Systematic Review and Meta-Analysis of the Diagnostic Accuracy of a Graded Gait and Truncal Instability Rating in Acutely Dizzy and Ataxic Patients

Abstract

Abstract Background In patients presenting with acute prolonged vertigo and/or gait imbalance, the HINTS [Head-Impulse, Nystagmus, Test-of-Skew] are very valuable. However, their application may be limited by lack of training and absence of vertigo/nystagmus. Alternatively, a graded gait/truncal-instability (GTI, grade 0–3) rating may be applied. Methods We performed a systematic search (MEDLINE/Embase) to identify studies reporting on the diagnostic accuracy of bedside examinations in adults with acute vestibular syndrome. Diagnostic test properties were calculated for findings using a random-effects model. Results were stratified by GTI-rating used. Results We identified 6515 articles and included 18 studies (n = 1025 patients). Ischemic strokes (n = 665) and acute unilateral vestibulopathy (n = 306) were most frequent. Grade 2/3 GTI had moderate sensitivity (70.8% [95% confidence-interval (CI) = 59.3–82.3%]) and specificity (82.7 [71.6–93.8%]) for predicting a central cause, whereas grade 3 GTI had a lower sensitivity (44.0% [34.3–53.7%] and higher specificity (99.1% [98.0–100.0%]). In comparison, diagnostic accuracy of HINTS (sensitivity = 96.8% [94.8–98.8%]; specificity = 97.6% [95.3–99.9%]) was higher. When combining central nystagmus-patterns and grade 2/3 GTI, sensitivity was increased to 76.4% [71.3–81.6%] and specificity to 90.3% [84.3–96.3%], however, no random effects model could be used. Sensitivity was higher in studies using the GTI rating (grade 2/3) by Lee (2006) compared to the approach by Moon (2009) (73.8% [69.0–78.0%] vs. 57.4% [49.5–64.9%], p = 0.001). Conclusions In comparison to HINTS, the diagnostic accuracy of GTI is inferior. When combined with central nystagmus-patterns, diagnostic accuracy could be improved based on preliminary findings. GTI can be readily applied in the ED-setting and also in patients with acute imbalance syndrome.