Abstract
Background
Nummular headache and epicrania fugax are two primary headaches that fall under the heading of epicranial headaches.
Methods
This article reviews the epidemiological and clinical features of nummular headache and epicrania fugax, proposed pathogenic mechanisms and state-of-the-art management, according to the literature.
Results and conclusions
Nummular headache and epicrania fugax are generally viewed as rare headache disorders, but no population-based epidemiological data are available. Nummular headache is characterized by continuous or intermittent head pain, which remains circumscribed in a round or oval area of the scalp, typically one to six centimeters in diameter. Epicrania fugax manifests with brief paroxysms of pain that move along the surface of the head, following a linear or zigzag trajectory through different nerve territories. Nummular headache and epicrania fugax are mostly primary headaches, but some secondary cases have been reported. The pathogenesis of these headaches is not fully understood. Nummular headache could probably originate in epicranial tissues or adjacent intracranial structures, while the origin of epicrania fugax could be extracranial or intracranial. Diagnostic assessment requires careful examination of the symptomatic areas. Underlying disorders should be excluded by additional investigations, including neuroimaging and appropriate blood tests. No controlled clinical trials have been conducted in nummular headache or epicrania fugax. Analgesics and anti-inflammatory drugs, botulinum toxin and gabapentin are currently the most recommended treatment options for nummular headache. In epicrania fugax, the most used treatments are gabapentin, lamotrigine, and other antiseizure medications.
Nummular headache
Introduction
The term nummular headache (NH) is derived from the Latin word for coin (nummus) and refers to the most distinctive feature of this entity: NH is a primary headache that remains restricted to a small coin-shaped area. In 2002, Pareja et al. published the first series of patients with this type of pain (1). Two years later, NH was added to the Appendix of the 2nd edition of the International Classification of Headache Disorders (ICHD-II) with provisional diagnostic criteria (2), based on the characteristics of those first patients. NH soon became a universally recognized condition. Numerous case descriptions and case series led to the updating of the diagnostic criteria for NH, as well as its inclusion in the main body of the 3rd edition of the International Classification of Headache Disorders (ICHD-3, Table 1) (3).
Diagnostic criteria for nummular headache (ICHD-3).
ICHD-3: International Classification of Headache Disorders, third edition, 2018 (3).
Epidemiology
NH may be considered an unusual type of headache, but it is not exceptional. To date, about 540 different cases have been described in full-length papers, with some of these cases being included in more than one paper (1,4–66). The gender distribution has shown a female predominance, with a ratio of 1.6:1. The age of onset has been extremely variable, ranging from four to 86 years, with a mean age around 48 years. Similarly, the duration of symptoms before diagnosis has been highly variable, ranging from less than one month to 50 years.
There are currently no population-based epidemiological data, although some estimates of the frequency of NH are available. In the largest published series, which included 225 patients attending a Spanish headache clinic over 10 years, patients with NH accounted for 4.1% of headache patients seen during the inclusion period (58). Another Spanish series involving 83 patients diagnosed with NH in a neurology department over five years yielded estimates of 6.64/100,000 for annual incidence and 0.03% for prevalence in the general population (59). Still, it is to be expected that many people affected by NH do not consult a doctor or are not referred to a neurologist, particularly those with less persistent or less severe pain.
Clinical features
NH pain is confined to a small cranial area, which patients can easily demarcate with their finger. The shape of the painful area is typically round (in more than 80%) but can also be oval. According to the ICHD-3 diagnostic criteria (3), its diameter ranges from one to six centimeters, although atypical cases outside this range have been described, with diameters between 0.6 and 10 centimeters (15,24,37). Remarkably, there is no pain outside this particular circumscribed area, and both the shape and diameter of the painful area remain fixed over time.
Any part of the cranial vault can be affected, but the most frequent location is the parietal region, specifically the parietal tuberosity, its most convex area (Figure 1); this was the location in more than 40% of the cases described. The occipital location is also relatively frequent, with locations in the temporal or frontal regions or in the border territories being less common. Normally the pain affects one of the two sides of the head; however, more than 50 cases have been reported with the painful area over the midline (4,6,7,9,15,24,32,37,55,58,59,62,65). Thirty-six patients have also been reported with bifocal NH, with two separate areas of pain (22,23,30,32,34,38,58,60). In these cases, the temporal sequence has been heterogeneous, either migratory (with transfer of symptoms from one area to another), additive (with aggregation of symptomatic areas) or synchronous (with simultaneous evolution of symptoms in all locations) (22,32). In addition, two cases of multifocal NH have been described, one with three separate symptomatic areas (28), and one with four (50).
Nummular headache is most frequently located in the parietal tuberosity.
The quality of NH pain is variable. It is mostly described as pressing (in about 40% of cases), but can also be stabbing, burning, or throbbing and, less frequently, boring, cutting, or electric. Pain can change in character from one moment to another in the same patient. The intensity of pain is also variable. In the initial descriptions, the intensity was mild or moderate (1,3), so that patients consulted more to ensure the benignity of their condition than for pain relief. However, patients with severe pain (9,12,33,37,57,62,65) and even refractory to multiple treatments (18,29,37,39,48,56,63) were subsequently described. In the most extensive series published, the mean pain intensity was 5.2 on a 0–10 numerical rating scale (58).
Exacerbations of baseline pain are frequent (4,9,12–14,16,18,21–23,26,32,33,40,53,56,58–60,63,64,66). These exacerbations sometimes occur spontaneously and sometimes are triggered by the application of light stimuli to the symptomatic area (touch, combing, etc.), or by other types of stimuli, such as head movements or Valsalva maneuvers. In a large series focusing on these events, more than half of the patients reported exacerbations. The usual length of exacerbations was seconds or minutes, their quality was mainly stabbing, and their intensity was on average 2 points higher than baseline pain (33).
The temporal profile of NH is heterogeneous. Among the published cases with a defined temporal pattern, three-quarters had persistent pain or pain occurring more than 15 days per month, and one-quarter had pain less than 15 days per month. On symptomatic days the pain may be continuous or intermittent; when pain is intermittent, it is usually present for several hours a day (4,6,17,22,23,25,47,48,57,62), but some cases have been described with very brief paroxysms of pain, lasting minutes (23,27,55,64) or even seconds (12,23,40,64). In the long term there may be spontaneous remissions, with or without relapses. In the largest series published to date, one-third of patients followed for more than one year had treatment-free remissions lasting more than three months (58). The possibility of spontaneous remissions should be kept in mind when assessing treatment responses.
In addition to spontaneous pain, there may be sensory disturbances in the affected area. The description given for these disturbances is diverse, but in one form or another they can be found in many cases of NH (1,4,6,10–14,16,20–27,29,31,35,39,41,42,47,48,51,53,57–60,63–66). In one of the largest series, sensory symptoms or signs were detected in more than 40% of cases (33). Patients may report numbness or hypersensitivity whereby innocuous stimuli, such as touching or brushing, may trigger abnormal sensations or an exacerbation of pain. Physical examination may reveal hypoesthesia for tactile and/or painful stimuli or, conversely, hyperesthesia including allodynia and/or hyperalgesia. In clinical practice, a detailed sensory examination can be used to precisely delineate the painful area in patients with NH (Figure 2). Symptoms and signs of sensory dysfunction may be present even while the patient is not experiencing spontaneous pain.
Allodynia and hyperalgesia in a circular area 4 cm in diameter in the right parietal region, corresponding to the painful area reported by a patient with nummular headache. The area of allodynia and hyperalgesia was delimited by applying successive touch and painful stimuli from the surrounding regions to the area of pain. In cases where an area of hypoesthesia has to be delimited, the stimuli are applied in the opposite direction, from the affected area outwards.
Trophic changes have been identified at the site of pain in a minority of patients. These changes consist of skin atrophy and/or hair loss, sometimes accompanied by redness and a local increase in temperature. Trophic changes diminish or disappear when the pain subsides or when it is controlled with therapy. Overall, 10 cases with obvious trophic changes can be found in the literature (1,16,28,39,57), but careful palpation of the symptomatic area could probably detect subtle skin depression in a higher proportion of NH patients. Local changes in temperature (warm or cold) have also been reported in three patients with no other skin abnormalities (55). There is also a description of a child who started to present NH at the age of four years in an area of the scalp that had a lighter tuft of hair since birth (27).
NH is not usually associated with general or systemic symptoms. However, some cases of NH recruit features of migraine. Four patients have been described with typical migraine accompaniments, i.e., photophobia, phonophobia and/or nausea (12,25,49,57). One of them had a typical migraine trigger (menstruation) (25). In another of these patients, pressure algometry revealed the presence of generalized cranial hypersensitivity, such as can be found in patients with migraine (49). It has been suggested that patients with migraine features might benefit from anti-migraine therapies.
A possible variant: Linear headache
In recent years, linear headache (LH) has been described as a new type of headache whose characteristics are like those of NH, but with a linear-shaped painful area. LH manifests as a well-demarcated continuous or intermittent pain with a fixed linear shape, shows no movement or irradiation along the line of pain, and is not confined to the territory of any single cranial nerve (67).
To date, 37 cases of LH have been reported in the literature (67–70), with a mean age of onset in the fourth decade of life, and with a female predominance. As in NH, a significant proportion of patients with LH exhibit symptoms and/or signs of sensory dysfunction within the painful territory. In addition, some patients have typical migraine symptoms, such as nausea, photophobia or phonophobia. It remains to be established whether LH may be a variant of NH, migraine, or another primary headache, or whether it is a separate entity.
Association with other headaches
NH frequently coexists with other primary headaches, particularly migraine (1,4,5,13,19–21,23–25,35,53,60,62,66). However, when this occurs, migraine and NH present independently and without a clear causal relationship. At times NH coexists with epicrania fugax (EF), and in these cases there does appear to be a connection. As will be discussed below, a coin-shaped pain with features of NH may sometimes persist at the point of origin of EF paroxysms (29,36,71–76).
Psychological factors
Initially, some authors suggested that NH could be a psychogenic condition (7,77). However, no differences have been found between the levels of anxiety and depression of patients with NH and those of the general population (53,78). Moreover, the emotional state of patients with NH has not been associated with the clinical parameters of pain intensity and frequency, or with the size of the painful area (78).
Secondary forms
NH is mostly considered a primary headache. Nevertheless, several cases of “secondary NH” have been reported in which the symptoms were attributed to an underlying structural lesion. Almost all the cases described showed a close anatomical relationship between the pain and the underlying lesion. Furthermore, in some cases resection or disappearance of the lesion was followed by resolution of the pain. Some of these lesions were in the skin or subcutaneous tissue: two patients had superficial fusiform aneurysms (48), one had a varicella-zoster rash (35), one had a calcified hematoma (44), one had linear scleroderma (51), and one had a subcutaneous cyst (64). Other cases had adjacent bone lesions: two suggestive of Paget’s disease (43,64), two eosinophilic granulomas (54), two intraosseous hemangiomas, one osteoma, one cholesterol cyst, one arachnoid granulation (64), one fibrous dysplasia (21), and one case of bone deformities due to craniosynostosis (45). Finally, some cases of NH have been associated with the presence of intracranial lesions. Most of these lesions were close to the skull – two arachnoid cysts (20), one meningioma (10) and one cavernoma (64) – but one had a deeper location, in the suprasellar region – a pituitary adenoma (41).
On the other hand, some patients link the onset of NH to a precipitating event. The most common precipitant is head injury; there are more than 30 cases reported in the literature (1,4,9,16,58). In addition, two patients started having pain after cranial surgery (21,42), and one patient related his symptoms to an insect bite in the affected region (16). In certain cases, NH could be caused or triggered by autoimmune conditions. This question has not been extensively analyzed, but markers of autoimmunity were detected in 16 out of 23 patients in one case series; 15 of them were finally diagnosed with Sjögren syndrome, rheumatoid arthritis, or antiphospholipid antibody syndrome (34).
Pathogenesis
The pathogenesis of NH is not fully understood. The restriction of symptoms and signs to a limited area of the cranial region indicates that it is not a generalized process. Indeed, unlike in patients with migraine or tension-type headache, standardized palpation in NH patients does not show diffuse hypersensitivity of pericranial structures (79). Moreover, pressure algometry has demonstrated a decrease in pain thresholds only at the symptomatic point, and not in the rest of the scalp (80,81). These data suggest that NH is a localized process, and that its origin is probably peripheral rather than central (82).
Initially, the possibility was raised that NH was a neuralgia of one of the terminal branches of a cutaneous nerve (1,2). However, in some patients the painful area is just in the midline (4,6,7,9,15,24,32,37,55,58,59,62,65), which contradicts this hypothesis. Moreover, no consistent improvement has been achieved in cases where anesthetic infiltration of the painful area has been performed (1,7,9,13,15,18,19,21,56,59,63,64). The origin of NH could be a local lesion or dysfunction of the terminal branches of the pericranial nerves, but it could also be localized at a deeper level. It has been suggested that NH could originate in the epicranial tissues, i.e., in any layer of the scalp or the skull (83). Secondary forms attributed to structural lesions support this concept, since almost all have been linked to extracranial lesions (35,44,48,51,64), bone lesions (21,43,45,54,64), or intracranial lesions adjacent to the skull (10,20,64). Post-traumatic cases (1,4,9,16,58) could also be due to contusion or injury of epicranial tissues. The cause of the idiopathic or primary forms of NH remains unknown, but it is very likely that the origin of the pain lies at these same levels. In both rodents and humans, the presence of nerve fibers passing through the skull via emissary channels and fissures has been observed by fluorescence microscopy and electron microscopy (84,85). Moreover, in experimental animals it has been shown that collaterals penetrating the skull can be activated by noxious stimuli from the intracranial and extracranial sides, i.e., from the dura mater and periosteum (86,87). It seems plausible that these fibers may be involved in the transmission of painful impulses in NH.
Diagnosis
As with other primary headaches, the diagnosis of NH is based on the identification of very well-defined clinical characteristics, in accordance with the ICHD-3 diagnostic criteria (Table 1). In addition, the presence of a lesion or disease that may be the cause of the pain must be excluded in all cases (3).
Careful assessment of the symptomatic area should be carried out, with inspection and palpation of the skin, and testing for sensitivity to touch and pain. As detailed above, the NH area may show trophic changes and/or signs of sensory disturbance (Figure 2). The skull itself, pericranial nerves, arteries, muscles and the trochlear region should also be palpated for other possible sources of pain. A complete neurological examination should also be performed to rule out the presence of neurological deficits. In all cases, neuroimaging studies must be obtained, including cranial computed tomography (CT) and/or magnetic resonance imaging (MRI) (64) and, in selected cases, ultrasound imaging or echo-Doppler of the scalp (48). Blood tests including sedimentation rate and/or C-reactive protein, and probably an autoimmunity profile, should be ordered for a complete evaluation (34). In patients with local skin changes a biopsy can occasionally be diagnostic of an underlying disease, as in the case of scleroderma (51).
NH must be differentiated from other conditions. First, the restriction of symptoms to a circumscribed area may resemble painful neuropathies or neuralgias affecting the pericranial nerves, such as supraorbital neuralgia (88), auriculotemporal neuralgia (89) or occipital neuralgia (90). In these cases, however, the pain is not limited to a circular area but extends over the nerve territory. Moreover, it is often accompanied by hypersensitivity to palpation of the nerve. The coincidence of a painful territory with that of a terminal branch nerve and the finding of hypersensitivity at the nerve's point of emergence should always raise the possibility of cranial neuropathy or neuralgia. In any case, if pain topography and examination are inconclusive, an anesthetic nerve block may be used for diagnostic purposes.
There are also certain similarities between NH and primary stabbing headache (PSH), as both are perceived at a specific point of the head (91,92). NH remains confined to the same area, and usually presents with chronic, continuous pain, whereas PSH is characterized by very brief paroxysms of pain (one or a few seconds), which are often multifocal. However, there are patients with PSH who have all their stabs in the same location and there are cases of NH in which the pain is not continuous and even presents with brief paroxysms. In these cases, the key to differentiating between them may be the patient’s delimitation of the circular area in which the pain occurs, as well as the demonstration of a certain degree of hyperesthesia or other sensory signs within the symptomatic area between the painful paroxysms; if these circumstances are present, the diagnosis should be NH, and not PSH.
There are other conditions that occasionally cause more or less circumscribed pain and should be included in the differential diagnosis of NH; these include temporal arteritis (93), myofascial syndrome (94), rhinosinusitis (95) and trochlear headaches (96). In cases of cutaneous involvement NH must be distinguished from dermatological diseases, such as alopecia areata (97) or nummular dermatitis (98), which do not usually present with pain. There is an entity known as trichodynia (99), defined by pain in the hair follicles or the scalp, which overlaps with NH; circumscribed forms of trichodynia could possibly be variants of NH.
Treatment
Some patients with NH do not require treatment and are content to learn of the benign nature of their condition. In patients in whom pain is severe or is a cause for concern, treatment should be prescribed (82). No controlled clinical trials have yet been conducted and there are no established protocols. Therefore, recommendations for the management of these patients are based on the clinical experience of several authors with single cases or case series.
Analgesics and non-steroidal anti-inflammatory drugs (NSAIDs), including indomethacin, have often been administered as symptomatic treatments during exacerbations of pain, or on a regular basis. The therapeutic response has been varied, with no response (1,9,12,13,21,23,25,57,58,61,62,64,65), partial response (6,7,11,22,56,58,62,64), or complete response (4,5,9,14,17,26,58,62,64). Analgesics and NSAIDs have proved beneficial in more than half of the cases in which they have been tried. These drugs can be considered for patients who have mild discomfort most of the time and only need analgesic therapy at certain times. They can also be added to preventive therapies for the acute treatment of pain crises.
Triptans were effective as acute treatment in one of the patients with migraine features; this patient also responded to migraine preventives, namely topiramate (49). In addition, oral triptans have been effective in the two patients with NH associated with superficial arterial aneurysms (48); in these latter cases, a vascular mechanism was suggested as a rationale for benefit.
Subcutaneous anesthetic injection has been performed in several patients (1,7,9,13,15,18,19,21,56,59,63,64). Nevertheless, this approach has rarely resulted in partial or transient improvement and has usually shown no effectiveness at all.
In patients with moderate to severe pain and in those who do not respond to symptomatic treatment, preventive treatment should be considered. In large hospital series, preventives have been necessary in most patients (58,59,62). Currently, the two preventive treatments with the highest level of evidence in NH are botulinum toxin injections and gabapentin, although no controlled clinical trials have been conducted so far.
There are several case descriptions and case series of NH patients successfully treated with botulinum toxin type A injections (13,19,23,39,64,100). The largest series included 53 patients with more than 10 days of pain per month, most of them with previous failures of one or more oral preventives. Patients received 25 U of onabotulinumtoxinA (OnabotA), which was distributed between one central and four peripheral points over the painful area. After injection, mean pain days, severe pain days and symptomatic medication days decreased; furthermore, in three quarters of the patients the reduction in pain days was greater than 50%, and in half of them it was at least 75%. A minority of patients had mild adverse effects (such as injection site pain, muscle pain or local bleeding), but overall tolerance was excellent (100). Treatment with OnabotA should be considered when there is a lack of response or tolerance to oral preventives; it can also be considered as a first option, given that no other treatment is supported by controlled clinical trials. Generically, it has been proposed to inject at five sites, at a rate of 2.5–5 U per site, four at the perimeter of the painful area and one in the center (Figures 3 and 4). However, the number of injection sites and the dose of OnabotA could be adjusted depending on the size and location of the symptomatic area (101).
Diagram of botulinum toxin infiltration procedure for nummular headache, with injection sites evenly distributed along the periphery and center of the painful area.
Botulinum toxin infiltration in a patient with nummular headache. The painful area was round, 3 cm in diameter, and located in the right parietal region. A total dose of 25 U of onabotulinumtoxinA was administered (5 points with 5 U each).
Among oral preventives, the most widely used drug has been gabapentin (6,8,11,14,21,23,32,35,42,46,50,54,55,58–60,62,64,65), with a response rate of over 60%; the response rate appears to be most favorable at medium doses, around 800 mg/day, while low doses may not be sufficient and high doses may not be well tolerated (58). Tricyclic antidepressants can also be used as an alternative to gabapentin, but results are less consistent; beneficial effects have occasionally been obtained with amitriptyline (15,21,58,59,62,64), nortriptyline (12,62) and clomipramine (17). Other drugs that have been reported as effective include pregabalin (20,51,58,62), lamotrigine (58), carbamazepine (15,64,65), oxcarbazepine (61,62), topiramate (23,49), duloxetine (23,62), metoprolol (57), cyclobenzaprine (25,62), neurotropin (31,38) and palmitoylethanolamide (52). A case has recently been reported that showed an absolute response to monthly subcutaneous injections of galcanezumab (66).
Less conventional treatments have occasionally been used. Two of the published cases benefited from neurostimulation procedures: one patient received transcutaneous electrical nerve stimulation (TENS) (18), and another received percutaneous electrical nerve stimulation (PENS) (63). One patient had pain relief after focal scalp excision of the affected area, but eventually had a recurrence (37). In a series of 49 patients, a beneficial effect on NH of surgical arterectomy under local anesthesia was reported; however, data on diagnosis and therapeutic response in these patients should be considered with some reservation, as most of them also underwent plastic surgery with a presumed effect on migraine (102). In cases of secondary NH, good results have been achieved by addressing the underlying cause. For example, the patient with NH associated with a meningioma went into complete remission after surgery (10); one of the patients with a superficial aneurysm (48) and the two patients with underlying Langerhans’ histiocytosis (54) also responded to surgical excision.
Epicrania fugax
Introduction
Epicrania fugax (EF) is a primary headache characterized by episodes of brief, paroxysmal pain that runs across the surface of the head, in a linear or zigzag trajectory, through territories of different nerves. In 2008, Pareja et al. first described EF as a paroxysmal headache that started at a certain point at the back of the head and rapidly radiated forward, with a linear or zigzag trajectory, reaching the ipsilateral forehead, eye, or nose within one or a few seconds (103). Two years later, an anteroposterior variant of EF was recognized, in which the pain moved in the opposite direction (104). In subsequent years, new cases and new series were published, which gave consistency to the syndrome and to the two variants described at that time, the posteroanterior and the anteroposterior (29,36,71–74).
In 2013, EF was included in the Appendix of the ICHD-3, in its beta version, with provisional diagnostic criteria. The dynamic topography of the pain was identified as a distinctive attribute of EF which could serve to differentiate this condition from other headaches and neuralgias (105). In the 2018 final version of the ICHD-3, EF has remained in the Appendix, with the same diagnostic criteria (Table 2) (3). However, in future editions it may eventually move into the main body of the classification with updated diagnostic criteria, as new cases of EF have been reported and its clinical spectrum has broadened.
Diagnostic criteria for epicrania fugax (ICHD-3).
ICHD-3: International Classification of Headache Disorders, third edition, 2018 (3).
Epidemiology
EF is viewed as a rare headache disorder, although its incidence and prevalence in the general population are unknown. To date, 82 cases with a typical picture of EF have been published (29,36,71–74,103,104,106–111). This figure excludes cases with atypical features and cases attributed to an underlying disease, as well as repetitions of individual patients in more than one publication. The cases described have been predominantly female (1.6:1), and the age of onset has ranged from 19 to 84 years, with a mean age of 45 years; no pediatric cases have been reported so far. The time elapsed from the onset of symptoms to diagnosis has also been highly variable, ranging from two days to 38 years.
Since this is a recently described syndrome, no population-based epidemiological data are available. We do have data from the registry of the Headache Unit of a Spanish university hospital, where 1.6% of the patients seen over seven years were diagnosed with EF (112). Although this is certainly not a high proportion, the number of patients with EF recorded in this register far exceeded those of other well-known headaches, such as short-lasting unilateral neuralgiform headache attacks with conjunctival injection and tearing (SUNCT) or with cranial autonomic symptoms (SUNA). In any case, it is likely that many people affected by EF do not consult or do not reach a headache specialist and are therefore undiagnosed.
Clinical features
Clinically, EF is characterized by paroxysms of dynamic pain. The patient feels the pain moving over the surface of the head, typically along one hemicranium, starting and ending at different dermatomes (3). Pain movement usually follows a linear trajectory, but some patients (<10%) depict it as a zigzag. The whole sequence is very fast, lasting between one and several seconds: generally, between one and 10 seconds (90%), and in nearly all cases between one and 15 seconds (>95%); longer durations of up to 50 seconds have been recorded in very few cases (104,106,111).
Depending on the direction of movement two main clinical forms have been distinguished: posteroanterior or forward EF, which is more frequent (two thirds of the cases described), and anteroposterior or backward EF (Figure 5). In the posteroanterior form the pain starts in a specific area of the occipital, parietal, or temporal region, and then radiates forward, reaching the ipsilateral forehead, temple, eye, or nose (29,36,71,72,74,103,107–111); in the anteroposterior form the pain follows a reverse path: it originates in the frontal region, eye, or nose, and goes backwards, towards the parietal or occipital region (36,73,74,104,106,110). Normally the attacks are strictly unilateral with no side shift. However, in about a quarter of patients, pain may occur on one side or the other and, in some rare cases, pain sometimes occurs on both sides of the head simultaneously (74,108).
The two main clinical forms of epicrania fugax. (a) Forward or posteroanterior and (b) Backward or anteroposterior.
EF pain takes an electric or stabbing quality in about 50 and 40% of cases, respectively. A minority of patients perceive it with other characteristics, e.g., as a lancinating, burning or oppressive pain. Pain intensity varies from patient to patient and, in the same patient, may vary from episode to episode. It is usually rated as moderate to severe. On a numerical scale from 0 to 10, patients have graded their pain from level 3 to the maximum level of 10 (113).
During painful paroxysms, 30% of patients may have cranial autonomic accompaniments ipsilateral to the pain (29,36,71–74,103,106,107,109,111). The most common autonomic symptom is lacrimation; other possible autonomic symptoms are conjunctival injection, a gritty sensation in the eye, eyelid edema, eyelid ptosis, and rhinorrhea. In an individual patient, these symptoms can occur in isolation or in combination.
EF attacks are mostly spontaneous, but more than 20% of patients identify triggers, either one or more than one (72–74,103,110,111). Triggers that have been recognized include tactile stimuli at the point of origin of the pain, neck movements, emotional stress, physical activity, Valsalva maneuvers, and eye movements.
Between the paroxysms, half of the patients experience tenderness or pain around the point of origin, either continuously or intermittently. This interictal pain is usually mild to moderate in intensity, and may have an oppressive, stabbing, burning, or throbbing quality. Some patients feel it over an area with poorly defined borders, but there are also patients who describe a well-circumscribed focal pain, which fits well with the diagnostic criteria for NH (29,36,71–74).
The frequency of attacks is extremely variable, ranging from less than one episode per year to multiple episodes per day (113); some patients exhibit a cluster of attacks on the same day, repeated at intervals of seconds or minutes (71,72,74,103,107,108). On the other hand, some patients keep a stable frequency of episodes, while others have an irregular frequency, with symptomatic periods alternating with phases of remission, or with exacerbations. In the long term, the temporal pattern is quite heterogeneous (113). In cases lasting more than three months, about half of the patients have painful paroxysms on most days, while the remaining patients have paroxysms on less than 15 days per month. Spontaneous remissions are possible: at least 12 of the published cases had spontaneous remissions in the first year after onset, and five of them had symptomatic periods of less than one month (71,72,74,103,104,111). There are also patients whose symptoms remit with drug therapy, and in some of them treatment can eventually be discontinued without relapse (74,103).
Atypical presentations
In recent years, some atypical forms of EF have been reported, which do not strictly conform to the current ICHD-3 diagnostic criteria. One of the reported cases had persistent linear pain throughout the pain course (Figure 6a). This patient presented with day-long episodes of continuous linear pain in her right hemicranium, together with superimposed paroxysms of EF, lasting seconds, and moving forward along the painful area (114). Another published case had a sagittal trajectory (Figure 6b). In this patient, the paroxysms of pain originated at a midpoint in the occipital region and radiated over the midline of the skull until they reached the interciliary space (115). Four cases have also been published in which the pain could travel in either direction, forwards or backwards (Figure 6c) (116). As another variant, three cases have been reported in which the pain moved along two parallel pathways simultaneously (Figure 6d) (117).
Some atypical forms of epicrania fugax. (a) Persistent linear pain between the paroxysms. (b) Sagittal trajectory. (c) Pain moving either forwards or backwards. (d) Pain moving along two parallel pathways simultaneously. (e) Coronal trajectory and (f) Pain moving in multiple possible directions.
Episodes of transverse or coronal irradiation, from one side of the head to the other, have been described in four patients (Figure 6e); one of them also had typical EF episodes with posteroanterior irradiation (118). One of the most striking published cases had paroxysms with zigzag trajectories in multiple directions (Figure 6f). Some of her episodes had a classic irradiation from the occipital to the frontal region, on either side of the head, while others followed coronal trajectories, in one direction or the other, or diagonal trajectories (119).
In addition, a facial variant of EF has been characterized in two series involving five and eight patients, respectively. These patients presented with paroxysms of pain that crossed several dermatomes of the face, either upwards or downwards (Figure 7), with or without extension into the cranial region (120,121). Moreover, two additional cases have been reported that combined episodes of facial and cranial EF (122); this finding supports the idea that both types of EF are just two distinct clinical presentations of the same disease.
Two forms of the facial variant of epicrania fugax. (a) Pain moving upwards and (b) Pain moving downwards.
Finally, among the variants described, five patients presented with EF-type pain that started in the head and radiated linearly to other regions of the body, reaching the neck, shoulder, or ipsilateral extremities. Interestingly, in three of these patients with extracephalic irradiation the pain originated in an area of NH (75).
Association with other headaches
As noted above, between the paroxysms of EF some patients present with a circumscribed headache, which remains restricted to a small coin-shaped area, and which according to ICHD-3 criteria would correspond to NH (29,36,71–76). In all published cases the area of NH was located at the point of origin of EF, so we could hypothesize that there may be a pathogenic connection between the two entities. It is not surprising that the Classification Committee decided to include NH and EF in the same chapter of the ICHD-3 (specifically in Chapter 4, “Other primary headaches”) and, furthermore, within the same category (the so-called epicranial headaches, headaches that are perceived on the surface of the skull) (3).
There may also be comorbidity between EF and other headache and facial pain conditions. Specifically, associations with migraine (71–75,103,106,108,110,115–117,119,120), tension-type headache (73,74,103,110,121), cluster headache (106), trigeminal neuralgia (74) and auriculotemporal neuralgia (111) have been reported. Episodes of other headaches do not usually coincide in time with EF, and patients tend to clearly differentiate between the two types of pain. However, in two of the cases described, paroxysms of EF were immediately followed by typical migraine episodes (106,108), and in one patient they were followed by cluster headache attacks (106).
Psychological factors
It is known that psychological factors may play a role in the maintenance or exacerbation of various headaches, but so far only two studies have analyzed the presence of psychological variables in patients with EF (110,123). One of these papers included controls without headache. When comparing the whole group of patients with the control group, EF patients showed higher levels of anxiety and depression. However, when two subgroups of patients were separated according to the frequency of episodes, only the low-frequency subgroup maintained the differences with respect to controls. Moreover, in this subgroup the differences extended to perceived stress and some of the variables that increase vulnerability to stress (123).
According to these data, patients with EF who have a lower frequency of headache episodes seem to have greater emotional disturbances. Therefore, it does not seem that these emotional disturbances are a direct consequence of the pain. Perhaps in patients with a low frequency of episodes, psychological factors may play a more relevant role as triggers or aggravators of EF. This is consistent with our own experience that some patients with infrequent episodes have symptom exacerbations coinciding with stressful life situations. In any case, these findings would need to be confirmed by further studies, and with a sufficient sample size.
Secondary forms
Theoretically, EF is a primary headache, so it cannot be attributed to any underlying disease. This is how it has been defined in the ICHD-3 (3), and this is the case in almost all patients. However, in a minority of patients presenting with an identical clinical picture there is an associated structural lesion. In such cases we may speak of “secondary” or “symptomatic EF”.
In one patient, backward-radiating EF-type pain started after radiation therapy for a skull base meningioma that contacted the trigeminal root. In this case the symptoms could have been due to the compressive effect of the tumor on the trigeminal nerve or, more likely, to the radiation (124). Additionally, vascular contact or compression of the trigeminal nerve has been demonstrated on the symptomatic side in some cases of facial EF (121). Yet, as neurovascular contacts can also occur in asymptomatic populations, their pathogenic role is uncertain.
A case of facial pain with the characteristics of EF followed Wallenberg syndrome caused by a lateral medullary infarction. Interestingly, the paroxysms of pain occurred on the ipsilateral side of the face (i.e., the side affected by the stroke) and could be triggered by painful stimuli applied to the contralateral upper limb (i.e., the upper limb affected by the stroke) (125). In another patient, the appearance of forward-radiating EF coincided with the appearance of an ipsilateral cerebellar abscess. The causal relationship was demonstrated by the disappearance of pain after treatment of the abscess with antibiotics and surgical drainage (126).
All the above cases have in common the presence of an intracranial lesion in the posterior fossa, but the cause could also be more superficial and lie in the epicranial tissues. A case has recently been described whose EF episodes originated from a keloid scar on the scalp resulting from excision of a subcutaneous hemangioma. In this patient, the keloid was associated with two types of pain: firstly, a persistent and circumscribed pain, with the characteristics of NH, and secondly, a paroxysmal pain radiating forward from the scar with the characteristics of EF (76).
Pathogenesis
The pathogenesis of EF is unknown, but some hypotheses can be put forward. Pathogenic mechanisms should explain the three phenomena that characterize EF episodes: first, the onset of pain; second, the spread of pain and, finally, the autonomic symptoms that accompany pain in a significant proportion of patients (113).
It was initially proposed that the EF pain generator might be epicranial, possibly affecting some of the nerve endings of the pericranial nerves (103). This hypothesis would justify the fact that the painful paroxysms start in a focal area. It would also be in line with the frequent coexistence of NH at this stemming point (29,36,71–76). Furthermore, this hypothesis would be supported by one of the reported cases of secondary EF, where the symptoms could be related to the development of a keloid scar after local scalp surgery (76).
However, there is evidence that the origin of EF may also be intracranial, involving nerve roots or directly affecting the central nervous system (CNS). As mentioned above, several cases have been described secondary to posterior fossa lesions, including a meningioma adjacent to the trigeminal root (124), a brainstem infarction (125), and a cerebellar abscess (126). The hypothesis of a possible central generator would be reinforced by the existence of EF cases in which pain can initiate at different points and follow different trajectories (75,116,118,119,122). Conceivably, EF is neither a purely peripheral nor a purely central disorder; the origin of pain could probably lie in both peripheral (extracranial or intracranial) and central generators.
Regardless of its origin, the spread of pain characteristic of EF should be explained by central mechanisms. In EF, pain irradiation follows a broad trajectory, crossing different dermatomes, which can hardly be explained by purely peripheral mechanisms. Sensory afferents from the occipital region are known to converge with trigeminal afferents at the highest levels of the spinal cord, in the so-called trigemino-cervical complex (127). This trigemino-cervical complex may well function as a central highway allowing the propagation of pain stimuli between the different cranial and facial territories. Additionally, in EF patients presenting with autonomic symptoms a trigemino-parasympathetic reflex must be activated, which necessarily passes through the CNS (128).
Diagnosis
EF was introduced in the Appendix of the ICHD-3 with provisional diagnostic criteria, based on the characteristics of the first patients (Table 2) (3). However, the phenotype of EF has expanded so that some of the reported cases do not fully meet those criteria. Today we propose new, more inclusive criteria that would fit almost all the cases described and that would remain specific. The proposed diagnostic criteria are shown in Table 3.
Proposed new diagnostic criteria for epicrania fugax.
ICHD: International Classification of Headache Disorders.
EF is, in principle, a primary headache, and its diagnosis requires the exclusion of any underlying disorder (76,124–126). Assessment of the patient should include careful inspection, palpation, and sensitivity testing of the area where the pain arises; this area may remain tender between attacks but should not show signs of another disease. Other cranial structures such as the superficial temporal arteries, pericranial nerves, trochlea and muscles should also be palpated. The physical examination should be completed with a thorough neurological examination for any pathological signs suggestive of secondary headache. Neuroimaging tests (MRI or CT) should be performed in all patients to rule out structural lesions. In addition, blood tests that include erythrocyte sedimentation rate and C-reactive protein should be requested in selected patients to exclude temporal arteritis and other systemic diseases.
Differential diagnosis must be made with other headaches and neuralgias with which EF could be confused, mainly NH, PSH, SUNCT and SUNA, trigeminal neuralgia and pericranial neuralgias or painful neuropathies. In NH the pain remains restricted to a small coin-shaped area of the skull and is usually continuous. NH pain does not move beyond the limits of the symptomatic area, even when exacerbations occur. PSH is characterized by a very brief stab, or series of stabs, at one or more points on the head. In any case, each stab is felt at a single point and the pain does not move in a straight line or a zigzag (91,92). SUNCT and SUNA present with attacks of pain in the trigeminal territory lasting seconds to minutes, which are always accompanied by prominent autonomic symptoms. Some patients with EF have autonomic symptoms and may meet criteria for SUNCT or SUNA. However, the pain described in SUNCT and SUNA has always been more diffuse and does not radiate in a linear or zigzag pattern (129,130). Trigeminal neuralgia manifests with paroxysms of severe, brief pain on one side of the face, commonly triggered by mechanical stimuli. Yet, trigeminal neuralgia primarily affects a single dermatome and, although it may eventually spread to other dermatomes, it does not have the vertical irradiation described in the facial forms of EF (131,132). Finally, in pericranial neuralgias or painful neuropathies, pain is limited to the territory innervated by the corresponding nerve branch (88–90).
Treatment
There are patients presenting with EF who do not need any treatment, either because they experience spontaneous remissions or because their attacks are infrequent. We should consider treating patients who have numerous attacks that do not remit (113). Among the published cases, more than half of those with cranial EF (>50%) and the vast majority of those with facial EF (>90%) have required some treatment. No clinical trials have been conducted so far and there are no established protocols. Therapeutic recommendations are based on the clinical experience of various authors with isolated cases or case series.
The drugs most used in EF are antiseizure medications; satisfactory therapeutic responses have been reported for several drugs in this group, especially gabapentin (30,71–74,109,116,120–122,124) and lamotrigine (71–75,116,120,121,124) but also pregabalin (74,121), carbamazepine (73,103,117,121), oxcarbazepine (120,121), eslicarbazepine (107,121,125), lacosamide (121), and levetiracetam (74). Amitriptyline (73,75,116,122), duloxetine (120), and indomethacin (74,75) have also been used with effect in a minority of patients. If the response to initial treatment is unsatisfactory, another treatment may be tried, as patients who do not respond to the first option may find relief with subsequent options. Likewise, if one drug produces partial relief, polytherapy may be considered.
Other therapeutic strategies that have been used in a limited number of patients with partial or complete response are anesthetic blocks of the greater occipital (72,74,103) supraorbital (74,103) or auriculotemporal nerves (111), steroid injection of the trochlear region (74), and botulinum toxin injection at the point of origin of the EF paroxysms (75,76,120,121). In the particular case secondary to a brain abscess, the pain disappeared with antibiotics and surgery (126).
In cases where NH and EF coexist, preventive treatment for both disorders may be necessary. In clinical practice, we have encountered patients in whom treatment controls the radiating pain of EF, while localized pain with features of NH persists at the point of origin of EF paroxysms. However, there are also patients in whom a single oral drug has been useful in controlling both types of pain (36). In addition, cases have been reported in which OnabotA injections at the NH site suppressed both local NH pain and radiating EF pain (75,76).
Article highlights
Nummular headache (NH) and epicrania fugax (EF) are not exceptional diagnoses in a headache clinic. NH is a well-established condition and is already included in the main body of the International Classification of Headache Disorders (ICHD). However, some issues about NH remain to be resolved, such as the lack of knowledge of its true incidence and prevalence and the need for well-designed randomized controlled trials. EF was described more recently and was added to the ICHD Appendix in its latest edition. It is to be expected that EF will eventually move to the main body of the ICHD with updated diagnostic criteria, since new cases are being reported and its clinical spectrum has broadened.
Footnotes
Acknowledgements
The author would like to thank Esperanza González-Perlado for drawing the figures.
Declaration of conflicting interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
