Abstract
BACKGROUND AND OBJECTIVES: The interpretation of video head-impulse tests (video-HITs) can often be complicated, limiting their clinical utility in acute vestibular syndrome. We aimed to determine video-HIT findings in patients with posterior circulation strokes (PCSs) and vestibular neuritis (VN). METHODS: We retrospectively analyzed the results of video-HITs in 59 patients with PCS. Irrespective of the actual lesion revealed later on MRIs, ipsilateral and contralateral sides were assigned according to the direction of slow phase of spontaneous nystagmus (SN). Then, the patterns of video-HIT findings were classified according to the vestibulo-ocular reflex (VOR) gain for the horizontal canals; (1) ipsilaterally positive, (2) contralaterally positive, (3) bilaterally normal, and (4) bilaterally positive. The abnormal responses were further defined into (5) wrong-way saccades, (6) perverted, and (7) early acceleration followed by premature deceleration. We also analyzed the asymmetry of the corrective saccadic amplitude between the sides, calculated from the sum of cumulative saccadic amplitudes on both sides. The results were compared with video-HIT results from 71 patients with VN. RESULTS: Video-HITs were normal in 32 (54%), ipsilaterally positive in 11 (19%), bilaterally positive in 10 (17%), and contralaterally positive in 6 (10%) patients with PCS. Wrong-way saccades were more frequently observed in VN than in PCS (31/71 [44%] vs 5/59 [8%], p < 0.001). Saccadic amplitude asymmetry was greater in VN than in PCS (median 100% [interquartile range 82-144, 95% CI 109-160] vs 0% [-29 to 34, -10 to 22, p < 0.001]). When differentiating VN from PCS, the sensitivity was 81.7%, and specificity was 91.5% at the cutoff value of 71% for saccadic amplitude asymmetry with an area under the curve (AUC) of 0.91 (95% CI 0.86-0.97). The AUC for saccadic amplitude asymmetry was larger than that for the ipsilateral VOR gain (p = 0.041) and other parameters. DISCUSSION: Patients with PCS may show various head-impulse responses that deviate from the findings expected in VN, which include normal, contralaterally positive, and negative saccadic amplitude asymmetry (i.e., greater cumulative saccadic amplitude contralaterally). A thorough analysis of corrective saccades in video-HITs can improve the differentiation of PCS from VN even before MRIs.