|Home | About | Journals | Submit | Contact Us | Français|
Although sex differences in anxiety sensitivity or the specific tendency to fear anxiety-related sensations have been reported in adults with clinical pain, there is a dearth of relevant research among children. This study examined sex differences in anxiety sensitivity across unselected samples of 187 children with chronic pain (71.7% girls; mean age = 14.5) and 202 non-clinical children (52% girls; mean age = 13.6). Girls in the chronic pain and non-clinical samples reported elevated anxiety sensitivity relative to boys irrespective of clinical status. Girls with chronic pain also reported heightened fears of the physical consequences of anxiety compared to non-clinical girls but there were no such differences for psychological or social concerns. Among boys, anxiety sensitivity did not differ between the chronic pain and non-clinical groups. Future longitudinal research may examine whether specific fears of anxiety-related somatic sensations constitutes a sex-based vulnerability factor in the development of chronic pain.
Increasing evidence supports a robust association between chronic pain and anxiety disorders among adults (1, 2) and children (3-6). Epidemiological findings indicate that adult women are at higher risk of developing chronic pain disorders (7, 8), as well as anxiety disorders (9-11). However, this female predominance is less clear among younger populations. Studies of children's clinical pain generally show no sex differences until adolescence (12-16). Longitudinal research in the anxiety disorders similarly points to sex differences in separation anxiety (17) and generalized anxiety disorder (18) that emerge around 11-13 years of age. These findings suggest that there may be latent vulnerability factors in girls that emerge with age or pubertal maturity that lead to greater risk for development of chronic pain and anxiety disorders in adolescence that persist into adulthood.
It has been suggested that anxiety sensitivity (AS), or the specific tendency to view anxiety sensations as dangerous (19, 20) may constitute a key dispositional construct that influences the development and maintenance of clinical anxiety (20) and chronic pain (21, 22). In longitudinal work, AS has been found to predict the experience of panic attacks in adolescents over a 4 year period (23), indicating that AS is a vulnerability factor for the experience of severe anxiety. Relatively few studies have examined AS in children with chronic pain. One report found elevated AS in pediatric patients with noncardiac chest pain compared to controls (24).
Higher AS among females compared to males has been widely reported in non-clinical samples of adults (25) and children (26-29). The factor structure of AS however, appears similar across male and female adults (25) and children (27, 29). The AS factor structure consists of a higher order global factor and inter-related lower order factors including physical concerns, psychological concerns, and social concerns (see (30) for review). Research examining the lower order factors points to a particularly pronounced sex difference for AS physical concerns among adults (25) and children (26, 29), suggesting that fears of physical sensations associated with anxious arousal may be particularly salient for females. Despite the potential importance of AS in understanding the female predominance in chronic pain, there have been few relevant investigations in clinical pain samples. An adult study found that women with chest pain reported elevated global AS compared to men; sex differences in the AS lower order factors were not tested (31).
Not only is there a paucity of research on sex differences in AS among children with chronic pain, to the authors' knowledge, no existing work has directly tested the magnitude of sex differentiation in AS across chronic pain and non-clinical samples. Such research may assist in the identification of sex-specific vulnerability factors for the development of chronic pain. To address this gap in the literature, the current study tested the following hypotheses in a sample of children with chronic pain and a comparison sample of non-clinical children: 1) Girls will report higher levels of global AS and AS physical concerns compared to boys across the pain and non-pain groups; 2) The pain group will report higher levels of global AS and AS physical concerns compared to the non-pain group across sex. These primary hypotheses focused on global AS and AS physical concerns since these constructs have demonstrated the strongest sex differences in previous research. Exploratory analyses examined sex differences in AS psychological concerns and AS social concerns across the pain and non-pain samples.
The chronic pain sample consisted of 225 children (159 girls, 70.7%) presenting for treatment at a multidisciplinary, tertiary clinic specializing in pediatric chronic pain. From this larger sample, the present study included 187 children (age range = 9 – 18 years; M = 14.5, SD = 2.4) with complete data on the main measure of interest, i.e., the Childhood Anxiety Sensitivity Index (CASI; see detailed description below). Demographic information for the chronic pain sample is shown in Table I. Pain diagnoses for the chronic pain sample are displayed in Table II. The non-clinical sample was drawn from a larger study on laboratory pain that included 244 children without chronic pain (124 girls; 50.8%; age range = 8 to 18 years; M = 13.6, SD = 2.6) (32, 33). From this larger sample, 202 non-clinical children in the same age range as the chronic pain sample were included in the present study. Demographic data for the non-pain sample are displayed in Table I. The UCLA Institutional Review Board approved the study. Written informed consent forms were completed by parents, and children provided written assent.
For the chronic pain sample, detailed description of the procedures for the administration of the questionnaire data examined in the current study is available elsewhere (34, 35). Briefly, prior to patients' initial clinic evaluation, two questionnaire packets, one for the child and one for a parent, were mailed to the home following verbal consent from a parent obtained via telephone. Written informed consent from parents and written assent from children were obtained either at the initial clinic evaluation or prior to an in-home qualitative interview for those families who agreed to be interviewed. The qualitative interview focused on the effect of chronic pain on the child's self-perceptions and on the child's and family's daily functioning. The questionnaire packets contained instructions that parents and children were to complete them separately, without consulting each other. A research assistant reviewed the questionnaires with parents and children at the initial clinic evaluation to clarify ambiguous or missing responses. The questionnaires assessed demographic and general health information including measures of the child's pain, anxiety and functioning. Only those measures relevant to the present study are discussed below.
Recruitment and study procedures for the non-pain sample and the laboratory pain study are described in detail elsewhere (32, 36). In brief, participants were recruited through mass mailing, posted advertisements, and children's classroom presentations. Advertisements were posted around the UCLA campus and other sites (e.g., community centers) in the greater Los Angeles area to enhance enrollment of children from low-income and minority neighborhoods. After confirmation of eligibility and verbal consent from a parent by telephone, written informed parental consent and child assent forms were mailed for review and signature. On the day of the laboratory session, participants completed questionnaires in a quiet room prior to undergoing experimental pain procedures. Only questionnaires relevant to the present study are discussed below. Participants received a $30 gift certificate and a T-shirt for participation.
Locally developed questionnaires completed by the parent assessed demographic information for children and parents including age, sex, race/ethnicity, and education.
(CASI) (37) is an 18-item instrument that assesses the specific tendency to view anxiety sensations as dangerous (e.g., “It scares me when my heart beats fast”). Items are scored on a 3-point scale (none = 1, some = 2, a lot = 3); total scores are calculated by summing all items. The CASI has demonstrated good internal consistency (Cronbach's alpha = .87) and adequate test-retest reliability (range = .62 - .78 over two weeks) (37). The CASI correlates well with measures of trait anxiety (r's = .55 - .69) but also explains variance in fear that is not attributable to trait anxiety (38). Cronbach's alpha for the CASI total scores in the pain group were 0.88, 0.81, and 0.87 for girls, boys, and the total sample, respectively. Cronbach's alpha for the CASI total scores in the non-pain group were 0.82, 0.72, and 0.78 for girls, boys, and the total sample, respectively.
Although there is controversy regarding the precise number of CASI lower order factors (27, 39-41), the subscale scores pertaining to the three lower order factors identified in prior research, i.e., physical concerns, psychological concerns, social concerns, were used to be consistent with prior work on sex differences in the CASI dimensions (28). Cronbach's alpha in the pain group for the CASI physical concerns subscale were 0.86, 0.84, 0.86 for girls, boys, and the total sample respectively. Cronbach's alpha in the non-pain group for the CASI physical concerns subscale were 0.81, 0.71, 0.77 for girls, boys, and the total sample respectively. Cronbach's alpha in the pain group for the CASI psychological concerns subscale were 0.72, 0.77, 0.73 for girls, boys, and the total sample respectively. Cronbach's alpha in the non-pain group for the CASI psychological concerns subscale were 0.63, 0.48, 0.57 for girls, boys, and the total sample respectively. Cronbach's alpha in the pain group for the CASI social concerns subscale were 0.65, 0.76, 0.69 for girls, boys, and the total sample respectively. Cronbach's alpha in the non-pain group for the CASI social concerns subscale were 0.54, 0.55, 0.55 for girls, boys, and the total sample respectively.
Independent t-tests and chi-square tests for continuous and categorical data, respectively, were used to preliminarily examine sex differences in demographic characteristics across the pain and non-pain groups, as well as sex differences in clinical characteristics in the pain group. Pooled-variance t-tests were employed if Levene's tests indicated unequal variance across groups. To examine sex differences in AS, total CASI and CASI subscale scores were compared using a series of 2 (group: non-pain, pain) × 2 (sex: girls, boys) ANCOVAs, controlling for child age and race/ethnicity (Caucasian vs. non-Caucasian). Following detection of significant interaction effects, simple-effects analyses were conducted on adjusted means. For the primary hypotheses of sex differences in global AS and AS physical concerns, a Bonferroni correction was used to protect against inflation of Type I error. To evaluate the two sets of hypotheses, alpha was set at .025 (two-tailed). For exploratory analyses of AS psychological and social concerns, as well as the preliminary analyses of sex differences in demographic and clinical characteristics described above, a standard probability level of .05 (two-tailed) was used.
Following established procedures for the detection of outliers (42), 2 outliers (1 girl in the non-pain group and 1 boy in the pain group) were identified and excluded. These outliers reported total CASI scores greater than three standard deviations above the mean for their sex. Thus, the analyses were conducted on a final sample of 186 pain patients (132 female; 70.6%) and 201 non-pain participants (104 female; 51.7%). Because cell sizes were unequal with more girls than boys in the pain group compared to the non-pain group (χ21 = 15.20, p < .001), potential violations of assumptions were carefully investigated. To ensure robustness of ANCOVA results with unequal sample sizes for two-tailed tests, the recommended ratio of the largest to smallest cell size should not exceed 4:1 and the ratio of the largest to smallest variance should not exceed 10:1 (42); page 328). For the current analyses, these ratios were 2.4:1 and 2.6:1, respectively.
Table II displays the presenting pain diagnoses, presence of multiple pain disorders (yes/no), as well as average duration of pain symptoms for girls, boys, and the total pain group. Pain diagnoses were categorized as follows: headaches (migraines; myofascial, vascular, tension, stress-related or other type of headaches), functional neurovisceral pain disorder (functional bowel, uterine, or bladder disorder), myofascial pain (of any part of the body excluding headaches), fibromyalgia, complex regional pain syndrome, type 1 or type 2 (CRPS-I; CRPS-II), and arthritis. As shown in Table II, over 30% of the patient sample presented with more than one pain diagnosis. In the pain group, there were no sex differences on the following demographic variables: child age, child race/ethnicity (Caucasian vs. non-Caucasian), or parent education. Similarly, no sex differences were found for the presence of multiple pain disorders. However, girls reported significantly shorter pain duration compared to boys (t95.8= 2.10, p < .05). There were no sex differences in the presence of any individual pain disorders except fibromyalgia (χ21 = 4.83, p < .03); girls were more likely to present with this condition than boys.
There were no sex differences in the non-pain group on any of the following demographic variables: child age, child ethnicity (Caucasian vs. non-Caucasian), or parents' education. Comparisons on demographic characteristics between the non-pain and pain groups indicated the following: the mean age of the non-pain group was significantly lower than that of the pain group (t420 = -3.72, p < .001), despite that fact that both groups were comprised of children in the same age range. In addition, there were significantly more non-Caucasian children in the non-pain group (60.9%) compared to the pain group (35.0%) (χ21 = 27.90, p < .001). Thus, child age and child race/ethnicity were used as covariates in the multivariate analyses examining sex differences in CASI scores across the two groups.
Frequencies for the CASI total and CASI subscale scores for boys, girls, and the total sample for the pain and non-pain groups are displayed in Table III. Analyses comparing total CASI scores across the pain and non-pain groups indicated a significant main effect for sex (F1,380 = 10.73, p< .01), such that across groups, girls exhibited higher total scores than boys. There was no main effect of group (F1,380 = .82, p = .37). The group by sex interaction did not reach significance (F1,380 = 3.82, p = .052). Inspection of the total CASI scores indicated that girls in the pain group tended to score higher than non-pain girls whereas boys in the pain group tended to score similarly to non-pain boys.
Results for AS physical concerns subscale indicated a significant main effect for sex (F1,378 = 10.73,p < .01, η2 = .04), such that girls reported higher scores than boys. In addition, there was a significant group by sex interaction (F1,378 = 5.31, p = .022). This interaction is illustrated in Figure 1. Simple effects analysis indicated that girls in the pain group scored significantly higher on AS physical concerns compared to girls in the non-pain group (F1,231 = 7.00, p < .01), but there was no such difference among boys. There was no main effect of group for AS physical concerns (F1,378 = 2.78, p = .10). Exploratory analyses of the AS social and AS psychological concerns subscales indicated no significant main effects of sex or group and no interaction effects.
Consistent with our first main hypothesis, girls in both the chronic pain and non-clinical groups reported elevated global AS and AS physical concerns relative to their male counterparts. These findings of sex differences irrespective of clinical status held after controlling for child age and child race/ethnicity (Caucasian vs. non-Caucasian). Contrary to our second main hypothesis, children with chronic pain did not report higher global AS or AS physical concerns relative to non-clinical children. However, these findings were qualified by a significant group by sex interaction effect which indicated that girls with chronic pain reported elevated AS physical concerns relative to non-clinical girls but no such group differences for boys (see Figure 1). This interaction did not reach significance for global AS and the results of the exploratory analysis of AS social and psychological concerns did not indicate any differences based on sex or group, nor any interaction effects. Taken together, these findings indicate that AS physical concerns reflect the most salient aspect of sex-based differentiation in AS for chronic pediatric pain. Thus, girls with chronic pain evidenced greater fears of the physical consequences of anxious arousal compared to girls without chronic pain but these differences did not extend to fears of the social or psychological consequences of anxiety sensations. Boys with chronic pain did not differ from non-clinical boys with respect to any such fears of anxious arousal.
To our knowledge, this is the first study that has specifically investigated sex differences in AS among children with chronic pain and a comparison sample of non-clinical children. One prior study that did not compare boys and girls found that children with noncardiac chest pain endorsed heightened global AS as well as elevated AS physical and psychological (but not social) concerns compared to controls (24). Our findings of elevated global AS in non-clinical girls relative to non-clinical boys agree with the results of four previous studies with non-clinical children (26-29). In an earlier report, with a subset of roughly half the current non-clinical sample, we found no sex differences in global AS (36) but this may have been due to restricted sample size or the inclusion of younger children (i.e., aged 8 years). Similarly, in a previous study in a subset of less than half the present chronic pain sample (N = 87), we did not find any sex differences in global AS (43), possibly due to low power. It should be noted however, that neither of our prior analyses explicitly tested for sex differences in AS physical concerns and this subscale has revealed the strongest sex-based differences in previous research (25, 26, 29) as well as in the current study.
It is unclear why our findings of heightened fears of anxiety-related somatic sensations in chronic pain vs. non-clinical girls were not evident among boys. Analyses of the clinical characteristics in the pain sample indicated that girls reported shorter duration of pain than boys, suggesting that in this respect, boys were more severely impacted than girls. On the other hand, there were no differences between boys and girls in the presence of multiple pain diagnoses or any individual pain diagnosis with the exception of fibromyalgia. Girls were more likely to present with fibromyalgia than boys and it is possible that this condition is more strongly linked with heightened AS. Fibromyalgia involves widespread pain and numerous tender points throughout the body as well as fatigue and disruptions in sleep, mood and cognition. Recent population-based estimates indicate that women are 1.64 times more likely than men to have fibromyalgia (95% CI = 1.59-1.69)(44). It is possible that fibromyalgia, because of its diffuse nature, demonstrates a more robust association with fears of physical sensations related to anxious arousal compared to pain conditions with a more discrete focus (e.g., headaches; abdominal pain). This possibility is speculative and requires further study. Because approximately 30% of our pain sample presented with more than one pain diagnosis, comparisons between individual diagnoses on AS were not conducted.
Reasons for the observed sex differences in AS are not well-understood. One possibility is that girls are more willing to report fears compared to boys (45). Research in twins indicates that AS is heritable in women but not in men (46) suggesting that genetic susceptibility to AS in females may be a factor. A recent report (47) found that for women, the more severe the AS, the more strongly it was influenced by genetic factors. In women, each of the three AS dimensions was influenced by both genetic and environmental factors. Among men, the AS dimensions were influenced by environmental but not genetic factors. Thus, the authors concluded that AS dimensions appear to arise from both dimension-specific and non-specific etiologic factors, the salience of which varied as a function of sex and severity. In accord, we recently found evidence for possible sex differentiation in routes of transgenerational transmission for AS in two studies conducted with a subset of the non-clinical sample included in this report. In the first study, we found that whereas parent AS was significantly positively associated with child AS in girls, no such parent-child AS relationship existed for boys (48). In the second study, we found that in girls only, parent AS was positively linked with child AS, which in turn positively predicted children's experimental pain intensity (49). These findings are consistent with the view that parent AS may operate via girls' own fear of anxious arousal to influence their responses to pain. One limitation of these findings is that the majority of parents in the two studies were mothers and an important consideration for future research is the examination of possible differences in father-daughter/son vs. mother-daughter/son AS and pain associations.
Limitations to the present findings should be mentioned. The cross-sectional nature of this work precludes any conclusions regarding causality. The current study relied on a self-report measure of AS; corroboration of the findings with behavioral measures of AS (e.g., laboratory stress tasks of anxious arousal) would increase confidence in the results. It has been reported that children with chronic pain endorse levels of anxious symptomatology comparable to that of children with clinical anxiety (4). Thus, it is possible that the findings of sex differences in the pain sample could have been due to higher levels of anxious symptoms among girls with chronic pain. Although we did not include a measure of anxiety symptoms, additional analyses of the chronic pain sample indicated that our findings held even after we controlled for the presence of possible anxiety disorder as assessed by a clinical psychologist or child psychiatrist during the initial clinic evaluation. Although formal clinical assessment of anxiety disorder(s) was not conducted, the possible presence of an anxiety disorder was noted on the patients' chart by the evaluating psychologist or psychiatrist. Notably, there were no sex differences in the frequency of possible anxiety disorder in our pain sample. Future studies of AS in chronic pain samples should include established measures of anxious symptomatology and/or formal clinical assessment of anxiety disorders. Nevertheless, research on vulnerability characteristics for chronic pain among children is still nascent (45) and few modifiable risk factors have been identified for the development of chronic pain and pain-related disability. Given that AS is a potentially modifiable risk factor that appears to be particularly salient among girls, these findings may point the way to sex-specific interventions targeting AS in vulnerable populations.
In sum, the findings point to heightened global fears of anxious arousal as well as elevated specific fears of the negative physical consequences of such arousal in girls with chronic pain and non-clinical girls relative to their male counterparts. Moreover, girls with chronic pain reported increased specific fears of anxiety-related physical sensations compared to non-clinical girls whereas such differences were not apparent among boys. These results indicate that specific aspects of AS focused on physical concerns are associated with the experience of chronic pain in girls but not in boys. Longitudinal studies are warranted to examine the extent to which AS may predict the development of chronic pain problems in a sex-dependent manner across adolescence and adulthood. Additional work may also examine whether interventions aimed at reducing AS prevents the development of chronic pain and/or reduces acute pain sensitivity among girls. Research on sex-specific childhood learning experiences may also shed light on the role of AS and related factors in the development of the adult female predominance in chronic pain.
This study was supported by R01MH063779, awarded by the National Institute of Mental Health (PI: Margaret C. Jacob), by 2R01DE012754, awarded by the National Institute of Dental and Craniofacial Research (PI: Lonnie K. Zeltzer), and by UCLA General Clinical Research Center Grant MO1-RR-00865 (PI: Lonnie K. Zeltzer).