Screening for celiac disease, by endomysial antibodies, in patients with unexplained articular manifestations
- First Online:
- Cite this article as:
- Ghozzi, M., Sakly, W., Mankaï, A. et al. Rheumatol Int (2014) 34: 637. doi:10.1007/s00296-013-2906-x
- 269 Views
Celiac disease (CD) is an autoimmune systemic disease characterized by not only gastrointestinal but also extraintestinal manifestations. The aim of our study was to do a serological screening for CD, by IgA endomysial antibodies (EmA), in patients with unexplained articular manifestations. Two hundred and eleven patients suffering from arthritis or arthralgia without evident cause were studied. EmA were determined by indirect immunofluorescence on human umbilical cord. Two thousand and five hundred blood donors served as control group. Out of 211 patients, 5 had EmA (2.37 %). The frequency of EmA in our patients was significantly higher than in the control group (2.37 vs. 0.28 %, p < 0.01). All patients with positive EmA were female. EmA were significantly more frequent in female patients than in female healthy subjects (3 vs. 0.4 %, p < 0.01). Medical records revealed: diarrhea (one patient), short size (one patient), anemia (three patients), weight loss (two patients) spontaneous abortion (three patients), secondary amenorrhea (one patient), early menopause (one patient) and early baby death (one patient). Biochemical analysis showed decreased level of calcium (one patient), vitamin D (one patient) and cholesterol (one patient). Unexplained liver cytolysis was observed in two patients. Radiological examination showed demineralization of two hands in one patient. Bone osteodensitometry done in one patient out of five revealed lumbar osteopenia. The articular manifestations of the five patients did not respond to corticosteroid treatment. CD must be considered among the differential diagnosis in a patient with arthritis or arthralgia.
KeywordsArthritisArthralgiaEndomysial antibodiesCeliac diseaseAdults
Celiac disease (CD) is an autoimmune systemic disease elicited by gluten in genetically susceptible individuals. CD is strongly associated with specific HLA class II genes known as HLA-DQ2 and HLA-DQ8. The susceptibility loci are shared with those for other autoimmune diseases . IgA antiendomysial antibody (EmA) is considered the “gold standard” in the diagnostic of CD, with both a high sensitivity and specificity [2, 3]. The EmA target transglutaminase type 2 (TG2) which is an ubiquitous enzyme. The TG2 manifests in all tissues with both intra- and extra-cellular localization; hence, CD can affect any organ including joints . In fact, the association between CD and articular manifestation was described [4–8]. Articular manifestations are more frequent in celiac patients than the general population . CD, in some reported cases, was revealed by arthritis which recovered on a gluten-free diet (GFD) [7, 10, 11]. Since the frequency of CD in patients with unexplained arthralgia or arthritis has not been reported previously, we decided to do a serological screening for CD, by EmA, in a series of patients with articular manifestations without evident cause.
Patients and methods
In our retrospective and multicenter study, we included two hundred and eleven consecutive sera collected, via the database of our immunology laboratory, from adult patients (45 males, 166 females, median age 39 years 2 months, age range 19–89 years) with arthritis or arthralgia associated or not with myalgia without evident cause. No specific arthritis neither rheumatoid arthritis (RA) nor systemic lupus erythematosus (SLE) was identified, and there was no evidence of ankylosing spondylitis or psoriatic arthritis in all patients included in this study according to classification criteria of each disease [12–15]. Patients were admitted for articular manifestations in the center of Tunisia during the period from January to December 2012.
Serological screening for CD has been done by EmA, and patients with positive EmA were invited for small bowel biopsy (SBB). All sera were stored at −80 °C until the use. Ethical committee of our hospital gave approval for this study.
Two thousand and five hundred blood donors from the center of Tunisia, who were screened for CD in our previous study, served as a control group .
EmA were detected by indirect immunofluorescence (IIF) using human umbilical cord cryostat sections (4 μm thick) as we have described previously . Briefly, the cryosections were incubated with serum diluted at 1/10. The slides were washed and rinsed in phosphate-buffered saline. Investigation of IgA class antibodies was done using fluorescein labeled antihuman IgA antibodies (Bio-Rad®, Marnes-La Coquette, France). A positive result of EmA was recorded if the connective tissue surrounding the muscle cells fluoresced brightly in a honeycomb pattern.
The comparison of EmA frequency between patients and controls was performed using Chi-square or Fisher’s exact test; 95 % confidence interval (CI) was calculated. All reported p values were two-tailed, and the level of significance was set at p < 0.05.
Frequency of EmA in patients and in control group
(n = 211)
(n = 2500)
p < 0.01
Frequency of EmA according to sex
(n = 211)
(n = 2500)
(n = 166)
(n = 45)
(n = 750)
(n = 1750)
Articular manifestations of five patients with EmA
Causes of admission to the hospital
Polyarthralgia (Large and small articulations)
Polyarthralgia (large and small articulations)
Limitation of her walking ability
Polyarthralgia (large and small articulations)
Arthralgia (large and small articulations)
Polyarthralgia (large and small articulations)
Characteristics of five patients with EmA
Medical records obtained after serological screening for CD
Radiological examination: demineralization of two hands
Secondary amenorrhea since 3 months
Blood calcium decreased
Unexplained liver cytolysis
Bone osteodensitometry: lumber osteopenia
Liver cytolysis without known etiology
Short size: 1.42 m
One abortion and 8 children died before the age of 6 months
Vitamin D deficiency
In the present study, we have done a serological screening for CD in 211 consecutive patients with unexplained articular manifestations. Since 1980, Prier et al.  reported two cases of CD revealed by arthritis. Then, other case reports were described. The cause–effect relation was proved by an improvement or complete remission of the symptoms after starting a GFD. In patients under gluten challenge, by interrupting their GFD, an exacerbation of joint pain and swelling was observed confirming that this manifestation was due to CD [4–6, 10, 18]. In the other hand, the frequency of CD in arthritis has been done in a large serious of patients with known rheumatic diseases . Another study has done the vice versa that means the frequency of arthritis in CD patients . Polyarthralgia is a common symptom of CD that often antedates the intestinal manifestation. It has been demonstrated that arthritis was non-erosive, predominantly involving the large joints and occurring as mono-oligo or polyarthritis [4, 6]. The frequency of CD in patients with unexplained arthralgia or arthritis has not been done previously. Given that most cases of CD are undiagnosed , we decided to study this at high-risk population.
We found that five patients out of 211 (2.37 %) had EmA. These five patients were informed that they have to undergo SBB. However, it has been recently demonstrated that SBB is not always required to diagnose CD . In fact, the specificity of EmA is 100 % as we and others demonstrated previously [17, 22]. Furthermore, the new recommendations for CD suggested that EmA are the gold standard for the diagnosis . According to all these data, the frequency of CD in our patients is 2.37 %. The frequency of CD was significantly higher in patients than in healthy subjects studied in our previous publication (2.37 vs. 0.28 %, p < 0.01). Since the distribution of males and females was substantially different in our cases and control cohort (frequency of female in case cohort was 78.6 %, whereas only 30 % of controls were female), we compared the frequency of CD in males and females between patients and control group. We found that the frequency of CD was significantly higher in female patients than in female controls (3 vs. 0.4 %, p < 0.01).
In the present study, we did not measure total IgA levels in sera of our patients with unexplained articular manifestations. Although CD occurs with increased frequency in those with selective IgA deficiency, screening studies of the general population suggest that very few cases will be missed by not routinely measuring IgA levels as part of the screening regimen. Thus, the strategy of routinely determining serum IgA levels or adding IgG-based serology as part of a panel to screen asymptomatic individuals in the general population is not warranted [23, 24].
We tried to know whether really the five patients, in whom we have discovered CD fortuitously thanks to this serological screening, had an asymptomatic CD. So, we had access to medical records and to our surprise three patients out of five had manifestations of classical CD, but they remained undiagnosed until we decided to do the present study. It is known that the majority of CD patients are not diagnosed . The five patients were females, and their mean age was 54 years 6 months. The peak incidence of adult CD is in the fourth and fifth decades . In our study, two patients out of five with CD were elderly. In fact, increasing prevalence and high incidence of celiac disease in elderly were reported . The first case, a 70-year-old woman, was admitted to the hospital for arthralgia, arthritis and morning stiffness. She had unexplained liver cytolysis. One of the most commonly presented liver problems among celiac patients is isolated hypertransaminasemia with non-specific histological changes . The other elderly patient was an 80-year-old woman admitted because of arthralgia. After CD screening, medical records revealed that she had classic symptoms of CD such as diarrhea, anemia, short stature, vitamin D deficiency and unexplained weight loss. Also, this patient had one abortion among 18 pregnancies, and eight of her children were dead before the age of 6 months. Hepburn et al.  reported a case of an 80-year-old woman who was referred to the hematology department for anemia, tests showed a deficiency in iron, folic acid and vitamin B12. She developed, later, a diarrhea, unexplained weight loss and she complained from painful swollen ankles and chest pain. CD was suspected and confirmed by a SBB. Two months after starting GFD, the ankle pain and swelling disappeared.
In the present study, one patient has demineralization, one had lumber osteopenia and decreased blood calcium and another had vitamin D deficiency. These abnormalities are due to malabsorption syndrome, and CD was not suspected in these three patients who have also other signs of CD. It is not surprising to diagnose CD when patients have severe symptoms due to malabsorption syndrome. In fact, we have previously described severe osteomalacia due to undiagnosed CD in three women .
In our study, four patients, out of five with EmA, had gynecological or obstetrical disorders which were spontaneous abortion, early menopause, secondary amenorrhea since 3 months and early baby death. Reproductive changes have been described, including impaired fertility and adverse pregnancy outcomes possibly related to immune-mediated mechanisms or nutrient deficiency. Other possible pathogenetic factors that may alter placental function include maternal CD autoantibodies binding to placental transglutaminase and genetic mutations that may facilitate microthrombus formation . We have previously published a case of CD in a 15-year-old girl who had secondary amenorrhea which recovered on GFD .
Oral aphthosis was observed in one patient. In fact, oral manifestations such as dental enamel hypoplasia, aphthous ulcers and delayed eruption of teeth have been found significantly associated with CD. It should be noted that aphthous ulcers or oral ulcers, a non-specific characteristic for CD, regressed on GFD .
Four patients, out of five with EmA, had no gastrointestinal symptoms. It may be of importance to the clinicians to be aware of predominant extraintestinal symptoms that can exist for several years in CD patients without any significant gastrointestinal symptoms. A diagnosis of non-classical CD might spare the patient for useless investigations and insufficient treatment and make the way for appropriate treatment with GFD. Our five patients received medical treatment for their articular and/or muscular manifestations but without successful remission. In the same way, we have previously reported a case of CD, in a 9-year-old girl, revealed by uveitis and joint pain. The patient was treated by topical and systemic corticosteroids without any improvement. Within 2 months after starting the GFD, both uveitis and joint pain disappeared . Lubrano et al.  demonstrated that arthritis was significantly more frequently noted in patient who had not yet started a GFD than in those who were already on GFD.
A limitation of our study is that it is a retrospective one. That means it did not include the follow-up of patients but according to all the data above, which means recovery of arthralgia and arthritis on GFD in previous studies, the fact that all our patients had clinical or biological signs suggestive of CD and that their articular manifestation was resistant to medical treatment, we could say that our five patients, will give importance to advantage of GFD.
We conclude that the routine tests for CD in patients with unexplained arthralgia or arthritis may increase the proportion of diagnosed cases, thereby reducing the complication rate. Further studies are urgently needed to know whether serological screening for CD must be added in the exploration of arthritis if another cause is not found. CD is frequent in our country in which the staple food is bread. So, prevention for CD must be undertaken. Introduction of gluten to the infant’s diet when infant is still being breastfed at the age of 6 months reduces the risk of CD in early childhood [32–34].
This study is supported by: Unité de recherche: Auto-immunité et Allergie (03/UR/07-02), Faculté de Pharmacie de Monastir, Tunisia.
Conflict of interest
None of the authors have conflicts of interest to declare.