Systematic review and meta-analysis of the diagnostic accuracy of spontaneous nystagmus patterns in acute vestibular syndrome

OBJECTIVES: For the assessment of patients presenting with acute prolonged vertigo meeting diagnostic criteria for acute vestibular syndrome (AVS), bedside oculomotor examinations are essential to distinguish peripheral from central causes. Here we assessed patterns of spontaneous nystagmus (SN) observed in AVS and its diagnostic accuracy at the bedside. METHODS: MEDLINE and Embase were searched for studies (1980–2022) reporting on the bedside diagnostic accuracy of SN-patterns in AVS patients. Two independent reviewers determined inclusion. We identified 4,186 unique citations, examined 219 full manuscripts, and analyzed 39 studies. Studies were rated on risk of bias (QUADAS-2). Diagnostic data were extracted and SN beating-direction patterns were correlated with lesion locations and lateralization. RESULTS: Included studies reported on 1,599 patients, with ischemic strokes (n = 747) and acute unilateral vestibulopathy (n = 743) being most frequent. While a horizontal or horizontal-torsional SN was significantly more often found in peripheral AVS (pAVS) than in central AVS (cAVS) patients (672/709 [94.8%] vs. 294/677 [43.4%], p < 0.001), torsional and/or vertical SN-patterns were more prevalent in cAVS than in pAVS (15.1 vs. 2.6%, p < 0.001). For an (isolated) vertical/vertical-torsional SN or an isolated torsional SN specificity (97.7% [95% CI = 95.1–100.0%]) for a central origin etiology was high, whereas sensitivity (19.1% [10.5–27.7%]) was low. Absence of any horizontal SN was more frequently observed in cAVS than in pAVS (55.2 vs. 7.0%, p < 0.001). Ipsilesional and contralesional beating directions of horizontal SN in cAVS were found at similar frequency (28.0 vs. 21.7%, p = 0.052), whereas for pAVS a contralesional SN was significantly more frequent (95.2 vs. 2.5%, p < 0.001). For PICA strokes presenting with horizontal SN, beating direction was ipsilesional more often than contralesional (23.9 vs. 6.4%, p = 0.006), while the opposite was observed for AICA strokes (2.2 vs. 63.0%, p < 0.001). CONCLUSIONS: (Isolated) vertical and/or torsional SN is found in a minority (15.1%) of cAVS patients only. When present, it is highly predictive for a central cause. A combined torsional-downbeating SN-pattern may be observed in pAVS also in cases with isolated lesions of the inferior branch of the vestibular nerve. Furthermore, in cAVS patients the SN beating direction itself does not allow a prediction on the lesion side.