OUP user menu

Headache Plus: Trigeminal and Autonomic Features in a Case of Cervicogenic Headache Responsive to Third Occipital Nerve Radiofrequency Ablation

Kathryn Giblin MD, Jordan L. Newmark MD, Gary J. Brenner MD, PhD, Brian J. Wainger MD, PhD
DOI: http://dx.doi.org/10.1111/pme.12334 473-478 First published online: 1 March 2014


Objective To describe a case of cervicogenic headache with associated autonomic features and pain in a trigeminal distribution, all of which responded to third occipital nerve radiofrequency ablation.

Design Single case report.

Setting Massachusetts General Hospital Center for Pain Medicine.

Patients A 38-year-old woman with history of migraines and motor vehicle accident.

Interventions Right third occipital nerve diagnostic blocks and radiofrequency lesioning.

Outcome Measures Pain reduction; physical findings, including periorbital and mandibular facial swelling, tearing, conjunctival injection, and allodynia; and use of opioid and non-opioid pain medicines.

Results The patient had complete relief of her pain and autonomic symptoms, and was able to stop all pain medications following a dedicated third occipital nerve lesioning.

Conclusions This case illustrates the diagnostic and therapeutic complexity of cervicogenic headache and the overlap with other headache types, including trigeminal autonomic cephalgias and migraine. It represents a unique proof of principle in that not only trigeminal nerve pain but also presumed neurogenic inflammation can be relieved by blockade of cervical nociceptive inputs. Further investigation into shared mechanisms of headache pathogenesis is warranted.

  • Cervicogenic Headache
  • Trigeminal Autonomic Cephalgia
  • Radiofrequency Ablation
  • Third Occipital Nerve


In cervicogenic headache, pain is classically present in the occipital, parietal, and auricular regions innervated by cervical nerves, but can also occur in the frontal and orbital regions subserved by the V1 branch of the trigeminal nerve. Anatomically, the basis of cervicogenic headache is in the spinal trigeminocervical complex, where the spinal trigeminocervical nucleus receives convergent primary nociceptive inputs from both cervical (C1, C2, and C3) and trigeminal (V1 spinal tract) sensory afferents, and then gives rise to the second-order ascending trigeminothalamic tract [1]. Due to this dual input from C1–C3 and V1 onto the second-order nociceptive neurons, cervical pathology can result in referred pain in cervical and occipital as well as trigeminal distributions, including the dura. This wide distribution of referred pain has been validated experimentally. For example, noxious focal stimulation with hypertonic saline of cervical structures innervated by C1, C2, and C3, such as the suboccipital and posterior cervical muscles, atlanto-occipital, lateral atlanto-axial, and C2–C3 zygapophysial joints, elicits occipital, parietal, frontal, and orbital pain [2,3]. The greater occipital nerve arises from the medial branch of the dorsal ramus of C2, while the third occipital nerve (TON) stems from the medial branch of the dorsal ramus of C3 and runs superficially. The C2–C3 zygapophyseal joint is innervated predominantly by the TON, along with articular branches from the C2 dorsal ramus and infrequently with modest innervation from the greater occipital nerve. The C3–C4 zygapophyseal joint is innervated by the deep portion of the C3 medial branch, after it gives off the TON, as well as the C4 medial branch [4].

Although the pathophysiological basis for cervicogenic headache has been well elucidated, particularly in comparison to other headache types, diagnostic criteria and the role of cervical procedural interventions in management are widely debated. Purely clinical diagnostic criteria for cervicogenic headache include unilateral pain, reduced range of neck movement, ipsilateral shoulder and arm discomfort, and mechanical precipitation of attacks by awkward neck positions or external pressure against occipital structures [5]. These criteria, however, have poor interrater reliability and specificity [6,7]. Others argue that in order to attain validity, a cervical source of the pain must be identified via diagnostic blocks [1]. Reflecting this disagreement, the International Headache Society (IHS) criteria include evidence of causation via multiple different modalities, including both clinical maneuvers and diagnostic blockade (Table C). This creates a certain tension between definitional criteria that include response to therapy and the prospect of performing invasive diagnostic interventions in patients who likely do not have cervicogenic headache, particularly as neck pain is a prominent feature of migraine in over 73% of cases [8], and as such neck injections are usually not warranted. The extent of the diagnostic morass is highlighted in the IHS definition, in which clinical diagnostic tests “must have demonstrated reliability and validity. The future task is the identification of such reliable and valid operational tasks.” Moreover, the clinical overlap between cervicogenic headache and other primary headache syndromes further complicates the criteria for the diagnosis of cervicogenic headache.

View this table:
Table 1

International Classification of Headache Disorders 3rd edition (ICHD-3) (beta) criteria

11.2.1: Cervicogenic Headache
A.Any headache fulfilling criterion C
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 able to cause headache
C.Evidence of causation demonstrated by at least two of the following:
1.Headache has developed in temporal relation to the onset of the cervical disorder or appearance of the lesion.
2.Headache has significantly improved or resolved in parallel with improvement in or resolution of the cervical disorder or lesion.
3.Cervical range of motion is reduced and headache is made significantly worse by provocative maneuvers.
4.Headache is abolished following diagnostic blockade of a cervical structure or its nerve supply.
D.Not better accounted for by another ICHD-3 diagnosis

We present a case of cervicogenic headache with trigeminal distribution of pain and facial swelling that was fully—albeit not permanently—relieved by radiofrequency lesioning of the TON at the C2–C3 facet joint. This serves as proof of principle that referred CN V distribution pain and even associated autonomic features can be ameliorated by disruption of the appropriate nociceptive cervical pain afferents relayed via the spinal trigeminocervical complex. A unique feature of this case was the significant associated facial swelling, which we postulate was due to trigeminal nerve-mediated neurogenic inflammation. Given the clinical similarities seen in this case between cervicogenic headache and trigeminal autonomic cephalgias (TACs), we discuss the role of unified diagnostic criteria to differentiate cervicogenic headache from overlap syndromes with migraine and TACs, as well as the use of cervical radiofrequency ablation in other primary headache syndromes. Finally, we advocate for the superiority of dedicated TON lesioning over the C3 medial branch lesioning.


A 38-year-old right-handed woman presented with 3 years of right-sided facial and neck pain since a motor vehicle accident with whiplash injury. The pain was cyclical, with several months of severe pain followed by months of relatively lessened pain. At its worst, the pain would keep her from working as a respiratory therapist. The pain was described as a burning, pressure-like sensation located in the right neck, ear, jaw, eye, and right nostril, with allodynia and burning sensation provoked by wind or hair touching her on the right face and occasionally with breathing in the right nostril. Associated with the pain exacerbations, she experienced right facial swelling, erythema, and right eye tearing; she did not report eye drooping. The pain was increased by head rotation, Valsalva, coughing, sneezing, and swallowing. The pain was not relieved by carbamazepine, amitriptyline, cyclobenzaprine, metaxalone, or topical lidocaine. The combination of gabapentin and hydrocodone/acetaminophen provided limited benefit.

Past medical history included well-controlled migraines characterized by throbbing pain in either temple with associated phonophobia, photophobia, and nausea lasting a few hours, occurring a few times a year, and relieved by over-the-counter medications. Thus, the new side-locked headache and associated symptoms fundamentally differed from those associated with her prior migraines. Her medical history was also notable for a congenital tracheoesophageal fistula status post thoracotomy, with presumed recurrent laryngeal nerve lesion resulting in chronic left vocal cord paralysis and mild hoarseness.

Her physical examination was notable for facial asymmetry, with right-sided predominantly periorbital and mandibular facial swelling, but symmetric facial movements. Extraocular movements, sensation to all modalities, and hearing were intact. Speech was chronically hoarse, but fluent, with no lingual, guttural, or buccal dysarthria. The tongue and palate moved in the midline. There was no change in taste. Gag was chronically absent. Shoulder shrug and head turn strength were full. There was no pronator drift. Deep tendon reflexes were brisk and symmetric, and Babinski response was absent bilaterally. There was no ataxia, and station and gait were normal. Spurling's test was normal; however, there was marked pain with palpation in the region of the upper cervical facets on the right. There was no restriction of neck range of motion, although palpation of the upper right cervical facets produced pain radiating to the right arm and occiput. Head magnetic resonance imaging (MRI) and magnetic resonance angiogram were unremarkable.

Based on the focal unilateral facet tenderness, it was felt that she had a combination of cervical nerve and trigeminal involvement due to a presumed cervicogenic headache. She first underwent diagnostic medial branch blocks with 0.5 mL of 1% lidocaine at the right C3, C4, and C5 levels to determine if facet arthropathy at any of these levels was acting as a primary pain generator. These blocks resulted in 90% relief of her right cervical and occipital pain that lasted for several hours, the expected duration of the local anesthetic. To confirm efficacy, the same diagnostic blocks were repeated a week later, again with 90% relief in her right cervical and occipital pain. Based on the positive diagnostic blocks, it was felt that she likely did have cervicogenic pain originating in part from facet disease, and that she might benefit from cervical medial branch radiofrequency lesioning. The impedance was between 300 and 800 ohms. Sensory stimulation at 50 Hz and motor stimulation at 2 Hz were carried out prior to each cycle of lesioning, which elicited discomfort but no motor stimulation. Denervation was performed for 60 seconds at 80°C with single lesions adjacent to the C3, C4, and C5 articular processes; she had marked but only temporary improvement of her pain and right facial swelling. Repeat MRI with and without contrast with thin cuts through the skull base to look at the jugular foramen was normal, as was MRI of the anterior neck. Cervical spinal MRI showed minimal degenerative changes. She underwent a second radiofrequency lesioning of the right C3, C4, and C5 medial branches a year after her initial procedure, with 50% improvement in her right-sided pain and swelling. The pain returned after 5 months, and she had a third right C3, C4, and C5 medial branch radiofrequency lesioning at that time, which provided no relief. As it was felt that she might still have cervicogenic pain involving the TON, she underwent a diagnostic right TON block with 0.5 mL of 1% lidocaine over the C2 and C3 facet pillars with excellent pain relief, which was confirmed on repetition a week later. She proceeded to radiofrequency lesioning of the right TON via multiple [3] focal lesionings, again for 60 seconds at 80°C, at and just inferior to the border of the C2 and C3 facet pillars. Following the procedure, she had complete resolution of her symptoms, including her auricular, ocular, and nasal pain, as well as her right facial pain and swelling, and she was able to titrate off of gabapentin (2,400 mg/d) and hydrocodone/acetaminophen (5 mg/500 mg Q6 hours) and return to work. This improvement lasted a year and a half, before her symptoms returned, and she had repeat right C2–C3 TON radiofrequency lesioning, again with abolition of her pain, facial swelling, tearing, and injection.


Here, we discuss a patient with cervicogenic headache relieved by cervical blocks and ultimately radiofrequency ablation. There are a number of unique features of this case. The headache syndrome was accompanied by autonomic features, which were also relieved by both local anesthetic blockade and radiofrequency neurolysis. The overlap between cervicogenic headache and the TACs seen in this case highlights several issues: the complexity of appropriate and useful diagnostic criteria for cervicogenic headache, the role of cervical injection for cervicogenic headache, and the potential overlap syndromes with TAC and migraine. Lastly, the greater efficacy of TON lesioning over the C3 medial branch lesioning observed in this patient despite the fact that the TON arises from the C3 medial branch highlights the importance of careful selection of interventions and potentially the effects of anatomic variability among individuals.

Another notable aspect of this case was the patient's profound periorbital and mandibular swelling, and its relief by cervical medial branch blockade and lesioning. Ipsilateral eyelid edema has been described in early literature on cervicogenic headache [9]. In the Cervicogenic Headache International Study Group criteria, it falls under category VI, rarely occurring associated phenomena, and is specifically noted to be present all the time, not only during attacks as the described patient experienced [5]. In TACs, periorbital edema is relatively common, and hypotheses to explain this phenomenon have included trigeminal discharge [10], central autonomic dysregulation, and parasympathetic hyperactivity associated with hypothalamic disturbance [11], as well as compression of pericarotid sympathetic fibers due to perivascular edema from parasympathetic overactivity [12]. While the TACs are felt to originate in the hypothalamic–pituitary axis [13,14], the spinal trigeminocervical complex implicated in cervicogenic headache would seem less likely to cause direct hypothalamic disturbance or pericarotid parasympathetic overactivity. The convergence of spinal and trigeminal afferents in the spinal trigeminocervical complex may also indirectly activate the superior salivatory nucleus mediating parasympathetic activity [15] as experimental stimulation of the greater occipital nerve branch of C2 has resulted in ipsilateral tearing, ptosis, and conjunctival injection in rats [16]; however, the potential anatomical connections for these effects are less clear. Alternatively, the trigeminal-distribution edema seen in cervicogenic headache could be due to neurogenic inflammation, whereby the stimulation of primary trigeminal nociceptive central terminals causes antidromic propagation in the trigeminal first-order nociceptors, peripheral release of neuropeptides calcitonin gene-related peptide and substance P, and subsequent vasodilation and increased capillary permeability [17,18]. However, such a mechanism would also need to explain the activation of the primary central trigeminal nociceptive terminals sufficient to generate the neurogenic inflammation. We theorize that in cervicogenic headache, the first-order trigeminal nociceptive neuron may be stimulated by nitric oxide from either the postsynaptic trigeminocervical complex neuron via retrograde synaptic transmission [19], or from the parallel occipital neuron activated by cervical pathology [20]. Such a mechanism, based on the production of postsynaptic or perisynaptic nitric oxide, and the development of neurogenic inflammation in the presynaptic trigeminal neuron, would require in vitro validation.

Although periorbital edema has been documented in cervicogenic headache, the potential clinical overlap in this case with the TACs illustrates the difficulty of diagnosis of cervicogenic headache. Significant overlap also exists in the symptomatology of cervicogenic headache, migraine, and hemicrania continua (HC). Per Sjaastad et al., 20% of cervicogenic headache is associated with photophobia and other features that are classically thought of as “migrainous,” including phonophobia, nausea, dizziness, and ipsilateral blurred vision, although “to a lesser degree” than in migraine [5]. Further complicating matters, neck pain is strikingly prevalent in migraineurs: in a study of 113 migraineurs who had been examined by headache specialists and felt to have pure migraine, neck pain was actually more prevalent in migraine than nausea, with up to 73% of migraines occurring with concomitant neck pain [8]. Studies have found increased neck muscle girth in migraineurs [21], as well as increased antagonist muscle activity on electromyogram during neck movements [22]. Adding to the complexity, orbital, facial, and auricular swelling has been reported in HC [23], which can be primary or posttraumatic [24]. Moreover, an overlap syndrome of cervicogenic headache and HC has been reported [25].

While the pathophysiology of cervicogenic headache is relatively well established, the clinical entity of cervicogenic headache is strikingly convoluted. There are multiple sets of diagnostic criteria as well as differing opinions about the role of interventional approaches in diagnosis and treatment. As noted above, the clinical criteria for cervicogenic headache suffer from poor interrater reliability and specificity with regard to range of motion and precipitation by external pressure [6]. On the interventional end of the diagnostic spectrum, the IHS criteria for the diagnosis of cervicogenic headache include abolition of pain via blockade of nerves that innervate responsible cervical lesions. Certainly, performing cervical medial branch blocks on large populations of refractory migraineurs with neck pain is inappropriate. Given reported magnitude of the placebo effect in interventional pain studies, assessing non-placebo benefits of cervical and occipital procedures is particularly challenging. Nonetheless, there is evidence that suboccipital steroid injection in the area of the greater occipital nerve may reduce the frequency of cluster headache [26,27], and greater occipital nerve blocks may be efficacious for short-lasting unilateral neuralgiform headache with conjunctival injection and tearing (SUNCT syndrome) [28], HC [29], and migraine [30]. Occipital nerve stimulation is being investigated for a number of primary headache syndromes [31].

These findings raise the question of whether cervical pathology may contribute to these other headache types and whether interruption of cervical somatosensory input may provide benefit in the absence of cervical pathology. Thus, potential analgesic response from medial branch and TON blockade in non-cervicogenic primary headache syndromes may confound the specificity of cervical facet blocks in diagnosing cervicogenic headache.

While the described patient had a history of cervical whiplash injury and a physical exam finding of precipitation of pain with palpation of right cervical facets, both consistent with a clinical diagnosis of cervicogenic headache, she had less therapeutic benefit from C3, C4, and C5 medial branch lesioning than from TON lesioning. This emphasizes the importance of appropriate choice of level of cervical medial branch lesioning in the treatment of headache. This is concordant with the findings of Govind and colleagues that more aggressive lesioning was required for TON block compared with earlier studies of lower cervical medial branch radiofrequency ablation [32,33]. While six passes in third occipital lesioning (as performed in Govind et al.) may not be well tolerated in all patients with only local anesthetic and without monitored anesthesia care, it may be that dedicated TON radiofrequency ablation using a smaller number of lesions can suffice.

In conclusion, despite having well-defined anatomic pathophysiology, the diagnosis and treatment of cervicogenic headache present a clinical conundrum. There is significant overlap with the TACs and migraine, which can hinder clinical diagnosis. Interventional diagnosis may be confounded by differences in technique and by benefit from intervention in primary headache syndromes other than cervicogenic headache. This case documents that trigeminal pain can be a component of cervicogenic headache, and that it can be relieved by cervical medial branch and TON neurolysis. The case also represents a unique proof of principle in that not only CN V pain but also presumed neurogenic inflammation in a CN V distribution can be relieved by blockade of cervical nociceptive inputs. Finally, multiple components of cervicogenic headache, TACs, and migraine may be present in individual patients, thus underscoring the importance of future investigation into shared mechanisms of headache pathogenesis.


  • Conflict of interest/disclosures: None.


View Abstract