ISSN 0017-8748

doi: 10.1111/j.1526-4610.2012.02210.x
Published by Wiley Periodicals, Inc.

Headache
2012 American Headache Society

Research Submission
Headache Disorders in Children and Adolescents:
Their Association With Psychological, Behavioral,
and Socio-Environmental Factors
head_2210

1387..1401

Birgit Krner-Herwig, PhD; Jennifer Gassmann, PhD

Objective.This cross-sectional study on a randomly drawn population sample of children and adolescents (n = 3399; aged
9 to 15) aimed at the assessment of patterns of associations between psychosocial variables and primary headache disorders like
migraine (MIG) or tension-type headache. A headache-free group served as a control.
Methods.Data on headache and psychological trait variables (eg, internalizing symptoms), behavioral factors (eg,
physical activities), and socio-environmental factors (eg, life events) were gathered by questionnaire. Logistic regression
analyses were conducted with headache types (MIG, tension-type, and non-classifiable headache) as dependent variables.
Results.The pattern of correlations was largely congruent between the headache disorders. Associations were closest
regarding maladaptive psychological traits (in particular internalizing symptoms with an odds ratio > 4 regarding MIG)
compared with socio-environmental factors and particularly the behavioral factors. Unfavorable psychological traits and
socio-environmental strains demonstrated distinctly stronger associations with MIG than tension-type headache and explained
more variance in the occurrence of pediatric headache disorders than parental headache. Sex-specific analyses showed similarities as well as differences regarding the correlations, and in general, the associations were stronger in girls than boys.
Conclusions.A common path model as posited by several researchers in the field may explain the parallelism in
biopsychosocial vulnerability regarding the different headache disorders.
Key words: headache disorder, psychosocial associations, traits, behavioral factors, socio-environmental factors, child
(Headache 2012;52:1387-1401)

A number of studies have reported links between

headaches in general and psychological dysfunctional
traits, behaviors and features of the social environment, without differentiating between the different
disorders.1 The question whether primary headache
types like migraine (MIG) or tension-type headache
From the Department of Clinical Psychology and Psychotherapy, Georg-Elias-Mller Institute of Psychology, University of Gttingen, Gttingen, Germany.
Address all correspondence to B. Krner-Herwig, PhD, MD,
Department of Clinical Psychology and Psychotherapy, GeorgElias-Mller Institute of Psychology, Georg-August University
Gttingen, Gosslerstr. 14, 37073 Gttingen, Germany, email:
bkroene@uni-goettingen.de
Accepted for publication June 5, 2012.

(TTH) differ in their pattern of associations, which is

the focus of this large-scale population-based study,
has only been examined in a few studies rendering
mixed results.
Gladstein and Holden2 analyzed 3 specifically
defined headache types (eg, transformed MIG) in a
small clinical sample (n = 33) in regard to internalizing and externalizing symptoms and found no differences between the groups.
Kawautz and colleagues3 assessed various psychosocial stressors in pediatric patients and
Conflict of Interest: The authors report no conflict of interest.
Funding: The research has been funded by a grant from the
German Federal Ministry of Education, Research and Science
within the German Headache Consortium.

1387

1388
headache-free controls. Their data showed TTH
patients to be subject to more stressors than MIG (eg,
divorce of parents) and headache-free controls,
whereas MIG-affected children did not differ from
headache-free individuals in this respect. The authors
did not assess psychological dysfunctions that are of
particular interest regarding their correlation to
headache disorders.
Anttila and colleagues4 studied children with
MIG and TTH in a population sample and found
higher levels of internalizing symptoms as well as
family and social problems in the MIG group than in
those without headache, but no difference to children
with TTH in this respect. They did not analyze
whether differential correlations regarding the disorders existed.
Mazzone and colleagues5 conducted a study on a
clinical sample of children with MIG or TTH and a
pain-free control group and compared them on
several psychological variables including internalizing and externalizing symptoms. They found differences in most variables assessed, not only regarding
the healthy control group, but also between headache
types. In general, TTH-afflicted children showed significantly higher scores in the psychosocial variables
than the MIG group, indicating a lesser degree of
psychological adjustment. Regression analyses, suitable for determining the explained variance in the
dependent variable (headache type) and comparing
headache disorders in this respect, were not conducted while this is our main objective.
The largely contradictory results and some methodological shortcomings of the earlier mentioned
studies motivated the current comprehensive study
aiming at the assessment and comparison of associations between different headache disorders and psychosocial factors. The selected variables were
grouped into 3 domains, the domain of (unadaptive)
psychological traits, habitual behaviors, and features
of the social environment. We examined by means of
regression analysis whether the psychosocial variables from these domains statistically predicted the
occurrence of different headache types with a
headache-free status being the reference. We wanted
to examine whether the pattern of associations was
specific to the type of headache or whether a common

October 2012
pattern of associations between headache and psychosocial variables emerged. The identification of
differences regarding certain features could have
important consequences for diagnosis and treatment
of pediatric patients: eg, if a particular proneness to
depressive or anxiety symptoms is found in a particular headache syndrome, it would suggest a comprehensive psychological diagnostic endeavor for the
afflicted individual child and also a multifaceted treatment approach.
Furthermore, we wanted to establish whether the
variables of the 3 psychosocial domains contribute
differently to the prediction of headache, ie, whether
psychological dysfunctions, adverse behaviors of the
individual, or stressful socio-environmental factors
contribute differently to the prediction of headache
syndromes.
From the wealth of psychosocial factors potentially being associated with headache, we selected
several variables on the basis of existing empirical
findings regarding their link with headache in general
and/or on the basis of reasonable hypotheses mainly
anchored in the stress concept.6 External or internal
factors that contribute to an emotional and physiological stress response were assumed to influence
recurrent headache or pain.

DOMAIN OF PSYCHOLOGICAL
(UNADAPTIVE) TRAITS
A large number of studies provided evidence
that depression/anxiety-like symptoms, labelled Internalizing Symptoms (Int-Symp) and Externalizing
Symptoms (Ext-Symp), a cluster of features of
hyperactivity, anger, and aggression, are associated
with headache in general, the latter with lessconsistent evidence, and other pain syndromes.1,4,7-10
A link between pain and Anxiety Sensitivity (AnxSens), indicating a disposition to feel threatened by
physical symptoms of arousal and anxiety, was shown
in adults.11-13 So far, however, this variable has not
been thoroughly studied in pediatric pain, although in
recent studies,14-17 an association of Anx-Sens with
headache was observed. Somatosensory Amplification
(Som-Amp),18 a tendency to attend to and emphasize
physical sensations, surprisingly was only examined in
pain research on adults,19,20 and not on children. Thus,

Headache
further exploration was needed. Self-Acceptance
(S-Accept), with a close resemblance to what other
authors denote as a positive self-concept or selfesteem, has been shown to be negatively correlated
with headache and recurrent abdominal pain.21-24 On
the basis of the stress coping model, Dysfunctional
Coping (DYS-COP) was assumed to function as a
predictor of headache, because unsuccessful coping
should lead to a failure in alleviating the impact of
stress. This assumption has received preliminary
support.25,26

DOMAIN OF BEHAVIORAL FACTORS

A low level of Physical Activities (Phys-Act)
potentially combined with too much Electronic
Media Utilization such as TV watching, etc (TV) is
commonly expected to be a risk factor for headache
in general.27 The associations have been found to be
inconsistent and even contradictory for various pain
syndromes, however,27-32 so further examination was
needed. We also assessed the amount of Free Time for
Leisure Activities (not pre-regulated or planned;
Leisure) mainly because our clinical impression of
children in treatment for headache indicated that
spending little or no time on self-regulated activities
could be a risk factor.7 Time needed for School
Homework (T-Home) was selected as a potential
stressor because homework can create a lot of strain
for children,30,31,33 especially within the German
school system.
DOMAIN OF SOCIO-ENVIRONMENTAL
FACTORS
We examined the variable Life Events (LE)
which has been shown to be associated with pediatric
headache and other recurrent pain in earlier
studies.34-43 School-Related Stress (Sch-Stress),
defined by problems with peers, teachers, and cognitive demands, was assumed to be a potent source of
strain in children and adolescents, and thus, an association with headache types was expected.21,44,45
Another stressor of assumed impact, Conflict in the
Family (Con-Fam), had been identified in some studies4,10,37,46 as a correlate of headache or other recurrent
pain. Regarding the family system, we were further
interested in the association of Dysfunctional Parent-

1389
ing (Dys-Parent), consisting of frequently applying
blame, relying on restrictive provisions, and being
inconsistent with headache.47 Financial Strain (FinStrain) in the family is expected to influence the interactional and affective climate within the family and to
exert an influence on the child in different aspects of
his/her social life and thus should be associated with
headache as it was suggested by earlier studies.46,48
It is known that parental headache is rather
closely connected to pediatric headache (eg, 23,49,50).
Hence, it was chosen as a biological reference variable for the evaluation of the relative predictive
power of the psychosocial factors of each domain.
Apart from expecting significant associations
between psychosocial variables with the headache
syndromes, we had no specific hypotheses regarding
the described main research questions.Thus, the study
was mainly exploratory.

MATERIALS AND METHODS

Participants.The study sample consisted of a
general population cohort of families (German citizens) with at least one child in the age range of 7 to 14
in the first wave of the survey. A sample of 8800
families residing in 4 districts of Southern Lower
Saxony was drawn randomly from the described subgroup whose data were registered in community files.
The study was conducted via a postal survey, and the
families were prepared for the questionnaires by an
introductory letter informing them of a research
project on the status of pediatric health, and headache in particular.
The survey comprised 4 annual waves, of which
wave 2 (2004) delivered data for the current study
(for further details see51). The sample analyzed consisted of those juvenile participants who had
answered the questionnaire on their own (9-15-yearolds) and for whom the parents questionnaires were
also available. It comprised 3399 respondents. About
one third (37.3%, n = 1267) of the children did not
report any headache during the last 6 months and
constituted the reference group (NoH). Sufficient
data to allow for the classification of headache disorders were available in 2132 responders. The number
of n differed between analyses, dependent on missing

1390
data in the specific variables analyzed. The mean age
of the children was 10.83 years (standard deviation = 1.98); boys constituted 49.4% of the sample.
The Ethics Committee of the German Society of
Psychology approved the study.
Headache Assessment and Classification.
Children were asked to report headaches experienced in the past 6 months. If they had experienced
any headache during this period, they had to
comment on further features like frequency, intensity,
and so on. Questions were conceptualized according
to International Classification of Headache Disorders, 2nd edition (ICHD-II)52 and analyzed to identify
MIG and TTH according to the rules of the diagnostic
system.53 Several studies have validated this procedure of disorder classification.49,54
Some 7.8% of respondents were classified as
suffering from MIG (n = 314), 25% from TTH
(n = 1010), and 20% could not be allocated to either
(non-classifiable headache; NCH: n = 808). The NCH
group was similar to the MIG group as most of these
children showed a type of headache that fulfilled the
MIG criteria except for one or two.52 Each headache
group comprised more girls than boys (see Table 1),
whereas the headache-free group included more boys
(56.3%).
Parental headache was assessed by self-report
with a question concerning occurrence in the last 6
months and the type of disorder.The parent who filled
in the parents questionnaire was also asked about the
headache of his/her partner. On the basis of sociodemographic data, we could decide whether the person
who filled in the questions and his/her partner were the
biological parents of the examined child. The variable
Parental Headache comprised 4 categories: the reference group Parent-NoH (no headache in either of the
parents), Parent-MIG (MIG at least in one parent),
Parent-TTH (TTH in at least one parent), and ParentMIG + TTH (both types of headache in at least one
parent), and other constellations.
Psychosocial Variables or Statistical Predictors:
Assessment Instruments.Domain of Psychological
Traits.Complete psychometric tests measuring the
psychological traits could not be adopted because of
the comprehensiveness of the questionnaire. A subset
of items was selected from validated instruments, and

October 2012
in most cases, item selection was based on the criterion largest item-scale correlation coefficients or
largest factor loadings. This criterion ensured that
those items were used that correlated most strongly
with the total score of the scale, thus warranting
high content validity and homogeneity of the short
scales. As expected, the correlations between the
short and the complete versions of tests were large
(0.74 r 0.96).55
Int-Symp and Ext-Symp were measured by
selected items from the Youth Self-Report (YSR56)
originally validated for youths between 11 and 18.
Int-Symp (example: I felt guilty) comprised 8 items.
Ext-Symp (example:I cannot sit still for a long time)
was assessed by 6 items. Contrary to the original form
(3-point rating scale), responses were scaled on a
5-point scale (<never>-<always>) to ensure a better
comparability with other scales and higher grade of
differentiation because less items were used than in
the original scales. Cronbachs alpha demonstrated
good homogeneity of the short scales: Int-Symp:
a = 0.86, Ext-Symp: a = 0.78. Five items (a = 0.71)
measured Anx-Sens57 (example: I am frightened
when my heart beats fast) and were also rated on a
5-point scale. The variable Som-Amp18 comprised 5
items (example: I am very sensitive to pain) and was
also rated on a 5-point scale (<never>-<very much>).
It was the only variable with a less-than-satisfactory a
(0.57). The variable S-Accept (example: How satisfied are you with your life in general?) was measured
by 6 items (a = 0.75) from the Child Health Questionnaire58 (5-point rating scale:<very satisfied>-<not satisfied at all>). DYS-COP was assessed by 5 items from
the Stress Coping Inventory59 (example: When I am
under stress. . . . I tend to pretend I am sick; 5-point
rating scale <never>-<always>) and showed good reliability (a = 0.76).
Domain of Behavioral Variables.They were
either measured by one item only or defined by the
number of items reaching a certain response criterion.
Phys-Act was determined by 11 items of the SelfAdministered Physical Activity Checklist60 (example:
to go swimming, with a 5-point rating scale: <less
than once/week>-<5 times/week>; score 2). The
variable TV was measured by the item: How long do
you spend watching TV/playing video or computer

Headache

1391
Table 1.Descriptive Data on Predictors, Outcome of Analysis of Variance (ANOVA) and Effect Size

Variables
Psychological Traits

Anxiety sensitivity (ANX-S)

Dysfunctional Coping (DYS-COP)
Externalizing symptoms (E-S)
Internalizing symptoms (Int-Symp)
Self-acceptance (S-Accept)
Somatosensory Amplification

ANOVA
(df = 3)/Chi-square
Effect Size(d):
F/Chi-square (P .01) Highest-Lowest

MIG
M (SD)

TTH
M (SD)

NCH
M (SD)

NoH
M (SD)

1.57 (0.59)
2.4 (0.76)
1.92 (0.56)
1.93 (0.60)
1.97 (0.72)
2.09 (0.70)

1.31 (0.41)
2.16 (0.70)
1.74 (0.54)
1.67 (0.50)
1.77 (0.64)
1.76 (0.56)

1.47 (0.51)
2.32 (0.73)
1.84 (0.54)
1.81 (0.57)
1.86 (0.66)
1.91 (0.59)

1.25 (0.39)
1.96 (0.69)
1.65 (0.52)
1.50 (0.45)
1.60 (0.56)
1.59 (0.51)

72.54
62.01
31.25
89.30
45.57
93.47

0.64/0.15
0.61/0.29
0.50/0.17
0.81/0.36
0.57/0.28
0.82/0.32

3.77 (0.80)
1.65 (0.72)
1.09 (0.44)
3.63 (1.15)

3.83 (0.71)
1.49 (0.65)
1.06 (0.44)
3.44 (1.0)

3.79 (0.72)
1.52 (0.65)
1.06 (0.40)
3.48 (1.1)

3.96 (0.67)
1.44 (0.61)
1.09 (0.45)
3.79 (0.72)

13.66
10.01
1.59 ns
3.03 ns

-0.19/-0.26
0.31/0.08
-0.070/0
-0.40/-0.17

4.02 (0.88)
2.11 (0.60)
24.2%
1.86 (1.91)
2.06 (0.47)
11.2
57%

4.11 (0.87)
1.95 (0.57)
19.2%
1.67 (1.75)
1.91 (0.43)
10.88
52.3%

4.08 (0.88)
2.01 (0.56)
23.8%
1.77 (1.75)
1.99 (0.43)
10.74
57.3%

4.26 (0.83)
1.90 (0.57)
18.2%
1.39 (1.57)
1.82 (0.44)
10.72
43.7%

15.70
13.76
Chi-square = 12.77 (3)
12.11
40.79
8.49
Chi-square = 48.18 (3)

-0.28/-0.18
0.36/0.09

0.27/0.17
0.53/0.21

16.1%
18.4%
32.1%
15.7%
17.7%
100%

17.9%
11.9%
38.7%
10.0%
21.5%
100%

Chi-square = 174.1(12)

Behavioral

Free time for leisure

Time for homework
Physical activities
Media use (TV/video, etc)
Social Environment

Conflict in family
Dysfunctional parenting
Financial strain (ref. = yes)
Life events
School stress
Age
Sex ( female)
Parental Headache

Parent-NoH
ParentMIG
Parent-TTH
Parent-MIG + TTH
Others

18.4%
13.1%
39.0%
9.4%
20.1%
100%

36.2%
10.5%
29.4%
5.6%
18.2%
100%

ANOVA: testing significance of main effect of headache type/chi-square testing unequal distributions; exact score are not given
because only the pattern of significant effects is of interest here and the following regression analyses give quantitative statistics on
these effects.
The higher the score, the less conflict.
= no such tests performed; df = degrees of freedom; M = mean; MIG = migraine; NCH = non-classifiable headache; NoH = no
headache; ns = nonsignificant; SD, standard deviation.

games each day?61 (6 response categories: <never>-<4

hours/day>). Leisure was assessed by the item: How
often do you have time to play or free time for yourself? on a 5-point rating scale: <never>-<always>THome was rated on a 4-point scale (<less than 1 hour><more than 3 hours> per day).

Domain of Socio-Environmental Variables.LE

was assessed using data from the parents questionnaires (waves 1 and 2) asking for the occurrence
of 8 critical LEs like loss of a family member, change
of school, etc. Five items assessing different aspects
of stress (problems with peers, teachers,

1392
performance) (a = 0.69) were taken from Karwautz
and colleagues3 5-point rating scale: <never>-;
<always>example: I get bullied or tormented by children at my school). The occurrence of Con-Fam was
assessed by the question How often has there been
quarrelling in your family in the last 3 months?;
5-point rating scale: <never>-<daily>Dys-Parent was
measured by 6 items [a = 0.73] from the Inventory of
Parenting Styles.62 The items represent parental
restriction, inconsistency, and blame (example: My
mother/father tells me I am a nuisance); 5-point rating
scale; <never>-<always>). One item from the Mannheimer Parent-Interview63 assessed Fin-Strain
(<no>or <yes>) by parental report.
Statistical Analysis.If an item of a scale showed a
missing value, it was replaced by the mean item score.
If equal to or more than 50% of the items of a scale
were not responded to, the variable was treated as
missing in the subjects dataset. Analysis of variance
(ANOVA) or chi-square tests were conducted to
examine differences between the headache disorders
in each of the predictor variables (testing the main
effect of headache type). Logistic regression analyses64
(hierarchical; control variables sex and age entered
first, predictor variables second) were conducted with
each of the headache types as criterion or dependent
variable. In each analysis, the NoH group was used as a
reference. In a first step, each of the predictors was
examined in a separate analysis controlling for sex and
age.A multiple regression analysis with all variables of
the specific domain followed. SPSS18 (IBM, Armonk,
NY, USA) was used in the statistical analyses, and the
level of significance was set at P .01 to compensate
for the large n of the study and reduce type 1 error.
Because the analysis was mainly exploratory, no Bonferroni correction was used.

RESULTS
Differences Between the Headache Disorders.
As shown by ANOVA, the mean age of the subjects
differed between headache disorders with the posthoc Scheffs test, indicating that children with MIG
were significantly older than children in the other
groups. A chi-square test showed a significant differential distribution of boys and girls in the 3 headache
groups, indicating that girls were more often affected

October 2012
by headache of any type, and particularly MIG (see
Table 1). Thus, the sociodemographic variables were,
as planned in advance, used as control variables in the
regression models with the psychosocial predictors.
Also, parental headache was analyzed and was
shown to be unevenly distributed in the 3 headache
categories as indicated by a significant chi-square.
ANOVA produced significant F-scores (Table 1;
descriptive data and significance of main effects)
regarding the factor headache disorder (3-level
factor) in all but the following variables: Phys-Act and
TV. Post-hoc Scheffs tests were conducted but are
not documented here because the individual test score
is not of interest but, instead, the general pattern of
differences. Furthermore, the odds ratios (ORs) of the
regression analyses (Table 2) represent the differences
between headache syndromes. Findings demonstrated
significant differences between MIG and TTH in all
cases except for 2 earlier mentioned variables, MIG
and NCH, which did not differ significantly in several
variables. In most variables, scores were ranked as
follows: MIG NCH > TTH > NoH (Table 1), indicating in general the highest level of dysfunction or
stress impact in MIG.
Regression Analyses.Age and Sex.In a first
step, separate regression analyses of the control variables age and sex regarding headache types were conducted for comparison with the ORs found in the
analysis with psychosocial factors. Only in the children with MIG did age significantly predict headache
(OR = 1.18, P < .001; confidence interval CI [95] =
1.11-1.26). As the results of ANOVA had shown
already, being female was a significant predictor
for all 3 types of headache (MIG: OR = 1.81; CI
[95] = 1.41-2.34/NCH: OR = 1.76; CI [95] = 1.47-2.10/
TTH: OR = 1.45; CI [95] = 1.23-1.71; all P < .001).
Psychological Trait Variables.Each single psychosocial variable was examined in a model with sex
and age included (Table 2). Each of the trait variables
(see Table 1 for descriptive data) was significantly
associated with each of the headache types (Table 2).
All ORs (each b 2) were largest regarding MIG,
closely followed by NCH. The largest OR (4.05) was
found for the variable Int-Symp predicting MIG. It
explained more than 16% of the variance of MIG
occurrence.

Headache

1393

Table 2.Odds Ratio (OR) and Confidence Interval (95%) for Each Headache Disorder and Each Single Predictor of All 3
Domains (Reference: No Headache [NoH]) and Parental Headache (Reference: No Headache in Either Parent [Parent-NoH])
When Controlling for Age and Sex

MIG

TTH

NCH

Domain 1: Psychological Traits

Anx-Sens*
3.66 (2.78-4.73)
1.37 (1.11-1.69)
DYS-COP
2.24 (1.88-2.68)
1.42 (1.26-1.60)
Ext-Symp
2.20 (1.80-2.87)
1.35 (1.15-1.59)
Int-Symp
4.05 (3.13-5.23)
1.96 (1.63-2.36)
S-Accept
2.41 (1.99-2.92)
1.59 (1.38-1.83)
Som-Amp
3.65 (2.92-4.56)
1.78 (1.51-2.09)
Max. explained variance in a regression model (m = 3 6): Nagelkerke = 16.6% (MIG; Int-Symp)
*All ORs significant : P .01

3.00 (2.41-3.74)
1.95 (1.71-2.24)
1.95 (1.64-2.31)
3.16 (2.60-3.86)
2.23 (1.88-2.64)
2.80 (2.35-3.33)

Domain 2: Behavioral Variables

Leisure*
1.38 (1.16-1.64)*
1.27 (1.13-1.44)
T-Home
1.54 (1.27-1.86)
n.s.
Phys-Act
n.s.
n.s.
TV
n.s.
n.s.
Max. explained variance in a regression model (m = 3 4): Nagelkerke = 2.5% (NCH; TV)
*All ORs significant : P .01

1.40 (1.23-1.60)
n.s.
n.s.
1.75 (1.46-2.09)

Domain 3: Socio-Environmental Factors

Con-Fam *
1.32 (1.15-1.52)
1.23 (1.11-1.35)
Dys-Parent
1.75 (1.41-2.16)
n.s.
LE
1.18 (1.10-1.26)
1.11 (1.06-1.70)
Fin-Strain
n.s.
n.s.
Sch-Stress
3.04 (2.30-4.03)
1.70 (1.40-2.06)
Max. explained variance by each model (m = 3 5): Nagelkerke = 10.4% (MIG; Sch-Stress)
*All significant OR: P .01

1.26 (1.14-1.40)
1.43 (1.22-1.85)
1.15 (1.09-1.22)
1.47 (1.24-1.74)
2.52 (2.03-3.11)

Parental Headache

Parent-MIG
3.85 (2.49-5.96)
2.92 (1.69-3.12)
Parent-TTH
2.50 (1.72-3.63)
2.66 (2.11-3.34)
Parent-MIG + TTH
6.53 (4.03-10.57)
3.63 (2.54-5.19)
Other constellations
2.19 (1.43-3.34)
2.40 (1.85-3.11)
Max. explained variance in a regression model (m = 3 6): Nagelkerke = 11.8% (MIG; Parent-MIG + TTH)
**All significant OR: P .01

2.43 (1.76-3.36)
2.65 (2.07-3.38)
3.33 (2.27-4.89)
2.17 (1.64-2.87)

Domain 1: Anx-Sens = anxiety sensitivity; DYS-COP = Dysfunctional Coping; Ext-Symp = externalizing symptoms; IntSymp = internalizing symptoms; S-Accept = self-acceptance; Som-Amp = somatosensory amplification. Domain 2: leisure = free
time for leisure; n.s. = nonsignificant; T-Home = time for homework; Phys-Act = physical activities; TV = time with TV/video/
computer. Domain 3: Con-Fam = conflict in family; Dys-Parent = dysfunctional parenting behavior; Fin-Strain = financial strain;
LE = life events; Sch-Stress = school stress. Headache Groups: MIG = migraine; NCH = non-classifiable headache; TTH = tensiontype headache. Parent-MIG = migraine in at least one parent; Parent-TTH = tension-type headache in at least one parent; ParentMIG + TTH = both types of headache in at least one parent.

1394

October 2012

In the multiple regression analysis with all variables of the domain entered (Table 3), Int-Symp,AnxSens,Som-Amp,and S-Accept remained significant,ie,
independent predictors of headache with a few exceptions (MIG:S-Accept = nonsignificant [ns]/TTH:AnxSens = ns). In contrast, DYS-COP and Ext-Symp
failed to maintain their predictive power in the context
of the other variables. The total variance explained by
all variables of the domain amounted to nearly 24% in
MIG, and about 18% in NCH; far less was explained in
TTH (6%).
Behavioral Variables.When the behavioral
variables were analyzed, less strong associations with
the headache disorders were found in general
(Table 2; see Table 1 for descriptive data). Leisure was
significantly related to all headache types, signifying
that less free time (reference category much free
time) contributed to the occurrence of headache of
any kind. Phys-Act did not function as a predictor
regarding any of the headache disorders. A differential association was found in 2 variables (TV, T-Home;
Table 2).
The multiple regression analysis of behavioral
variables (Table 3) consistently showed Leisure to be
a predictor of headache of any type. Specific associations were again seen for T-Home and TV. The
explained variance was rather minute, reaching a

maximum of 2.5% for the single predictors (Table 2)

and about 7% regarding the combined influence of all
predictors. As seen in the analysis of psychological
dysfunctions, variables of this domain explained most
variance in MIG.
Socio-Environmental
Variables.Regression
analyses of the socio-environmental variables (see
Table 1 for descriptive data) revealed a more homogeneous pattern (Table 2). Sch-Stress, Con-Fam, and
LE showed significant correlations with each of the
headache types, Dys-Parent with MIG and NCH only.
The regression with Sch-Stress produced by far the
highest ORs. The comprehensive analysis of socioenvironmental factors confirmed the significance of
Sch-Stress and LE for all headache types (Table 3)
and a specific association of Con-Fam with TTH. The
explained variance varied between 4% (TTH) and
nearly 12% (MIG), ie, the closest associations were
found regarding MIG.
Separate Regression Models for Girls and Boys.
We separately analyzed the subsamples of girls and
boys within the multiple regression models regarding
the occurrence of MIG only (Table 4), where we had
found the highest amount of variance explanation in
the total sample. Anx-Sens and Som-Amp turned out
to be predictors for both sexes in the psychological
trait domain. Int-Symp showed to be significant in

Table 3.Outcome of Comprehensive Regression Analyses With All Variables of Each Domain for Each Type of Headache:
Significant Variables Contributing to the Model (P .01)

Domains
Headache Type

Predictor variables

Explained variance
(Nagelkerke)

Psychological

Behavioral

Social Environment

MIG

TTH

NCH

MIG

TTH

NCH

MIG

TTH

NCH

Anx-Sens*
Int-Symp

Som-Amp

23.7%

Int-Symp
S-Accept
Som-Amp

6.2%

Anx-Sens
Int-Symp
S-Accept
Som-Amp
Age
Sex
18.2%

Leisure
T-Home
TV

Age
Sex
7.4%

Leisure

Leisure
T-Home

4.4%

LE
Sch-Stress

11.9%

Con-Fam
LE
Sch-Stress

4.1%

LE
Sch-Stress

Sex
8.8%

Sex
2.2%

*All variables noted: significant at P .01.

Domain 1: = not identified as a significant predictor; Anx-Sens = anxiety sensitivity; Int-Symp = internalizing symptoms;
S-Accept = self-acceptance; Som-Amp = somatosensory amplification. Domain 2: leisure = free time for leisure; T-Home = time for
homework; TV = time with TV/video/computer. Domain 3: Con-Fam = conflict in family; LE = life events; Sch-Stress = school stress.
Headache Groups: MIG = migraine; NCH = non-classifiable headache; TTH = tension-type headache.

Headache

1395

Table 4.Separate Analyses of Girls and Boys With Migraine Comprehensive Regression Analyses With Each Domain
of Predictors

Domain
Sex

Predictors*

Nagelkerke

Psychological

Behavioral

Socio-Environmental

Boys

Girls

Boys

Girls

Boys

Girls

Anx-Sens

Som-Amp
15.2%

Anx-Sens
Int-Symp
S-Accept
Som-Amp
29.3%

T-Home

8.2%

Leisure

TV

10.3%

Sch-Stress

8.4%

Sch-Stress
Dys-Parent

13.9%

*All predictors significant at P .01.

= not identified as a significant predictor; Anx-Sens = anxiety sensitivity; Int-Symp = internalizing symptoms; leisure = free time
for leisure; S-Accept = self-acceptance; Sch-Stress = school stress; Som-Amp = somatosensory amplification; T-Home = time for
homework; TV = media use.

girls only, as did S-Accept. The analysis of behavioral

variables produced a pattern of correlations with no
overlap regarding predictors for girls and boys (see
Table 4). Results were more homogeneous regarding
the socio-environmental factors, showing Sch-Stress
to be a significant predictor for MIG in both sexes.
For both sexes, the Nagelkerke indicated that the
highest amount of variance was explained by the psychological trait variables. For all domains, the amount
of explained variance was largest in girls (29.3%),
with the biggest difference to boys (15.2%) evidenced
in the psychological trait domain.
Parental Headache.We found ORs regarding
the association of parental and childrens headache
ranging beyond 2 to nearly 7 (Table 2). The relative
risk of being afflicted by any of the headache disorders was highest when a parent showed combined
headache, particularly in children with MIG (OR =
6.53). The correlation of TTH in parents with pediatric headache was about equally strong for every headache category, whereas Parent-MIG correlated more
closely with pediatric MIG than with any other
disorder.
The highest OR regarding parental headache was
larger than any from the domain of psychosocial variables. The Nagelkerke index, however, was lower for
parental headache (all Parent combinations included:
11.8%) than for the personality variable Int-Symp
(16.3%).

DISCUSSION
Headache Classification.The questionnairebased computerized classification of headache disorders led to plausible prevalence rates regarding MIG
(~8%) and TTH (25%), which are comparable with
those from other studies.65 This procedure, which had
been validated by interview data in earlier studies,54
can consequently be utilized when no clinical decision
is associated with the classification.
The Prediction of Headache Disorders.Domain
of Psychological Traits.To our main question regarding the specificity of the pattern of psychosocial factors
associated with the different disorders, we found a
rather clear answer regarding the psychological trait
variables: ORs in the single predictor analyses signal
unanimously positive associations between elevated
psychological scores characterizing maladaptive functioning and the occurrence of headache regardless of
type of disorder. An increase in the dysfunctional psychological trait level increased the probability of children being afflicted by one of the 3 headache disorders
by at least 35% up to nearly 300% depending on the
predictor and the particular disorder.
There are, however, differences in the size of
the correlations. They all showed their most prominent magnitude in regard to MIG and less distinct
links with TTH. The most powerful predictor for
MIG turned out to be Int-Symp. They are also highly
predictive of the NCH group in which a large

1396
proportion of children reported MIG symptoms but
did not fulfill all the criteria.These results partly agree
with those reported by Anttila,66 who found significant differences in Int-Symp between MIG and a
no-headache group but not TTH in their populationbased study. A recent study27 confirmed our finding of
a higher level of psychological dysfunctions in MIG
compared with TTH-afflicted subjects also assessed in
a population-based study.
They contrast in some respects with the findings
of Mazzone and colleagues,5 who also found IntSymp to be higher in MIG than in a no-headache
group, but with significantly higher scores in TTH.
The latter authors attribute this to the special selection of their clinical sample (tertiary care), with the
TTH group being troubled by additional psychosomatic symptoms, not to be expected in headacheafflicted children from the general population. We are
inclined to interpret our results in agreement with
Mazzone and colleagues in that the level of overall
illness severity can be assumed to be a moderator
of the strength of the association between psychological variables and the headache disorder. MIG as the
condition with highest illness severity defined by
disability in our sample, as data presented elsewhere
show,50,52 has the strongest links with the dysfunctional psychological traits.
Our finding that the variable Int-Symp is the
strongest predictor for each of the headache disorders is in line with results of numerous other studies.
Significant correlations of this variable with headache
in general and also with other functional pains were
consistently found (see review,1 meta-analysis67). IntSymp, as a genetically influenced, early observable
temperamental characteristic also represents a predisposition for the manifestation of psychological disorders in later life.68 Thus, the variable marks a
susceptibility to somatic disorders like headaches and
other functional pain syndromes as well as mental
disorders.69
Also, the variables Anx-Sens and Som-Amp
retained their predictive power in the multiple regression analysis, which showed their specific and independent contribution to the prediction of headache
disorders. They are of special interest as they seem to
mark a transition between somatic and cognitive-

October 2012
emotional processing. They denote the process of
responding to bodily sensations in a hypersensitive
way and endowing them with a negative appraisal.
Thus, together with the dispositional negative affectivity, as an integral part of the Internalizing Syndrome, which may also be associated with lack of
Self-Accept, they link biological and psychological
functioning.
Behavioral Characteristics.The associations of
behavioral factors with the headache disorders were
clearly less close than those of psychological traits,
explaining only between 2 and 7% of the variance in
the comprehensive regression analysis, compared
with the 8-24% of the psychological traits.The pattern
of associations between these variables and the headache disorders was also less consistent over headache
types. This may be the case because the correlations
overall were weak and statistically not very robust.
A single interesting observation was that the variable Leisure showed the only consistent significant
correlation with all headache types, signifying that
children who have no or little time for self-regulated
leisure activities were at considerable risk of experiencing headache, irrespective of its syndromatic features. This finding has recently been corroborated.27
The amount of Phys-Act and the amount of
Media Use, representing passive activities exerted
little or no influence on headache, including TTH, was
assumed to be particularly susceptible to this behavioral pattern. Findings suggest that the analyzed
behavioral dispositions are of little importance for
headache, which had been already shown in a few
epidemiological studies.28,69
Socio-Environmental Variables.In general, the
variables of this domain showed closer links to the
different headache types than the behavioral factors,
although less close than the psychological trait variables. There was much congruence between the different types of headache regarding the associations.
Sch-Stress was the strongest predictor for all 3 types,
supporting the findings of other studies.21 Con-Fam70
and the number of LE,34 all reflecting stressful
changes in social relationships, were consistently
associated with each headache disorder.
As in the other domains, MIG showed the closest
links with socio-environmental factors: about 12% of

Headache
the variance in MIG occurrence was explained by
them in the multiple regression analysis. This result is
again compatible with the hypothesis that the size of
the covariation with psychosocial factors is moderated by the level of illness severity.
Sex Differences in the Prediction of MIG by Psychosocial Variables.In both girls and boys, we found
3 identical predictors (psychological traits: Anx-Sens,
Som-Amp, socio-environmental: Sch-Stress). The
trait variable Int-Symp, representing the most
common predisposition of psychological maladaptation in girls, reached significance only in the female
sample.The factor Ext-Symp as a trait, predominantly
predisposing for psychological disorders in boys,69
however, did not reach significance in either group,
contradicting findings of some earlier studies14,18
which, however, did not differentiate between sexes.
In the domain of behavioral variables, we found no
common variables correlating with the headache
disorders.
It is evident that the associations of MIG with
psychosocial variables are more intimate in girls
(reaching nearly 30% of explained variance in the
domain of psychological traits) than in boys. Hence,
the associative net between psychosocial factors and
headache is more densely knit in females.
Parental Headaches as Predictors of Pediatric
Headache.The analysis of parental headache was
conducted mainly for purposes of comparison with
the psychosocial variables. The occurrence of combined headache in at least one parent, presumably
suggesting a strong unfavorable genetic burden, constituted the largest single risk for pediatric MIG as
the most disabling headache type. Although most
single ORs were higher for Parental Headache
parameters (2.2-6.6), the total explained variance was
lower compared with the psychological domain (11.8
vs 23.7%), presumably based on the relatively large
CIs. The ORs found in our study compare very well
with those reported by Aromaa et al46 (OR = 3.5) and
Bener and colleagues49 (OR = 3.35), who used a
rather non-specific predictor (a family history of
headache: yes/no) in their studies of young children.
Integration of Study Findings Into the Body of
Empirical Evidence and Theoretical Assumptions.
The main findings of our study show

1397
1. That the pattern of associations with psychosocial
factors was quite congruent between the headache
disorders,
2. That psychological trait variables explained the
largest amount of variance. These were characterized by negative affect and also hypersensitivity to
somatosensory stimuli, and
3. That the associations were, in general, strongest
with the MIG type of headache.
These findings shall be evaluated in the context of
further empirical findings on pediatric pain. Some
recent epidemiological data have weakened the
assumption of the predominant importance of a specific pathological background even for phenotypically
different functional pains like back pain or headache.
It has been shown that more children and adolescents
are characterized by multiple recurrent pains than by
pain at a single site,7 which we also observed in the
epidemiological sample analyzed here.71 Thus, pains
tend to occur together, a fact that suggests some
common background. It also had been suggested long
ago that headache disorders like MIG and TTHs
should not be considered as different clinical entities,
but might differ mainly on the dimension of severity
of the disorder.72
This general idea of a common background was
lately advanced by Kato and colleagues.73 In their
recent paper on a population-based twin study, they
pleaded for a so-called common path model. Their
empirical findings suggest that probably 2 paths lead
to the development of functional pain, explicitly
including headache disorders, one characterized by
sensory dysfunctions like hypersensitivity of the
central stimuli processing system and a second path
characterized by dysfunctional affective processing.
In our study, we recognized a disposition for hypersensitivity especially in the traits Anx-Sens and SomAmp (including affective components), which were
associated with all headache disorders. Negative
effect was best represented by the Int-Symp score, a
variable also closely linked with all headache disorders. Kato and colleagues73 found the latter path also
to be strongly associated with the occurrence of manifest psychological disorders. Both background factors
are considered to be influenced by shared genetic and

1398
environmental factors. Sex differences, especially
regarding Int-Symp being predictive of headache
only in girls, may be explained by the special significance of the affective path in females.
This model integrates our findings regarding the
congruence and character of psychological traits associated with the different headache disorders. It also
explains the co-occurrence of psychological and
somatic dysfunctions, which were even found in this
population sample where the majority of children
presented somatic and psychological dysfunctions
outside a pathological range.70 The closeness of bonds
between psychological dysfunctions and MIG is best
accounted for by the severity of the condition.
Limitations.A limiting factor especially regarding the psychological variables is that we could not
include the complete test versions but had to reduce
the number of items. The high intercorrelations
between the partial and the complete tests, and in
most cases, high reliability, mitigate this weakness,
however. The deviation from the original format of
the assessment instruments (eg, YSR) also has the
consequence that the grade of abnormality of
symptom scores cannot be evaluated.
Furthermore, the headache classification was not
based on professional clinical judgment, and criteria
B and F (see ICHD-II) could not be checked. A clinical interview, giving the chance to repeat or reformulate questions, might have reduced the number of
children with non-classified headache.
Regarding the analysis of the influence of parental headache, our data were incomplete as the combination of headache diagnoses of both parents could
not be used. Also, diagnoses were only assessed by
self-report and not by a clinical interview.
In addition, there is some arbitrary voluntary
selection involved regarding the variables chosen for
analysis. Despite the fact that we selected our independent variables mainly on the ground of empirical
data and theoretical assumptions, other psychosocial
factors could have been considered. These decisions
influence the findings because in multiple regression
analyses, results depend on the set of variables
entered.
Because a cross-sectional assessment was conducted, assumptions regarding the direction of effects

October 2012
underlying the given associations cannot be empirically corroborated.

CONCLUSIONS
Our results indicate that somatic and psychological vulnerability are closely linked, which is in line
with arguments being put forward by pain researchers like Powers and colleagues,1 Diatchenko et al,74
von Baeyer and Champion,75 and especially Kato and
colleagues,73 who assume a common ground for
somatic dysfunctions such as pain comprising primary
headaches and psychological disorders.
The common path for different functional pains is
suggested to be negative affectivity as an emotional
disposition and hypersensitivity in the processing of
internal and external stimuli. These assumptions are
at least compatible with the strong and common associations found in our study. MIG as the most severe
pain condition shows the most intimate association
with psychological dysfunctions. Some differences
seen between boys and girls in the associative pattern
suggest that the common path model could be
extended to include interactions with sex.
REFERENCES
1. Powers SW, Gilman DK, Hershey AD. Headache
and psychological functioning in children and adolescents. Headache. 2006;46:1404-1415.
2. Gladstein J, Holden EW. Chronic daily headache in
children and adolescents: A 2-year prospective
study. Headache. 1996;36:349-351.
3. Karwautz A, Wber C, Lang T, et al. Psychosocial
factors in children and adolescents with migraine
and tension-type headache: A controlled study and
review of the literature. Cephalalgia. 1999;19:32-43.
4. Anttila P, Sourander A, Metshonkala L, Aromaa
M, Helenius H, Sillanp M. Psychiatric symptoms
in children with primary headache. J Am Acad Child
Adolesc Psychiatry. 2004;43:412-419.
5. Mazzone L, Vitiello B, Incorpora G, Mazzone D.
Behavioral and temperamental characteristics of
children and adolescents suffering from primary
headache. Cephalalgia. 2006;26:194-201.
6. Lazarus RS, Folkman S. Stress, Appraisal, and
Coping. New York: Springer; 1984.
7. Larsson B, Sund AM. Emotional/behavioral, social
correlates and one-year predictors of frequent pains

Headache

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

among early adolescents: Influences of pain characteristics. Eur J Pain. 2007;11:57-65.

Virtanen R, Aromaa M, Koskenvuo M, et al. Externalizing problem behaviors and headache: A
follow-up study of adolescent Finnish twins. Pediatrics. 2004;114:981-987.
Strine TW, Okoro CA, McGuire LC, Balluz LS. The
associations among childhood headaches, emotional
and behavioral difficulties, and health care use. Pediatrics. 2006;117:1728-1735.
Kashikar-Zuck S, Lynch AM, Slater S, Graham TB,
Swain NF, Noll RB. Family factors, emotional functioning, and functional impairment in juvenile fibromyalgia syndrome. Arthritis Rheum. 2008;59:13921398.
Norton PJ, Asmundson GJ. Anxiety sensitivity, fear,
and avoidance behavior in headache pain. Pain.
2004;111:218-223.
Keogh E, Cochrane M. Anxiety sensitivity, cognitive
biases, and the experience of pain. J Pain. 2002;3:320329.
Greenberg J, Burns JW. Pain anxiety among chronic
pain patients: Specific phobia or manifestation of
anxiety sensitivity? Behav Res Ther. 2003;41:223240.
Vulic-Prtoric A, Galic S, Coha R, et al. Anxiety in
children with headaches. Psihologijske Teme. 2007;
16:201-224.
Tsao JC, Allen LB, Evans S, Lu Q, Myers CD,
Zeltzer LK. Anxiety sensitivity and catastrophizing:
Associations with pain and somatization in nonclinical children. J Health Psychol. 2009;14:10851094.
Fuss S, Pag G, Katz J. Persistent pain in a
community-based sample of children and adolescents. Pain Res Manag. 2011;16:303-309.
Martin AL, McGrath PA, Brown SC, Katz J. Anxiety
sensitivity, fear of pain and pain-related disability in
children and adolescents with chronic pain. Pain Res
Manag. 2007;12:267-272.
Barsky AJ, Goodson JD, Lane RS, Cleary PD. The
amplification of somatic symptoms. Psychosom
Med. 1988;50:510-519.
Gregory RJ, Manring J, Wade MJ. Personality traits
related to chronic pain location. Ann Clin Psychiatry. 2005;17:59-64.
Raphael KG, Marbach JJ, Gallagher RM. Somatosensory amplification and affective inhibition are
elevated in myofascial face pain. Pain Med. 2000;
1:247-253.

1399
21. Gordon KE, Dooley JM, Wood EP. Self-reported
headache frequency and features associated with
frequent headaches in Canadian young adolescents.
Headache. 2004;44:555-561.
22. Rhee H. Prevalence and predictors of headaches in
US adolescents. Headache. 2000;40:528-538.
23. Stanford EA. The prevalence, trajectories and predictors of recurrent pain among adolescents: A
population-based approach. Dissertation Abstracts
International: Section B:The Sciences and Engineering 2007;68:2672.
24. Raymer D, Weininger O, Hamilton JR. Psychological problems in children with abdominal pain.
Lancet. 1984;1:439-440.
25. Luka-Krausgrill U, Reinhold B. Kopfschmerzen bei
Kindern: Auftretensrate und Zusammenhang mit
stress, Stressbewltigung, Depressivitt und sozialer
Untersttzung. Zeitschrift fr Gesundheitspsychologie. 1996;4:137-151.
26. Weickgenant AL, Slater MA, Patterson TL, Atkinson JH, Grant I, Garfin SR. Coping activities in
chronic low pain: Relationship with depression.
Pain. 1993;53:95-103.
27. Milde-Busch A, von Kries R, Thomas S, Heinrich S,
Straube A, Radon K. The association between use of
electronic media and prevalence of headache in adolescents: Results from a population-based crosssectional study. BMC Neurol. 2010;10:1-10.
28. Sundblad GB, Jansson A, Saartok T, Renstrm P,
Engstrm L-M. Self-rated pain and perceived health
in relation to stress and physical activity among
school-students: A 3-year follow-up. Pain. 2008;136:
239-249.
29. Oksanen A, Metshonkala L, Anttila P, et al.
Leisure activities in adolescents with headache. Acta
Paediatr. 2005;94:609-615.
30. Mohseni-Bandpei MA, Bagheri-Nesami M,
Shayesteh-Azar M. Nonspecific low back pain in
5000 Iranian school-age children. J Pediatr Orthop.
2007;27:126-129.
31. Skoffer B, Foldspang A. Physical activity and lowback pain in schoolchildren. Eur Spine J. 2008;17:
373-379.
32. Shehab DK, Al-Jarallah KF. Nonspecific low-back
pain in Kuwaiti children and adolescents:Associated
factors. J Adolesc Health. 2005;36:32-35.
33. Larsson B, Melin L. The psychological treatment of
recurrent headache in adolescents: Short-term
outcome and its prediction. Headache. 1988;28:187195.

1400
34. Reynolds DJ, Hovanitz CA. Life event stress and
headache frequency revisited. Headache. 2000;40:
111-118.
35. DeBenedittis G, Lorenzetti A. The role of stressful
life events in the persistence of primary headache:
Major events vs. daily hassles. Pain. 1992;51:35-42.
36. Waldie KE. Childhood headache, stress in adolescence, and primary headache in young adulthood.
Headache. 2000;41:1-10.
37. Liakopoulou-Kairis M, Alifieraki T, Protagora D,
et al. Recurrent abdominal pain and headache: Psychopathology, life events, and family functioning.
Eur Child Adolesc Psychiatry. 2002;11:115-122.
38. Passchier J, Schouten J, van der Donk J, van
Romunde LKJ. The association of frequent headaches with personality and life events. Headache.
1991;31:116-121.
39. Boey CC, Goh KL. Stressful life events and recurrent abdominal pain in children in a rural district in
Malaysia. Eur J Gastroenterol Hepatol. 2001;13:401404.
40. Lampe A, Doering S, Rumpold G, et al. Chronic
pain syndromes and their relation to childhood
abuse and stressful life events. J Psychosom Res.
2003;54:361-367.
41. Macfarlane GJ, Pye SR, Finn JD, et al. Investigating
the determinants of international differences in the
prevalence of chronic widespread pain: Evidence
from the European male ageing study. Ann Rheum
Dis. 2009;68:690-695.
42. Wigers SH. Fibromyalgia outcome: The prediction
values of symptom duration, physical activity, disability pension, and critical life events A 4.5 year prospective study. J Psychosom Res. 1996;41:235-239.
43. Feuerstein M, Sult S, Houle M. Environmental stressors and chronic low back pain: Life events, family
and work environment. Pain. 1985;22:295-307.
44. Metsahonkala L, Sillanpaa M, Tuominen J. Social
environment and headache in 8- to 9-year-old
children: A follow-up study. Headache. 1998;38:222228.
45. Hjern A, Alfven G, stberg V. School stressors, psychological complaints and psychosomatic pain. Acta
Paediatr. 2008;97:112-117.
46. Aromaa M, Sillanpaa M, Rautava P, Helenius H.
Pain experience of children with headache and their
families: A controlled study. Pediatrics. 2000;106:
270-275.
47. Evans S, Shipton EA, Keenan T. The relationship
between maternal chronic pain and child

October 2012

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

adjustment: The role of parenting as a mediator. J

Pain. 2006;7:236-243.
Isik U, Topuzoglu A, Ay P, et al. The prevalence of
headache and its associations with socioeconomic
status among schoolchildren in Istanbul, Turkey.
Headache. 2009;49:697-703.
Bener A, Uduman SA, Quassimi M, et al. Genetic
and environmental factors associated with migraine
in schoolchildren. Headache. 2000;40:152-157.
Aromaa M, Rautava P, Helenius H. Factors of early
life as predictors of headache in children at school
entry. Headache. 1998;38:23-30.
Krner-Herwig B, Heinrich M, Morris L. Headache
in German children and adolescents: A populationbased epidemiological study. Cephalalgia. 2007;27:
519-527.
International Headache Society Classification Subcommittee. International Classification of Headache
Disorders, 2nd edition. Cephalalgia. 2004;24(Suppl.
1):1-160.
Heinrich M, Morris L, Krner-Herwig B. Self-report
of headache in children and adolescents in
Germany: Possibilities and confines of questionnaire
data for headache classification. Cephalalgia. 2009;
29:864-872.
Laurell K, Larsson B, Eeg-Olofsson O. Headache
in schoolchildren: Agreement between different
sources of information.Cephalalgia.2003;23:420-428.
Krner-Herwig B, Morris L, Heinrich M. Biopsychosocial correlates of headache: What predicts pediatric headache occurrence? Headache. 2008;48:529544.
Achenbach TM. Manual for the Youth Self-report
and 1991 Profile. Burlington: University of Vermont,
Department of Psychiatry; 1991.
Blais MA, Otto MW, Zucker BG, et al. The anxiety
sensitivity index: Item analysis and suggestions for
refinement. J Pers Assess. 2001;77:272-294.
Landgraf JM, Abetz L, Ware JE. The CHQ: A Users
Manual. Boston, MA: The Health Institute, New
England Medical Centre; 1996:1.
Hampel P, Petermann F, Dickow B. SVF-KJ: Stressverarbeitungsfragebogen Von Janke Und Erdmann,
Angepasst Fr Kinder Und Jugendliche; Manual.
Gttingen; Bern; Toronto; Seattle, WA: Hogrefe,
Verl. fr Psychologie; 2001.
Prochaska JJ, Sallis JF, Griffith B, Douglas J.
Physical activity levels of Barbadian youth and
comparison to a U.S. sample. Int J Behav Med.
2002;9:360-372.

Headache
61. Lampert T, Sygusch R, Schlack R. Nutzung elektronischer Medien im Jugendalter. Ergebnisse des
Kinder- und Jugendgesundheitsurveys (KiGGS).
Bundesgesundheitsblatt. 2007;50:643-653.
62. Krohne HW, Kiehl GE, Neurser KW, Pulsack A.
Das Erziehungsstil-Inventar (ESI): Konstruktion,
Psychometrische Kennwerte Gltigkeitsstudien.
Diagnostica. 1984;30:299-318.
63. Esser G, Blanz B, Geisel B, Laucht M. Mannheimer
Elterninterview MEI: Strukturiertes Interview Zur
Erfassung Von Kinderpsychiatrischen Aufflligkeiten. Weinheim: Beltz; 1989.
64. Greenland S. Regression methods for epidemiologic
analysis. In: Ahrens W, Pigeot I, eds. Handbook of
Epidemiology. Berlin: Springer; 2005:627-691.
65. Zwart JA, Dyb G, Holmen TL, Stovner LJ, Sand T.
The prevalence of migraine and tension-type headaches among adolescents in Norway. The NordTrndelag Health Study (Head-HUNT-Youth), a
large population-based epidemiological study.
Cephalalgia. 2004;24:373-379.
66. Anttila P. Tension-type headache in children and
adolescents. Curr Pain Headache Rep. 2004;8:500504.
67. Koetting OByrne K. Psychosocial factors in pediatric tension and migraine headache: A meta-analysis.
P.h.D. University of Kansas City, Missouri; 2003.

1401
68. Gladstone G, Parker GB. Is behavioral inhibition a
risk factor for depression? J Affect Disord. 2006;
95:85-94.
69. Dunn KM, Jordan KP, Mancl L, Drangsholt MT, Le
Resche L. Trajectories of pain in adolescents: A prospective cohort study. Pain. 2011;152:66-73.
70. Calkins SD, Blandon AY, Williford AP, Keane SP.
Biological, behavioral, and relational levels of resilience in the context of risk for early childhood
behavior problems. Dev Psychopathol. 2007;19:675700.
71. Krner-Herwig B, Gassmann J, van Gessel H, Vath
N. Multiple pains in children and adolescents: A risk
factor analysis in a longitudinal study. J Pediatr
Psychol. 2011;36:420-432.
72. Celentano DD, Stewart WF, Linet MS. The relationship of headache symptoms with severity and duration of attacks. J Clin Epidemiol. 1990;43:983-994.
73. Kato K, Sullivan PF, Evengard B, et al. A
population-based twin study of functional somatic
syndromes. Psychol Med. 2009;39:497-505.
74. Diatchenko L, Nackley AG, Slade GD, et al. Idiopathic pain disorders pathways of vulnerability.
Pain. 2006;123:226-230.
75. von Baeyer CL, Champion GD. Commentary: Multiple pains as functional pain syndromes. J Pediatr
Psychol. 2011;36:433-437.