Abstract
The aim of the study was to examine the influence of concomitant neck pain (NP) on the outcome of headache (HA) frequency and HA type in adolescence. A population-based sample of 13-year-olds with or without HA (n = 228) was followed for 3 years. NP was evaluated at the beginning of the follow-up on the basis of recorded muscle tenderness and self-reported symptoms. During the 3 years of follow-up, changes in both HA type and frequency were common. NP interfering with daily activities at the age of 13 years predicted change from non-frequent (0-1/month) to monthly HA (>1/month), especially in boys (P = 0.03 boys, P = 0.06 girls). The use of physiotherapy predicted persistence of monthly HA in boys (P = 0.004). The changes in HA type were not predictable by NP. In conclusion, the risk of worsening HA in adolescence is more probable if the HA is associated with NP interfering with daily activities.
Introduction
In population-based studies both neck pain (NP) and headache (HA) have been found to be common in adolescents (1–7), and often concomitant (8–11). At the age of 7 years, tenderness in occipital muscle insertion areas has been more common in HA sufferers than in controls (10). At the age of 9–12 years, 62% of children with weekly NP have also reported weekly HA (11). In 10-year-old children NP has been found associated with migraine (9), and at the age of 13–16 years the frequency of neck pain has correlated with the frequency of HA (8). At the age of 18–27 years, pericranial muscle tenderness has been associated with frequent HA, but not with HA type (12). Muscular factors have also been associated with HA in adults (13–20). In our previous study of children aged 12 and 13 years, muscle tenderness at palpation of neck muscles was associated with migraine and self-reported NP was associated with episodic tension-type HA and migraine (21, 22). However, it is not known if the appearance of NP is of prognostic importance for the outcome of HA.
A better understanding of the determinants of HA outcome in adolescence could provide tools for the prevention of development of chronic HA in adults. In our previous report, significant predictors for the onset or persistence of monthly HA (>1/month) were female gender, frequent use of analgesics, consistent migraine and high basic education of one parent (23). Other studies have found associations of gender, age at onset of HA, anxiety, depressive disorders and maternal HA with the outcome of childhood or adolescent HA (1, 5, 24–30)
The objective of the present study was to examine the possible influence of neck pain on the outcome of HA in adolescence.
Subjects and methods
The present study is part of a population-based follow-up study of HA in schoolchildren (Figure 1). The original source population consisted of all 12-year-old schoolchildren in the city of Turku (total population of 175 000), south-western Finland. The questionnaire on HA was acceptably completed by 1135 (81%) of the 1409 eligible children (21).
Flow chart of study population.
At the age of 13 years, 70 children were randomly selected by computer from each of the following HA groups: migraine (IHS 1.1–1.2), tension-type HA (IHS 2.1–2.2), migrainous disorder (IHS 1.7) and healthy controls. In addition, all 47 children with tension-type HA not fulfilling the criteria (IHS 2.3) were included in the study (31). If a child was unable to participate (24% of children), another child of the same sex, randomly selected in advance, was recruited. The total number of participants was 311; their HA type was reclassified in a face-to-face interview and clinical examination between 12 October 1998 and the end of February 1999. The details of the study design have been published previously (21, 22). As part of the clinical examination, manual palpation, measurement of the pressure pain threshold of neck muscles, and stomatognathic examination were carried out. All examiners were blinded to the participants' pain history. Participants reporting HA at least once a month at the interview were given a headache diary for making notes over the following 2 months. Sixty-four per cent of the diaries (98/154) were returned acceptably completed. All participants with HA occurring at least twice a month according to the HA diary (n = 43) were offered participation in relaxation training given by a nurse in the Department of Child Neurology. Nineteen (15 girls and 4 boys) children participated in relaxation training.
At the age of 16 years, from the beginning of February to the end of April 2002, all the 311 adolescents were invited and 228 (73%; 82% of girls and 65% of boys) participated in a clinical re-examination. The number of dropouts varied from 20 to 27% in the HA groups and was non-significantly higher (P = 0.31) in the healthy controls (31%).
The outcome of HA was measured using two end variables, outcome of HA frequency and outcome of HA type. In analysing the outcome of HA frequency, improved HA (decrease in headache frequency from >1/month to 0–1/month) was compared with unchanged monthly HA (>1/month both at the age of 13 and 16 years). Worsened HA (increase in frequency from 0 to 1/month to >1/month) was compared with unchanged non-frequent HA (0–1/month). The cut-off point of HA frequency (>1/month, ≤1/month) was chosen partly because of clinical relevance and partly because of statistical needs. The numbers of adolescents in each group and the detailed outcome of HA frequency were presented in our previous article (23). The mean HA frequency at the beginning of follow-up was 4.5 times a month (SD 4.2) for monthly HA and 0.39 times a month (SD 0.38) for non-frequent HA. Half (49%) of the 13-year-olds with monthly HA, showed improvement of HA to non-frequent HA. Of adolescents with non-frequent HA, more than one-quarter (28%) showed worsening of HA to monthly HA (23).
In analysing the outcome of headache type, adolescents with two or more types of HA at the age of 13 or 16 years (n = 19) were classified according to their most frequent HA type. Tension-type HA (IHS 2) evolving to migraine (IHS 1) was compared with unchanged tension-type HA and migraine evolving to tension-type HA was compared with unchanged migraine. The classification of the HA type was based on a neurological examination including a structured interview and the criteria of the Committee of the International Headache Society (31). As the second, current edition of the International Classification of Headache Disorders (32) had not yet been published at the time of data collection, IHS 1988 criteria were used (31).
Self-reported neck symptoms and measured muscle tenderness were used as predictive NP factors. The questionnaire included frequency of NP in the preceding 6 months (response choices: no, <1/month, 1–3/month, weekly or more often). Children were asked about the occurrence of tenderness or stiffness of the neck region during daily activities (response choices: never, sometimes, often), and about the use of physiotherapy for NP or HA during the 6 months before the first study visit (reply choices: yes/no). The use of physiotherapy prescribed before onset of the study was considered a predictive NP factor because it was thought to describe the intensity of NP at the beginning of the follow-up. Muscle tenderness was recorded by manual palpation and pressure dolorimeter. A trained physiotherapist carried out a structured manual palpation test on tenderness at seven neck sites (frontal and temporal muscles, the suboccipital muscle insertion, anterior aspect of C5-C7, origin of the supraspinatus muscle, midpoint of the upper border of the trapezius muscle, and insertion of the levator scapulae muscle) bilaterally, graded on a four-step scale (0 = no pain, 1 = no visible reaction but reporting mild pain, 2 = reporting pain and distorting the face, 3 = reporting considerable pain and withdrawing) (33). Accordingly, the total tenderness score could vary between 0 and 42 (22). A dolorimeter (34) was used for the measurement of the pressure pain threshold from five pericranial/neck sites bilaterally (the frontal and temporal muscle, suboccipital muscle insertion, midpoint of the upper boarder of the trapezius muscle, and insertion of the levator scapulae muscle). Pressure pain was measured in kg/cm2 units (22). The mean dolorimeter scores for the pressure pain threshold were calculated. The higher the score, the less muscle tenderness.
Confounding factors
We wanted to exclude causation of the found association of NP with HA outcome by any other factor. As such possible confounding factors, signs of temporomandibular disorder, reported depressive and stress symptoms, relaxation training, leisure sport activities and working/playing with the computer were analysed. The predictive factors for the outcome of HA frequency found in our previous study (use of analgesics, basic educational level of parents, non-headache pain, change in headache type) were also analysed for exclusion as confounding factors (23).
In a stomatognathic examination carried out by a dentist, the severity of temporomandibular disorder (TMD) was graded using a TMD score. TMD score values increase from 0 to a maximum of 35 with severe symptoms. The TMD score included pain on palpation of the masticatory muscles or temporomandibular joints, pain on opening the mouth or on jaw movements, pain or stiffness on guiding the mandible, and joint sounds (35). DSM IV criteria for depression (36) were used to detect depressive symptoms (21). The following criterion features occurring almost every day for 2 weeks were used: depressed mood; loss of interest or pleasure; tearfulness; weight loss or gain, poor or increased appetite; insomnia or hypersomnia; psychomotor agitation or retardation; fatigue or loss of energy; feelings of worthlessness or excessive guilt; diminished ability to concentrate or indecisiveness; and recurrent thoughts of death or a suicide attempt. The sum score could vary between 0 and 10. Symptoms of school stress were adjusted by the frequency of the following situations/problems: fear of school, fear of teachers, fear of failure, exposure to bullying, loneliness, behavioural problems, learning difficulties, problems with other pupils, feeling of insufficiency and stress in school. Questions were asked about conflict situations at home or during participation in hobbies (choices: often, quite often, sometimes, hardly ever).
Participation in relaxation training (n = 19) at the age of 13 years for those with monthly headache was regarded as a confounding factor, because the inclusion criterion for the training was the frequency of HA (>1/month) and the training could have influenced NP. This relaxation training included general information on headaches and five 30-min, nurse-assisted relaxation sessions including muscle contraction relaxation and relaxation with visualization. Leisure sport activities were adjusted by the time the children spent in sports activities outside school hours so intensively that they were breathless or sweating (choices: not at all, half an hour, 1 h, 2–3 h, 4–6 h, ≥7 h a week). The children were also asked about the number of days per week spent on the Internet or in playing computer games.
Statistical analysis
The descriptive values were expressed as medians and interquartile ranges, or frequencies and percentages. Associations between categorical variables were evaluated using the chi-square test or Fisher's exact test. Differences in continuous variables between the HA groups were calculated using the Mann–Whitney U-test. The association of predictive NP factors (recorded muscle tenderness, NP frequency, NP in daily activities, use of physiotherapy) and confounding factors (relaxation training, temporomandibular disorder, leisure sport activities, using the Internet or playing computer games, stress and depressive symptoms, use of analgesics, basic education of parents, non-headache pain, change in HA type) with the outcome of HA frequency and HA type was first analysed for both genders together. Interactions between predictive NP factors and gender or HA type were analysed with the logistic regression (37) outcome of HA frequency as a dependent variable. As gender was associated with the outcome of HA, the associations of predictive NP and confounding factors with the outcome of HA were further analysed separately for girls and boys. With genders analysed separately, interactions between NP during daily activities and confounding factors on the outcome of non-frequent HA were analysed using logistic regression. P-values <0.05 were considered statistically significant. Statistical computations were done using the SAS System for Windows, release 8.02 (SAS Institute, Cary, NC, USA).
The study design and the informed consent procedures were approved by the Joint Ethics Review Committee of the Turku University Medical Faculty and the Turku University Central Hospital.
Results
Outcome of headache type
During follow-up, the HA type changed in 65% of participants (Table 1). In migrainous (IHS 1) children, HA remained as migraine, evolved to tension-type HA, or disappeared in 43%, 50% and 7% of the girls, and in 37%, 37% and 26% of the boys, respectively. The difference in changes was not significant between the genders (P = 0.06). In children with tension-type HA (IHS 2), HA remained as tension-type HA, evolved to migraine, or discontinued at the follow-up in 54%, 33% and 13% of the girls, and in 70.5%, 8% and 21.5% of the boys, respectively. The difference in changes was significant between the genders (P = 0.008). Of the nine adolescents with two or more types of HA at the age of 13 years, one became HA-free, three had only migraine, and five had only tension-type HA at the age of 16 years.
Changes in headache types (IHS 1988) from 13 to 16 years of age
Migraine (= IHS 1.1–1.2), MD = migrainous disorder (IHS 1.7), TT not crit = headache of tension-type not fulfilling the criteria (IHS 2.3), TTHA = tension-type headache (IHS 2.1–2.2), Sec = headache associated with infection (IHS 9), NC = headache not classifiable (IHS 13), No HA = no headache in the preceding 6 months.
Neck pain and outcome of HA frequency
At the beginning of follow-up, monthly HA was associated with a high frequency of NP (P = 0.04) and NP during daily activities (P = 0.003). When analysing the influence of predictive NP factors on the outcome of HA frequency for both genders together, NP was not associated with the outcome of monthly HA (Table 2) or with the outcome of non-frequent HA (Table 3). A high frequency of NP during daily activities was associated with worsened HA in boys (P = 0.03), but not in girls (P = 0.06) (Table 4). When analysing interactions of predictive NP factors and gender or HA type with the outcome of HA frequency, the only significant interaction was found to occur between the use of physiotherapy and gender (P = 0.02) with the outcome of monthly HA. Unchanged monthly HA was associated with use of physiotherapy in boys (P = 0.004) (Table 5). The confounding factors (temporomandibular disorder, reported depressive and stress symptoms, relaxation training, leisure sport activities and using the Internet or playing computer games) were not associated with the outcome of HA frequency when the genders were analysed together. Frequent use of computers was associated with unchanged monthly HA only in boys (P = 0.04). When interactions of NP during daily activities and confounding factors with the outcome of non-frequent HA were analysed separately for boys and girls, no significant interactions were found.
Univariate analysis of studied factors and outcome of monthly headache from 13 to 16 years of age
Univariate analysis of studied factors and outcome of non-frequent headache from 13 to 16 years of age
Univariate analysis of neck pain in daily activities and outcome of non-frequent headache by gender from 13 to 16 years of age
Univariate analysis of physiotherapy and outcome of monthly headache by gender from 13 to 16 years of age
Neck pain and the outcome of HA type
Reported or measured NP was not associated with the persistence of tension-type HA or persistence of migraine during 3 years follow-up. Gender was associated with the outcome of tension-type HA: tension-type HA evolved to migraine more often in girls than in boys (P = 0.004). Of confounding factors when the genders were analysed together, temporomandibular disorder, reported depressive and stress symptoms, relaxation training, leisure sport activities and using the Internet or playing computer games were not associated with the outcome of HA type. Temporomandibular disorder was associated with unchanged tension-type HA in boys (P = 0.03) and the use of analgesics was associated with unchanged migraine when the genders were analysed together (P = 0.02) and in girls (P = 0.02).
Discussion
Both the frequency and type of headache (HA) change substantially during adolescence. The determinants of the outcome of HA in adolescence are not known, although they would be important for the prediction and prevention of chronicization of the HA problem. Our present interest was in the predictive role of concomitant neck pain (NP) in adolescent HA. This role was suspected as HA and NP often co-occur (8, 9, 11, 12) and have similar epidemiological features: increasing prevalence with age in girls (4, 6–8, 38, 39) and recent increases in prevalence (3, 38, 40).
Our study showed that self-reported NP during daily activities predicted worsening of HA frequency and that a history of physiotherapy was associated with the persistence of monthly HA in boys. This does not mean that the use of physiotherapy leads to persistent HA but it is a sign of a more serious or interfering NP and HA problem. Adolescents with more disabling neck pain use more physiotherapy. NP interfering with daily activities is also probably more intensive, more consistent or more disabling than NP with a similar frequency but not experienced as interfering with daily activities. The intensity of NP could be a more important indicator of the sensitization process of pain than the frequency of NP alone. The importance of the intensity of NP is supported by the finding that the frequency of NP was not associated with the outcome of HA. As measured muscle tenderness had no predictive value for the outcome of HA, we could assume that muscle pain in this age group is a temporary phenomenon with no permanent structural changes in neck muscles.
NP interfering with daily activities and the use of physiotherapy were significantly associated with the outcome of HA only in boys. This gender difference could not be explained by interactions between NP and confounding factors. The difference between genders could partly depend on possible differences in the consistency of NP between genders in this age group. In preadolescents, however, NP has been more permanent in girls than in boys (11). The small number of girls may have limited the reliability of the analysis of the association of the use of physiotherapy and the outcome of HA in girls.
The co-occurrence of HA and NP could suggest either common pathogenesis, a causal association, or a common confounding factor. Two main hypotheses of the possible causal association of NP and HA have been presented: pericranial tenderness with HA provides a peripheral focus or cause or it is a manifestation of a central mechanism. Both these theories might explain the predictive role of concomitant NP.
Firstly, NP has been regarded as a peripheral cause or trigger for HA or for HA becoming more chronic (16, 17, 19, 20, 41, 42). Epidemiological studies of adolescents support this hypothesis, because NP has been shown to be the most permanent pain syndrome in adolescence, with a high tendency to convert to widespread pain on follow-up and also influencing NP in early adulthood (11, 43–45). Episodic tension-type HA has been regarded as a multifactorial disorder with several pathophysiological mechanisms, with extracranial myofascial nociception constituting one of them. In more chronic tension-type HA, secondary, segmental central sensitization and/or impaired supraspinal modulation of incoming stimuli seems to be involved, and prolonged nociceptive stimuli from muscles in the neck-shoulder region could be important for the conversion of episodic HA into chronic HA (19, 46).
In the second hypothesis of the association of HA and NP, HA has been thought to precede NP. NP has been found to be a common part of migraine attacks (47), improving with the treatment of HA. NP could partly be a consequence of pain extended from the head to other parts of the body also in other types of HA. Central neuroplastic changes in HA could affect the regulation of peripheral mechanisms and lead to increased pericranial muscle activity (46). Both hypotheses may be partly true, and NP could be a peripheral cause of HA, but also a manifestation of central mechanisms. This theory is supported by many studies showing that nociceptive afferents from the meninges and upper three cervical nerves from cervical structures converge to the same second-order neurones in the trigeminovascular complex (41, 48, 49) and inputs both from the periphery and from central nerve structures may modify the experience of pain. The two hypotheses of the possible causal association of NP and HA can only be ascertained using a prospective follow-up study design.
Factors that we know to have an independent predictive role in the outcome of HA (use of analgesics, consistent migraine, basic educational level of parents and non-headache pain) (23) did not explain the independent role of NP. The signs of temporomandibular disorder, reported depressive and stress symptoms, relaxation training, leisure sport activities and using the Internet or playing on the computer, did not interfere with the association between NP and the outcome of HA.
In this population-based study, the proportion of chronic and severe HA sufferers was small and our results can not be directly applied to patients suffering from chronic HA. One reason for the high number of changes in HA type could be the difficulty of accurate diagnosis of adolescent HA according to the applied criteria of IHS (31). Using the IHS criteria of 1988 (31) instead of current ICHD criteria (32) did not influence the results, because the one-digit criteria for migraine and tension-type HA are identical in these two versions. If adolescents have two different kinds of HA, they probably recall the most frequent one at an interview, and the nature of other HA types remains obscure without a HA diary (50).
Conclusion
Substantial changes occur in the frequency and type of headache in adolescence. The risk of worsening HA is more probable if the HA is associated with NP interfering with daily activities.
Acknowledgements
This study was supported by grants from the city of Turku, University of Turku and University Hospital of Turku. Olli Kaleva BSc is appreciated for skilful computation of the statistical analyses and Mrs Inger Vaihinen for practical assistance and for being quizzical.