Prevalence of autoantibodies in patients with juvenile idiopathic arthritis: results from the German inception cohort ICON-JIA

Juvenile idiopathic arthritis (JIA) is the most common chronic rheumatic disease in childhood. According to the current International League of Associations for Rheumatology (ILAR) classification, 7 categories can be differentiated based on clinical and laboratory parameters [1]. The pathogenesis is unclear, but it is often referred to as autoimmune arthritis, especially for oligoarthritis and seropositive and negative polyarthritis.

The co-occurrence of JIA with other autoimmune disease is a matter of debate [2]. However, individual studies come to different results regarding the prevalence of autoimmune diseases in JIA patients, so that screening examinations are not routinely carried out. This can partly be explained by the fact that autoimmune diseases are initially asymptomatic. They develop over a long period of time, while laboratory markers that can indicate the presence of an autoimmune disease are often only used for diagnosis when irreversible tissue damage has already occurred [3].

Data from a single-center analysis in Italy with 79 patients showed that 15.2% of JIA patients had at least one autoimmune disease in addition to JIA. Autoimmune thyroid disease was found to be most common (10.1%) [4]. Another study (n = 151) reported a 7-fold increased risk for celiac disease and a high prevalence of autoimmune thyroiditis (11.9%) together with a high rate of subclinical hypothyroidism (9.3%) in JIA [5]. In an Austrian study, JIA patients (n = 95) were found to have a 14-fold increased risk of developing celiac disease [6]. A large cross-sectional study using two United States administrative healthcare claims databases compared the prevalence of multiple autoimmune diseases of more than 29,000 JIA patients with that of more than 134,000 matched children with attention deficit hyperactivity disorder (ADHD). Almost all investigated autoimmune diseases were more prevalent in patients with JIA, and especially psoriasis and uveitis were significant comorbidities [7]. Similar findings were reported from a comparison of patients with JIA with a control group from the general pediatric patient population at the Cincinnati Children’s Hospital Medical Center [8]. Also a German study showed, that type 1 diabetes is significantly more frequent in patients with JIA [9].

On the other hand, there are also studies showing that other autoimmune diseases, especially celiac disease, are not more prevalent in JIA patients than in the normal population. In a Dutch study, 62 children with JIA were tested for celiac disease. With a prevalence of 1.5%, the results were close to the prevalence of the normal population (Dutch children) [10]. A study from Iran also tested 53 children for anti-tTG IgA (anti-tissue transglutaminase), of which only one child (1.8%) had elevated levels [11]. Another study found no child with elevated anti-tTG levels among 96 JIA patients [12].

The aim of our cross-sectional study was to quantify the presence of autoantibodies in patients with established JIA. We used serum samples from the biobank of the prospective, multicenter inception cohort of children newly diagnosed with JIA (ICON-JIA) in Germany to analyse thyroid and celiac disease-specific antibodies, as well as antibodies with reasonable specificity for connective tissue disorders. Age and gender differences as well as other influencing variables were taken into account.

An association of JIA with other autoimmune diseases is suspected. However, only few and partially conflicting data on the co-existence of other autoimmune disorders are available. The prevalence of autoantibodies in patients with JIA in Germany is not known. We therefore took advantage of biosamples stored in the ICON-JIA study to analyse laboratory parameters that can indicate autoimmune phenomena in patients with JIA. In the present study, the frequency of thyroid antibodies, celiac serology abnormalities and CTD antibodies in JIA patients in Germany was systematically investigated for the first time.

While there have been a few studies for other countries and ethnic groups investigating the association between autoimmune diseases and JIA, no data were available for the German population. However, since it is known that there are different prevalences for different ethnic groups and also within Europe itself, the results cannot be transferred to the all German JIA patients. In addition, against this background it is not surprising that the individual studies came to very different results. Some of them also included only small samples with a size of 50–150 participants. The large prospective, multicenter cohort studies (ReACCh-Out, CAPS, Nordic Cohort Study), which are comparable to the ICON study, investigated the prognosis, treatment or even influencing factors of JIA very precisely, but not the connection of JIA with other autoimmune diseases. Our study population was comparable to that of other multicenter inception cohorts from Canada, Great Britain and Scandinavia (Denmark, Finland, Sweden, Norway) [21‐23]. The sample size, the age at onset and the proportion of female participants are within the range of the other cohorts. The distribution of the JIA categories also largely corresponds to the expected distribution. Minor deviations were noted in this work (ICON-JIA) for systemic arthritis (slightly less frequent at 3%) and seronegative polyarthritis (slightly more frequent at 28%).

Overall, most ICON-JIA values were thus in the range of the general population values. However, a more frequent occurrence in the JIA patients would have been expected, since a connection of JIA was most frequently investigated with celiac disease and in some cases a significantly increased risk was found [5, 6]. Even in the studies that did not find an increased risk of celiac disease, the prevalence was 1.5% [10] and 1.8% [11], which is 3–4 times higher than in our study. This could be in part due to the fact that different celiac antibodies were studied. For example, George et al. studied antigliadin, antireticulin, and antiendomysium antibodies, but not anti-tTG [10]. In addition, most of the samples were very small, ranging from about 50–150 participants. Furthermore, the control groups differed in the different studies. For example, Simon et al. used ADHD patients as a control group, which does not reflect the entire healthy population [7]. Besides, it has to be considered that the patients of this study were at different treatment at time of sampling. This may influence the presence of autoantibodies resulting in a lower level. Anti-tTG IgA and IgG were even higher in controls than in JIA patients. For the CTD screen, no comparable control group was available.

The correlation of JIA and celiac disease is a matter of special interest. However the ESPGHAN guideline (European Society for Paediatric Gastroenterology, Hepatology and Nutrition) defines a risk group (people with certain underlying diseases) with an increased probability of (asymptomatic) celiac disease [24]. JIA is not listed in this category. Since the published studies provided conflicting results, a higher prevalence of celiac disease in JIA patients cannot be clearly assumed. Notably, the samples sizes are partly very small (< 100 study participants) and differ with regard to their ethnicity, and this can influence the prevalence of autoimmune diseases observed even within a region such as Europe [17, 25]. Recently, Lovell et al. compared 2026 patients with JIA and 41572 general pediatric patients via ICD-9 and ICD-10 codes for autoimmune diseases and demonstrated that 14 autoimmune diseases had a significantly higher prevalence in the JIA cohort [24]. In contrast to our study, this study analyzed clinical diagnoses of autoimmune diseases, not autoimmune phenomena. However, the prevalence of clinical autoimmune thyreoiditis among JIA patients was even lower (16/1332; 1.21%) than in our cohort (8/499; 1.60%). Therefore, it appears rather unlikely that this comorbidity was underrecognized in the ICON-JIA patients. For Germany, no results are available for the occurrence of autoantibodies in patients with JIA other than ANA and anti-CCP. It has been discussed whether an association between celiac disease and JIA should prompt us to apply a screening by antibody testing accordingly.

It needs to be noted that the presence of autoantibodies does not prove a clinical diagnosis in patients. However, autoimmune diseases often develop over a longer period of time. It is hypothesized that in genetically predisposed individuals specific autoimmune phenomena can be triggered. In a subclinical phase, autoantibodies can become present. Progressive tissue and organ damage only occurs in the later clinical phase [26]. On the other hand, autoimmune disease may be present even without laboratory proof, and vice versa positive autoantibodies may be present without clinical relevance. In our cohort, a total of 76 patients had either clinically diagnosed autoimmune comorbidity or elevated autoantibodies (15.2%). Of 21 patients with clinical autoimmune comorbidity (especially thyroid autoimmunity), only 8 were also serologically positive at the time of testing, while 55 patients had autoantibodies without clinical diagnosis. In most patients with comorbidity but without autoantibodies, the comorbidity was already noted at inclusion. Autoantibodies may have become negative during therapy. It is conceivable that anti-tissue glutaminase antibodies become negative in celiac disease patients adhering to diet. In our study only one out of three patients with celiac disease was serologically positive, while antibodies were also found in two patients without confirmed celiac disease. On the other hand, a significant number of patients had anti-TPO or anti-TG antibodies despite no thyroid disease recorded. This may warrant a suspicion for the development of autoimmune features in JIA patients as suggested in other reports [3, 27].

There are several limitations to our study. As a major drawback we lack a matched control population. The ICON-JIA control group consisted of non-diseased peers of the JIA patients. For ethical reasons, however, blood sampling from the healthy and especially underage control participants was rejected for study purposes and literature values were used instead. The literature values however, could neither be matched for age nor for ethnic background. It is known that in different ethnic groups partly different prevalence for individual autoimmune diseases may exist [25]. Children and adults are also not equally affected. Autoimmune diseases manifest themselves primarily in adulthood and occur significantly more frequently in this age group (40–50 years of age) [28]. In addition to these demographic differences, each study also has different methodological approaches. For example, the reference values for the individual antibodies are not always identical, which would, however, be more suitable for a direct comparison. Moreover, we only analysed one serum sample per patient, and we might have missed visits at which autoantibodies were present. We choose rather late-stage time-points as we assumed that the development of autoantibodies may evolve over time. In addition, the overall screening of ANA-IFT titers could be considered. We only had ANA-IFT results reported by the physicians at inclusion. Most patients with autoantibodies as tested by us at later time points were ANA-positive at inclusion, but ANA-IFT was not reported for them at the later follow-up time-points. We could only retest those with positive CTD-screen and saw that most of them had elevated ANA, but we do not know the frequency of patients without positive CTD-screen but positive ANA-IFT in the overall cohort. It is conceivable that the CTD-screen is detecting ANA positivity that is often seen in young JIA patients, as opposed to a real predisposition to a connective tissue.

In conclusion, the frequencies of autoantibodies in JIA patients later than one year in the course of the disease did not differ remarkably from literature data on general populations. A strong general correlation between JIA and laboratory-proven autoimmune phenomena could not be confirmed. It must be noted that there was no matched control group and literature data were used instead. Future studies will be needed to further analyse the relevance of coexisting autoimmune phenomena in JIA patients.