Using an algorithmic approach to acutely dizzy patients, physicians can often confidently make a specific diagnosis that leads to correct treatment and should reduce the misdiagnosis of cerebrovascular events. Emergency clinicians should try to become familiar with an approach that exploits timing and triggers as well as some basic “rules” of nystagmus. The gait should always be tested in all patients who might be discharged. Computed tomographic scans are unreliable to exclude posterior circulation stroke presenting as dizziness, and early MRI (within the first 72 hours) also misses 10% to 20% of these cases.
Summarize the current diagnostic paradigm and its origins.
Promote a new evidence-based algorithmic approach that better serves clinicians.
Discuss the extent of and reasons for misdiagnosis.
Describe physical examination techniques that aid bedside diagnosis.
Clarify limitations of diagnostic imaging.
Dizziness, the chief complaint in approximately 3% of emergency department (ED) patients, has a broad differential diagnosis. Although most of the causes are benign, serious cerebrovascular conditions account for a small but significant proportion. Emergency physicians are tasked with accurately distinguishing benign from serious causes while being parsimonious with resource utilization.
The traditional diagnostic paradigm is faulty and may actually contribute to misdiagnosis. Brain computed tomography (CT) is often used but has extremely limited utility. Recent studies show that physical examination can be extremely useful in distinguishing benign from serious causes of dizziness. Because treatment follows a correct diagnosis, this article’s focus includes the following:
Summarize the current diagnostic paradigm and its origins
Promote a new evidence-based algorithmic approach that better serves clinicians
Discuss the extent of and reasons for misdiagnosis
Describe physical examination techniques that aid bedside diagnosis
Clarify limitations of diagnostic imaging
History, definitions, and background
The traditional “symptom quality” approach to diagnose patients with dizziness starts by asking the patient, “What do you mean by ‘dizzy’?” Their response (true vertigo vs lightheadedness vs imbalance/disequilibrium vs “other”) drives the differential diagnosis and the evaluation. This diagnostic paradigm stems from a study published in 1972.
The investigators established a “dizziness clinic” to which patients were referred and underwent several days of clinical evaluation. Based on these results, the lead author assigned a diagnosis. Despite serious flaws and no prospective validatation, , the “symptom quality” approach became the prevailing paradigm. Newer evidence shows that its underlying logic is faulty and that a newer paradigm based on “timing and triggers” is more consistent with current evidence.
Relevant anatomy, physiology, and pathophysiology
Dizziness from general medical causes may be mediated by neural dysfunction due to factors such as fever, hypotension, anemia, medication side effects, electrolyte abnormalities, and others. Dizziness due to vestibular and neurologic causes have more clear-cut mechanisms that require some familiarity of basic anatomy and physiology.
The end organs in the labyrinth, the peripheral vestibular apparatus including semicircular canals, utricle, and saccule (organ of balance), and the cochlea (organ of hearing) lie in the temporal bone ( Fig. 1 ). The vestibular apparatus includes 3 paired semicircular canals that sense rotational motion and the utricle and saccule that sense linear motion ( Fig. 2 ). These interconnected structures are filled with endolymph. Hair cells in the utricle and saccule are covered by a gelatinous otolithic membrane in which calcium carbonate particles (otoliths) are embedded. As fluid moves in a semicircular canal, it displaces the cupula (see Fig. 2 ), which generates the sense of motion.
The vestibular nerves supply the end organs in the labyrinth.
With linear head movement (see Fig. 2 ), gravity causes the heavier otoliths to move, displacing the hair cells in the utricle (vertical movement) and saccule (horizontal movement). With angular motion, fluid motion displaces the cupula that lies within the dilated end portion of the semicircular canals (ampulla). This displacement of the cupula is transduced into electrical energy, which is transmitted to the brainstem via the vestibular nerve.
Benign paroxysmal positional vertigo (BPPV) is caused by dislodged otoliths from the utricle that migrate into one of the semicircular canals, most commonly 1 of the 2 posterior canals because they are the most gravity dependent, therefore simulating motion that is nonexistent ( Fig. 3 ).
The eighth cranial nerve comprises the vestibular and the cochlear nerves. Signals from the vestibular labyrinth travel through the vestibular nerve to the vestibular nuclei in the brainstem, then connects with the cerebellum, oculomotor system, cortex, and spinal cord.
The connections with the oculomotor system contain a reflex arc, the vestibulo-ocular reflex (VOR), that helps to fix one’s gaze on a visual target when the head is moving, an important survival trait. The horizontal head impulse test (HIT), which was first described in 1988, tests the VOR. This arc does not loop through the cerebellum, which is why the HIT is “negative” in cerebellar stroke. However, the cerebellum does modulate the VOR due to connections between the vestibular nuclei and the cerebellum.
The posterior circulation is fed by paired vertebral arteries that ascend in the neck and fuse to form the basilar artery. Just before fusing, they give off the posterior inferior cerebellar artery (PICA), the anterior inferior cerebellar artery (AICA), and the superior cerebellar artery. These are the major branches of the basilar artery before it splits into terminal branches—the paired posterior cerebral arteries. The PICA nourishes the lateral medulla and inferior cerebellum. The AICA nourishes the lateral pons including the vestibular nerve root entry zone. A branch of the AICA, the labyrinthine artery, supplies the peripheral labyrinth ( Fig. 4 ). This explains why strokes of the lateral pons and the labyrinth are associated with a “positive” or falsely “reassuring” HIT.
The evidence base for diagnosis of dizzy patients is weak and because it specifically relates to routine emergency medicine practice, even weaker. In some of the most important articles about the diagnosis of dizzy patients, the interventions were done by neuro-otologists. That said, the evidence base for the diagnosis of dizziness is growing.
Differential diagnosis and misdiagnosis
The individual causes of dizziness are too numerous to be clinically useful. Instead, one must use an organized diagnostic method. In an analysis of 9472 patients from a large National Hospital Ambulatory Medical Care Survey database of ED patients, the causes of dizziness listed in the charts by the attending emergency physicians were as follows:
General medical (toxic, metabolic, and infectious) conditions: 49%
Otologic or vestibular conditions: 33%
Cardiovascular causes: 21%
Respiratory conditions: 12%
Neurologic diseases: 7%
Cerebrovascular causes: 4%
Predefined “dangerous” diagnoses (mostly serious cardiovascular, cerebrovascular, and general medical conditions) accounted for 15% of cases and were twice as likely in patients older than 50 years of age.
Misdiagnosis of patients with acute dizziness, especially misdiagnosis of cerebellar and brainstem stroke, is a common problem not restricted to emergency physicians. In a German study of 475 ED patients with dizziness assessed by a neurologist, nearly 50% of diagnoses were changed by a second neurologist (blinded to the initial diagnosis) on follow-up. Importantly, evolution of the clinical course over time after the initial evaluation (obviously not available to physicians who are first diagnosing the patients) was a factor in 70% of the misdiagnosed patients.
Patients with anterior circulation strokes often present with lateralizing weakness, which scores more points on the National Institute of Health Stroke Scale and receives more attention in the medical literature than posterior circulation strokes. The latter are misdiagnosed more than twice as often as anterior circulation events. Our use of stroke heuristics emphasizes lateralizing deficits, which are often subtle or absent in patients with posterior circulation strokes.
Another reason for misdiagnosis is the “needle in the haystack” phenomenon. Very few ED patients with dizziness are having strokes. In one study of 1666 adult ED patients with dizziness, fewer than 1% of those with isolated dizziness had a cerebrovascular cause. Other studies found that of ED patients with dizziness who are discharged with a peripheral vestibular diagnosis, less than a half of 1% (range = 0.14–0.5%) are subsequently hospitalized with a stroke. Although low, given the high number of dizzy patients, the absolute number is significant.
Other studies that “look backwards,” analyzing patients who are ultimately diagnosed with strokes (only focusing on the needles), show that between 28% to 59% of patients with cerebellar strokes are missed in the ED. In one study of 240 patients with cerebellar strokes, 10% (25 patients) presented with isolated dizziness that mimicked peripheral lesions. Misdiagnosis rates look very different depending on whether or not the study looks “forward” or “backward.” Factors associated with misdiagnoses of patients with dizziness include use of the traditional symptom quality approach, lack of familiarity with eye movement examinations, overweighting of age and other traditional vascular risk factors, and overreliance on CT scanning, as well as younger age, vertebral dissection as a cause and a presentation of dizziness. , ,
Lack of familiarity with some of the eye findings is an important knowledge gap. Nystagmus in particular, is often underutilized or incorrectly utilized by emergency clinicians. The presence or absence of nystagmus is important, but the details of the nystagmus are far more important diagnostically. In a study of 1091 dizzy patients in US EDs, physicians used templates to document the presence or absence of nystagmus in 887 (80%). Nystagmus was documented as present in 185 (21%), of which, diagnostically meaningful information was recorded in only 10 (5.4%). Of patients given a peripheral vestibular diagnosis, the nystagmus description often conflicted with that diagnosis in 81%.
Finally, it should be emphasized that CT scanning, the typical ED “go to” brain imaging test, is of extremely limited utility in these patients, often giving false reassurance. In a study of ED patients with dizziness who were discharged with a benign ICD-9 “dizzy” diagnosis and followed-up for 30 days, patients who returned with a stroke were 2.3 times more likely to have had a CT on the first visit, suggesting that physicians were correctly determining worrisome patents but then applying the wrong diagnostic test.
According to traditional “symptom quality” paradigm, patients will select either vertigo, lightheadedness, disequilibrium, or other. The first group has vestibular problems; the second group, cardiovascular or general medical problems; the third group, neurologic problems; and the last group, psychiatric disease.
For traditional approach to work, patients must be able to consistently select a single “type” of dizziness, and each dizziness type is tightly associated with specific diagnoses. Neither proposition is true. In a study of ED patients with dizziness, researchers asked questions about the type of dizziness and the timing and triggers of the dizziness. When they reasked the same questions but in a different sequence an average of 6 minutes later, 50% of patients changed the “type” of dizziness that they had selected just minutes before. Frequently, they selected 2 or 3 different dizziness types. However, they were far more consistent about the timing of the dizziness and the factors that triggered it.
In addition, the use of one term versus another (eg, vertigo vs lightheadedness) is not diagnostically meaningful. In the ED study of 1666 acutely dizzy patients, use of the word “vertigo” was not associated with a cerebrovascular diagnosis. In another study of patients with BPPV, 27% endorsed “dizziness” and not “vertigo,” far more commonly in elderly patients. In another review of patients with cardiovascular conditions with dizziness, nearly 40% endorsed vertigo (not lightheadedness, which the traditional paradigm would predict).
Timing and Triggers Approach
Not only is a timing and triggers paradigm more consistent with current evidence but it is the way we take histories from every other patient. One would never base the differential diagnosis of a chest pain patient solely on the descriptor that the patient uses. One elicits the onset, the evolution, the constant or intermittent nature, and factors that trigger or alleviate the pain. For example, it is less important that a patient endorses sharp or dull or even tearing chest pain but rather if it has been intermittent, only occurring with exertion and is relieved by rest.
Rather than basing a differential diagnosis on the descriptive word a patient uses (vertigo or lightheadedness or imbalance), a more logical strategy is based on the timing and the triggers of the dizziness. , ,
The “ATTEST” algorithm uses an evidence-based systematic approach to diagnose acutely dizzy patients. The first three letters in the ATTEST pneumonic ( A ssociated symptoms, T iming, and T riggers) refer to historical information: “What happened?” “When?” “Is the dizziness continuous or intermittent?” “Are there associated symptoms?” “What is the broader context?” ( Fig. 5 ).
Patients without an obvious general medical cause usually fall into 1 of 3 categories: the acute vestibular syndrome (AVS) (acute onset persistent, continuously present dizziness), the triggered episodic vestibular syndrome (t-EVS) (brief episodes of dizziness caused by some obligate trigger), and the spontaneous episodic vestibular syndrome (s-EVS) (spontaneous episodes of variable duration dizziness not triggered by anything).
A = associated symptoms
TT = timing and triggers
ES = examination signs
T = (confirmatory) testing
Although this approach has not been validated in routine ED practice, it is consistent with the way other chief complaints are approached, is rooted in vestibular and neurologic physiology and pathophysiology, and is consistent with current evidence.
Asking about timing and triggers yields 3 acute vestibular categories that are tightly associated with a specific differential diagnosis ( Table 1 ). , , , Patients may have an AVS, an s-EVS, or a t-EVS (see Table 1 ).
|Syndrome||Description||Common Benign Causes||Common Serious Causes||Important Rare Causes|
|AVS||Acute, continuous dizziness lasting days, accompanied by nausea, vomiting, nystagmus, head motion intolerance, and gait unsteadiness||Vestibular neuritis |
|Posterior circulation ischemic stroke||Multiple sclerosis |
Drug/medication side effects or toxicity
|s-EVS||Episodic dizziness that occurs spontaneously, is not triggered, and usually last minutes to hours||Vestibular migraine |
|Posterior circulation TIA||Cardiac dysrhythmia |
|t-EVS||Episodic dizziness triggered by a specific, obligate trigger (typically a change in head position or standing up), and usually lasting <1 min||BPPV |
Orthostatic hypotension caused by benign problems
Orthostatic hypotension due to serious medical illness
|Superior canal dehiscence |
Postural tachycardia syndrome
Vertebral artery rotation (Bow Hunters syndrome)
For practical purposes in patients with an AVS, the major distinction to be made is neuritis versus stroke because approximately 95% of patients with AVS have 1 of those 2 diagnoses. Two to three percent have an initial presentation of multiple sclerosis. , The remainder may have a long list of very uncommon diagnoses. One important uncommon cause is Wernicke encephalopathy (thiamine deficiency).
For patients with an s-EVS, the most common diagnosis by far is vestibular migraine but the important serious diagnosis is TIA. Although posterior circulation TIA presenting as isolated dizziness was long thought to not exist, mounting evidence demonstrates that it does. In one study of 1141 stroke patients, brief episodes of symptoms occurring within the 2 days before the stroke that could be ascribed to posterior circulation ischemia were 36 times more likely in patients who had vertebrobasilar strokes compared with those with anterior circulation strokes.
For patients with a t-EVS, BPPV and non–life-threatening causes of orthostatic hypotension are the common benign causes, and central paroxysmal positional vertigo (CPPV) and serious causes of orthostasis are the life-threatening causes. CPPV is caused by small lesions (mass, multiple sclerosis, or tumor) in the region of the fourth ventricle that can mimic BPPV. ,
General Medical Causes
History and vital signs usually suggest the roughly 50% of patients whose dizziness is caused by general medical causes. The particular descriptive word used by the patient (eg, lightheadedness or vertigo or imbalance) to describe their dizziness is not diagnostically useful.
Consider the following histories in patients who endorse dizziness plus:
Heavy ibuprofen use and black stools
New antihypertensive or anticonvulsant medication use
Moderate mechanism motor vehicle crash
Abdominal pain, vaginal bleeding, and positive pregnancy test
Chest pain and dyspnea in a patient with Factor V Leiden
Each situation suggests a diagnosis or group of diagnoses that would require confirmatory testing. Similarly, the vital signs inform this diagnostic process. If a general medical diagnosis is likely, the authors recommend a brief diagnostic “STOP” that takes less than 1 minute to perform (see Fig. 5 ). , In order to identify disorders that might mimic a general medical condition, the 3 components of the “STOP” are testing for worrisome nystagmus (described in detail later), arm dysmetria, and truncal ataxia. To test for truncal ataxia, simply have the patient sit up in the stretcher without holding onto the side rails. If the “STOP” is reassuring, proceed with management for the presumed condition. If it is worrisome, consider various vestibular or central conditions.
Acute Vestibular Syndrome
If the history does not suggest a general medical condition (or if the “STOP” is worrisome), then the next question to pose is, “is the dizziness persistently present and still present at the time of ED evaluation?” A “yes” answer identifies patients with the AVS, who have the abrupt or rapid onset of dizziness that has lasted hours to days and is still present at the time of examination even when the patient is lying still. The dizziness may decrease when lying still and worsen with head movement, a common occurrence that does not mean that the dizziness has a peripheral cause.
Although the strict definition of the AVS includes the presence of nystagmus, some patients who otherwise fulfill the AVS definition (such as many with cerebellar stroke) do not have nystagmus. Its presence or absence is a key distinction because it affects how one interprets the HIT. , , , Bedside eye findings in these symptomatic patients to distinguish central (stroke) from peripheral (neuritis) causes, referred to as Head Impulse–Nystagmus–Test of Skew (HINTS) testing in patients with AVS. An important caveat is that most of the studies that examine the utility of HINTS have been done with neuro-otologists performing the examinations. , A potential source of confusion, the HINTS acronym, is similar to but different from the abbreviation for the HIT, a component of HINTS.
Although one study found that vascular neurologists can be trained to effectively use the HINTS examination and other European studies of emergency physicians (who received 12 hours of special training using Frenzel lenses to interpret the eye findings) showed the same, , HINTS accuracy in routine practice has not been fully validated. The authors therefore recommend 2 additional components to the examination of patients with the AVS , —a targeted posterior circulation examination and gait testing.
Five questions are posed when doing the physical examination (intentionally in the following sequence) ( Fig. 6 ):
Is there a central pattern of nystagmus?
Is skew deviation present?
Is the HIT worrisome for a central process (ie, absent corrective saccade)?
Are there central nervous system (CNS) findings on the targeted posterior circulation examination?
Can the patient sit up or walk without assistance?