Health-related quality of life in adolescents with screening-detected celiac disease, before and one year after diagnosis and initiation of gluten-free diet, a prospective nested case-referent study

Celiac disease (CD) is a chronic disorder in genetically predisposed individuals in which a small intestinal immune-mediated enteropathy is precipitated by dietary gluten [1]. The prevalence has generally been suggested to be around 1% [2‐4], however more recently it has been shown that the prevalence has risen [3, 5, 6] and most of the people with CD are actually undiagnosed [7‐10]. CD can be difficult to diagnose because the signs and symptoms may be absent, subtle, or not recognized as CD related and therefore not prompt testing within routine clinical practice [11]. The best strategy for finding those with unrecognized CD is still debated, but options include: active case finding through routine clinical practice with a broader consideration for variety in presentation, testing groups at high risk (e.g., first degree relatives of those with CD or those with other autoimmune disorders), or a public health intervention that would involve screening the general population [3, 7, 10, 12, 13].

Recently, Aggarwal et al. presented a review of literature that addresses how screening, diagnosing, and treating (gluten-free diet) asymptomatic individuals affects their quality of life [3]. Their overview illustrates that there is a lack of consensus on whether or not the gluten-free diet (GFD) improves the quality of life in “asymptomatic screening-detected” individuals [3]. Research addressing the quality of life (QoL) and health-related quality of life (HRQoL) for people with CD is mostly based on those diagnosed through routine clinical practice or selected for screening because they are considered at high risk [14‐28]. The HRQoL for individuals who have received their CD diagnosis from a screening of the general population could differ from these groups and this study focuses on adolescents diagnosed with CD as a result of screening study. The aim of this study is to investigate the HRQoL of adolescents with screening-detected CD before and one year after diagnosis and treatment.

The HRQoL for the adolescents with screening-detected CD, as reported on the EQ-5D instrument, is similar to the HRQoL of the referents, both before and one year after diagnosis and initiation of the GFD, except for in the dimension of pain at follow-up, where fewer cases reported problems than referents (12.6% and 21.9% respectively, Adjusted OR 0.50, 95% CI 0.27-0.94) (Table 1). However, when this dimension was stratified by sex, it was revealed that the difference was between boy cases and referents at follow-up. In the sex stratified results there was a significant difference between the boys in the pain dimension at follow-up, where fewer of the screening-detected boys reported problems (4.3%) than the boy referents (18.8%) (Adjusted OR 0.17, 95% CI 0.04-0.73) (Table 2).

In the anxiety dimension, both the cases and referents had an increase (from baseline to follow-up) in the proportion of adolescents who reported problems, although these were not significant changes (Table 1). When stratified by sex, there was no significant difference in proportion of problems reported for anxiety for the boys or girls between the cases and referents (Table 3).

Adolescents with screening-detected CD had a median VAS score of 91 at baseline and 90 at follow-up (Table 4), which was not a significant change (Wilcoxon signed rank test, p value = 0.92). The referents median VAS score was 90 at baseline and at follow-up (Table 4). Comparisons of cases to referents at baseline and at follow-up showed no significant differences (Mann–Whitney U test, not shown).

When comparing boys and girls within the case and referent groups, the only significant difference was between the boys and girls in the referent group at follow-up (Mann–Whitney U test, p value = 0.01, not shown), in which both report a median of 90 but for the boys and girls the 25^th percentile is 85 and 75 (respectively) and the 75^th percentile is 98.75 and 97 (respectively).

The HRQoL for adolescents with screening-detected CD, as self-reported on the EQ-5D, is similar to that of their peers, both before and one year after diagnosis and GFD, except in the dimension of pain at follow-up. A sex stratified analysis revealed that this is due to the difference between boy cases and referents at follow-up, in which 4.3% of the screening-detected boys report problems and 18.8% of the referent boys report problems (Adjusted OR 0.17, 95% CI 0.04-0.73) (Table 2).

We present a unique study in which adolescents with screening-detected CD report their HRQoL before and one year after diagnosis and treatment. A strength of our study is that the adolescents report baseline HRQoL before knowledge of their CD diagnosis, unlike many studies in which patients are asked to recall how they felt at the time of their diagnosis. We have previously published baseline data, including participants from the same screening study, in which there was also no significant difference in HRQoL for those with unrecognized CD compared to their peers without CD at the time of the screening [34]. However, in this current study we explore the adolescents’ HRQoL before and after diagnosis and treatment.

Another strength of this study is the fact that the adolescents had their CD detected as a result of a screening study, and not clinically or because they were considered high risk. This means that our results may be more reflective of those living with unrecognized CD, which has been seen as a limitation in other studies that attempt to address the QoL for those with undiagnosed CD [3]. In a review of literature that addresses how screening, diagnosing, and treating “asymptomatic screening-detected individuals” affects their QoL, done by Aggarwal et al. [3], most of the studies involved CD patients who were identified from high risk groups [19, 26] or were compared based on what type of symptoms led to their diagnosis; i.e., typical, not typical, or those who reported they had not experienced any symptoms [23, 24, 28]. However, two of the studies included individuals that could be considered as screened from the general population [3, 35, 36]. In those studies, the individuals who had “typical” symptoms showed an improvement in QoL scores after one year on the GFD, while those who were supposedly symptom free had scores comparable to healthy controls at baseline and at follow-up, [35, 36] similar to the screening-detected adolescents in our study.

Even though the EQ-5D instrument is a validated tool and we have used the Swedish child-friendly pilot version, this tool may have limitations in the context of this study. Perhaps, it is not the ideal tool for capturing problems with subtle symptoms or feelings caused by unrecognized CD. Also, these adolescents may have adapted to their current health situation as normal and at baseline rated their health status as high as possible and similar to their peers. If they did experience improvement after diagnosis and treatment, they would be unable to demonstrate improvement from the high health status previously reported. These screening-detected adolescents were also invited to participate in another follow-up study, where they participated in focus group discussions (n = 31) and wrote narratives (n = 91), and when they were asked specifically about change in well-being after diagnosis, 53.8% reported feeling better [37]. In that study, it was also shown that some of the screening-detected children only realized they had been experiencing symptoms after they had been diagnosed and treated [37].

Although it is beyond the scope of this study to provide an explanation, the finding that fewer screening-detected boys (4.3%) reported problems at follow-up in the dimension of pain than the boy referents (18.8%) is interesting to consider and warrants further investigation. In a previous Swedish study, a doubled risk for symptomatic CD in girls compared to boys was found [38]. In this screening study, the male to female ratio for those with clinically diagnosed CD was 1:2 compared to 1:1 for those with screening-detected CD [8]. This difference reveals that, at the time of the screening, a larger proportion of girls compared to boys had already been diagnosed with CD in routine clinical practice [8]. It could be that the boys with unrecognized CD were further progressed in their disease at baseline because they were not as likely as the girls to have been found in clinical practice. One could speculate that, even though they may not have realized the extent of their problems at baseline, they experienced benefit from the treatment resulting in fewer of these boys reporting problems with pain than the boys in the referent group, a phenomenon which was not captured for the girls. In studies involving adults, men and women have been shown to experience the burden of CD differently [27, 39, 40]. There is also research that shows men and women access health care differently [41‐43] and perhaps the boys were less likely to seek or receive care. It is also a possibility that clinicians expect girls to have a higher risk of developing CD and more readily recognize and diagnose girls.

In our study there were no significant differences in the anxiety dimension between cases and referents, suggesting that for these screening-detected adolescents the CD diagnosis and GFD have not caused excess anxiety (at least at one year follow-up). Other studies have shown that those with a strict GFD can have the same HRQoL as healthy children/adolescents [44, 45]. However, in the qualitative study mentioned previously [37], it was revealed that some of these adolescents perceived the GFD and lifestyle changes as inconvenient and causing feelings of stigma while others adapted well to the disease.

The findings of this study suggest that the adolescents with unrecognized CD experience similar HRQoL as their peers without CD, both before and one year after the diagnosis and initiation of the GFD, except for boys in the dimension of pain at follow-up. The screening-detected boys seem to benefit, because fewer report problems with pain than the boy referents at follow-up, one year after diagnosis and initiation of the GFD. However, further research is needed to explore why the screening-detected boys seem to have a different experience from the referent boys and from the screening-detected girls. If considering general population screening for CD in the future, more research is also needed to learn more about the benefits or drawbacks of early diagnosis, long term consequences of untreated CD, and the health economic implications of population screening.