Celiac disease (CD) is a permanent, immune-mediated enteropathy caused by gluten ingestion in genetically susceptible subjects. It is characterized by various degrees of villous atrophy in presence of gluten-dependent autoantibodies [1, 2].

The prevalence of CD is currently increasing compared to our experience in the past. Serological findings, such as anti-endomysium (EmA) and anti-tissue-transglutaminase antibodies (tTG-Ab), are very useful in increasing our diagnostic capacity [3–5], but are not always able to predict the histological features [6–8].

The pathogenic cascade that causes the typical histological lesions, characterized by flat mucosa with tissue destruction and reorganization of the intestinal picture, is still partially unknown. In this respect, a role of cytoskeleton has been described: the gluten ingestion has been reported to induce a rapid alteration of the actin network on intestinal mucosa of CD patients [9]. Gliadin rapidly increases actin polymerization leading to rearrangement of actin filaments, especially in the intracellular subcortical compartment [10]. It is likely that newly generated actin polymers may be exposed to gut-associated lymphatic tissue, causing the production of IgA antibodies against actin filaments (IgA-AAA).

Previous studies have described that the presence of antibodies against actin filaments is associated with severe degrees of mucosal damage and that IgA-AAA may also contribute to exacerbate the villous' cytoskeleton damage [11–14]. It has also been suggested that the presence of IgA-AAA may, in some patients, overcome the need of the intestinal biopsy [9].

The aims of this study were to evaluate, using two different assays (immunofluorescence (IFI) and ELISA), the prevalence of IgA-AAA in a group of newly diagnosed CD patients and to verify the relationship between these serological tests and degrees of intestinal lesions. Finally, we verified the reliability of our tTG-Ab IgA test in predicting intestinal mucosal damage.

CD diagnosis was confirmed in 82 subjects: 18 patients presented IIIA lesion, 29 patients presented IIIB lesion and 33 patients had IIIC lesion; two patients had type 1 lesion in presence of positive tTG-Ab and abdominal complaints. The gastrointestinal manifestations of the latter patient can be referred to as minimal immunopathological changes in the intestine exposed to gluten even in the absence of overt CD enteropathy: this condition is called "gluten sensitivity" and is defined by some morphological, immunological, or functional disorder that generally respond to gluten exclusion [18]. All patients started a gluten free diet following the diagnosis.

Eight patients with positive tTG-Ab presented normal mucosa (type 0 M/O). All 20 biopsied controls with negative tTG-Ab had normal histology. The tTG-Ab serum concentration of patients with IIIC M/O (66.14 ± 33.8 AU) or IIIB M/O (53.16 ± 30.76 AU) were higher than IIIA intestinal lesion (32.36 ± 29.98 AU, F = 3.4; p = 0.0123). The 8 patients with normal mucosa and the two ones with minor lesions (type I M/O) presented a low tTG-Ab titre (15.85 ± 11.50 AU).

IgA-AAA tested by IFI and ELISA were detected in 35 (43.7%) and in 38 (47.5%) untreated celiac patients, respectively.

The rate of IgA-AAA positivity by IFI and ELISA in histologically proven celiac disease patients, were 5.5% and 25% patients in IIIA, 27.5% and 34.4% patients in IIIB, 78.8% and 75% in IIIC patients, respectively: showing an high sensitivity in predicting severe intestinal damage.

IgA-AAA were not detectable in control subjects; just one biopsied subject resulted positive to IgA-AAA tested by ELISA.

tTG-Ab assay was plotted against mucosal lesion degree and results are shown in figure 2 (Significant differences between groups were shown by analysis of variance F = 6.7; p = 0.0021).

The diagnosis of celiac disease relies on the demonstration of intestinal histological lesions associated with presence of positive gluten-autoantibodies and or clinical improvement after gluten free diet.

Endoscopic biopsy is an useful tool, but it is often perceived as invasive and dangerous. IgA-AAA has been proposed as a serological marker of intestinal mucosa damage as it seems to correlate with histological lesions.

Recent studies suggest that IgA-AAA test could be considered a useful tool in diagnosis of CD, being a reliable marker of severe intestinal mucosal damage. In this respect the immunofluorescence assay has been suggested as an useful method and ELISA has been demonstrated to be an accurate assay for their determination [11–13].

To add our experience to these data, we studied IgA-AAA antibodies by two different assays (IFI and ELISA), that resulted positive in a similar proportion in newly diagnosed CD patients (43.7% and 47.5%, respectively).

Our data confirm that serum IgA-AAA were more frequently positive in presence of total villous atrophy than in patients with subtotal or mild villous atrophy. Patients with normal mucosa, regardless of tTG-Ab status, presented negative IgA-AAA by IFI assay, thus showing a specificity of 100%.

Because previous reports [19–22] indicated a possible association between IgA-AAA positivity and inflammatory bowel disease or autoimmune/infective hepatic disorders, all patients were investigated for inflammatory indices and hepatic function. None of the tested subject had altered biochemical indices.

These results are apparently in contrast with data previously described in literature [11] that showed higher sensitivity of IgA-AAA due to the presence of IgA-AAA reactivity also in subtotal mild mucosal involvement. This discordance could be explained by some technical bias related to measurements methods, as recently reported [23]. It has been underlined that the sensitivity of IgA-AAA assay can significantly be enhanced by heating or chelating with calcium the serum samples before performing IgA-AAA test. We did not perform such treatment because we strictly followed the manufacturer's instructions.

Although IFI assays are operator-dependent techniques, we believe that this case does not apply to our study because the operator that performed IFI test is a skilled one and the results obtained by IFI are not different from data obtained by ELISA.

On the other hand, tTG-Ab were able to distinguish patients with severe lesions (M/O type 3) from those with milder ones (M/O type 0-1); as recently described [23, 24], we can confirm that IgA-AAA assay does not add to tTG-Ab assay better performances in sensibility, specificity and correlation with the degree of intestinal lesion.

In conclusion, histology remains the corner stone for diagnosis of CD and our data demonstrate that the IgA-AAA assay is not useful in the standard work-up for the diagnosis of CD, particularly because it is not able to identify mild and subtotal villous atrophy. Nevertheless, the IgA-AAA presence suggests that the ingestion of gluten had already caused advanced intestinal mucosal lesions and that IgA-AAA measurement could have a role to support the diagnosis of CD when the histology interpretation is difficult (i.e. patchy distribution), when the biopsy is avoided for clinical reasons (i.e. when a biopsy or anesthesia represents a life-threatening risk), or in case of negative parents' consensus to endoscopy.