Long-term follow-up of a community sample of adolescents with frequent headaches

Headache is one of the most common health and somatic complaints reported by adolescents in the general population [1‐5]. In two reviews of epidemiological surveys conducted in various countries and cultures, the estimated mean prevalence rates of unspecified headaches among children and adolescents were 54.4% [4] and 58.4% [4, 5]. A striking increase in the prevalence of unspecified headaches and migraine occurs at the onset of puberty among girls but not boys [4‐8]. Such complaints can negatively affect the quality of life and lead to impairments in daily functioning such as school performance, absence and recreational activities [9]. Severe or frequent headaches in children and adolescents are also related to emotional problems; in particular, anxiety, depression, behavioral problems [10, 11], other somatic symptoms, social impairment, problems interacting with peers, pain comorbidity, and parental headache [12, 13].

The persistence of headaches among children and adolescents experiencing unspecified headaches or migraine is an important aspect and burden of these complaints. Several retrospective and prospective follow-up studies of clinical samples of children and adolescents experiencing primarily migraine have reported on both the short-term and longer perspectives [14‐17]. However, such samples typically include young people with more complicated neurological symptoms, frequent attacks, or impairment compared with those who are not referred [18].

In community surveys conducted in Scandinavia and Germany, the course of unspecified headaches in these age groups has been investigated primarily over the short term covering 1–3-year periods [14, 19‐25] for which the overall persistence of headaches in the general population has been found to be high. For example, at the 1-year follow-up evaluation, slightly more than half (57%) of children and adolescents with headaches still experienced such complaints, as reported by parents [19]. In another similar longitudinal study, one-third of early adolescents reported frequent headaches (at least once a week) [20]. In a 3-year follow-up of school children, most of those with headaches (80%) still reported headaches, and female gender and the frequency of headaches predicted the persistence of these complaints [22].

To date, the existing information on the long-term prognosis of headaches in children and adolescents in community populations is limited. In a pioneering survey conducted in the 1950s, Bille followed a subsample of school-aged children with non migrainous headaches for a 6-year period [26] and another group with pronounced migraine at various time points for an extended period up to 40 years [27, 28]. Although about half of the children with migraine still had attacks around the age of 50 years, the prognosis was better for men than for women. In a school-based sample, Özge and collaborators followed 1155 children for a 6-year period up to mid-adolescence [29]. The overall headache prevalence increased from 45.2 to 78.7%, and most students had stable headaches, although the headache diagnoses showed a high transition rate over time. In this sample, female gender and having a parent with headaches predicted headache stability. In a 5- and 10-year follow-up of 8-year old children in a community study, Schmidt and colleagues found a high stability of headaches across the two reassessment points up to late adolescence (73 and 47%, respectively) [30].

In a 13-year follow-up of a sample of 335 school children into young adulthood (21–27 years of age), Brattberg [31] reported that about one-quarter still had headaches, while the same percentage represented incident cases. A headache frequency of at least once a week predicted the persistence of headaches. A longitudinal 15-year study of headaches severe enough to disturb daily life in the past 6 months collected data for children starting school (age 7 years) and again at the ages of 13–14 and 22 years [32]. The overall headache prevalence was virtually unchanged during and after puberty in this study. However, these estimates were surprisingly low, probably because of the conservative definition of recurrent headaches associated with impairment used in the study. In a general population study of migraine and other headaches among children starting school from the same research group in Turku, Finland, Anttila and collaborators examined long-term trends at two separate time points in adulthood when the participants were 22 and 35 years of age in separate cohorts [33]. They found an increase in the incidence of frequent headache and migraine in both boys and girls over a 30-year period. In another extensive longitudinal study of a large sample (N = 11,407) from the general population in the UK, parent-reported headaches among children starting school were associated with the frequency of headaches, multiple somatic symptoms, and psychiatric morbidity at the age of 33 years [8].

We have previously reported on the incidence, course, and 1-year outcomes of frequent headaches among early adolescents in a representative sample of the general population [19].We also assessed the influence of several potential predictors of outcomes: gender, age, relationships with peers, parental divorce, pain comorbidity and impairment related to disease or pain, and behavioral and emotional problems; in particular, depressive symptoms. In the present study, we investigated these aspects further in a long-term (14-year) study of the participants in our original sample.

In the present long-term follow-up study, early adolescents (aged 12–15 years) in a representative general population sample were reassessed about 14 years later in adulthood (mean age 27.2 years) with a focus on the prevalence, incidence, course, and outcome predictors of frequent versus infrequent headaches.

Our prevalence estimates for these young adults and a report of headaches at least once a month at the extended follow-up was low—17.5%—and only 8.4% of participants reported frequent headaches occurring once a week or more. These relatively low estimates of headache frequency may reflect our introductory phrasing of the question, “Do you suffer from regular pain?” This item was used to allow comparisons of prevalence estimates and changes across the follow-up period with the rates reported by early adolescents at the baseline and our previous 1-year follow-up evaluation [20]. The present prevalence estimates are similar to those of a large epidemiological survey of adults conducted in one of the counties and the same health region as included in the present baseline study. In the previous study, 11.6% of females and 4.4% of males reported frequent headaches (“Have you suffered from headaches during the past year?”) [38]. Furthermore, our estimate of daily or almost daily headaches (2.4%) was identical to that reported in the previous survey of adults. Although this latter survey was conducted some years before our present follow-up evaluation, our estimates seem to be reliable and are likely to reflect the true prevalence rates of headaches in this age group of young adults.

Over a short-term perspective of 1–3 years, the persistence of headaches has been shown to be high; that is, between 33% and nearly 80% of children and adolescents still experienced headaches at reassessment; in particular, frequent headaches [19, 20, 22, 23]. High persistence rates have also been reported in follow-up studies for longer periods (5–10 years) in which half to most (47–83.5%) of the school children still had headaches in adolescence [29, 30], whereas about one-quarter of young adults still reported headaches in longer follow-up evaluations (13–15 years) [31, 32]. To date, the longest follow-up studies of school-aged children were performed by Bille, who examined the course of pronounced migraine over 30–40-year intervals when about 50% of the participants were about 50 years of age and still experienced migraines [28].

In the present study of early adolescents with frequent headaches, who were followed for about 14 years, 19% of the participants reported such headaches, but there was a negligible gender difference. Our findings seem to be similar to those of Aromaa and colleagues, who performed a 15-year follow-up study of frequent headaches in school children as they became young adults [32]. However, their definition of frequent headaches during the past 6 months also included impairment of the child and specific areas in daily life, which may have contributed to very low prevalence rates and smaller numbers in subgroups.

While annual incidence rates of headaches were reported for children and early adolescents in two community-based surveys [19, 20], our previous study estimated the 1-year incidence for frequent headaches as 6.5%. In a reassessment of school children with no headache at baseline, Özge and colleagues [29] reported that three-quarters had developed some form of headache disorder at the 6-year follow-up. In longer follow-up studies of 14 and 26 years, 25.1% [8] and 12.2% [29] of school children had developed unspecified headaches, respectively. In a 16-year follow-up of children in a headache-free control group, 11% had developed migraine [28]. However, we found no extended follow-up studies using the same criterion for frequent headaches as used in the present study, in which 7.4% of participants had developed frequent headaches over the 14-year period. Importantly, we found a strong gender difference: females had a higher risk than males (11.3 and 1.5%, respectively) for developing frequent headaches in adulthood. A previous epidemiological survey conducted in the same health region [38] reported a similar gender difference in prevalence rates of frequent headaches (at least six episodes per month) among 20–29-years olds: 11.6 and 4.4% for females and males, respectively.

The predictors of outcomes have been examined further in long-term follow-up studies. For example, a 6-year follow-up of school children into adolescence in Turkey studies the predictors of incident cases and persistence of headache [29]. In that study, having a sibling with headache and a working mother increased the risk for developing headaches in adolescence for headache-free children, whereas having a mother with headache and female gender were associated with persistence of headaches. In the 15-year follow-up by the Turku, Finland, research group, headaches at school entry also predicted headache and migraine in early adulthood [32]. In the longer 26-year follow-up of children starting school in the UK, frequent headaches predicted headaches, multiple physical symptoms, and mental morbidity in adulthood [8].

In their extensive review of the potential risk factors that may differentiate between various headache disorders in cross-sectional clinical and school-based samples, Karwautz and collaborators [35] reported that parental divorce, low number of friends, and school absence were important discriminators between headache disorders such as migraine and tension-type headaches compared with headache-free controls. A previous study of adolescents in a community sample also found a strong relation between pubertal development and an increase in headaches and depressive symptoms in girls, which suggested a possible link to hormonal changes during puberty in girls [39].

Whether these potential indicators found in cross-sectional studies also predict short- and long-term outcomes in prospective outcome studies of children and adolescents with recurrent headaches in community-based studies is an important issue. We evaluated the importance of these short-term predictors and others in our previous 1-year follow-up study of early adolescents in the original sample [20]. In that study, unspecified frequent headaches at the baseline, gender (worse prognosis for girls), impairment of leisure-time activities and seeing friends, and high levels of depressive symptoms contributed significantly to the outcome. Of particular note is that the same set of baseline factors also predicted the frequency of headaches in the present 14-year follow-up evaluation of some of the original sample. This established model shows that the frequency of headaches, presence of comorbid depressive symptoms, and social impairment are important predictors of the short- and long-term prognosis for frequent headaches in early adolescents in unselected community samples.

The reasons for these variables having such a significant effect both in short term and long-term perspectives on frequent headaches are likely to be linked to relationships between vulnerability and resilience factors in adolescence. Increasing demands both developmentally, socially and academically might produce a stress pattern influencing the development and persistence of frequent headaches in youth. They may also lead to depressive symptoms, and their poor regulation often co-occuring with poor social skills and having few friends and avoidance of social activities, contribute to a negative and accumulative spiral into adulthood.

We note that many of these young people with such headaches do not seek or receive clinical treatment but rather learn to endure their headaches without complaint. One important factor increasing the likelihood that adolescents with frequent headaches also seek clinical treatment is degree of social disability as reflected by levels on a standardized measure such as the PedMIDAS [40] developed to assess social functioning among children and adolescents with migraine in a tertiary clinical sample. Here frequency of headaches was positively correlated to levels of disability, but less so for duration. While substantially lower disability levels have been observed in community samples of children and adolescents with frequent headaches [41, 42], those with a chronic migraine disorder have reported significantly higher disability levels.

We found that a sizable proportion of adults aged 26–28 years reported frequent headaches to the same extent as they did in early adolescence. This and other findings of short- and long-term follow-up studies of community samples of children and adolescents suggest that there is strong risk for continuation of frequent headaches commonly associated with higher levels of emotional problems and impairment in social activities. In the long-term perspective, they are likely to have developed into chronic tension-type or migraine headaches or a combination. Although the prevalence of headaches improves in a substantial proportion of adolescents as they move into adulthood, the persistence of frequent headaches, particularly among girls, along with higher levels of depressive symptoms and impairment, emphasizes the need to provide effective psychological and pharmacological treatments to reduce frequent headache complaints and their associated social burden.

In future longitudinal research, more frequent and repeated assessment over time will reveal whether changes in persistence, improvement rates, and relapses occur in adolescents with recurrent headaches. This information will improve the identification of individuals experiencing persistent frequent headaches over extended periods. More importantly, for these people, the potential influence of gender, presence of depressive symptoms, and extent of impairment should be tested in the context of controlled treatment trials to examine whether these factors also contribute to changing the outcome.