Abstract
70124-5_summary.png)
Optokinetic nystagmus (OKN) testing is one method to determine central vestibular dysfunction. OKN may be elicited by partial visual field stimulation with a light bar (OKN-ENG) or by full visual field stimulation with rotating stripes in a rotational chair test booth (OKN-RVT). OKN-ENG and OKN-RVT were elicited in 36 healthy subjects and 48 patients with known peripheral or central vestibular disorders. Abnormal test results suggested central pathology in 29 of 36 healthy subjects with OKN-ENG versus 1 of 36 with OKN-RVT. Twenty-eight of 33 patients with peripheral pathology demonstrated abnormal OKN-ENG findings, whereas 4 of 33 had abnormal OKN-RVT results. Thirteen of 15 patients with central vestibular disorders had abnormal OKN-ENG, whereas 7 of 15 had abnormal OKN-RVT. Sensitivity and specificity of OKN-ENG were 86.7% and 17.4% versus 46.7% and 92.7%, respectively, for OKN-RVT. These findings were statistically significant (P < 0.00001). OKN elicited by full visual field stimulation (OKN-RVT) is a more accurate indicator of central disease than OKN elicited by partial visual field stimulation (OKN-ENG). The use of OKN-ENG to identify central vestibular dysfunction is questionable. (Otolaryngol Head Neck Surg 1999;121:52-6.)
Optokinetic nystagmus (OKN), the phylogenetic precursor to the smooth pursuit system, 1 interacts with the smooth pursuit, vestibular and saccadic systems to ensure target fixation, visual acuity, and spatial orientation. OKN is a reflexive response elicited by visual stimuli that relies on intact visual-ocular reflexes. A moving visual surround in the periphery of the visual field elicits the reflex, 2 although inclusion of the fovea enhances the response. 3 When a subject looks at a moving surround, the surround induces a slow conjugate drift of the eyes in the same direction as the moving stimulus, known as the slow component of OKN. As the visual target moves to the periphery of the visual field, there is a saccade-like conjugate movement of the eyes back to the center position. This is known as the fast component of nystagmus, and according to conventional terminology, the direction of this fast component determines the direction of OKN. 4 Although OKN has been described as a combination of smooth pursuit (slow component) and saccade (fast component) eye movements, 5 OKN is a different system than either the pursuit or saccade systems and shares anatomic features with each. 6
The central vestibular system is composed of neural pathways within the brain and brain stem, such as the vestibular nuclei, vestibulo-ocular nuclei, vestibulo-ocular connections, visual-ocular connections, vestibulospinal connections, and vestibulocerebellar connections. The peripheral vestibular system is composed of the components extrinsic to the brain and brain stem, such as the vestibular end organs within the inner ear and the vestibular nerve. OKN is processed by 2 different pathways: cortical and subcortical. 7 The cortical pathway extends from the foveal retina to the lateral geniculate body, occipital lobe, cerebellar flocculus, pontine paramedian reticular formation, and oculomotor neurons. The subcortical pathway is primarily elicited by having the patient stare directly ahead at the moving targets during OKN testing (stare nystagmus) rather than having the patient fixate on an individual target. 7 The subcortical pathway travels from the peripheral retina to the contralateral accessory optic system by way of the accessory optic tract and to the nucleus of the optic tract, nucleus reticularis tegmentum pontis, and vestibular nuclei. From the vestibular nuclei, the signals pass to the medial longitudinal fasciculus, tract of Deiters, and oculomotor nuclei, where they cause an OKN response. 7
OKN testing has been used as a means of determining the nature of vestibular pathology. It is generally accepted that OKN abnormalities reflect central vestibular pathology rather than peripheral vestibular pathology. 8 Peripheral vestibular pathology, in theory, should not affect OKN, with the exception of acute unilateral lesions of the inner ear or vestibular nerve. 9 – 12 OKN recordings should therefore provide a basis by which to classify causes of vertigo as either central or peripheral.
OKN-ENG and OKN-RVT results in patients with central disorders
70124-5-table1.png)
-, Normal test result; +, abnormal test result.
Different methods of eliciting OKN vary in the directions given to the patient, extent of visual surround provided, and type of visual target used. OKN testing has been traditionally included as part of the electronys-tagmography (ENG) battery. 2 In this method of testing, referred to hereafter as OKN-ENG, OKN is elicited by a light bar that can display a moving pattern of light targets at various speeds. The rotary chair vestibular test (RVT) provides another means of eliciting and measuring OKN. In this test, hereafter referred to as OKN-RVT, the rotary chair chamber is an enclosed, light-impermeable area that can provide full visual field stimulation. It has been suggested that devices such as light bars, because of the partial visual field stimulation, do not elicit true OKN, but instead elicit a combination of OKN and smooth pursuit. 2 Because OKN-RVT provides a full visual surround, it might provide a better stimulus for eliciting OKN. This study was designed to determine whether a difference exists between the OKN elicited by the use of the partial visual field stimulus in ENG and that elicited by the full visual field stimulus of RVT, by comparing the results in healthy subjects and patients with vestibular disorders.
METHODS AND MATERIAL
This study retrospectively identified 48 subjects with vestibular pathology and recruited 36 healthy subjects who were seen at the Balance Disorder Center at the Yale University School of Medicine, Section of Otolaryngology. The 36 subjects were enrolled in an investigational protocol that had been reviewed and approved by the institutional review board. Informed consent was obtained for all subjects before testing. Of the 48 patients with vestibular pathology, 15 had central vestibular pathology, and 33 had peripheral vestibular pathology alone. Six patients with both central and peripheral vestibular pathology were included in the group of those with central pathology because all central lesions should be detected regardless of additional peripheral pathology. Any patients with dizziness of unclear or nonvestibular origin or disconjugate gaze were excluded. Healthy subjects were screened for any history of vertigo or symptoms of vestibular pathology.
Age, diagnosis, OKN findings by ENG, and OKN findings by RVT were recorded for all subjects. The diagnoses were made by the senior author (J.F.K.) on the basis of clinical examination, history, imaging studies, surgical pathology reports, and vestibular test battery results excluding OKN testing. The diagnosis of central or peripheral vestibular dysfunction was made in symptomatic patients who displayed abnormalities of the vestibular test battery other than OKN testing. Any ambiguous or conflicting diagnoses were excluded from the sample.
An audiologist specializing in vestibular testing administered the complete vestibular battery (ENG, RVT, and dynamic platform posturography) to all patients. This test battery included testing for OKN activity with both OKN-ENG and OKN-RVT methods. All tests were performed with the ICS ENG Mastr, ICS ENG Chartr, and ICS Rotary Chair testing apparatuses. Subjects all received standard instructions regarding test administration and preparation for testing by avoiding vestibular suppressants for at least 72 hours. Testing devices were calibrated for each patient. Patients were allowed to rest between tests according to their comfort level. Administration of OKN-RVT and OKN-ENG was random and occurred within the overall test battery.
OKN-ENG and OKN-RVT results in patients with peripheral disorders
70124-5-table2.png)
-, Normal test result; +, abnormal test result.
OKN-ENG and OKN-RVT results in healthy subjects
70124-5-table3.png)
-, Normal test result; +, abnormal test result.
During the OKN-ENG test, subjects were seated in a dark, enclosed room 122 ± 15 cm from a wall-mounted light bar. Bitemporal electrodes were placed to record eye movement. Stimuli consisted of 6 red circular lights, each less than 0.5 arc degrees in diameter, which traveled 16 degrees to the right and then to the left at target velocities of 20 degrees/second and 40 degrees/second. Patients were instructed to keep their heads perfectly still and look directly at the center of the light bar, fixating on each light only as it passed through the center of the bar (stare nystagmus, the recommended technique for eliciting OKN 5 ). During the examination subjects received verbal feedback and reinstruction as needed by the audiologist.
During the OKN-RVT test, subjects sat in a chair situated in a light-proof booth. Bitemporal electrodes were placed on the subject. Full visual field stimuli consisted of 5-cm-wide vertical light strips spaced 18 cm apart, which were projected from a centrally placed projector. The light strips extended from the ceiling of the booth to the floor. During testing, the strips were rotated at target velocities of 20 degrees/second and 40 degrees/second to the right and then the left. The subjects received the same instructions as during the OKN-ENG tests to elicit stare nystagmus and were monitored by an infrared camera. The subjects received verbal commands and feedback from the audiologist during the test.
Raw test data and computer-generated results were reviewed by the audiologist and senior author (J.F.K.). A value of less than 80% gain ratio (defined as the ratio of slow phase component eye velocity in degrees per second to target velocity in degrees per second) during either the slow or fast velocity OKN tests was considered an abnormal response. In addition, an asymmetry of 20 degrees/second or more between responses to rightward and leftward stimuli was considered an abnormal response. This was in accordance with the conventional parameters used at this institution as well as with guidelines established in the literature. 4
RESULTS
All patients were tested by OKN-ENG and OKN-RVT. Of the 15 patients with central pathology, 13 had positive OKN-ENG findings, and 7 had positive OKN-RVT findings (Table 1). Of the 33 patients with peripheral vestibular pathology, 28 had positive OKN-ENG findings, and 4 had positive OKN-RVT findings (Table 2). Thirty-six healthy subjects without vestibular pathology were studied. Of these 36 subjects, 29 had positive OKN-ENG findings, and only 1 had positive OKN-RVT findings (Table 3).
An analysis of sensitivity and specificity including all 3 study groups showed that OKN-ENG testing had an 86.7% sensitivity for detecting central vestibular pathology (13 positive tests in 15 patients with central disease). However, it also had a 17.4% specificity for detecting patients without disease (12 true-negative results in 69 patients without central disease). The OKN-RVT method showed a sensitivity of 46.7% (7 positive tests in 15 patients with central disease) and a specificity of 92.7% (64 true-negative results of 69 patients without central disease). These values translate into a positive predictive value and negative predictive value for OKN-ENG of 18.5% (13 true-positive results of 70 total positive results) and 85.7% (12 true-negative results of 14 total negative results), respectively. The OKN-RVT method had a positive predictive value of 58.3% (7 true-positive results of 12 total positive results) and negative predictive value of 88.9% (64 true-negative results of 72 total negative results).
Statistical analysis of these results was done with McNemar's χ2 test. This test revealed that the OKN-RVT and OKN-ENG results differed significantly to a P value of < 0.00001. This indicates that the differences between results obtained by each method are statistically relevant. The effect of age on test results was evaluated with 1-way analysis of variance tests. The following test groups were compared for any age-related effects on their test results: (1) patients with central pathology versus healthy subjects; (2) patients with central pathology versus those with peripheral pathology; (3) patients with vertigo versus healthy subjects; (4) those testing positive by OKN-ENG versus those with negative by OKN-ENG; (5) those testing positive by OKN-RVT versus those testing negative by OKN-RVT; and (6) those with no tests positive versus those with 1 test positive versus those with 2 tests positive. The 1-way analysis of variance revealed that age did not bias the results of any test modalities used or the composition of the study populations.
DISCUSSION
OKN testing is frequently used to evaluate patients with vertigo. This usage has been based on its presumptive accuracy in separating vestibular disorders of central origin from those of peripheral origin. As the results of this study indicate, however, OKN is not necessarily a reliable means of determining such a distinction. This study has shown that the method by which OKN is elicited significantly affects the gain ratio achieved. According to our results, the ENG method of eliciting OKN with the stare technique provided little help in identifying patients with central vestibular pathology. In healthy individuals the ENG method of OKN testing produced nearly all false-positive results. Furthermore, almost all patients tested positive regardless of whether they had central or peripheral pathology. Hence, in both healthy subjects and patients with vestibular pathology, OKN testing by ENG was not diagnostically significant.
The RVT method of eliciting OKN, in contrast, appears to provide a more useful method of identifying central vestibular pathology. The main advantage is demonstrated by the improved specificity of OKN-RVT in comparison with OKN-ENG. Although OKN-RVT identified central pathology with a sensitivity of only 46.7%, it also had a specificity of 94.4% for identifying subjects without central disease. Given the option of either the RVT or ENG method of OKN testing, the RVT method is clearly superior to the ENG method. If only the ENG method is available, OKN testing is unlikely to be of value.
These data reveal that the 2 popular methods of OKN testing considered here are simply not perfect in identifying patients with central vestibular pathology, although this has been one of their main clinical uses. To date we have no information regarding the test-retest reliability of the OKN results. These factors therefore have important implications for the clinical treatment of patients with vertigo. Clearly, a clinical history and examination are crucial elements in proper decision-making. Vestibular test-battery results are most effective as a supplement to the information gained from a patient's clinical context. This study provides a basis by which to evaluate the meaning of 2 vestibular tests. 13
In summary, we have shown that OKN-ENG does not appear to be helpful in clarifying the cause of vestibular pathology. This most likely reflects the lack of full visual field stimulation provided by OKN-ENG with the light bar method, as compared with OKN-RVT. We have shown furthermore that OKN-RVT testing, while better than OKN-ENG, has a poor sensitivity. Therefore OKN-RVT alone will not properly identify all patients with central vestibular pathology and should be used in a test battery with other tests of central vestibular function. In conclusion, a proper understanding of the information provided by these vestibular tests is needed for their correct use in patient treatment.
We thank Dr Linda Bartoshuk and Dr Jim Jekel for their statistical analyses of the data presented here.