Review

Cervicogenic headache: an assessment of the evidence on

clinical diagnosis, invasive tests, and treatment
Nikolai Bogduk, jayantilal Govind*

Cervicogenic headache is characterised by pain referred to the head from the cervical spine. Although the International
Headache Society recognises this type of headache as a distinct disorder, some clinicians remain sceptical. Laboratory
and clinical studies have shown that pain from upper cervical joints and muscles can be referred to the head. Clinical
diagnostic criteria have not proved valid, but a cervical source of pain can be established by use of fluoroscopically
guided, controlled, diagnostic nerve blocks. In this Review, we outline the basic science and clinical evidence for
cervicogenic headache and indicate how opposing approaches to its definition and diagnosis affect the evidence for its
clinical management. We provide recommendations that enable a pragmatic approach to the diagnosis and
management of probable cervicogenic headache, as well as a rigorous approach to the diagnosis and management of
definite cervical headache.

Introduction
Cervicogenic headache is pain referred to the head from
a source in the cervical spine. Unlike other types of
headache, cervicogenic headache has attracted interest
from disciplines other than neurology, in particular
manual therapists and interventional pain specialists,
who believe that they can find the source of pain among
the joints of the cervical spine. Neurologists differ in
their acceptance of this disorder. The International
Headache Society recognises cervicogenic headache as
a distinct disorder1 and one chapter in a leading
headache textbook acknowledges that injuries to upper
cervical joints can cause headache after whiplash,2
although another chapter indicates that this concept is
not fully accepted.3
In terms of basic sciences, cervicogenic headache is
the best understood of the common headaches. The
mechanisms are known, and this headache has been
induced experimentally in healthy volunteers. In some
patients, cervicogenic headache can be relieved
temporarily by diagnostic blocks of cervical joints or
nerves. However, a matter that remains contentious is
how cervicogenic headache should be diagnosed. Some
neurologists maintain that this headache can be
diagnosed on clinical features; others are not convinced
of the validity of such diagnosis. Manual therapists use
manual examination of vertebral motion segments,
whereas interventional pain specialists use fluoroscopically guided diagnostic blocks.
In this Review, we provide a synopsis of the available
evidence on cervicogenic headache. We summarise the
basic mechanisms, analyse the evidence on diagnosis
and treatment, and provide recommendations on
management.

Mechanism of pain referral

Cervicogenic headache is referred pain from the cervical
spine. Physiologically, this pain is analogous to pain felt
in the shoulders, chest wall, buttocks, or lower limbs
that is referred from spinal sources; hence its familiarity
to pain specialists.
www.thelancet.com/neurology Vol 8 October 2009

Lancet Ncuro/ 2009; 8:959-68

See Reflection and Reaction
page 875
Newcastle Bone and Joint
Institute, Royal Newcastle
Centre, Newcastle,
New South Wales, Australia
(N Bogduk MD); Faculty of
Health Sciences, University of
Newcastle, Callaghan,
New South Wales, Australia
(N Bogduk); Pain Management

The mechanism underlying the pain involves

convergence between cervical and trigeminal affrents
in the trigeminocervical nucleus (figure I).45 In this
nucleus, nociceptive affrents from the Cl, C2, and C3
spinal nerves converge onto second-order neurons that
also receive affrents from adjacent cervical nerves and
from the first division of the trigeminal nerve (V), via
the trigeminal nerve spinal tract. This convergence has
been shown anatomically and physiologically in
laboratory animals.5-9 Convergence between cervical
affrents allows for upper cervical pain to be referred to
regions of the head innervated by cervical nerves
(occipital and auricular regions). Convergence with
trigeminal affrents allows for referral into the parietal,
frontal, and orbital regions.
Such patterns of referral have been elicited in healthy
volunteers by experimental, noxious stimulation of
cervical structures. Early studies targeted the
suboccipital and posterior cervical muscles,1011 and
investigators have shown that noxious stimulation of
more rostral structures in the cervical spine elicited
referred pain in the occipital region and more distant
regions, such as the frontal region and orbit. By contrast,
stimulation of more caudal structures elicited pain in
the neck, which could be referred to the occipital
regions, although not to distant regions of the head
(figure 2). Results from later studies have shown that
noxious stimulation of the atlanto-occipital and lateral
atlanto-axial joints,12 the C2-3 zygapophysial joint,13 and
the C2-3 intervertebral disc1415 can produce pain in the
occipital region (figure 3).
Complementary studies have mapped the distribution
of pain that could be relieved in patients by controlled
diagnostic blocks of the lateral atlanto-axial joint or the
C2-3 or C3-4 zygapophysial joints.16 Patients with pain
from a particular joint do not have exactly the same
distribution of pain, but there are similarities in the
distribution. Pain from the lateral atlanto-axial joint
(Cl-2) tends to be focused on the occipital and
suboccipital regions, and tends to be referred to the
vertex, orbit, and ear (figure 4). Pain from the C2-3

Unit, Canberra Hospital,

Woden, Australian Capital
Territory, Australia
(J Govind MBChB); and School
of Medicine, Australian
National University, Canberra,
Australian Capital Territory,
Australia (J Govind)
*Dr Govind died on June 16,
2009
Correspondence to:
Nikolai Bogduk, Newcastle Bone
and Joint Institute, Royal
Newcastle Centre, PO Box 564J,
Newcastle, New South
Wales 2300, Australia
[email protected]

959

Review

Australia and North America and was based on

establishing a cervical source of pain in patients with
headache by use of controlled diagnostic blocks.

Midbrain

Trigeminal
nerve (V)

Trigeminothalamic
tract
<^J

"

i Spinal tract of trigeminal nerve

Cl spinal nerve

C2 spinal nerve

Trigeminocervical
nucleus
~)

C3 spinal r

Figure 1: Mechanism of pain referral from the cervical spine to the head
Nociceptive affrents of the trigeminal and upperthree cervical spinal nerves
converge onto second-order neurons in the trigeminocervical nucleus in the
upper cervical spinal cord. This convergence mediates the referral of pain signals
from the neck to regions of the head innervated by cervical nerves orthe
trigeminal nerve.

zygapophysial joint also occurs in the occipital region

and spreads across the parietal region to the frontal
region and orbit. Pain from the C3-4 joint can be
referred to the head, but is more commonly focused in
the upper and lateral cervical region (figure 4).
These data show that the structures capable of
producing referred pain to the head are those innervated
by the Cl, C2, and C3 nerves. No experimental studies
have shown that structures innervated by lower cervical
nerves are capable of directly causing headache.

Epidemiology
Estimates of the prevalence of cervicogenic headache
differ according to the populations studied and the
criteria used to make the diagnosis. When clinical
criteria have been used, the prevalence of cervicogenic
headache has been estimated to be 1%, 2 S%,17 or 4-1%18
in the general population and as high as 17 5% among
patients with severe headaches.17 The prevalence is as
high as 53% in patients with headache after whiplash.19

Diagnosis
The diagnosis of cervicogenic headache has been driven
by two schools of practice. The clinical diagnosis
approach arose in Europe and was based on the belief
that cervicogenic headache had distinctive clinical
features by which it could be diagnosed. The approach
of interventional diagnosis by pain medicine arose in
960

Clinical diagnosis
The first set of clinical diagnostic criteria, published in
199020 and revised in 1998,21 defined cervicogenic
headache as a unilateral headache associated with
evidence of cervical involvement through provocation
of pain by movement of the neck or by pressing the
neck; concurrent pain in the neck, shoulder, and arm;
and reduced range of motion of the neck, with or
without other features. Results from subsequent studies
have shown that these clinical features were either not
unique to cervicogenic headache22-24 or were
insufficiently different from features in healthy
individuals,25,26 thereby precluding them from being
valid diagnostic features.52728 Similarly, either there are
no radiographie abnormalities in patients said to have
cervicogenic headache29,30 or the radiographie features
overlap with those seen in healthy individuals.5,31
When tested for agreement between observers, the
proposed clinical features of cervicogenic headache
differed in their reliability.528 The most reliable features
were pain that starts in the neck and radiates to the
fronto-temporal region; pain that radiates to the
ipsilateral shoulder and arm; and provocation of pain
by neck movement.32,33 Agreement was poor about other
features, such as restricted range of motion and
pressure pain on palpation.5,28
Some investigators have proposed a less emphatic
clinical approach to diagnosis. These clinicans reduced
the clinical criteria to a list of seven features (panel I),34
and qualified the certainty of diagnosis. These authors
proposed that possible cervicogenic headache could be
diagnosed if patients had unilateral headache and pain
that starts in the neck. If any three additional criteria
were fulfilled, the diagnosis was advanced to probable
cervicogenic headache. By use of these operational
guidelines, the authors felt that they could confidently
distinguish cervicogenic headache from migraine. The
clinical features most strongly indicative of cervicogenic
headache were pain that radiates to the shoulder and
arm, varying duration or fluctuating continuous pain,
moderate, non-throbbing pain, and history of neck
trauma.
Although some investigators have defended the
clinical diagnostic criteria for cervicogenic headache,
they have examined only their nosological validity (ie,
the extent to which the criteria distinguish cervicogenic
headache from migraine and tension-type headache).
No studies have established that patients who fulfil
these diagnostic criteria actually have a cervical source
for their pain. Eundamental to the concept of
cervicogenic headache is that it constitutes pain referred
to the head from a cervical source. Therefore, proving
such a source is essential for the diagnosis.
www.thelancet.com/neurology Vol 8 October 2009

Review

The revised criteria of the International Headache

Society1 reflect the controversy between clinical
diagnosis and objective testing for cervicogenic
headache (panel 2). Evidence of a cervical source of pain
is required, but the explanatory notes declare that
clinical features that have little reliability or validity are
not acceptable. With out other evidence, controlled
diagnostic blocks become the only means of establishing
the diagnosis.
An idiosyncrasy of the International Headache Society
criteria is criterion D, which describes resolution of
pain 3 months after treatment. This criterion was not
designed for diagnosis before treatment, but to promote
rigour. The criterion requires that, if a cause is found,
relief of pain should ensue if that cause is successfully
treated. A corollary of this criterion is that partial or
short-lasting relief after treatment does not qualify as a
diagnostic criterion.
Interventional diagnosis
Practitioners of interventional pain medicine use
fluoroscopically guided, controlled diagnostic blocks to
test whether particular structures are the source of pain
in patients with suspected cervicogenic headache.
Studies have focused on three structures: the lateral
atlanto-axial joint can be anaesthetised by use of
intr a-articular blocks (figure 5);35"37 the C2-3
zygapophysial joint can be blocked by anaesthetising
the third occipital nerve where it crosses the joint and
supplies it with articular branches;19,38 and the C3-4
zygapophysial joint can be anaesthetised by blocking
the medial branches of the C3 and C4 dorsal rami.38
Complete relief of headache after such blocks, under
controlled conditions, provides objective evidence of a
cervical source of pain.
The best available studies indicate that the C2-3
zygapophysial joints are the most common source of
cervicogenic headache,1619 39,40 accounting for about 70%
of cases.13 Although data are not available for the
prevalence of lateral atlanto-axial joint as a source of
cervicogenic headache, this joint seems to be quite
commonly involved.13,35 The C3-4 zygapophysial joint
has only occasionally been implicated in cervicogenic
headache.16 The C2-3 intervertebral disc can also be a
source of this headache, but its prevalence as a causative
factor is not known.41
In patients whose headaches have been relieved by
controlled diagnostic blocks of upper cervical joints,
there are no distinctive clinical features. The pain is
typically dull and aching in quality. The range of
movement of the head might be restricted, but not in
any characteristic manner or to any characteristic
degree. Tenderness over the C2-3 joint is not a
diagnostic feature as this sign has a positive likelihood
ratio of only 2 1 to I.19
All studies that have implicated the C2-3 zygapophysial
joint as a source of cervicogenic headache have been
www.thelancet.com/neurology Vol 8 October 2009

Interspinous C3-4

Interspinous C4-5

Interspinous C5-6

Figure 2: Referred pain patterns after noxious stimulation of basal occipital periosteum and interspinous
muscles at Cl-2, C2-3, C3-4, C4-5, and C5-6
The morecephaladthesiteof stimulation, the more likely that pain Is referred to distant regions of the head. The
numbers indicate the percentage of individuals who reported pain in the area shown after stimulation at each
segmental level. The arrows indicate the approximate site of stimulation. Adapted from Campbell and Parsons, with
permission from Lippincott Williams &Wilkins.10

Atlanto-occipital joint
O-Cl

C2-3 zygapophysial joint

C2-3 intervertebral disc

\\
^

Lateral atlanto-axial joint

Cl-2

Figure 3: Referred pain patterns after noxious stimulation of upper cervical

joints and the C2-3 intervertebral disc
Based on data from Dreyfuss and colleagues,12 Dwyer and colleagues,13 Schellhas
and colleagues,14 and Grubb and Kelly.15

conducted in patients with a history of trauma.16,19,39,40 No

patients with spontaneous onset of headache have had
a cervical source of pain. This observation reinforces
one feature in the clinical approach to diagnosis, in
which history of neck trauma is an important criterion.
A complement to this observation is that no studies
that have used controlled diagnostic blocks have shown
complete relief of pain in patients with migrainous
961

Review

95-100%

70-94%

45-69%

20-44%

C2-3

C3-4

Figure 4: Areas of pain relief in patients who underwent controlled blocks of the synovial joints at Cl-2, C2-3, and C3-4
The density of shading is proportional to the number of patients who perceived pain in the particular area indicated. Adapted from Cooper and colleagues, with
permission from Blackwell Science.16

Panel 1: Clinical criteria for the diagnosis of cervicogenic

headache
1

Unilateral headache without side-shift

Symptoms and signs of neck involvement: pain triggered

by neck movement or sustained awkward posture and/or

the blocks were not controlled. In a later study, which

used a large sample and had controlled diagnostic
blocks, manual examination had a high sensitivity but
no specificity and, therefore, had no diagnostic validity.43
Further evidence is required before a scientific basis for
manual diagnosis can be confirmed.

external pressure of the posterior neck or occipital region;

ipsilateral neck, shoulder, and arm pain; reduced range of
motion
3

Pain episodes of varying duration or fluctuating

continuous pain

Moderate, non-excruciating pain, usually of a

non-throbbing nature

Pain starting in the neck, spreading to

oculo-fronto-temporal areas

Anaesthetic blockades abolish the pain transiently

provided complete anaesthesia is obtained, or occurrence
of sustained neck trauma shortly before onset

Various attack-related events: autonomie symptoms and

signs, nausea, vomiting, ipsilateral oedema and flushing
in the peri-ocular area, dizziness, photophobia,
phonophobia, or blurred vision in the ipsilateral eye

Satisfying criteria 1 and 5 qualifies for a diagnosis of possible cervicogenic headache.

Satisfying any additional three criteria advances the diagnosis to probable cervicogenic
headache. Adapted from Antonaci and colleagues,34 with permission from
Wiley-Blackwell.

features, such as photophobia and vomiting. In our

experience, such patients do not respond to diagnostic
blocks of cervical joints.

Diagnosis through greater occipital nerve blocks

Some investigators use greater occipital nerve blocks as

a diagnostic test either for greater occipital neuralgia or
for cervicogenic headache.44 However, there is no clear
rationale for this practice. Diagnostic blocks relieve
pain from a source innervated by a nerve distal to where
it is blocked. Greater occipital nerve blocks are executed
near to where the nerve crosses the superior nuchal
line. Distal to this point, the nerve supplies only the
skin of the occipital region. There are no known
disorders of the scalp that might cause persistent pain.
Therefore, greater occipital nerve blocks cannot be used
to diagnose cervicogenic headache as these blocks do
not establish a cervical source of pain. Furthermore, no
studies have shown that controlled blocks of the greater
occipital nerve consistently produce complete relief of
headache. At best, greater occipital nerve blocks have a
type of partial, neuromodulatory effect on headache
mechanisms, whether the headaches have a cervical
source or not. Greater occipital nerve blocks relieve
pain, temporarily, in substantial proportions of patients
with migraine, cluster headache, and hemicrania
continua.454* A positive, greater occipital nerve block,
therefore, cannot be a specific test for cervicogenic
headache.47

Manual diagnosis

Manual therapists contend that they can diagnose

cervical sources of headache by manual examination of
upper cervical joints. However, this practice has not
been validated. An early study compared the diagnosis
made by a manual therapist with that made with
diagnostic blocks,42 but the sample size was small and
962

Differential diagnosis

Several disorders share certain features of cervicogenic

headache, such as pain in the neck and head. These
disorders can be difficult to distinguish, unless (or
until) additional features emerge. Other disorders form
the differential diagnosis of cervicogenic headache only
www.thelancet.com/neurology Vol 8 October 2009

Review

notionally because they affect cervical structures and

might cause headache, but these disorders have
distinctive features.
The most crucial differential diagnosis of cervicogenic
headache is dissecting aneurysms of the vertebral or
internal carotid arteries, which can present with neck
pain and headache.48-50 These aneurysms are indicated
by the onset of cerebrovascular features, which typically
emerge within 1-3 weeks. If this differential diagnosis
is not considered, there is a risk of patients being
treated with cervical manipulation, with fatal
consequences due to aggravation of the aneurysm.
The second most important differential diagnosis is
lesions of the posterior cranial fossa, as the dura mater
and vessels of the posterior fossa are innervated by
upper cervical nerves. These lesions are distinguished
by the onset of neurological features or systemic illness.
Meningitis of the upper cervical spine can be
distinguished from cervicogenic headache by the
presence of systemic illness and neck rigidity.
Additionally, herpes zoster can produce pain in the
occipital region during its prodromal phase; however,
the eruption of vesicles distinguishes this disease from
cervicogenic headache.
Because the C2 spinal nerve runs behind the lateral
atlanto-axial joint and is accompanied by its durai sleeve
and a substantial plexus of veins, two distinctive
disorders can be confused with cervicogenic headache.51
First, neck-tongue syndrome occurs when rapid
turning of the head subluxates the lateral atlanto-axial
joint posteriorly. Tension in the joint capsule causes
ipsilateral occipital pain, while compression of the
C2 spinal nerve produces numbness of the tongue.5253
This syndrome is distinguished by its precipitating
factor and accompanying features. Second, C2 neuralgia
can be caused by various disorders. Inflammatory
disorders or injuries of the lateral atlanto-axial joint can
result in the adjacent nerve becoming incorporated in
the fibrotic changes of chronic inflammation.54,55 The
C2 spinal nerve can be compromised by a meningioma,56
neurinoma,57 anomalous vertebral arteries,58 and several
other vascular anomalies.54'58,59 Nerves affected by
vascular abnormalities have several features indicative
of neuropathy, such as myelin breakdown, chronic
haemorrhage, axon degeneration and regeneration, and
increased endoneurial and pericapsular connective
tissue.58 Unlike the dull, aching pain of cervicogenic
headache, the features of C2 neuralgia are intermittent,
lancinating pain in the occipital region associated with
lacrimation and ciliary injection.54,5859
Greater occipital neuralgia is an outdated diagnosis,
used before the concept of somatic referred pain was
widely understood, when physicians believed that any
pain in a particular region was due to some affliction of
the nerve that ran through that region. Accordingly,
pain in the occipital region was attributed to greater
occipital neuralgia. However, no pathology of the
www.thelancet.com/neurology Vol 8 October 2009

Panel 2: Diagnostic criteria for cervicogenic headache, as proposed by the

International Headache Society1
Diagnostic criteria
A Pain referred from a source in the neck and felt in one or more regions of the head
and/or face, fulfilling criteria C and D
B Clinical, laboratory, and/or imaging evidence of a disorder or lesion within the cervical
spine or soft tissues of the neck known to be, or generally accepted as, a valid cause of
headache1
C Evidence that the pain can be attributed to the neck disorder or lesion based on at
least one of the following: 1) evidence of clinical signsthat implicate a source of pain
in the neck;2 or 2) abolition of headache after diagnostic blockade of a cervical
structure or its nerve supply with placebo or other adequate controls3
D Pain resolves within 3 months after successful treatment of the causative disorder
or lesion
Notes
1 Tumours, fractures, infections, and rheumatoid arthritis of the upper cervical spine
have not been validated formally as causes of headache, but are nevertheless
accepted as valid causes when proven to be so in individual cases. Cervical
spondylosis and osteochondritis are not accepted as valid causes fulfilling criterion B.
When there are myofascial tender spots, the headache should be coded under 2
(tension-type headache)
2 Clinical signs acceptable for criterion Cl must have shown reliability and validity. The
future task is the identification of such reliable and valid operational tests. Clinical
features such as neck pain, focal neck tenderness, history of neck trauma, mechanical
exacerbation of pain, unilaterality, coexisting shoulder pain, reduced range of motion
in the neck, nuchal onset, nausea, vomiting, and photophobia are not unique to
cervicogenic headache. These can be features of cervicogenic headache, but they do
not define an association between the disorder and the source of the headache
3 Abolition of headache means complete relief of headache, indicated by a score of
zero on a visual analogue scale. Nevertheless, a >90% decrease in pain to a level of
<5 on a 100-point visual analogue scale is acceptable to fulfil criterion C2

greater occipital nerve has ever been proven to explain

occipital pain. The proposition that the greater occipital
nerve could be compressed between the posterior arch
of the atlas and the lamina of the axis is incompatible
with the anatomy and biomechanics of those vertebrae60
and was retracted by one of the authors who originally
proposed it.61 Entrapment of the greater occipital nerve
where it leaves the posterior neck muscles has not been
distinguished from the normal anatomy of the nerve
penetrating a fibrous sling.5,62 Liberation of the nerve
has also not proved to be an effective, lasting
treatment.5,63 Lancinating pain in the occipital region is
more likely to be C2 neuralgia, in which case the
pathology affects the C2 spinal nerve rather than the
greater occipital nerve. Deep aching pain is incompatible
with a neuralgia: this pain is more likely to be somatic
referred pain from an upper cervical joint.1,5
Another outdated diagnosis is cervical migraine (also
known as Barr-Lieou syndrome).6465 In this disorder,
headache was purportedly due to irritation of the
vertebral nerve, which causes spasm of the vertebral
artery. In physiological studies, the vertebral artery is
remarkably inert to electrical stimulation of the vertebral
963

Review

AOJ

LAAJ

C2-3 ZJ

C3-4ZJ'

Lateral view

Figure 5: Posterior and lateral views of the upper cervical spine, showing the leading articular sources of
cervicogenic headache, the related nerves, and where needles are placed for diagnostic blocks of these
structures
Red labels and needles point to target sites for diagnostic blocks. AOJ=atlanto-occipital joint. C3 DMB=C3 deep
medial branch block. C4mb=medial branch of the C4 dorsal ramus. dmb=deep medial branch of the C3 dorsal
ramus. LAA IAB=intra-articular block of the lateral atlanto-axial joint. LAAJ=lateral atlanto-axial joint. ton=third
occipital nerve.TONB=third occipital nerve block. ZJ=zygapophysial joint.

nerve66 and even inert to intra-arterial infusions of

vasoactive drugs.67 These properties rule out the
purported mechanism of migraine cervicale.5
Lower cervical disorders
Some investigators have studied the proposition that
lower cervical disorders can cause headache. These
researchers refer to circumstantial evidence that some
patients with lower cervical radiculopathy also have
headache, of whom various proportions are relieved of
headache when the radiculopathy is treated surgically.68,69
Although these data imply an association, they do not
indicate a direct association between headache and
lower cervical disorders. Neuroanatomically, there is no
direct link between lower cervical affrents and the
trigeminocervical nucleus. Intermediate mechanisms,
such as muscle tension and secondary kinematic
abnormalities that affect upper cervical joints70 might
be involved.

Treatment
Although there have been many treatments suggested
for cervicogenic headache, few have been tested and
even fewer have been proven successful. Among the
determinants of effectiveness are whether the headache
was diagnosed clinically or whether a cervical source
was proven.
Clinical diagnosis
No drugs are effective for cervicogenic headache.
Transcutaneous electrical nerve stimulation has been
investigated, but not in a controlled study. About 80%
of patients obtained at least a 60% decrease in their
headache index with this technique, but only at
1 month after treatment.71
964

For manual therapy, most publications are case

reports or case series.72 The few randomised, controlled
studies provided follow-up of only 1 or 3 weeks,73-75 and
gave conflicting results.72 The largest and most recent
study showed that treatment with manual therapy,
specific exercises, or manual therapy plus exercises was
significantly more effective at reducing headache
frequency and intensity than was no specific care by a
general practitioner.76 Manual therapy alone, however,
was not more effective than exercises alone, and
combining the two interventions did not achieve better
outcomes than either intervention alone. About 76% of
patients achieved a more than 50% decrease in headache
frequency and 35% achieved complete relief at the
7-week follow-up. At 12 months, 72% had a more than
50% decrease in headache frequency, but the proportion
that had complete reliefwas not reported. Corresponding
figures for decrease in pain intensity were not
reported.
In patients with a clinical diagnosis of cervicogenic
headache, some investigators have targeted the greater
occipital nerve for treatment. In one series, 169 of
180 patients (94%) obtained relief after an injection of
160 mg of depot methylprednisolone and 34 mL of
1% lidocaine, but only for a mean duration of 23 5 days
(range 10-77 days).77 In an unrelated study of 50 patients,
surgical "liberation" of the nerve initially relieved
headache in about 80% of cases, but for a median
duration of only about 3-6 months.63 Excision of the
greater occipital nerve provided relief in about 70% of
patients, but for a median duration of only 244 days.78
Specific diagnosis
In one study, patients were selected for surgery if they
fulfilled the clinical criteria for cervicogenic headache
and obtained relief of headache from diagnostic
blockade of the C2 spinal nerve.79 Patients underwent
decompression and microsurgical neurolysis of the
C2 spinal nerve, with excision of scar, and ligamentous
and vascular elements that compressed the nerve. 14 of
31 patients were rendered pain-free at a mean follow-up
of 16 months. Details on the remaining patients are
incomplete, but 51% gained what was called "adequate"
relief, and 11% suffered a recurrence.
For patients with pain stemming from the lateral
atlanto-axial joints, two options are available. In an
observational study, 26 of 32 patients obtained
immediate relief after intra-articular injection of
steroids.80 About one in five patients obtained greater
than 50% relief from their headache for 3 months, and
one in eight obtained complete relief lasting 9 months.
However, such outcomes have not been confirmed in a
controlled trial and, therefore, cannot yet be attributed
to the injection of steroids. It is possible that a placebo
effect might have influenced results. The other option
is arthrodesis of the joint. The surgical reports attest to
success with this procedure, albeit in small numbers of
www.thelancet.com/neurology Vol 8 October 2009

Review

patients, with complete relief of pain lasting over

2 years.81"83
For patients with pain stemming from the C2-3
zygapophysial joint, investigators from one study
reported that some patients could obtain relief from
intra-articular injection of steroids.84 At 19 months after
such injections, two of 18 patients (11%) were free of
pain. A further 50% had a reduced frequency of
headaches. This study, however, was not controlled and
placebo effects were not excluded. Nevertheless, without
other alternatives, intra-articular injection of steroids
would seem to be a safe and expedient intervention that
could benefit some patients.
Radiofrequency neurotomy
The most extensively studied treatment for cervicogenic
headache is percutaneous radiofrequency neurotomy.
The rationale for this procedure is that if headache can
be relieved temporarily by controlled diagnostic blocks
of the nerve (or nerves) that innervate a particular
cervical joint, then interrupting the pain signal along
that nerve, by coagulating it, should provide long-lasting
relief. This neurosurgical procedure is particularly
applicable for the treatment of headache stemming
from the C2-3 zygapophysial joint, in which case the
target nerve is the third occipital nerve, which innervates
that joint.
Three studies have reported that radiofrequency
neurotomy is not effective.85-87 In these studies, patients
were selected on the basis of clinical criteria, and
neurotomy was done at all levels from C3 to C6.
Diagnostic blocks were done in one study,86 but the
results were not used as an indication for treatment. In
the first study, only one of 15 patients achieved complete
relief of pain;85 in the second study (n=12), outcomes
were no different in patients who received active lesions
from those who received sham lesions;86 and in the
third study (n=30), outcomes from neurotomy were no
different from those of an injection of local anaesthetic
into the greater occipital nerve.87
However, there are three potential difficulties with
these studies of radiofrequency neurotomy. First, at no
stage was the source of pain established. Second, the
technique used for neurotomy has never been validated.
Third, neurotomy was done at segmental levels (C3-6)
that have rarely, if ever, been implicated as a source of
headache. Nerves not proven to mediate the patient's
pain were disrupted by use of techniques not proven to
denervate them.
Opposite results were reported in two studies in
which the diagnosis was carefully established with
controlled diagnostic blocks and meticulous surgical
techniques were used.88,89 For patients in whom effects
of diagnostic blocks indicate that the C2-3 zygapophysial
joint is the source of pain, that joint can be denervated
percutaneously by radiofrequency neurotomy of the
third occipital nerve. The procedure involves placing an
www.thelancet.com/neurology Vol 8 October 2009

electrode parallel and close to the nerve where it crosses

the joint, and using the electrode to disrupt the nerve.90
Under these conditions, complete relief of pain was
achieved in 88% of patients,88,91 with a median duration
of relief of 297 days.88 For patients in whom headaches
recur, relief can be reinstated by repeating the
neurotomy. By undertaking repetition as required,
some patients have been able to maintain relief of their
headache for longer than 2 years.88 The results of a
randomised, placebo-controlled trial indicate that
responses to radiofrequency neurotomy are not due to
placebo effects (p=0-03).91 The successful treatment of
third occipital headache in this study cannot, therefore,
be dismissed as a placebo effect.

A pragmatic clinical approach

The degree to which practitioners might manage
cervicogenic headache depends on the facilities available
to them. If clinicans can undertake fiuoroscopically
guided diagnostic blocks, they can establish a cervical
source of pain and thereby fulfil the diagnostic criteria
for cervicogenic headache as set by the International
Headache Society.1 If physicians are restricted to clinical
diagnosis only, they cannot fulfil those criteria.
Nevertheless, a working diagnosis of possible or
probable cervicogenic headache can be established
based on the criteria listed in panel I.34
For probable cervicogenic headache, exercises with or
without manual therapy seems to be the best option
among conservative therapies.76 All other treatment
strategies are entirely speculative.
If diagnostic blocks or discography can be applied, a
source of pain might be established in the lateral
atlanto-axial joint, the C2-3 intervertebral disc, or the
C2-3 zygapophysial joint. For pain stemming from the
lateral atlanto-axial joint, arthrodesis is the only
treatment for which there is any evidence of
effectiveness.81"83 Intra-articular injections of steroids
are more conservative but their efficacy has not yet been
shown. For pain stemming from the C2-3 intervertebral
disc, anterior cervical fusion can be effective.41 No
alternative is known for discogenic pain. For pain
stemming from the C2-3 zygapophysial joint,
intra-articular injection of steroids is a low-risk
treatment from which some patients can benefit.84
However, a placebo effect has not been excluded. The
only definitive treatment for headache stemming from
the C2-3 zygapophysial joint is radiofrequency
neurotomy.8889 However, this procedure has to be done
with meticulous accuracy and is indicated only if
patients obtain complete relief of headache after
controlled blocks of the third occipital nerve.8889

Conclusions
Neurologists are accustomed to diagnosing headache
on the basis of clinical features, supplemented in some
cases by medical imaging or other tests. Cervicogenic
965

Review

Search strategy and selection criteria

References for this Review were identified from the personal
libraries of the authors, supplemented by the reference lists of
recent reviews and book chapters and by a search of PubMed
with the search terms "cervicogenic headache", "cervical
headache", "headache", "cervical vertebrae", and "neck pain",
between 1950 and July, 2009- Papers published in English,
French, German, Italian, Spanish, and Slavic languages were
reviewed. Seminal articles that introduced new concepts or
provided original data were included. Articles that only
reiterated established concepts or data were excluded. Review
articles have been cited to cover older and extensive literature
on neuroanatomy and neurophysiologyorwhen reviews
cover several articles on particular interventions.

4
5

9
10

headache does not lend itself to this approach. As a

result, tensions and controversies have arisen within
the field. On the one hand, some experts have insisted
that cervicogenic headache can be defined by clinical
criteria, butthe evidence shows otherwise. Consequently,
owing to the absence of valid clinical criteria, some
neurologists have doubted the diagnosis or consider it
to be too often carelessly applied. On the other hand,
pain specialists have developed invasive techniques
whereby the diagnosis can be established objectively.
But fluoroscopically guided, controlled diagnostic
blocks are not among the conventional armamentarium
of neurologists, nor are they widely available.
One of two developments is required for cervicogenic
headache to become more commonly accepted. Pain
specialists, or interventional radiologists, could
collaborate with neurologists to add further studies to
the published works showing the effects of diagnostic
blocks in patients with suspected cervicogenic headache
and the effects of target-specific interventions. Such
collaboration would be in line with the call to reduce
iatrogenic discomfort resulting from incorrect diagnosis
and inappropriate treatment of headaches in general.92
Alternatively, physicians intent on clinical diagnosis
need to develop criteria that are not only reliable but
also valid for a cervical source of pain. Doing so would
allow elimination of the footnotes in the criteria of the
International Headache Society, which currently
prohibit clinical diagnosis.1

11

12
13
14

15
16
17
18
19

20
21
22

23

24

25

Contributors

NB drafted the Review. JG reviewed, improved, and finalised the paper,

and designed the figures.

26

Conflicts of interest

We have no conflicts of interest.

27

References

966

International Headache Society. The International Classification

of Headache Disorders, 2nd edn. Cephalalgia 2004;
24 (suppl 1): 115-16.
Kasch H, Ramadan NM. Headache attributed to whiplash injury.
In: Olesen J, Goadsby PJ, Ramadan NM, Tfelt-Hansen P,
Welch KMA, eds. The Headaches, 3rd edn. Philadelphia:
Lippincott Williams & Wilkins, 2006: 879-83.

28

29

Gobel H, Edmeads JG. Disorder of the skull and cervical spine.

In: Olesen J, Goadsby PJ, Ramadan NM, Tfelt-Hansen P,
Welch KMA, eds. The Headaches, 3rd edn. Philadelphia:
Lippincott Williams & Wilkins, 2006: 1003-11.
Bogduk N. The neck and headaches. Neurol Clin N Am 2004;
22: 151-71.
Bogduk N, Bartsch T. Cervicogenic headache. In: Silberstein SD,
Lipton RB, Dodick DW, eds. Wolff's Headache, 8th edn.
New York: Oxford University Press, 2008: 551-70.
Bogduk N. Cervicogenic headache: anatomic basis and
pathophysiologic mechanisms. Curr Pain Headache Rep 2001;
5: 382-86.
Bartsch T, Goadsby PJ. Stimulation of the greater occipital nerve
induces increased central excitability of durai afferent input.
Brain 2002; 125: 1496-1509.
Bartsch T, Goadsby PJ. Increased responses in trigeminocervical
nociceptive neurons to cervical input after stimulation of the dura
mater. Brain 2003; 126: 1801-13.
Goadsby PJ, Ratsch T. On the functional neuroanatomy of neck
pain. Cephalalgia 2008; 28 (suppl 1): 1-7
Campbell DC, Parsons CM. Referred head pain and its
concomitants./ Nerv Ment Dis 1944; 99: 54451.
Feinstein B, Langton JBK, Jameson RM, Schiller F. Experiments
on referred pain from deep somatic tissues./ Bone Joint Surg
1954; 36A: 981-97
Dreyfuss P, Michaelsen M, Fletcher D. Atlanto-occipital and
lateral atlanto-axial joint pain patterns. Spine 1994; 19: 1125-31.
Dwyer A, Aprill C, Bogduk N. Cervical zygapophysial joint pain
patterns I: a study in normal volunteers. Spine 1990; 15: 453-57
Schellhas KP, Smith MD, Gundry CR, Pollei SR. Cervical
discogenic pain: prospective correlation of magnetic resonance
imaging and discography in asymptomatic subjects and pain
sufferers. Spine 1996; 21: 300-12.
Grubb SA, Kelly CK. Cervical discography: clinical implications
from 12 years of experience. Spine 2000; 25: 1382-89.
Cooper G, Bailey B, Bogduk N. Cervical zygapophysial joint pain
maps. Pain Medicine 2007; 8: 34453.
Evers S. Comparison of cervicogenic headache with migraine.
Cephalalgia 2008; 28 (suppl 1): 16-17.
Sjaastad O. Cervicogenic headache: comparison with migraine
without aura; Vg study. Cephalalgia 2008; 28 (suppl 1): 18-20.
Lord S, Barnsley L, Wallis B, Bogduk N. Third occipital
headache: a prevalence study. / Neurol Neurosurg Psychiatr 1994;
57:1187-90.
Sjaastad O, Fredriksen TA, PfafFenrath V. Cervicogenic headache:
diagnostic criteria. Headache 1990; 30: 725-26.
Sjaastad O, Fredriksen TA, PfafFenrath V. Cervicogenic headache:
diagnostic criteria. Headache 1998; 38: 442-45.
Leone M, DAmico D, Frediani F, Torri W, Sjaastad O, Bussone G.
Clinical considerations on side-locked unilaterality in long-lasting
primary headaches. Headache 1993; 33: 381-84.
DAmico D, Leone M, Bussone G. Side-locked unilaterality and pain
localization in long-lasting headaches: migraine, tension-type
headache, and cervicogenic headache. Headache 1994; 34: 526-30.
Leone M, DAmico D, Moschiano F, Frainotti M, Filippini G,
Bussone G. Possible identification of cervicogenic headache
amongst patients with migraine: an analysis of 374 headaches.
Headache 1995; 35: 461-64.
Bovim G. Cervicogenic headache, migraine, and tension-type
headache. Pressure-pain threshold measurements. Pain 1992;
51: 169-73.
Jull G, Barrett C, Magee R, Ho P. Further clinical clarification of
the muscle dysfunction in cervical headache. Cephalalgia 1999;
19:179-85.
Leone M, DAmico D, Grazzi L, Attanasio A, Bussone G.
Cervicogenic headache: a critical review of the current diagnostic
criteria. Pain 1998; 78: 1-5.
Bogduk N. Distinguishing primary headache disorders from
cervicogenic headache: clinical and therapeutic implications.
Headache Currents 2005; 2: 27-36.
Fredriksen TA, Fougner R, Tengerund A, Sjaastad O. Cervicogenic
headache. Radiological investigations concerning head/neck.
Cephalalgia 1989; 9: 139-46.

www.thelancet.com/neurology Vol 8 October 2009

Review

30

31
32

33

34

35
36

37

38

39
40
41

42

43

44

45

46
47

48

49

50

51

52
53
54

55

Pfaffenrath V, Dandekar R, Pollman W. Cervicogenic headache

the clinical picture, radiological findings and hypotheses on its
pathophysiology. Headache 1987; 27: 495-99.
Zwart JA. Neck mobility in different headache disorders.
Headache 1997; 37: 6-11.
van Suijlekom JA, de Vet HCW, van den Berg SGM, Weber WEJ.
Interobserver reliability of diagnostic criteria for cervicogenic
headache. Cephalalgia 1999; 19: 817-23.
van Suijlekom HA, de Vet HCW, van den Berg SGM,
Weber WEJ. Interobserver reliability in physical examination
of the cervical spine in patients with headache. Headache 2000;
40: 581-86.
Antonaci F, Ghirmai S, Bono S, Sandrini G, Nappi G.
Cervicogenic headache: evaluation of the original diagnostic
criteria. Cephalalgia 2001; 21: 573-83.
McCormick CC. Arthrography of the atlanto-axial (C1-C2) joints:
technique and results./ Intervent Radiol 1987; 2: 9-13.
Busch E, Wilson PR. Atlanto-occipital and atlanto-axial injections
in the treatment of headache and neck pain. Reg Anesth 1989;
14 (suppl 2): 45.
Aprill C, Axinn MJ, Bogduk N. Occipital headaches stemming
from the lateral atlanto-axial (Cl-2) joint. Cephalalgia 2002;
22:15-22.
International Spine Intervention Society. Cervical medial branch
blocks. In: Bogduk N, ed. Practice Guidelines for Spinal
Diagnostic and Treatment Procedures. San Francisco:
International Spinal Intervention Society, 2004: 112-37
Bogduk N, Marsland A. On the concept of third occipital
headache./ Neurol Neurosurg Psychiatr 1986; 49: 775-80.
Bogduk N, Marsland A. The cervical zygapophysial joints as a
source of neck pain. Spine 1988; 13: 610-17.
Schofferman J, Garges K, Goldthwaite N, Kosetler M, Libby E.
Upper cervical anterior diskectomy and fusion improves
discogenic cervical headaches. Spine 2002; 27: 2240-44.
Jull G, Bogduk N, Marsland A. The accuracy of manual diagnosis
for cervical zygapophysial joint pain syndromes. MedJ Aust 1988;
148: 233-36.
King W, Lau P, Lees R, Bogduk N. The validity of manual
examination in assessing patients with neck pain. Spine J 2007;
7: 22-26.
Bovim G, Berg R, Dale LG. Cervicogenic headache: anaesthetic
blockades of cervical nerves (C2-C5) and facet joint (C2/C3).
Pain 1992; 49: 315-20.
Afridi SK, Shields KG, Bhola R, Goadsby PJ. Greater occipital
nerve injection in primary headache syndromesprolonged
effects from a single injection. Pain 2006; 122: 126-29.
Goadsby PJ, Lipton RB, Ferrari MD. Migrainecurrent
understanding and treatment. N EnglJ Med 2002; 346: 257-70.
Leone M, Cecchini AP, Mea E, Tullo V, Bussone G. Epidemiology
of fixed unilateral headaches. Cephalalgia 2008;
28 (suppl 1): 8-11.
de Sousa JE, Halfon MJ, Bonardo P, Reisin RC,
Fernndez Pardal MM. Different pain patterns in patients with
vertebral artery dissections. Neurology 2005; 64: 925-26.
Saeed AB, Shuaib A, Al Sulaiti G, Emery D. Vertebral artery
dissection: warning symptoms, clinical features and prognosis in
26 patients. Can J Neurol Sei 2000; 27: 292-96.
Campos CR, Calderaro M, Scaff M, Conforto AB. Primary
headaches and painful spontaneous cervical artery dissection.
/ Headache Pain 2007; 8: 180-84.
Bogduk N. Local anaesthetic blocks of the second cervical
ganglion: a technique with application in occipital headache.
Cephalalgia 1981; 1: 41-50.
Lance JW, Anthony M. Neck tongue syndrome on sudden turning
of the head./ Neurol Neurosurg Psychiatr 1980; 43: 97-101.
Bogduk N. An anatomical basis for neck tongue syndrome.
/ Neurol Neurosurg Psychiatr 1981; 44: 202-08.
Jansen J, Markakis E, Rama B, Hildebrandt J. Hemicranial attacks
or permanent hemicraniaa sequel of upper cervical root
compression. Cephalalgia 1989; 9: 123-30.
Poletti CE, Sweet WH. Entrapment of the C2 root and ganglion by
the atlanto-epistrophic ligament: clinical syndrome and surgical
anatomy. Neurosurgery 1990; 27: 288-91.

www.thelancet.com/neurology Vol 8 October 2009

56
57

58

59

60
61
62
63

64

65

66

67

68
69

70
71

72
73

74

75
76

77

78
79

80

81

Kuritzky, A. Cluster headache-like pain caused by an upper

cervical meningioma. Cephalalgia 1984; 4: 185-86.
Sharma RR, Parekh HC, Prabhu S, Gurusinghe NT, Bertolis G.
Compression of the C-2 root by a rare anomalous ectatic vertebral
artery. / Neurosurg 1993; 78: 669-72.
Jansen J, Bardosi A, Hildebrandt J, Lucke A. Cervicogenic,
hemicranial attacks associated with vascular irritation or
compression of the cervical nerve root C2. Clinical manifestations
and morphological findings. Pain 1989; 39: 203-12.
Hildebrandt J, Jansen J. Vascular compression of the C2 and C3
rootsyet another cause of chronic intermittent hemicrania?
Cephalalgia 1984; 4: 167-70.
Hunter CR, Mayfield FH. Role of the upper cervical roots in the
production of pain in the head. Am/ Surg 1949; 78: 743i9.
Mayfield FH. Symposium on cervical trauma. Neurosurgical
aspects. Clin Neurosurg 1955; 2: 83-90.
Bogduk N. The anatomy of occipital neuralgia. Clin Exp Neurol
1980; 17: 167-84.
Bovim G, Fredriksen TA, Stolt-Nielsen A, Sjaastad O. Neurolysis
of the greater occipital nerve in cervicogenic headache. A follow
up study. Headache 1992; 32: 175-79.
Barre N. Sur un syndrome sympathique cervicale postrieure et
sa cause frquente: l'arthrite cervicale. Revue du Neurologie 1926;
33: 1246-48.
Tamura T. Cranial symptoms after cervical injury. Aetiology and
treatment of the Barre-Lieou syndrome. / Bone Joint Surg 1989;
71B: 283-87.
Bogduk N, Lambert G, Duckworth JW. The anatomy and
physiology of the vertebral nerve in relation to cervical migraine.
Cephalalgia 1981; 1: 1-14.
Lambert GA, Duckworth JW, Bogduk N, Lance JW. Low
pharmacological responsiveness of the vertebro-basilar
circulation in macaca nemestrina monkeys. EurJ Pharmacol 1984;
102: 451-58.
Jansen J. Surgical treatment of cervicogenic headache. Cephalalgia
2008; 28 (suppl 1): 41-44.
Diener HC, Kaminski M, Stappert G, Stolke D, Schoch B. Lower
cervical disc prolapse may cause cervicogenic headache:
prospective study in patients undergoing surgery. Cephalalgia
2007; 27:1050-54.
Amevo B, Aprill C, Bogduk N. Abnormal instantaneous axes of
rotation in patients with neck pain. Spine 1992; 17: 748-56.
Farina S, Granella F, Malferrari G, Manzoni C. Headache and
cervical spine disorders: classification and treatment with
transcutaneous electrical nerve stimulation. Headache 1986;
26: 431-33.
Haldeman S, Dagenais S. Cervicogenic headaches: a critical
review. Spine J 2001; 1: 31^-6.
Nillson N. A randomized controlled trial of the effect of spinal
manipulation in the treatment of cervicogenic headache.
/ Manipul Physiol Ther 1995; 18: 435-40.
Nillson N, Christensen HW, Hartvigsen J. The effect of spinal
manipulation in the treatment of cervicogenic headache.
/ Manipul Physiol Ther 1997; 2: 326-30.
Vernon HT. Spinal manipulation and headaches of cervical origin.
/ Manipul Physiol Ther 1989; 12: 455-68.
Jull G, Trott P, Potter H, et al. A randomized controlled trial of
exercise and manipulative therapy for cervicogenic headache.
Spine 2002; 27: 1835-43.
Anthony M. Cervicogenic headache: prevalence and response to
local steroid therapy. Clin Exp Rheumatol 2000;
18 (suppl 19): S59-64.
Anthony M. Headache and the greater occipital nerve.
Clin Neurol Neurosurg 1992; 94: 297-301.
Pikus HJ, Phillips JM. Characteristics of patients successfully
treated for cervicogenic headache by surgical decompression of
the second cervical root. Headache 1995; 35: 621-29.
Narouze SN, Casanova J, Mekhail N. The longitudinal
effectiveness of lateral atlantoaxial intra-articular steroid injection
in the treatment of cervicogenic headache. Pain Med 2007;
8: 184-88.
Joseph B, Kumar B. Gallie's fusion for atlantoaxial arthrosis with
occipital neuralgia. Spine 1994; 19: 45455.

967

Review

82

83
84

85

86

87

968

Ghanayem AJ, Leventhal M, Bohlman HH. Osteoarthrosis of the

atlanto-axial jointslong-term follow-up after treatment with
arthrodesis. ) Bone Joint Surg 1996; 78A: 1300-07.
Schaeren S, Jeanneret B. Atlantoaxial osteoarthritis: case series
and review of the literature. Eur Spine J 2005; 14: 501-06.
Slipman CW, Lipetz JS, Plastara CT, Jackson HW, Yang ST,
Meyer AM. Therapeutic zygapophyseal joint injections for
headache emanating from the C2-3 joint. Am J Phys Med Rehabil
2001; 80: 182-88.
van Suijlekom HA, van Kleef M, Barendse GAM, Sluijter ME,
Sjaastad O, Weber WEJ. Radiofrequency cervical zygapophyseal
joint neurotomy for cervicogenic headaches: a prospective study
of 15 patients. Fund Neurol 1998; 13: 297-303.
Stovner LJ, Kolstad F, Heide G. Radiofrequency denervation of
facet joints C2-C6 in cervicogenic headache: a randomised,
double-blind, sham-controlled study. Cephalalgia 2004;
24: 821-30.
Haspeslagh SR, van Suijlekom HA, Lame IE, Kessels AG,
van Kleef M, Weber WE. Randomised controlled trial of cervical
radiofrequency lesions as a treatment for cervicogenic headache.
BMC Anesthesiol 2006; 6: 1.

88

Govind J, King W, Bailey B, Bogduk N. Radiofrequency

neurotomy for the treatment of third occipital headache.
J Neurol Neurosurg Psychiatr 2003; 74: 88-93.
89 Barnsley L. Percutaneous radiofrequency neurotomy for chronic
neck pain: outcomes in a series of consecutive patients. Pain Med
2005; 6: 282-86.
90 International Spine Intervention Society. Percutaneous
radiofrequency cervical medial branch neurotomy. In: Bogduk N,
ed. Practice Guidelines for Spinal Diagnostic and Treatment
Procedures. San Francisco: International Spine Intervention
Society, 2004: 249-84.
91 Lord SM, Barnsley L, Wallis BJ, McDonald GJ, Bogduk N.
Percutaneous radio-frequency neurotomy for chronic cervical
zygapophysial-joint pain.N EnglJ Med 1996; 335: 1721-26.
92 Lipton RB, Silberstein SD, Saper JR, Bigal ME, Goadsby PJ. Why
headache treatment fails. Neurology 2003; 60: 106470.

www.thelancet.com/neurology Vol 8 October 2009