Celiac disease (CD) is a chronic inflammatory bowel disease triggered by the ingestion of gluten in genetically susceptible individuals, who test positive for human leukocyte antigen (HLA) DQ2 and/or DQ8. Its prevalence is about 1%, and since the small intestine is its main target organ, CD can have gluten-related gastrointestinal manifestations, such as bloating, abdominal pain, diarrhea, and constipation [1
]. What is more, CD is a multisystem disorder, and patients can also complain of extraintestinal signs and symptoms [4
]. CD diagnosis in adults is usually based on positive results of specific serological tests for anti-endomysial antibodies (EMA) and anti-tissue transglutaminase (tTG) antibodies performed during a free diet and then confirmed by the finding of intestinal villous atrophy on histological examination of duodenal biopsies. The only treatment currently available is a lifelong and strict gluten-free diet (GFD) but, nevertheless, many CD patients complain about the persistence or relapse of symptoms even during GFD [6
]: in this case, interviews with gastroenterologists and nutritionists are necessary in order to investigate a proper adherence to the GFD; moreover, repetition of serological tests and duodenal biopsies are mandatory to exclude ongoing intestinal damage and gluten exposure. When persistent damage in the duodenal mucosa is found despite a correct GFD, refractory CD and possible complications such as intestinal lymphoma must be investigated [7
Moreover, in clinically non-responding CD, other possible overlapping diagnoses should be considered, such as inflammatory bowel diseases (IBD), and irritable bowel syndrome (IBS)-like disorders (e.g., lactose intolerance), but many cases seem to remain unsolved [8
]. Recently, a diet low in fermentable oligo-, di-, and monosaccharides and polyols (FODMAPs) has been proposed as an ex adiuvantibus treatment to reduce IBS-like symptoms in CD patients following a GFD, although there is no specific indication or supporting diagnostic test [9
More recently, nickel (Ni) allergic contact mucositis (ACM), which is linked to the ingestion of Ni-rich foods, has been added to IBS-like disorders. Together with Ni allergic contact dermatitis (ACD), Ni ACM is an expression of “systemic Ni allergy syndrome” (SNAS) and can have both gastrointestinal and extraintestinal manifestations. According to the European Surveillance System on Contact Allergy (ESSCA), the prevalence of an epicutaneous patch test positive to Ni may reach 30% in some European countries, but Ni ACM prevalence may even be higher [11
]. Patients affected by Ni ACM show a low-grade intestinal inflammation with a local adaptive response to Ni-containing foods: this mucositis seems to be characterized by increased lymphocyte trafficking (type IV immune response) [12
]. Ni ACM diagnosis is currently based on a Ni oral mucosa patch test (omPT), which has already proved good sensitivity and specificity [13
], and a low-Ni diet (LNiD) can be thus suggested in this condition in order to significantly reduce both Ni-related gastrointestinal and extraintestinal symptoms [14
shows the main foods with the highest Ni content, and it is easy to observe that many of them (e.g., corn) are consumed in large quantities by CD patients on a proper and strict GFD [15
]. It is therefore possible that a high load of alimentary Ni may induce or exacerbate a “Ni sensitivity” in predisposed subjects, especially in CD patients on a long-term GFD.
On these premises, our aims were (1) to study the prevalence of Ni ACM in CD patients in serological and histological remission with relapsing symptoms; (2) to evaluate the effects of an LNiD on gastrointestinal and extraintestinal symptoms in these patients.
The persistence or recurrence of gastrointestinal and/or extraintestinal symptoms in CD patients during GFD is a very common condition and is a topic of great relevance. The causes of this problem are to be initially searched in an incorrect adherence to GFD and this is what can happen in those patients who still show persistently positive antibody titers and significant duodenal histological alterations, even despite quite a long period of GFD (>12 months). Furthermore, the possibility of refractory CD is always to be taken into consideration [4
This issue becomes even more difficult to decipher and solve when CD patients reach serological and histological remission, but symptoms are still present or show a new peak despite a correct GFD. Furthermore, some of these patients may even report the appearance of new symptoms never complained about before. Other overlapping disorders, such as IBS or IBS-like disorders may be the causes of these symptoms and, in this regard, encouraging results have been obtained by ex adiuvantibus use of a low-FODMAP diet, although this approach revealed some limitations that will be discussed further [9
On the other hand, many gluten-free foods consumed by CD patients are high in Ni content. Therefore, once a CD diagnosis has been obtained, the progressive Ni load induced by the GFD can trigger a relapse of symptoms in subjects predisposed to Ni allergy. Ni is present in many foods with different concentrations. It can be responsible for SNAS, which can have both gastrointestinal and extraintestinal manifestations. Specifically, Ni ACM is estimated to be one of the most common IBS-like disorders, and its diagnosis can rely on an Ni omPT, more sensitive and specific than the epicutaneous patch test [11
]. What is more, excellent clinical results have already been obtained by Ni omPT and LNiD in the management of IBS-like and extraintestinal symptoms of women suffering from endometriosis who were still symptomatic despite different treatments [15
Based on these considerations, we investigated the prevalence of Ni ACM in still symptomatic CD subjects after appropriate GFD and studied the effects of an LNiD on their gastrointestinal and extraintestinal symptoms.
Firstly, we selected symptomatic CD patients on a proper GFD who had no more serological and histological signs of disease activity. Then, we excluded those who did not meet the minimum clinical criterion by means of the GSRS questionnaire: in this way, we eliminated the less disabling and most confounding clinical pictures, even if this has led to a reduction in the number of patients studied. Other possible overlapping confounding pathologies have been excluded, such as lactose intolerance, HP infection and IBDs.
Our results showed an Ni ACM prevalence of 100% in the final 20 patients actually enrolled (Figure 3
a,b): this percentage may appear extraordinarily high, but it is mandatory to consider not only the high prevalence of Ni ACM in the general population (estimated to be even greater than 30%) but also the strict exclusion criteria previously applied [11
]. These 20 patients with Ni omPT positive results should be contextualized among the 85 CD patients who had a significant symptomatic picture: thus, Ni ACM should have a prevalence of at least 23.5% in our study. This percentage could have been higher considering not only the dropouts but also those patients affected by other pathologies who had been excluded in recruitment phase: in fact, Ni ACM can also easily overlap with other disorders, especially lactose intolerance, and in our study we excluded 54 lactose intolerant patients (about 63.5% of the 85 patients with a significant clinical picture) [11
Afterward we focused on the effects of the different diet regimens on the symptoms.
First, we confirmed a general clinical improvement after CD diagnosis and a correct GFD (T0 vs. T1): once serological and histological remission were achieved, about 80% of the 24 total symptoms improved (almost 60% was also statistically significant).
On the other hand, a prolonged strict GFD (T1 vs. T2) resulted in a general clinical relapse involving more than 80% of all symptoms (the worsening was statistically significant in more than 40% of the symptoms). This negative change may be attributed to a GFD-related load of Ni in already sensitive or predisposed subjects.
This theory seemed to be confirmed after a balanced restriction of Ni-rich foods: in only three months, GFD plus LNiD (T2 vs. T3) induced an improvement of more than 80% of the symptoms and in the half of the cases the improvement was statistically significant, including the most complained about and disabling symptoms, such as abdominal pain, swelling, increased evacuations, and loose stools. Moreover, many of them got even better compared to the initial GFD alone (T1 vs. T3), although this difference was not statistically significant. As regards extraintestinal symptoms in the T2 vs. T3 range, dermatitis, headache, and fatigue statistically improved. Dermatitis deserves a special mention, as it showed a very peculiar trend: at the beginning (T0 and T1) it was almost totally absent, then it was significantly exacerbated reaching an acute peak at T2 and finally significantly reduced/resolved after GFD plus LNiD at T3. The curve of dermatitis’ clinical course can further suggest the interference of an “alimentary trigger factor” during prolonged GFD, which is not related to gluten contamination, as demonstrated by the negative serological and histological results. The dietary profile of CD patients on a GFD and the impressive results of both Ni omPT and LNiD, would confirm that alimentary Ni overload is able to unmask/exacerbate not only gastrointestinal but also systemic symptoms, over a medium to long-term time period.
It should be emphasized that no symptoms significantly improved or worsened by comparing T0 (free diet) with T2 (Ni overload during GFD), suggesting a close clinical similarity between these two times: this is what often makes gastroenterologists think that the cause of the clinical relapse is a new gluten contamination.
The comparison between free diet and GFD + LNiD (T0 vs. T3) led to an improving trend of more than 70% of the symptoms: this is a very good percentage, although slightly lower than the almost 80% obtained from GFD alone, before Ni overload (T0 vs. T1). This may mean that, although LNiD is very effective in achieving a new clinical remission, Ni-sensitive patients cannot completely eliminate Ni, therefore, Ni-related symptoms, from their GFD.
The comparison between the well-being obtained by the initial GFD alone and the well-being obtained by the addition of the LNiD (T1 vs. T3) is also interesting: these two stages showed no statistically significant difference in symptoms’ intensity perceived. This means that after the relapse peak in T2 (Ni load), a correct dietary intervention (LNiD) is able to completely restore well-being again.
The absence of a trial design was a limit of our study. Furthermore, it was carried out in a single center, the final sample size was quite small, and finally resulted in including only female patients. In addition, the very high prevalence of Ni omPT positive results (100%) may seem misleading.
Firstly, our results must be contextualized in the initial larger pool (102 total patients): the choice to exclude from the study those patients with less marked symptomatic pictures certainly led to the underdiagnosis of many other Ni-sensitive patients. In addition, it should be considered that Ni ACM can coexist with other disorders and can overlap with them from the symptomatic point of view [11
]: in our study many HP-positive and lactose intolerant patients (almost 60% of the 102 patients initially recruited) were excluded for methodological correctness, and, thus, many other Ni-sensitive patients were probably lost among them.
The fact that the final 20 patients studied were only females is probably due to the greater prevalence of females in CD: in literature the female/male ratio is estimated to be about 3:1 and this proportion is approximately preserved in the 102 patients recruited at the beginning [27
]. Furthermore, it has been described that Ni can act as a metalloestrogen and, thus, may have a greater influence in women with both extraintestinal and gastrointestinal clinical manifestations [15
Given the strict differential diagnosis previously performed, this would explain such a high percentage (100%) of Ni omPT positive results. Moreover, the specific and successful treatment by LNiD seemed to confirm the appropriateness of our assumptions and supported the Ni omPT positive results: as above mentioned, more than 80% of symptoms improved after LNiD and about the half of them were statistically significant. In this regard, if we had not used the Bonferroni correction (p < 0.05 instead of p < 0.0083), the improvements of some other important symptoms (borborygmus, foggy mind, muscle pain, and joint pain) would have resulted statistically significant. We hope that future trials with larger populations will be able to confirm these preliminary observations.
Another weak point of our study, as well as possible obstacle for future trials, may be the impossibility to accurately measure Ni contained in foods and biological samples from patients studied. There is some variability of Ni content in foods and, to date, there are still no standard methods to measure it routinely: if these methods existed, we would have the possibility to prescribe highly personalized diets, more effectively monitor the intake of Ni-rich foods, and make even more appropriate comparisons. As recently demonstrated, we can successfully overcome this limit by prescribing patients a balanced LNiD on the base of an estimated average content of Ni in foods and under direct control of trained dieticians. A daily dietary diary and a detailed interview were also used to verify correct adherence to the GFD and LNiD.
It has already been described in literature that a low-FODMAP diet as an ex adiuvantibus treatment can benefit still symptomatic CD patients on a GFD. However, this dietary intervention has not so far been supported by specific diagnostic tests [9
]. Moreover, it is known that many foods with estimated high content of FODMAPs may also cause other IBS-like disorders, such as lactose intolerance. It is therefore possible that during a low-FODMAP diet, other underlying and unrecognized diseases are treated. In this regard, we also observed a significant overlap between FODMAP-rich and Ni-rich foods (Figure 1
), especially corn and other gluten-free foods consumed by CD subjects. It is therefore possible that the benefits of a low-FODMAP diet can depend on a concomitant involuntary LNiD in unrecognized Ni-sensitive patients. Given the high prevalence of Ni ACM and our results, this is more than only a hypothesis. On the other hand, our study can claim among its strengths an accurate preliminary differential diagnosis with other common IBS-like disorders and organic diseases. In addition, a targeted LNiD was prescribed after a reliable specific diagnostic test (Ni omPT), thus avoiding unnecessary dietary exclusions [15
It may also be interesting to discuss the possible effects of close contact with patients during follow-up: frequent and extensive dietary interviews can have a placebo effect in clinical setting, capable even of inhibiting a symptom. On the other hand, they are essential methodological tools for an adequate quality assessment of the diet followed, as well as for the determination of the patient’s clinical status. Close clinical contact also seemed to play a relevant role in supporting such “delicate” patients who had to follow two strict diet regimens: GFD and LNiD. This was confirmed not only by the encouraging clinical results obtained but also by the very low number of drop-outs. In addition, close clinical contact appeared even more necessary for such a study spread over a long period of time: the long time span of observation could have led to misestimation of the beneficial or harmful effects of dietary interventions. Finally, the inclusion of a significant number of symptoms in the GSRS standardized test for clinical evaluation has most probably helped to further reduce possible placebo/nocebo effects and misestimation of the results.