Key words
celiac disease - hypothalamic-pituitary-thyroid axis - thyroid function tests - thyroiditis
List of abbreviations
CI: confidence interval
IU: international unit
SD: standard deviation
SPINA: structure parameter inference approach
TgAb: thyroglobulin antibodies
TPOAb: thyroid peroxidase antibodies
Introduction
Hashimoto’s thyroiditis, an autoimmune disorder destroying thyroid follicular cells
by cell and antibody-mediated immune processes, is the most common thyroid disease
in iodine-sufficient areas and one of the most common human disorders [1]
[2]. The disease is caused by replacement of follicular cells by lymphocytic infiltrate
and fibrosis, and is characterized by the presence of thyroid antibodies, particularly
of thyroid peroxidase antibodies (TPOAb) and thyroglobulin antibodies (TgAb) [3]
[4].
The results of many studies evidenced an association between Hashimoto’s thyroiditis
and coeliac disease or an asymptomatic increase in anti-tissue transglutaminase antibodies
[5]
[6]
[7]
[8]. Autoimmune thyroiditis is the most prevalent coexisting autoimmune disorder in
patients with celiac disease [9]. A pooled analysis, including 6024 patients with autoimmune thyroiditis, found a
markedly increased prevalence of biopsy-confirmed coeliac disease, allowing the authors
to conclude that all patients with autoimmune thyroiditis should be screened for the
presence of coeliac disease [10]. The association between autoimmune thyroid disease and coeliac disease may be explained
by low selenium [11] or vitamin D [12] status secondary to their malabsorption, the interaction of tissue transglutaminase-2
IgA antibodies to thyroid follicles and thyroid extracellular matrix [13], or by a shared immunogenetic make-up [14]. Both exogenous vitamin D (cholecalciferol) [15] and selenium [16] seem to play a role in the development and progression of Hashimoto’s thyroiditis.
To best of our knowledge, no previous study investigated whether restriction of gluten
intake affect thyroid autoimmunity and function. Therefore, the purpose of our study
was to investigate whether a gluten-free diet has an impact on thyroid autoimmunity,
hypothalamic-pituitary-thyroid axis activity and thyroid function tests in drug-naïve
women with Hashimoto’s thyroiditis.
Materials and Methods
Patients
The participants of the study were selected among young women, aged between 20 and
45 years, with recently diagnosed and previously untreated autoimmune thyroiditis.
To be admitted to the study, they were required to have (a) positive TPOAb (>100 U/mL), (b) the reduced echogenicity of the thyroid parenchyma on thyroid ultrasonography;
(c) normal thyroid function (thyrotropin levels in the range between 0.4 and 4.5 mU/L,
free thyroxine in the range between 10.0 and 21.0 pmol/L and free triiodothyronine
in the range between 2.6 and 6.5 pmol/L) and (d) incidentally found positive anti-tissue
transglutaminase antibodies without clinical symptoms of coeliac disease. All patients
ingested gluten in their diet prior to starting the study.
We excluded women with symptomatic coeliac disease, positive antibodies against thyrotropin
receptor, diabetes or other endocrine disorders, impaired renal or hepatic function,
any acute and chronic inflammatory processes, any other serious disorders, pregnancy
or lactation, as well as women receiving any chronic treatment.
The research was conducted in accordance with the Helsinki Declaration. The study
was approved by the local review board and written, informed consent was obtained
from each participant.
Study design
Before the study, all participants were informed about the benefits and risks of following
a gluten-free diet. Based on patient preference, the participants were allocated to
one of two groups. Patients belonging to group A (n=16) were instructed by a physician
and a professional dietitian with the help of leaflets to follow a gluten-free diet.
Patients from group B (n=18) were not prescribed any special dietary recommendations.
Participants were seen every two months to ensure adherence to the diet and to boost
compliance with the protocol. During each visit, women were asked to complete a questionnaire
evaluating how often in the past two months they had consumed each of the twenty most
commonly used meals of Polish cuisine. Food intake frequency was assessed in terms
of the following six categories: everyday, 5-6 times per week, 3-4 times per week,
1-2 times per week, less than once per week and never. The questionnaire included
also questions on the use of gluten-free products. Moreover, each patient was required
to provide the packaging and labels of the gluten-free products he or she consumed.
Laboratory assays
Laboratory assays were performed in duplicate (to minimize analytical errors) at baseline
and 6 month later (at the end of the treatment period). Venous blood samples were
collected from the antecubital vein between 8 and 9 a.m. (to avoid possible circadian
fluctuations in the parameters studied) after an overnight 12-h fasting. Serum levels
of thyrotropin, free thyroxine and free triiodothyronine and titers of TPOAb and TgAb
were measured by direct chemiluminescence using acridinium ester technology (ADVIA
Centaur XP Immunoassay System, Siemens Healthcare Diagnostics, Munich, Germany). Serum
25-hydroxyvitamin D levels were assayed by competitive immunoassay using Roche Diagnostic
commercial kits and a multichannel automatic analyzer (Roche Cobas e 411, Mannheim,
Germany). Immunoglobulins A against the tissue transglutaminase antigen were detected
using the enzyme-linked immunosorbent assay method (Euroimmun, Lübeck, Germany). Jostel’s
thyrotropin index, the structure parameter inference approach (SPINA)-GT index as well as the SPINA-GD indices were calculated by the investigators
based on thyrotropin and free thyroid hormone levels using SPINA-Thyr 4.0.1 for Windows
software in accordance with the formulas described previously [17]
[18]
[19].
Statistical analysis
Values for hormones, antibodies and thyroid function tests were natural log-transformed
to yield a normal distribution for statistical analyses. Treatment groups were compared
using the t test for independent samples. The differences between the means of variables
within the same treatment group were analyzed with Student’s paired t test. The clinical
importance of the result was assessed based on the 95% confidence interval. A t statistic
and two sample means were used to generate an interval estimate of the difference
between two population means. Categorical variables were analyzed by χ2 test, while Pearson’s r-tests were used to calculate the significance of each correlation.
Differences were described as statistically significant if 95% confidence intervals
did not include the null value and/or two-tailed p values were below 0.05.
Results
At study entry, both groups of patients were comparable with respect to age, body
mass index, smoking, the percentage of women giving birth in the past, the number
of deliveries, as well as serum levels of hormones and 25-hydroxyvitamin D, serum
titers of thyroid antibodies, Jostel’s thyrotropin index, the SPINA-GT index and the
SPINA-GD index ([Table 1], [2]). The gluten-free diet was well tolerated and no patient prematurely terminated
the study. With the exception of gluten-free and gluten-containing products, the study
groups did not differ in the frequency of particular food consumption. During the
follow-up period, no patient developed clinical symptoms of coeliac disease. At the
end of the study, positive anti-tissue transglutaminase antibodies were present in
sera of 6 patients (38%) from group A and in all 18 patients from group B.
Table 1 Baseline characteristics of patients.
|
Variable
|
Gluten-free diet (Group A)
|
Gluten-containing diet (Group B)
|
Difference [95% CI]
|
|
Number of patients [n]
|
16
|
18
|
-
|
|
Age [years; mean (SD)]
|
30 (5)
|
31 (6)
|
1 [−3, 5]
|
|
Smokers [%]
|
25
|
22
|
-
|
|
Body mass index [kg/m2; mean (SD)]
|
22.9 (2.3)
|
23.1 (2.1)
|
0.2 [−1.3, 1.7]
|
|
Past deliveries (%)/Number of deliveries [n; mean (SD)]
|
69/1.1 (0.9)
|
67/1.2 (0.9)
|
0.1 [−0.5, 0.7]
|
CI: confidence interval; IU: international unit; SD: standard deviation; SPINA: structure
parameter inference approach; TgAb: thyroglobulin antibodies; TPOAb: thyroid peroxidase
antibodies
Table 2 The effect of the gluten-free diet on thyroid antibody titers, hormones, thyroid
function tests and 25-hydroxyvitamin D levels in euthyroid women with Hashimoto’s
thyroiditis.
|
Variable
|
Gluten-free diet (Group A)
|
Gluten-containing diet (Group B)
|
Difference [95% CI]
|
|
TPOAb [U/mL; mean (SD)]
|
|
Baseline
|
925 (265)
|
891 (242)
|
−34 [−211, 143]
|
|
After 6 months
|
705 (206)#
|
920 (280)
|
215 [42,388]*
|
|
Change
|
−200 (105)
|
29 (25)
|
229 [177,281]&
|
|
TgAb [U/mL; mean (SD)]
|
|
Baseline
|
832 (311)
|
792 (274)
|
−40 [−244, 164]
|
|
After 6 months
|
629 (240)#
|
845 (324)
|
216 [15,417]*
|
|
Change
|
−203 (120)
|
53 (58)
|
256 [194,318]&
|
|
Thyrotropin [mIU/L; mean (SD)]
|
|
Baseline
|
2.7 (1.0)
|
2.9 (0.8)
|
0.2 [−0.4, 0.8]
|
|
After 6 months
|
2.4 (0.8)
|
2.6 (0.9)
|
0.2 [−0.4, 0.8]
|
|
Change
|
−0.3 (0.2)
|
−0.3 (0.2)
|
0.0 [−0.1, 0.1]
|
|
Free thyroxine [pmol/L; mean (SD)]
|
|
Baseline
|
14.9 (2.3)
|
15.3 (2.7)
|
0.4 [−1.4, 2.2]
|
|
After 6 months
|
16.1 (2.4)
|
15.0 (2.3)
|
−1.1 [−2.7, 0.5]
|
|
Change
|
1.2 (1.4)
|
0.3 (0.6)
|
−0.9 [−1.9, 0.1]
|
|
Free triiodothyronine [pmol/L; mean (SD)]
|
|
Baseline
|
3.2 (0.6)
|
3.1 (0.6)
|
−0.1 [−0.5, 0.3]
|
|
After 6 months
|
3.6 (0.7)
|
3.2 (0.7)
|
−0.4 [−0.9, 0.1]
|
|
Change
|
0.4 (0.3)
|
0.1 (0.3)
|
−0.3 [−0.1, 0.5]
|
|
Jostel’s thyrotropin index [mean (SD)]1
|
|
Baseline
|
3.0 (0.2)
|
3.1 (0.3)
|
0.1 [−0.1, 0.3]
|
|
After 6 months
|
3.0 (0.2)
|
3.0 (0.2)
|
0.0 [−0.1, 0.1]
|
|
Change
|
0.0 (0.1)
|
−0.1 (0.2)
|
−0.1 [−0.3, 0.1]
|
|
SPINA-GT index [pmol/s; mean (SD)]2
|
|
Baseline
|
2.28 (0.38)
|
2.26 (0.34)
|
−0.02 [−0.27, 0.23]
|
|
After 6 months
|
2.62 (0.40)#
|
2.30 (0.37)
|
−0.32 [−0.59, −0.05]*
|
|
Change
|
0.34 (0.12)
|
0.04 (0.06)
|
−0.3 [−0.37, −0.23]&
|
|
SPINA-GD index [nmol/s; mean (SD)]3
|
|
Baseline
|
19.86 (3.04)
|
18.73 (2.53)
|
−1.13 [−3.08, 0.82]
|
|
After 6 months
|
20.68 (2.41)
|
19.73 (2.81)
|
−0.95 [−2.79, 0.89]
|
|
Change
|
0.82 (0.25)
|
1.00 (0.31)
|
0.18 [-0.02, 0.38]
|
|
25-hydroxyvitamin D [ng/mL; mean (SD)]
|
|
Baseline
|
20 (6)
|
21 (5)
|
1 [−3, 5]
|
|
After 6 months
|
25 (6)#
|
20 (5)
|
−5 [−9, −1]*
|
|
Change
|
5 (3)
|
−1 (2)
|
−6 [−8, −4]&
|
CI: confidence interval; IU: international unit; SD: standard deviation; SPINA: structure
parameter inference approach; TgAb: thyroglobulin antibodies; TPOAb: thyroid peroxidase
antibodies; 1Reference range: 1.3-4.1 [17]; 2Reference range: 1.4-8.7 pmol/s [18];
3Reference range: 20-60 nmol/s [18]; *statistically significant difference between both groups; #statistically significant difference between post-treatment and baseline values in
the same group; &statistically significant difference between the changes in both treatment groups.
The gluten-free diet reduced serum titers of TPOAb and TgAb, as well as increased
serum levels of 25-hydroxyvitamin D and the SPINA-GT index ([Table 2]). The gluten-free diet did not affect thyrotropin, free thyroid hormones, Jostel’s
thyrotropin index and the SPINA-GD index. In group B, serum thyroid antigens, thyrotropin,
free thyroxine, free triiodothyronine and 25-hydroxyvitamin D as well as Jostel’s
thyrotropin, the SPINA-GD and the SPINA-GT indices remained at the similar levels
throughout the study. At the end of the study, there were significant differences
between both groups in thyroid antibody titers, 25-hydroxyvitamin D levels and the
SPINA-GT index ([Table 2]).
At entry, TPOAb titers correlated with TgAb titers (r=0.60, p<0.001). Antibody titers
inversely correlated with 25-hydroxyvitamin D levels (TPOAb: r=−0.37, p<0.01; TgAb:
r=−0.25, p<0.05). Titers of TPOAb and TgAb correlated also with the SPINA-GT index
(TPOAb: r=−0.35, p<0.01; TgAb: r=−0.29, p<0.05). In group A, the impact of treatment
on TPOAb titers correlated with the changes in TgAb titers (r=0.52, p<0.001) and with
the changes in 25-hydroxyvitamin D levels (r=0.34, p<0.05). The gluten-free diet-induced
changes in antibody titers correlated also with the impact on the SPINA-GT index (TPOAb:
r=0.40, p<0.001; TgAb: r=0.32, p<0.05). No other correlations were found.
Discussion
The major finding of our study is that the gluten-free diet reduced thyroid autoimmunity
and slightly increased thyroid output in euthyroid women with Hashimoto’s thyroiditis.
This action seems to be a specific effect of the gluten-free diet because in a significant
proportion of patients (62%) the decrease in thyroid antibody titers and changes in
thyroid function tests were accompanied by disappearance of anti-tissue transglutaminase
antibodies. The obtained results cannot be explained by seasonal fluctuations or time-dependent
changes in antibody titers and thyroid function tests, because serum levels of thyrotropin
and free thyroid hormones, serum titers of TPOAb and TgAb, as well as Jostel’s thyrotropin,
the SPINA-GT and SPINA-GD indices in individuals not limiting gluten intake remained
at the similar levels during the whole study period. Considering very strict inclusion
and exclusion criteria, the consequence of which was obtaining a homogenous group
of patients with autoimmune thyroiditis, the results of our study cannot be attributed
to coexisting disorders or to the action of any drugs taken by patients. Finally,
the obtained results do not seem to be a consequence of the improvement in the diet
composition because analysis of individual food-frequency questionnaires revealed
no significant differences in particular food consumption, including fish or milk
intake.
It is difficult to explain mechanisms responsible for the favorable effect of the
gluten-free diet on thyroid autoimmunity. Our research indicates that one of them
is an improvement in vitamin D status. In line with this hypothesis, our population
of patients was characterized by relatively low serum levels of 25-hydroxyvitamin
D, being is the best marker determining vitamin D status [20]. Moreover, the gluten-free diet, but not a normal diet, increased circulating levels
of 25-hydroxyvitamin D, and this action correlated with the changes in TPOAb titers.
Interestingly, exogenous vitamin D preparations reduced thyroid antibody titers in
women with thyroid autoimmunity, and this effect was stronger for TPOAb than for TgAb
[21], which probably reflects the fact that TPOAb are characterized by higher sensitivity
and at least equal specificity to TgAb in the diagnosis of autoimmune thyroid disease
[22]. Taking into account that correlations between TPOAb titers and 25-hydroxyvitamin
D levels were moderate, it seems that also other mechanisms contribute to the favorable
effect of the gluten-free diet. One of them may be an improvement in selenium status,
because preparations of this micronutrient were found to reduce thyroid autoimmunity,
at least in part by inhibiting a secretory function of human T cells [16]. Apparently healthy women inhabiting the Upper Silesia, where the study was conducted,
are characterized by low selenium status [23], which might have been even worse in the participants of our study. Because no small
intestinal biopsy was performed, it is possible that a significant proportion of our
patients might have had subclinical (asymptomatic) coeliac disease. Coeliac disease,
even mild, is characterized by selenium depletion [24], more pronounced in subjects poorly complying with the gluten-free diet than in
well-complying patients [25]. We did not measure serum or whole-blood selenium levels because in the adult population
they do not correlate well with dietary consumption of this micronutrient, reflecting
better intake of selenomethionine and selenium-enriched yeast supplements than intake
of the remaining forms of selenium [26]. However, relatively low baseline values of the SPINA-GD index in both groups of
patients indirectly support selenium deficiency in the study population. The gluten-free
diet may also directly inhibit activity of inflammatory cells. In line with this explanation,
dietary gluten was found to alter the balance of pro-inflammatory and anti-inflammatory
cytokines in T cells of mice towards a more inflammatory cytokine profile [27], while the gluten-free diet reduced circulating levels of proinflammatory cytokines
[28].
The gluten-free diet induced an increase in the SPINA-GT index, being a marker of
the maximum secretion rate under stimulated conditions [18]
[19]. Its retest reliability is higher than that of thyrotropin, free thyroxine and free
triiodothyronine [18], and this fact may explain why, despite affecting the SPINA-GT index, the dietary
treatment had a neutral effect on serum hormone levels. The impact on thyroid output
was relatively small, because during the whole period of the follow-up values of the
SPINA-GT index remained within the reference range. Interestingly, treatment-induced
changes in the SPINA-GT correlated with both baseline serum titers of TPOAb and TgAb
and with the reduction in thyroid antibody titers. This finding allows us to draw
two conclusions. Firstly, the inhibitory effect of the gluten-free diet on thyroid
autoimmunity may participate in the improvement of thyroid secretory capacity. Secondly,
the effect on antibody titers was strongest in patients with the highest antibody
titers. Considering a predictive value of thyroid antibodies and their titers as risk
factors of the development of hypothyroidism [29]
[30], the gluten-free diet may delay the development of thyroid hypofunction in euthyroid
women with Hashimoto’s thyroiditis. Despite increasing the SPINA-GT index, the gluten-free
diet exerted no effect of the other calculated parameters of thyroid homeostasis.
A neutral effect on the SPINA-GD index, estimating the sum of activity of peripheral
deiodinases [19], means that the gluten-free diet does not affect peripheral metabolism of thyroid
hormones. Finally, no change in Jostel’s thyrotropin index, a quantitative marker
for pituitary thyrotropic function [17], indicates that the gluten-free diet is devoid of a direct effect on thyrotropic
cells.
Several shortcomings of our study bear mentioning. The most important limitation of
the study is its non-randomized nature, a small number of participants and a short
period of therapy. Moreover, the study population was characterized by low selenium
status [23] and adequate iodine intake [31]. It is not certain that the impact of the gluten-free diet is similar in selenium-sufficient
and/or iodine-depleted areas. Furthermore, because the study included only euthyroid
women, the question whether the gluten-free diet affects antibody titers and thyroid
function tests in subjects with thyroid hypofunction remains unanswered. Finally,
it cannot be ruled out that the effect of the gluten-free diet may be different in
women treated with agents reducing thyroid antibody titers (levothyroxine, vitamin
D and selenium), who were not included in our study.
Summing up, the gluten-free diet reduced serum titers of TPOAb and TGAb in euthyroid
women with Hashimoto’s thyroiditis, which correlated with the increase in the SPINA-GT
index. This finding indicates that the gluten free-diet may bring clinical benefits
to euthyroid women with Hashimoto’s thyroiditis, who, because of markedly elevated
thyroid antibody titers, are at high risk of the development of hypothyroidism. Because
of study limitations, our study should be regarded as a pilot one, and larger prospective
trials are required to support our findings.
Institutional approval: The study was approved by the Bioethical Committee of the Medical University of
Silesia.
Funding
The study was supported by the statutory grant of the Medical University of Silesia
(KNW-1-062/N/7/0). The experiments comply with the current law of Poland.
