Head Impulse, Nystagmus, and Test of Skew Examination for Diagnosing Central Causes of Acute Vestibular Syndrome
Cochrane Database of Systematic Reviews
Gottlieb, M., Peksa, G. D., et al. (2023).
Cochrane Database of Systematic Reviews, 11(11), Cd015089.
This systematic review and meta-analysis investigates the diagnostic accuracy of the head impulse, nystagmus, test of skew (HINTS) and HINTS Plus (HINTS combined with hearing loss) examinations for individuals with acute vestibular syndrome.
Cochrane Collaboration
From database inception to September 26, 2022
Primary diagnostic test accuracy studies including prospective and retrospective studies
19
Meta-analysis from 12 studies of clinical HINTS estimated a summary sensitivity of 94.0% (95% CI = 82.0%-98.2%) and specificity of 86.9% (95% CI = 75.3%-93.6%) with low certainty of evidence. Video-assisted HINTS was evaluated in three studies; summary sensitivity was calculated to be 96.3% (95% CI = 87.6%-100%) with low certainty of evidence and specificity was calculated to be 75.7% (37.1%-100%) with very low certainty of evidence. Five studies evaluated the clinical HINTS plus. Pooled sensitivity was 95.3% (95% CI = 78.4%-99.1%) and specificity was 72.9% (95% CI = 44.4%-90.1%) with low certainty of evidence. For the video-assisted HINTS plus, summary sensitivity was 90.6% (95% CI = 81.1%-100%) and summary specificity was 37.6% (95% CI = 29.2%-46.0%) with moderate certainty of evidence from two studies.
Summary estimates for the HINTS subcomponent, head impulse were sensitivity 68.8% (95% CI = 42.0%-87.1%) and specificity 92.5% (95% CI = 75.8%-98.0%) with low-certainty evidence. Video head impulse had a sensitivity of 61.1% (95% CI = 24.7%-97.5) and a specificity of 66.7% (95% CI = 36.7%-96.7%) with very low-certainty evidence. For HINTS subcomponent nystagmus, sensitivity was 56.8% (95% CI = 40.2%-71.9%) with low-certainty evidence and specificity was 98.8% (95% CI = 96.0%-99.7%) with moderate-certainty evidence. Sensitivity for the test of skew was 15.0% (95% CI = 6.6%-30.5%) with low-certainty evidence and specificity was 98.7% (95% CI = 97.1%-99.4%) with moderate-certainty evidence.
For the clinical HINTS, one study reported a time frame of 24 hours or fewer from symptom onset to presentation. Sensitivity was 71% (95% CI = 49% - 87%) and specificity was 100% (95% CI = 92% - 100%). For studies with longer than a 24-hour time frame, there was a pooled sensitivity of 52.9% (95% CI = 0-100%) and a pooled specificity of 44.4% (95% CI = 0-100%) for clinical HINTS. For video-assisted HINTS with time from symptom onset to presentation longer than 24 hours, the sensitivity was 93.5% (95% CI = 79.0%-100%) and the specificity was 63.7% (95% CI = 15.0%-100%). For clinical HINTS plus, one study reported on accuracy when presenting at fewer than 24 hours (sensitivity: 100%, 95% CI = 29%-100%; specificity: 75%, 95% CI = 51%-91%) and one study reported on accuracy when presenting at more than 24 hours (sensitivity: 80%, 95% CI = 56%-94%; specificity: 43%, 95% CI = 34%-53%).
Pooled analysis found a sensitivity of 95.7% (95% CI = 87.7%-98.6%) and specificity of 89.1% (95% CI = 76.5%-95.4%) when using a reference standard of advanced imaging for clinical HINTS. For video-assisted HINTS using advanced imaging as a reference standard, summary specificity was 93.8% (95% CI = 82.7%-100%). For the clinical HINTS plus, sensitivity was 93.6% (95% CI = 77.3%-100%) and specificity was 65.1% (95% CI = 40.7%-86.5%) with advanced imaging as a reference standard.
For clinical HINTS in those with ischemic stroke, pooled sensitivity was 91.3% (95% CI = 66.8%-98.2%) and specificity was 79.9% (95% CI = 63.2%-90.2%. For clinical HINTS plus when underlying etiology was reported to be ischemic stroke, the pooled sensitivity was 95.4% (95% CI = 57.9%-99.7%); specificity was 60.0% (95% CI = 39.1%-77.9%).
The meta-analysis for the outpatient setting found a sensitivity of 94.4% (95% CI = 81.0%-98.6%) and a specificity of 85.4% (95% CI = 72.7%-92.8%) for clinical HINTS.
