NCGS is still an undefined syndrome, with several unsettled issues despite the increasing awareness of its existence . One of the most critical aspects is the absence of biomarkers, making it difficult to reach a clear-cut diagnosis. Consequently, NCGS can be suspected only on the basis of clinical response to gluten ingestion and withdrawal, followed by gluten challenge after both CD and WA have been excluded. The gold-standard assay for confirming NCGS requires dietary elimination, followed by a double-blind, randomized, placebo-controlled (DBPC) food challenge. This procedure is difficult to adopt routinely in clinical practice, and to date, only a few DBPC studies have been performed, showing that in a variable proportion of patients classified as having NCGS, their symptoms were caused by a nocebo effect [15–18]. Indeed, in a number of patients classified as being gluten-sensitive and put on a gluten-free diet (GFD), the double-blind reintroduction of gluten did not result in recurrence of the symptoms, which were instead evoked by the ingestion of placebo. The results of DBPC studies have been controversial and, in some cases, even contradictory, with the same authors who in a first study validated the existence of this syndrome with the central role of gluten as trigger concluding after more recent research that this syndrome is unrelated to gluten and was improved by a low-FODMAPs diet [16, 18].
The above data emphasize how difficult it is to be sure of NCGS diagnosis and, consequently, how its prevalence in the general population can only be defined empirically. Indeed, studies in USA have reported a large variability for NCGS prevalence with figures ranging from 0.6% in primary care to 6% in tertiary care [10, 11]. This high figure of NCGS is likely to be influenced by the fact that many patients decide to avoid wheat and gluten themselves without taking medical advice, because they feel better if they do not eat gluten-containing foods. In one recent study, a reassessment of patients who had decided by themselves to avoid gluten allowed an alternative diagnosis to be made in about 30% of cases; these diagnoses included small intestinal bacterial overgrowth, fructose or lactose intolerance, microscopic colitis, gastroparesis, and pelvic floor dysfunction . In a recent paper from the UK, it was stated that the self-reported prevalence for NCGS was 13% . The mean age of individuals with self-reported gluten sensitivity was 39.5 years (range 18 to 75 years), and 79% were females. About 5% of children and adolescents in New Zealand were reported to be avoiding gluten in their diet, without any evidence for NCGS in many of these cases . Therefore, it is likely that the number of people worldwide following a GFD is far higher than that of patients with true NCGS.
Previous papers on NCGS have shown that this syndrome is much more frequent in adulthood than in childhood, and also in female than in male patients, with a clinical picture characterized by a large combination of both gastrointestinal and extraintestinal symptoms [2, 12, 17, 22].
To our knowledge, our study is the first multicenter prospective survey on patients with suspected NCGS, and it provides a picture of this syndrome in Italian centers for gluten-related disorder diagnosis. On one hand, it must be emphasized that this study has major limitations because the diagnosis of NCGS was highly presumptive in all our patients, being based exclusively on clinical criteria and on exclusion of CD and WA. None of our patients classified as having NCGS has had his/her diagnosis confirmed by the gold-standard method of testing (DBPC) and, as already reported, it is likely that a proportion of patients identified as suspected NCGS improved after GFD only because of the placebo effect. However, the strengths of this study include the large number of patients investigated and the high skill level of the physicians working in the referral centers involved in this survey. All the patients were prospectively evaluated clinically by experts who planned the diagnostic investigation based on blood tests and endoscopy to identify CD or NCGS.
Our study has a number of important findings. First, our results confirm that suspected NCGS is much more frequent in females than in males (5.4 to 1, respectively), and that this syndrome can have its onset at any age, but is much more common in adulthood, with the maximum peak in the fourth decade of life. Based on our data, suspected NCGS was found in 3.19% of a clinical group at risk for gluten-related disorders. Obviously, this prevalence does not reflect that of suspected NCGS in the general population, but, because the established prevalence of CD in the general population is 1%, and the ratio between suspected NCGS and CD in our study was 1.15 to 1, we can extrapolate that the likely NCGS prevalence in the general population should be slightly higher than 1%. In the four participating pediatric centers, the prevalence of suspected NCGS was far lower than that of CD (0.29 to 1), confirming that this syndrome is rare in childhood . The clinical picture of NCGS emerging from our study is a combination of IBS-like symptoms, including abdominal pain, bloating, bowel habit abnormalities (either diarrhea or constipation), and systemic manifestations such as tiredness, headache, fibromyalgia-like joint/muscle pain, leg or arm numbness, 'foggy mind', dermatitis or skin rash, depression, anxiety, aphthous stomatitis, and anemia (both iron and folate deficiency anemia). From a clinical point of view it is relatively difficult to differentiate NCGS from CD because both conditions share the same clinical features. In the large majority of our patients with suspected NCGS, the time lapse between gluten ingestion and the occurrence of symptoms varied from a few hours to 1 day, whereas in CD this interval is much longer (up to weeks or years). In line with the increased awareness of NCGS by specialists, it should be emphasized that the diagnosis was suspected by a gastroenterologist in a large number of cases. However, in more than half of the cases, NCGS was a self-diagnosis made by the patients, who started to avoid gluten in their diet by themselves. In such cases, gluten reintroduction for at least 6 weeks is highly recommended in order to detect clear-cut histological changes consistent with CD. Recently, it has been suggested that a 3-day gluten challenge can disclose possible CD in patients with self-prescribed GFD by evaluating HLA-DQ2-tetramer staining for gluten-specific T cells . Because HLA typing displays an almost absolute negative predictive value for CD, the absence of HLA-DQ2 and DQ8 can exclude CD in patients with suspected NCGS without the need of gluten challenge .
As already emphasized by many papers [16, 17, 24], NCGS is closely related to IBS, which was found in about half of our NCGS subjects. In addition to IBS, one-third of patients with suspected NCGS displays another food intolerance (most frequently lactose intolerance). A subgroup of patients with suspected NCGS (around 20%) have an IgE-mediated allergy with specific IgE to food and inhalants, the most frequent allergens involved being mites, graminaceae, cat/dog hair, and shellfish. Similar to patients with CD, a relatively high number of patients with suspected NCGS develops nickel allergy. Published studies have previously reported that patients with suspected NCGS did not seem to present major autoimmune comorbidities [11, 12]. Unlike these reports, our present study highlights that there is a high prevalence of Hashimoto thyroiditis in patients with suspected NCGS and that other autoimmune conditions such as psoriasis and Graves’ disease seem to be relatively common in these patients. In this respect, NCGS might behave in a similar way to CD, which is characterized by a close association with a wide spectrum of autoimmune disorders.
As already demonstrated , NCGS was commonly detected in first-degree relatives of celiacs. However, the genetic typing excluded any relationship between NCGS and HLA-DQ2 and/or HLA-DQ8, which were found in patients with suspected NCGS at a prevalence very close to that of the general population. Interestingly, laboratory findings indicated that more than 20% of patients with suspected NCGS showed biochemical signs of malabsorption, with low levels of ferritin or folic acid or vitamin D. These biochemical deficiencies, which were less pronounced than in CD, might be due to small bowel inflammation reported in NCGS [2, 14]. In line with this hypothesis, an increased number of intraepithelial lymphocytes, consistent with a type 1 lesion according to Marsh-Oberhuber classification, has been documented in the intestinal biopsy of about 30% of patients with suspected NCGS. In the majority of the suspected NCGS subgroup identified at Bologna center (by UV), a typical linear distribution of CD3+ T lymphocytes in the deeper part of mucosa and clusters of CD3+ T lymphocytes in the superficial epithelium were observed, in accordance with the pattern identified as a possible histological marker of NCGS .
Although no biomarker has been identified to date, previous studies have reported that about 50% of patients with suspected NCGS show positivity for AGA, mainly of the IgG class [11, 12, 17]. These antibodies are not specific for NCGS, being also found in CD, autoimmune liver disorders, connective tissue disease and IBS,as well as in healthy controls , but their finding in patients with a clinical picture consistent with NCGS has been regarded as an element supporting this diagnosis . In this study, AGA IgG were confirmed as the most frequent serological marker found in patients with suspected NCGS, but AGA prevalence was lower than that reported in previous papers (25% vs 50%) [11, 12, 17]. The lower AGA finding in our study can be explained by the fact that many of our patients had diagnosed themselves with a gluten disorder, thus they were already on a GFD when tested, and AGA tend to disappear very quickly in patients with NCGS after GFD . Positivity for AGA IgA, DGP IgG and DGP IgA were found in only a few patients (5% to 6%).