Background
Epilepsy and primary headache disorders affect individuals of all ages worldwide. Several studies have been performed to attest if there is a relationship between the two conditions, in order to verify the existence of a causal association or if the two disorders can occur in the same individual by chance. In the last century Gowers first advanced the clinical hypothesis of a relationship between epilepsy and migraine [
1] since the two conditions show a well-recognized clinical, pathophysiological and therapeutic overlap [
2]. Studies on the association between epilepsy and other types of primary headache are difficult to perform as tension-type headache (TTH) is extremely common in the general population [
3] whereas cluster headache is very rare [
4]. For this reason most studies analyzing the prevalence of headache in patients with epilepsy focused only on migraine and results remain controversial (Table
1) [
5‐
21].
Table 1
Literature prevalence of migraine in patients with epilepsy
| 100 | 39/61 | 32 mean | Interview | 9 % |
Ottman and Lipton 1994 [ 6] | 1948 | 40 %/60 % | ≥18 | Structured telephone interviews + medical records review for 60 % of probands | 24 % |
Ito and Shachter 1996 [ 7] | 162 | 82/80 | 19-65 range | Questionnaires mailed to the subjects + medical records review | NAa
|
| 109 | 36/73 | 38 ± 12 meanb
| Questionnaire + interview + medical records review | 12.8 % |
Velioglu and Ozmenoglu 1999[ 9] | 412 | 212/200 | 15-70 range | Interview with a standardized questionnaire | 14 % |
| 341 | 154/187 | 40 ± 15 mean | Interview with a standardized questionnaire | 18.2 % |
Karaali Savrun et al. 2002 [ 11] | 135 | 80/55 | ≥10 | Questionnaire administered to patients | 14.8 % |
Förderreuther et al. 2002 [ 12] | 110 | 69/41 | 35.2 mean | Semi-standardized interview | 10 %b
|
| 364 | 163/201 | 12-81 range | Structured interview with standardized questionnaire | 8 % |
Syvertsen et al. 2007 [ 14] | 109 | 44/65 | 20-71 range | Questionnaire + semi-structured telephonic interview | 20 % |
| 227 | 98/129 | 36.0 ± 11.3 mean | Interview with standardized questionnaire + seizures and headache diary over the 3-month observation period + final interview | 6.6 %b
|
HELP Study Group 2010 [ 16] | 597 | 348/249 | ≥13 | Questionnaire at initial visit | 12.4 % |
| 492c
| 154/338 | ≥18 | Direct interview with questionnaire | 18.3 %b
|
Duchaczek et al. 2012 [ 18] | 201 | 106/95 | ≥18 | Semi-structured interview | 11 % |
| 730d
| 285/445 | ≥12 | Telephone or in-person interview + medical record abstraction | 25.2 %e
|
Gameleira et al. 2013 [ 20] | 304 | 141/163 | 4-88 range | Patients evaluated at the epilepsy clinic | 32.9 %f
|
| 1109 | 607/502 | ≥18 | Self-administered questionnaire + standardized semi-structured telephone interview | 12.53 % |
According to its temporal relationship with epileptic seizures, headache can be classified as inter-ictal (inter-IH) or peri-ictal (peri-IH). Inter-IH is not temporally related to seizures, whereas peri-IH manifests in their time frame (pre-ictally, ictally, post-ictally) [
7,
8]. Literature data on the relation between inter-IH (in particular migraine) and peri-IH are controversial owing to the methodological heterogeneity of previous studies [
5‐
21]. Moreover, in the context of seizure-related headaches, entities identified as “migralepsy” or “epileptic headache” are still matter of discussion [
22].
The aim of this study was to estimate the prevalence of headache in adult patients with epilepsy, describing its clinical features and temporal relationship with seizure occurrence.
Methods
The institutional review board of the IRCCS Institute of Neurological Sciences of Bologna approved the project. Clinical investigations have been carried out in accordance with the Helsinki Declaration adopted by the 18th World Medical Assembly in Helsinki, in 1964, as last amended by the World Medical Assembly.
Study design and participants
This is a cross-sectional study conducted at the outpatient clinic of our tertiary Epilepsy Center between March and May 2011 and March and July 2012. Patients aged ≥ 17 years were consecutively asked to participate in the study and a self-report form was administered to those who accepted. This form dichotomously ruled out patients who reported a lifetime presence of headache and patients who had never suffered from headache. If patients confirmed the occurrence of headache, trained physicians (GM, CL, LF), blinded to the patient’s diagnosis, conducted a semi-structured interview characterizing the type of inter-IH and peri-IH, if present. Headache data were revised by headache experts (SC, GG, PC), who validated the diagnosis according to ICHD-II criteria [
23]. Expert epileptologists (FB, PT, PA, MS) classified epileptic seizures and syndromes according to the 2010 International League Against Epilepsy (ILAE) Commission report [
24].
All patients with a diagnosis of epilepsy were included in the study. We excluded patients who had arrived for a first visit and proved not to be affected by epilepsy (i.e. psychogenic non-epileptic seizures, sleep disorders, syncope, dystonia), patients with a doubtful epilepsy diagnosis, patients who had only a single seizure, and patients with a severe mental retardation.
Data collection
The self-report form administered contained socio-demographic data and a preliminary question regarding the lifetime presence of headache. Patients who answered affirmatively also had to indicate their age at headache onset, headache in the previous three months, the frequency of attacks and the use of analgesics in a month. On the same form, patients who reported headache also answered the already validated ID migraine, a three-item instrument for migraine screening, and then underwent an ad hoc semi-structured interview with trained physicians concerning inter-IH and peri-IH. The formulation of this semi-structured interview was the product of a collaboration between the Epilepsy and Headache Centers. The clinical data collected to evaluate inter-IH concerned headache in other family members, age at headache onset, lateralization of headache, quality of pain, duration, intensity, frequency in a month, worsening with physical activity, use of analgesic treatment, associated symptoms like photophobia, phonophobia, nausea, vomiting, and presence of aura. For patients who presented peri-IH, the interview also included questions on timing of onset (pre-ictal, ictal or post-ictal headaches), duration, and features of peri-IH. Finally expert epileptologists filled out the last section of the form reviewing patients’ clinical records and collecting data on epilepsy syndrome, seizure types, frequency of seizures, epilepsy etiology, age at epilepsy onset, disease duration, current antiepileptic medications, and photosensitivity. The interview had previously been tested on a series of 50 consecutively recruited patients with epilepsy, showing that the diagnosis gathered from the semi-structured interview correlated with that of the headache experts.
Definitions and classifications
According to the temporal relationship with seizures, peri-IH was divided into pre-ictal headache (pre-IH), ictal and post-ictal headache (post-IH). Pre-IH was defined as appearing within 24 h before the seizure [
18,
22,
25]. Ictal headache was present exclusively during the seizure [
22]. Post-IH was defined according to the ICHD-II as a “headache which develops within 3 h following a partial or generalized seizure and resolves within 72 h after the seizure” [
23]. We collected data on the lifelong presence of headache and verified if headache attacks had occurred in the three months prior to the interview. We defined as “inter-ictal headache” all headaches that manifested within a time period of the epileptic disease and whose attacks were not temporally related to an epileptic seizure. According to ICHD-II criteria [
23], inter-IH was divided into migraine (with or without aura), probable migraine, TTH, probable TTH, cluster headache, and other primary headaches. Secondary headaches in structural epilepsies were ruled out by means of imaging studies and were not considered in our analyses.
Statistical analysis
Clinical variables in patients with inter-ictal migraine, inter-ictal TTH and without inter-IH were analyzed. In addition we compared clinical features between patients with post-IH and those without post-IH and between patients with pre-IH and those without pre-IH. We performed χ2-test and t-test to compare categorical variables. T-test and Mann–Whitney test were performed to evaluate continuous variables with a symmetrical and asymmetrical distribution respectively. A logistic regression model was used to calculate OR and a 95 % confidence interval CI to assess the association between dependent and independent variables. Adjustment for the possible effect of confounding variables such as age, sex and migraine prophylactic therapy when appropriate was performed through a multivariable-adjusted logistic regression analysis. Statistical significance was set at p value < 0.05. Statistical analyses were performed with STATA® version 12.0.
Discussion
Headache is a significantly frequent symptom both in patients with epilepsy and in the general population. This is why the prevalence of primary headaches in adult patients with epilepsy remains a matter of debate. Differences in target populations, age range, data collection, diagnostic and classification criteria, methodological study design, and a lower reliability of retrospective studies, may account for such controversial results. Moreover most previous reports focused only on migraine, excluding the other types of primary headache (Table
1) [
5‐
21].
The close collaboration between the Epilepsy and Headache Centers in our study allowed an accurate diagnosis of both epilepsy syndromes and primary headaches.
Among our patients
migraine was the most frequent type of headache occurring in 26 % of patients, in accordance with previous studies in which the prevalence of migraine ranged from 9 % to 30 %. Although the association between these two diseases has been reported in several epidemiologic studies their relationship has not been clarified yet [
5‐
21]. Previous Ottman and Lipton’s studies on 1948 patients with epilepsy demonstrated a twofold higher risk for migraine in patients with epilepsy compared to their first degree relatives without epilepsy, and they also showed a nearly twofold risk of migraine compared to controls (24 % vs 12 %). However, data from Brodtkorb et al. in a Norwegian population of epilepsy patients failed to confirm a statistically significant association between migraine and epilepsy [
6,
26].
Unlike the incidence of migraine we reported
TTH in only 19 % of our cases, a lower prevalence compared to the general population. To date few studies have addressed TTH and epilepsy and no hypotheses on the relation between the two diseases have been put forward. We could speculate that our result is in line with the type of headache. As a mild condition in spite of seizure disorder, TTH could be underestimated in patients with epilepsy [
11,
12,
14,
17,
18,
20,
21].
In any case it is interesting that in our cases both patients with migraine and tension-type inter-IH reported seizure-related headache with features similar in quality to their habitual headache attacks.
Analyzing accurately each different types of headaches in relation to seizures (
peri-IH), we found a slightly lower lifelong prevalence of headaches temporally related to seizures than in the literature (23 % vs 28-50 %) [
18]. We actually think it is probably due to the well-controlled epilepsy in most of our patients, with only very rare seizures, or no seizures at all.
Pre-IH, usually with migrainous features, was reported in 6 % of our population and only in patients with interictal migraine, in line with previous studies [
14,
16,
18,
21,
25]. Only one of our patients with pre-IH had a migraine with aura. In this case the headache attack started about five hours before epileptic seizure onset, and hence did not fall within the diagnostic criteria for
“migralepsy” that anyway is still a controversial entity [
22].
Ictal headache occurred only in three patients (0.8 %) without inter-ictal headache. Two of them have temporal lobe epilepsy and the other one a myoclonic epilepsy. All of them reported a tightening quality of headache with a sense of head pressure and throbbing pain during the seizure or a few seconds before. However, as an ictal EEG recording was lacking in these cases, a definition of
“Epileptic Headache” was not corrected [
9,
12,
27]. There is considerable confusion regarding the definition of “Epileptic Headache”, in both headache and epilepsy classifications (ICHD-II and ILAE). The ICHD-II classification (2004) defines “Epileptic Headache” as a headache with migraine features while the patient also has a focal epileptic seizure. These cases are extremely rare, and the term is not used in the current ILAE and ICHD classification [
23,
24]. In our study we found that a lower number of adults have ictal headache (recognized as a headache lasting from seconds to days, with evidence of ictal epileptiform EEG discharges) than children. We can speculate that it is strictly due to different symptomatology that children have both in epilepsy and in headache. Children are in fact more likely to have autonomic symptomatology attacks, with long-lasting ictal autonomic manifestations, while adults often have other sensory or motor ictal signs. Thus, it is more probable that many cases are genuine seizures imitating migraine, easily recognizable by an EEG recording [
28].
At last in our study
post-IH was the most frequent type of peri-IH, occurring in 19 % of our patients: 37/74 patients reported a migraine type, 30/74 a TTH, and only in seven cases the quality of headache was unclassified. It is not straightforward to compare our data with previous reports in literature in which post-IH prevalence ranges from 12 % to 52 % [
5,
7,
10‐
14,
16,
18,
20,
21,
29,
30]. Several variables need to be considered. First of all most studies are retrospective [
13,
14,
16,
25] and the few prospective studies available were conducted either for a brief period [
15] or on pediatric populations [
31,
32]. Moreover post-IH prevalence could be influenced by the type and intensity of epilepsy, with a higher prevalence in samples of patients with drug-resistant epilepsy [
12,
30]. In our population post-IH was more frequently reported in patients on polytherapy, suggesting a severe epilepsy phenotype, and in patients with a higher seizure frequency (monthly/daily) suggesting that the seizures act as a trigger for headache attacks, as reported in literature [
18,
29].
Our study has some limitations. Firstly the retrospective assessment of headaches can lead to a recall bias, especially for peri-IH in patients who had their last seizure years before the interview. For this reason the prevalence of both pre-IH and post-IH may be underestimated in the current study. Secondly we recruited patients from a tertiary care center, thus there is a possible selection bias. However the features of our sample resemble those of population-based epidemiological studies on epilepsy [
33]. This is mainly due to the fact that patients with severe encephalopathies were excluded, and the catchment area of our Institute comprised the entire city of Bologna and its hinterland, thus resembling a population. Finally the absence of a control group limits the significance of our results.
Over the last decade, possible pathogenetic mechanisms common to epilepsy and migraine have been investigated in depth. The two disorders seem to be genetically inter-related and are comorbid in several clinical syndromes [
34]. An altered membrane channel function and an imbalance between excitatory and inhibitory factors seem to have a central role in the pathogenesis of the two disorders [
28,
35‐
37]. Cortical spreading depression (CSD) is believed to underlie migraine with aura attacks and, according to some evidence, also migraine without aura [
28]. Our study disclosed a relatively low occurrence of migraine with aura (1.5 %) that seems not to support this hypothesis. It is likely that both CSD and other mechanisms, such as different environmental or individual factors (genetic or otherwise), are implicated in the link between migraine and epilepsy.
By lowering the trigger threshold in the epileptic focus, CSD could increase the risk of seizures, explaining the onset of pre-IH [
2]. Similarly, recurrent seizures may predispose a patient to CSD, inducing post-IH [
37]. We suggest that, in analogy with migraine patients, the stressful event represented by an epileptic seizure is a trigger for a headache attack in subjects with migraine or inter-ictal TTH [
36,
38,
39]. However, further studies are required to clarify the mechanisms underlying the two conditions.
Competing interests
The authors declare that they have no competing interests.
Author’s contribution
FB and SC conceived the study and helped to draft the manuscript. PT, PC, GM, GG, LZ, CL, MB, LF, MS, AP and PA took part in the design of the study. GM, CL and LF collected data. GG performed the statistical analysis. GM and LZ participated in writing the manuscript. All authors agreed to accept equal responsibility for the accuracy of the content of the paper. All authors read and approved the final version of the manuscript.