Background Thyroid autoimmunity (TA) is often associated with coeliac disease (CD).
Objective To evaluate, in children and adolescents with CD on a gluten-free diet (GFD): (1) the prevalence of TA; (2) the impact of TA on growth and the need for L-thyroxine (L-T4) treatment, during a longitudinal survey.
Method Between January and December 2005, 545 patients with CD, prospectively followed up until December 2007, and 622 controls were screened for TA. Antithyroperoxidase and antithyroglobulin antibodies were assayed and, if positive, serum free tri-iodothyronine, free thyroxine and thyroid-stimulating hormone (TSH) assays and thyroid ultrasound were performed. L-T4 was started if TSH was >5.5 mU/ml at two successive measurements.
Results There was no significant difference in TA prevalence between patients with CD on a GFD (10%) and controls (8.2%). Duration of GFD differed significantly in coeliac patients with TA in comparison with those without TA (7.9±0.9 and 10.2±0.3 years, p<0.001), but no significant difference was found for weight and height gain (1.8±1.0 vs 3.7±1.5 and 2.1±1.2 kg/year vs 4.0±1.1 cm/year, respectively). At the end of the follow-up an increase of 7% in the prevalence of patients with CD with TA requiring L-T4 was found.
Conclusions TA seems no more common in paediatric and adolescent patients with CD on a GFD than in controls; its clinical evolution does not seem to impact on growth. Therefore, a long-term regular screening programme for thyroid disease may not be necessary for all patients with CD on a GFD, but only for those who are suspected of having thyroid diseases.
Statistics from Altmetric.com
Autoimmune diseases are associated with coeliac disease (CD): their reported prevalence is up to 10-fold that in the general population.1,–,4 Among patients with CD or dermatitis herpetiformis, 4–10% have autoimmune thyroid diseases.5,–,9 It is not known if the gluten-free diet (GFD) can reduce the likelihood of developing autoimmune disorders. A recent paper on a related subject10 reports increased prevalence of thyroid-autoantibody positivity in patients with active CD and positive antitransglutaminase antibodies (tTG), noting a positive correlation between levels of tTG and of antithyroperoxidase antibodies (TPOAb).
The aim of our study was to evaluate, in children and adolescents with CD who adhered to a GFD (1) the prevalence of thyroid autoimmunity (TA) in CD compared with controls; (2) the impact of TA on growth and the need for L-thyroxine (L-T4) treatment in patients with CD, during a longitudinal survey.
What is already known on this topic
▶ The association between coeliac disease (CD) and thyroid autoimmunity (TA) may justify routine screening for thyroid dysfunction in this population, so as to facilitate early diagnosis.
▶ It is not clear if a gluten-free diet (GFD) can reduce the likelihood of developing autoimmune disorders in these patients.
What this study adds
▶ The course of TA in patients with CD is slow and seems similar to that seen in children with TA who did not have CD.
▶ Our study found no differences in growth between patients with and those without TA.
▶ TA seems no more common in paediatric and adolescent patients with CD receiving a GFD than in controls.
Materials and methods
The study was conducted at the gastroenterology unit of Bambino Gesù Children's Hospital in Rome. It was approved by the ethics committee of our hospital and informed consent was requested and obtained from parents/guardians and/or patients and control subjects.
Patients and control subjects
Between January and December 2005, 692 consecutive patients with biopsy-proven CD on a GFD were enrolled in the study. The diagnosis of CD was according to the accepted criteria.7 11,–,14 Only patients with CD with negative tTG and on a GFD for at least 12 months were included.
In December 2005, 558/692 (81%) patients were considered eligible (165 boys and 393 girls), included in the study and prospectively followed up through December 2007. Of these 558, 545 (161 boys, 384 girls) completed the 2 years of the regular screening programme.
At study entry the median age of patients with CD was 10.9 years (range 1.9–24) and median disease duration was 2.8 years (range 1–14.8).
Between January and December 2005, 705 consecutive patients, referred for minor surgical procedures to the department of surgery, were enrolled as controls. Subjects with positive tTG, family or clinical history positive for CD and/or autoimmune disorders were excluded.
In December 2005, 622/705 (88%) controls (330 girls and 292 boys), with a median age of 10.1 years (1.75–22.5), were considered eligible and included in the study.
Study design and end points
All patients with CD underwent annual tests of IgA-class tTG, to evaluate adherence to the GFD, and annual thyroid status assessments. In controls, serum IgA and tTG and thyroid status tests were performed at study entry. Patients with CD and controls with IgA deficiency were screened for CD by IgG-class tTG tests, as recommended.12
In all subjects TPOAb and antithyroglobulin antibodies (TgAb) were assayed: in patients with positivity for one or two antibodies, serum free tri-iodothyronine (fT3), free thyroxine (fT4) and thyroid-stimulating hormone (TSH) assays and thyroid ultrasound scans were performed.
Primary end points were to assess:
The prevalence of TA in patients with CD on a GFD and in controls.
The impact of TA on growth (weight and height) in all prepubertal patients with CD on a GFD.
The prevalence, in patients with CD with TA, of hypothyroidism requiring hormone replacement.
The secondary end point was to determine the differences in age and symptoms at CD diagnosis, in family history of autoimmune diseases and in human leucocyte antigen (HLA) assessment (DQ2 and DQ8) between patients with CD with TA and those patients without TA.
CD serology assays
The tTG were measured with an ELISA kit (Eu-tTG; Eurospital, Trieste, Italy); (normal value (NV)<4 IU/ml; >7 IU/ml was considered positive, 4–7 IU/ml was considered borderline). Patients' total serum IgA was determined using a Boeringer Mannheim Hitachy 912 Analyzer, with Tina-quant reagents (Roche Diagnostics GmbH, Mannheim, Germany). To determine HLA DQ2–DQ8, type Eu-DQ (Eurospital) was used as indicated by the manufacturer.
Thyroid function assessment
A chemiluminescent immunometric assay (Advia Centaur; Siemens, Munich, Germany) was employed for testing all thyroid functions: TgAb and TPOAb: NV<60 IU/ml; TSH: NV 0.4–5.5 µIU/ml; fT3: NV 2.2–5.1 pg/ml; fT4: NV 0.70–1.80 ng/dl. Thyroid function values were considered abnormal when they were above or below the range and at least twice the coefficient of variation.
Patients were classified as: (1) euthyroid, if serum TSH, fT3 and fT4 were within the normal range; (2) subclinically hypothyroid, if TSH was from 4 to 10 μIU/ml and fT3 and fT4 were normal; (3) hypothyroid, if TSH was >10 μIU/ml and fT4 was low.
L-T4 treatment was started for hypothyroid patients and for those with subclinical hypothyroidism if TSH was >5.5 mU/ml in at least two successive measurements.
Ultrasonography of the thyroid gland was performed with a commercially available real-time scanner (Siemens Acuson S-2000, Siemens Medical Solutions USA, Inc., Mountain View, California, USA) equipped with a 7.5–13 MHz linear probe, by two sonographers with experience in thyroid imaging. The sonographic images were classified according to their patterns, in accordance with Sostre and Reyes15 and Marcocci et al.16
Statistical data were generated with the SPSS version 13.1 for Windows (SPPS, Chicago, Illinois, USA). To ascertain whether possible differences in TA between patients and controls could be clinically meaningful and hence worth detecting, we had previously estimated the required sample size by calculating the α and β probabilities of type I and type II errors, respectively. By assuming that the expected prevalence of TA was 8% in the general population (and therefore among our controls), and 14% among patients with CD, and accepting a power of 0.80 and a 0.05 two-sided type I error, we calculated that 457 patients and 457 controls should be enrolled.
Continuous variables were summarised as median and range or as mean and SD. Categorical variables were summarised as the percentage of the entire group. Between patients with CD and control subjects and within the CD group, the χ2 test (or Fisher's exact test, where indicated) was used to compare differences in categorical variables, while the t test (paired or unpaired, as appropriate) was used for continuous variables. For all statistical analyses a two-tailed p value of 0.05 was considered significant.
At inclusion, the percentage of subjects with TA was not significantly higher in patients with CD (56 patients (10%)) than in controls (51 patients (8.2%)). The markers of thyroid function as well as the results of thyroid ultrasound were similar in patients with CD and controls (table 1).
After 2 years of follow-up, the patients with CD with TA numbered 65 (12%): 43 (8%) had been positive since 2005, 10 (2%) since 2006 and 12 (2%) since 2007. Thus, at the end of the study, the TA prevalence in patients with CD (12%) was not significantly increased in comparison with baseline (10%). Ten patients out of the 56 who had been positive for thyroid autoantibodies at inclusion, and 12 of the 64 who had been positive during the second year of the study, became negative during follow-up; among these, the percentage of patients with CD on a GFD for <6 years (77%) was significantly higher than that of patients with CD on a GFD for >6 years (23%) (p<0.05).
Growth in patients with CD with TA did not differ significantly from that in patients without TA (table 2).
Six out of the 65 (9.2%) patients with CD plus TA developed subclinical hypothyroidism requiring L-T4 treatment, as has been recommended.17 The proportion of L-T4-treated patients at the end of follow-up in comparison with baseline was increased, although not significantly (9% vs 2%, respectively).
Thus the yearly incidence rate of hypothyroidism requiring treatment was 3.5% among patients with CD with TA and 0.4% among all patients with CD.
No differences were noted between patients with CD with and those without TA in age or symptoms at CD diagnosis, in family history of autoimmune diseases or in HLA assessment (see table 2 and figure 1).
The main findings of our study were (1) a similar prevalence of TA in patients with CD (10%) and in control subjects (8.2%); (2) a high rate of subsidence of TA, especially in patients with CD on a GFD for <6 years; (3) no differences in growth rate between patients with CD with and without TA; (4) a low, but not negligible, incidence rate, during follow-up, of hypothyroidism requiring hormone replacement (3.5% in patients with CD with TA). Furthermore, there were no differences in age, presenting symptoms, family history of TA or HLA assessment between patients with CD who had TA and those who did not.
Previous studies have reported wide ranges of TA in paediatric and adult patients with CD—that is, from 1.2% to 41%5 18,–,23 in children and from 5.4% to 16.2% in adults.7 8 24 In our study the prevalence of TA was 10% at baseline, considerably lower than that found in an earlier Italian paediatric multicentre study17 (26%); conversely, TA prevalence in control subjects (10%) in that study was similar to ours (8.2%). Meloni et al23 report a prevalence of TA in patients with CD similar to that found in our survey (10.5%), although the prevalence of TA in their controls was significantly higher than in ours. Guariso et al4 found a prevalence of 10% of TA in patients with CD but of only 1% in controls. Unlike all previous studies, we included only patients with CD on a strict GFD for at least the preceding year. In the light of the evidence7 18 22 of a remission of TA following the start of GFD, our inclusion criteria may have contributed in decreasing the actual risk of developing TA in our patients with CD. This observation could explain the discrepancy in TA prevalence between our survey and two earlier Italian studies,17 23 which included patients on a GFD as well as patients with a baseline diagnosis of CD who had not yet started GFD.
A noteworthy finding is the subsidence of thyroid autoantibodies in about a quarter of all patients with TA during the follow-up, especially during the first 6 years of GFD. Two previous longitudinal surveys22 25 of paediatric patients with CD have also demonstrated that TA can regress with a gluten-free regimen. These two studies report a prevalence of autoantibody cessation of 33% and 85% in a series of 6625 and 90 patients with CD,22 respectively, during the first 2 years after starting GFD.
Our data show that the course of TA in CD subjects is quite slow: hypothyroidism requiring hormone replacement was found in 2% of patients with TA at study entry and in a further 7% after 2 years of follow-up. Thus 91% of patients with TA were still euthyroid for at least 2 years, and in some cases 3, after the original diagnosis. The course of TA in our study seems similar to that seen in children with TA who did not have CD.22 In that survey, Radetti et al22 found that more than half of the children with TA were still euthyroid after 4 years of follow-up.
Growth evaluation is a great concern in the treatment of paediatric patients with chronic diseases. Our study did not find growth retardation in patients with CD with TA, according to that reported in children with acquired hypothyroidism and TA, not affected by CD.26 27 Thus we suggest that if hypothyroid children have normal growth, regular screening programmes in asymptomatic children affected or not by CD may not have clinical advantages. This suggestion, nevertheless, will need of large prospective study to be confirmed.
Additionally, the usefulness of ultrasound findings for either treatment or outcome was negligible in our sample, as found by other paediatric studies.26
TA seems no more common in paediatric and adolescent coeliac patients on a GFD than in controls; its clinical evolution is slow and it seems to have no impact on growth. Thyroid autoantibodies tend to disappear in a quarter of all coeliac patients on a GFD. A long-term regular screening programme for thyroid disease may not be necessary for all coeliac patients on a GFD, but only for those who, for clinical reasons, are suspected of having thyroid disease.
Competing interests None.
Ethics approval Our study was approved by the ethics committee of our hospital and informed consent was requested and obtained from parents/guardians and/or patients and control subjects.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.