Introduction
Vertigo and dizziness are common complaints, with considerable healthcare and personal impact. Therefore, the assessment of vertigo requires effective examinations which can provide a definitive diagnosis. Examinations of limited diagnostic value which may unnecessarily burden patients and drain healthcare resources should be restricted.
The diagnostic approach to the patient suffering from vertigo always begins with a detailed history and clinical neurotological examination. However, although fundamental and irreplaceable, these are not always enough to establish a diagnosis.
Electronystagmography (ENG) has long been used in the investigation of peripheral and central oculo-vestibular system disorders. However, with the continuous development of more precise and sensitive imaging techniques, the value of conventional laboratory examinations is now being questioned. In clinical practice, magnetic resonance imaging (MRI) is now often regarded as the examination of choice in patients with unclear history and symptoms of vertigo. However, a review of the current, relevant literature reveals that studies which compare the diagnostic value of MRI and ENG in such patients appear to be lacking. There are several case studies which have demonstrated correlations in certain conditions, such as multiple sclerosis, Arnold–Chiari malformation and vestibular migraine,Reference Williams, Roland and Yellin1–Reference Cass, Furman, Ankerstjerne, Balaban, Yetiser and Aydogan3 and a few other studies which have examined MRI findings in patients who had already undergone ENG and were found to present abnormalities indicative of central nervous system (CNS) dysfunction.Reference Stoddart, Baguley, Beynon, Chang and Moffat4–Reference Kinney, Wallace, Ross, Kinney and Hamid6
The aim of the current study was to determine the most effective test battery for the aetiological diagnosis of balance disorders in patients with an inconclusive history and clinical examination regarding the peripheral or central origin of vertigo. The dilemmas regarding such investigation are summarised as follows. Should MRI precede, replace or follow ENG in the diagnostic assessment of such patients? Which is the examination that meets more frequently the need for an aetiological diagnosis? How often do structural abnormalities seen on MRI provide the final aetiological diagnosis in patients in whom the history and clinical examination provide limited indication of peripheral or central vertigo? To what extent do MRI findings verify the indications of ENG-derived indicators for central oculo-vestibular system dysfunctions?
Patients and methods
We conducted a four-year, retrospective study (covering 2001–2005) in order to compare the diagnostic value of ENG and MRI in patients with an inconclusive history and clinical examination regarding the peripheral or central origin of vertigo and unsteadiness.
Patients were usually referred to our ENT department by their physician, neurologist or otolaryngologist. Patients not responding to medical treatment and/or vestibular rehabilitation, and who had persistent or deteriorating dizziness, imbalance, tinnitus or headache for at least two months, were included in the study. Patients who were not able to give valid information about their anamnesis and vertigo or unsteadiness comprised a small percentage of this population. Patients with a history of cancer and increased possibility of CNS metastasis were also included in the study.
All patients underwent a full otorhinolaryngological clinical evaluation, personal history-taking and neurotological assessment, including pure tone audiography, auditory brainstem response (ABR) testing and ENG. Electronystagmography consisted of the following groups of subtests: saccade analysis, gaze fixation and caloric irrigations. Positioning (Dix–Hallpike) testing was also performed in all patients before ENG.Reference Shepard and Katz7 A full ENG examination was in some cases too time-consuming for routine practice within our busy neurotological department. In an attempt to decrease the duration of the test battery, some ENG tests were omitted from our protocol. Saccade testing was included in the protocol because it is thought to provide differentiating information concerning brainstem versus posterior cerebellar vermis involvement. Suggestions for possible frontal or parietal lobe involvement can also be obtained from saccadic testing. Optokinetic stimulation was performed in our department only in selected cases, rather than as a routine examination. This test was omitted from the routine study protocol, because it is considered the least sensitive test, probably due to the combination of both foveal and peripheral retinal stimulation. Similarly, smooth pursuit, which is considered the most sensitive of the ocular-motor tests, was not included in our test battery, because it provides poor lesion site localisation within the multiple vestibulo-cerebellar pathways involved in pursuit generation.Reference Shepard and Katz7 Electronystagmography recordings were performed using Life-Tech Model 3002 equipment (Life-Tech Instruments, Inc., Houston, TX).
Of the clinical signs of a central lesion, vertical nystagmus was the only one diagnosed in this population. In terms of ENG findings, the abnormalities indicative of central lesions which were identified in our series were square waves, abnormal saccades in terms of accuracy and a low fixation index in caloric responses.Reference Stoddart, Baguley, Beynon, Chang and Moffat4 A directional preponderance may be present in both central and peripheral lesions.Reference Shepard and Katz7 Pathological findings indicating possible retrocochlear pathology on ABR testing included: prolonged I–V or I–III delays, an interaural wave V delay (IT5) of more than 0.3 milliseconds, an abnormal V/I amplitude ratio, and disorganised waveforms. All patients were additionally advised to undergo an imaging study, preferably brain and cerebellopontine angle MRI.
Results
We analysed the medical records of 105 patients. One female patient had refused to undergo any imaging studies because of claustrophobia, and two patients had undergone computed tomography (CT) scanning rather than MRI due to metallic implants in their extremities following orthopaedic surgery. Brain and cerebellopontine angle MRI recordings were available in 102 patients. Sixty-one of the patients were women, aged 20 to 82 years (mean = 55.27, standard deviation (SD) = 15.44), and 41 were men, aged 26 to 82 years (mean = 58.00, SD = 13.45). These differences were not statistically significant (t-test, p = 0.36). Tinnitus was reported by 19 patients. In this population, conventional pure tone audiometry revealed sensorineural hearing loss in 23 patients.
In another 46 patients (45 per cent), the presence or absence of spontaneous nystagmus and the clinical interpretation of positioning testing and caloric irrigation results provided the final diagnosis of peripheral system involvement with insufficient CNS compensation. The caloric test provided a measure of unilateral labyrinthine hypofunction and/or directional preponderance and was interpreted as indicating peripheral vestibular dysfunction in 41 patients. Benign paroxysmal positional vertigo (BPPV) was identified in another five patients. Pathological central ENG findings were identified in seven patients (6 per cent). A low fixation index and a saccade test with abnormal accuracy were recorded in two patients; multiple sclerosis was confirmed by MRI in one of these patients. A low fixation index was additionally identified in four patients and was interpreted as a possible indication of insufficient CNS blood supply.Reference Hood and Korres8 Additionally, square waves were obtained from one further patient, but MRI failed to identify any CNS dysfunction. Among the seven patients with abnormal ENG findings indicating a lesion of central origin, only one patient had congruent structural abnormalities on MRI (Tables I and II).
Table I Patients with ENG abnormalities indicating lesions of central origin
* Patient 3 also had abnormal findings on magnetic resonance imaging (MRI). Pt = patient; yrs = years; ENG = electronystagmography; PTA = pure tone audiography; ABR = auditory brainstem response testing; M = male; F = female; mod = moderate; SNHL = sensorineural hearing loss; mid = middle
Table II Patients with abnormal MRI scans
* Patient 1 had findings indicative of CNS lesions in both ENG and MRI. MRI = magnetic resonance imaging; pt = patient; yrs = years; ENG = electronystagmography; PTA = pure tone audiometry; ABR = auditory brainstem response testing; F = female; CP = cerebellopontine; SNHL = sensorineural hearing loss
Structural lesions identified on MRI included: acoustic neuroma (two patients), demyelinating disorder (one patient) and meningioma of the cerebellopontine angle (one patient). Therefore, ENG contributed to the diagnostic procedure of 53/102 patients (52 per cent), compared with four of 102 patients (3.9 per cent) for MRI (Table III). Although MRI demonstrated a high specificity in diagnosing central vestibular dysfunction, its contribution to the aetiological diagnosis of this population was small (four of 102 patients). Electronystagmography, on the other hand, demonstrated a significantly higher diagnostic effectiveness, as it indicated both peripheral (46/102 patients) and central (seven of 102 patients) vestibular dysfunction. Electronystagmography was normal in one patient with acoustic neuroma, while caloric irrigation revealed unilateral vestibular hypofunction in one patient with acoustic neuroma and also in another patient with meningioma of the cerebellopontine angle. Auditory brainstem responses revealed a prolonged I–V delay and an interaural wave V delay in all three patients with lesions of the internal auditory canal and cerebellopontine angle. Waveform disorganisation was found in the ABR examination of the patient with multiple sclerosis. False positive ABR findings were obtained from eight patients; atypical waveforms and an interaural wave V delay indicated CNS lesions, but the patients' MRI examinations were normal.
Table III Contribution of ENG and MRI to aetiological diagnosis of vertigo and unsteadiness*
* In 102 patients with inconclusive history and clinical examination regarding the peripheral or central origin of dysfunction. ENG = electronystagmography; MRI = magnetic resonance imaging
Discussion
The one-year prevalence of vertigo (as a symptom) in adults aged 18–79 years has been estimated at 7.1 per cent for women and 2.6 per cent for men, according to a German population survey using validated neurotological interviews.Reference Neuhauser, von Brevern, Radtke, Lezius, Feldmann and Ziese9 Electronystagmography has for many years been considered the ‘gold standard’ in the diagnosis of vertigo, especially in cases in which thorough history and clinical neurological and otological examination are inconclusive. However, nowadays several healthcare institutions advocate direct or sole use of MRI in such cases. The increasing demand for aetiological diagnosis and specialised treatment, on the one hand, and the need for prudent use of healthcare budgets, on the other, encourages healthcare professionals constantly to critically appraise the actual diagnostic contribution of laboratory examinations, and to redesign test batteries to improve effectiveness.
This study was conducted in order to evaluate the contribution of both ENG and MRI to the final diagnosis of patients with vertigo, and thus to designate the most cost-effective test battery. Our population and methods differ from those of the few previous, relevant studies in more than one respect.
Firstly, MRI was performed in all patients with a diagnostic dilemma, not just those with previous abnormal ENG findings suggestive of a central lesion. This methodological difference may be crucial in enabling identification of patients with central lesions who may otherwise escape diagnosis because they have undergone only ENG.Reference Stoddart, Baguley, Beynon, Chang and Moffat4, Reference Casselman, Kuhweide, Dahaene, Ampe and Devlies5 In addition, through this procedure, one may estimate the number of patients with functional abnormalities of the vestibulo-ocular system which have escaped MRI diagnosis, either because they are peripheral or because they are not presented as detectable, concise, structural lesions.
Secondly, being significantly more recent than previous reports, this study may have used more modern MRI techniques.
Magnetic resonance imaging and ENG were thus compared regarding their efficiency in contributing to the final diagnosis. To make this point more obvious, we divided our population into two subgroups: (1) patients who reached a final aetiological diagnosis after MRI; and (2) patients reaching a final diagnosis through ENG.
Vertigo patients diagnosed via magnetic resonance imaging
Magnetic resonance imaging contributed to the final diagnoses of only four out of 102 patients (namely, acoustic neuroma in two patients, multiple sclerosis in one patient and meningioma of the cerebellopontine angle in one patient). The literature lacks similar studies investigating patients with positive MRI findings who additionally undergo ENG. In Shepard's study,Reference Shepard and Katz7 in which 134 patients with clearly defined brainstem or cerebellar lesions diagnosed on MRI (namely, multisystem atrophy, olivopontocerebellar atrophy, multiple sclerosis or Friedreich's ataxia) were examined with ENG, a combination of gaze fixation, smooth pursuit and saccadic testing provided evidence for central vestibulo-ocular pathway dysfunction in 87 per cent. In this study, the diagnostic capacity of ENG was reinforced by the performance of the smooth pursuit examination.
In the current study, abnormal ENG findings were present in three of the four cases with pathological MRI findings, while ABR recordings were abnormal in all four cases (Table II). However, in our experience, normal ENG and ABR results cannot always exclude the possibility of abnormal MRI findings. Magnetic resonance imaging is indisputably the examination of choice for the identification of a CNS lesion. However, the advent of MRI in clinical practice does not alter the indications for ENG. The fact that, in the current study, in a population referred due to possible central vestibular disorders, aetiological diagnosis was reached through MRI in only four of 102 patients needs to be highlighted and taken into consideration in routine clinical practice. On the other hand, as a result of their MRI, the remaining 98 patients were reassured that they did not have a CNS lesion, which is very significant for those patients' recovery and quality of life. Nevertheless, the absence of CNS pathology is not in itself a diagnosis, and after the initial relief patients are bound to enquire about the cause underlying their alarming symptoms of vertigo and imbalance.
Vertigo patients diagnosed via electronystagmography
Eye movement recordings observed during ENG provide valuable information on the function of both the peripheral vestibular system and the ocular-motor pathways in the brainstem and cerebellum which are required for the production of the vestibulo-ocular reflex.Reference Shepard and Katz7, Reference Korres, Balatsouras and Ferekidis10 Benign paroxysmal positional vertigo was identified in the current study in five patients. Positional (Dix–Hallpike) testing can provide considerable evidence regarding both the site of lithiasis (i.e. involving the horizontal, posterior or anterior semicircular canal) and (when performed by specialist neurotologists) the particular characteristics of lithiasis (e.g. cupulolithiasis, canalithiasis or canalith jam).Reference Epley11 This information is crucial, not only because it represents a full, detailed diagnosis, but most importantly because it will provide the basis for aetiological treatment through canalith repositioning manoeuvres.Reference Korres, Balatsouras and Ferekidis10 Information concerning unilateral labyrinthine hypofunction and/or directional preponderance (n = 41 out of 102, in the current study) is very important for vestibular rehabilitation.
Saccadic hypermetria is the hallmark of cerebellar disease and brainstem lesions, and failure of suppression by optic fixation may provide evidence of a floccular lesion.Reference Hood and Korres8 In combination with the other ENG tests, saccadic analysis can provide a criterion for the diagnosis of brainstem, posterior cerebellar vermis, frontal or parietal lesions.Reference Shepard and Katz7 In the current study, gaze fixation indicated a general suggestion of brainstem or cerebellar dysfunction in seven patients. One of these patient's MRI scans revealed demyelinated lesions. In the other six patients, the functional abnormalities detected by the ENG were not confirmed by congruent structural lesions in the MRI. Stoddart et al. reported that, of 23 patients undergoing MRI after having demonstrated abnormal ENG recordings of central origin, only seven (30 per cent) had pathological MRI findings.Reference Stoddart, Baguley, Beynon, Chang and Moffat4 Similarly, Kinney et al. calculated the diagnostic sensitivity of MRI to be only 30 per cent, using ENG as the gold standard.Reference Kinney, Wallace, Ross, Kinney and Hamid6 The hypothesis given for this disparity was either that the 70 per cent of cases with normal MRI had changes which were functional rather than structural, or that the structural changes were outside the resolution capacities of the MRI. Higher resolution MRI or functional MRI may in the future alter these results in favour of MRI. There are interesting reports of patients with intralabyrinthine schwannoma, ischaemia of the membranous labyrinth, intralabyrinthine haemorrhage and viral labyrinthitis diagnosed using more modern MRI scans enhanced by various special techniques.Reference Mafee12, Reference Bruzzone, Grisoli, De Simone and Regna-Gladin13 Unfortunately, such technology is currently available in only a limited number of specialised hospitals.
• Many clinicians tend to underestimate the contribution of electronystagmography to the aetiological diagnosis of vertigo and unsteadiness
• Magnetic resonance imaging (MRI) is indisputably the examination of choice for the diagnosis of central nervous system lesions
• However, in a population prone to central vestibular disorders, MRI failed to contribute to the aetiological diagnosis of 98/102 patients
• Electronystagmography, on the other hand, contributed to the aetiological diagnosis of 53/102 patients; thus, its diagnostic indications, recently underestimated, should not be affected by introduction of MRI into routine clinical practice
In the current study, the number of patients obtaining a final diagnosis though ENG was considerably higher than that diagnosed through MRI (53 versus four, respectively). A patient complaining of vertigo or dizziness is much more likely to have peripheral rather than central dysfunction. Brandt et al. found that central, vestibular vertigo was diagnosed in 13.1 per cent (n = 699) of 5353 patients referred to a neurological dizziness unit for vertigo or dizziness.Reference Brandt and Strupp14 Similarly, Stoddart et al. reported identification of a central vestibular disorder in 34 patients (4.7 per cent), out of a total of 725 cases.Reference Stoddart, Baguley, Beynon, Chang and Moffat4 By the initial use of ENG, Bakr and Saleh identified central lesions in 21/260 (8.1 per cent) patients, and mixed central and peripheral disorders in 26 (10 per cent).Reference Bakr and Saleh15 With the subsequent addition of MRI, CT, and vertebral and carotid artery Doppler imaging, the number of CNS lesions diagnosed was increased to 60/260 patients (23.1 per cent), with most of this increase being due to vascular disorders. Abnormal ENG findings are more frequent in patients with BPPV of the horizontal canal–superior division of the vestibular nerve (64.2 per cent), compared with patients with BPPV of the vertical canals (40.9 per cent).Reference Korres, Balatsouras and Ferekidis10
One important disadvantage of ENG is that it provides no specific information about the exact site or morphology of the lesion in the cerebellum, the brainstem or, in some cases, the cortex. Patients with indications of a dysfunction in the oculo-vestibular system have to be referred for MRI in order to precisely visualise the site and the nature and dimensions of the lesion. However, false negative results do not occur only for ENG; MRI seems to have false negative results also, although documentation of such cases is much more difficult.Reference Stoddart, Baguley, Beynon, Chang and Moffat4, Reference Kinney, Wallace, Ross, Kinney and Hamid6
Conclusion
Electronystagmography remains an irreplaceable tool in the aetiological diagnosis of patients complaining of vertigo, even in the post-MRI era. Magnetic resonance imaging is indisputably the examination of choice for identifying the morphology and topography of CNS lesions. However, its overall contribution to the aetiological diagnosis of such patients is limited. The current study found that, in patients with an increased probability of CNS dysfunction according to history and clinical examination findings, the actual percentage with vertigo and unsteadiness of central origin, confirmed by MRI, was still very small (3.9 per cent). If the diagnostic procedure is interrupted at this point, a large majority of patients remain with practically no information on the aetiology of their symptoms. The current study found that the aetiological diagnosis of vertigo and imbalance was still achieved through neurotological testing, including ENG, in a considerably larger number of patients than those diagnosed via MRI (53 versus four, respectively). As a result of these observations, we may need to re-evaluate the diagnostic importance of ENG, and restore it to its rightful place in the routine diagnostic assessment of vertigo and unsteadiness.
