Risk factors for migraine and tension-type headache in 11 year old children

Headache is one of the most frequently reported pain complaints in children and adolescents [1‐3], with the prevalence increasing throughout childhood and peaking at 11–13 years of age [4]. It has been estimated that around 6.1% to 13.6% of children suffer from migraine and 9.8% to 24.7% suffer from tension-type headache (TTH) [5‐8]. Despite the high prevalence, there is a paucity of data on childhood headache relative to our knowledge of adult-onset headache epidemiology. Understanding factors relating to migraine and tension-type headache prior to puberty is important because environmental factors such as stress and other family variables may not be independent of the possible effects of subsequent hormonal changes on the developing teenager [9].

Childhood headache disorders should be recognised as a significant health concern due to the considerable impact on the child and the family [10, 11]. Compared to headache-free children, concerns include greater rates of absence from school [12, 13], fewer and poorer peer relations including bullying [14‐16], and general impairments in home life, school and leisure activities [17‐20]. Although less common, migraine is considered to be more severe and disabling than TTH and involves higher rates of medication use [1].

Obesity, sleep disturbance, behavioural and psychiatric problems have all been linked with childhood migraine and/or TTH. As body mass increases, headache frequency and disability due to headache also increases [21‐23]. Difficulties falling asleep and maintaining sleep, sleep breathing disorders and disorders of arousal have been particularly associated with migraine [24‐26]. Behavioural problems include concentration difficulties, hyperactivity, and conduct problems [27, 28]. Children with headache also have more internalizing disorders, particularly major depression [29] and anxiety [30, 31], than headache-free controls. It is not clear, however, whether these associations are different between the two types of headache in children, given their potentially distinct underlying pathophysiology [32].

Toward this end, we investigated risk factors for primary headache in preadolescent children using data collected from the Auckland Birthweight Collaborative study (ABC), a longitudinal, case–control study of appropriate for gestational age (AGA) and small for gestational age (SGA) individuals born at term. SGA has been associated with numerous adverse outcomes related to cognition, health and behaviour [33‐35]. As such, we studied the association of birthweight and several other antenatal, early life, and childhood risk factors with migraine and TTH at age 11 using the International Headache Society (IHS) classification system [36]. The use of a prospective longitudinal study meant that we were able to investigate factors earlier in life that may be associated with later headache diagnosis, without the problems of retrospective recall such as distortion or forgetting which may introduce bias.

Of the 617 participants in this study, 42.8% reported having experiencing at least one headache lasting 30 minutes or more in the past year (n = 264). From these, 10.5% were classified as experiencing migraines (n = 65) and 18.6% were classified with TTH (n = 115).

Table 1 presents descriptive statistics for sex, pain duration, and parental socio-demographic information. Males were slightly more likely to experience both migraine (n = 34, 52.3%) and TTH (n = 67, 58.3%) than females, but this difference was not statistically significant. Males reported significantly greater pain intensity during headache than females (t₍₂₆₆₎ = 2.17, p = .03) but there were no sex differences in headache duration. As expected, those categorised as having migraine were significantly more likely to have been given a medical diagnosis of migraine in the past than those with TTH. Children with migraine were also significantly more likely to complain of stomach aches than those with TTH. There were no differences between groups in family socioeconomic status, maternal school leaving age or marital status.

Migraine is an episodic and disabling neurological condition affecting about 14% of the adult population [40] while TTH is even more prevalent than migraine [41]. Relative to adults, little is known about the epidemiology of primary headache in children. We found that migraine was associated with both reduced sleep duration and problem behaviour. TTH was associated with antenatal maternal smoking, problem behaviour, higher body fat, and being bullied.

Just under half (48%) of 11 year old ABC study members had reported experiencing at least one headache lasting half an hour or more in the past year. From this, the prevalence of primary headache was 10.5% for migraine and 18.6% for TTH. These rates accord well with other studies [1‐8]. Males and females were almost equally likely to report symptoms of migraine and TTH. Our lack of sex difference is consistent with previous research [1, 4, 31]. Headache prevalence in females tends to increase with age [42]. After about 12 years of age the female to male ratio is 1:3 for migraine and 4:5 for TTH [41‐43]. This is most likely due to hormonal changes during puberty [44, 45] but may also be related to gender differences in cognitive and social reactions to pain [46].

Family socio-economic status was not significantly associated with either migraine or TTH, possibly due to the small sample sizes in the lower socio-economic group or because the components of SES likely to be contributing to headache is more accurately captured by other factors. This finding is consistent, however, with a number of studies that failed to establish a link between socioeconomic status and headache [2, 17, 28, 31]. A population based study of Finnish school children did find a higher risk of TTH for those children whose fathers were lower level white collar workers as opposed to independent traders/employees, but found no association between episodic TTH and family unemployment, economic problems and single parent families [7]. We similarly found no association between headache and marital status.

Little is known about the influence of prenatal events on the development of migraine and TTH in children. Though studies have found evidence of a link between maternal stress during pregnancy and negative outcomes in children [47], we did not find any association between perceived stress during pregnancy and offspring headache. There was also no difference in the rates of primary headache between the SGA and AGA groups.

Maternal smoking during pregnancy was significantly associated with TTH but not with migraine. Arruda and colleagues [48] similarly found that exposure to prenatal tobacco was a risk factor for chronic daily headache in children. It is unclear whether this association is due to socioeconomic factors or is specifically health-related. Women who continue smoking throughout pregnancy are generally of lower age, socio-economic status, level of education and occupational status than non-smokers [49], and offspring are exposed to higher levels of social disadvantage and family dysfunction throughout childhood [50].

It may also be that smoking during pregnancy has affected the child’s neonatal environment, predisposing them to develop headaches later in life. Smoking during pregnancy has been linked to increased risk of numerous childhood problems including asthma [51], attention deficit disorder [52] and conduct disorder [50].

Percentage body fat at age 11 was associated with TTH. These findings are supported by research by Hershey and colleagues [53], who found that body mass was positively associated with headache frequency in 913 children recruited from paediatric headache centres. Furthermore for those children who were defined as overweight in the first visit, drops in weight were significantly associated with decreases in headache frequency at both 3 and 6 months follow up. Earlier studies have also found associations between obesity and migraine frequency among children [21‐23] but there are no known reports of increased frequency or prevalence of TTH among overweight/obese children or adults. It may be that children who have more body fat may also have higher stress due to a lower self-esteem. It should be noted that this variable didn’t remain significant in our multivariable analysis, possibly due to low statistical power. Further research with larger samples is needed to determine the relationship between body weight and headache disorder and the mechanism behind it.

One possible consequence of being a child with higher body fat might be increased susceptibility to bullying by peers. We found that bullying was not associated with migraine but was significantly associated with TTH. Children with TTH tend to be shier and less sociable than children with migraine [28]. Further research is needed to determine whether severity of bullying is associated with headache prevalence or frequency and whether bullying is consistently associated with TTH rather than migraine.

Children with migraine were more likely to sleep for less than 10 hours each night when they were 7 years of age compared to those with no headache disorder. Several prior studies have also demonstrated the co-occurrence of sleep disturbances and migraine in children [2, 24]. Although children with migraine in our study slept less than children without headache, this did not seem to be related to increased time watching television. Others have similarly found no relationship between migraine or TTH and hours spent using electronic equipment such as television/computer/PlayStation [17]. One possible reason that our migraineurs are getting less sleep is that they may also suffer from a periodic limb movement disorder [54]. This possibility will be tested at the next ABC medical assessment.

The co-morbidity of sleep disorders and migraine in childhood has led to the suggestion that they may have a common genetic or pathophysiological mechanism [26]. The serotonergic system is likely to be involved, as it has been implicated in the physiology of migraine and plays an important role in the initiation and maintenance of sleep activity [26]. Serotonin also plays an important role in emotional disorders such as depression, which has also been implicated in the physiology of migraine [31] and studies with adults have found that depression precedes migraine onset [55]. Although we did not find depression was related to childhood migraine or TTH, analyses with larger samples are needed to confirm this and to determine whether serotonin links the underlying pathophysiology between migraine, sleep problems and emotional disorders.

Though we didn’t find an association between headache and depression, our results indicate that many other areas of psychological functioning were significantly associated with childhood headache. This is consistent with a growing body of literature suggesting a link between childhood headache disorders and emotional and behavioural difficulties [27‐31]. In our study, psychological functioning was assessed by parent report at ages 3.5, 7, and 11 (and self-report at 11) using the SDQ. This questionnaire assesses emotional symptoms, conduct problems, hyperactivity/impulsivity, and peer problems. Scores from these subgroups are summed to provide a total difficulty score.

More specifically, migraine and TTH were both significantly associated with self-reported borderline/abnormal score even after controlling for other risk factors. An earlier clinical study of behavioural and emotional difficulties found that children with TTH have more behavioural, emotional and temperamental difficulties than those with migraine [28] while a later study from a paediatric headache centre found the reverse [31]. In particular, migraine was associated with a borderline/abnormal score on the hyperactivity/impulsivity subscale of the SDQ, consistent with an earlier study [56]. Despite some inconsistencies, findings suggest that primary headache in childhood is complicated by, or occurs due to, a combination of a predisposition and psychosocial and behavioural difficulties.

Our findings support the view that migraine and TTH should be considered separate entities in childhood [but see 17 for an alternate view]. While both migraine and TTH were associated with more self-reported total difficulties at age 11 years, we found a number of independent risk factors. TTH was associated with smoking during pregnancy, parent-reported total difficulties at age 3.5 years, higher percent body fat, and being bullied at age 11. Migraine was independently associated with reduced sleep duration at 7 years, consistent with reports that psychological problems and sleep issues are often linked with migraine.

A number of important limitations of our study need to be discussed. Firstly, although the overall sample size was large (n = 617), the combination of the prevalence of headache type and other factors with low prevalence such as depression resulted in small case sizes for those factors. Secondly, our participants were from New Zealand European families and did not include Maori and Pacific families. This was due to low response rate from non-European families during the earlier follow-up phases. Thirdly, we did not have access to medical information from first degree relatives and therefore could not include family history of headache in our analyses. Fourthly, our findings are correlational and should not be interpreted as factors that necessarily cause headache disorder.

Despite the many limitations, there were some particular strengths of our study. For example, much of the research into childhood headaches is conducted using clinical samples from speciality headache centres. Results from these studies may therefore not generalize to the general population. Moreover, many studies do not differentiate between general headache and disorder subtypes, or they use non-standardised classification systems to diagnose headache, making it difficult to compare results across studies. Broadly representative cohort studies using internationally recognized headache diagnostic criteria might more reliably determine if differences exist between children who suffer from primary headaches and from those who do not and whether differences exist between headache groups.

In summary, a number of health-related and psychosocial risk factors were associated with childhood headache in our cohort study. Future studies are planned to determine if there is an association between primary headache and common illnesses/diseases in childhood. Further work to determine risk factors of migraine and TTH will help ensure that earlier diagnosis and effective treatment occurs.