Timothy C. Hain, MD • Page last modified: August 14, 2021
Dizziness accounts for 3.5% to 11% of ED visits (Crespi, 2004; Lammers et al, 2011). Of these visits, the majority are from cardiac (heart) or general medical causes. The proportion of costs for ER visits is similar (4% according to Saber et al ,2013). The table below comes from Dr. Newman-Toker's study, and provides the proportion of diagnoses in ED patients. (Newman-Toker et al, 2008)
Newman-Toker and associates (2008) report was based on on dizziness visits to US emergency departments: cross-sectional analysis 9472 dizziness cases in the US. sampled over 13 years. This is about 2 orders of magnitude more patients than were previously studied.
Concerning costs of ED care, HCUP-CCS key diagnostic groups for those presenting with dizziness and vertigo included the following (fraction of dizziness visits, cost-per-ED-visit, attributable annual national costs): otologic/vestibular (25.7%; $768; $757 million), cardiovascular (16.5%, $1,489; $941 million), and cerebrovascular (3.1%; $1059; $127 million) (Saber et al, 2011). According to Newman-Toker, the costs of US E.D. dizziness presentations now exceed $10 billion/year, largely owing to a combination of frequent neuroimaging obtained in about half, and admissions for nearly 20% (Newman Toker, 2016). This figure presumably includes all of the causes above -- many of which are general medical rather than vestibular or stroke.
It is well understood that Emergency Departments are not very accurate in diagnosing dizziness. This is not surprising as they are in a hurry, and ED practitioners must be generalists as well. EDs are also incredibly expensive, and the process of evaluating dizziness can be extraordinarily expensive compared to seeing an expert in the outpatient arena.
Kerber (2009) commented concerning the source of error in ED departments that "a common theme among these misconceptions is an overreliance on the patient's description of symptoms and an overreliance on CT scans (Kerber, 2009). Royl et al (2011) noted that on follow-up 43% of all ER diagnoses (of dizziness) were corrected: 6% of benign ER diagnoses were corrected to serious diagnoses, 23% of serious ER diagnoses were revised to benign. The most frequent corrections concerned patients with an ER diagnosis of stroke or vestibular neuronitis. It is not only ED physicians -- neurology consultants make many mistakes too. Misdiagnosis or diagnostic uncertainty occurred in over one-third of all neurological consultations in the emergency department setting". (Moeller., et al., 2008).
It is not just CT scans - - It is very common for doctors in the ED to obtain a CT angiogram and/or an MRI for patients with dizziness. For example, a patient of mine developed severe vertigo recently. He went to the ED, and instead of examining him for dizziness (perhaps they don't know how ?), they obtained a CT angiogram, and an MRI. He subsequently saw a physical therapist, who noticed that his eyes were jumping (nystagmus), and his HIT test was positive. The physical therapist diagnosed him with vestibular neuritis, which I confirmed in the outpatient setting. Why do ED doctors do this ? Perhaps they are in a hurry, and sending the patient for a test saves their time. It may also reduce their malpractice risk - -getting an MRI means they won't miss a stroke. Patients expect "tests", and this fulfills their need to have their anxiety treated. This is not always good for their patients though - - here there were about $5000 of tests, for a disorder that can and was detected easily with a physical examination.
According to the literature, another reason ED errors are made so oftenbecause of inappropriate emphasis of ED physicians on symptom quality. According to Stanton et al (2007), "In a multivariate model, those ranking quality first (particularly resident physicians) more often reported high-risk reasoning that might predispose patients to misdiagnosis (e.g., in a patient with persistent, continuous dizziness, who could have a cerebellar stroke, resident physicians reported feeling reassured that a normal head computed tomogram indicates that the patient can safely go home) (odds ratio, 6.74; 95% CI, 2.05-22.19). " Newman-Toker also criticized using "quality" of symptom historical measures to make diagnoses in dizziness patients (2007).
While we agree that this is perhaps part of the problem, we think there are more obvious reasons such as the ED doctors may not know much about dizziness, they have no compelling motivation to save other people's money, they are usually in a hurry, they are very motivated to reduce malpractice risk, and perhaps they hope to alleviate anxiety with tests.
We think that the problem is not really solvable for the ED physician -- there are always going to be mistakes, and there is always going to be a need to save time and avoid risk of litigation. Here are the reasons:
ED physicians should do better with vestibular neuritis --this is not so hard. (see following)
Next we will review common categories of dizziness, and then offer an approach to dizziness in the ED.
OTOLOGIC DIZZINESS (i.e. dizziness from the ear -- about 1/3 of all dizzy patients in the ED)
Otologic dizziness is generally safe, and these patients can almost always be discharged from the ED, possibly with some symptomatic treatment.
The key to recognizing otologic vertigo is recognizing nystagmus (involuntary jumping of the eyes), and excluding people with neurological symptoms (such as weakness of the body on one side) and severe headache. Kerber (2011) reported that even when the ED documents nystagmus, it is usually useless. The documentation of nystagmus (including all descriptors recorded) enabled a meaningful inference about the localization or cause in only 10 of the 185 (5.4%) visits.
Nystagmus in brief
There are just a few common types of otologic vertigo:
- Involuntary eye movement
- It begins with a slow movement followed by a fast rapid resetting phase (corrective saccade)
- Beats towards the more active ear and away from the "bad" ear.
- Named by the direction of the fast phase right, left, up, or downbeating, torsional
- BPPV (Benign paroxysmal positional vertigo, 5% of population/year). This one is easy to diagnose, and ED should do better.
- Acute but not very severe dizziness
- Recurrent attacks
- "bed spins"
- No hearing symptoms and no neurological symptoms.
- BPPV is overdiagnosed in emergency settings -- ED physicians often don't check positional testing, and just call everyone "BPPV" because it is common.
- Vestibular Neuritis (a common source of errors in the ED). ED physicians should do better here.
- Usually first ever severe dizzy spell ("lightning strike")
- Acute severe vertigo, nausea, vomiting and disequilibrium
- Vertigo lasts approximately 48 - 72 hours
- Spontaneous and gaze evoked nystagmus will beat towards the good ear. ED physicians generally are very bad about recording the direction of nystagmus, and get it right only about 5% of the time.
- Positive head thrust (HIT test) to the bad ear side.
- Most gradually improve within a few weeks to months
- May be associated with sudden unilateral hearing loss (labyrinthitis)
- Meniere's disease attack (not a common source of errors -- probably because it is rare).
- Acute, severe
- Recurrent attacks
- Vertigo AND hearing loss/tinnitus
- History of reduced hearing on one or both sides.
BPPV in the Emergency Department -- missed and untreated
BPPV is a gigantic cause of dizziness in the general population, with a yearly incidence of about 2%. One would think that the ED should be very good at diagnosing and treating BPPV, but the literature suggests the contrary. Kerber (2013) noted that out of A total of 3522 visits for dizziness, a DHT (Dix Hallpike test) was documented in only 137 visits (3.9%). Experts (such as the author of this page), do DHT in 100% of dizzy patients, because the "hit rate" is very very high. A CRM (canalith repositioning maneuver) was documented in 8 visits (0.2%). Even among patients diagnosed with BPPV, a DHT was documented in only 21.8% (34 of 156) and a CRM in 3.9% (6 of 156). This is not a good record!
NEUROLOGICAL DIZZINESS IN THE EMERGENCY DEPARTMENT
"Neurological causes account for about 11% of ED visits for dizziness (Newman-Toker et al, 2011). So if we do the math -- 10% of ER visits are for dizziness, and 10% are neurological, we get about 1%. Not so much. It may be even less -- Cheung (2011) reported 6% in Hong Kong.
The core problem is that neurological dizziness is rare (1% of dizzy patients in ED), dangerous, and difficult to diagnose. This means neurological dizziness is a high risk area for the ED. The goal here should be to sort out patients into safe ones (e.g. migraine), dangerous ones (e.g. stroke), maybe dangerous (usually refer), probably dangerous (usually admit and let someone figure out). Practically speaking, there will be many mistakes in the ED in this group (see above). Even if you make a mistake, if you document your logic and follow accepted algorithms (see ABCD2 and HINTS below), you did the right thing.
According to Hanson (2011), Neurologists are more often called in to consult when dizziness is present than in other ED presentations. Of the 500 patients analyzed in their study, the most common chief complaints were focal weakness (22%), headache (18.2%), dizziness or vertigo (16%), and seizure (14.2%). As dizziness accounts for roughly 5% of all ED visits, the dizzy patient is 3 times more likely to have a neurology consult called in than other groups.
Overview of neurological dizziness in ER.
- Neurological dizziness accounts for -- Roughly 0.3 to 1% of all ED visits.
- Serious neurological causes – about 5% of ED visits for dizziness (e.g. about 0.5%) (Navi et al, 2012b; Cheung et al, 2011)
There are just a few (uncommon) causes of neurological dizziness (recall that these account for only 1% of ED visits, and only 11% of dizzy cases in ER).
Migraine associated vertigo (1% of entire population has this -- this is the main one)
- Recurrent attacks of dizziness
- Vertigo and headache (not necessarily at the same time)
- Light and sound sensitivity
- Usually woman with estrogen fluctuations (i.e. not very common in post-menopausal women).
Strokes: Strokes are rare but dangerous causes of dizziness. (3.2% of patients with dizziness in ED) (Kerber, 2009). About 0.7% of patients with stroke had "isolated" dizziness. The risk of an early stroke after discharge from the ER with a diagnosis of peripheral vertigo is very low (0.18%) but still about 9 times greater than matched patients with renal colic (Atzema et al, 2016)
- ABCD2 score is a way of determining risk of stroke for anyone (whether dizzy or not) (Johnson et al, 2007). We think ED physicians (and neurologists) should be good at this. The ABCD2 score does not diagnose stroke -- just risk.
- Age > 60 (1 point)
- Blood pressure > 140 or 90 (1 point)
- Clinical signs
- Weakness (2 points)
- Speech impaired (1 point)
- Duration (> 1 hour -- 2 points, 10-50 minutes 1 point)
- Add them all up
- 0-3 points = Low risk
- 4+ points = high risk.
- HINTS score is another way of determining risk of stroke, just in dizzy patients( Kattah, 2009).
- HI -- Head impulse test (also known as rapid dolls). If normal, it suggests central dizziness. One would expect the impulse test to be sensitive to complete unilateral peripheral loss (which is not hard to diagnose in any context, and might also provide false-negatives in persons who have old lesions).
- NT - nystagmus test-- positive for stroke if nystagmus is gaze evoked, rather than unidirectional, it suggests central. Of course, the lack of a horizontal unidirectional nystagmus in the face of significant dizziness is a strong suggestion of central. As everyone has some gaze-evoked nystagmus if you wait long enough, there is some judgement required here.
- S -- skew -- if patients have vertical diplopia, it suggests central. If there is skew, it takes priority over the head impulse test result. Skew is rare and adds specificity.
- The "HINTS+" includes a screening for hearing (finger rubbing).
- So the way you use the HINTS test is:
- If positive HIT test AND strong unidirectional nystagmus--> not stroke, probable vestibular neuritis.
- If strong gaze-evoked nystagmus, without lateralizing signs --> not vestibular neuritis, more likely stroke.
- If skew deviation --> more likely stroke
- The HINTS test is essentially just subspecialty neurology clinical judgement, boiled down to a bare minimum. Because the HINTS tests contains some "judgement calls" -- and because the ED rarely has the basic equipment needed to monitor nystagmus (i.e. portable Frenzel goggles), we doubt that this can be successfully adopted by ED physicians. It is something that the neurology consultant should know however.
- Our thought is that by adding the Dix-Hallpike test for BPPV, these criteria could be made much stronger. If someone has BPPV, it's not likely to be central. However, ED doctors don't seem to be very good about this.
- One might argue that other useful tests for stroke would be the time-honored general neurological exam, checking the fundus (for papilloedema), looking for a Babinski sign -- in other words -- if you are thinking that a HINTS test is needed -- practically you will likely need a neurology consultant.
- Does the HINTS+ work ? Recent papers say no.
- The HINTS test requires a fair amount of subspecialty knowledge --one must understand how to do the HIT test, have some judgement about nystagmus, and know how to examine for skew. To us, this doesn't seem like something an ED physician would be likely to be good at. Perhaps a Neurology consultant activity ? This general idea was borne out by a large chart review by Dmitriew et al (2021), which found "A total of 2,309 patients met criteria for inclusion in the study. In no case was dizziness due to a central cause identified using the HINTS exam. " This is not encouraging.
- Choi et al (2021) reported that "Bedside examination and routine magnetic resonance imagings have a limitation in diagnosing strokes presenting with ATVS, and perfusion imaging may help to identify strokes in ATVS of unknown cause. Associated craniocervical pain and focal neurological symptoms/signs are the useful clues for strokes in ATVS." (ATVS is acute transient vestibular syndrome. the bedside exam they are referring to is the HINTS)
The rest of neurology
- Almost any brain disorder can cause dizziness.
- B12 deficiency
- Brain tumors
- Multiple sclerosis
- Chiari malformation
- Spinocerebellar degeneration
- Hydrocephalus, low CSF pressure
- etc. etc. etc.
- Don't try to diagnose these in the ED (i.e. don't do a CTA). Refer if they are not very sick, call your neurology consultant or admit if they are very sick
Most dizziness in the ED is due to "medical" causes. Items below are abstracted from Newman-Toker, 2008. Note that almost any of these causes exceeds the frequency of stroke in dizzy patients.
- Low blood pressure, arrhythmia (e.,g. heart attack, anemia, medication overdose) -- 21%
- respiratory problems (e.g. low oxygen) -- 11.5%
- metabolic issues (e.g. low blood sugar) -- 11.0%
- Injury/poisoning -- 10.6%
- digestive (e.g. stomach flu) -- 7%
- genitourinary (e.g. bladder obstruction) -- 5.1%
- infections (e.g. "flu"). -- 2.9%
This is certainly the place where the ED should do well with diagnosis and treatment. These are all very common problems, and for the most part, they are not dangerous (other than cardiac).
UNDIAGNOSED DIZZINESS IN THE EMERGENCY DEPARTMENT
There are are many many times when the diagnosis is "undiagnosed" -- nobody can figure out why a patient is dizzy. Even in subspecialty dizzy clinics, there are many patients where no diagnosis is attained. These people are often given "wastebasket" diagnoses such as "chronic subjective dizziness". This is not a diagnosis -- it is just another way of saying that the clinician doesn't know what is going on. Perhaps the BPPV patient is no longer dizzy, perhaps the brain tumor patient is just not severe enough to be diagnosed. Perhaps they have something difficult to figure out. Once you are sure they are safe (e.g. no life-threatening condition), this is a "refer" situation for the ED.
Rather than organizing diagnosis down lines of mechanism, dizziness in emergency departments is best thought of in terms of the situation.
Kerber suggested an algorithm to sort Dizziness in the ED (Kerber KA, 2009). Kerber said -- "understanding three peripheral vestibular disorders--vestibular neuritis, benign paroxysmal positional vertigo, and Meniere's disease--is the key to the evaluation and management of vertigo and dizziness presentations in the emergency department". In other words, Kerber is a "lumper".
The general process suggested by Dr. Kerber in his 2009 article for ED physicians is to "rule in" vestibular neuritis, BPPV or Meniere's using a combination of one's clinical judgements or in the case of BPPV, a specific bedside test, and to consider all else to be a potential stroke. (note that the suggested treatment for vestibular neuritis (steroid burst), has now dropped out of favor).
Kerber (2009) suggested that ED patients should be sorted out into the following categories:
- Acute-severe -- with nausea/vomiting/imbalance -- including vestibular neuritis and stroke
- Recurrent positional -- examples being BPPV, cerebellar disorders, and degenerative ataxias.
- Recurrent vertigo -- (unprovoked) -- examples being Meniere's disease and TIA.
There are many critical details left out of this algorithm. For example, "chronic dizziness" is not considered.
The author of this page (Dr. Timothy Hain) suggests the following algorithm for evaluation of dizziness in the ED:
History of Dizziness in the ED. Key points to quickly find out (these are from the author of this page, not Dr. Kerber)
- Is it acute ?
- Supine Position – 85% BPPV
- Upright Position –cardiac, low blood pressure.
- Hearing loss or tinnitus (acute) -- ear source of vertigo
- Weakness/numbness – CNS source of vertigo (stroke or migraine).
Rapid examination of dizzy patients in ED. Here we are trying to be quick and safe.
Laboratory testing of dizzy patients (If not obvious from exam -- i.e. they have BPPV)
- Take blood pressure standing up, make sure heart is regular. If BP is too low or pulse is irregular- think cardiac. You might need to admit.
- Check for hearing loss and unidirectional nystagmus. If acute, then this is almost certainly an ear problem, treat and refer.
- Check reflexes and power and for Babinski sign (for stroke). If anything is positive, go through the ABCD2 algorithm.
- Check for positional vertigo (do a Dix-Hallpike test). Treat or refer and then discharge if Dix-Hallpike is positive.
- Do check EKG (for heart)
- Do check CBC, blood sugar and electrolytes (for anemia and diabetes).
- Do get a Tox screen if the situation appears suspicious
- Don't get a CT or MRI on dizzy patients unless either the ABCD2 score or HINTS test (from Neurology consultant) is in favor of stroke. Be more aggressive in older patients, patients with headache with vomiting (could be migraine), and of course patients with focal neurological findings.
- With respect to CT scans, Newman-Toker (2016) points out that CT scans usually miss strokes of the posterior fossa (80% of the time), which are the strokes that cause dizziness. In other words, if someone is hemiplegic -- get a CT and admit. If they are dizzy -- don't get a CT.
- Don't get a CTA in ED either-- this is too much risk (radiation, dye, financial cost to medical system) for the benefit (almost nothing). Just because you can get a CTA at 3AM as an emergency, doesn't mean it is a good idea. In the ED, you don't have enough time to think this through -- you don't have their previous medical records or have time to read them if you are lucky enough to have them-- generally you are just "shooting from the hip". If you really think they are having a stroke, admit them and let the inpatient physician make this decision.
Treat or refer the safe ones
- Inner ear disorders:
- BPPV - -Epley maneuver
- Vestibular neuritis or Meniere's-- hydrate, meclizine, refer
- Orthostatic hypotension -- hydrate, refer
- Migraine -- migraine abortives, refer
- Many medical illnesses (e.g. dehydration)
Admit the dangerous ones
- High stroke risk
- Some medical illnesses.