Introduction

Celiac disease (CD) is an immune-mediated chro nic enteropathy, produced by the interaction of certain haplotypes (genes) of the host that provide the genetic susceptibility¹ and the environment, which contributes the prolamins of gluten in the diet². Other factors, such as epigenetics and host microbiota, would participate in the pathogenesis of CD through less-known mecha nisms³. It is currently recognized as a high-frequency autoimmune condition, which appears at any age⁴, re aching a global prevalence of 1-3% of the population⁵. It is diagnosed by serology (measurement of antibodies against tissue transglutaminase 2(TTG), anti-endomysial (EMA) antibodies or deamidated gliadin peptides (DGP) and histological confirmation in duodenal biopsies, which show various degrees of lymphocytic infiltration, crypt hyperplasia, and villous flattening^4,6. By now, the only effective treatment for CD is the glu ten-free diet (GFD), strict, permanent and lifelong⁴. This diet eliminates wheat, rye, and barley from the diet, included in both natural and processed foods⁷. Oats are also usually eliminated because it has a high probability of being contaminated⁸. The main limita tion of this treatment is the lack of adherence of pa tients to GFD.

The treatment of CD includes patients follow-up, periodic and long-term, to evaluate the clinical course and adherence to GFD. This evaluation is difficult be cause there is no reliable and reproducible method to measure the actual gluten intake. The current strategy is based on the consensus of specialists⁴. They recom mend periodic check-ups that estimate the adheren ce based on what the patient reports and the measu rement of antibodies that reflect autoimmunity. The most used are TTG^9,10. It is expected that, by not eating gluten, patients will show blood clearance and progres sive reduction of the intestinal mucosa damage and autoimmunity phenomena. However, the usefulness of these antibodies for follow-up has currently under discussion^11-13, exposing again the problem of how to ensure that the information provided by the patient is accurate and reliable. Foreign data show that adhe rence to GFD in children is 80-95% and in adolescents and adults 36-52%^14,15. Assessing adherence through interviews by specialized professionals and validated questionnaires is giving promising results, but not yet definitive ones^12,16. In Chile, studies indicate that pediatric patients who adhere to GFD vary between 24- 34%^17-19, without significant variations over time. The World Health Organization (WHO) has described five types of factors that influence the adherence to a chro nic treatment: related to the disease, the patient, the treatment, the health system or health team, and the socioeconomic characteristics. The information about how these factors affect the celiac patient and the GFD is insufficient. Considering that the available data in our country are scarce, show low degrees of adherence and that in the national public health system the fo llow-up is by serology (TTG and occasionally EMA), the objective of this study was to evaluate the factors that influence in adherence to GFD. Additionally, the concordance between the measured antibodies (TTG and EMA) and the nutritional adherence questionnai re was evaluated.

Patients and Methods

Results

The universe of patients (2008-2016) with diag nosis of CD was 101, which represents 0.04% of the new consultations made at the HRR during the period²³. 86/101 patients met the inclusion criteria and 65 (75.6%) agreed to enter the study. 61.5% were girls, a percentage higher than in the general popula tion in that age group²⁴. 69.2% of patients were eutrophic, 10.8% at risk or malnourished, and 20% were overweight or obese. In the assessed age group, the percentage of eutrophic/at nutritional risk was higher and the percentage of obesity was lower than in the general population²⁵. 86% of the patients lived in the Metropolitan Region, 93.8% was using Fonasa (public health system) as health system and 86% was in middle socioeconomic level (SEL), according to Esomar 2000 classification. 50.8% of the patients were preschool children or had completed 4th grade or less. 66.1% of mothers and 55.4% of fathers had completed secon dary education or higher. Table 1 shows the variables related to the patient and his/her disease and table 2 the knowledge, disease management, and lifestyle. 52.3% of the interviewed patients (or guardians) stated that the GFD was handled correctly. There was 100% of concordance between the responses of adolescents and their parents or guardians. 43.1% (28 patients) received at least one medication permanently, mainly levothyroxine, insulin, zinc sulfate, polyvitamins, fe rrous sulfate, folic acid, and valproic acid. 6.2% repor ted receiving food in religious ceremonies, 10.8% did not commune due to fear of getting sick, and 43% had doubts about their first communion.

Table 1 Variables related to the child and his/her illness. 

Table 2 Variables related to knowledge on celiac disease and life style. 

Graph 1 shows the frequency of adherence/transgression to the GFD measured by the presence of antibodies EMA +/-, TTG +/-, EMA and TTG +/-, and EMA or TTG +/-. According to the adherence ques tionnaire to the GFD (scores 3 and 4 of Biagi), 30.8% adhered strictly to the GFD. There was concordance between the questionnaire and both tests performed and their combinations, being all P significant with modera te concordance force (table 3). In the analysis of adherence defined by serology (chi-square), only three variables were significant: “age of disease onset” (p = 0.049), “perception of doing well the GFD” (p = 0.002) and “behavior of the pa tient in front of food in social events”(p = 0.005). Assessed by serology at the time of the interview, those diagnosed be fore two years of age, between 2-10 years and after 10 years old, 60.7%, 28% and 33.3% were adherent to the GFD, res pectively. At the time of the interview, 63.3% of those who “think that they are correctly following GFD” had negative both antibodies. Among patients whose guardian thinks that they are making mistakes in the GFD, 75.9% had some positive serology, but 24.1% of the patients, although they stated to transgress the regimen, maintained both blood antibodies negative. The third significant variable was the “habits that patients have at birthdays and/or social events” (p = 0.005). Those who choose only gluten-free foods or those who take food from their homes had a significantly higher frequency of negative TTG and EMA (adherence) than those who do not follow the diet on social occasions. 63.2% of those who choose “only gluten-free” and 57.1% of those who take food from home, had negative serology. 8/9 (88.9%) patients who “eat everything” had some positive test (EMA or TTG). Meanwhile, 90% of patients who do not attend birthdays had positive serology (EMA or TTG). In relation to EMA, there were two significant variables: “read ingredients in the nutrition labeling” (p = 0.041) and “feel welcomed by the foundations supporting the ce liac patient” (p = 0.045), but these variables were not significant for the combination EMA-TTG. The self perception analysis of the GFD and the Biagi question naire showed 72% of concordance. 10/11 patients in whom there was no concordance thought they were doing well with the diet and the Biagi questionnaire classified them as transgressors.

Graph 1 Percentage of patients adhering or not adhering gluten free-diet (GFD), defined by negative serology (TTG* and EMA**). *antitransglutaminase 2 antibo dies. **anti endomysial antibodies. 

Table 3 Concordance serology/autoimmune antibodies EMA and TTG and Biagi's adherence questionnaire*. 

The logistic regression provided a predictive model that included three variables: “perception of following well the GFD” (p = 0.003), “patient’s behavior in front of foods in social events” (p = 0.003) and “age of disea se onset” (p = 0.049). In the logistic regression model only two out of the three variables reached statistical significance (table 4).

Table 4 Logistic Regression with EMA-TTG, OR, adjusted. 

The GFD Knowledge Score did not show signifi cant differences for adherence by serology. When divi ding patients into those with a score of < 3 and those with a score of 4 in the Biagi questionnaire, half of the patients were grouped in each group. Patients with Score < 3 had 1.5 times more risk of transgressing the diet. Although this variable was not significant in the chi-square analysis, due to the importance of patient education, it was included in the multivariable logistic regression analysis.

Discussion

The adherence to GFD in this study is lower than the one published in the international literature for this age group^14,19, but similar to the ones described in our country in previous studies^17-19, both by measurement of TTG and EMA antibodies and the questionnaire de veloped by Biagi. This indicates that the adherence to GFD among our pediatric patients is not only low but has not changed significantly in the last two decades. It is interesting that both methodologies showed concor dance since other studies have not shown it¹¹.

The variable “perception” of following the GFD correctly, was the one that showed the best associa tion with negative markers, being always statistically significant. The patient who believes that he/she is not doing a correct diet has 3.4 times higher risk of not adhering than the patient that believes that he/she fo llows GFD correctly The second t variable best asso ciated with adherence was age of diagnosis. Patients diagnosed before two years of age showed significantly higher adherence to GFD than those diagnosed at later ages. One can think that the earlier CD is diagnosed, the easier it is to incorporate and accept the limita tions imposed by GFD⁸. The “attitude in birthdays” was significant when making a univariate analysis, but it lost significance in the logistic regression of associa tion, except in the group that “does not attend birth days”, which showed a greater risk of “transgressing the GFD”. This implies that not taking celiac children to birthday parties does not protect them from trans gressions to diet. That 90% of patients who did not attend a birthday presented positive serology (EMA or TTG) is against the expected; it could be interpre ted as a behavioral option to avoid the opportunity to transgress²⁶.

The guardian and/or the patient’s “perception” of “being doing GFD properly” shows that two out of three patients who believe that they are following the diet, are really achieving it. This contrasts with the variables referring to knowledge itself about diet and gluten-free foods, which did not achieve statistical sig nificance in any of the analyses. On the other hand, in patients whose guardian thinks that they make mis takes in GFD, 75.9% have some positive serology. It is interesting that 24.1% of patients (seven children), despite thinking that they are doing poorly, either be cause they eat gluten in an obvious way or because they do not have clear their indications of feeding, main tained negative both antibodies. This supports the idea that serology is not a good indicator for follow- up^11,27-29; it also opens the possibility that some patients have certain tolerance to small amounts of gluten without arising a detectable autoimmune response³⁰. Another possibility is that the time elapsed between the transgression of the diet and the measurement of mar kers was insufficient to produce a detectable change. Undoubtedly, it is necessary to improve the capacities to measure gluten intake in celiac patients along time. Measurements of gluten in stools and/or urine seem an interesting option, but the currently available results are not sufficient to propose them as marker of follow up in t CD^31-35.

In this respect, it is relevant to remind that if the re are doubts about whether a patient has elevated antibodies because he/she does not adhere to GFD or because he/she does not respond to treatment, a new duodenal biopsy is useful, as well as a genetic study if it shows absence of risk haplotypes. Unfortunately, genetic studies are not usually practiced in our country. It is interesting that the patient or guardian that reads the nutritional labeling of packaged foods showed signifi cant association with negative EMA. This relation with adherence could reflect a behavior of greater com mitment to leading a good management of the disease.

Results show that for health professionals to pro vide effective counseling to celiac patients, in addition to know the biological aspects of the disease, it is ne cessary they manage the emotional and social aspects imposed by chronic diseases. It must be considered that treatments that involve restrictions in daily life are difficult to maintain in the long-term and can affect quality of life^26,36,37. This is important to highlight sin ce we must recognize the enormous benefit that most patients who follow GFD have; they manage to stay asymptomatic and avoid the risks of long-term com plications. In this sense, the chronic or sporadic consumption of medications and the intake of foods re lated to religious ceremonies gain interest. In the first case, it is important to choose gluten-free medicines and in case of not having safe alternatives, put in the balance the risk/benefit of each treatment. In the case of foods consumed in religious ceremonies, alternati ves suitable for celiacs must be offered to maintain ad herence to GFD, such as communion wafer for celiacs, as is the case in our country, where wafers are certified by the Archbishopric of Santiago. These aspects favor the integral recovery of the patient (physical, emotio nal and spiritual).

Due to the importance of patient education in the treatment of CD^38,39, the multivariable logistic re gression analysis was adjusted by the knowledge sco re, although this was not significant in the univariate analysis. The fact that the “perception” of being doing correctly the GFD maintain its significance, seems to us a very relevant finding since it can be used to modify/reinforce behaviors and improve the follow-up and adherence to the treatment of celiac patients⁴.

Ethical Responsibilities

Human Beings and animals protection: Disclosure the authors state that the procedures were followed according to the Declaration of Helsinki and the World Medical Association regarding human experimenta tion developed for the medical community.

Data confidentiality: The authors state that they have followed the protocols of their Center and Local regulations on the publication of patient data.

Rights to privacy and informed consent: The authors have obtained the informed consent of the patients and/or subjects referred to in the article. This docu ment is in the possession of the correspondence author.

Financial Disclosure: INTA, Universidad de Chile. This study had no exter nal funding.

Conflicts of Interest: Authors declare no conflict of interest regarding the present study.

Acknwoledgements: Authors thank all professionals of the Gastroenterolo gy Unit, Hospital Roberto del Río, Santiago Chile that helped us conducting this study.