Migraine is not a disorder of the inner ear, but of the brain. Thus one would expect VEMP testing to be normal in the sense that they should exist and be of average intensity. On the other hand, Migraine lowers the thresholds in general for many types of sensory inputs. For this reason, it would be reasonable to suppose that Migraine individuals might have generally lower thresholds. This idea so far has not been confirmed, and at this point, we would have to say that VEMP testing has no rationale in Migraine, other than to find other conditions.
Interestingly, when researchers have found that results conflict with the dogma about migraine expressed above, they sometimes propose that the migraine patients that they studied, actually have a different disease, such as Meniere's disease (e.g. Baier et al, 2009). This is certainly possible, although we would not limit the disease options to Meniere's disease. It does not appear that VEMPs can diagnose Meniere's disease however, so this inference doesn't really make much sense. It may be that vestibular migraine (VM) is a proxy for "undiagnosed dizziness with headaches", and what these investigators are really saying is that they have confused migraine with other illnesses that injure the ear.
Fife et al (2017) in a practice parameter published in Neurology, stated that "VEMP may not be used to assist in VM diagnosis or management".
cVEMPS in Migraine
Jung et al (2015) reported "Abnormal caloric, VEMP, and vestibular ratio measurements were found in 25%, 29%, and 58%, respectively. " They used the term "vestibular ratio" for posturography tests with a "vestibular pattern". They did not use a control group, and the degree to which these findings differ from chance is unclear.
Hong et al (2011) reported cVEMPS were abnormal in some of 30 subjects with migrainous vertigo. This is "strange" because migraine is a disorder of the brain characterized by reduced threshholds, and one would not expect a reduction in any vestibular response. We think that a study of more subjects is needed.
Similarly, Liao and Young (2004) reported cVEMPS were abnormal in about a third of 20 patients with basilar artery migraine. This paper also reported abnormal OKN testing, caloric tests, and directional preponderance. Again, we are dubious. As migraine is extremely common, we again would like to see a much larger study.
Baier and Dietrich (2009) reported that patients had reduced cVEMPS - 68% of migraine patients had reduced amplitudes compared to age-matched controls. We find this puzzling.
Moallemi et al (2015) found no difference in amplitude ratios for cVEMP in migraine vs. normal subjects. We think this is reasonable and applaud the Acta Med Iran journal where this was published for publishing a negative paper.
One possible explanation for the results above is that "migraine" is not a single disorder, and that an appreciable fraction of them also have inner ear damage.
oVEMPs in Migraine --
Zaleski et al (2015) reported that migraine patients have more common absence of oVEMPS as well as more asymmetry. We find this puzzling, as we would expect stronger responses in migraine patients rather than weaker responses.
Kim et al (2015) studied both cVEMPs and oVEMPS in 38 migraine patients, 30 tension headache patients, and 50 healthy control subjects. The reported that longer oVEMP latencies were found in the migraine group, but there were no differences in cVEMPs. We are dubious that oVEMP latencies are relevant to migraine, and we suspect that this finding was a statistical anomaly.
Inoue et al (2016) studied 28 patients with vestibular migraine, Meniere's and controls. They reported that asymmetry ratios for cVEMPs and oVEMPs were significantly larger in VM than controls, and the ARs for oVEMPs and caloric tests were smaller in VM than Meniere's disease. It would seem to us that this is hardly diagnostic.
As of 2016, there is little evidence that any type of VEMP is useful in Migraine. Their main role is to diagnose other illnesses that might be confused with Migraine. We agree with Fife et al (2017), who in a practice parameter published in Neurology, stated that "VEMP may not be used to assist in VM diagnosis or management".