Characteristics of Headache Associated with Intractable Partial Epilepsy
Address correspondence to Dr. F. Andermann at Montreal Neurological Hospital and Institute, 3801 University Street, Montreal H3A 2B4, Quebec, Canada. E-mail: firstname.lastname@example.org
Summary: Purpose: The association between headache (HA) and epilepsy is well known. However, few previous studies characterized HA types and head sensations (HSens) in large populations of individuals with well-defined forms of epilepsy.
Methods: To analyze the incidence of HA in such a group, we compare HA and non-HA patients to identify special predictive factors for HAs or migraine. We also investigate the pathologically verified group for possible correlations with HAs or migraine. One hundred consecutive patients undergoing presurgical evaluation for pharmacologically intractable partial epilepsy were interviewed. For each HA type, we inquired about lateralization, localization, quality of HA, and results of treatment.
Results: Periictal HAs were reported by 47 patients. Of those, 11 had preictal HA (PIHA), and 44 had postictal HA (PostHA). Eight patients had both PIHA and PostHA. Interictal HAs (InterHAs) were reported by 31 patients. Twenty-nine (62%) of 47 patients had frontotemporal HAs. Twenty-five patients had migraine-like HA without aura: 18 (60%) of 30 patients with temporal lobe epilepsy (TLE) and seven (41%) of 17 with extratemporal epilepsy (ETE). No correlation between pathology and presence of HA was found in 59 pathologically verified patients, except in four who had arteriovenous malformations (AVMs): three had and one did not have HAs. Eighteen patients had, in addition, poorly localized and ill-described HSens other than HAs.
Conclusions: We confirm an association between focal epilepsy and HAs, including migraine without aura. This is true for both TLE and ETE. PIHA and even prodromal HA may be related to the epileptic discharge and may have lateralizing value. This association is not recognized by the current International Headache Society (IHS) classification. The presence of HA and migraine is not related to the underlying epileptogenic pathology except in patients with AVMs.
The association of headaches (HAs) with seizures is well known but poorly understood. It was classically described by Gowers (1). Periictal HAs were studied in patients with different seizure patterns and syndromes. They may be preictal (PIHA), ictal (IHA), or postictal (PostHA). All types may represent migraine (2–8), as defined by the diagnostic criteria of the International Classification of Headache Disorders (HIS) (9). PIHA has been previously reported (2,7,10), and in a recent series, its lateralizing value and its distinction from an epileptic aura (e.g., starting >5 min before seizure onset) was emphasized (Yankovsky et al., unpublished data). Lateralizing value and migrainous features of periictal HAs have been documented as well (5). Ictal HAs were described by Young and Blume (11) and documented by stereo-EEG (12,13). PostHA is a common symptom after generalized tonic–clonic seizures (GTCSs) (2,3) and also occurs after complex partial seizures originating in the temporal lobe (2–5). Patients are usually more concerned by their seizures, and HAs associated with epilepsy are often neglected, not only by the patient, but also by the physician. Few previous studies characterized HA types and HSens in large populations of individuals with well-defined forms of epilepsy (10). We are not aware of studies looking systematically for a correlation between pathologically verified partial epilepsy and the characteristics of associated HAs.
The aim of this study was to analyze the incidence of HA in such a group, compare HA and non-HA patients to identify special predictive factors for migraine or other HAs, and investigate the pathologically verified group for possible correlations with HAs or migraine.
We interviewed, according to a standard form, 100 consecutive patients (45 male subjects; mean age, 33 years) undergoing presurgical evaluation for pharmacologically intractable partial epilepsy (5). The interview began with a verbatim description of the HA by the patient. For each HA type, questions inquired about lateralization, localization, quality, and results of treatment. Interictal HA was defined as not temporally related to seizures. Severity was assessed by using the visual analogue scale (VAS) of pain (0–10; 0, no pain, and 10, worst possible pain). Each section contained an open-ended question, with the answer recorded verbatim, followed by the specific questions. Migrainous character of the HA was determined according to the diagnostic criteria of the IHS, except that in some cases, duration was <4 h (9). Family history of recurrent HAs or migraine was also documented.
Demographic and clinical data were obtained through interviews with patients and their relatives and by review of hospital charts. Diagnosis and lateralization of the seizure focus was based on a comprehensive evaluation including seizure history and semiology, neurologic examination, video-EEG telemetry with scalp electrodes, and neuropsychological evaluation in all. The seizure focus was determined by predominantly ipsilateral interictal epileptic abnormalities and by unequivocal seizure onset. In all patients, magnetic resonance (MR) images were acquired on a 1.5-T scanner with a T1-fast field echo, proton-density, T2-weighted, and fluid attenuated inversion recovery images.
Fifty-nine (59%) patients (31 females) reported recurrent HAs. Mean age was 31.8 years (range, 8–52 years). InterHAs were reported by 31 patients. Periictal HAs were reported by 47 patients. Of those, 11 had PIHA, and 44 had PostHA. Eight patients had both PIHA and PostHA. Twenty-nine (62%) of these 47 patients had frontotemporal HAs. Twenty-five had migraine-like HAs without aura, according to the diagnostic criteria of the IHS: 18 (60%) of 30 patients with TLE and seven (41%) of 17 with ETE, including four with frontal lobe epilepsy (FLE) and three with occipital lobe epilepsy (OLE). Twenty-nine patients reported a family history of recurrent HA or migraine (precise diagnosis was impossible because of the lack of direct confirmation). The characteristics of HA variables are presented in Table 1.
Table 1. Characteristics of PIHA, PostHA, and InterHA in 59 patients with intractable epilepsy and recurrent HAs
|Occurrence with seizures||75%||20 pts, almost 100%; 12, 75%; 7, 50%; 5, 25%||—|
|Side of HA||R, 6; L, 5||R, 12; L, 14; Bil, 18||R, 1; L, 2; Bil, 23; R or L, 13|
|Location||F-T in all; middle, 2||F-T, 32; middle, 16; posterior, 3||F-T, 26; middle, 9; posterior, 5|
|Spread||Ipsi, 2||Ipsi, 8||Ipsi, 8|
| ||Contra, 1||Contra, 4||Contra, 1|
|Intensity (VAS)||7.1 cm||6.7 cm||6.2 cm|
|Use of analgesics||4 (27%)||36 (81%)||25 (80%)|
|Analgesics||One, 2||One, 29||One, 21|
| ||Two, 4||Two, 4|
|Positive effect||Always, 2; sometimes, 2||Always, 11; sometimes, 18; never, 4||Always, 8; sometimes, 17; never, 2|
|Duration||EarlyPIHA||5–60 min, 13;||5–60 min, 6;|
| ||(5–30 min, 7)||60 min–12 h, 19;||60 min–12 h, 18;|
| ||ProdPIHA||longer or shorter, 12||longer or shorter, 7|
| ||(30 min–24 h, 4)|| |
| ||Daily, 2;|
|Frequency||Related to Sz||Related to Sz||<1/wk, 9;|
| ||1/wk–1/mo, 14;|
| ||1/mo–6/mo, 5|
Eleven (11%) patients had PIHA. Seven had early PIHA within 30 min before seizure onset, and four had HA 24 h to 30 min before seizure onset (i.e., prodromal PIHA). All but one had TLE. Four (36%) of the 11 patients with PIHA had a pattern compatible with migraine without aura. In two, the duration was <4 h. Eight patients had PIHA continuing to PostHA. Intensity of PIHA was moderate (7.1 cm). Four (36%) patients used over-the-counter analgesics: two always responded, and two others only at times.
Forty-four (44%) patients had PostHA. Duration was 1–12 h in 19 patients and from 5 to 60 min in another 13. Twelve patients had HAs lasting <5 min or >24 h. Maximum duration was 48 h. Intensity of PostHA was moderate (6.7 cm). Thirty-six (81%) patients used over-the-counter analgesics: 11 always responded, 18 sometimes, and four never benefited.
Thirty-one patients had InterHAs. Two reported migraine with aura unrelated to seizures. The other 29 had unclassifiable HAs according to IHS criteria. They were either bilateral (23 patients) or at times changing sides from attack to attack (13 patients). Location was frontotemporal in 28 patients. Intensity was moderate (6.2 cm). Twenty-five (80%) patients used over-the-counter analgesics: eight always improved, 17 sometimes, and two never showed an effect. Duration of InterHA was several hours in most patients.
Lateralization and location of HAs
In 29 (62%) of the 47 patients with periictal HAs, these were frontotemporal. They were ipsilateral to the epileptic focus in 27 (90%) of 30 patients with TLE and in only two (12%) of 17 with ETE. Lateralization of PIHA was ipsilateral to the epileptic focus in nine TLE patients, contralateral in one with TLE and in one with frontal seizure onset.
Twenty-five patients had periictal migraine-like HAs without aura, according to the diagnostic criteria of the IHS (9), except that in some cases, duration was <4 h. Eighteen patients had TLE, and seven had ETE, including four with FLE and three with OLE. Four (36%) of 11 patients with PIHA had migraine, and 21 (48%) of 44 with PostHA did. Another two patients reported migraine with aura only in the interictal period. None of the patients was treated by triptans or any other specific antimigraine medication.
Pathologically verified group
We found adequate pathologic specimens in 59 of 68 patients who were operated on and compared the HA with the non-HA groups (Table 2). Of 37 TLE patients, 36 had hippocampal sclerosis (HS), and one had a glioma. No correlation was found between HS and HA: 20 had HAs, and 16 did not. Of 22 ETE patients, eight had focal cortical dysplasia (FCD) (five with and three without HA), five had focal atrophy (three with and two without HA), four had AVMs (three with and one without HA), and one had a calcification and HA. One patient with an occipital glioma had HA. Three patients (one with and two without HA) had normal tissue. The analysis of 59 pathologically verified patients did not show any significant difference in HA occurrence in relation to the lobar epileptogenic area: 29 (63%) of 46 in TLE, six (66%) of nine in FLE, five (71%) of seven in OLE, and one (25%) in four in parietal lobe epilepsy (PLE).
Table 2. Comparison between HA and non-HA groups in 59 pathologically confirmed patients of 68 operated on
|Pathology (n = 59)|| (n = 34)|| (n = 25)|
|HS||20 (59%)||16 (64%)|
|FCD||5 (15%)|| 3 (12%)|
|Glioma|| 1 (3%) || 1 (4%)|
|AVM|| 3 (8.5%)|| 1 (4%)|
|Focal atrophy|| 3 (8.5%)|| 2 (8%)|
|Normal|| 1 (3%)|| 2 (8%)|
|Sz-onset lateralization (n = 68)||(n = 43)||(n = 25)|
|R||22 (51%)||13 (52%)|
|L||21 (49%)||12 (48%)|
Other head sensations
Eighteen patients had, in addition, poorly localized and ill-described HSens other than HAs: five as a prodrome, nine in early preictal, three in postictal, and four in the interictal phase. Most of the patients described their HSens as having a constant-pressure quality. Intensity according to the visual analogue scale was always <3. None used medications. Duration was short: seconds to minutes. Frequency was variable. No clear lateralization to the seizure focus was found in this group of patients.
In our study of 100 patients with medically intractable partial epilepsy, we found that 59 (59%) had recurrent HAs. Periictal HAs were reported by 47; either PIHA or PostHA. InterHAs were described by 31 patients. Most previous studies discussed PostHAs only, with a range of incidence from 37 to 51% (2,3,6). These findings are similar to our 44%. PostHA is the most frequent HA type associated with seizures. PIHAs have rarely been reported (2,5,7,8), and their characteristics have not been thoroughly studied. We found that 11 (11%) patients had PIHA. Little attention has been paid to PIHA lasting minutes to hours (2,10,14). Recent studies of cerebral blood flow (CBF) in patients with TLE detected an increase several minutes before seizure onset (15,16). This increase in CBF may be related to changes in neuronal activity long before the obvious EEG or clinical seizure onset and provoke HA through the trigemino-vascular system, causing the release of vasoactive sensory neuropeptides that increase the pain response (17). PIHA was a lateralizing preictal clinical symptom in patients with TLE, because it was ipsilateral to the seizure focus in nine of 10 TLE patients. One patient with TLE and one with ETE had contralateral HA. A larger number of patients must be studied prospectively to confirm the lateralizing value of PIHA. In eight patients, PIHA continued into the postictal period. This may point to a similarity in the pathogenesis of both forms.
The rate of unclassifiable HA in the InterHA group was high: 29 (94%) of 31 patients. In the majority, it was mostly of pressure or throbbing type (24 and 14 patients, respectively), either uni- or bilateral, and had no sufficient features to fulfill IHS criteria of migraine or tension HA. They were also not of any other primary HA type. Nevertheless, 80% of patients with unclassifiable HA used medications, and therefore these may be considered clinically significant.
Migrainous HAs occurred in 27% of our total population and in 46% of patients with recurrent HAs. Ito and associates (6) reported lower numbers of these symptoms (10% and 26%, respectively), considering only PostHAs in patients with partial epilepsy. Their patients were recruited from epilepsy clinics, and their epilepsy was probably less severe than that of our group of surgical candidates (6). It is not clear whether intractability itself is a factor leading to the migrainous quality of HA. Of our 11 patients with PIHA, four had migrainous HA. In the study of Ottman and Lipton (18), the frequency of migrainous HAs in epilepsy patients was similar to ours, and in the study of Leniger and colleagues (8), a frequency of 55.7% was reported.
None of our patients was treated with triptans or other antimigraine medications, but they used over-the-counter medications for any type of HA with variable effect. This is probably related to lack of awareness and of specific inquiry by the patients' physicians. It also may be related to the lack of a special seizure-related category of HA in the previous IHS classification. Two patients with postictal migraine were reported to benefit from oral sumatriptan, 100 mg (19). Both had an excellent response; within 30 min in one patient and 2–3 h in another. No systematic study has been conducted in patients with periictal migraine. Such a study is warranted because this condition is frequent and at times might be more disabling than the seizures themselves. Prophylactic treatment with sodium valproate (20–22) and topiramate (23,24) for migraine is well established. These and other AEDs may help both migraine and the seizures.
The frequency of occurrence of HAs with seizures was typically high and predictable in both PIHA (75% of attacks) and PostHA, where it occurred in 75–100% of attacks in 32 of 44 patients. Most HAs were frontotemporal. The consistent lateralization in individual patients was striking. Only 18 of 44 who had PostHA had bilateral HAs; the others described them as strictly unilateral. It seems that seizure foci tended to induce consistently lateralized HAs. These were ipsilateral in 90% of TLE patients and in only 12% of ETE patients. Studying the exposed human cerebral cortex during epileptic seizures, Penfield and Jasper (25) observed widespread vasodilation and reactive hyperemia in the region of the discharging epileptic focus. Studies of CBF in patients with TLE detected an increase several minutes before seizure onset (15,16) and during the seizure (26). The ipsilaterality of the HA may be explained by activation of the trigemino-vascular system and neurogenic inflammation at the side of the seizure focus in susceptible individuals.
The intensity of PostHA was comparable with that of PIHA (6.7 and 7.1 cm), but more patients used analgesics postictally (81% vs. 27% of PIHA patients). These medications were over-the-counter and had not been specifically prescribed. Antimigraine preparations were not prescribed for patients with migrainous HAs. This is in agreement with the finding of Forderreuther et al. (10), who reported a similar average intensity of the HAs. Thirty percent of their patients self-administered over-the-counter analgesics on a regular basis. Another 66% of their patients reported taking medications occasionally. The higher number of patients taking analgesics for PostHA compared with PIHA probably reflects the longer duration of PostHAs and therefore their more disabling nature.
We compared pathologic findings in HA and nonHA groups. Hippocampal sclerosis, focal cortical dysplasia, glioma, AVMs, focal atrophy, and calcification were found in similarly decreasing frequency in both groups without any significant differences in prevalence. It seems that pathology does not have an effect on HA occurrence, except for AVMs. Three of our four patients with an AVM had HAs, and only one did not. Although the number of patients with AVMs is small, the association between HAs and AVM is well known, and HA with or without migrainous features is a frequent presenting symptom of AVMs.
We looked at epilepsy syndromes in HA and nonHA groups (Table 3). Although 37 (62%) of 60 patients with TLE had HA, 23 (38%) of 60 did not. It appears that in patients with TLE, both PIHA and PostHA may have lateralizing value. Six of eight patients with OLE had HA, and only two did not. This is in agreement with previous studies emphasizing the association between OLE and HA or migraine (27). Interestingly, three of our six OLE patients with HA had migraine. Furthermore, of seven patients with migraine and ETE, three had OLE, which emphasizes the strong association of OLE and migraine.
Table 3. Correlation of HA and epileptic localization
|TLE||37 (63%)||23 (56%)|
|FCLE||14 (24%)||12 (29%)|
|OLE|| 6 (10%)||2 (5%)|
|PLE||2 (3%)|| 4 (10%)|
Other than pain, HSens did not have any specific features, localizing or lateralizing value, and were of low intensity. They occurred in all periictal phases, and their nature remains uncertain. This group probably included patients with a cephalic aura.
In conclusion, this study confirms an association between focal epilepsy and HA, including migraine without aura. This is apparent not only with occipital epilepsy, which is well known, but also with TLE and other forms of ETE. HA is probably underrecognized and underreported, and insufficient attention is paid to it by physicians. It adds significantly to the burden of epilepsy. PIHA and even prodromal HA may be related to the epileptic discharge and may provide valuable, although not absolute, lateralizing information in TLE. The mechanism of activation of migraine without aura by seizures, mostly temporal, remains uncertain.