Despite the progressive awareness of its existence, NCGS is still a condition with many unanswered questions. In contrast to CD, the prevalence of NCGS is far from being established since few reliable epidemiologic studies have been so far published [3, 10]. Indeed, the National Health and Nutrition Examination Survey has identified 49 cases of NCGS over 7,762 subjects (age range 6-80 years) in the 2009-2010 period with a weighted prevalence of 0.55% . In a tertiary care center for celiac research, the criteria for NCGS were met by 347 over 5,896 patients observed between 2004 and 2010 with a prevalence of 5.9% . This latter figure should be cautiously taken as a patient selection bias may have occurred. In our last year (2012) experience at the Celiac Disease Centre of St. Orsola-Malpighi Hospital in Bologna, the ratio between the new NCGS and CD cases was 1.6 to 1, confirming a slightly higher prevalence of NCGS vs. CD .
A further aspect of distinction between CD and NCGS is given by the pathogenetic mechanisms underlying these two diseases. Indeed, while both adaptive and innate immunity are well known to have a major role in CD, only innate immunity has been thought to be activated by gluten proteins in NCGS . However, even adaptive immunity may play a role in NCGS as recently suggested . In support of this possibility, previous papers have shown that antibodies, such as AGA, can be detected in more than half of NCGS patients [3, 9, 12, 15]. Although AGA cannot be considered a specific marker for NCGS (as these antibodies can be found in many other conditions including autoimmune disorders and even in healthy people), their positivity in the presence of gluten-related symptoms can be helpful for confirming a diagnosis of NCGS [3, 12, 15, 19]. However, there are no data concerning the effect of the GFD on AGA in NCGS and whether AGA persistence after GFD correlates with a low compliance to gluten-withdrawal and clinical manifestations in this new gluten-related syndrome. In this context, our study provides new findings showing that GFD can improve significantly both gastrointestinal and extra-intestinal symptoms in patients with NCGS. A possible link between gluten and symptoms has been suggested in NCGS patients , although other components, mainly contained in the wheat (i.e. proteins), may trigger symptoms in NCGS. For example, wheat amylase- and trypsin-inhibitors, a complex of proteins triggering innate immunity, could contribute to symptom generation in NCGS . Similarly to IBS, it is likely that FODMAPs may also play a role in evoking gastrointestinal (as well as extra-intestinal) symptoms in patients with NCGS [8, 20]. In this line, recent evidence by Bisierkieski et al. showed that a diet low in FODMAPs resulted in an improvement of the clinical picture of NCGS in IBS patients, thus supporting a major role of these dietary factors, rather than gluten .
In a previous series of untreated NCGS patients we demonstrated that about 50% of cases resulted positive for AGA IgG . Those patients were followed up and retested for AGA in the present study where we highlighted that NCGS and CD have a different immune response to gliadin after GFD. Interestingly, concerning NCGS, the detection of AGA IgG identifies a possible subgroup of such patients. The vast majority of NCGS patients showed AGA IgG disappearance after gluten withdrawal, whereas the same antibodies persisted in 40% of CD patients after GFD. In NCGS patients the negativization of AGA IgG after gluten withdrawal was significantly related to the strict compliance to the diet as demonstrated by the antibody persistence in only one out of the 41 patients with a strong adherence to the diet and in two of the 3 who did not adhere strictly to GFD. Moreover, AGA IgG disappearance in treated NCGS patients was closely related to the good response to GFD with a significant improvement of the clinical picture (i.e., symptom score decreased more than a half compared to that of gluten-containing diet). None of the 39 NCGS patients classified as good responders to the diet maintained AGA IgG positivity, whereas these antibodies persisted in 3 out of the 5 NCGS patients assessed as mild responders to the diet. In contrast to NCGS, AGA IgG in celiac patients persisted after GFD regardless the adherence and the clinical response to gluten withdrawal. The different behaviour of AGA IgG identified in patients with CD vs. NCGS in response to GFD may reflect the different pathogenic mechanisms underlying these two disorders. In fact, CD is a well-recognized autoimmune disease, whereas NCGS is likely a gluten hypersensitivity without an established involvement of autoimmunity [2, 5]. The persistence of AGA IgG in a large proportion of CD patients following GFD can be regarded as an expression of the immunological memory of the autoimmune disorder, whereas the gluten withdrawal in NCGS switches off the immune process and this effect is supposed to lead to the rapid disappearance of AGA . In contrast with AGA IgG, AGA IgA disappearance can be viewed as the result of a strict GFD as well as of a good clinical response in both NCGS and CD patients. Indeed, these antibodies disappeared in all the NCGS patients with a strict adherence to GFD and a good clinical response, remaining positive only in one patient with admitted low dietary compliance and mild clinical response to the diet. Similarly, the finding of AGA IgA in the celiac group on GFD was closely related to the low compliance with the diet and the mild clinical response to the dietary treatment. Our study focused on a specific subset of NCGS patients, i.e. those with AGA IgG positivity and whether GFD may have an effect on NCGS-related symptoms in the "antibody-negative" subset of gluten sensitive patients remains unknown. It is tentative to speculate that even in AGA-negative NCGS patients gluten withdrawal can improve the symptom profile. Clearly, further studies testing this hypothesis are eagerly awaited. The fluctuation of AGA with dietary changes (i.e. gluten withdrawal and re-challenge) in NCGS patients remains another very interesting aspect. In a recent double blind, placebo controlled study, Biesiekierski et al. have shown that NCGS patients in gluten-withdrawal for 6 weeks and re-challenged with high dose of gluten (16 g/day for one week) had an increase of AGA IgG and IgA in 8% and 21% of cases, respectively . These are lower proportions of NCGS patients compared to those detected in patients during a gluten containing diet. Thus, the data provided by Biesiekierski et al. appear quite different from our results, indicating about 50% of NCGS patients with positive AGA IgG. A possible explanation for such discrepancy may be the very short period (one week) of gluten challenge used by Biesiekierski et al. . Probably, this short-time gluten exposure in NCGS patients was not enough to evoke the reactivation of the adaptive immunity to produce AGA in response to gliadin. This exciting possibility should deserve further study.