Relation of infarction location and volume to vertigo in vertebrobasilar stroke

Abstract Objective Vertigo is a common presentation of vertebrobasilar stroke. Anecdotal reports have shown that vertigo occurs more often in multiple than in single brainstem or cerebellar infarctions. We examined the relation between the location and volume of infarction and vertigo in patients with vertebrobasilar stroke. Methods Consecutive patients with vertebrobasilar stroke were prospectively recruited. The infarction location and volume were assessed in the diffusion‐weighted magnetic resonance imaging. Results Fifty‐nine patients were included, 32 (54.2%) with vertigo and 27 (45.8%) without vertigo. The infarction volume did not correlate with National Institute of Health Stroke Scale (NIHSS) score on admission (Spearman ρ = .077, p = .56) but correlated with modified Rankin Scale (ρ = .37, p = .004) on discharge. In the vertigo group, the proportion of men was lower (53.1% vs. 77.8%, p = .049), fewer patients had focal neurological deficits (65.6% vs. 96.3%, p = .004), patients tended to present later (median [IQR] was 7.5 [4–46] vs. 4 [2–12] hours, p = .052), numerically fewer patients received intravenous thrombolysis (15.6% vs. 37%, p = .06), and the total infarction volume was larger (5.6 vs. 0.42 cm3, p = .008) than in nonvertigo group. In multivariate logistic regression, infarction location either in the cerebellum or in the dorsal brainstem (odds ratio [OR] 16.97, 95% CI 3.1–92.95, p = .001) and a total infarction volume of >0.48 cm3 (OR 4.4, 95% CI 1.05–18.58, p = .043) were related to vertigo. In another multivariate logistic regression, after adjusting for age, sex, intravenous thrombolysis, serum level of white blood cells, and atrial fibrillation, vertigo independently predicted a total infarction volume of >0.48 cm3 (OR 5.75, 95% CI 1.43–23.08, p = .01). Conclusion Infarction location in the cerebellum and/or dorsal brainstem is an independent predictor of vertigo. Furthermore, larger infarction volume in these structures is associated with vertigo. A considerable proportion of patients with vascular vertigo present without focal neurological deficits posing a diagnostic challenge. National Institute of Health Stroke Scale is not sensitive for vertebrobasilar stroke.


| INTRODUC TI ON
Dizziness is a major public health problem and an independent predictor of increased mortality (Corrales & Bhattacharyya, 2016). Dizziness and vertigo are more common among women and elderly population (Neuhauser et al., 2008;Rieger et al., 2014).
Affected patients do not usually receive the adequate medical attention and are more likely to consult a general practitioner or an internist rather than a neurologist or an ear, nose, and throat (ENT) specialist (Neuhauser et al., 2008). Stroke is the underlying etiology in 17%-25%, and 4% of patients presenting with acute onset isolated vertigo (Norrving, Magnusson, & Holtas, 1995;Zuo et al., 2018) and dizziness (Navi et al., 2012), respectively.
Of note, 18.9% of patients with vertebrobasilar stroke suffer from vertigo in comparison with only 1.7% of those with stroke in the anterior circulation (Tao et al., 2012). Lesions affecting the following structures are related to the development of vascular vertigo: vestibular nuclei in the dorsolateral portion of the rostral medulla, nucleus prepositus hypoglossi in the dorsal brainstem, dorsal insular cortex as well as cerebellar tonsil, flocculus, nodulus, and inferior cerebellar peduncles (Kattah, Talkad, Wang, Hsieh, & Newman-Toker, 2009;Kerber, Brown, Lisabeth, Smith, & Morgenstern, 2006;Kim, Kim, & Kim, 2017;Neuhauser et al., 2008;Rieger et al., 2014;Saber Tehrani et al., 2014). However, one study showed that more than one-fifth of patients with isolated lateral medullary infarction have no vertigo and patients with lateral medullary infarction plus additional extralateral medullary infarctions are more commonly associated with vertigo than those with isolated lateral medullary infarctions (Kang et al., 2018). Another study reported vertigo in all patients with multiple unilateral pontine infarctions and in less than half of those with a single pontine infarction (Kumral, Bayulkem, & Evyapan, 2002).
Moreover, small lesions ≤10 mm in axial diameter were detected in only 14% of patients with acute vascular vertigo (Saber Tehrani et al., 2014). We aimed to examine the relationship of infarction volume and location to vertigo in patients diagnosed with a vertebrobasilar stroke.

| Study design
Consecutive patients admitted to the Department of Neurology (University Hospital of Würzburg) with the diagnosis of vertebrobasilar stroke were prospectively recruited between February and October 2018. Patients were included only if they could communicate the presence or absence of vertigo, and a magnetic resonance imaging (MRI) could be done within 4 days of admission and showed brain infarction. National Institute of Health Stroke Scale (NIHSS) on admission and modified Rankin Scale (mRS) and Barthel index on discharge were used as clinical scales. Patients were stratified into two groups: group 1 with vertigo (vertigo +) and group 2 with no vertigo (vertigo −). Vertigo was defined according to the definition of the Bárány Society as follows: "the feeling of self-motion, when no selfmotion is occurring" (Bisdorff, Von Brevern, Lempert, & Newman-Toker, 2009).

| MRI imaging
Magnetic resonance imaging scanners with a field strength of 3-Tesla were used according to our standardized stroke acquisition protocols with a slice thickness of 5 mm and an interslice gap of 0.5 mm. The infarction location and volume were assessed in the strong (b = 1,000) diffusion-weighted images (DWI) in our picture archiving and communication system (PACS) by a single investigator (AME), who was nonblinded to the clinical data. The infarction was manually delineated in our PACS to obtain the infarction area in each slice separately. The latter was multiplied by the slice thickness and interslice gap. Finally, the sum of all slices was calculated to obtain the infarction volume (Inaba et al., 2017;Jung, Kwon, Lee, & Kang, 2010).

| Statistical analyses
Qualitative data were expressed in absolute values and percentages, while quantitative data were expressed using median and range. To check for normality, we used Q-Q plot, histogram, and the We tested the fitness of this model using a Hosmer-Lemeshow K E Y W O R D S brain stem, cerebellum, infarction volume, stroke, vertebrobasilar insufficiency, vertigo "goodness-of-fit" test. Data were analyzed in SPSS software package version 25 (SPSS). p-values < .05 were considered statistically significant.

| Baseline characteristics
Fifty-nine patients were included in the study. Baseline characteristics are illustrated in Table 1. There was a statistically significant less proportion of men among vertigo (+) patients in comparison with vertigo (−) patients (53.1% vs. 77.8%, respectively, p = .049).

| Relation between infarction location and volume and the presence or absence of vertigo
An infarction location either in the cerebellum or in the dorsal brainstem was significantly more common among vertigo (+) patients in comparison with vertigo (−) patients (90.6% vs. 40.7%, respectively, p < .001). The total infarction volume was much larger among vertigo (+) patients with a median of 5.6 cm 3 versus 0.42 cm 3 among vertigo (−) patients (p = .008, Figure 1b Table 2. The Hosmer-Lemeshow "goodness-of-fit" test showed a nonsignificant difference between the observed and expected results with p = .88. In another multivariate logistic regression, after adjusting for age, sex, intravenous thrombolysis, serum level of white blood cells, and atrial fibrillation, vertigo independently predicted a total infarction volume >0.48 cm 3 (OR 5.75, 95% CI 1.43-23.08, p = .01) as shown in Table 3 (Hosmer-Lemeshow "goodness-of-fit" test showed a p-value of .2).

| Relation between the clinical scales and the infarction volume
The total infarction volume did not correlate with NIHSS score on admission (ρ = .077, p = .56) but correlated with mRS (ρ = .37, p = .004) as well as Barthel index on discharge (ρ = .33, p = .011).

| Relation of infarction location and volume to vertigo
In the vertigo (+) group, the proportion of men was lower, fewer patients had focal neurological deficits, and the total volume of infarction was larger than in nonvertigo group. Similar to previous studies (Kattah et al., 2009;Kerber et al., 2006;Kim et al., 2017;Neuhauser et al., 2008;Rieger et al., 2014;Saber Tehrani et al., 2014), we found that an infarction location in the cerebellum or dorsal brainstem was significantly related to the development of vertigo. Moreover, the infarction volume, especially for cerebellar infarction, was related to the occurrence of vertigo among our patients. To our knowledge, no similar reports exist in literature. We speculate that larger infarcts, especially in the cerebellum, mediate the development of vascular vertigo through affection of several brain structures and interconnections. On the other side, the development of focal neurological signs results from small strategically located brain infarctions. This might be comparable to patients with small capsular infarctions having moderate or severe hemiparesis, in spite of the small infarction volume. Similar to our results, previous authors identified small lesions ≤10 mm in axial diameter in only 14% of patients with vascular vertigo (Saber Tehrani et al., 2014). The authors found focal neurological signs among 27% of their patients with small lesions, whereby the inferior cerebellar peduncle and the lateral medulla were most often involved. Another study reported vertigo in 92% of patients with lateral medullary infarction plus additional extralateral medullary lesions in comparison with 78.9% of patients with pure lateral medullary infarction (p = .008), which is in line with our findings (Kang et al., 2018). Additionally, in one series vertigo was found among all patients with multiple unilateral pontine lesions and in less half of those with a single pontine lesion (Kumral et al., 2002).

| Vertigo may mask an underlying vertebrobasilar stroke
In our vertigo (+) group, fewer patients had focal neurological deficits, and hence, the patients tended to present later than the vertigo (−) group. Therefore, numerically fewer patients in the vertigo (+) group received intravenous thrombolysis in comparison with the vertigo (−) group (15.6% vs. 37%, p = .06). In concordance with our findings, other authors showed that 37% of patients with vertebrobasilar stroke versus 16% of patients with stroke in anterior circulation were misdiagnosed (p < .001) and the presence of focal neurological signs helped to pave the way for the accurate diagnosis (Arch et al., 2016). In a previous cohort, only 42% of patients with vertebrobasilar stroke presenting with vertigo had obvious neurological signs (Kattah et al., 2009). Vertigo was frequently found to be related to misdiagnosis of ischemic stroke (Arch et al., 2016;Newman-Toker, 2016;Savitz, Caplan, & Edlow, 2007). Atrial fibrillation, n (%) 6 (22.2) 9 (28.1) .6
b In patients with wake-up stroke or those found with stroke, the time, when the patient was found with stroke, was used to denote the time of symptom onset. c The patient presented with transient episodes of blurring vision in both eyes. In addition, the patient had nausea and vomiting without vertigo. MRI brain showed a small right cerebellar infarction.

TA B L E 2 Predictors of vertigo in the binary logistic regression models
presenting with severe vertigo may have an NIHSS score of 0 but a mRS score of 3, if he requires some help for the activities of daily living. In our cohort, the infarction volume did not correlate with NIHSS score on admission but correlated with mRS and Barthel index on discharge.

| Patient sex, vertigo, and brain infarction
The incidence of stroke, both in the anterior or posterior circulation, in men is around 32% higher than in women (Giroud et al., 2017). Furthermore, the odds ratio for the development of vertebrobasilar stroke rather than stroke in the anterior circulation was even found to be higher in men (Subramanian et al., 2009).
Contrarily, men represent around one-third of the vertigo population (Neuhauser et al., 2008;Rieger et al., 2014). In our cohort, we

| One-fifth of our patients met the old TIA definition of the WHO and at the same time the modern AHA/ASA stroke definition
Historically, the World Health Organization (WHO) defined transient ischemic attack (TIA) as a transient focal neurological deficit lasting <24 hr (Albers et al., 2002). In 2009, the American Heart Association/American Stroke Association (AHA/ASA) revised this definition, stating that TIA is "a transient episode of neurological dysfunction caused by focal brain, spinal cord, or retinal ischemia, without acute infarction" (Easton et al., 2009). The implementation of this AHA/ASA definition "moved" a significant proportion of patients from the TIA to the stroke category (Kvistad et al., 2013). In the current work, 12/59 (20.3%) of the patients had no focal neurological deficits, yet they had a brain infarction on MRI.

| Clinical importance
Our study has several clinical implications.

| Study limitations
There are limitations of this study. The nonrandomized nature of this single-center cohort should be kept in mind before a conclusion can be drawn from our results. We recommend the conduction of large randomized multicenter studies in this regard. Another limitation of this study is that the infarction volume and location were assessed by a single nonblinded investigator (AME). However, the assessment methods used in this study are more or less objective.

| CON CLUS ION
Cerebellar or dorsal brainstem infarctions are strong predictors of vertigo in vertebrobasilar stroke. Larger infarction volume in these structures may also be related to vascular vertigo. Vertigo without any focal neurological deficit is not uncommon among patients with vertebrobasilar stroke and represents a diagnostic challenge.
National institute of health stroke scale has limited value in assessing vertebrobasilar stroke compared with anterior circulation stroke.

ACK N OWLED G M ENTS
This publication was funded by the German Research Foundation (DFG) and the University of Wuerzburg in the funding program Open Access Publishing.

CO N FLI C T O F I NTE R E S T
None.

AUTH O R S' CO NTR I B UTI O N S
All authors made a substantial contribution to the conception, design, and revision of the work. AME and/or FF examined all the patients. AME collected the data, performed the measurements and the statistical analysis, and wrote the first draft. MA, JV, FF, and DE supervised the work, provided consultations, and revised the manuscript. All authors were involved in the final approval of the final version to be published.

E TH I C A L A PPROVA L
Data collected within routine clinical care were used. The study was approved from the University Hospital Würzburg Ethics Committee (AZ 223/16). The patients or their next available kin signed an informed consent prior to the inclusion in the study.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.