Practice Diagnosis in General Practice


BMJ 2009; 339 doi: (Published 22 September 2009) Cite this as: BMJ 2009;339:b3493
  1. Kevin Barraclough, general practitioner1,
  2. Adolfo Bronstein, professor of neuro-otology2
  1. 1Painswick GL6 6TY
  2. 2Neuro-otology Unit, Division of Neurosciences and Mental Health, Imperial College London
  1. Correspondence to: K Barraclough k.barraclough{at}

Strategies for improving the pattern recognition involved in making a correct diagnosis amount to forcing yourself to use analytic reasoning (doi:10.1136/bmj.b3490); diagnosis of vertigo is an example

A 58 year old woman presented to her general practitioner stating that she woke that day feeling that the room was moving. She had vomited twice. She seemed anxious, was slightly unsteady on her feet, and was hyperventilating. She did not have fever but had a sore throat, slight difficulty swallowing, slight hoarseness, and a red throat. The Hallpike test induced vertigo and nystagmus; the nystagmus was sustained. The general practitioner thought the likely diagnosis was either viral labyrinthitis with pharyngitis, or benign paroxysmal positional vertigo.

Hallpike positioned manouvre: Instruct the patient to keep their eyes open and stare at the examiner's nose. In each position, observe eyes for up to 30 seconds for nystagmus. With the patient sitting upright on a couch, head turned 45 degrees to the right, lie the patient down rapidly until head is dependent. Return patient to upright position.

The diagnostic dilemma

In the 1970s a classification of dizziness was developed, based on an analysis of 125 patients attending a dizziness clinic.1 2 Vertigo was defined as the illusory sense of movement or orientation, and indicates disorders of the labyrinth or brainstem; presyncope was defined as a sense of near faint, typically due to transient hypotension. Disequilibrium of the elderly is often described as a non-specific slight “unsteadiness,” particularly on turning, and indicates poor balance and strength,3 4 and lightheadedness is often associated with dysfunctional breathing or anxiety.

To distinguish vertigo from non-rotatory dizziness, this validated question may be asked: “Did you just feel light headed or did you see the world spin around as though you had just got off a playground roundabout?”5

In our case scenario, the patient has vertigo, a common condition in the community. In a postal questionnaire study of general practice patients aged 18 to 65, 7% of the 2064 responders reported true vertigo in the past year.6 A full time general practitioner may expect to see 10-20 cases of vertigo a year.7

Eliciting the symptom of vertigo narrows the differential diagnosis to disorders of the labyrinth or its central connections. With new onset of vertigo, the diagnostic challenge for the general practitioner is to distinguish the (common and easily treatable) peripheral causes of vertigo from the (very uncommon and more serious) central causes such as a brainstem stroke. The key causes of vertigo are outlined in the table; the clinical features to be elicited are shown in box 1.

Causes of acute and recurrent vertigo

View this table:

Box 1 Minimum clinical features to be elicited in the presence of acute vertigo

  • No hearing loss or tinnitus; no headache

  • Onset and duration of vertigo; whether vertigo is positional or sustained

  • Nystagmus during Hallpike manoeuvre

  • Eye movements normal (except horizontal nystagmus)

  • No Horner’s syndrome (miosis with partial ptosis)

  • Consider using the head thrust test

Ears and face
  • Tympanic membranes normal

  • No new onset deafness (rub fingers to each ear)

  • No facial weakness, dysphagia, or dysphonia

Body and limbs
  • Must be able to walk (albeit with subjective unsteadiness)

  • No limb paresis, objective ataxia, or sensory loss

In primary care we often identify the patient who has clinically important disease because he or she feels very unwell, or we recognise that within our professional lifetime of experience, this patient “does not fit” with our initial hypothesis of minor disease.19 However, this “intuitive” approach may fail with vertigo. The symptom is often alarming for both the patient and the assessing doctor, yet it is usually due to self limiting disease. Equally, the features that differentiate between major and minor illness may be subtle unless verified by analytical reasoning.

The diagnostic solution: iterative diagnosis

As in many areas of medicine, the list of potential diagnoses in a patient with vertigo is exhaustive (and exhausting).20 The process of iterative diagnosis involves formulating a few likely hypotheses early on in the assessment (benign positional vertigo or acute vestibular neuritis), but then recurrently (iteratively) testing the hypothesis to check that the presumptive diagnosis is safe.

This analytical process involves brief testing for evidence that would refute the presumptive diagnosis (the “red flags”—particularly for a brainstem stroke; box 2) as well as looking for confirmatory evidence. The Hallpike manoeuvre and the “head thrust test” are confirmatory tests that can enable general practitioners to confidently make a positive diagnosis of vestibular as opposed to brainstem disease. The Hallpike test is widely described and used, but the head thrust test (sometimes called the head impulse test; see fig 3) will be new to most non-specialists and is potentially a useful new test for general practitioners.

If applied carefully to the patient with vertigo, the iterative method (both “ruling out” central vertigo and “ruling in” benign positional vertigo or acute vestibular neuritis) can be quick and safe and can reduce unnecessary referral and investigation. Specialist examinations such as magnetic resonance imaging are often of little use in establishing the cause of isolated vertigo.21

Box 2 “Red flags” for urgent referral in a patient with acute vertigo

  • Any central neurological symptoms or signs22

  • New type of headache (especially occipital)23 24

  • Acute deafness25

  • Vertical nystagmus

A diagnostic strategy that has been derived from guidance on imaging patients with acute vertigo in secondary care (derived from the authors’ experience and case series) is shown in figure 1.25


Fig 1 Diagnostic strategy with acute vertigo

Vertigo: the red flags

Other neurological symptoms or signs

Because structures in the brainstem are closely packed together, vertigo in the absence of any other cranial nerve feature (such as diplopia, facial weakness, facial numbness, dysphagia, dysphonia) or long tract symptom (such as weakness or numbness of the limbs) is unlikely to have a central cause. This reasoning is partly confirmed by a large study of patients with vascular brainstem disease.26 Fewer than 1% of 407 patients with posterior circulation strokes in the New England Medical Centre posterior circulation registry presented with a single isolated symptom. Another study of 1666 patients aged over 44 presenting to a US emergency department with “dizziness” found that 53 (3.6%) were due to stroke or transient ischaemic attacks, and among patients with dizziness without other symptoms or signs, only 0.7% had had a stroke or transient ischaemic attack.22 Although case series indicate that 5% of new presentations of multiple sclerosis do have vertigo, a first episode of demyelination is unlikely to present with vertigo only. Acoustic neuroma is also a very rare cause of isolated acute vertigo.13

However, vertigo associated with any other neurological features such as visual blurring, diplopia, facial weakness or numbness, dysphagia, dysphonia, limb weakness or ataxia is likely to be due to brain stem disease such as lateral medullary or cerebellar stroke. These patients should be referred urgently.


Vertigo in association with a headache may be due to migraine. Vestibular migraine (migrainous symptoms associated recurrently with unsteadiness or vertigo) may be commoner than previously recognised in patients with recurrent symptoms. Two large secondary care studies found that 7% of a total of 544 chronically dizzy patients attending dizziness clinics had vestibular migraine.27 28 A population questionnaire study of 8318 participants found that 29.5% reported an episode of vertigo or severe dizziness. Of 243 participants with vertigo who were studied further, 27 women and 6 men (14%) had vestibular migraine.29 Case studies indicate that vertigo does not necessarily follow the normal time course of other migrainous “auras.” 30 The vertigo may occur before, during, or after the characteristic headache and vomiting. The symptoms may last many days, rather than the usual 5-60 minutes of a typical migrainous aura.30

The diagnosis of vestibular migraine needs to be made with caution, particularly in the acute setting, because brainstem vascular events can present in an almost identical way. Prospective studies indicate that between 10% and 34% of all stroke patients experience headaches around the time of the stroke. Headaches are commoner in patients with posterior circulation strokes.23

Vestibular migraine is largely a diagnosis of exclusion. Most patients with new onset headache and clear vertigo will need admission to hospital, unless there is a history of recurrent similar episodes.

New hearing loss

Sudden onset of complete unilateral deafness associated with vertigo suggests acute ischaemia of the labyrinth or brainstem,31 and patients presenting in this way require immediate admission. Other rare causes of unilateral hearing loss with vertigo are an acoustic neuroma or (after straining or trauma) a perilymphatic fistula. All of these diagnostic possibilities should be referred to a specialist. Ménière’s disease may also present with hearing loss and vertigo.

Vertical nystagmus

Vertical nystagmus indicates brainstem or cerebellar disease. The nystagmus of benign positional vertigo is torsional (“rotatory”) and not sustained. The nystagmus of acute vestibular neuritis is unidirectional (fast phase to one side) and horizontal.

Vertigo without red flags: confirmatory tests

Features characterising benign positional vertigo

Right sided benign positional vertigo with torsional (rotatatory) nystagmus that is not sustained

Embedded Image

In benign positional vertigo, the history is of a sudden onset of positionally induced vertigo which attenuates after 10 to 60 seconds and which is reproducible by head positioning. Often it first occurs when turning over in bed, and 20% of patients may have had a minor head injury. The vertigo is intense and nausea is common but vomiting is unusual.32 The Hallpike manoeuvre will reveal no other symptoms and no signs apart from torsional (rotatory) nystagmus that has a latency of a few seconds and disappears within about 30 seconds (fig 2; video).33 34


Fig 2 Examining the right ear for positional vertigo or nystagmus with the Hallpike positional manoeuvre. Top—the traditional way with, the patient’s head hanging off the couch; bottom—in the sideways Hallpike manoeuvre the patient starts from a sideways position on the couch rather than with the feet up along the couch; this allows patients to be examined when there is insufficient room for the patient’s head (if the couch is positioned between two filing cabinets, for example). The final head position is the same in both procedures. Note how the doctor can help the patient open the eyes fully for optimal viewing of a nystagmus

Benign positional vertigo is due to free floating crystals within the lumen of the semicircular canals (usually the posterior canal). It is easily and successfully treated by one of two simple repositioning manoeuvres: the popular Epley manoeuvre or the less well known but simpler Semont manoeuvre. These are illustrated by video clips on

Acute vestibular neuritis

Acute vestibular neuritis is a common diagnosis, defined by sustained (non-positional) vertigo with unidirectional, predominantly horizontal nystagmus (no hearing loss or tinnitus, and no other neurological symptoms or signs).35 The cause is uncertain; nausea and vomiting are common. The patient is usually young or middle aged and presents with sustained vertigo rather than positional vertigo. Some patients have symptoms of a preceding viral respiratory infection. Vertigo may have an abrupt onset (73% of cases) or increase over a few hours (27%),36 or it can occur on waking. The severity usually peaks within a day and resolves gradually over several days.

A problem for the general practitioner is that, rarely, cerebellar strokes can present with isolated, sustained vertigo. In one case series of 240 patients with cerebellar strokes, 25 had sustained vertigo as the sole feature.37 In elderly patients at high risk of vascular disease this may be a noteworthy possibility if the patient has sustained vertigo.

Head thrust to the left is normal - the eyes continue to fixate on the visual target (usually the examiner's nose). Head thrust to the right is abnormal - the eyes make a catchup saccade to re-fixate on the visual target.

In recent years a simple test, the head thrust test, has been described (fig 3; video). This fast movement of the head tests the vestibular ocular reflex (which is also the basis of the doll’s eye test in comatose patients). The test always has abnormal results in patients with acute vestibular neuritis, and it was normal in 24 out of 25 patients with isolated vertigo due to cerebellar stroke.37 The vestibular ocular reflex fails with movement of the face towards the side of the affected labyrinth.


Fig 3 The head thrust test assesses the vestibular ocular reflex.38 Reprinted with permission

No red flags, no confirmation of benign positional vertigo or acute vestibular neuritis

If the vertigo is persistent and the head thrust test has negative results, consider admission to hospital for imaging.25 If the vertigo is not persistent and there is a history of recurrent episodes, consider the possibility of vestibular migraine. Ask specifically about photophobia and headache (even mild headache) and any history suggestive of migraine. Non-urgent referral or a trial of treatment would be reasonable.

If there is a suggestion of recurrent cochlear symptoms (“fullness” in the ear, transitory deafness, or tinnitus), consider the possibility of Ménière’s disease. Ménière’s is a rare cause of vertigo, but in the early stages it can be characterised by either paroxysmal vertigo or episodes of fluctuating aural symptoms without vertigo. In one case series of 243 patients with Ménière’s, vertigo was the only initial feature in 22%.13

Discussion of case scenario

Vertigo is alarming for the patient, and anxiety can be expected. On review of this patient the general practitioner felt less confident that the diagnosis was “viral labyrinthitis” (acute vestibular neuritis) or benign positional vertigo. He had “prematurely anchored” the diagnosis, relied on a confirmatory test (the Hallpike manoeuvre), and ignored the data that did not fit—the hoarseness, difficulty swallowing with minimal signs of pharyngitis, and the fact that the positionally induced nystagmus was sustained. More careful questioning elicited facial numbness and slight clumsiness of the left hand. On admission to hospital the patient had Horner’s syndrome on the left side of the face and sustained nystagmus on positioning. Magnetic resonance imaging of the brain showed a left lateral medullary (brain stem) infarct. In this case clinical review allowed the general practitioner’s cognitive errors to be corrected.

Key points

  • Vertigo needs to be distinguished from the three other categories of dizziness: presyncope, disequilibrium of the elderly, and “lightheadedness”

  • Most isolated vertigo is due to benign positional vertigo or acute vestibular neuritis (sustained vertigo)

  • Isolated vertigo and nystagmus (without other features such as hearing loss, tinnitus, diplopia, dysphagia, facial weakness, or long track symptoms or signs) is unlikely to be due to central causes (vascular, demyelination, or tumour).

  • A positive diagnosis of benign positional vertigo or acute vestibular neuritis can be made with the Hallpike manoeuvre or the head thrust test


Cite this as: BMJ 2009;339:b3493


  • This series aims to set out a diagnostic strategy and illustrate its application with a case. The series advisers are Kevin Barraclough, general practitioner, Painswick, and research fellow in community based medicine, University of Bristol; Paul Glasziou, professor of evidence based medicine, Department of Primary Health Care, University of Oxford; and Peter Rose, university lecturer, Department of Primary Health Care, University of Oxford.

  • We thank Cambridge University Press for permission to use the video clips.

  • Contributors: KB wrote the first draft. AMB reviewed and contributed to rewriting. KB is guarantor.

  • Competing interests: None declared.

  • Provenance and peer review: Commissioned; externally peer reviewed.

  • Patient consent not required (patient anonymised, dead, or hypothetical).