Abstract
Premise
This review summarises the pain syndromes that overlap between headache and facial pain and overlap between pain and cranial nerve lesion.
Problem
These syndromes share two features in common. First, they show both cranial nerve impairment (e.g. palsy, autonomic dysfunction) and pain; second, they have inflammatory (and/or small vessel) processes as the underlying mechanism. A typical representative of these syndromes is recurrent painful ophthalmoplegic neuropathy, which was previously called ophthalmoplegic migraine and was regarded as a migraine subtype. It is now supposed that this syndrome is caused by an inflammation of one of the ocular motor nerves. Other syndromes discussed in this review are optic neuritis, Tolosa-Hunt syndrome, ischaemic ocular motor nerve palsy, and the very rare Raeder’s syndrome. Treatment of all these syndromes is mainly based on steroids.
Potential solution
Management of possibly underlying disorders such as multiple sclerosis or atherosclerosis should also be considered.
Keywords
Introduction
This article reviews orbital/periorbital pain syndromes associated (and thus overlapping) with probably inflammatory or vascular lesions of periorbital cranial nerves. The aim is to help clinicians with differential diagnosis and give an overview on the state of the art in these syndromes. These syndromes are of particular interest, since they show both pain and focal neurological features. They are painful, which interests craniofacial pain specialists, and on the other hand most of them have inflammatory/immunomediated aetiology. Furthermore, the history of these syndromes is remarkable since some of them show changes of description and pathophysiological explanation over time. On the other hand, the anatomical basis of these syndromes provides valuable knowledge for other headache and pain syndromes and cranial nerve disorders.
Headache and facial pain syndrome with involvement of cranial nerves others than nerve V.
Optic neuritis
Definition
Optic neuritis can be divided into four subtypes: Retrobulbar neuritis; papillitis; perineuritis; and neuroretinitis (2). This is a pain behind one or both eyes caused by demyelination (or other inflammatory processes) of the optic nerve(s) and is accompanied by impairment of central vision. However, the exact clinical definition should also include other causes of optic neuritis, such as infection.
Symptomatology
Patients with autoimmune optic neuritis are typically healthy young adults, and most patients present between 20 and 45 years of age. The headache or orbital pain develops in temporal relationship to optic neuritis and is localised in retro-orbital, orbital, frontal and/or temporal regions. It is usually aggravated by eye movement. Vision of colours is impaired (in particular vision of red colours). The loss of visual acuity increases over hours to days; after two weeks improvement should take place, and after eight weeks visual acuity should be normalised, otherwise the underlying cause should be carefully revised.
Epidemiology
Optic neuritis is often a manifestation of multiple sclerosis. The rough incidence is about 5 in 100,000; the prevalence is about 115 in 100,000 in the northern hemisphere with a female preponderance of three to one (2). Pain may precede impairment of vision. Clinical series report the prevalence of head pain in optic neuritis to be about 90%. There is a high incidence (90%) of pain with eye movement when there is an orbital segment enhancement in cranial magnetic resonance imaging (MRI), and a high probability (70%) of no such pain when there is no enhancement. About 25% of all patients with optic neuritis show disability after 10 years, and about 50% of these patients had started multiple sclerosis treatment in this time period (3). In the Vaga Study, one out of 1,863 inhabitants reported painful optic neuritis (4). Optic neuritis can also occur in children; then, it is more often bilateral (about two thirds of all cases) and shows a more severe loss of visual acuity.
Pathophysiology
The typical optic neuritis is a demyelination of the optic nerve between the chiasma opticum and the bulb disc. This is part of the multiple sclerosis and the neuromyelitis optica spectrum and will, therefore, not be discussed in detail in this review. Other causes of optic neuritis include autoimmune disorders (e.g., Sjögren’s syndrome; neurosarcoidosis; vaccination) and infections (e.g., borreliosis; syphilis; varicella zoster virus) (5–9). The pain originates, at least in part, from involvement of trigeminal nerve fibres in the optic nerve and in its surrounding tissues (in particular trigeminal meningeal fibres).
Diagnostic considerations
The diagnosis of optic neuritis is a clinical one. If it is regarded as necessary to prove optic neuritis, this is best shown by contrast enhancement in the optic nerve or by thickening of the optic nerve in the MRI scan. Clinical, electrophysiological, and laboratory evidence can also confirm the presence of optic neuritis; visually evoked potentials might be useful to monitor the long-term improvement and to classify the demyelination and/or axonal damage. There might be secondary optic neuritis due to infection or other autoimmune disorders. Therefore, cerebrospinal fluid (CSF) analysis is mandatory, at least in the first episodes of optic neuritis. More than 50% of all patients with first autoimmune optic neuritis will develop multiple sclerosis. Therefore, optic neuritis can be the first step in the diagnostic procedure and treatment of multiple sclerosis.
Treatment
The treatment of (painful) optic neuritis depends on the pathogenesis. In typical autoimmune (or idiopathic) optic neuritis, steroids are the treatment of first choice (10,11), although the published evidence is poor (12). There is an ongoing debate about which way steroids should be given. Usually, 500 to 1,000 mg prednisolone or prednisone are given for three to five days. In case of unsuccessful treatment, another steroid treatment for three days with 2,000 mg is recommended. If this is also not helpful, plasmapheresis can be considered. Furthermore, optic neuritis as a clinically isolated syndrome might be an indication to start a specific immunomodulatory treatment of multiple sclerosis. In the case of symptomatic optic neuritis, the underlying disorder should be treated (e.g., antibiotic treatment in optic neuritis due to borreliosis).
Headache attributed to ischaemic ocular motor nerve palsy
Definition
This headache syndrome is a unilateral frontal and/or periorbital pain caused by and associated with other symptoms and/or clinical signs of ischaemic paresis of the ipsilateral third, fourth, or sixth cranial nerve. This also includes diabetic ophthalmoplegia, since this is regarded as a consequence of the affection of very small vessels (vasa vasorum), including inflammatory mechanisms.
Symptomatology
There is no systematic review of the typical symptoms of this disorder. Patients often complain of diplopia. The headache develops in close temporal relation to the motor nerve palsy and is localised around the ipsilateral brow and eye. Headache attributed to ischaemic ocular motor nerve palsy can occur prior to or concurrently with the onset of diplopia. The majority of ocular motor nerve palsies are painful, regardless of the presence or absence of diabetes. Pain is most frequent in patients with third nerve palsies, less so in sixth nerve paresis and least frequent in cases of fourth nerve paresis. Ischaemic lesions of the oculomotor nerve usually only affect the motor fibres, but in a very few cases can also affect the autonomic fibres. Therefore, mydriasis can sometimes be a sign of this syndrome as well.
Epidemiology
Prevalence and incidence of this facial pain syndrome are unknown. The distribution of ocular motor nerve palsies to the single cranial nerves is roughly about 20% (third nerve), 20% (fourth nerve), and 60% (sixth nerve), respectively. However, pain is most often seen in third nerve palsy; there is a male preponderance of 60% and a median age of onset of 64 years (13). Interestingly, migraine patients show a 2.5-fold increased risk of developing a palsy of the ocular motor nerves over a three-year observation period than healthy control subjects (14).
Pathophysiology
In most cases, the ischaemic lesion causing the nerve palsy and the pain is due to embolic infarction or a small vessel disease. The underlying disorder of these conditions is very often diabetes mellitus. A monoparesis of the oculomotor nerve is a typical feature of diabetes mellitus and is – in contrast to other causes of such a monoparesis (e.g. ophthalmoplegic aneurysm) – usually painful. However, several other reasons for ocular motor nerve palsy have been detected: Altogether 10% are not of vascular origin such as large vessel disease, giant cell arteritis, neoplasm, or inflammation (12,15). The pain in this syndrome might originate from trigeminal nerve fibres, which have been detected even in the pure oculomotor nerves (16).
Diagnostic considerations
The ischaemic lesion can in some cases be seen in an MRI scan. This is, however, not always possible, since the lesion can be rather small. If the lesion cannot be shown, this diagnosis is a diagnosis of exclusion. A generalised atherosclerosis, as often occurs in diabetes mellitus, can support the assumption of a vascular origin, the glucose serum level and HbA1c should be measured immediately and might support a diabetic painful ophthalmoplegia. Inflammatory causes of the cranial nerve palsy should be excluded by CSF analysis; an ophthalmoplegic aneurysm should be ruled out by imaging.
Treatment
There is no specific treatment for this condition. A diabetes mellitus should be treated strictly. The pain can be treated with simple analgesics. This condition, if of proven vascular origin, should also lead to an unlimited treatment with a platelet aggregation inhibitor. In more than 50%, the palsy disappeared within three weeks (17).
Tolosa-Hunt syndrome
Definition
Tolosa-Hunt syndrome is a unilateral orbital pain associated with paresis of one or more of the third, fourth and/or sixth cranial nerves caused by a granulomatous inflammation in the cavernous sinus, in the superior orbital fissure, or in the orbit. The clinical features of the syndrome were fully described by Tolosa in 1954, while Hunt, in 1961, emphasised the efficacy of steroid treatment (18). The definition according to the IHS classification (1) requires proof of a granulomatous inflammation; however, this can only be sufficiently achieved by biopsy. MRI can reveal inflammation, but not really the granuloma. Therefore, it was suggested that three subtypes of Tolosa-Hunt syndrome should be differentiated: Idiopathic/benign (i.e. normal MRI scan); inflammatory (i.e. lesion seen in MRI and/or biopsy); and symptomatic (18). Some authors differentiate between the benign type (i.e. without MRI lesion) and inflammatory type (i.e. inflammation shown in MRI). The semiological and epidemiological features of these two types are, however, very similar (19).
Symptomatology
The headache precedes the paresis of the third, fourth and/or sixth nerve by less than two weeks (on average by eight days), or develops with it; the third nerve is most commonly affected. The headache is localised around the ipsilateral brow and eye. Some reported cases of Tolosa-Hunt syndrome had additional involvement of the fifth nerve (commonly the first division) or optic, seventh, or eighth nerve. Sympathetic innervation of the pupil is affected in about 30% of patients (20). Most cases are unilateral, but rarely are also bilateral. Tolosa-Hunt syndrome usually occurs only once. However, recurring and side alternating patients have been described occasionally (21).
Epidemiology
Prevalence and incidence of this facial pain syndrome are completely unknown. The mean age at onset is 45 years, the sex ratio is balanced. The duration of the Tolosa-Hunt syndrome episodes lies between three days and eight months, with a mean of about 50 days (18).
Pathophysiology
Granulomatous inflammation of the cavernous sinus, of the superior orbital fissure, or of the orbit, demonstrated by MRI or biopsy, is the cause of the inflammatory subtype of Tolosa-Hunt syndrome. However, structural lesions of the cavernous sinus such as lymphoma (22) can also mimic Tolosa-Hunt syndrome. A lot of symptomatic cases have been described in the literature.
Diagnostic considerations
Careful follow-up is required to exclude other causes of painful ophthalmoplegia, such as tumours including lymphoma, vasculitis, basal meningitis, carotid artery dissection and sarcoidosis, or diabetes mellitus. The MRI scan reveals inflammation in about 50% of all cases with a clinical Tolosa-Hunt syndrome (17). Therefore, it has been suggested that Tolosa-Hunt syndrome should be made a diagnosis of exclusion, since the MRI can be normal. Another aspect of the differential diagnosis of Tolosa-Hunt syndrome is that other entities or syndromes exist that might also mimic Tolosa-Hunt syndrome, such as Eales disease (23) and Gradenigo syndrome (24).
Treatment
Pain and paresis of Tolosa-Hunt syndrome resolve within three days when it is treated adequately with corticosteroids. In some cases, long-term immunosuppression is necessary. Besides drug treatment, radiotherapy has also been applied successfully in some patients (25,26). However, there is little experience with radiotherapy, in particular with respect to eye damage. Therefore, this is not recommended unless corticosteroid treatment is absolutely impossible.
Paratrigeminal oculosympathetic syndrome (Raeder’s syndrome)
Definition
Raeder’s syndrome is a constant, unilateral pain in the distribution of the ophthalmic division of the trigeminal nerve, sometimes extending to the maxillary division, accompanied by Horner’s syndrome and caused by a disorder in the middle cranial fossa or of the carotid artery. It was first described by Raeder in 1918, and fully published in 1924 as an anatomic subtype of Horner’s syndrome (27). The term paratrigeminal oculosympathetic syndrome was suggested by Goadsby in 2002 in order to elucidate the pathoanatomical background of this disorder (28).
Symptomatology
The headache is strictly unilateral and localised to the distribution of the ophthalmic division of the trigeminal nerve, with or without spread to the maxillary division. The pain is aggravated by eye movement and accompanied by ipsilateral Horner’s syndrome and by ipsilateral trigeminal sensory deficits. Anhidrosis is absent secondary to interruption of the postganglionic oculosympathetic pathway (29,30).
The clinical features and the physical examination findings can be associated with intracranial pathology involving the middle cranial fossa. Understanding this pathway is important in distinguishing Raeder’s syndrome from Horner’s syndrome, as the presentation of the former is now recognised to accompany a number of other disease entities in the head and neck region.
Epidemiology
Since there are no epidemiological studies and not even any case series published, the epidemiology of this syndrome is completely unknown.
Pathophysiology
The original description of the paratrigeminal oculosympathetic syndrome suggested an involvement of oculopupillary sympathetic fibers, indicating a lesion of the middle cranial fossa. Whether the term Raeder’s syndrome should be used today is heavily debated, but painful Horner’s syndrome is still considered by some authors to be a diagnostically useful indication of a middle cranial fossa lesion or of carotid artery dissection. In particular, carotid dissection, aneurysm, and stenosis have been verified by imaging as a cause of the clinical picture. In addition, a chronic inflammatory sinus disease that was supported by brain imaging was the cause of a clinical Raeder’s syndrome (29).
Diagnostic considerations
Imaging of the carotid artery, in particular the intracranial section, is mandatory to detect the underlying vessel disorder. Raeder’s syndrome can be misdiagnosed as hemicrania continua or paroxysmal hemicrania and vice versa. It is important to perform a trial of indomethacin, which is usually not helpful in Raeder’s syndrome (31), although the transition of Raeder’s syndrome into indomethacin responsive hemicranias continua has been described (32). The response to indomethacin to such pain syndromes is however unspecific (33). Recognition that periarterial inflammation can cause Raeder’s syndrome should alert the clinician to request relevant imaging studies. The spectrum of symptomatic cases includes vascular damage, tumours, and infections (e.g. borreliosis).
Treatment
If a clear underlying vessel disorder has been detected, this should of course be treated adequately. Most cases of idiopathic Raeder’s syndrome were self-limiting or responded to steroids. One case of successful treatment with onabotulinumtoxin A was published (31).
Recurrent painful ophthalmoplegic neuropathy
Definition
Recurrent painful ophthalmoplegic neuropathy (RPON) is a disorder that changed its definition throughout the history of headache classification. It was first described by Charcot as a subtype of migraine (34). The actual definition in the IHS classification (subtype 13.9) is “repeated attacks of paresis of one or more ocular cranial nerves (commonly the 3rd), with ipsilateral headache” (1). In the first IHS classification from 1988, this syndrome war regarded as a subtype of migraine and called ophthalmoplegic migraine (35). However, it soon became clear that there is a different pathophysiological background. Therefore, this syndrome changed its place in the second edition of the IHS classification in 2003 and became syndrome 13.17 within the chapter of cranial neuralgias and facial pain, but was still called ophthalmoplegic migraine (36). In the years thereafter, it was obvious that this syndrome is most probably an inflammatory neuropathy and it is now called RPON.
Symptomatology
A unilateral headache is accompanied by ipsilateral paresis of one, two, or all three ocular motor nerves. Orbital, parasellar, or posterior fossa lesions have to be excluded by appropriate investigations. The headache itself can develop up to 14 days prior to ocular motor paresis.
The headache features are typical for migraine (i.e., severe, throbbing, aggravated by physical activity) and accompanying symptoms (photophobia in 65%; phonophobia in 56%; nausea in 66%; vomiting in 69%) are frequent but can be absent (37). The paresis is most often seen in the functions of the oculomotor nerve and only rarely in the abducens or trochlear nerve. The neurological dysfunctions normally last much longer than a typical migraine aura. The attacks usually recur but some patients have only one attack in their life (which makes the diagnosis only probable). The frequency is up to one per week, most often less frequently (38). After puberty, the attack frequency usually decreases. The headache is unilateral and can change in a very few cases (but is always ipsilateral to the cranial nerve signs).
The ocular nerve paresis is often complete, in a few cases the pupillomotor fibres of the oculomotor nerve can be spared (39–41). It is debated whether an intermittent unilateral mydriasis, which can occur in a typical migraine attack, is a subtype of migraine or should be regarded as an incomplete RPON (42–44). The ocular nerve paresis fades away over days to weeks in single cases with frequent attacks; persistent mild cases have been described.
Epidemiology
RPON is extremely rare and can occur at any age. However, the highest prevalence is in children under the age of 12. The median age at onset is about eight years, and ranges from seven months to 50 years of age (37). There have been only few investigations into the incidence. One Danish study showed an incidence of 0.7 per million people (45). In an older study of 5,000 headache patients, eight cases of ophthalmoplegic migraine were detected (46). Men are probably more often affected than women.
Pathophysiology
The exact pathophysiology of RPON is unknown. In one patient, a post-mortem autopsy revealed an atrophy of the oculomotor nerve (47). There is no evidence for a genetic background to this syndrome; however, many patients with RPON have a positive family history of migraine (37). One previous hypothesis postulated an oedema in the vessel wall of the carotid artery in the cavernous sinus, which leads to a compression of the cranial nerve (48,49). Other theories postulated inherited anomalies of the oculomotor or abducens nerve with a perforation of a small artery (50). In modern angiography, patients with RPON normally show unremarkable results. However, modern MRI techniques show a reversible enhancement of contrast media with a thickening of the oculomotor nerve in its cisternal part (51–53). A balanced debate on the arguments pro and contra a migrainous background of RPON has been published (37).
Diagnostic considerations
Recent data suggest that Gadolinium enhancement or nerve thickening can be demonstrated using MRI. Therefore, MRI with contrast medium, including thin slices of the orbita and of the chiasma opticum, should be performed on all patients, if possible, during the days of an attack. However, the MRI can be completely unremarkable. Then, RPON is a diagnosis of exclusion. The differential diagnosis comprises all types of inflammatory or space occupying lesions in the parasellar region and in the orbita (45,54–58). The differentiation between RPON and Tolosa-Hunt syndrome can in particular be difficult (59,60). It is typical for RPON that the pain develops before the cranial nerve palsies and lasts for days, whereas in most other similar syndromes the pain and the cranial nerve palsy occur together. Furthermore, the recurrent time pattern of RPON supports the diagnosis. A subarachnoid haemorrhage can sometimes only be ruled out by conventional angiography. Laboratory diagnostics comprise exclusion of vasculitis, of endocrinological disorders, and of diabetes.
Treatment
In previous times, RPON was treated like a typical migraine attack (i.e., with analgesics, even triptans, prophylactic drugs such as betablockers or antidepressants), but most often these drugs failed to be effective. Treatment with corticosteroids is beneficial in a variety of patients, but there are only case reports or case series (37,45,61). There are also clear non-responders to steroids.
Expert opinion: Open questions and burning desires
The group of syndromes described in this review share the involvement of a cranial nerve that normally is not involved in pain processing. Therefore, the main question is why pain is one of the most prominent symptoms in these disorders although trigeminal structures are not primarily involved. We have to look for sensorimotor reflex mechanisms that are only poorly understood in the cranial motor or optic nerves. It is probably too easy just to search for sensory nerve fibres in the motor or optic nerves. More likely, there are reflex mechanisms in the brainstem that lead to pain when efferent nerves such as the oculomotor nerve are affected.
Another problem is the type of inflammation that causes the different painful orbital syndromes. Most of the syndromes have a subtype with an autoimmune (or idiopathic) inflammation of the cranial nerve. We do not know whether this inflammation is triggered by infection, whether it is primarily a demyelination or an axonal loss, and whether there is long-term nerve damage after recurrent inflammations. The mechanisms of these inflammations should be further investigated. It might be that the recent advances in the detection of multiple sclerosis mechanisms can also contribute to the detection of painful cranial nerve inflammation, as both can go alongside demyelination and subsequent axonal damage.
From a clinical and systematic point of view, it is still to be determined whether we want to classify only idiopathic disorders or whether we combine idiopathic and symptomatic types. This is an old dilemma in headache classification, since we try to strictly differentiate between idiopathic and symptomatic headache disorders. However, many headache types (in particular group 4 of the IHS classification and the syndromes discussed in this review) have an idiopathic and a symptomatic pathogenesis leading to the same clinical picture. It would be helpful to allow symptomatic subtypes to be classified among the clinical syndromes in the IHS classification. Otherwise, we would not have a place for these symptomatic facial pain or headache syndromes. The IHS classification therefore has to be opened for other aetiologies. For example, other causes than ischaemic lesions as the underlying mechanism of direct ocular motor nerve palsies should be accepted in the IHS classification.
Expert opinion: Where the field needs to go
One step forward would be to have more information about the inflammatory mechanisms underlying these overlapping syndromes. CSF analysis with methods known from multiple sclerosis research could be helpful for this approach. Histoanatomical studies would also be useful, but are of course very difficult since it is nearly impossible to obtain the relevant tissue.
Epidemiological research is also warranted in these syndromes. This is of course difficult, given their low frequency. However, sampling painful orbital syndromes in a multicentre prospective and well-controlled registry could result in useful information about frequency, demographic distribution, and prognosis of these syndromes. This would be helpful for consulting patients and their relatives. In particular, RPON is a syndrome that affects children and adolescents and is a cause of sincere concern among parents. Better understanding of the natural course of this disorder would support families in coping with it.
From a therapeutic point of view, it would be useful to have more evidence for the different treatment procedures. Steroids are given in most of the syndromes discussed here, and many experts regard this as the most efficacious treatment. However, randomised placebo controlled trials are lacking, and the analysis of the studies available surprisingly shows that steroids are not necessarily better than placebo (12). A multicentre database for the different painful orbital syndromes could be helpful in establishing clinical trials on steroids, and other treatment procedures if steroids are not helpful or contraindicated.
Clinical implications
The diagnosis of these syndromes depends on the association of cranial nerve disorders with focal neurological signs and typical features of facial pain or headache. Most of these syndromes are rare, and many neurologists or pain specialists do not see such cases in their working life. It might be necessary to refer these patients to a specialised headache centre. The pathophysiological background of most of these syndromes is a non-infectious inflammation that is not characterised in detail yet. Therefore, treatment is based on the suppression of inflammatory processes. However, several symptomatic cases have been published and make a diagnostic work-up including brain imaging necessary in all these cases. There is no published evidence on the best treatment. Usually, steroids are given in different doses and for different time periods. Many of these syndromes are self-limiting. However, sometimes long-term immunosuppression might be necessary.
Footnotes
Declaration of conflicting interests
The authors declared the following potential conflicts of interest with respect to research, authorship, and/or publication of this article: SE has received honoraria for speaking and serving on the advisory board of Allergan, Mundipharma, Pfizer, Reckitt Benckiser (past two years).
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.