Abstract
This paper will discuss evidence which supports a link between ‘side locked’ migraine (SLM) and the trigeminal autonomic cephalgias (TACs). Recent papers brought strictly unilateral primary headaches into focus, proposing new classification and discussing pathophysiological mechanisms. We reviewed those proposals and present evidence that SLM falls in between the well-defined TACs and side shifting migraine (SSM). It is difficult to differentiate SLM from the recently proposed headache subtype called hemicrania generis incerti (i.e. hemicrania continua unresponsive to indomethacin). We also present cases that may exemplify the considerations made in the paper.
Keywords
Introduction
Usually ‘side locked’ headaches, i.e. strictly unilateral headaches without shifting sides, are easy to diagnose when they are brief and have the typical association of autonomic features. However, the diagnosis becomes difficult when a patient presents side locked headache that shares typical features of more than one primary headache disorder or lacks typical features.
For a long time associations of more than one primary headache syndrome have been recognized, and are referred to as cluster migraine, cluster tic, cluster headache–chronic paroxysmal hemicrania, and chronic paroxysmal hemicrania–tic (1, 2). In the same way the presentation of a primary headache disorder plus one feature of a second primary headache disorder has been described such as seasonal migraine, migraine with autonomic symptoms, cluster headache with visual aura or hemiplegic symptoms, hemicrania continua (HC) with photo- and phonophobia, and HC with aura are some of the examples (3, 4).
The current International Headache Society (IHS) classification system divides the primary headaches into four distinct groups. Group 1 migraine headache, group 2 tension-type headache, group 3 cluster headache and chronic paroxysmal hemicrania, and group 4 miscellaneous headaches, unassociated with structural lesion, such as idiopathic stabbing headache and benign cough headache (5).
A number of papers proposing diagnostic criteria and addressing the inclusion of HC in the IHS classification system have been published (6, 7), arguing convincingly that HC shares enough clinical and pathophysiological features to be placed together with cluster headache and chronic paroxysmal hemicrania. Pareja and colleagues in recent papers (7, 8) proposed the use of the term hemicrania generis incerti (HGI) to classify the HC patients unresponsive to indomethacin. Peres and colleagues (9) recently presented a series with 34 new HC patients, proposing that HC is not that rare, and that the condition has been under-diagnosed since its description.
Among the possible differential diagnoses for HC (both hemicrania continua vera and hemicrania ‘generis incerti’), we find the ‘side locked’ (i.e. strictly unilateral without shifting sides) migraine to be the most important. We consider this topic deserves more discussion.
Considering the fact that primary headache disorders with overlapping clinical features are still poorly studied, we will address, through the report of four recent cases, the evidence of the link between side locked migraine (SLM) and the trigeminal autonomic cephalgias (TACs), and the importance of that link in considering the diagnosis of some primary headache disorders with clinical overlap. Considerations concerning the existence of the HGI and the difficult to differentiate SLM from the TACs will also be addressed.
Case reports
Case 1
A 36-year-old Caucasian male patient with strictly unilateral, right-sided headache, occurring frequently in the last 9 months. The pain is described as throbbing, of sudden onset, located in the retroocular region, with two distinct patterns. It is preceded most of the time by a visual aura, right visual hemianopia, lasting between 20 and 30 min Then, after 10 min the throbbing pain starts, associated with intense photophobia, phonophobia, nausea, and vomiting every time, lasting between 24 h and 48 h. He also describes at least two exacerbation periods, every time he has a headache attack, lasting between 3 h and 4 h, with non-intense autonomic features like eyelid drooping, redness in the eye, and nasal congestion, ipsilateral to the pain. These episodes recur at least four times a month. The pain can be triggered by alcohol intake.
He had tried triptans, including subcutaneous sumatriptan, with poor response that was also associated with bad chest symptoms. Because he has a personal and familial history of high blood pressure and Marfan syndrome, he was advised to avoid this class of medication. Prophylactic treatments with propranolol 80 mg a day for 2 months, and verapamil 360 mg/day for 6 weeks were not helpful. He had a normal computed tomography (CT) and magnetic resonance imaging (MRI) of the brain, and normal EEG. During our evaluation he experienced the onset of the pain, and achieved complete relief with oxygen therapy. He was diagnosed as having migraine with aura and autonomic features.
Case 2
A 34-year-old Caucasian male patient with strictly unilateral, right-sided headache that started 10 years ago. The pain is described as throbbing most of the time, and pressure-like a few times, located in the retroocular region, that usually spreads to temporal and occipital regions 2–3 h after the start of the pain. The total duration is about 12 h. He describes exacerbation periods, lasting about 3–6 h, associated with intense phonophobia, mental confusion, and few episodes of expression aphasia. He also experienced some déjà vu episodes, of short duration, a few times. Autonomic features are modest, consisting of redness of the eye, and curiously, an increase in the palpebral fissure ipsilateral to the pain. During the exacerbation period he mentions that there is a change in the kind of pain, but is not able to describe it adequately.
He is able to identify some triggers to the headache such as prolonged fasting, alcohol consumption, and sleep deprivation. Initially these episodes were infrequent, but became daily in the last 7 years. Therapeutic response to triptans was poor. Inhalation of oxygen did not help. Prophylactic medications divalproate sodium 1000 mg/day for 6 months, verapamil 360 mg/day for 2 months, and nadolol 40 mg/day for 4 months were tried unsuccessfully. Finally, he exhibited partial but consistent relief with indomethacin 100 mg a day, and achieved complete relief with 150 mg a day. Familial history of migraine was positive. He was investigated with normal CT, MRI, and MRA of the brain, and normal EEG.
Case 3
A 41-year-old Caucasian female patient with headaches for the past 11 years. She describes her headaches as throbbing and stabbing type, mostly in the anterior region, always at the left side. Pain in the eyeball and behind the eye is continuous. It is a mild pain most of the time, 2–3 on a 1–10 scale. Associated symptoms are photophobia, phonophobia, and aggravation with physical exercise. She also presented one to two periods per day of worsening, when the pain intensity rises to 8 or 10, associated with redness in the left eye, tearing, and discrete eyelid droop. These episodes last between 15 min and 2 h, most often 1 h. She has a medical history of Type 2 diabetes which started 1 year ago. Her father had headaches for many years before he died of a heart attack. Previous treatment trials with triptans did not help her. Indomethacin 150 mg/day for 3 weeks was ineffective. Prophylactic treatments with topiramate 150 mg/day for 6 months, and propranolol 120 mg/day for 3 months were unhelpful. She tried oxygen therapy with a good consistent response during the exacerbation periods. Brain MRI and MRA were normal, as well as EEG.
Case 4
A 27-year-old Hispanic female patient. She began presenting headaches when she was 14, always at the right side. Usually her headaches were dull or throbbing, moderate in intensity, grade 5 on a 1–10 scale. Photophobia, phonophobia, and aggravation by physical exercise were always present. Lasting between 24 h and 36 h, she commonly had one exacerbation period, where the pain became throbbing or stabbing, with an intensity of 10, associated with ptosis, redness of the eye, tearing, and nasal congestion.
She reported five episodes where she developed a sensation of numbness that started in the left face, spreading to the neck region and left arm, associated with the more intense exacerbation period, that lasted about 1 h. There is no familial history of hemiplegic migraine, but a consistent history of migraines; one of her sisters experienced migraine with autonomic features also. She had very good response to injectable and oral triptans, a partial response to indomethacin 150 mg day for 2 weeks, and no response to oxygen. Propranolol 80 mg/day for 6 weeks proved not to be helpful. Brain MRI and MRA were normal.
All the reported cases were admitted to receive in-patient treatment with IV Di-Hydro-Ergotamine with good, consistent response. Salient features of the four cases are summarized in Table 1.
Clinical features of side locked headaches
Headache type: throb, throbbing; stab, stabbing; pres, pressure. Localization: O, ocular; RO, retrocular; F, frontal; T, temporal. Treatment: –, no response; +, positive; ++, strong positive response; Identified triggers: A, alcohol; Fa, fasting; L, light; MSG, monosodium glutamate; SD, sleep deprivation.
Discussion
Patients presented in this report manifest as ‘side locked’ headache with migraine features, associated with an exacerbation pattern commonly described in HC patients, with few autonomic features, some with typical aura symptoms.
The diagnosis in case 1 was migraine with aura and autonomic features. The clear temporal pattern of migraine with identifiable trigger factors, associated symptoms and visual aura favours the above diagnosis. Atypical features were the presence of autonomic symptoms, and the responsiveness to oxygen inhalation in the exacerbation periods. Points against the diagnosis of cluster headache in this case were the less intense autonomic features, absence of periodicity, the temporal pattern, and the kind of pain; HC was ruled out because of the lack of responsiveness to indomethacin, the presence of trigger factors, and the occurrence of a non-continuous headache.
Case 2 was diagnosed as having migraine with aura responsive to indomethacin. Favouring the migraine diagnosis were the temporal pattern of duration, the throbbing pain, the moderate to severe intensity of baseline pain, the presence of trigger factors. Atypical features were the presence of autonomic symptoms. Presence of trigger factors, baseline severe pain, and aphasic aura diminish the possibility of HC diagnosis (8). Cluster features were not present in this patient. Interestingly, this patient, who had a good response to indomethacin, reported clearly identifiable precipitating factors also described by other patients presented here, considered an uncommon feature in hemicrania continua patients.
Case 3 was diagnosed as migraine without aura with autonomic symptoms. The presence of a low-grade daily headache with associated features such as photophobia and phonophobia, and the presence of identifiable trigger factors are more common to a migraine diagnosis. Atypical were the presence of autonomic features and a consistent response to oxygen inhalation during the exacerbation periods. Lesser intensity of the autonomic symptoms, and occurrence of a daily continuous baseline headache went against the cluster diagnosis. Lack of responsiveness to indomethacin, presence of trigger factors, restlessness diminished the possibility of diagnosis of HC.
Case 4 was diagnosed as migraine with aura with autonomic symptoms. The initial presentation of migraine type headache, the temporal pattern, the presence of common migraine associated features, the presence of paraesthesic symptoms identified as aura manifestation, and a good response to triptans favour the diagnosis of migraine. Atypical was the presence of autonomic features. Against the cluster headache diagnosis were the less intense autonomic features, the lack of responsiveness to oxygen inhalation therapy, the duration of exacerbation periods. Although she presented a partial relief of the symptoms with indomethacin in a good therapeutic dosage, 150 mg/day for 1 week, the presence of trigger factors, the responsiveness to triptan, the associated migraine features out of the exacerbation periods diminished the possibility of HC diagnosis.
Additionally, all of our patients presented migrainous features and had identifiable trigger factors, a feature necessarily absent in HC patients according to all existent diagnostic criteria.
In all four patients investigations ruled out structural causes of ‘side locked’ headaches such as intracranial arterio-venous malformation, nasal septum deviation with contact points or sinusitis.
Evidence of the association between SLM and TACs
Goadsby and Lipton (6) proposed that the trigeminal-autonomic reflex could account for many of the features of the TACs, as well as the painful exacerbations of HC. This reflex pathway consists of a brainstem connection between the trigeminal nerve and facial parasympathetic outflow. Stimulation of the trigeminal ganglion leads to local release of trigeminal (CGRP and substance P) (10, 11) and parasympathetic (VIP) marker peptides in cats and humans. For both the extracerebral and the cerebral components, the major part of the vasodilator response is mediated via a reflex connection with the facial nerve. The trigeminal afferent arm of this reflex is clearly active in migraine while the parasympathetic arm is active in cluster headache (12).
Perhaps SLM patients represent a group of migraineurs with an overactive trigeminal parasympathetic reflex, accounting for the parasympathetic autonomic symptoms, even though ptosis is difficult to explain.
Unilaterality is an expected feature in the TACs, being part of the classification criteria for the TACs in the current IHS system, as well as in the various proposed diagnostic criteria for those syndromes not currently included in the IHS system, such as hemicrania continua, episodic paroxysmal hemicrania, and SUNCT syndrome (5–7, 13).
The strict unilaterality in migraine was described in the patient series of Leone and colleagues (14). They showed occurrence of a strictly unilateral pattern in 15.1% of the migraine without aura patients and in 16.9% of the migraine with aura patients. Only 4.1% of the tension-type headache patients presented the strictly unilateral pattern. Greater occurrence was observed in the atypical facial pain group with 31.6% of the total, and also in the non-classifiable headache group, 26.7%. At that time the authors suggested that this non-classifiable headache group must have been experiencing a distinct form of migraine, which does not completely fit the IHS criteria for migraine.
Unilateral cranial symptoms seem to be associated with strictly unilateral migraine as shown by Barbanti and colleagues in a prospective study (4). About 45% of their migraine patient sample, n = 177, presented at least one autonomic symptom, and stronger association was observed to those patients with unilateral and severe migraine attacks. These findings are important to corroborate the idea of the under-recognition of cranial autonomic symptoms in migraine patients. They conclude that headache that occurs in migraine with autonomic symptoms differs from typical migraine, being more severe and strictly unilateral. Because in most the sample was composed of episodic migraine patients they found it easy to differentiate those migraine patients from cluster headache. However, when facing chronic migraine with autonomic features one can find it difficult to differentiate, for example, from HC patients.
Pareja and colleagues state that a strictly unilateral headache syndrome means that only one – and always the same – side is symptomatic, and that it differs completely from side shifting headaches (8).
Wheeler and colleagues call attention to the fact that the exacerbations often observed in the HC patients met the IHS criteria for migraine in 60.2% of a review series. Moreover, 74.5% of the entire group had episodic migraine or migrainous disorder before HC ever developed. They concluded that severe headache attacks, e.g. exacerbations, in HC patients meet migraine diagnostic criteria and are easily confused with episodic migraine when daily background headache is not identified (15). In four indomethacin-resistant cases reported there were migrainous features present in at least three (16).
In the report of Peres et al. (9) 34 new patients with HC diagnosis were described, and a detailed analysis of associated features was made. They showed that migrainous features such as nausea, vomiting, photophobia or phonophobia were common during the exacerbation periods (70.6% of all patients), and that they met IHS criteria for migraine during the exacerbation period. Compared with Wheeler's series, none of Peres’ patients met the IHS criteria for migraine, except during the exacerbation period or prior to developing HC. No mention of occurrence of identifiable trigger factors was made in those review series.
Evers and colleagues described an interesting case with a consistent history of coexistent familial hemiplegic migraine (FHM) and HC. They emphasized some aspects such as unilaterality of both kinds of pain, the increase of autonomic symptoms, and the increase of the pain of HC during the migraine attack. They concluded that the close relationship of symptoms in that patient, and experimental findings on the shared pathophysiological mechanisms of migraine and TAC syndromes, suggest that the coexistence of HC and FHM may be more than just chance, although sumatriptan did abort the migraine pain but not the pain of HC in that particular case (17).
More evidence linking migraine and the TACs are clinical reports of the presence of typical visual aura in HC patients (18), and the development of a similar pattern of HC pain in a migraine patient after analgesic overuse (19). The apparent overlap that occurs between migraine and HC could explain the good indomethacin responsiveness of some HC patients, unresponsiveness to indomethacin in others, and the sometimes beneficial effects of drugs used in migraine treatment for HC patients (20).
It is difficult to explain what determines the presence of strictly unilateral headache in these migraine patients, similar to the great majority of TAC patients. In the past some authors (21, 22) tried to correlate the localization of the headache with the presence of unilateral mechanical or transient trigger conditions such as sinus disease and nasal septum deviation. However, no clinical studies of natural disease processes which confirm Wolff's experimental observations can be found, and recent clinical studies (23, 24) suggest there is limited value in attaching too much significance to headache localization resulting from sinus diseases or alterations. Moreover, no history of head trauma was reported by our patient series. Normal brain and sinus CT scan ruled out the possibility of sinus disease as well as nasal septum deviation in our sample.
The above cases are examples which highlight the difficulty in differentiating SLM from HC unresponsive to indomethacin, and some other primary headache disorders with overlapping clinical features. At this time there is no evidence of the presence of a biological marker that allows us to differentiate the primary headache disorders. Until there is, tentative classification based on clinical aspects should be done. More than that, we must search for conclusive answers that can explain the variability of clinical presentations in migraine patients, and consider the use of indomethacin when facing a side locked headache patient with migrainous features.
We suggest that the proposal for using HGI to classify HC (7, 8) patients unresponsive to indomethacin should be reviewed. It is not clear at this time if the unresponsiveness to indomethacin is a feature strong enough to be considered an exclusion factor from diagnostic criteria for HC. Bigal and colleagues (25) comparing the Goadsby and Lipton with Pareja classification system showed that the former is more clinically useful, considering that a variety of medications have been reported to be effective in patients with HC. While commenting on the diagnostic criteria proposed by Pareja and colleagues, Dodick (26) reminds us that the strict responsiveness to indomethacin may exclude some HC typical cases, while other cases may eventually respond to other medications including NSAIDs. Moreover, there is no specific or exclusive responsiveness to indomethacin in HC patients. Otherwise, secondary cases of HC have been described that responded initially to indomethacin.
We propose, based on our small sample and on a recent literature report (4), that both episodic (case 1) and chronic (cases 2, 3 and 4) forms of SLM can be considered together as a particular group of migraine patients, since they can represent a different clinical manifestation than that of typical migraine patients. Li and Rozen (27) recently described the clinical characteristics of 56 patients diagnosed as new daily persistent headache (NDPH). Their findings showed that most of their sample presented bilateral headache, and no reference to strictly unilateral headache was made. Even though 23% of the patients reported presence of lacrimation, no clear description of autonomic symptoms is found in that report. Maybe NDPH, a headache disorder clinically similar to chronic migraine, can manifest as a side locked headache as well as SLM, but all of our cases presented here evolved from a typical episodic pattern of headache occurrence.
For this sample of SLM patients, it seems that the presence of trigger factors, the presence of migraine-associated features out of the exacerbation periods (photophobia, phonophobia, nausea, and vomiting), and aura symptoms, present in all of our cases, were more important corroborators in considering the differential diagnosis between SLM and HC than responsiveness to indomethacin, observed in 2/4 of patients. We understand that the presence of these migrainous features favoured the diagnosis of SLM, while indomethacin responsiveness was observed in half of our patients.
Probably the excessive activity of the trigeminal-autonomic reflex in this subset of migraine patients (SLM) and the consequent clinical similarities observed between them and the TACs, should be considered in the future organization of the nosologic structure of the primary headaches.
Conclusion
In light of recent discussions about the diagnosis and treatment of trigeminal autonomic cephalgias, the increasing reports of primary headache disorders with overlapping clinical features, and recent description of autonomic features in migraine patients, we believe that ‘side locked migraine’ falls in between two well-defined primary conditions, the TACs and side shifting migraine (SSM).
Future systematic observations using a larger series of patients should be done to allow us determine if the incidence of autonomic symptoms is higher in the SLM when compared with SSM, and also to determine if the responsiveness or lack of responsiveness to indomethacin in unilateral headache may be crucial in differentiating between SLM and HC patients.
Footnotes
Acknowledgement
This paper was the result of a clinical fellowship programme undertaken by the first author at the Houston Headache Clinic, Houston, TX. He was supported financially by Conselho Nacional de Desenvolvimento Gentítiro e Tecnológico – CNPq−, a Brazilian agency for scientific and technological development.