The March edition of Journal Club reviewed the HINTS exam for acute vestibular syndrome. Much has been written/podcasted in the FOAM world about this controversial article, with the HINTS exam being touted as a godsend by some, while others recommend cautious use. In addition to the usual critical appraisal of the article, journal club leaders Anthony Seto, Vanessa Potok and Andrew McRae also had us review the recommendations of 4 FOAM resources based on the HINTS article. The FOAM recommendations were highly variable, to say the least. This post should be enlightening to many of us, and highlights the importance of actually reading articles yourself and critically appraising them, rather than implicitly taking advice based on the recommendations of a blog.
Article: HINTS to Diagnose Stroke in the Acute Vestibular Syndrome: Three-Step Bedside Oculomotor Examination More Sensitive Than Early MRI Diffusion-Weighted Imaging
Image courtesy emjclub.com
Before getting into the nitty gritty, these are a few major reasons we should be cautious with use of the HINTS exam.
1)The definition of the acute vestibular syndrome (AVS) is continuous vertigo for > 24 hours. Ie: Not intermittent vertigo. There is a very small subset of patients who actually have AVS, and in general we would likely have a higher index of suspicion for ominous causes in patients with continuous vertigo. At journal club we agreed that many of us were applying this to patients with intermittent vertigo. These patients were not included in the study.
2)This was done by a single neuro-ophthamologist using specialized equipment (Frenzl goggles) on a small number of patients (n=101) who were referred to a tertiary care stroke center from 25 community centres. Most of us are not likely using Frenzl goggles in the ED, nor have we had specific training in this exam. The study needs to be validated by EPs performing the exam in a typical group of patients presenting primarily to the ED.
Image courtesy takenotenigeria.com
*May 20 Edit: Comments by Andrew McRae*
One novel aspect of this journal club was to examine the treatment of the HINTS exam by various FOAM resources. FOAM resources are rapidly becoming the most commonly used source of new medical information by both trainees and experienced providers. We examined the discussion of the HINTS exam on several emergency medicine blogs, chosen primarily because of their wide readership.
The FOAM resources all discussed key aspects of the inclusion criteria of the study, namely that patients all had acute vestibular syndrome with persistent vertigo, and emphasized that the HINTS exam should not be used in patients with intermittent vertigo.
However, JC participants noted wide variability in conclusions on the clinical utility of the HINTS exam for ED patients. FOAM resources varied from urging caution in using HINTS given the lack of data on inter-rater reliability of the exam and uncertain generalizability to emergency physicians without specialized equipment, to enthusiastic recommendation of incorporation of HINTS into practice with links to how-to videos and links to slow-motion camera apps that have not been rigorously evaluated.
FOAM resources included both thoughtful critical appraisal of the internal validity and generalizability of the HINTS exam studies (eg. EMNerd) as well as more practical “how-to” material. One challenge JC participants identified was that the how-to resources also referred to the supporting literature (some even describing HINTS as being supported by “VERY STRONG” evidence). This can lead to the impression that FOAM authors based their endorsement on rigorous critical appraisal (ie. producing a “how-to” could imply that the evidence supporting an intervention is solid), and participants had a difficult time differentiating “how-to” FOAM from rigorous critical appraisal.
The remainder of the JC discussion focused on how physicians can be savvy users of FOAM resources. Suggestions included examining several resources and comparing/contrasting their conclusions, as well referring back to and critically appraising the primary source literature. Participants agreed that FOAM resources can be excellent tools for identifying important new EM literature, but that it is ultimately the practitioner’s responsibility to appraise the evidence for him/herself prior to incorporating new evidence into practice. Participants agreed that it would be helpful for FOAM providers to be more explicit as to whether the material they were presenting is critical appraisal or a more practical “how-to”.
Now let’s critically appraise the article.
METHODS/SETTING/DESIGN/NUMBER OF PATIENTS INCLUDED
- Prospective, cross-sectional study performed at single urban, academic hospital serving as regional stroke referral centre for 25 community hospitals
Patients were recruited from the emergency department (n=59), other institutions (n=37), admitted patients with cerebellar infarctions (n=4), and one outpatient (n=1). A total of 101 patients studied, collected over 9 years. 65% were men with mean age 62 (range 26 – 92 years old).
- Single neuroophthalmologist did HINTS exam on patients
- HINTS exam was performed between 1 hour to 9 days from symptom onset (mean 26 hours). 75% were examined within 24h of symptom onset.
- All patients underwent neuroimaging generally after bedside HINTS. If they had imaging prior to HINTS, examiner was blinded to these results at time of assessment. 70% were imaged within 6h of symptom onset. 97% were imaged within 72h of symptom onset.
- All patients (including those with suspected acute peripheral vestibulopathy) were admitted for observation and serial daily examinations
- Stroke diagnosis: MRI with DWI (97%) or CT
- Acute peripheral vestibulopathy diagnosis: absence of stroke in brainstem/cerebellum on MRI with DWI, lack of neurological signs on serial exam, and characteristic clinical course, +/- caloric testing
- Eight patients with initial negative MRI later underwent repeat MRI for unexplained neurological signs suggesting brainstem localization
- One neuroophthalmologist conducted HINTS exam
- HI = Head Impulse
- N = Nystagmus
- TS = Test of Skew
- “Benign” HINTS = abnormal “HI”, direction-fixed “N”, and absent “TS”
- “Dangerous” HINTS = normal/untestable “HI” or direction-changing “N”, or present/untestable “TS”
- Core features of acute vestibular syndrome (rapid onset of vertigo, N/V, and unsteady gait +/- nystagmus). Note that duration of symptoms was not important to be included in the study, some patients had ~ 1 hr of symptoms prior to HINTS exam.
- At least 1 stroke risk factor
- Atrial fibrillation
- Hypercoagulable state
- Recent cervical trauma
- Prior stroke
- Prior MI
- History of recurrent vertigo +/- auditory symptoms
- Determine diagnostic accuracy of skew deviation for identifying stroke in acute vestibular syndrome: compare proportions with skew deviation in peripheral versus central cases
- Determine added value of skew deviation beyond h-HIT: stratify results by horizontal head impulse test findings
- Compare HINTS exam’s sensitivity and specificity for presence of stroke
- Skew present in 17% of the 101 studied subjects and was associated mainly with brainstem lesions
- Skew in 4% (n=1 of 25) with acute peripheral vestibulopathy, 4% (n=1 of 24) with pure cerebellar lesions and 30% (n=15 of 50) with brainstem lesion (chi-squared, P=0.003)
- Skew correctly predicted lateral pontine stroke in 2 out of the only 3 cases in which an abnormal horizontal head impulse test erroneously suggested peripheral localization (Table 3, below)
- Dangerous HINTS = 100% sensitive and 96% specific for central lesion (Table 4, below)
- Skew deviation is insensitive for central pathology but fairly specific for brainstem involvement for patients with acute vestibular syndrome
- Skew may help identify stroke when a positive head impulse test falsely suggests a peripheral lesion
- HINTS appears more sensitive (100% for HINTS) and specific (96% for HINTS) than MRI with DWI in detecting stroke in first 24-48 hours after symptom onset
- Internal validity
- An appropriate reference standard was used initially: neuroimaging
- Partially unmasked examiner: although masked to results of imaging, was not masked to patient’s clinical history, general neurological exam, or obvious oculomotor findings
- Selective MRI follow-up scans: MRI follow-up was only obtained for selected cases based on evolution of new neurological signs or atypical subtle oculomotor signs, potentially leading to the misclassification of strokes and acute peripheral vestibulopathy, increasing apparent sensitivity
- Observer bias in interpretation of eye findings may inflate HINTS sensitivity
- Unclear whether radiologist was blinded to HINTS exam results
- Unclear whether there may have been other eye exam components that would be useful for predicting stroke and a validation study would be required if the current study functions as just derivation data
- Relatively small sample size (N=101)
- External validity
- HINTS is not very time consuming to do, and it is free
- HINTS may be useful for the right patient population, i.e. active symptoms to avoid labeling a normal head impulse test as “central” in normal patients
- Acute vestibular syndrome in the literature is defined as >24 hours. The study enrolled patients between 1 hour to 9 days. It seems to suggest that HINTS can be used for patients presenting as early as 1 hour after symptom onset.
- Generalizability of exam technique: HINTS was performed by a neuroophthalmologist and not by emergency physicians
- Single examiner: unsure if testing style can be replicated by others
- Restricted enrollment to high-risk patients: unclear whether results can be extrapolated to lower-risk population. However, there were some younger patients in the mix (15 were less than 50 y.o. and 6 were less than 40 y.o.). It was not clear which patients of what age had which risk factors, and if those combinations would make one at “lower” or “higher” risk.
It may have been useful to stratify patients into ataxic versus non-ataxic, since we are more concerned about what to do about patients who have less obvious neurological findings.
UCalgary Journal Club Group
These points were discussed during Journal Club on March 19, 2015:
- In a patient with acute vestibular syndrome, who had a normal CTA and reassuring HINTS, would you send them home? 43% would consult neurology first. 57% would send the patient home.
- Would you incorporate HINTS as an additional part of your physical examination? Most people agreed they would consider incorporating the HINTS components as part of their examination and, in particular, examining for high-risk nystagmus and skew deviation. There appeared to be agreement that the head impulse test may be challenging to perform, and we may not have adequate inter-rater reliability with a single, experienced neuro-ophthalmologist.
- If you do use HINTS, ensure you use it on the right patient population: i.e. someone who has active persistent vertigo (acute vestibular syndrome patient) and not an intermittent vertigo like BPPV.
- You may consider downloading a slow-motion camera application on your mobile device in order to detect the subtle ocular symptoms. However, the safety and accuracy of this practice has not been rigorously evaluated. Evidence in support of slow-motion applications is anecdotal at best, and its use is not recommended without evidence of its utility.
- It may be challenging to differentiate between a normal versus abnormal head-impulse test. Moreover, an abnormal head-impulse test, although should imply a peripheral cause, can still mean a central cause. Therefore, direction-changing nystagmus and present skew deviation may be elements of HINTS less challenging to interpret.
- Although there are many limits to the generalizability of the HINTS exam, if it can be conducted on the right patient population using appropriate technique, it may add to your overall diagnostic evaluation for patients with acute vestibular syndrome. Consider reviewing the HINTS exam: http://empablogsite.com/2014/08/hints-exam/
Thanks again to Anthony Seto and Vanessa Potok for compiling this excellent summary.