Key points

Introduction

Approximately 3.5% of emergency department (ED) visits are for dizziness. Numerous conditions, some benign and self-limiting and others extremely serious, can present with dizziness. This is classic emergency medicine — sorting out the large majority of patients with a given chief complaint who have a self-limiting or easily treatable condition from the smaller number who have life-threatening, limb-threatening, or brain-threatening problems. Increasingly, physicians are charged with performing this diagnostic process using fewer resources. As of 2013, the direct ED-related costs of care for patients with dizziness in the United States was estimated to approach $4 billion. In addition to economic cost, there is additional cost in terms of patient-experienced anxiety and falls, attributed to dizziness, with their resultant morbidity.

Compared with patients without dizziness, in the ED, dizzy patients undergo more testing and more imaging, have longer ED lengths of stay, and are more likely to be admitted. The large majority of brain imaging is CT, which has little diagnostic value in patients with dizziness. In 2011, approximately 12% of the estimated $4 billion is related to brain imaging, three-quarters of which was due to CT.

The existing paradigm for diagnosing dizziness is based on symptom quality (ie, asking the question, “What do you mean ‘dizzy’?”). This approach is taught in nearly all review articles and textbooks across specialties; however, newer research has shown that its scientific basis and its internal logic lack foundation.

Currently, misdiagnosis in patients with dizziness is a problem in an environment that is paying increasing attention to diagnostic errors. Misdiagnosis of patients with cerebellar stroke can have disastrous consequences. This article reviews the differential diagnosis of acute dizziness in adult patients, analyzes the origin of the traditional symptom quality approach to dizziness, reviews newer data on an approach to diagnosis of dizziness, and suggests a new approach.

The new approach places a heavy emphasis on history and physical examination. Using these techniques, emergency physicians can improve the care of patients with dizziness by more frequently and confidently making a specific diagnosis. When a confident diagnosis is made of a peripheral problem, time-consuming consultation, expensive imaging, and hospitalization become unnecessary. When the evaluation suggests a central problem, especially stroke, steps can be taken to diagnose and treat the offending vascular lesion and institute secondary prevention measures.

This new approach to the ED patients with dizziness should improve diagnostic accuracy, reduce length of stay and resource utilization, and likely improve overall patient outcomes.

Introduction

Approximately 3.5% of emergency department (ED) visits are for dizziness. Numerous conditions, some benign and self-limiting and others extremely serious, can present with dizziness. This is classic emergency medicine — sorting out the large majority of patients with a given chief complaint who have a self-limiting or easily treatable condition from the smaller number who have life-threatening, limb-threatening, or brain-threatening problems. Increasingly, physicians are charged with performing this diagnostic process using fewer resources. As of 2013, the direct ED-related costs of care for patients with dizziness in the United States was estimated to approach $4 billion. In addition to economic cost, there is additional cost in terms of patient-experienced anxiety and falls, attributed to dizziness, with their resultant morbidity.

Compared with patients without dizziness, in the ED, dizzy patients undergo more testing and more imaging, have longer ED lengths of stay, and are more likely to be admitted. The large majority of brain imaging is CT, which has little diagnostic value in patients with dizziness. In 2011, approximately 12% of the estimated $4 billion is related to brain imaging, three-quarters of which was due to CT.

The existing paradigm for diagnosing dizziness is based on symptom quality (ie, asking the question, “What do you mean ‘dizzy’?”). This approach is taught in nearly all review articles and textbooks across specialties; however, newer research has shown that its scientific basis and its internal logic lack foundation.

Currently, misdiagnosis in patients with dizziness is a problem in an environment that is paying increasing attention to diagnostic errors. Misdiagnosis of patients with cerebellar stroke can have disastrous consequences. This article reviews the differential diagnosis of acute dizziness in adult patients, analyzes the origin of the traditional symptom quality approach to dizziness, reviews newer data on an approach to diagnosis of dizziness, and suggests a new approach.

The new approach places a heavy emphasis on history and physical examination. Using these techniques, emergency physicians can improve the care of patients with dizziness by more frequently and confidently making a specific diagnosis. When a confident diagnosis is made of a peripheral problem, time-consuming consultation, expensive imaging, and hospitalization become unnecessary. When the evaluation suggests a central problem, especially stroke, steps can be taken to diagnose and treat the offending vascular lesion and institute secondary prevention measures.

This new approach to the ED patients with dizziness should improve diagnostic accuracy, reduce length of stay and resource utilization, and likely improve overall patient outcomes.

Differential diagnosis of acute dizziness

Part of the problem is that numerous disorders and conditions that span multiple organ systems can present with acute dizziness. Many of these diagnoses are benign whereas others are life threatening. A study from the National Hospital Ambulatory Medical Care Survey (NHAMCS) database of patients seen in all varieties of hospitals over a 13-year period identified 9472 patients with dizziness. These data suggest that most patients have general medical (including cardiovascular) diagnoses (approximately 50%), otovestibular diagnoses (approximately 33%), and neurologic (including stroke) diagnoses (approximately 11%). This breakdown has some face validity to practicing emergency physicians, for whom general medical conditions outnumber vestibular and neurologic diagnoses in real-life practice.

Studies done on large administrative databases have the limitation that the accuracy of the charted diagnosis is unknown. In the NHAMCS study, 22% of patients received a symptom-only diagnosis (eg, dizziness, not otherwise specified). Although assigning a diagnosis of the presenting symptom is not uncommon in emergency medicine practice, a symptom-only diagnosis was almost 2 times more common in dizzy patients than in all other patients (22.1% vs 8.4%, odds ratio [OR] = 3.1, P <.001). In addition, even if a specific vestibular diagnosis is made, such as BPPV or an acute peripheral vestibulopathy, use of imaging and treatment with medications is not in accordance with best evidence.

In the NHAMCS study, prospectively defined “dangerous” diagnoses (various cardiovascular, cerebrovascular, toxic, metabolic and infectious conditions in which the possibility of a poor outcome without treatment was likely) were found in 15% of patients and the proportion increased with age (21% dangerous diagnoses in patients >50 years vs 9.35 in patients ≤50, P <.001). Among 15 dangerous causes analyzed, the most commonly recorded were fluid and electrolyte disturbances (5.6%), cerebrovascular diseases (4.0%), cardiac arrhythmias (3.2%), acute coronary syndromes (1.7%), anemia (1.6%), and hypoglycemia (1.4%). Some dangerous causes of dizziness, such as adrenal insufficiency, aortic dissection, carbon monoxide intoxication, pulmonary embolus, and thiamine deficiency, are treatable causes that are important but rare.

How does this study compare to others? One older single-institution study analyzed 125 patients prospectively identified over a 16-month period ; 43% had a diagnosis of a peripheral vestibular problem and 30% had a “serious” diagnosis. Another much larger prospective single-institution Chinese study of adult ED patients with dizziness reported results of 413 patients recruited over just 1 month. A central nervous system (CNS) cause was found in 23 patients (6%).

Two retrospective studies also provide relevant data. One study was done in a German ED of 475 consecutive dizzy patients who were seen by a neurologist during the index ED visit. The initial diagnoses assigned by the neurologists were benign in 73% of cases and serious (mostly cerebrovascular and inflammatory CNS disease) in 27% of cases. Overall, the 2 most common diagnoses were BPPV (22%) and stroke (20%). In follow-up by a neurologist blinded to the ED diagnosis, 44% of diagnoses (previously made by a neurologist in the ED) were changed. More than half of these diagnostic changes were from a serious to a benign diagnosis, which errs toward patient safety but is more resource intensive than necessary. In approximately 1 patient in 7, the error was from benign to serious (5 patients diagnosed with vestibular neuritis and 1 with vestibular migraine, all reassigned to stroke), a dangerous misdiagnosis.

The other study analyzed patients who had an ED triage diagnosis of dizziness, vertigo, or imbalance as a primary symptom, collected over a 3-year period, and identified 907 patients (only 0.8% of all ED patients over that period of time), suggesting a very targeted selection compared with other large studies. Of the 907 patients, 1 in 5 was admitted (68% to an ICU). The most common admitting services were medicine (41% of admissions), cardiology (32%), and neurology (24%). Of the 907 patients, most had either benign conditions, such as peripheral vestibular problems (32%), orthostatic hypotension (13%), and migraine (4%) and in a full 22% could not be specifically diagnosed. Serious neurologic disease was found in 49 patients (5%), of which 37 were cerebrovascular. Finally, only 2 patients with serious neurologic disease presented with isolated dizziness.

Many of these studies were based on the emergency physician diagnosis. In a Swiss study of 951 patients referred (not all from an ED) to a multidisciplinary dizziness clinic, there was a significant change in the final diagnoses of dizzy patients. The final diagnosis of “undetermined” fell by 60% and BPPV, multisensory dizziness, and vestibular migraine were all significantly underdiagnosed by the referring doctors. This, coupled with the fact that emergency physicians often assign a symptom-only diagnosis, suggests an inherent, but unknown, incorrect diagnosis rate of dizzy patients in the ED.

The incidence of important CNS disease in adult ED patients with dizziness is approximately 5%. The high-end outlier is Royl and colleagues’ study that reported that 27% of patients have serious CNS causes, skewed by the fact that the study was conducted in a neurologic ED. Various studies have tried to identify risk factors for ED dizzy patients with CNS causes. One ED study of dizzy patients found that abnormal gait and subtle neurologic deficits on neurologic examination were associated with a CNS cause. Overall, the risk factors include increasing age, vascular risk factors, history of previous stroke, complaint of “instability,” and focal neurologic findings ( Table 1 ).

Taken as a whole, these data suggest the following conclusions:

• 1.
  

  Most adult patients who present to the ED with acute dizziness have general medical or cardiac conditions.

  
  
• 2.
  

  Although benign vestibular diseases are much more common than CNS causes of dizziness, when emergency physicians make these (benign) diagnoses, their use of imaging and meclizine is not in accordance with best available evidence.

  
  
• 3.
  

  Of the CNS causes, acute cerebrovascular disease (ischemic stroke or transient ischemic attack [TIA]) is the most common cause and misdiagnosis in the ED is not uncommon in these patients.

Because some of the underlying reasons for this situation have to do with the use of the prevailing traditional symptom quality approach to dizziness, an in-depth analysis of its origin is important to this discussion.

Origin of the symptom quality approach to diagnosing dizziness

Prior to the publication of their article in 1972 (by a neurologist, David Drachman, and an ear, nose, and throat surgeon, Cecil Hart), there was no well-accepted organized algorithmic approach to the diagnosis of an acutely dizzy patient. Drachman and Hart created an outpatient dizziness clinic. Over a 2-year period, they enrolled 125 patients who had to be available to return to their clinic for 4 half days of testing. The study suffers from several shortcomings ( Box 1 ).

The small number of subjects enrolled over a 2-year period suggests a highly select group of individuals. These were not representative ED patients with acute blood loss, ectopic pregnancy, sepsis, pneumonia, dehydration acute strokes, arrhythmias, and other symptoms due to toxic, metabolic, or infectious conditions. One-quarter of the 125 enrolled patients were rejected for various reasons (see Box 1 ).

The patients were first asked questions to describe their “subjective experience of dizziness” to separate and classify “all complaints of dizziness into 4 types: (1) a definite rotational sensation, (2) a sensation of impending faint or loss of consciousness, (3) disequilibrium or loss of balance without head sensation and (4) ill defined ‘lightheadedness’ other than vertigo, syncope or disequilibrium.” The methods then state, “Once the type of complaint has been sorted out, secondary inquiries were sought to identify related neurological, otological, cardiac, psychiatric, gastrointestinal, visual or other symptoms.”

Patients were then asked about positional triggers and the timing of their dizziness. Finally, a comprehensive battery of physical examination testing occurred, including detailed vital sign testing; neurologic, otologic, and ophthalmologic examinations; and a variety of tests to provoke the dizziness at the clinic. Given the era, the only brain imaging that was done was plain films of the skull. At the end of the process and unblinded to the data and dizziness category, the first author made a final diagnosis.

There was no long-term follow-up or validation of the final diagnoses, which were intrinsically linked to the dizziness category that the patients were assigned at the onset. Once the type of dizziness was known, so too was its differential diagnosis. There were other issues related to diagnoses not recognized at the time the research was done (see Box 1 ). Although this was a landmark study published in a prominent journal in 1972, it has important limitations to current use in undifferentiated patients. In fairness, the investigators concluded, “this study should serve as the point of departure for further investigation.” The medical world, however, largely accepted this symptom quality approach by perpetuating that the first question to be asked of the dizzy patient should be, “What do you mean, ‘dizzy’?” and that the response generated a particular differential diagnosis (eg, vestibular problems if vertigo, cardiovascular disorders if lightheadedness or near-syncope, neurologic issues for disequilibrium, and largely psychiatric causes if other).

Reasons why the symptom quality approach lacks scientific validity

For the symptom quality approach to work, 2 facts must be true. First, patients should be able to reliably and consistently chose 1 (and only 1) dizziness type. Secondly, each symptom type should be tightly linked with a given differential diagnosis. Both facts are demonstrably false.

Patients do not chose a single dizziness type. Sensory symptoms are difficult for many patients to describe. Patients with dizziness may use words like, “dizzy,” “lightheaded,” “spinning,” “rocking,” “vertigo,” “giddy,” “like I’m going to faint,” “off balance,” “spacey,” and others to describe what they feel. For this article, I use the word dizziness in a general way (incorporating all of these descriptors).

In 2007, researchers published a study about how dizzy patients in an ED respond when asked a series of questions related to symptom quality. Research assistants asked a series of ED patients with dizziness a battery of questions aimed at determining symptom quality and timing and triggers of the dizziness. The questions were asked and then reasked an average of 6 minutes later (the second time in a different sequence). More than 60% of the patients chose more than 1 dizziness type. In response to the same questions (the second time), more than 50% of the patients changed their primary dizziness type. The responses to timing and triggers of dizziness were more consistent and reliable between the first and second responses.

This 1 study severely undercuts the logic of a diagnostic process based on symptom quality. How should a patient be evaluated who endorses both lightheadedness and vertigo? If a patient answers “vertigo” the first time, then “disequilibrium” the second time, how should the work-up proceed?

A patient with chest pain is not evaluated differently if the pain is described as “sharp,” “dull,” “discomfort,” or “pressure.” Pain described as “sharp” may be more likely to be pulmonary embolism or pleurisy and ”dull” more likely an acute coronary syndrome, but the descriptor of the pain is not used in a binary way. In a patient with chest pain, it is the timing and triggers that are more important in rank-ordering a differential diagnosis.

Pain that is brought on by exercise and rapidly resolves with rest suggests angina (or aortic stenosis or right-sided heart strain due to obstruction from a pulmonary embolism). Pain that is constant for 10 hours is more likely to be a myocardial infarction (or an aortic dissection or a pulmonary infarction) rather than unstable angina. Chest pain regularly triggered by swallowing is more likely an esophageal problem than a cardiac or pulmonary one. I take histories of virtually all chief complaints using the concept of timing and triggers.

Another concept that physicians use regularly to construct a differential diagnosis is that of context. The presence of associated symptoms is also important. Patients are thought of differently with chest pain associated with (1) leg swelling and dyspnea, (2) productive cough and fever, or (3) hypotension, unilaterally diminished breath sounds, and distended neck veins. The same logic applies for patients with headache, abdominal pain, or dyspnea. It is not simply the word that the patient uses that informs a differential diagnosis but also the timing, triggers, and associated symptoms and epidemiologic context. It should be no different with dizziness.

Finally, the differential diagnosis is not tightly linked with a given use of the descriptors. The use of the word “vertigo” was not associated with a higher incidence of stroke in a large series of ED patients with dizziness. Patients with a cardiovascular cause of dizziness do endorse “vertigo” in almost 40% of cases. Patients with BPPV often say they feel lightheaded and not vertiginous, especially elderly patients. The reality is that the differential diagnosis should NOT be based on the word but rather on the timing, triggers, associated symptoms, and epidemiologic context. Yet physicians often use a generalized approach to a patient with “vertigo” without considering the timing and triggers.

For all these reasons, the symptom quality approach to dizziness is not based on strong science. Nevertheless, it is the predominant paradigm used across specialties. Despite this, or to some extent, because of it, significant misdiagnosis of dizzy patients exists.

Misdiagnosis of patients with dizziness and resource utilization

Misdiagnosis of patients with dizziness is common. In the German ED study, neurologists seeing patients made diagnostic errors in 44% of patients. The investigators of that study found 3 factors that contributed to misdiagnosis. First, subsequent clinical course evolved, making the ultimate diagnosis more clear. This factor played a role in 70% of misdiagnoses. This is a regular event in emergency medicine, in which patients whose symptoms evolve in a variable way are seen. What is obvious the next day, or even a few hours later, is not always clear on initial presentation. The other 2 factors were insufficient brain imaging (mostly MRI, found to be a factor in half of cases) and failure to screen for vascular risk factors using advanced testing, such as echocardiography, long-term telemetry, or Doppler ultrasound of cervical arteries (24% of cases). There has never been a head-to-head comparison of emergency physicians versus neurologists diagnosing patients with dizziness at the same phase of care (and likely never will be), but this German study clearly shows that dizziness is complicated, even to those with specialized training and focus.

In another study of 1091 dizzy patients in US EDs, emergency physicians documented comments about nystagmus in 887 (80%), of whom nystagmus was documented to be present in 185 (21%). No other information beyond presence or absence was recorded in 26% of the 185 patients and sufficient information to be diagnostically useful was only recorded in 10 patients (5.4%). Of patients given a peripheral vestibular diagnosis, the description of the nystagmus conflicted with that diagnosis. This illustrates a knowledge gap in emergency physicians’ understanding of nystagmus: what to look for, how to report it, and, most importantly, how to use the findings to their advantage.

Reporting presence or absence of nystagmus in a dizzy patient is akin to reporting simple presence or absence of electrocardiographic abnormalities in a chest pain patient. The mere presence or absence of abnormalities is not the key finding. In the latter example, 3 patients with chest pain, 1 with sinus tachycardia and new right bundle branch block, another with PR segment depression with diffuse ST segment elevations, and a third with focal ST segment elevations in leads 2, 3, and aVF with reciprocal changes laterally all have very different abnormalities that have a different significance. Similarly, in a patient with an AVS ( Table 2 ), the findings of direction-fixed horizontal nystagmus versus direction-fixed vertical nystagmus versus direction-changing nystagmus have different significance (discussed later). A recent review illustrates how to use the physical examination in dizzy patients.

Multiple studies find that patients with an AVS that superficially appears to be a peripheral process in fact have posterior circulation strokes. In 1 study, almost 3% of patients referred to a ear, nose, and throat clinic for vertigo had a missed cerebellar stroke. There are 2 major reasons that missed stroke is an important misdiagnosis. The first is that the underlying vascular mechanism goes untreated, leaving the patient vulnerable to another stroke and the second is that some patients develop posterior fossa edema that can be fatal. Although lost opportunity for thrombolysis is often suggested as a third negative consequence of missing a posterior circulation stroke, many of these patients have minor deficits and are not thrombolysis candidates. Some even have an National Institutes of Health Stroke Scale score of zero.

Younger age and dissection as a cause were found risk factors for missed cerebellar stroke. Posterior circulation location is a risk factor for stroke misdiagnosis in general. To put these data into some context, only a small number (0.18%–0.63%) of patients who are seen in the ED, diagnosed with a benign or peripheral vestibular diagnosis, return to the ED within 30 days and are hospitalized with a cerebrovascular diagnosis. On a relative basis, this is a small number. Because dizziness is so common, however, this small fraction of a large number suggests that many thousands of patients have a missed diagnosis of an acute cerebrovascular syndrome (stroke TIA) each year.

The other side of the coin is a lack of recognition of common peripheral vestibular problems, the most common of which are BPPV and vestibular neuritis (or labyrinthitis if hearing is also involved). Lack of familiarity with these diagnoses may lead to undertreatment, incorrect treatment, and resource overutilization. In addition, rarely, a patient who seems to have BPPV actually has a central mimic, called central paroxysmal positional vertigo (CPPV), discussed later.

A recent review of misdiagnosis of patients with dizziness suggested 5 common pitfalls. These are over-reliance on a symptom quality approach to diagnosis, underuse of a timing and triggers approach, lack of familiarity with key physical examination findings, overweighting traditional factors such as age and vascular risk factors to screen patients, and over-reliance on CT. Although stroke is more common in older individuals, young patients do have strokes, a fact that may contribute to misdiagnosis.

A new paradigm to diagnose patients with acute dizziness – ATTEST

A new diagnostic paradigm is based on the timing, triggers and context of the dizzy symptoms. It is possible that this new paradigm will reduce misdiagnosis. Although the use of this new paradigm has not been proved to reduce misdiagnosis, I have been using and teaching it for 6 years, and my experience is that it allows confidently making a specific diagnosis more frequently than the traditional paradigm (for which there are also no quality data that it helps make a specific diagnosis). Different mnemonics have been used but the basic idea is that it is the timing, triggers, evolution, and context of symptoms that should drive the work-up rather than the specific words that a patient uses to describe dizziness. I currently favor the mnemonic, ATTEST – A (associated symptoms), TT (timing and triggers), ES (bedside examination signs), and T (additional testing as needed).

This new paradigm may seem like a radically new way of approaching a dizzy patient, but this is only because the traditional symptom quality approach is so deeply engrained in how this subject has been taught. Using a timing and triggers approach is no different from taking a history in any other patient, for example, someone with chest pain. A differential diagnosis is not created based only on the descriptor of the pain (dull, sharp, pressure, or tightness) but rather on timing and triggers of the pain: “Is the pain intermittent or continuous?” “Did the pain start abruptly or gradually?” “Does the pain increase with deep breath?” and “Is it triggered by physical exertion, or by eating?” These are some of the components of the history that drives the differential diagnosis. It is no different with a dizzy patient.

Using this paradigm, there are 4 timing and triggers categories that are important for emergency physicians (see Table 2 ). In the traditional paradigm, a patient who endorsed “vertigo” would get an evaluation to try to diagnose peripheral vestibular CNS causes of dizziness. This has led to confusion, for example, using imaging for diagnosis and meclizine for treatment no matter what the diagnosis is. That is, the traditional paradigm tend to treat all patients with peripheral vertigo the same whereas the 2 most common by far being BPPV and vestibular neuritis should be treated differently. The following sections review the presentation, differential diagnosis, and appropriate testing to make a specific diagnosis for each of the timing and triggers categories.

Acute vestibular syndrome

Spontaneous AVS is defined as the acute onset of persistent dizziness in association with nausea or vomiting, gait instability, nystagmus, and head-motion intolerance that lasts days or weeks. Patients are usually symptomatic at the time of assessment and focused physical examination is often diagnostic. The most common cause is an acute peripheral vestibulopathy known as vestibular neuritis (dizziness only) or labyrinthitis (dizziness plus hearing loss or tinnitus). The most frequent dangerous cause is posterior circulation ischemic stroke, generally in the cerebellum or lateral brainstem. A distant third most common cause is multiple sclerosis. Uncommon causes of an isolated AVS include cerebellar hemorrhage and several rare, but often treatable, autoimmune, infectious, or metabolic conditions. The spontaneous AVS is to be distinguished from a triggered AVS, which is not discussed further in this article because the cause is usually obvious, such as posttraumatic dizziness or diphenylhydantoin toxicity.

An important concept is that patients with an AVS generally experience worsening of their symptoms with head movement or during provocative testing (such as if performing a Dix-Hallpike maneuver). These exacerbating features should not be mistaken for head movement triggers that facilitate diagnosis in episodic vestibular syndrome (EVS) patients. Confusion on this point probably contributes to difficulty differentiating BPPV from vestibular neuritis. Acute BPPV patients occasionally complain of more persistent symptoms that may be due to repeated triggering symptoms with small, inadvertent head movements or anticipatory anxiety about moving. This can usually be teased out by careful history taking. When such patients lack obvious features of vestibular neuritis or stroke, the Dix-Hallpike and supine roll tests should be performed to assess for an atypical, AVS-like presentation of BPPV.

Vestibular neuritis is a benign, self-limited, presumed viral or postviral inflammatory condition affecting the vestibular nerve and causing spontaneous AVS. This can be thought of as similar to Bell palsy but involving the vestibular portion of the 8th nerve rather than the 7th nerve. Some cases are associated with inflammatory disorders (eg, multiple sclerosis or sarcoidosis), but most are idiopathic and possibly linked to herpes simplex infections. The idiopathic form is generally monophasic, although 25% have a single, brief prodrome in the week prior to the attack, and approximately 16% recur months or years later. High-field-strength MRI with high-dose gadolinium has demonstrated vestibular nerve enhancement in select cases, but typical vestibular neuritis shows no abnormality on routine MRI with contrast. The diagnosis is, therefore, clinical and requires excluding other causes. A related condition, herpes zoster oticus (Ramsay Hunt syndrome type 2), may present with AVS, usually in conjunction with hearing loss, facial palsy, and a vesicular eruption in the ear or palate.

Posterior fossa strokes may present with spontaneous AVS mimicking vestibular neuritis (or labyrinthitis if auditory symptoms are present). The prevalence of cerebrovascular disease in patients presenting to the ED with dizziness is 3% to 6%, but among AVS presentations it is estimated at approximately 25%. Almost all (96%) are ischemic strokes, rather than hemorrhages. CT sensitivity for acute ischemic stroke is low (perhaps as low as 16% in the first 24 hours, possibly less in the posterior fossa). Therefore, CT cannot rule out ischemic stroke in AVS, a fact often contributing to misdiagnosis. Even MRI with diffusion-weighted imaging misses 10% to 20% of strokes in spontaneous AVS during the first 24 to 48 hours after symptom onset, and repeat delayed imaging (3–7 days post–symptom onset) may be required to confirm the presence of a new infarct.

Fortunately, a physical examination can help make the distinction between vestibular neuritis and posterior circulation stroke with greater sensitivity than early MRI. These 2 studies were done by neuro-otologists performing a targeted oculomotor examination consisting of 3 components — the head impulse test (HIT) and testing for nystagmus and skew deviation (head impulse, nystagmus, and test of skew [HINTS]). One other study showed similar accuracy when performed by stroke neurologists (not neuro-otologists). Preliminary evidence suggests that components of this approach (nystagmus testing and HIT) can be successfully used by specially trained emergency physicians. My own anecdotal experience also suggests that with some training, emergency physicians can perform and interpret this examination. Because this approach has not been fully validated when used by nonspecialists, I have added 2 additional components that should be a part of the basic evaluation of the acutely dizzy patient — the general neurologic examination and testing of gait.

I do not perform these tests in the order of the HINTS mnemonic but rather in the following order:

• 1.
  

  Nystagmus testing

  
  
• 2.
  

  Skew deviation

  
  
• 3.
  

  HIT

  
  
• 4.
  

  General neurologic examination, focusing on cranial nerves, including hearing, cerebellar testing, and long-tract signs

  
  
• 5.
  

  Gait testing

There are 2 reasons for my preferred sequence. First, I try to start with the least intrusive parts of the examination and, second, nystagmus testing is the component that helps the most, in part by its presence or absence and in part by its quality. Once any 1 component tests positive for a central cause, then the patient’s disposition (admission for further neurologic evaluation) is clear. Although all 5 components are part of a complete examination for a patient with an AVS, worsening a patient’s symptoms with further intrusive testing (eg, testing gait that provokes vomiting) does not change the disposition but causes the patient’s worst symptoms.

Furthermore, nystagmus helps to anchor and inform the rest of the process. All or nearly all patients with an AVS due to a vestibular cause have nystagmus if examined within the first days, so its absence should make 1 question a peripheral or central vestibular process. To be sure that nystagmus is truly absent, it should be tested with visual fixation removed. Experts state that the absence of nystagmus in a patient examined with visual fixation removed essentially rules out a vestibular cause for the dizziness. Subspecialists typically use Frenzel lenses to remove visual fixation, neither available nor common practice in emergency medicine practice. There are, however, easy bedside alternatives for emergency physicians to take away visual fixation. A newly described one uses a new less bulky set of glass lenses. Another method is to use a penlight to reduce visual fixation. To do this, cover 1 eye and intermittently shine a light close to and directly at the other eye, telling the patient not to look at the light, just to stare off in the distance. Observe for nystagmus appearing when the light is blocking fixation. Finally, simply take a piece of white paper and place it close to the patient’s eyes (telling the patient to “look through the paper”) and examine the nystagmus from the side. Note that these other steps are only needed if there is no nystagmus with the basic examination.

If nystagmus is truly absent, then this is unlikely to be a vestibular process and, therefore, HIT is probably not useful and may yield false information.

Another potential issue is that the degree (or amplitude) of nystagmus can fluctuate markedly over hours. To some extent, this may represent the natural history of the underlying pathology as the CNS accommodates to the abnormal physiology from vestibular neuritis. Medications that are often (appropriately) used in the ED, however, to reduce symptoms, such as ondansetron or a benzodiazepine, may affect the rate at which the nystagmus dampens.

With all that said, the basic clinical test for nystagmus is simple. Have the patient look straight in neutral or primary gaze and observe for eye movements. By convention, the direction of nystagmus is named by the direction of the fast component. For example, if the eyes drift leftward and snap back horizontally to the right, this is termed right-beating horizontal nystagmus. With some practice this is easy to see and describe because it is the details of the nystagmus, not simply its presence, that is most diagnostically important. After observing for nystagmus in primary gaze, test for gaze-evoked nystagmus by having the patient look to the right and then to the left, each for several seconds, and observe for the presence of nystagmus and the direction of its fast-beating component. The patient only needs to move the eyes 15° to 20° off center when testing for gaze-evoked nystagmus because many normal individuals have a few beats of horizontal nystagmus on full end gaze. This physiologic nystagmus is generally very low amplitude and extinguishes quickly. Table 3 shows the typical findings for patients with the oculomotor examination for patients with the AVS. Direction-changing gaze-evoked nystagmus or nystagmus that is pure torsional or vertical should be considered central in origin (and in the setting of a AVS, a stroke).

Table 3

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

Neurologic Emergencies Treatment of Acute Ischemic Stroke Headache in the Emergency Department Acute Nontraumatic Back Pain Diagnosis and Management of Acute Intracerebral Hemorrhage Subarachnoid Hemorrhage