Abstract
We investigated 260 consecutive patients classified as migraine cases aged 3-69 at two tertiary headache centres, one for children and adolescents and the other for adults to evaluate the relationship between age and clinical features of migraine cross-sectionally. We only included subjects with definite migraine without or with aura and we excluded subjects with coexisting tension-type headache, medication overuse and/or other clinically relevant disorders. The percentage of males decreased markedly from childhood to adulthood and this affected the evalution of age-related changes in male patients, as only large differences reached the level of statistical significance. In females, the headache duration and the prevalence of unilateral, pulsating pain, photophobia and phonophobia increased, whereas the prevalence of aggravation by physical activity decreased with age. In conclusion, this cross-sectional, clinic-based study on a strictly defined sample of 260 consecutive patients with definite migraine covering a wide range of age from the very young to the old suggests marked age-related differences of the clinical features of migraine in females and failed to demonstrate similar differences in males due to the small number of adult male migraineurs.
Introduction
Migraine is a frequent disease with a prevalence of 3–10% in children and adolescents (1–4) and 10–18% in adults (5–10). It may occur at all ages, from the very young to the old, but it is most common in the 3rd to 5th decades. Some migraineurs have recurrent attacks without remission periods; others experience symptom-free intervals lasting several years; a third group becomes free of attacks for the rest of their life (11). Apart from that, the clinical features of migraine seem to differ with age. Migraine attacks in children and adolescents differ from those in adults, as the younger often have attacks of shorter duration and bilateral location. In addition, several other differences have been reported (2–4, 7, 8, 12, 13). Surprisingly, the available papers focus either on children and adolescents or on adults and none of the papers included subjects from all age groups. Therefore, we were interested to evaluate the symptomatology of migraine cross-sectionally in a clinic-based sample of migraineurs covering a wide range of age from the very young to the old.
Patients and methods
We examined 260 consecutive patients suffering from migraine with and/or without aura. The patients were seen at one of the two headache centres, i.e. the Department of Neuropsychiatry of Childhood and Adolescence (patients aged 2–18) and the Department of Neurology (patients older than 18 years) of the University Hospital of Vienna.
To record the headache history and symptomatology cross-sectionally, we used a semistructured interview published elsewhere in detail (12). Briefly, we asked for biographical data, duration of the headache history, the frequency of headaches, and the total number of headache attacks or episodes in the case of short history or low frequency headaches. Further, we recorded the duration, location, quality and intensity of the headache, the effects of physical activity, the occurrence of nausea, vomiting, photo- and phonophobia, and the occurrence and duration of aura symptoms. Using this questionnaire in 437 children and adolescents presenting to a headache centre (12) revealed that the IHS critieria of migraine are highly specific but less sensitive and that the intensity of pain, aggravation of headache by physical activity, nausea and vomiting are most important for differentiating migraine from tension-type headache. In the present paper, we modified the question about intensity by replacing ‘learning and playing’ with ‘working and leisure activities’ for interviews with adults. Otherwise the questionnaire remained unchanged.
Apart from the headache characteristics, we asked about head trauma, internal diseases and a series of other disorders to get information about a possible underlying organic disease. In children, we followed the child's report, provided this was not contradicted by the mother; if so we relied on the mother's report. All patients underwent obligatory neurological and optionally other clinical examinations. In some subjects, additional cranial computed tomography or magnetic resonance imaging were performed.
Headaches were classified according to the two-digit diagnosis of the International Headache Society (IHS) (14). One physician (ÇW-B.) was responsible for the final diagnosis, thus reducing interobserver variability. Patients diagnosed with migraine (IHS 1.1, 1.2) were included, while those with migrainous disorder (IHS 1.7) as well as those with coexisting tension-type headache, current or previous medication overuse and/or clinically relevant other disorders were excluded.
The SAS © system was used for calculations and statistical analyses. Continuous variables are reported as medians, lower and upper quartiles. Wilcoxon 2-sample tests were used to compare men and women regarding age, duration of headache history, migraine frequency and headache duration. The prevalence of the pain characteristics was compared using Chi-squared tests. Spearman correlation coefficients were calculated for analysis of the association between age and continuous variables. The relations of age with pain characteristics and with associated symptoms were analysed with univariate logistic regression analyses. P-values < 0.05 were considered as statistically significant.
Results
Patient characteristics
Of 260 patients with migraine, 169 were female and 91 were male. The mean age was 20.7 ± 15.4 years (Interquartile range (IQR) 10–28; median 14). Females were significantly older than males (P = 0.0001). Details are given in Table 1.
Patient characteristics, duration of headache history, and headache characteristics in 260 migraine patients
IQR, Interquartile range.
Headache characteristics
Considering the different age of females and males in this cross-sectional study, we analysed the headache characteristics not only in the entire group of patients, but also separately for both sexes and we looked for differences related to gender (Table 1, Fig. 1). The headache history was longer in females than in males (P = 0.002) and showed a close relation with age within the entire group and in both females and males (all patients, r = 0.72, p = 0.0001; females, r = 0.83, P= 0.0001; males, r = 0.35, P= 0.0006).
Headache characteristics separated by age in 169 females and 91 males with migraine. a, frequency; b, duration; c, pulsating quality; d, intensity; e, unilateral localisation; f, aggravation. □≤14 years (n= 147, 73 F); ░ 15–40 years (n= 75, 64 F); ▪>40 years (n= 38, 32 F). ∗∗P= 0.001; ∗P= 0.01.
The median of the days with migraine per week was 1.0 (IQR 0.25–2.0). The days with migraine increased with age in the entire group of patients (r = 0.17, P= 0.007), and it increased in males (r = 0.37, P= 0.0003), but not in females (r = 0.03, P= 0.7). There was a trend towards more attacks in females which was on the border of statistical significance (P = 0.0503).
The median of the headache duration was 6.0 h (IQR 4.0–24.0). The duration increased with age in all study groups (all patients, r = 0.51, P= 0.0001; females, r = 0.54, P= 0.0001; males, r = 0.38, P= 0.0002) and it was longer in females than in males (P = 0.0005).
The pain characteristics showed different relations to age in both females and males (Table 1, Fig. 1). In females, the prevalence of unilateral pulsating pain increased (P = 0.0004 and P= 0.005, resp.) and the prevalence of aggravation by physical acitivity decreased with age (P = 0.0001), whereas, in males, these parameters were not related to age. The association between age and pain intensity could not be assessed since only 4 males and 6 females reported mild pain. Comparing females and males regarding the prevalence of the four pain characteristics showed that unilateral pulsating pain occurred more often in females (P = 0.01 and P= 0.0001, respectively), whereas intensity and aggravation by physical activity did not differ between females and males.
The associated symptoms were not related to age in males and showed some age-associated differences in females, i.e. an increase in photo- and phonophobia with increasing age (P = 0.002 and P= 0.0001, respectively). Nausea and vomiting showed no consistent differences associated with age (Table 1, Fig. 2).
Prevalence of (a) nausea, (b) vomiting, (c) photophobia, (d) phonophobia, (e) both photo- and phonophobia, (f) aura symptoms separated by age in 169 females and 91 males with migraine. □≤14 years (n= 147, 73 F); ░ 15–40 years (n= 75, 64 F); ▪>40 years (n= 38, 32 F). ∗∗P= 0.001; ∗P= 0.01.
Aura symptoms were found most often among subjects aged 15–40 years (Table 1, Fig. 2), and less often among patients younger and older, but the relation to age was not statistically significant. Similarly, the prevalence of aura symptoms did not differ in males and females.
Discussion
At two headache centres – one for children and adolescents and the other for adults – we evaluated 260 consecutive patients cross-sectionally for their symptomatology of migraine. We included only patients fulfilling the IHS criteria for migraine without or with aura and we excluded subjects with coexisting tension-type headache, current medication overuse and/or clinically relevant other disorders in order to obtain a sample of migraineurs as homogeneous as possible and to diminish a bias due to coexisting disorders which can be expected more often in adults than in children. A similar protocol has been used previously in a study of tension-type headache (15).
As expected from epidemiological studies (1–10), the proportion of males decreased in our study from 50% in children to 15% in adults. Accordingly, the males where significantly younger than the females and therefore, the data were analysed separately for these two gender groups. The small number of males among adult migraineurs affected the evalution of age-related changes in male patients, as only large differences reached the level of statistical significance.
The frequency of migraine attacks increased with age (in the entire group of patients and in males), but from a clinical point of view it seems even more important to stress that more than 75% of the children with migraine had one or more attacks per month. In an epidemiological study, 40% of the subjects had at least one attack per month (2).
The duration of migraine attacks increased with age. The present study is the first one demonstrating this well known observation within one study using a well defined population and a homogeneous study design. Up to now, this increase could only be demonstrated by comparing different studies dealing either with children and adolescents or with adults, or by extrapolating findings of longitudinal studies in young patients to adults (16–18).
Concerning the pain characteristics, the lack of age-related changes in male patients might be explained by the small number of adult male patients in our study. The evolution from bilateral nonpulsating pain to unilateral pulsating pain with increasing age is also a well-known observation derived from studies in children and adolescents compared with those in adults (6, 12), and it was confirmed within this cross-sectional study by including subjects covering a wide range of age.
The finding that the prevalence of photo- and phonophobia was significantly lower in children than in adults might indicate a true change in associated symptoms. We cannot exclude, however, that children (and their parents) underreport less severe associated symptoms such as photo- and phonophobia. The peak of aura symptoms in the group aged 15–40 years might be caused by a selection bias. The first occurrence of visual, sensory and/or speech disturbances probably provides a strong motivation to visit a headache centre.
Our study is limited by the fact that it was performed in a clinic-based population. However, physicians are confronted with these patients and not with those who do not go to the doctor. Therefore, studies in referred patients are relevant for daily clinical practice, even though they do not provide information which is generally valid, as many migraine patients never consult a physician (9). A second limitation lies in the fact that the study was performed cross-sectionally and not longitudinally. To compensate for this as far as possible, we applied strict inclusion and exclusion criteria in order to diminish a bias due to confounding coexisting disorders which can be expected more often in adults than in children. The restriction to patients with definite migraine and the exclusion of subjects with migrainous headache and/or coexisting tension-type headache was necessary, since different types of headache occurring in the same patient cannot be evaluated cross-sectionally, but require longitudinal studies using headache diaries (19). Accordingly, the results of our study cannot be generalized to the whole population of patients with migraine.
In conclusion, this cross-sectional, clinic-based study in a strictly defined sample of 260 consecutive patients with definite migraine covering a wide range of age from the very young to the old suggests marked age-related changes of the clinical features of migraine in females and failed do demonstrate similar changes in males due to the small number of adult male migraineurs.