Abstract
Objective
To compare comorbidities between migraine and tension headache in patients treated in a tertiary pediatric headache clinic.
Methods
Files of patients with migraine or tension headache attending a pediatric headache clinic were retrospectively reviewed for the presence of organic comorbidities. Additionally, patients were screened with the self-report Strengths and Difficulties Questionnaire to identify nonorganic comorbidities. If necessary, patients were referred to a pediatric psychiatrist, psychologist or social worker for further evaluation.
Results
The study cohort comprised 401 patients: 200 with migraine and 201 with tension headache. The main organic comorbidities were atopic disease, asthma, and first-reported iron-deficiency anemia; all occurred with statistical significance more often with migraine than with tension headache (Familial Mediterranean fever was six times more frequent in the migraine group than in the tension headache group, but the difference was not statistically significant. Nonorganic comorbidities (psychiatric, social stressors) were associated significantly more often with tension headache than with migraine (48.3% versus 33%; p = 0.03).
Conclusions
Children and adolescents with migraine or tension headache treated in a dedicated clinic have high rates of organic and nonorganic comorbidities. In this setting, patients with migraine have significantly more organic comorbidities, and patients with tension headache, significantly more nonorganic comorbidities.
Introduction
The most common types of headache in the general population of children and adolescents are tension headache and migraine (1). A German epidemiology study of children aged 7–14 years from 8000 households randomly selected from community registers reported a 7.5% rate of migraine and 18.5% rate of tension-type headache (1).
In a detailed review of several studies, an association was noted between primary pediatric headache and various organic and nonorganic comorbidities, including atopic disease, cardiovascular disease, epilepsy, depression, anxiety, sleep disorders, attention-deficit hyperactivity disorder, and Tourette syndrome (2). The nationwide survey of Arruda et al. (3) is one of the few to compare comorbidities between migraine and tension headache in the general pediatric population, but further data are needed, in other settings.
The aim of the present study was to compare comorbidities in children and adolescents with migraine or tension headache attending a tertiary pediatric headache clinic.
Methods
The study group included children aged 2–18 years diagnosed with migraine or tension headache who attended the headache clinic of a tertiary, university-affiliated, pediatric medical center in 2010–2015. The diagnosis of tension or migraine headache was based on the criteria of the International Classification of Headache Disorders of the International Headache Society (ICDIII-beta version) (4). Exclusion criteria were age outside the predefined range, diagnosis other than migraine headache or tension headache, and missing data.
At the first visit to the headache clinic, older children and the parents of younger children routinely complete a standard questionnaire based on the 2013 revised International Headache Society criteria for children. As part of the anamnesis, the clinic physician reviews the patient’s computerized medical files from hospital departments, subspecialist pediatric clinics, and the primary pediatrician for data on background and clinical parameters. (In our country, these files are linked and can therefore be viewed at a tertiary center.) The parents are interviewed to gather further information on any medical disorders in the child, and the patients are interviewed by the head of the clinic. They are also asked about problems with school, peers, or family, socioeconomic issues, learning disorders, and school achievements. Patients aged 11–17 years, and the parents of all patients, complete the short form of a self-report Strengths and Difficulties Questionnaire (SDQ) (5). Parents completed the SDQ for younger patients by themselves. The SDQ (Self Diagnosis Questionnaire) is a behavioral screen that includes 25 items assessing emotional symptoms, conduct problems, hyperactivity/inattention, peer problems, and prosocial behavior problems (6).
All patients also undergo a general physical examination. Body mass index (BMI) (weight (in kilograms)/height (in meters squared)) is calculated together with the body mass index standard deviation score (50th percentile BMI value for age and sex of the Centers for Disease Control (CDC) and Prevention/standard deviation of the CDC value) (7). Some of the patients who present to the headache clinic are already being followed by a pediatric neurologist for background neurological diseases diagnosed before admission. A pediatric neurologist sees others, after presentation for evaluation of the presence of epilepsy (diagnosed by electroencephalography) and other neurologic diagnoses, such as attention deficit disorder (with or without hyperactivity), and a pediatric rheumatologist evaluates familial Mediterranean fever (FMF).
Children found to have a social problem by anamnesis or abnormal SDQ scores are referred to a pediatric psychiatric and psychologist and a social worker. All psychiatric diagnoses are based on DSM-IV criteria (8). Anemia is diagnosed on the basis of the medical records, according to accepted norms (9), and atopic diseases according to the criteria for allergic diseases (10). FMF is diagnosed according to published criteria (11).
The decision to prescribe migraine pharmacologic treatment is made after evaluation for all the possible organic diseases and consultation with the psychiatrist, psychologist, and social worker, based on their findings.
Data for the present study were collected from the headache clinic files. The study was approved by the hospital’s Research Ethics Board (approval number RMC- 534-14). Owing to the study’s retrospective design, the Research Ethics Board waived the need for informed consent.
Statistical analysis
With a proposed sample size of 199 in each arm to detect less frequent diseases and comorbid states, the study would have a power of 80% to yield a statistically significant result. This computation was based on the assumption that the difference in proportions is 0.1. The criterion for significance (alpha) was 0.050.
Data were managed and analyzed with BMPD software (10). Continuous variables are presented as means and standard deviations. Because of their non-gaussian distribution, continuous variables were compared between groups with the nonparametric Kruskal-Wallis test. We also applied Levene's test for variance; if it showed a significant difference between variances, we applied the equality of means test in which variances are assumed to be unequal. Discrete variables were compared between groups with Pearson's chi-square test or Fisher’s exact test, as appropriate. A p value of <0.05 was considered significant. Stepwise logistic regression was used to identify the variables most significantly associated with a migraine versus tension headache.
Results
Patient characteristics (Table 1)
Demographic and clinical parameters of patients with migraine or tension headache.
Data are presented as n (%), mean ± SD, or mean ± SD (median).
Disease characteristics
In the whole cohort, the mean time elapsed from onset of the headaches to admission to the headache clinic was 18.35 ± 21.94 months (range 1–120 months, median 12 months), and the mean number of headache attacks per month was 15.87 ± 10.47 (range 1–30 attacks per month, median 12 per month). Chronic headache, defined as a frequency of ≥15 attacks per month, was diagnosed in 188 (46.82%); the remainder were categorized as having episodic headache. The mean duration of the headaches was 8.12 ± 16.66 hours (range 0.5–168 hours, median 3 hours). The breakdown of these parameters by group is shown in Table 1. There was no between-group difference in time from headache onset to clinic admission or in duration of headache attacks. Patients with tension headache had a significantly higher frequency of headache attacks per month than patients with migraine (18.4 ± 10.39 versus 13.69 ± 10.1 for migraine, p < 0.001) and more chronic headache (37.5% versus 56.5% for migraine, p < 0.001). Patients with migraine had significantly more nausea and dizziness, phonophobia, and photophobia.
SDQ scores for all study patients were significantly higher in the tension-headache than the migraine group, both in self and parent reports (p < 0.001) (p < 0.001).
Organic comorbidities (Table 2)
Organic diseases in patients with migraine or tension headache.
All data are presented as n (%).
Normal weight – body mass index standard deviation score −1 to 1, equivalent to the 5th percentile to ≤85th percentile; overweight – body mass index standard deviation score >1 to 2, equivalent to the 85th to ≤95th percentile; obese – body mass index standard deviation score >2; equivalent to the >95th percentile.
Calculation of the body mass index standard deviation score revealed an overweight/obesity rate of 21.19% in the cohort. There was no significant difference in obesity rate between patients with migraine or tension-type headache.
Nonorganic comorbidities and social stressors (Table 3)
Psychiatric comorbidities and social stressors in patients with migraine or tension headache.
All data are presented as n (%).
Some patients had more than one comorbidity.
Eating disorder, post-traumatic stress disorder, oppositional defiant disorder.
Recent immigration of the patient, family violence, patient is adopted.
Psychiatric comorbidities were diagnosed in 89 patients with tension headache (44.3%) and 59 patients with migraine (29.5.%). This difference was statistically significant (p = 0.03). Corresponding rates of comorbid social stressors were 44.27% and 23.4% (p < 0.001). There was also a higher percentage of patients with both psychiatric comorbidities and social stressors in the tension-headache group (n = 42, 20.9%) than in the migraine group (n = 22, 11%; p < 0.001). Specifically, significantly more children in the tension-headache group had divorced parents (p = 0.03) and/or peer problems (p = 0.002). Individual psychiatric disorders with significantly higher rates in the tension-headache group included depression (p = 0.03) and learning disorders (p = 0.04); the difference for anxiety was significant (p = 0.05).
Logistic regression analysis
Variables were included in the logistic regression for migraine versus tension headache in three steps: step 1: organic comorbidities; step 2: chronic versus episodic migraine; step 3: psychosocial comorbidities. Results were as follows: organic comorbidities – OR = 2.4, 95% CI 1.6–3.6; chronic migraine – OR = 0.46, 95% CI 0.31–0.7; psychosocial comorbidities: OR = 0.52, 95% CI 0.34–0.8.
Treatment
Sixty-seven patients with migraine (33.5%) received preventive pharmacologic treatment. It is our clinic policy not to treat tension headache pharmacologically.
Discussion
The present study shows that in children and adolescents migraine is associated with a significantly higher rate of organic comorbidities than tension headache in children attending a tertiary headache clinic. Tension headache is associated with a significantly higher rate of nonorganic (mental) disorders and psychosocial stressors.
Organic comorbidities
Atopic diseases
Among the organic disorders, atopy and asthma occurred more frequently in the migraine group than the tension-headache group. This finding agrees with the study of Özge et al. (13), in which atopic disorders were documented in 18.2% of 438 patients with migraine and 6.4% of 357 patients with tension headache (p < 0.001). It is further supported by the large epidemiological study from the UK based on the national database (all age groups), which reported a relative risk of 1.59 (95% CI 1.54–1.65) for asthma in patients with migraine (14).
Özge et al. (13) suggested, on the basis of their findings and earlier data (14), that the comorbidity between asthma and migraine may point to a common underlying immunological mechanism possibly related to a lowered threshold for neurogenic inflammation. Accordingly, in a review study, Theoharides et al. (15) observed that mast cells, which play a crucial role in the pathogenesis of asthma, are located perivascularly in close association with neurons, especially in the dura, where they can be activated to secrete vasoactive, pro-inflammatory, and neurosensitizing mediators by neuropeptides in the trigeminal nerve as well as the cervical or sphenopalatine ganglion.
Others have implicated fibrinogen in the asthma-migraine association. Fibrinogen is known to play a proinflammatory role in vascular wall disease such as stroke, spinal cord injury, brain trauma, multiple sclerosis, and Alzheimer’s disease. (16) A recent study suggested that fibrinogen binding to the ICAM-1CD11/CD18 ligand promotes mucin production in the airway, such that fibrinogen may also be involved in chronic airway diseases such as asthma, chronic obstructive pulmonary disease, and cystic fibrosis. (17) Fibrinogen activates proinflammatory pathways such as nuclear factor-κB by signaling through CD11b/CD18, which results in the local production of inflammatory cytokines (16). It also signals directly or indirectly other receptors, adhesion molecules, and cell-surface proteins that are involved in inflammatory processes to participate in the induction of macrophage activation and the release of several chemokines and cytokines, such as monocyte chemoattractant protein-1, interleukin (IL)-6, IL-8, tumor necrosis factor-α, and matrix metalloproteinase 1. The role of IL-6 and other interleukins (IL-8, IL-10) and of leukotriene E4 and prostaglandinF2a in pediatric and adult migraine has been reported in the past (15,18).
Together, these studies emphasize the possible common inflammatory mechanism of asthma and migraine, supported by the present results. We assume that other atopic conditions such as atopic dermatitis and allergic rhinitis were not found to be significantly more frequent in the migraine group because of their relatively low prevalence in the cohort as a whole. Our preliminary analysis showed that a larger sample size would be needed to yield statistically significant results for less prevalent diseases.
FMF
In the present study, FMF was found six times more often in the migraine group than the tension-headache group. The between-group difference was not statistically significant (p = 0.068), most probably because of the generally low prevalence of FMF in our cohort and the small sample size.
Uluduz et al. (19) conducted a multicenter study in Turkey including 378 patients with FMF under treatment in rheumatology departments. On interviewing the participants, they found that 29.5% had migraine, 37.6% probable migraine, and 32.9% tension-type headache. We calculated that the prevalence of migraine/probable migraine in this population (19) was significantly higher than the prevalence of tension headache (p < 0.001). It has been suggested that under normal conditions, inflammation is induced when pyrine domains (PYDs) in the proteins involved in the NLRP3-controlled inflammatory complex bind with the PYD in pyrine, cleaving the IL-1β precursor (20). In FMF, however, a mutation in the pyrine protein gene leads to an imbalance in this inflammation cascade (20) Thus, given the inflammatory process of FMF and the relationship of neuroinflammation in the trigeminal vascular vessel with migraine, the high prevalence of FMF in migraine (19) may be related to the common vascular immune-mediated pathogenic mechanism. By contrast, the tension headache reported in the study (19) was probably associated with the depression and anxiety caused by the chronic systemic disease, but was unrelated to FMF activity per se.
It is noteworthy that the methodology in the Turkish study (19) differed from ours, in that they examined the occurrence of different types of headache in pediatric patients known to have FMF whereas we examined the presence of FMF in pediatric patients diagnosed with headache in a headache clinic. We should also keep in mind that the type of FMF mutation affects the severity of the disease, and both the prevalence and type of FMF mutations differ between our country and Turkey (21,22), which may lead to a different impact in patients with migraine.
Anemia
Iron-deficiency anemia affects an estimated 30% of the world’s population (23). We found that iron-deficiency anemia was significantly more prevalent in the migraine group than the tension-headache group.
Lateef et al. (24) examined rates of recurrent headache and its comorbidities in 10,198 children in the United States. They found that the odds ratio for having anemia was 2.06 in children with headache compared to children without headache. The type of anemia and the type of headache were not analyzed, although owing to the higher frequency of iron-deficiency anemia compared to other types of anemia, we assume most of the children with anemia had this type. Our study is the first to link migraine headache specifically with iron-deficiency anemia. We found no such association for other types of anemia. We speculate that the two diseases may be linked by an underlying neuroinflammatory pathogenic mechanism. This is supported by the innovative study of Łukaszyk et al. (25), who showed that in iron-deficiency anemia the level of hepcidin, a 25-aminoacid peptide synthesized in hepatocytes and one of the key elements in iron metabolism (26), increases in response to iron overload and inflammation, preventing iron absorption. The increase in hepcidin production during iron-deficiency anemia is mediated by lipopolysaccharide, IL-6, and IL-1, all of which are also involved in local neuroinflammation and pain signaling in migraine (27). Indeed, our clinic is part of a tertiary medical center, so most of our patients with migraine are referred following hospital admission, and have severe disease. We speculate that in some cases the increased IL-6 and IL-1 levels induced by the migraine may lead to an increase in hepcidin level and a subsequent decrease in iron absorption and ultimately iron-deficiency anemia (18,27). However, primary iron deficiency anemia from other causes cannot be ruled out.
Although low iron levels are more prevalent in low socioeconomic sectors, we did not find any correlation between socioeconomic status and headache type.
Obesity
Several studies have reported a correlation between primary headache (migraine, tension) and obesity (28–31) and between headache frequency and obesity grade (31,32). The two disorders may be related pathophysiologically (28). An extensive central and peripheral overlap in the pathways regulating feeding and the pathways implicated in migraine is suggested by findings of hypothalamic activation and release of the same inflammatory neurotransmitters (e.g., serotonin) and immune modulators (e.g., adiponectin) in both disorders. Furthermore, inflammation regulators such as adiponectin, IL-6, adenosine diphosphate, C-reactive protein, IL-6, and tumor necrosis factor-alpha may be associated with headache severity (28).
The 2010 cross-sectional, general population study by Robberstad et al. (29) reported that compared to normal-weight adolescents, overweight or obese adolescents had a 60% greater risk of migraine (OR = 1.6; 95% CI: 1.4–2.2, p < 0.0001) and a 40% greater risk of tension-type headache. Accordingly, Ravid et al. (31) reported that obesity seems to be more of a risk factor for migraine than for tension-type headache. For both types of headache, a high BMI percentile was associated with increased headache frequency and disability (31).
Our study yielded contradictory findings. We found no significant correlation of headache type (migraine, tension) with obesity grade. Furthermore, the 21.96% rate of overweight/obesity in our cohort was in line with the estimated rate of pediatric obesity in the general population of Israel (26%) according to the 2010 report of the Organization for Economic Cooperation and Development (32). The difference from other studies in the literature may be related to differences in the methodologies used. Alternatively, migraine is known to be influenced by both genetic and environmental factors, and these may differ among populations. Our finding might be explained by a clinical spectrum of primary headaches of migraine and tension headache involving a single pathophysiological mechanism, as suggested by Cady (33). However, the correlation of other organic comorbidities such as asthma and iron-deficiency anemia with migraine in our study negates this theory.
Epilepsy
Epilepsy is known to have a high frequency (up to 17%) in individuals with migraine relative to the general population (0.5–1%). Additionally, a higher frequency of epilepsy (8–15%) has been reported in patients with migraine than in otherwise healthy individuals (34). These data are valid for both adults and children. Many of the studies of epilepsy and headache support the hypothesis that excessive neocortical cellular excitability is the main pathological mechanism underlying both diseases and that cortical spreading depression is their connecting point (35).
However, we found no significant difference in epilepsy rate between the migraine and tension-headache groups. This might be explained by the very low overall rate of epilepsy in our small cohort (1.9%). Our hospital treats young patients with epilepsy in the pediatric neurology department, and we assume most patients with headache and comorbid epilepsy are treated there as well. Further studies in a larger patient sample are needed to clarify this issue.
Nonorganic comorbidities
Significantly higher rates of psychiatric comorbidities and psychosocial stressors were found in the tension-headache group than in the migraine group.
Findings in the literature on the association of primary headache with psychological, behavioral, and socioenvironmental factors are controversial. Mazzone and colleagues (36) evaluated children referred to a headache clinic with migraine or tension-type headache and a pain-free control group for several psychological variables including internalizing and externalizing symptoms. They found differences in most variables between the headache groups and the control group and also between the migraine and the tension-headache groups. In general, children with tension headache had significantly higher scores for psychosocial problems than the migraine group, indicating a lesser degree of psychological adjustment. These results are supported by the present study.
Kröner-Herwig and Gassmann (35) performed a large cross-sectional study on children and adolescents aged 7–14 years derived from a randomly selected sample of families in the general population. They found that headache was highly associated with maladaptive psychological traits, particularly internalizing symptoms (odds ratio >4 for migraine), compared with socio-environmental factors, particularly behavioral ones. By headache type, unfavorable psychological traits and socio-environmental strains were associated more with migraine than tension-type headache. However, in a literature review of 11 studies, Amouroux and Rousseau-Salvador (37) reported that patients with migraine had slightly higher scores than controls on at least one of the anxiety or depression scales, but their average scores on all studies were still within the normal range established at validation of the instruments.
Arruda et al. (3) conducted a nationwide survey of 5671 Brazilian schoolchildren. Nine percent had episodic migraine, and 12.8% had episodic tension headache. Psychosocial adjustment was evaluated with the SDQ. Abnormal scores were found to be significantly more likely in children with episodic migraine relative to children without headaches and children with episodic tension headache, and they were significantly influenced by frequency of headache attacks, presence of nausea, poor school performance, prenatal exposure to tobacco, as well as phonophobia and photophobia. In our study, patients with tension headache had significantly more depression, learning disorders, and social/behavioral problems. The difference for anxiety approached significance.
The differences among the studies may be related to the differences in design. To identify nonorganic comorbidities, we used an interview and the SDQ, and on the basis of the findings, referred patients to a psychiatrist, psychologist, or social worker, as necessary. Others based the psycho-psychiatric diagnosis on direct interview with a psychologist or psychiatrist. Usually patients with tension headache are treated by their primary physician rather than a headache specialist because their pain is less severe than that of patients with migraine (according to the ICDIII criteria) (6). Thus, we assume that the patients with tension headache who are referred to our clinic may have a more severe disease than the general population of patients with tension headache, and this may affect the prevalence of nonorganic comorbidities relative to patients with migraine. This is supported by the greater headache frequency in the tension-headache group than the migraine group in our study. Accordingly, Arruda et al. (3) found that abnormal psychosocial adjustment scores were more likely in children with a greater frequency of headache attacks.
Zebenholzer et al. (38), in an innovative study in eight headache centers for adults in Austria, found that depression and anxiety were significantly more common in patients with chronic headache than episodic headache (64% vs. 41%, p < 0.0001). Others, in Taiwan, reported an association of chronic tension-type headache with increased rates of major depressive disorder and anxiety disorders (39). In our cohort, chronic headache was more prevalent in the tension-headache group than the migraine group, which may account for the significantly higher rate of depression and anxiety in the patients with tension headache. Family or school-related stress has been reported to be a common psychological predictor of headaches among children, particularly tension-type headaches (38). Family stressors included marital problems, family bereavement, poor relationships with parents, illness in the family, and low socioeconomic status (38). We too found higher rates of social problems in the tension-headache group, most with chronic headache (56%).
Our study has several limitations. First, since our hospital is a tertiary center, the results may be biased toward patients with more severe disease. Second, for some of the comorbidities, frequencies were lower than estimated by our power analysis, and a larger sample size would have helped to identify more comorbid states. Third, most of our patients had headache for a prolonged period before admission to the clinic (18.35 ± 21.94 months, median 12), which may have affected the prevalence of secondary nonorganic comorbidities, which increase with an increase in duration of pain. Fourth, only 51.1% of our patients were more than 11 years old. In the remainder, only the parents completed the SDQ, and we could not compare the findings with the children’s scores.
Conclusion
Pediatric patients with migraine and tension headache treated in the setting of a tertiary pediatric headache clinic have a high rate of organic and nonorganic comorbidities. Migraine headache is significantly associated with more organic comorbidities, and tension headache with more nonorganic comorbidities. A multidisciplinary-team approach, including experts in various pediatric subspecialties and a psychologist, child psychiatrist, and social worker, is needed to ensure proper and comprehensive treatment.
Footnotes
Clinical implications
The finding that migraine headache is significantly associated with more organic comorbidities, and tension headache with more nonorganic comorbidities in pediatric headache clinic settings can help to guide diagnosis and treatment.
The association of migraine with organic comorbidities (asthma, iron deficiency anemia, and most probably, familial Mediterranean fever) suggests a possible co-inflammatory mechanism.
Our study is the first to link migraine headache specifically with iron-deficiency anemia. The two diseases may be linked by an underlying neuroinflammatory pathogenic mechanism.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.