Accuracy and rEliabilitY of the vEstibuLo-ocular ExAmination Performed by inteRNs IN the emerGency Department

Unfortunately, the diagnostic evaluation of an isolated AVS is often frustrating because patients with vertigo tend to be uncertain and imprecise when describing their symptoms. Examination of the vestibulo-ocular pathway, by assessing eye movements in different head positions, is an essential step in the clinical examination to reliably and quickly predict the origin of a AVS. The 3-step clinical rule "HINTS" (Head Impulse, Nystagmus, Test of Skew) was developed in 2009 by "neuro-ophthalmologists" in the USA. It evaluates the vestibulo-ocular reflex by the Head Impulse Test (pathological with the presence of ocular saccades for a peripheral AVS), the characteristics of the nystagmus (multidirectional and/or vertical for a central AVS), and looks for ocular misalignment (Test of Skew with cover-test, pathognomonic of a central involvement). In patients at high cardiovascular risk, HINTS is 100% sensitive and 96% specific in predicting posterior fossa stroke within 2 minutes at the bedside, compared with the Gold-Standard. The "STANDING" (SponTAneous, Nystagmus, Direction, head Impulse test, STANDING) clinical rule is a structured bedside diagnostic algorithm in four clinical steps. It is based on the evaluation of two eye movements (Head Impulse Test and nystagmus), an evaluation of gait (search for ataxia) and the search for positional nystagmus (Dix Hallpike test for the posterior canal and Panigni test or "supine and roll test" for the lateral canal). In the ED, with patients at heterogeneous cardiovascular risk, the STANDING is more interesting than the HINTS. Indeed, it allows the detection of BPPV in pauci-symptomatic patients at the time of their clinical evaluation (i.e., those without spontaneous nystagmus). It had also shown excellent performance (sensitivity:95%, specificity:87-96%) in diagnosing any central cause of VAS in the hands of emergency physicians. However, the 2017 validation study of STANDING had an audit bias. The gold standard test, brain imaging, had only been performed when deemed appropriate (ie, in 34% of peripheral diagnoses, and for only 5% of these by brain MRI).

In the EYE-ECG study conducted at the GhPSJ SU from October 2019 to January 2021, the invesitgators determined the diagnostic performance of these two clinical rules performed by emergency physicians to 300 patients with isolated VAS, comparing it to the Gold-Standard. The manuscript of this study has been published in the Academic Emergency Medicine journal (DOI: 10.1111/acem.14337) in June 2021. Its originality was to evaluate the performance of the tests in the hands of emergency physicians and on a population not selected for its risk of stroke. The investigators showed that the HINTS and STANDING tests had excellent sensitivities (97% and 94% respectively) and negative predictive values (99% and 98% respectively) for predicting any central cause of VAS on brain MRI. They thus allowed emergency physicians to exclude a central cause in a rapid manner (on average 5±3 minutes) with very acceptable false-negative rates (3% and 6%, respectively), and with a potential impact on the reduction of unnecessary brain imaging (-33% and -32%, respectively). For predicting MRI normality, STANDING was more specific than HINTS (75% vs. 67%). This is explained by the addition of a clinical item dedicated to positional nystagmus in the STANDING, and by the high prevalence of BPPV among VAS of ENT origin in SU patients (40%). Indeed, among the false-positives of the HINTS test (erroneously predicted as a central cause for benign vestibular disease), one third of the diagnoses were BPPVs.

The major limitation of the EYE-ECG study was that it did not assess interindividual variability of the HINTS and STANDING tests between different emergency physicians. The STANDING validation study was the only study that assessed global and individual agreement of the algorithm item scores in a binary "central or peripheral" mode between two senior emergency physicians. The inter-observer agreement of the STANDING was good globally (K=0.83) and also at each step of the algorithm (distinction between spontaneous and positional nystagmus: K=0.83, interpretation of a multidirectional or vertical nystagmus: K=0.95, normal Head Impulse Test: K=0.74, recognition of a walking ataxia: K=0.81)18.

Each semester, an average of 7 medical students in general medicine and one in emergency medicine complete an internship at the GhPSJ. In routine practice, the medical students interview and examine their patients alone. They document their clinical examination in the patient's medical record and then present the patient's clinical record to a senior emergency physician. In a second step, the patient is clinically reassessed by the senior emergency physician who becomes responsible for the patient's diagnostic and therapeutic strategy. The accuracy of the vestibulo-ocular examination performed by medical students has never been described in the literature.

The hypothesis of the investigators is that after standardized theoretical training and with the clinical experience gained during their semester in the ED, medical students should be able to examine and interpret the vestibulo-ocular examination of patients with isolated AVS as accurately as seniors.