicating the detection of third window syndromes or vestibular neuronitis. This study aimed to better define the distributions of various cVEMP parameters by age to reassess in a set of otologically and neurologically normal individuals.

Methods
We collected cVEMP amplitude and threshold data from 444 otologically and neurologically normal individuals examined in the audiology department at Mayo Clinic Florida between 2018 and 2022. Distributions for P1 and N1 latency, amplitude, and inter-aural asymmetry were analyzed and fit to normal and log-transformed distributions stratified by decade. Additionally, regression models were developed to predict cumulative probabilities for each observed parameter by age.

Results & Conclusions
The sample was large enough to represent the observed distributions with reasonable granularity (acceptable goodness of fit). Normal reference values from the second to the ninth decades were generated. Latency data were well described by standard parametric methods (means and standard deviations). However, amplitude data were better described after log transformation. Specifically, cVEMP amplitudes that were smaller or larger than expected for age were more accurately predicted after log transformation. In some age ranges, cVEMP amplitudes appeared too large when modeled using untransformed mean and standard deviation limits but were in the normal range when modeled after log transformation. This implies that untransformed cVEMP amplitude limits may overidentify cases of suspected third window syndrome or superior canal dehiscence.