Celiac disease is one of the most common food-related chronic health conditions, affecting up to 1–3% of people worldwide (1). The only currently available treatment is a life-long gluten-free diet, entailing strict avoidance of wheat, barley, rye, and other dietary products containing added gluten (2). Initiation of the diet usually results in rapid clinical and histologic recovery, whereas continuous gluten exposure predisposes to persistent symptoms and various complications (3–6). Notwithstanding the benefits, the extensive dietary restriction increases the risk for unbalanced diet and nutritional deficiencies, and may also cause financial and social burden and have a negative effect on quality of life (7–11). These challenges may lead to dietary lapses, especially in subjects not experiencing symptoms after gluten exposure (12).

Patients diagnosed in childhood comprise a special group in this context. Parents and other adults have usually taken care of their gluten-free diet until adolescence and, in spite of having lived with the diet for many years, patients themselves may be uncertain about the original reason for their diagnosis. Accordingly, adolescents have comprised a particular risk group for difficulties with the diet and non-adherence (13–17). However, data on long-term dietary adherence and associated factors after the transition from pediatric to adult care remains limited (12, 18–20). This information would be especially important to improve treatment outcomes by focusing intensified follow-up and support on those in need (12).

Our aim was to evaluate the above-mentioned issues by comparing large well-defined cohorts of long-term adherent and non-adherent adults diagnosed in childhood with celiac disease.

In total, 237 (42%) currently adult celiac disease patients responded to the questionnaires and comprised the final cohort for the study. Based on patient record data, the responders were more often women (69% vs. 52%, p < 0.001) and relatives of celiac disease patients (56% vs. 44%, p = 0.035) and had less often concomitant T1D (9% vs. 16%, p = 0.029) than the non-responders (n = 322). The groups did not differ in other clinical and histological features at diagnosis (data not shown) or short-term adherence to a gluten-free diet after 6–24 months (strict diet 86% vs. 85%, p = 0.228).

Altogether 186 (79%) of the 237 celiac disease patients diagnosed in childhood reported strict long-term adherence to a gluten-free diet after a median of 18 (range 3–51) years in adulthood. Fifty-one (22%) patients reported dietary lapses and, in detailed analysis, 31 (13%) of them reported occasional lapses, 10 (4%) 1–5 lapses per month, four (2%) weekly lapses and five (2%) did not adhere to the diet at all.

Coexisting T1D at childhood diagnosis was more common in long-term non-adherent than adherent patients (18% vs. 4%, p = 0.003). Adherent and non-adherent patients did not differ in demographic, clinical, and histological data (Table 1), in the nature and severity of symptoms (data not shown), or in the median values of hemoglobin (123 vs. 123 g/l, p = 0.920) and serum autoantibodies (TGA 86 U/l vs. 71 U/l, p = 0.393; EmA 1:500 vs. 1:500, p = 0.912) at diagnosis. The groups also demonstrated comparable short-term dietary adherence (Table 1), frequency of treatment response (99% vs. 100%, p = 1.000), and presence of follow-up (91% vs. 95%, p = 0.534) 6–24 months after the diagnosis in childhood.

Long-term non-adherent patients reported more coexisting T1D, also in adulthood (Figure 1), and all the diagnoses had been set in childhood, either before or concomitantly with celiac disease. Gastrointestinal comorbidities were less common than in adherent patients, but the difference was not significant after adjusting for groups with concomitant T1D. The groups were comparable in other comorbidities in adulthood (Figure 1). Non-adherent patients reported higher median BMI, but this difference was likewise not significant after adjustment for T1D (Table 2). The groups did not differ in current age, time from diagnosis, working/studying status, membership of a celiac society, lifestyle, presence of offspring, and family risk for celiac disease (Table 2).

Long-term non-adherent patients found gluten-free diet more difficult to maintain and more often reported celiac disease-related symptoms, whereas experiences or concerns about health, restrictions caused by the diet and presence of regular follow-up were comparable (Table 3). They had also more overall gastrointestinal symptoms, especially constipation and abdominal pain measured by the GSRS, while there were no differences in other sub-scores or in health-related quality of life measured by the PGWB (Figure 2). Adjustment for T1D did not affect the results despite abdominal pain, in which the difference was no longer significant.

The main decisive factors in both groups for long-term dietary adherence were attempts to avoid symptoms and complications, but these were considered less important and the price of gluten-free products more important among the non-adherent patients (Figure 3). Presence of T1D did not affect the significance of difference in these motivational factors. Specific explanations for ongoing or earlier non-adherence included being a teenager or student (n = 15), poor availability or quality of gluten-free products (n = 14), financial issues (n = 10), absence of gluten-induced symptoms (n = 6), difficulties in adhering to or failure to understand the diet (n = 3), anger or embarrassment due to the diet (n = 6), urge to eat gluten/interest in tasting normal food (n = 5), living abroad (n = 3), work-related eating situations (n = 1) and military service (n = 1).

In a subgroup analysis, adult celiac disease patients with concomitant T1D presented with poorer adherence to a gluten-free diet already in childhood follow-up, as well as reported poorer dietary adherence and more difficulties with the diet, more significant role of availability and price of gluten-free products for their adherence and more regular celiac disease follow-up in adulthood compared to those without T1D (Supplementary Table). The groups did not differ in current or diagnosis age, sex distribution, health experiences or prevalence of persistent symptoms (Supplementary Table).

One of our main findings was the generally good long-term adherence to a gluten-free diet (79% on a strict diet) among adult patients with a childhood diagnosis of celiac disease. Moreover, most patients reported occasional lapses and only 4% reported non-restricted gluten consumption or weekly lapses. So far a few studies with follow-up in adulthood of originally pediatric patients have reported variation in dietary adherence between 36–81% (12, 18–20), whereas those diagnosed in adulthood have shown slightly better long-term figures of 40–98% (19, 26–29). However, comparison between studies is not straightforward given the wide variation in study designs and characteristics of study populations, as well as in methods of measuring dietary adherence (30, 31).

We found long-term non-adherence to be overrepresented among patients with concomitant T1D. Earlier research findings on this issue are inconsistent (32–39), and the primary aim in most of these studies has been to evaluate the impact of gluten-free diet on diabetic control (34–37, 39) rather than to estimate dietary adherence as such (32, 33, 38). It is logical that the co-existence of two chronic conditions requiring continuous monitoring and care would increase the risk for non-adherence. However, these challenges may not emerge before responsibility for treatment shifts from caregivers to patients themselves in adolescence (33, 40), although here the difference was seen already during childhood follow-up. Also, availability and price of gluten-free products seemed to play especially significant role for those with coexistent T1D. In contrast to T1D, presence of a gastrointestinal comorbidity in the present cohort predicted better adherence. This may be attributable to attempts to exclude celiac disease-related symptoms with a strict diet before evaluations of possible co-existing conditions.

Although the non-adherent patients reported more ongoing symptoms, their experienced general health and quality of life were comparable to those who were strictly adherent. This could be partly explained by the observation that they were less often afraid of complications of untreated celiac disease. Ongoing gluten intake is the most common reason for symptom persistence after celiac disease diagnosis (41, 42) and avoidance of symptoms has previously been reported to motivate patients to maintain a strict diet, whereas the role of fear of complications is less clear (12). The relatively young age of our participants may have affected these experiences, as the possible negative long-term consequences of dietary lapses may have not yet have become apparent. Both ongoing symptoms and non-adherence have previously been associated with poorer quality of life, but whether there is a true causality remains unclear (43, 44).

Here, long-term non-adherent patients also found maintaining a gluten-free diet more challenging and reported that the price of products affected their adherence more often. In contrast, availability of products was considered less important, possibly reflecting the good availability in Finland. The high price of gluten-free foodstuffs has also previously been reported to predispose to non-adherence (26, 45, 46). Whether celiac disease patients receive financial support to cover the higher costs varies considerably worldwide (47). In Finland, children with confirmed celiac disease are currently entitled to monthly subsidies, whereas this is not the case among adults. Additional factors possibly explaining differences in the adherence include patients' knowledge about celiac disease and their self-confidence in managing the diet (26, 45, 48).

Excluding T1D, long-term difficulties with a gluten-free diet could not be predicted at celiac disease diagnosis or during childhood follow-up. Although most diabetics were diagnosed by risk-group screening, the baseline clinical features were not associated with adulthood dietary adherence. This is important, as there has been a fear that screening-detected patients might have poorer motivation to accept the diet than those detected due to symptoms (20, 49, 50). However, this was not supported by our findings. In theory, being diagnosed in childhood could promote adherence, as many of these patients have been on a gluten-free diet for almost their whole lives (18). Conversely, an early diagnosis could also predispose to reduced treatment motivation if the reasons for maintaining a restrictive diet have remained unclear (15, 51). Transition to adult care is a special phase when adolescents should assume responsibility for their treatment and cultivate skills for good treatment for the rest of their lives. Therefore, it would seem logical to support patients at that specific point, although more data is needed about the best practices for the transition process (52). In line with our earlier studies (15, 28), presence of adulthood follow-up or membership of a celiac society were not associated with dietary adherence here, whereas in some other countries both of these have been reported to have a positive effect (19, 53). External support could play a greater role if the general circumstances for a strict gluten-free diet are challenging. On the other hand, patients with concomitant T1D reported poorer dietary adherence despite more regular follow-up, suggesting that the physicians might have focused more on T1D during the visits.

Despite the generally good adulthood adherence to a long-term gluten-free diet among these originally pediatric celiac disease patients, particularly those with a co-existing T1D, persistent symptoms and financial issues were at risk for non-adherence and challenges with the diet. These individuals should receive special attention during follow-up, especially during the transition from pediatric to adult care, when the foundations of commitment to life-long treatment are laid. Recognizing the patients' personal motivation factors toward the diet could provide further support for optimized long-term treatment and health outcomes.

The datasets presented in this article are not readily available because, privacy or ethical restrictions. Requests to access the datasets should be directed to laura.kivela@tuni.

The studies involving human participants were reviewed and approved by Regional Ethics Committee of Tampere University Hospital, Tampere, Finland. Written informed consent to participate in this study was provided by the participants or their legal guardian/next of kin.

LK, AE, HH, and KKu designed the study. LK and HH were responsible for statistical analyses. LK and KKu collected the data. LK, AE, and KKu drafted the manuscript and wrote the final version. MR, HH, and KKa reviewed the paper for important intellectual content. All authors interpreted the results, approved the final draft submitted, and agreed to be accountable for all aspects of the work.

This study was supported by the Foundation for Pediatric Research, the Competitive State Research Financing of the Expert Area of Tampere University Hospital, the Päivikki and Sakari Sohlberg Foundation, the Maire Rossi Foundation, the Maud Kuistila Foundation, the Mary and Georg Ehrnrooth Foundation, the Paolo Foundation, the Emil Aaltonen Foundation, the Finnish Celiac Society, the Sigrid Juselius Foundation, and the Academy of Finland.

LK, KKa, and KKu have received personal lecture fees from the Finnish Coeliac Society outside the submitted work and serve as members of the advisory committee of the Finnish Coeliac Society. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.