Jeroen Venhovens

The application of static SVV testing outside specialized tertiary dizziness clinics is also hampered by the high costs of the specialized equipment. Interestingly, Zwergal et al.32 studied a low-cost, easily performed, and reliable bedside static SVV test by using a selfmade modified bucket, see figure 1. The conversion of the bucket takes less than one hour and costs less than 5 dollars. The authors compared the bucket-test with conventional static SVV testing by using a hemispheric dome and they found an excellent inter- and intra-test reliability (Pearson correlation coefficients respectively 0.89-0.90 and 0.92) both in healthy subjects and in patients with acute neurovestibular disorders. Brodsky et al.33 recently repeated the bucket version of the static SVV test in paediatric patients with suspected neurovestibular disorders, however the authors modified the original bucket test by using a smartphone with a subjective vertical application which costs about 18 dollars. The authors also compared the smartphone-based bucket version of the static SVV test with conventional static SVV testing while using a hemispheric dome and found a moderate correlation (Pearson correlation coefficient 0.43). However, despite the moderate correlation with conventional static SVV testing smartphone-based bucket static SVV testing seems to be a useful screening test for acute neurovestibular disorders with a sensitivity of 75%, specificity of 100%, positive predictive value of 100%, and negative predictive value of 92%. However, as promising as these new methods seem, further research with additional comparisons of the (smartphone-assisted) bucket version with the conventional static SVV test is needed in larger groups in order to confirm the earlier research reports.

Figure 1. Bucket version of the SVV test, adapted from Zwergal et al. (2009)32

PART TWO
Neurovestibular analysis and falls in Parkinson’s disease and atypical parkinsonism. A baseline study and one-year follow-up study

chapter 5
Recurrent falls are a major problem in PD patients, and even more so in patients with AP. About 50% of the PD patients have recurrent fall episodes, and approximately 70% fall (at least once) annually34–37. PD patients have a 2.2-fold increase in fractures overall, and a 3 to 4-fold increase in hip fractures compared to a healthy age-matched population. Moreover, PD patients with hip fractures have higher overall mortality rates, and higher rates of surgical and post-operative complications38. Recently, a practical consensus-based overview concerning the risk factors for falling in PD patients was published39. However, vestibular function abnormalities were not addressed as an individual risk factor, which is surprising as the vestibular system is one of the key systems in maintaining balance by integrating multimodal sensory information (i.e. vestibular, proprioceptive, and visual sensory input) and by secondarily adjusting the outgoing motor response by means of the vestibulospinal reflexes37. Chapter 5 shows the results of a case-control study in which the vestibular system function in 25 healthy age-matched volunteers, 30 PD, and 14 AP patients (6 patients with multiple system atrophy, 3 with progressive supranuclear palsy, and 5 with vascular parkinsonism) was systematically investigated and compared to an age and gender-matched healthy control group37. Ninety percent of PD patients (27 of 30) and 100% of AP patients (all 14) showed signs of neurovestibular dysfunction, mainly with a central vestibular dysfunction profile, on laboratory examinations (i.e. VEMP, SVV, videonystagmography with additional caloric and rotatory chair stimulation). Moreover, vestibular testing abnormalities were correlated with an increased risk of falling when fallers among PD and AP patients were compared with non-fallers (P ≤ 0.001).

Abnormal neurovestibular laboratory function test results are seen remarkably often in both PD and AP patients (90% of the PD patients and 100% of the AP patients). About 66% of the patients with mild Parkinson’s disease had abnormal neurovestibular laboratory function test results in comparison to 96% and 100% of the patients with moderate and severe Parkinson’s disease respectively (however, only 1 patient with severe Parkinson’s disease was included). Therefore, abnormal neurovestibular laboratory test results are very common in early Parkinson’s disease, however, they seem to increase in advancing stages of the disorder. Also, patients with atypical Parkinsonism were clinically more affected in comparison to the patients with Parkinson’s disease (respectively Hoehn and Yahr stages 2.9±0.8 and 2.3±0.7; P = 0.0152). Persons with vestibular and or other neurological disorders (i.e. other than Parkinson’s disease or atypical parkinsonism) in their medical history were excluded from the study. Moreover, none of the patients had relevant actual complaints or signs during the neurological and neuro-otological examinations suggestive of other vestibular disorders. We were obviously not able to exclude all other (central) neurovestibular disorders, mainly because the absence of current cerebral imaging studies. However, we did not have any clues suggestive of other neurovestibular disorders in the absence of relevant current medical complaints or abnormalities during neurological and neurovestibular clinical testing.

These abnormal neurovestibular function test results mainly have a central vestibular dysfunction profile in 78-93% of the PD and AP patients, whereas only one PD patient (4%) had a profile suggestive of a peripheral vestibular disorder (the other vestibular dysfunction profiles were non-localizing). Also, 57-79% of the AP and PD patients did not have complaints of dizziness or vertigo. However, one can question whether these central neurovestibular laboratory testing abnormalities are truly asymptomatic. Indeed, 10-18% of the AP and PD patients with falls had neurovestibular testing abnormalities as the only identifiable risk factor for falling (after exclusion of orthostatic hypotension, postural instability, freezing of gait, and moderate-to-severe cognitive problems)37. However, we acknowledge that we were not able to exclude all possible and previously reported risk factors for falling in PD, but we did exclude the most relevant causes. Also, we may have missed additional contributions of, for example, mild orthostatic hypotension which can be variably present during clinical examination, and which can be missed during routine testing40. However, neurovestibular testing abnormalities were strongly correlated with an increased risk of falling when PD and AP patients were compared with non-falling patients (P≤0.001). Therefore, it was concluded that central neurovestibular dysfunction on vestibular laboratory testing is highly prevalent in both PD and AP patients compared to healthy volunteers, and is strongly associated with an increased risk of falling, which has not been reported earlier in the literature37. I acknowledge that the studied groups in the study were small, and also had a heterogeneous population (especially in the AP group, where subgroups were small). Therefore, the results must be interpreted cautiously and as hypothesis-generating for future research in larger groups. Moreover, it is currently unknown whether these strongly associated neurovestibular laboratory abnormalities are directly associated with falling or a secondary epiphenomenon, for instance as a result of deconditioning. Also, dopaminergic medication does not seem to be the cause of these abnormalities as Pötter-Nerger demonstrated that levodopa may partially decrease vestibular testing abnormalities41,42. However, the central vestibular system is one of the key systems for maintaining balance by integrating multimodal sensory input (i.e. the visual, peripheral vestibular, and proprioceptive input) and by subsequently adjusting the outgoing motor response by means of the vestibulospinal reflexes. Therefore, one could argue, purely speculative and based upon theoretical physiological evidence, that vestibular dysfunction is causally linked to falling as a possible cause and not an epiphenomenon.

These central vestibular testing abnormalities in PD patients mostly consisted of abnormal evoked potential test results (VEMPs and BAEPs), which showed significant prolongation of the latencies (p13, n1, and interpeak III-V latencies) on the symptomatic brainstem side compared to healthy volunteers (0.003 ≤ P ≤ 0.019). However, comparison of these latencies at the asymptomatic side of PD patients with those of healthy volunteers, or a side-to-side comparison within PD patients, did not yield significantly different results, with the exception of the p13 cervical VEMP latency side-to-side comparison (P = 0.020). These results point to an asymmetrical brainstem involvement in the neurodegenerative process, mainly affecting the symptomatic brainstem side, and subsequently traveling along with the affected fibre tracts (i.e. the affected brainstem side is ipsilateral to the patient’s resting tremors, rigidity, and bradykinesia at the pontomedullary junction below the level of the vestibular nuclei; and the affected fibre tracts cross over to the contralateral brainstem side above the level of the vestibular nuclei along with the decussation of the medial longitudinal fasciculus). There is just one prior study by Pötter-Nerger et al.41, which only shows asymmetrical abnormalities of the ocular VEMP tests involving the symptomatic brainstem side, whereas the cervical VEMP tests did not show these asymmetrical abnormalities. However, the differences between Pötter-Nerger et al.41’s study and ours can possibly be explained by the difference in population size (i.e. 30 PD patients in our study versus 13 patients in the study by Pötter-Nerger et al.41), as small differences between groups are more easily detected in a larger study population. Our study demonstrates that not only the central vestibular brainstem pathways are affected at the symptomatic brainstem side, but also that the central auditory brainstem pathways seem to be affected, which is illustrated by the prolonged III-V BAEP interlatencies at the symptomatic brainstem side. As discussed earlier, the main BAEP and VEMP abnormalities in our study consisted of severely prolonged cervical VEMP p13 latency, ocular VEMP n1 latency, and BAEP III-V inter-latency prolongations (i.e. p13 mean latency prolongation of 134% upper limit of normal (ULN); n1 mean latency prolongation of 128% ULN; mean III-V BAEP inter-latency prolongation of 120% ULN) with preserved amplitudes37. As discussed in paragraph 1.1 of this chapter, asymmetrically severely prolonged evoked potential latencies combined with relatively preserved amplitudes can theoretically be caused by a primary demyelinating neurodegeneration. However, the pathophysiological distinction between axonal and demyelinating central nervous system pathologies based upon evoked potential measurements is purely speculative and has to our knowledge not been studied earlier in literature. At this moment, clear neurophysiological criteria for this pathophysiological differentiation in central nervous system disorders are lacking, in contrast to the neurophysiological criteria that are present for this differentiation of the peripheral neuropathies. Further research concerning this pathological differentiation between axonal and demyelinating central nervous system disorders, based upon evoked potential measurements seems interesting, as this gives important insights into the pathophysiological hallmark of the underlying disease.

Parkinson’s disease is defined pathophysiologically by the presence of abnormal alpha-synuclein protein aggregations in neurons and glial cells, and by progressive neurodegeneration of selected brain regions (e.g. cholinergic pedunculopontine nucleus, dopaminergic substantia nigra and ventral tegmental area). These protein aggregates are called: a) Lewy bodies in neuronal perikarya, b) Lewy neuritis in neuronal processes, and c) coiled bodies in affected oligodendrocytes43,44. Also, multiple system atrophy (MSA) – one of the forms of atypical parkinsonism – is characterized by neuronal and glial alpha-synuclein aggregates, however, it is generally believed that neurons in MSA also have inclusions within their brainstem nuclei in contrast to PD44. Seidel et al.43 demonstrated Lewy bodies and Lewy neuritis in all cranial nerve nuclei, premotor oculomotor, precerebellar, and vestibular nuclei for the first time in PD patients; this contrasted with the general belief that such abnormalities were limited to MSA patients. Moreover, Lewy neuritis and coiled bodies were demonstrated in all brainstem fibre tracts. These abnormalities may be the result of a transneuronal disease spread along the anatomical pathways due to disturbed intra-axonal transport processes43. When we combine both the neuropathological and neurophysiological research data in PD patients, a transneuronal disease progression along the anatomical pathways, asymmetrically affecting the brainstem seems most likely (i.e. brainstem fibre tracts ipsilateral to the patient’s clinically most affected side at the medullopontine junction, and contralateral to the patient’s clinically most affected side above the motor decussation at the pontine and pontomesencephalic junction). However, asymmetrical brainstem degeneration spreading along with the brainstem fibre tracts below the level of the substantia nigra in the mesencephalon has not yet been confirmed in neuropathological studies and seems an interesting topic for future research. The neuropathological brainstem abnormalities suggest a neurodegenerative process with a primary axonal hallmark43, however the electrodiagnostic evoked potential abnormalities (i.e. severely delayed latencies with relatively spared amplitudes) suggest an important demyelinating neurodegenerative component. However, recent experimental evidence in different animal models suggests that synaptic dysfunction is an early and important feature in PD, which might also (partially) explain the evoked potential abnormalities found in VEMP and BAEP testing (apart from the possible concomitant primary axonal and or demyelinating features of the disorder). Future neuropathological brainstem studies in PD should therefore also focus on the myelination of affected brainstem fibre tracts containing Lewy neuritis and coiled bodies, with the primary question whether concomitant demyelination is an important part of the neurodegenerative process. Obviously, more studies are needed concerning the process of synaptic transmission to confirm whether the synaptic dysfunction, which is found in animal PD models, is present in PD patients.

chapter 6
The one-year follow-up results of the case-control study described above in paragraph 2 of this chapter, and which is discussed in chapter 5 in more detail, are discussed in Chapter 6. All 25 healthy volunteers, 14 AP, and 30 PD patients were contacted for a telephone interview one-year following the baseline examinations, and only 1 AP (MSA) patient was lost to follow-up (attrition bias of only 1.4%). All participants were questioned about their fall frequency based upon their fall diaries, and their subjective balance confidence (according to the ABC-16 questionnaire). The results were compared to the baseline results. The primary aim was to determine whether neurovestibular laboratory tests (collected at baseline) can predict future falls in PD and AP patients. Cervical and ocular VEMPs combined with clinical tests for postural imbalance were the best predictors of future fall incidents, both in PD and AP groups, with a sensitivity of 100%. The positive predictive value was 68.2% when only one VEMP test was abnormal, and 83.3% when both VEMP tests were abnormal. The fall frequency at baseline and after one year was significantly higher and the balance confidence scale (ABC-16) was significantly lower in both the PD and AP groups compared to healthy controls. Therefore, abnormal VEMP tests are associated with an increased risk of future fall incidents in PD and AP patients with postural imbalance in the absence of freezing of gait46.

A screening test is only useful in clinical practice when it has both a high sensitivity and a positive predictive value (PPV). Knowing that freezing of gait is a common risk factor for falls47,48, we examined the association between freezing at baseline and the subsequent risk of falls. In our study, assessment of freezing of gait (by qualitative assessment of straight walking in combination with the assessment of rapid clockwise and count-clockwise turning) was the clinical test with the highest PPV (namely 87.5%), but the sensitivity was only 46.9% (as only 7 of the 15 falling patients were detected). This makes testing for freezing of gait by itself inadequate for screening purposes, since half of the potential fallers are missed. Moreover, clinical testing for postural imbalance (using the retropulsion test) had a sensitivity of 100% for detection of falling among PD and AP patients, but the PPV was only 46.9%. So again, this clinical test is, by itself, unsuitable for screening purposes because about half of the patients with abnormal results will not fall in the following year. However, the combination of both cervical and ocular VEMP testing, combined with clinical testing for postural imbalance, seemed to be the optimal combination (yielding a sensitivity of 100% for detecting patients at risk of future falling, with a PPV of 68.2% for one abnormal VEMP test and a PPV of 83.3% when both VEMP tests are abnormal). However, the presence of freezing of gait is such a strong, but insensitive, predictor for future falls in both AP and PD patients that VEMP testing in these patients does not have any additional value (i.e. all patients with freezing of gait also had postural imbalance and abnormal VEMPs). Therefore, I concluded that cervical and ocular VEMP testing can possibly give additional information concerning the future fall risk in selected PD and AP patients (i.e. those patients with postural imbalance in the absence of freezing of gait)46. However, these results, as discussed above, must be interpreted cautiously as the studied groups contained small numbers of patients, with a heterogeneous population in the AP patient group.

One can speculate whether this additional information will aid in guiding future fall prevention therapies as PD and AP patients with postural imbalance already have a high risk for falling. Multiple meta-analyses and Cochrane reviews of physiotherapy in PD reported short-term statistically significant, but clinically modest positive effects concerning balance-related activity performance and improvement of gait. However, there was no significant evidence concerning the reduction of the risk of falling49–52. Moreover, one Cochrane review49 compared the different physical therapies and concluded that a formal qualitative comparison was not possible because of the heterogeneity of the studied techniques, small sample sizes of the studied groups, methodological flaws, and risk of publication bias. Therefore, there is no evidence to choose or advise one form of physical therapy over another, as was concluded by the authors49. High-quality prospective studies with large study groups and long-term follow-up are needed to assess the effect of physiotherapy in PD and AP concerning the reduction of the risk of falling, and to assess which form of physical therapy is superior in achieving this goal. Such studies should follow evidence-based guidelines for physiotherapy strategies in PD, and the interventions should ideally be delivered by therapists experienced in the management of patients with PD.

A recent single-case study described the motor and non-motor effects of repeated caloric vestibular stimulation in a patient with PD. The motor performance and the non-motor scores (i.e. Montreal cognitive assessment, hospital depression scale, and Epworth sleepiness scale) improved significantly during repeated caloric vestibular stimulation and persisted during follow-up 5 months after stimulation. Moreover, these improvements were not seen during the first phase of the study which consisted of sham-stimulation. Moreover, recent studies in PD patients provide preliminary evidence that (stochastic) galvanic vestibular stimulation can possibly significantly reduce postural instability, improve corrective postural responses, improve anterior bending posture, and is, above all, safe for short term use58–62. These studies, however, consist of single case-reports or small case series without further evidence, and should therefore in my opinion be considered as proof-of-concept evidence. Obviously, large scale randomized controlled double-blinded trials are needed in PD and AP patients to demonstrate whether (stochastic) galvanic vestibular or repeated caloric stimulation is able to clinically and statistically significantly reduce fall episodes and/or to improve motor and non-motor symptoms.

At this moment, I do not advise to perform VEMP testing in PD or AP patients as it does not have any treatment consequences. However, vestibular laboratory testing (i.e. VEMP, SVV, and or videonystagmography with caloric and rotatory chair stimulation) might be helpful in elucidating the underlying physiological changes secondary to repeated vestibular stimulation as mentioned earlier.

Highlights of the most important findings
- Abnormal VEMP and SVV laboratory test results can give important clues to the localization of the underlying neurovestibular problem, but are not disease specific.
- Approximately 90% of PD patients and 100% of AP patients have abnormal neurovestibular laboratory test results, with a predominantly central neurovestibular dysfunction profile.
- In PD patients, it is mainly the symptomatic brainstem side that is electrophysiologically affected, compared to healthy volunteers (i.e. the affected brainstem side is ipsilateral to the patient’s resting tremors, rigidity, and bradykinesia at the pontomedullary junction below the level of the vestibular nuclei; and the affected fibre tracts cross over to the contralateral brainstem side above the level of the vestibular nuclei along with the decussation of the medial longitudinal fasciculus).
- PD and AP patients with laboratory signs of neurovestibular dysfunction have a statistically significant increased risk of falling.
- Cervical and ocular VEMP testing in combination with clinical testing for postural imbalance can predict future fall incidents in PD and AP patients (sensitivity of 100%, PPV of 68.2% with one abnormal VEMP test, and a PPV of 83.3% when both VEMP tests are abnormal).