One-Year Outcomes Among Children Identified With Celiac Disease Through a Mass Screening Program

Abstract

BACKGROUND & AIMS:

Celiac disease (CD) mass screening remains controversial in part because of a paucity of data to support its benefit. The Autoimmunity Screening for Kids study is a mass screening study for pediatric CD and type 1 diabetes in Colorado.

METHODS:

This study prospectively follows up children ages 1 to 17 years who screened positive for tissue transglutaminase IgA autoantibodies in the Autoimmunity Screening for Kids study subsequently referred for diagnostic evaluation. Children diagnosed with CD by biopsy or serologic criteria were included in this study. Evaluation at baseline and 12 month follow-up evaluation included demographics, laboratory studies, symptoms, health-related quality of life, anxiety/depression, and gluten-free diet adherence. Paired Student t test, chi-square, and Wilcoxon signed-rank tests compared baseline and follow-up data. For symptom scores, odds of improvement were assessed.

RESULTS:

Of the 52 children with CD enrolled, 42 children completed 12-month follow-up evaluation. On the symptom questionnaire completed at diagnostic evaluation, 38 of 42 children reported 1 or more symptoms. CD mean symptom severity and frequency scores improved from baseline to follow-up evaluation (P < .001). Reported health-related quality of life scores improved among caregivers (P = .002). There was no significant change in reported anxiety or depression. Iron deficiency without anemia was common at baseline (21 of 24 children; 87.5%) and normalized at follow-up evaluation (11 of 21 children; 52.3%). Twenty-six of 28 families reported good or excellent gluten-free diet adherence.

CONCLUSIONS:

This novel study of children with CD identified through a mass screening program demonstrated improvement in symptoms, quality of life, and iron deficiency after 1 year follow-up evaluation. This demonstrates that there may be benefit to CD mass screening.

Celiac disease (CD) is a chronic autoimmune enteropathy triggered by exposure to dietary gluten in genetically predisposed individuals.¹ It is one of the most common pediatric autoimmune diseases with a prevalence as high as 3% in some regions.^2–4 Because many children now present with subclinical or unrecognized features, a case finding approach limited to the presence of symptoms often leads to diagnostic delay.^5–8

Despite increasing prevalence and an often delayed diagnosis, mass screening for CD remains controversial. Although several pediatric gastroenterology societies support targeted screening of asymptomatic individuals at higher risk of CD, the US Preventative Services Task Force has concluded that there is insufficient evidence to recommend screening for CD in asymptomatic individuals owing to the lack of North American studies on the risks and benefits of screening this particular population.^9–11

Several European studies have suggested a benefit to diagnosing and treating screening-identified asymptomatic individuals.^12–18 No such studies have been con-ducted in the United States. The Autoimmunity Screening for Kids (ASK) Study, a pediatric general population screening study for CD and type 1 diabetes (T1D) in Colorado, is designed to assess the benefits and cost effectiveness of the screening. ASK has reported previously that 2.4% of Colorado children had previously undiagnosed CD autoimmunity.¹⁹ The aim of the current study was to evaluate health-related outcomes of children identified with CD through ASK and followed up for 1 year on a gluten-free diet (GFD).

Methods

Since January 2017, the ASK study has screened for CD and T1D in more than 33,000 Colorado children ages 1 to 17 years. The screening takes place in the context of routine health care at private pediatric practices, and clinics of the Children’s Hospital Colorado and Denver Health networks. This protocol was approved by the Colorado Multiple Institutional Review Board. Eligible participants were screened for tissue transglutaminase IgA autoantibodies (tTGA) to detect CD autoimmunity and for islet autoantibodies to detect presymptomatic T1D.^20,21 Children who screened positive were invited back for a confirmation; children who had confirmed tTGA more than 2 times the upper limit of normal were referred to the Colorado Center for Celiac Disease for further diagnostic evaluation. The initial and confirmation screening protocol have been detailed in previous publications.¹⁹ The study described here included 52 children with a confirmed tTGA through the ASK study, who were seen at the Colorado Center for Celiac Disease between January 2019 and March 2022, and later were diagnosed with CD based on biopsy criteria as described by the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition or nonbiopsy criteria as described by the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition (ESPGHAN).^9,11 ESPGHAN criteria include a tTGA greater than 10 times the upper limit of normal and a positive endomysial IgA on a separate blood sample. See Supplementary Figure 1 for a flow chart describing the number of children who completed each of the measures outlined later included in this study. Enrollment was completed in conjunction with their first celiac clinic visit before their confirmed diagnosis and treatment with a GFD. Data were collected at a baseline assessment before treatment and 12 months later after treatment with a GFD. Exclusion criteria included those who were not fluent in English or Spanish or had developmental delays that precluded completion of the study surveys.

Demographic Data

Demographic data including age, family history of CD, and medical insurance were collected from the medical record. Caregiver education status, household income, race, and ethnicity were queried in participant surveys at the time of enrollment.

Gluten-Free Diet Adherence

Patients diagnosed with CD meet with a registered dietitian trained in the assessment of adherence to a GFD using the center’s established standardized dietitian evaluation (Supplementary Methods 1).^22,23 Total scores range from excellent (score, 0) to very poor adherence (score, ≥ 28). Descriptive statistics were used to report the proportion of children in each category.

Laboratory Tests

Baseline and follow-up laboratory tests were obtained per routine clinical care.²⁴ Included in this analysis were tTGA, ferritin, hemoglobin, mean corpuscular volume (MCV), aspartate aminotransferase (AST), alanine transaminase (ALT), 25-hydroxy(OH) vitamin D, and thyroid stimulating hormone (TSH) obtained within 3 months of the baseline and follow-up visit.

Symptoms

At the time of the first ASK screen, gastrointestinal symptoms including diarrhea, frequent stomachaches, constipation, vomiting, difficulty gaining weight, and poor growth were assessed over the prior 3 months. This symptom questionnaire was administered before the children or caregivers were aware of their autoantibody status.

An extended symptom questionnaire that queried both gastrointestinal and extraintestinal manifestations of CD was administered at the baseline clinic visit before diagnostic confirmation and again at the 12-month follow-up assessment. The families were aware of their positive test result through the ASK study at the time the extended symptom questionnaire was administered (Supplementary Table 1)

Psychosocial Measures

Family impact and health-related quality of life.

To evaluate the child and caregiver’s perceptions of the child’s physical and mental health over time, the standardized measure, the Pediatric Quality of Life Inventory 4.0 Generic Core Scale (PedsQL; child self-report and caregiver proxy-report) was chosen because of its established reliability and validity, as well as its general applicability leading to comparability across both different cultures and medical conditions (Supplementary Methods 1). The 23-item measure yields a total health-related quality of life score and subscale scores of physical, emotional, social, and school-related quality of life.²⁵

To evaluate the impact of the child’s condition on the family and caregivers, the PedsQL Family Impact Module (FIM) was chosen.²⁶ Higher scores indicate better health-related quality of life and functioning.

Mental health.

The Revised Children’s Anxiety and Depression Scale (RCADS; child self-report and caregiver proxy-report) was used to assess specific domains of anxiety and depression symptoms in participants ages 8 to 17 years old.²⁷ For the description of prevalence of anxiety and depression, T scores of 70 or higher were considered above the clinical threshold for anxiety and depression, T scores of 65 to 69 were considered borderline, and scores less than 65 were considered average. For comparison of anxiety and depression before and after the initiation of a gluten-free diet, this score was treated as a continuous variable.

Statistical Analysis

Descriptive analyses were used to examine baseline and follow-up data. In addition, to summarizing the mean, median, and proportion of pertinent features for the entire cohort, analyses also dichotomized patients with gastrointestinal symptoms at the initial ASK screen to those without symptoms at their initial screen. The proportion of children in each category of GFD adherence also was determined. Paired Student t test, chi-square, and Wilcoxon signed-rank tests were applied as appropriate. Baseline laboratory tests including tTGA, ferritin, hemoglobin, MCV, AST, ALT, 25-OH vitamin D, and TSH were evaluated both as categorical (abnormal vs normal based on the laboratory-provided upper limit of normal) and continuous variables.

To calculate an overall symptom frequency and severity score, the reported ordinal scale ranging from 0 to 4 (with 0 indicating not present) for each symptom frequency and severity was summed across all 15 symptoms for a total possible score of 60. The odds of improvement in these symptom measures were assessed using an ordinal logistic regression model. A generalized estimation equation was used to account for the correlation between baseline and 12-month measurements from the same child. This model was used to evaluate the odds of having an improvement in symptom severity or frequency with an odds ratio greater than 1 denoting that there was symptomatic improvement from the baseline to the follow-up visit.

The psychosocial measures (FIM, the PedsQL, RCADS) were scored using previously described methods. The FIM and PedsQL scores were treated as a continuous variable whereby higher scores indicated better functioning. Depending on the comparison, the RCADS scores were treated as both continuous and categorical variables whereby higher scores denoted worse functioning and also could be categorized as average, borderline clinical, and above the clinical threshold with respect to concerns for anxiety or depression. Statistical analyses were performed using SAS version 9.4 (SAS Institute, Inc, Cary, NC). Because of the exploratory nature of the study, no adjustments for multiple outcomes were used. All P values and 95% CIs are reported for all the a priori comparisons. P < .05 was deemed statistically significant.

Results

Cohort Characteristics

As of March 31, 2022, there were 52 children identified through the ASK study with confirmed CD enrolled in our follow-up protocol and 42 children completed 12 months of follow-up evaluation. Ten children were lost to follow-up evaluation (Table 1). Twenty-five children were confirmed to have CD by intestinal biopsy (Marsh score, ≥2) and 17 children by ESPGHAN serologic criteria.^9,11 Children who did not complete the follow-up evaluation were more likely to be of male gender (70% vs 33.3%, P = .03), were more likely to report non-White race (40% vs 4.8%; P = .005), were more likely to have a lower estimated household income (66.6% vs 35%; P = .02), were more likely to have public health insurance or to not be insured (70% vs 11.9%; P = .03), and to be older at the age of enrollment in the study (mean age, 12.5 vs 9.5 y; P = .03). There was no difference in the mode of diagnosis between those who did and did not complete follow-up evaluation (P = .45).

Table 1.

Baseline Characteristics of Children Who Did or Did Not Complete Follow-Up Evaluation

	Completed follow-up evaluation (n = 42)	No follow-up evaluation (n = 10)	P valuea
Mean age at enrollment, y (SD)	9.46 (3.99)	12.54 (3.84)	.0317
Female gender	28 (66.7%)	3 (30.0%)	.0337
Family history of celiac disease	10 (23.8%)	1 (10.0%)	.9410
Race
White	40 (95.2%)	6 (60.0%)	.0056
African American	1 (2.4%)	1 (10.0%)
Unknown	1 (2.4%)	3 (30.0%)
Household income
>$150,000	13 (32.5%)	3 (33.3%)	.0204
$100,000–$149,999	13 (32.5%)	0 (0%)
$50,000–$99,999	10 (25.0%)	2 (22.2%)
$25,000–$49,999	1 (2.5%)	3 (33.3%)
<$25,000	3 (7.5%)	1 (11.1%)
Public insurance/uninsured	5 (11.9%)	7 (70.0)%	.0329
Presence of gastrointestinal symptoms	13 (31.0%)	6 (60.0%)	.0865
Biopsy at diagnosis	25 (59.5%)	6 (60.0%)	.4517

^aContinuous variables were compared using the Student t test and categoric variables were compared using chi-square analyses.

Of the 42 children who completed follow-up evaluation, 13 children reported 1 or more symptoms on the ASK initial questionnaire. Ten children who completed follow-up evaluation also reported a family history of CD and 7 children reported a family history of T1D.

Gluten-Free Diet Adherence

Twenty-eight families completed the follow-up adherence assessment with the dietitian. Most (26 of 28) families reported good or excellent adherence and all of these children had a down-trending tTG IgA level at the time of follow-up evaluation (Supplementary Table 2).

Laboratory Tests

Most baseline clinical laboratory tests including hemoglobin (40 of 40; 100% normal), MCV (40 of 40; 100% normal), ALT (37 of 39; 94.9% normal), AST (34 of 39; 87.2% normal), vitamin D (35 of 40; 87.5% normal), and TSH (35 of 36; 97.2% normal) values were normal. Interestingly, 1 child had a negative tTG IgA value based on repeat clinical testing and a positive confirmatory biopsy. With respect to other celiac serologies, 24 of 26 children (92.3%) had a positive endomysial antibody IgA value at baseline.

Although many of the laboratory parameters showed statistically significant improvement from the baseline to the follow-up visit (Supplementary Figure 2), these improvements were not clinically relevant because most laboratory parameters were not outside normal cut-off values at baseline. Ferritin, which is indicative of iron deficiency, was low for a majority of children at the baseline visit (21 of 24; 87.5% abnormal) and did improve for most at the follow-up visit (11 of 21; 52.3% normalized at follow-up evaluation).

Symptoms

Of the 42 children who completed the follow-up evaluation, 29 children reported no gastrointestinal symptoms on their initial ASK screening questionnaire. However, 93% (27 of 29) of these children reported 1 or more symptoms on the extended symptom questionnaire at their baseline visit to the Colorado Center for Celiac Disease. There was no difference in baseline demographics, health-related quality of life (HRQoL), and anxiety or depression based on the presence of gastrointestinal symptoms on the initial screening questionnaire. Overall, 38 of 42 children who completed follow-up evaluation reported 1 or more symptoms on the initial extended screening questionnaire at their baseline assessment at the Colorado Center for Celiac Disease and 30 of 38 children reported an improvement in 1 or more symptoms at their follow-up assessment.

The mean change in the symptom severity and frequency scores improved from baseline to follow-up evaluation, denoting overall lower scores at the follow-up assessment (P < .001 and P < .001, respectively). Similarly, both symptom severity and frequency improved based on the overall odds of improvement from the ordinal logistic regression model (Figure 1, Supplementary Figure 3). Of children completing the symptom survey at both time points, 85.5% of children had improvement in symptom frequency and 79.5% had improvement in symptom severity at follow-up evaluation compared with baseline. When examining the odds of improvement of individual symptom frequency, joint pain, headaches, brain fog, itchiness/rash, gassiness, decreased energy, irritability, constipation, abdominal pain, and diarrhea all improved significantly. Headache, itchiness/rash, gassiness, decreased energy, irritability, constipation, abdominal pain, and diarrhea all improved significantly with respect to symptom severity.

Figure 1.

Forest plot demonstrating symptom improvement from baseline to follow-up evaluation. (A) Odds of improvement in symptom frequency with respect to the individual queried symptoms. A positive odds of improvement in this score (to the right of the dashed line) indicates improvement in this score from the baseline to the follow-up visit. (B) Odds of improvement in symptom severity with respect to the individual queried symptoms. Brain fog and vomiting were omitted from this graphical representation owing to large CIs.

Psychosocial Measures

Family impact and health-related quality of life.

Overall, caregivers (n = 39) reported no significant changes in their own physical, emotional, social, and cognitive functioning, or in their communication or worry about their child’s health (P = .0861). Addition-ally, no significant changes were endorsed regarding the impact of their child’s health on the family’s daily activities or familial relationships from baseline to follow-up assessment.

For screening-identified children enrolled in this study, their HRQoL demonstrated improvement from the baseline to follow-up visit. Caregivers reported a significant improvement in the child’s overall HRQoL and in almost all subdomains of functioning (physical, emotional, and school functioning; P < .0001, .004, and .012, respectively), except for social functioning (P = .107). Although the child self-reported scores all trended toward improvement from the baseline to the follow-up assessment, only the school functioning subscale was statistically significant (P = .030) (Figure 2).

Figure 2.

Forest plot of improvement in the Pediatric Quality of Life Inventory 4.0 Generic Core Scale (PedsQL) scores from baseline to follow-up evaluation. This forest plot shows the mean change in the PedsQL total and subdomain scores for child and caregiver reported measures. Scores to the right of the dashed line indicate improvement in quality of life.

Mental health.

Overall, caregiver proxy- and child self-reported symptoms of anxiety and depression did not differ from screening to follow-up evaluation (Figure 3). One child with a clinically significant depression score at baseline improved at their follow-up assessment and 1 child developed a clinically significant anxiety score at follow-up evaluation. These results did not change based on the presence or absence of symptoms on the initial ASK screening questionnaire.

Figure 3.

Spaghetti plots comparing Revised Child Anxiety and Depression Scores at baseline and follow-up evaluation. These spaghetti plots detail the changes in total anxiety and depression scores.

Discussion

Overall, this novel US study examining the outcomes of children with mass screening-identified CD demonstrated improvement after screening and treatment with respect to initially unrecognized symptoms, quality of life, and iron deficiency. Mass screening–identified children also reported excellent to good adherence to the GFD and most did not report experiencing an increase in anxiety or depression symptoms after treatment. This overall indicates that diagnosis and treatment of CD through a mass screening program was beneficial for enrolled children.

Although most children (29 of 42) initially were deemed asymptomatic, nearly all of them (27 of 29) had identifiable symptoms on an extended symptom screening questionnaire before treatment and many of these symptoms improved with treatment with a gluten-free diet. This study parallels the findings of European screening studies that highlighted the prevalence of un-identified symptoms and the limited utility of symptoms in identifying those with undiagnosed celiac disease.^14,18,28 It also corresponds with an earlier Dutch screening study that showed that screening-identified children experienced improvement in symptoms with a GFD.¹⁵ These children with subclinical, unrecognized symptoms would not be identified by our current case finding approach. Even with truly asymptomatic children, other studies have suggested that there still may be benefit with respect to long-term health outcomes such as bone health.^13,17 Our study found that children also may have iron deficiency, which subsequently improved with treatment with a GFD.

Health-related quality of life similarly improved after treatment with a GFD in children with screening-identified CD. This study of pediatric screening-identified CD in the United States indicates this improvement and supports prior findings from multiple pediatric and adult studies in Europe.^14,15,29 Attention to the perceived degree of improvement for caregivers and children and variation in the improvement in health-related quality of life based on the age of the child should be addressed in future analyses. Although not statistically significant, social functioning declined with the GFD in 10 caregivers and 7 children in our study. This could reflect the difficulty of the GFD in social settings, the availability and expense of gluten-free products, and social support networks.^14,30 These are surmountable barriers for which policy greatly could improve the quality of life for individuals with CD more broadly. Despite the social impacts of a GFD, the majority of families in this study reported good to excellent adherence, similar to reports of dietary adherence of screening-identified adults in the United States.²⁹ Screening and subsequent treatment of pediatric screening–identified CD also were not associated with increased anxiety and/or depression symptoms for most children.

This observational study had several limitations, including loss to follow-up evaluation from the initial screening visit to diagnosis and follow-up evaluation. There was a risk of attrition bias because the individuals who benefited most from screening and treatment may have been more likely to complete follow-up evaluation. In our study, minority children and those of lower socioeconomic status were more likely to be lost to follow-up evaluation. However, this should not detract from the overall finding that a majority of children who did complete follow-up evaluation improved after screening and treatment.

Another limitation was the timing of the coronavirus disease-19 pandemic. We were able to quickly convert and provide care by telehealth, but some outcomes such as anthropometric measures and laboratory parameters, which required in-person collection, were limited. Thus, there was the potential that some of the missing data created bias in our results. However, all of the questionnaires were able to be collected electronically, resulting in a relatively complete data set for health-related quality of life and celiac disease–related symptoms. The coronavirus disease-2019 pandemic also may have had unmeasurable impacts in symptoms, health-related quality of life, and mental health, which, unfortunately, cannot be accounted for in our analyses. With respect to the reporting of symptoms, knowledge of a positive test result may have influenced the reporting of symptoms and also may have contributed to initially asymptomatic individuals later reporting symptoms. Finally, longer-term health-related outcomes will need to be assessed and compared with those not screened to help distinguish the complete benefits of celiac disease screening.

Overall, the follow-up evaluation of children identified to have CD through the Colorado ASK screening program suggests that mass screening and treatment improves previously unrecognized symptoms, health-related quality of life, and iron deficiency while not increasing symptoms of anxiety or depression. Additional studies are needed to determine the most cost-effective approach to mass screening and how mass screening compares with current targeted screening practices with respect to outcomes before policy changes can be considered in support of celiac disease mass screening.

Supplementary Material

Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at www.cghjournal.org, and at http://doi.org/10.1016/j.cgh.2024.03.030.

What You Need to Know.

Background

There currently is insufficient evidence to recommend population screening for celiac disease. The Autoimmunity Screening for Kids study is a pediatric mass screening program for type 1 diabetes and celiac disease in Colorado.

Findings

One-year follow-up evaluation of children diagnosed with celiac disease through the Autoimmunity Screening for Kids study who were treated with a gluten-free diet demonstrated an improvement in symptoms, quality of life, and iron deficiency.

Implications for patient care

Pediatric mass screening for celiac disease may be beneficial, but more data are needed on long-term outcomes and the cost effectiveness of screening before mass screening can be implemented.

Abbreviations used in this paper:

ALT

alanine transaminase

AST

aspartate aminotransferase

ASK

Autoimmunity Screening for Kids

CD

celiac disease

ESPGHAN

European Society for Pediatric Gastroenterology, Hepatology, and Nutrition

FIM

PedsQL Family Impact Module

GFD

gluten-free diet

HRQoL

health-related quality of life

MCV

mean corpuscular volume

PedsQL

Pediatric Quality of Life Inventory 4.0 Generic Core Scale

RCADS

Revised Children’s Anxiety and Depression Scale

TSH

thyroid-stimulating hormone

tTGA

tissue transglutaminase IgA autoantibodies

T1D

type 1 diabetes