Anxiety Symptoms and Disease Severity in Children and Adolescents With Crohn Disease

Abstract

Objectives: 

Children and adolescents diagnosed as having Crohn disease (CD), a type of inflammatory bowel disease (IBD), have increased vulnerability for anxiety symptoms that may be related to disease-related processes. The aims of this article are 3-fold: to report the proportion of pediatric patients with CD whose self-reported anxiety symptoms are indicative of distress, to describe the constellation of anxiety symptoms, and to examine the relation between anxiety and disease symptoms.

Methods: 

Retrospective medical chart review was performed for 93 youths with CD (ages 9–18 years) who had completed the Screen for Child Anxiety Related Disorders during their gastroenterology visit. Medical records were reviewed for demographic and disease characteristics. the Harvey-Bradshaw Index (HBI) was used as a measure of CD activity.

Results: 

Thirty percent of the youths reported experiencing elevated anxiety symptoms (Screen for Child Anxiety Related Disorder score >20), and 50% had scored above the cutoff in 1 or more anxiety domains, with school anxiety, general anxiety, and separation anxiety symptoms reported most frequently. Youth rated with moderate/severe disease activity on the HBI (n = 4) self-reported more anxiety symptoms compared with youth with inactive disease (n = 78, P = 0.03). Greater school anxiety was significantly associated with decreased well-being (P = 0.003), more abdominal pain (P < 0.001), and the number of loose stools (P = 0.01). Having extraintestinal symptoms was significantly associated with higher somatic/panic anxiety (P = 0.01).

Conclusions: 

Implementing a brief anxiety screen in tertiary pediatric settings may be one approach to identify young patients with CD in distress. Health care providers should consider periodic assessment of school anxiety among youth with CD.

In recent years, there has been a growing body of research that has contributed to the understanding of how biopsychosocial factors are associated with well-being and health among children and adolescents diagnosed as having physical illness. Anxiety in particular has gained greater attention because the occurrence of anxiety is high among youth with chronic illness ⁽¹⁾, and the presence of anxiety symptoms has been found to negatively affect disease and functional outcomes in numerous pediatric medical conditions such as cystic fibrosis, juvenile fibromyalgia, type 1 diabetes mellitus, asthma, and allergies ^(2–6).

Children and adolescents diagnosed as having Crohn disease (CD), a type of inflammatory bowel disease (IBD), also seem to have increased vulnerability for anxiety symptoms ^(7–9). CD is characterized by chronic remitting and relapsing inflammation of the gastrointestinal tract; symptoms include abdominal pain, diarrhea, vomiting, and poor nutritional absorption. Findings suggest that anxiety may affect disease-related processes ⁽¹⁰⁾; however, little attention has been focused toward characterizing anxiety symptoms and assessing the link between specific types of anxiety symptoms and disease symptoms among pediatric IBD samples. Characterizing anxiety domains is clinically relevant given that anxiety symptoms are heterogeneous ^(11,12), and differing normative and at-risk trajectories have been documented for separate anxiety disorder symptom domains (generalized anxiety disorder, panic disorder, school anxiety, separation anxiety disorder, and social anxiety disorder) across adolescence among community samples ⁽¹³⁾. In pediatric patients with medical disease, more broadly considering separate anxiety domains, rather than a global, 1-dimensional approach to anxiety, allows for further elucidation of the impact of anxiety on functional and medical outcomes. Furthermore, a better understanding of anxiety patterns among patients with CD may help raise awareness among practitioners of the importance of screening for anxiety symptoms to improve the overall care of the patient with CD.

IBD consists predominantly of 2 broad phenotypic entities: CD and ulcerative colitis (UC). These 2 subtypes share common symptoms but are differentiated by their distinct underlying pathophysiology, and previous findings suggest that distinguishing between these 2 phenotypes is important when studying the impact of anxiety in IBD ^(11,14,15). Indeed, a recent meta-analysis examining psychological stress and disease in adults with IBD found that associations between stress and disease course were only present in those studies that assessed a single subtype (CD or UC) but not in mixed samples ⁽¹²⁾. Similarly, 2 prior studies ^(16,17) that used pediatric samples and included a measure of anxiety provide evidence that there may be a different relation between internalized symptoms and other indices of psychological and physical functioning in youth with CD compared with UC. Both studies used a global index of internalized symptoms that combine anxiety and depressive symptoms. The relation between anxiety and disease symptoms has yet to be studied in pediatric samples.

The aims of this article are 3-fold: to report the proportion of pediatric patients with CD whose self-reported anxiety symptoms are indicative of significant distress, to describe the constellation of anxiety symptom domains, and to examine the relation between anxiety symptom severity and CD symptom severity. This study is the first of which we are aware to examine anxiety symptoms (including particular constellations of symptom domains) and their specific relations with disease symptoms in children and adolescents diagnosed as having CD.

METHODS

All study procedures were approved by the institutional review boards of Brooklyn College of the City University of New York and Icahn School of Medicine at Mount Sinai.

Sample and Procedure

A retrospective medical chart review was performed of children and adolescents between the ages of 9 and 18 years who were seen for their CD at a pediatric gastroenterology practice in New York City and had completed an anxiety symptom questionnaire as part of a larger study ⁽²²⁾ between January 2008 and August 2010.

Front staff team members were instructed to provide all of the children and adolescents with IBD an anxiety screen to complete upon clinic check-in. Of the 424 unique patients seen at the clinic in this time frame, 94 youths were approached and only 1 family refused to complete the anxiety screen, capturing 22% of possible youths seen at the clinic. The children and adolescents not approached were randomly missed by front staff team members because of their busy schedules. There was no difference in age, sex, or visit type among those children and adolescents approached and not approached (Table 1).

TABLE 1: Descriptors of the youth with CD seen in clinic compared with study sample

Research staff collected data pertaining to sex, age, race, date of CD diagnosis, family history of IBD, and prescribed IBD and psychotropic medications from patient electronic medical charts. In addition, the Harvey-Bradshaw Index (HBI) ⁽¹⁸⁾ was completed by 2 pediatric gastroenterology medical fellows via retrospectively reviewing each patient's electronic medical chart. To establish reliability, given that record keeping and interpretation of medical records can vary, 1 fellow rated the HBI for 100% of charts, of which 50% were randomly selected for a second pediatric gastroenterologist to generate a second HBI score. Raters were blinded to the participants’ self-reported anxiety scores.

Measures

Anxiety Symptom Questionnaire

Children and adolescents self-report anxiety symptoms were assessed using the Screen for Child Anxiety Related Disorders (SCARED) ⁽¹⁹⁾. The SCARED is a widely used, validated self-report questionnaire that assesses anxiety symptoms in children and adolescents ages 9 to 18 years. The SCARED has been successfully implemented in various medical settings to screen for elevated anxiety in youth with a chronic illness ^(20,21) including IBD ⁽²²⁾. The SCARED consists of 41 items on a 3-point Likert scale that collectively measure the severity of anxiety symptoms in youths for a 2-week period. Scores range from 0 to 82, with higher scores signifying higher distress. The SCARED has excellent internal consistency in clinical samples (α = 0.93) and discriminates anxiety from other nonanxiety psychiatric conditions ⁽¹⁹⁾. A total sum score of 20 has been used as a cutoff score to identify children and adolescents with IBD in distress who may meet the criteria for a clinical anxiety disorder ⁽²²⁾. Based on principal component analyses, 5 domains have been identified: panic/somatic anxiety, general anxiety, separation anxiety, social anxiety, and school phobia. Factors have good internal consistency (α = 0.74–0.89), with established cutoff scores that distinguish between the 5 anxiety disorders (P < 0.01) (≥7 panic, ≥9 general anxiety, ≥5 separation anxiety, ≥8 social anxiety, and ≥3 for school phobia) ⁽¹⁹⁾.

Disease Activity Index

Gastroenterologists retrospectively completed the HBI based on the clinician documentation to generate a disease activity score at the time of the SCARED completion. The score derived from the HBI is based on 5 items that assess a patient's well-being status, abdominal pain, diarrhea frequency, the presence of abdominal mass, and the number of extraintestinal symptoms (eg, arthralgia, uveitis, erythema nodosum, pyoderma gangrenousum, apthous ulcer, anal fissure, new fistula, abscess). Inactive disease is defined as an HBI score of <5, mild disease is designated by a score of 5 to 6, and ≥7 indicates moderate or severe disease activity ⁽²³⁾. The HBI has been used in retrospective pediatric studies to gauge severity from medical records ^(24–27).

Statistical Analyses

The frequency and percentage of children and adolescents diagnosed as having CD who self-reported a total sum score of 20 or more on the SCARED are provided (aim 1). The total sum score and standard deviation for each anxiety domain, as well as the percentage of scores above the established domain cutoffs, are presented to characterize the constellation of anxiety symptoms experienced by youth diagnosed as having CD (aim 2). Given that the items of the HBI were measured on an ordinal and count scale, rather than an interval/ratio scale, nonparametric statistics were used to analyze these data. To elucidate the relation between anxiety and HBI disease severity symptoms (aim 3), we conducted a Kruskal-Wallis (K-W) test, the nonparametric equivalent of a 1-way analysis of variance, to examine group differences in total SCARED scores across each disease severity category (inactive, mild, and moderate or severe). In addition, 4 K-W tests were conducted to investigate differences in anxiety domain scores (ie, generalized anxiety disorder, panic disorder, school anxiety, separation anxiety disorder, and social anxiety disorder) on the SCARED for each disease symptom item (well-being, abdominal pain, number of loose stools, complications) on the HBI. Only 3 youths were rated as having an abdominal mass on the HBI; therefore, this item was not independently examined in relation to the anxiety domains but was included in the total HBI score. Given that the relation between anxiety and disease activity may also be affected by medications, we ran each K-W test described above again excluding individuals with prescribed steroid, biologic, or psychotropic medications (removed n = 26). Mann-Whitney (M-W) U tests, the nonparametric equivalent of independent-samples t tests, were used to follow up significant K-W test results. (No α adjustment is required when performing pairwise comparisons to follow up on an omnibus test with only 3 levels ⁽²⁸⁾.)

Spearman rank-order correlations (for age, age at CD diagnosis, and the duration of illness) and χ² tests (for sex and family history of IBD) were conducted to examine whether demographic and disease factors were related to the total SCARED or HBI scores.

Significance was set at an α level of 0.05 for our a priori analyses (ie, the relation between total SCARED scores and disease severity criteria), whereas for omnibus K-W tests of SCARED domains, a more stringent α level of 0.01 was used to correct for multiple analyses. All of the analyses were performed using IBM SPSS Statistics software version 20 (IBM SPSS Statistics, Armonk, NY).

RESULTS

Participants

Demographic details and illness characteristics are depicted in Table 2. This was mostly a sample (80%) consisting largely of white youth ages 12 years or older (89%), with roughly equal sex distribution (47% girls). Disease duration ranged from <1 month to 12 years. More than half of the youths had been diagnosed as having CD for >1 year (60%) and 25% were diagnosed within the last 6 months.

TABLE 2: Characteristics of sample (N = 93)

Anxiety Symptoms

The mean score of anxiety symptoms on the SCARED was 14.9 (SD 12.1) and was approximately normally distributed with no outliers. Internal consistency of SCARED items with this sample was excellent (α = 0.94). Table 3 presents the means, SDs, and percentage of participants scoring above the cutoff for each anxiety domain. Overall, 30% of the children and adolescents self-reported a total score of 20 or higher on the SCARED, and 50% scored above the established cutoff on 1 or more anxiety domains. Total SCARED scores were significantly associated with age (P = 0.02), with younger children reporting more anxiety symptoms; neither sex nor disease factors (ie, age at CD diagnosis, illness duration, or family history of IBD) were associated with anxiety symptoms.

TABLE 3: Youth self-reported anxiety symptoms on SCARED (N = 93)

Disease Activity Index

The mean HBI score was 2.0 (SD 2.3). Table 4 displays the mean, SD, and range of each HBI item. Overall, ratings were skewed; a large proportion of the sample was classified, based on review of clinician documentation, as having inactive disease (84%). Age, sex, age at CD diagnosis, and family history of IBD were not associated with total HBI scores; youth with shorter illness duration had higher disease severity ratings (P = 0.04). Interrater agreement on the HBI was satisfactory (r = 0.85).

TABLE 4: Patient total disease activity and symptoms on the HBI (N = 91^*)

Anxiety and Disease Activity

Evaluating the relation between total SCARED score and HBI disease activity categories (inactive, mild, and moderate/severe) with the K-W test revealed that anxiety significantly differed for disease activity categories, χ²(2) = 6.86, P = 0.03. Post-hoc comparisons showed that youth (n = 4) rated as having moderate or severe disease self-reported more anxiety symptoms (mean rank 39.0) compared with youth (n = 76), with inactive disease (mean rank 67.6, U = 43.5, P = 0.01). Children and adolescents categorized as experiencing mild disease activity (n = 12) did not significantly differ in self-reported anxiety scores compared with those with inactive or moderate/severe disease.

Further evaluation of the relation between individual anxiety symptom domains and each disease symptoms item on the HBI (Table 5) showed that only youth self-reported school anxiety was significantly related to well-being (χ²(2) = 11.76, P = 0.003); abdominal pain (χ²(2) = 20.64, P < 0.001), and number of loose stools (χ²(2) = 8.82, P = 0.01). As displayed in Figure 1, greater school anxiety was reported in children and adolescents with “poor” well-being (mean rank 43.9) compared with those rated as “very well” well-being (mean rank 27.2, U = 127.5, P = 0.002). Youth with “none” abdominal pain had lower school anxiety (mean rank 29.0) compared with those with “mild” abdominal pain (U = 405.0, P = 0.002) as well as “moderate” abdominal pain (mean rank 52.7, U = 80.5, P < 0.001). Finally, school anxiety was higher among children and adolescents rated as having 1 to 2 loose stools per day (mean rank 54.8) than those with 0 loose stools (mean rank 37.9, U = 288.5, P = 0.01). The comparison of those with 3 to 5 loose stools per day to those with 0 fell short of significance (P = 0.06).

TABLE 5: Summary of χ² values from K-W tests on anxiety domains and disease symptoms (N = 91^||)

FIGURE 1: School anxiety mean scores and error bars (standard deviation) for well-being, abdominal pain, and number of loose stools items on the Harvey-Bradshaw Index (HBI). (*) indicates significance at the 0.01 level on the K-W test. School anxiety was (A) higher for those rated as having poor well-being compared with high well-being, (B) higher for those with moderate abdominal pain, compared with those with mild and no pain, and (C) lower for those with no loose stools per day compared with those with 1 to 2 loose stools per day. K-W = Kruskal-Wallis.

Having 1 or more extraintestinal symptoms was significantly associated with higher somatic/panic domain scores (mean rank 64.5) than not having a complication (mean rank 44.2, U = 292.0, P = 0.01). Although social anxiety scores were higher when extraintestinal symptoms were present (mean rank 61.5) compared with not present (mean rank 44.6), this difference did not meet our more stringent α level for multiple comparisons (U = 331.5, P = 0.03) (Fig. 2). Separation anxiety and general anxiety domains were not significantly related to any HBI items (Table 5).

FIGURE 2: Displayed are the means and error bars (standard deviation) for somatic/panic anxiety symptoms as a function of the number of extraintestinal symptoms. Extraintestinal symptoms were reported for 11 patients, all of whom experienced arthralgia, whereas 2 patients also had an additional complication (ie, uveitis and erythema nodosum).

All of the results of these analyses remained similar when children and adolescents taking steroids, biologics, and psychotropic medications were removed.

DISCUSSION

This article reports on the proportion of pediatric patients who self-report elevated anxiety symptoms, including global and specific domains, as well as describes the association between anxiety domains and specific disease symptoms among children and adolescents diagnosed as having CD who were screened for anxiety symptoms in a tertiary specialty medical clinic.

The findings from this study showed that 30% of our sample was experiencing elevated global anxiety symptoms on the SCARED, indicative of the risk or possible presence of clinical anxiety. In addition to global anxiety symptoms, 50% of children and adolescents self-reported experiencing distress in at least 1 anxiety domain, with school anxiety, general anxiety, and separation anxiety the most common areas reported. Epidemiology studies have established that anxiety is the most common disorder in childhood, with the lifetime prevalence of “any anxiety disorder” in children or adolescents ranging from 15% to 20% ⁽²⁹⁾. Our findings are consistent with previous reports that children and adolescents diagnosed as having CD may have an increased vulnerability for developing anxiety ^(7,30); however, prevalence rates still need to be empirically established.

The high incidence of anxiety symptoms in this sample is noteworthy given that the majority of children and adolescents were rated, based on information in their medical chart, to have inactive disease at the time of the anxiety survey. These data suggest that for some, anxiety is likely present even in the absence of disease activity. Indeed, a history of intestinal inflammation ⁽³¹⁾ and anxiety disorders ⁽³²⁾ have been linked to increased risk of experiencing physical symptoms not directly related to active disease (including abdominal pain). Furthermore, anxiety can cause physical discomfort not related to inflammation by increasing visceral hypersensitivity or pain perception ⁽³³⁾. Accordingly, these issues are particularly important for youth managing CD because gastrointestinal symptoms associated with psychological distress, or conditioned functional pain, may be misinterpreted as disease symptoms ^{(9,16,34–36)}. The misperception of disease activity may lead to gastroenterologists attributing these symptoms to disease activity and can make disease management difficult for the patient, family, and physician.

In this study, the 4 youths rated with moderate-to-severe disease were experiencing more anxiety symptoms than youths with inactive disease. When anxiety domains were examined in relation to specific disease symptoms, only school anxiety was related to poor well-being, mild and moderate abdominal pain, and the presence of loose stools. It may be that managing diarrhea and pain can be anxiety provoking at school because such symptoms have the potential to interfere with concentration and school attendance, both of which may negatively affect academic performance ⁽³⁷⁾. School-related issues can also contribute to, and in turn be affected by, general anxiety and separation anxiety symptoms ^(38,39). Even when disease inflammation is absent, fear of bowel urgency may be present and can cause school-related anxiety ⁽²²⁾, suggesting broader school-related issues for children and adolescents as having CD. Health care providers should consider the potential merits of assessing school anxiety because these symptoms may be common among youth diagnosed as having CD.

The 11 children and adolescents with extraintestinal manifestations, all with arthralgia, self-reported greater somatic/panic anxiety, which is an index of anxiety surrounding experiencing physical symptoms. The arthralgia and arthritis that is seen with CD is a sign of inflammation and often associated with increased disease activity. It may be that additional symptoms outside the gastrointestinal tract may provoke more bodily concerns because they are not well explained and are an additional, nongastrointestinal, physical symptom. Alternatively, connections between chronic pain, such as arthritis, and anxiety disorders have been demonstrated in adults ^(40,41). Given that our sample was mostly inactive, arthralgia may be another pain manifestation, suggesting that perhaps anxiety may be associated with central processing of a variety of pain signals from various body sources.

Pediatric anxiety disorders have been shown to respond to psychosocial interventions ⁽⁴²⁾. Anxiety disorders are associated with significant impairment in school, social, and family life ⁽⁴³⁾ and, for youth with chronic illness, comorbid anxiety symptoms can have a negative impact on illness adjustment and medical outcomes ⁽¹⁾. Consequently, early recognition and appropriate treatment of anxiety conditions could significantly improve the quality of care for these children. Implementing a brief anxiety screen in gastroenterology tertiary pediatric settings is 1 entry point to identify young patients with CD at increased risk for comorbid anxiety. Once identified, medical staff can play a critical role in youth adjustment by promoting families’ engagement in nonmedical interventions that target anxiety and coping that augments medical services (eg, providing referrals to appropriate specialists and/or disseminating educational materials) ^(10,44).

The results of this preliminary study should be interpreted in the context of several limitations. The sample size, specific patient characteristics (mild disease, mostly white), and recruitment from a single medical center limit the ability to extend our findings to patients with different characteristics. Although the disease severity seen in our group of patients was similar to the distribution seen in the much larger cohort reported by the ImproveCare Now registry ⁽⁴⁵⁾, only 4 youths in our sample were experiencing moderate-to-severe disease activity. This small sample not only precludes a full examination of anxiety symptoms and the association between anxiety and disease activity but also increases the chance of false-negative findings. In addition, only 22% of the children and adolescents seen in the clinic were approached to complete the anxiety screen, although our sample did not appear to differ in age, sex, and visit type from those who were not approached. Given that this study was not designed to be an epidemiological study, information obtained on patients who were approached is useful in providing an estimation of anxiety. Furthermore, the cross-sectional and retrospective nature of this study prevents us from establishing causation. Therefore, prospective, longitudinal studies performed in multiple medical settings are needed to fully examine the interplay between anxiety, disease, and functional outcomes.

An additional consideration is the measurement of anxiety and disease symptoms that requires the use of more rigorous assessment methods beyond this study. Although our results provide an initial signal that the SCARED may have utility in assessing anxiety symptoms in this population, future studies are needed that include diagnostic interviews, and assess for additional anxiety domains (eg, specific phobias, obsessive-compulsive disorder, trauma, IBD-specific anxiety). Given that mood symptoms are often comorbid with anxiety ⁽⁴⁶⁾ and may also affect disease ⁽⁴⁷⁾, future investigations should include an assessment of depressive symptoms. Furthermore, there may be better retrospective indexes of disease activity than the HBI ^(48,49), and the inclusion of objective measures of disease activity such as laboratory inflammatory markers, fecal calprotectin, endoscopic findings, or mucosal biopsies would allow for the relation between anxiety and disease symptoms to be teased apart further.

More broadly, future scientific inquiry is necessary to elucidate the complex interaction, and distinction, between anxiety and disease symptoms for children and adolescents with CD. This includes further investigation of the relation between anxiety and disease factors for pediatric UC because this relation may differ. In this study 30% of children and adolescents had a family history of IBD; future studies may investigate the moderating role of family environment and disease experience on the relation between anxiety and disease symptoms. Finally, the systematic collection of school data (eg, number of absences, grades, rates of homeschooling), factors that impede school attendance (eg, surgery, hospitalizations) and details pertaining to areas of school anxiety (eg, managing pain, bathroom access, negative peer evaluation) would expand empirical understanding of the impact of IBD on school-related functioning.

In summary, these findings build on a growing body of literature within pediatric chronic illness that anxiety symptoms are prevalent and are associated with physical symptoms. This is one of the first studies to explore the role of anxiety in children and adolescents diagnosed as having CD. The results indicate that school anxiety may be an area of marked concern for children and adolescents managing CD and that these symptoms may be uniquely associated with functional (eg, well-being) and physical (eg, abdominal pain) symptoms that may (or may not) be disease related. More empirically rigorous, prospective investigations are needed to further tease apart the subtle interaction between anxiety and disease.