Strengths and limitations
This is the first study reporting the diagnostic accuracy of automated skew deviation measurements by a portable VOG system designed for the assessment of dizzy patients in the ED. We did not test, however, the specificity of the VOG device in measuring misalignments in subjects without inducing any skew. Unlike the APCT, it allows a precise quantification of eye misalignment with an objective measure of eye excursion.
Our investigative setup is not generalizable to all other systems on the market. We used VOG goggles, designed and approved for measuring static and dynamic vestibular reflexes such as the VOR. Any test using certified eye tracking systems for that purpose can be adapted using an alternating manual or automated cover to obtain skew deviation measurements. In order to assure meaningful saccade detection we recommend a frame rate of 250 Hz.
It has been shown in previous studies that skew deviation varies with upright and supine body position in patients with a vestibular imbalance.21 In our study, we tested participants solely in an upright sitting position.
We studied only healthy subjects and did not test patients with spontaneous nystagmus. Such vestibular nystagmus is predominately horizontal, however, a mixed torsional and upbeat component in incomplete, superior vestibular neuritis, might bias the clinical assessment. Small vertical eye movements from the underlying nystagmus might be falsely interpreted as vertical skew during the clinical APCT. Physiological skew deviation (<= 0.30 degrees4) in healthy participants, however, is usually not discernable.
We were not able to induce small skews in all healthy participants due to central adaptation (see condition 1- 4 PD in Figure 5). We therefore asked participants to keep their eyes closed before the measurements began in order to prevent as much as possible prior adaptation and to maximize the success of obtaining a valid skew measurement. Those participants with small induced skews (< 2 PD (1.14°)) were still detectable by the eye tracker, provided that the recorded eye traces were clean and free from artifacts (Figure 2).
Finally, during the measurements of large skews (10PD (5.7°)), there was selection bias due to a limited ordinal scale from 1-10PD (0.57-5.7°) compared to continuous results supplied by the VOG. Our current study investigated artificially induced skew on healthy participants. Thus, it has not yet been tested on patients with pathological skews. Future studies must be validated with patients suffering from conditions such as heterophoria, acute vestibular imbalance or other neuro-ophthalmological ailments affecting the alignment of the eyes.
Additionally, it would be interesting to use this quantitative measurement of skew alongside with other vestibular exams such as oVEMPs, for example, to confirm the involvement of the utricular pathway in transitory skews observed in patients with an acute vestibular syndrome.