Rant – More HINTS We Have a Problem

December 3, 2020

---

Gentle reader, this post is longer than usual. Some topics need more ink, and HINTS is one of them. Let’s jump carefully hold the rails and walk in.

---

Written by Clay Smith

Spoon Feed
HINTS was misapplied 97% of the time by emergency physicians in patients with dizziness. Five patients that clinicians thought had central vertigo by HINTS all did not. Four of six patients who had proven central vertigo all had intermittent symptoms that sounded more like benign positional vertigo. I give up…

Why does this matter?
HINTS is great in the right hands. But we covered an article recently that raised concern about the accuracy of HINTS when performed by emergency physicians. If we use HINTS on the wrong patient or misinterpret the results, it could lead to unnecessary testing or a misdiagnosis – possibly a missed stroke. Personally, I have to watch a YouTube video every time I want to use HINTS. Afterward, I just feel confused, intimidated, and dumber than usual. Then I just feel sad. Then I drown my sorrows in coffee. Then I start to develop nystagmus. Speaking of which, I can’t seem to remember which kind of nystagmus is bad? Was the skew really skewed? What’s a saccade again? An abnormal head impulse is good? Sheesh!

For the record, HINTS should only be used in patients with acute vestibular syndrome (AVS) defined as: “acute onset, continuous vertigo associated with gait unsteadiness, nausea and/or vomiting, and spontaneous or gaze-evoked nystagmus.” It should not be used in, “episodic, positionally-evoked vertigo, as is typical of BPPV.”

HINTS consists of three parts:

1. Head Impulse: Normally, when rapidly turning the head, the eyes will stay fixed on a target. With peripheral vertigo [causing AVS], rapidly turning the head leads to the eyes going off target, with a fast corrective saccade to get back on target. In central vertigo, head impulse is normal. So, it’s good when the head impulse is abnormal (which is why vertigo makes my head spin). The corrective saccade happens fast, is subtle, and depends on good exam technique in rapidly moving the head.

2. Nystagmus: Unidirectional nystagmus is good and suggests a peripheral cause; direction-changing nystagmus is bad and indicates a central cause. This can be very subtle. Was it just a rotary component or was that multi-directional nystagmus? This is not easy to differentiate sometimes.

3. Test of Skew: “Vertical deviation followed by correction on cover-uncover test,” is called skew deviation. This one happens fast but seems easier for me spot as abnormal.

For more, see this video by Dr. David Newman-Toker on how to do HINTS correctly. He is Master Yoda of HINTS.

And you must watch this video of Dr. Peter Johns, who appears to be the Luke Skywalker of HINTS. It is better than the video above.

So, what’s the diagnostic accuracy of HINTS among emergency physicians?

I’m not good at taking HINTS
This was a retrospective study of 2,309 dizzy, vertiginous, light-headed, or unsteady patients at a single center (truly a nightmarish patient population). At this hospital, clinicians used and documented HINTS commonly, in 450 of these cases. Unfortunately, emergency physicians misapplied HINTS 97% of the time. In other words, they used HINTS, but they shouldn’t have. The patients didn’t meet the criteria of AVS: continuous symptoms, nystagmus, and unsteady gait. Of these three, the most commonly absent clinical variable was nystagmus followed by presence of intermittent, non-continuous symptoms.

Of the 14 patients in which HINTS was used appropriately, 5 were found to have a central cause by HINTS. However, none of them proved to have central vertigo with additional testing. They were wrong about all five.

In almost half the patients (220/450), clinicians performed both HINTS and Dix-Hallpike, which makes no sense. Dix-Hallpike should only be used in patients with brief, intermittent, motion-triggered symptoms. But can we really blame the emergency physicians and say they misclassified these patients? Don’t patients with BPPV often say they feel “off” continuously, even though they only have bouts of vertigo intermittently? Of course they do [note the comment from Dr. Peter Johns below on this point]! So, was it intermittent or continuous? Who knows, especially in a retrospective study trying to lift these nuanced elements from chart review?

Six patients out of 450 actually had a central cause of vertigo: 4 had a stroke; 1 had a TIA; and one had MS. Yet none of them had documentation of AVS. Four had intermittent symptoms, and three lacked nystagmus, or at least it was not documented. That is truly scary.

This study is limited by its retrospective design. If the chart didn’t mention duration, for example, it was presumed to be continuous. This is not an ideal way to study the accuracy of HINTS. However, I think it shows we, as emergency physicians, are probably not as skilled as we think we are at this physical exam technique. It also shows that we have serious trouble applying HINTS to the right patient population and need more education to use this in practice.

My soul dies a little when I see “dizziness” as the chief complaint. I must say, this study doesn’t encourage me. Again, a retrospective study to extract variables from chart review about dizzy/ vertiginous/ lightheaded/ unsteady patients – the vaguest chief complaints on the planet – is not ideal. However, at present, I am only…mostly despondent about using HINTS. I will fully despair when a well designed, prospective study of emergency physicians shows we can’t perform the HINTS exam. Until then, I’m cautious about implementing HINTS in my practice. I think I understand HINTS, but I don’t want to be Dr. Dunning-Kruger. For those who use it, make sure you’re applying it in the right patients, and make sure you know what you’re seeing on exam. It is not a failure to order an MRI once in a while.

Source
Diagnostic accuracy of the HINTS exam in an emergency department: A retrospective chart review. Acad Emerg Med. 2020 Nov 10. doi: 10.1111/acem.14171. Online ahead of print.

Open in Read by QxMD