Fatigue in pediatric inflammatory bowel Explained by transdiagnostic and disease‐focused factors

Authors: Maartje D. Stutvoet, Anouk Vroegindeweij, Tessa Z. Toonen, Merel M. Nap‐Van der Vlist, Johanna W. Hoefnagels, Anemone van den Berg, Sanne L. Nijhof

Abstract

Objective

Fatigue is highly prevalent in children with inflammatory bowel disease (IBD), even during clinical remission. This suggests a role for transdiagnostic factors—lifestyle, psychological, and social influences not specific to the disease. This study aimed to assess the prevalence of severe fatigue in pediatric IBD and evaluate its associations with both IBD‐focused and transdiagnostic factors.

Methods

Children with IBD ages 8–18 from the PROactive cohort completed the Pediatric Quality of Life Inventory—Multidimensional Fatigue Scale. IBD‐focused clinical data were extracted from electronic health records. Transdiagnostic factors were assessed using validated patient‐reported outcome measures. Associations with fatigue were examined using linear regression.

Results

Among 127 patients (mean age 14.9 ± 2.7 years; 43% male), most were in clinical remission. One hundred six patients self‐reported fatigue, and 99 parents reported their child′s fatigue. Severe fatigue was self‐reported by 29%. Of IBD‐focused factors, only disease activity (β = −0.40) and comorbidity (β = −0.18) showed significant associations with fatigue. Of transdiagnostic factors, lower physical, emotional, and social functioning; poorer sleep quality; less physical activity; more pain, anxiety, and depressive symptoms; lower life satisfaction and self‐rated health; and increased school absence and pressure were significantly associated with more fatigue. Backward selection retained only transdiagnostic factors in the multivariate model, which explained 78% of fatigue variance.

Conclusion

Fatigue is common in children with IBD and is more strongly associated with transdiagnostic than disease‐focused factors. These findings highlight the importance of integrative care strategies that address modifiable psychosocial and lifestyle domains to reduce fatigue and improve functioning in children with IBD.

INTRODUCTION

Fatigue is a frequent and burdensome symptom in children with inflammatory bowel disease (IBD), ^1^ , ^2^ strongly linked to reduced quality of life. ^3^ , ^4^ Proposed mechanisms include subclinical inflammation, anemia, and sleep disturbances. ^5^ Fatigue has been associated with higher disease activity, ^3^ , ^6^ older age, ^3^ anxiety, depression, ^7^ , ^8^ and family dysfunction ^8^ in pediatric IBD. Children, caregivers, and healthcare professionals have jointly identified fatigue as a top research priority, ^9^ highlighting the need for better understanding and treatment. ^1^ , ^10^ , ^11^ , ^12^ , ^13^

Fatigue is common across chronic pediatric conditions, such as other autoimmune diseases and childhood cancer. ^14^ , ^15^ Its widespread occurrence suggests that fatigue often stems from transdiagnostic factors—shared psychological, lifestyle, and social influences—rather than solely disease‐specific mechanisms. ^14^ , ^16^ Fatigue during clinical remission further supports this perspective. ^1^ , ^2^ , ^4^ , ^5^ It shows the need to assess beyond IBD‐focused clinical factors, specifically, to include transdiagnostic factors.

Reported fatigue prevalence in pediatric IBD varies widely (8%–23%), ^4^ , ^17^ likely due to inconsistent definitions and measurement tools, and cohort differences. Prior studies have often relied on disease‐specific measures ^2^ , ^4^ or lacked a comprehensive assessment. ^1^ This study estimated the prevalence of severe fatigue using a validated, generic measure in a well‐characterized pediatric IBD cohort, and examined associations with both IBD‐focused and transdiagnostic factors. Understanding these associations may clarify whether IBD‐specific care is required, or whether integrative, transdiagnostic approaches addressing modifiable factors are more appropriate. ^16^

METHODS

Ethics statement

The institutional review board classified the study as exempt from the Medical Research Involving Human Subjects Act (METC no. 16‐707/C). This study adheres to all local laws and the Declaration of Helsinki.

Study design and patients

This is a preregistered (DOI: 10.17605/OSF.IO/CK8J5) single‐center, cross‐sectional study in pediatric patients (ages 8–18) with IBD. All patients with IBD were selected from the PROactive cohort between October 2017 and September 2024. ^18^ The PROactive cohort is a longitudinal study that collects transdiagnostic patient‐reported outcomes on fatigue, daily life participation, and psychosocial well‐being in children with chronic conditions. It is based at the Wilhelmina Children's Hospital in the Netherlands, an academic medical care center. Patients are generally invited 6–12 months after initial diagnosis, a timing that typically reflects a stable disease phase following completion of induction therapy. Before enrollment, patients and/or one of their parents provided informed consent to use clinical and patient‐reported data for research. Of the children approached for the PROactive IBD cohort, 127 (43%) provided informed consent and completed the questionnaires. Based on administrative records, the sample did not differ from nonparticipants in age. The most common reasons for nonparticipation were feeling preoccupied with participation in other research, no longer receiving care at the hospital, or unknown. More details about the PROactive cohort are described elsewhere. ^18^

Data collection

The PROactive cohort collected transdiagnostic factors through patient and/or parent questionnaires administered via the online portal “www.hetklikt.nu” before an outpatient visit. For this cross‐sectional study, the earliest available time point with patient‐reported fatigue was selected. If patient‐reported fatigue was unavailable, the earliest time point with parent‐reported fatigue was used. IBD‐focused clinical data were extracted from electronic health records (EHRs).

Measurements

The primary outcome was fatigue, measured using the general fatigue subscale of the validated Dutch version of the Pediatric Quality of Life Inventory—Multidimensional Fatigue Scale (PedsQL‐MFS). ^19^ This subscale has six items (e.g., “I feel tired,” “I feel too tired to do things that I like to do”) and yields a score ranging from 0 to 100, with lower scores indicating more fatigue.

IBD‐focused factors were selected to reflect disease characteristics such as disease activity. They were distinguished from transdiagnostic factors based on measures routinely used in IBD care. Factors indicating disease activity were fecal calprotectin (FC; ≥250 μg/g indicating active inflammation ^20^ ), C‐reactive protein (CRP; ≥5 mg/L considered elevated ^21^ ), and clinical disease activity. Clinical disease activity was based on the weighted Pediatric Crohn's Disease Activity Index (wPCDAI) for patients with Crohn's disease ^22^ and the Pediatric Ulcerative Colitis Activity Index (PUCAI) for those with ulcerative colitis. ^23^ Other IBD characteristics included age at and time since diagnosis, the presence of comorbidities (Supporting Information S1: Table S1), hemoglobin concentration (mmol/L), and current IBD medication use. Medication was categorized as biologicals (e.g., infliximab), immunosuppressants (e.g., azathioprine or methotrexate), and salicylates (e.g., mesalazine). All IBD‐focused factors were obtained from the EHR and matched to the questionnaire date as closely as possible, within a window of 3 months (Table 1).

Transdiagnostic factors were assessed across lifestyle, psychological, and social domains, and were patient‐reported unless indicated otherwise. Physical, emotional, and social functioning were measured using the corresponding subscales of the PedsQL‐Generic Core Scales (GCS). ^24^ Sleep/rest problems were assessed using the 6‐item “sleep/rest fatigue” subscale from the PedsQL‐MFS. ^19^ All PedsQL subscales range from 0 to 100, with higher scores indicating better functioning/fewer difficulties. The PedsQL‐GCS and ‐MFS demonstrate good internal consistency, with Cronbach's α values ranging from 0.73 to 0.88 and 0.79 to 0.80, respectively. ^19^ , ^24^ Symptoms of anxiety and depression were measured using the corresponding subscales of the 25‐item Revised Child Anxiety and Depression Scale (RCADS‐25), which have good internal consistency. The anxiety and depression subscales range from 0 to 45 and 0 to 30, respectively, with higher scores indicating more severe symptoms. ^25^ Perceived health, school pressure, and physical activity were evaluated using items from the Health Behavior in School‐aged Children questionnaire. ^26^ Physical activity was measured by the number of days in the past week the patient was physically active, based on World Health Organization (WHO) recommendations, ^27^ and has acceptable reliability. ^28^ Average pain in the last week was indicated on a visual analog scale from 0 to 10. ^29^ Obesity was defined as a body mass index (BMI) z score > 2 standard deviation (SD), calculated using normative data from the Third Dutch Growth study (1980) based on anthropometrics in the EHR. ^30^ School absence, the only parent‐reported factor, was assessed by the percentage of school missed over the previous 6 months.

Data analysis

Patient and IBD characteristics, and fatigue scores were summarized using descriptive statistics. For normally distributed variables, we reported means and SDs, and for non‐normally distributed variables, medians, and interquartile ranges (IQRs). Fatigue and severe fatigue were respectively defined as a score >1 SD and >2 SDs below the normative mean, based on reference values from healthy Dutch controls matched for age and sex, ^19^ factors known to influence fatigue. ^31^ Thresholds for severe fatigue ranged from 43.3 to 62.0. ^15^

To examine the associations between patient‐reported fatigue (continuous outcome) and IBD‐focused and transdiagnostic factors, we applied a stepwise analytic approach following multiple imputation of missing values in both outcome and predictor variables using the MICE package in R (MICE package, R). ^32^ The proportion of missing data ranged from 0% (e.g., medication use, age) to 23% for school absence (Supporting Information S1: Table S2).

First, we conducted exploratory Spearman's rank correlations to identify candidate factors (p < 0.05). Second, candidate factors were entered into separate linear regression models adjusted for age and sex. Third, we constructed a comprehensive multivariable model to explain fatigue variance using backward selection, beginning with variables showing p < 0.01 in the adjusted linear regressions. All assumptions for linear regression were verified. To evaluate the robustness of our findings, we performed a sensitivity analysis using the original (nonimputed) dataset.

RESULTS

The mean age of the 127 participating patients was 14.9 ± 2.7 years, and 43% were male (Table 1). Most were diagnosed with Crohn's disease (52%, n = 66), followed by ulcerative colitis (42%, n = 54) and unclassified IBD (6%, n = 7). The median time since diagnosis was 1.5 years IQR [1.0; 3.2]. The majority were on immunosuppressants (59%, n = 75), and a substantial proportion on biologicals (46%, n = 51). Nearly all patients were in clinical of the seven patients (6%) with moderate to severe clinical disease activity, only three (3%) had elevated FC levels (≥250 µg/g) as well. In total, elevated FC levels were observed in 23 patients (19%).

Fatigue

Patient‐reported fatigue was available for 106 patients, and parent‐reported fatigue for 99 86 patients had both. Almost half of the patients reported fatigue (46%, n = 49), with a substantial proportion experiencing severe fatigue (29%, n = 31) (Figure 1). Parent‐reported fatigue and severe fatigue prevalences were similar, at 45% (n = 45) and 26% (n = 26), respectively. The mean patient‐reported general fatigue score was 65 ± 25, with a corresponding z score of –0.97 ± 1.83.

Association of IBD‐focused factors with fatigue

Of the IBD‐focused factors significantly correlated with patient‐reported fatigue (Table 2), higher clinical disease activity (β = −0.40, p < 0.01) and the presence of comorbidity (β = −0.18, p = 0.048) remained significantly associated with more fatigue in the linear regression model adjusted for age and sex (Table 3). In contrast, age at diagnosis (p = 0.087) and time since diagnosis (p = 0.084) were no longer significantly associated with fatigue after adjustment. No other IBD‐focused factors showed significant associations with fatigue (Table 2).

Association of transdiagnostic factors with fatigue

Obesity was the only transdiagnostic factor not significantly correlated with patient‐reported fatigue (r = −0.052, p = 0.58). In linear regression models adjusted for the age and sex, lower fatigue levels were significantly associated (p < 0.001 unless indicated otherwise) with better physical (β = 0.71), social (β = 0.57), and emotional functioning (β = 0.48), better sleep/rest quality (β = 0.61), higher life satisfaction (β = 0.47) and more days of physical activity (β = 0.21, p = 0.040). Conversely, higher fatigue was significantly associated with more depressive (β = −0.75) and anxiety symptoms (β = −0.38), poorer self‐rated health (β = −0.60), more school absence (β = −0.52) and school pressure (β = −0.30, p = 0.006), and more pain (β = −0.48) (Table 3).

Full explanatory model

The final multivariable model explained 78% of the variance in fatigue (Table 3). It initially included all factors significantly associated with fatigue in univariable analyses (p < 0.01), adjusted for age and sex. Emotional functioning was excluded due to multicollinearity with depressive symptoms, and clinical disease activity and pain were removed during backward selection because their exclusion improved model fit. The resulting model retained only transdiagnostic factors. Depressive symptoms remained the strongest predictor (β = −0.50), while associations with most other factors were small. The direction of the association between anxiety symptoms and fatigue changed from negative in the univariable analysis to positive in the multivariable model. Post hoc inspection revealed no evidence of multicollinearity or suppression effects to explain this reversal.

Sensitivity analyses

Descriptive statistics were similar for original and imputed data (Supporting Information S1: Table S2). Characteristics also aligned between those with and without patient‐reported fatigue data (n = 21, Supporting Information S1: Table S3). Patients with missing IBD‐focused factors (n = 24) were less often male (21% vs. 49%, p = 0.025, Supporting Information S1: Table S4). Repeating analyses on the original data showed consistent directions of associations (Supporting Information S1: Tables S5 and S6). Univariate models showed no changes in direction or significance (Supporting Information S1: Table S6). In the multivariable model (n = 74), directions were stable with two small shifts in significance.

DISCUSSION

This study demonstrates that fatigue is highly prevalent in pediatric IBD, with 29% of patients reporting severe fatigue—despite most being in clinical remission. While clinical disease activity showed a modest univariable association, only transdiagnostic factors explained unique variance in the final model (78%).

The prevalence of severe fatigue in our pediatric IBD cohort (29%) was substantially higher than in the general population (2%). ^19^ It also appears slightly higher than in other chronic pediatric diseases (21%) (Figure 1), ^15^ as well as a previous Dutch IBD cohort (23%). ^4^ While these findings show a notable burden of fatigue in pediatric IBD, the apparent differences with other chronic diseases and the earlier IBD study should be interpreted with caution, given the relatively small sample sizes of both IBD studies. Direct comparisons across diagnoses are further complicated by the absence of a shared disease activity metric, though shared patterns in transdiagnostic associations suggest possible common mechanisms. ^16^ A potential difference with the previous IBD cohort could be attributable to differences in setting (only academic vs. also nonacademic) and the inclusion of patients regardless of disease activity. Notably, the mean general fatigue score in our cohort fell within the range reported in earlier studies (60.87–75.30). ^3^ , ^33^ Importantly, our findings confirm the presence of fatigue also during clinical remission, ^3^ , ^4^ a pattern also observed in other pediatric autoimmune diseases. ^14^ , ^15^ While our cross‐sectional design limits conclusions about the chronicity of fatigue, the persistence of fatigue in adults with IBD over time ^34^ underscores the importance of addressing fatigue in children to mitigate long‐term impacts.

In our study, among the IBD‐focused factors, only clinical disease activity was significantly associated with fatigue in univariate analysis (p < 0.01). This is consistent with earlier studies showing modest associations between clinical disease activity and fatigue ^2^ , ^3^ , ^4^ , ^33^ , ^34^ , ^35^ and no association with biochemical markers such as FC, CRP, and anemia. ^2^ , ^3^ , ^4^ , ^11^ , ^35^ The discrepancy between clinical and biological markers may reflect the inclusion of functional limitations in the clinical disease activity scores, and their weak correlation with measures of biological disease activity as FC. ^36^ It should be noted that IBD subtype‐specific associations with IBD‐focused factors may exist, ^35^ and we were unable to include all potentially relevant factors, such as prior bowel surgery and being underweight, due to limited variability. Additionally, our sample consisted largely of patients in clinical remission, which may have reduced the sensitivity to detect associations with biological disease severity. Yet, the lack of associations with objective disease markers mirrors findings in other pediatric conditions. ^14^ , ^16^ Taken together, the association with clinical disease activity shows that fatigue may be closely linked to perceived disease‐related limitations rather than inflammatory activity alone.

Transdiagnostic factors—including depressive symptoms, poor sleep, and social functioning—were strongly associated with fatigue, consistent with previous findings in pediatric IBD cohorts, ^3^ , ^4^ , ^33^ and other chronic diseases. ^14^ , ^16^ Interestingly, the inverse association between anxiety and fatigue in our multivariable model contrasts with adult IBD studies, where higher anxiety has been linked to greater fatigue. ^11^ However, a similar inverse trend has been reported in other chronic childhood diseases, ^16^ though the mechanism remains unclear. It is possible that contextual factors, such as support from parents and friends, need to be taken into account to properly interpret this association.

Our results emphasize the complex interactions among psychosocial factors and support a more integrative approach to fatigue assessment. This aligns with the biopsychosocial model of fatigue, which proposes that fatigue arises from dynamic interactions among biological, psychological, and social contributors. ^37^ While disease activity may initially precipitate fatigue, other factors—such as poor sleep or school pressure—can perpetuate it, even after inflammation resolves. ^38^ Although our cross‐sectional design precludes causal inference, the consistent associations across lifestyle, psychological, and social domains—including school absence and reduced quality of life—underscore the broad impact of fatigue and suggest that these factors may also contribute to its persistence. The prominence of transdiagnostic factors in our findings highlights their potential as modifiable intervention targets, particularly in patients in clinical remission.

Currently, no evidence‐based treatments for fatigue in pediatric IBD exist. ^39^ However, modifiable factors such as sleep, physical activity, and depressive symptoms represent promising targets. These are transdiagnostic factors on which children with IBD do not necessarily perform worse than healthy peers. ^4^ , ^40^ , ^41^ Yet, these factors have been associated with fatigue in IBD before, ^4^ , ^5^ , ^11^ and have been shown to longitudinally predict fatigue in adults with IBD, ^34^ in the general pediatric population, ^42^ , ^43^ , ^44^ and in children postcancer. ^45^ Given that these modifiable factors are not IBD‐specific, non‐IBD‐specific interventions—such as cognitive behavioral therapy and graded exercise—may be relevant in managing fatigue in pediatric IBD. ^46^ , ^47^

Our study has several limitations. First, the representativeness of our study sample is uncertain due to the low PROactive participation rate among the IBD population at our center and limited information on nonparticipants. Additionally, as the study was conducted in an academic hospital, the findings may not generalize to other clinical settings. Retrospective EHR‐based collection of disease activity indicators could have introduced bias if disease activity was highly fluctuating. A major impact is unlikely, given the short median 4‐ to 8.5‐day interval with the fatigue assessment, which reflects the preceding 2 weeks. Regarding model construction, we selected the final model based on adjusted R ^2^ rather than formal model fit comparisons. While a more parsimonious model might be preferable for clinical application, we aimed to provide a broader overview of fatigue‐associated factors. Nevertheless, given the sample size and data limitations, our overview is not exhaustive and could benefit from including additional constructs, such as family dysfunction. ^8^ Also, post hoc analyses were not feasible due to sample size. Finally, the cross‐sectional design precludes conclusions about causality. Future longitudinal research is needed to identify children most at risk for persistent fatigue and determine optimal timing for intervention.

On the other hand, this study has strengths. We used a validated fatigue questionnaire with reference values from the Dutch adolescent population, allowing for comparisons with both healthy peers and children with other chronic diseases. ^15^ Using generic instruments for assessing transdiagnostic constructs, rather than IBD‐specific quality of life measures, enhances comparability across conditions. Focusing on the general fatigue subscale of the PedsQL‐MFS allowed for distinguishing fatigue from sleep quality. Finally, our reliance on patient‐reported outcomes highlights the lived experiences of children with IBD, recognizing the well‐established discrepancy between child self‐report and parent proxy‐report. ^48^

CONCLUSION

Fatigue is common in pediatric IBD, even during clinical remission. IBD‐focused clinical factors, reflecting disease severity and activity, showed no or modest associations with fatigue, and did not contribute unique variance in the multivariate model. In contrast, strong associations with transdiagnostic factors such as depressive symptoms, sleep quality, and quality of life underscore the need to routinely monitor not only fatigue but also related psychosocial and lifestyle domains. Based on child self‐reports, our findings highlight the relevance of the patient perspective and support an integrative health model that extends beyond IBD‐focused follow‐up. In the absence of active disease, targeting modifiable transdiagnostic factors may be key to reducing fatigue and its impact. Parallels with other chronic pediatric diseases suggest that interventions developed for other contexts may also benefit children with IBD. Longitudinal research is needed to identify which children are at risk for persistent fatigue, when to intervene, and how transdiagnostic interventions can effectively improve fatigue‐related outcomes.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information