Introduction

Celiac disease (CD) is a life-long autoimmune condition [1] mainly of the gastrointestinal tract, affecting the small intestine of genetically susceptible individuals. Gluten, which is the storage protein of wheat and its alcohol soluble gliadins are the offending inducers of the disease together with structurally related molecules found in barley and rye. Nevertheless, additional environmental factors like infections might play a role in CD induction [2]. Tissue transglutaminase (tTG) is the autoantigen against which the abnormal immune response is directed to [3] and two main auto antibodies, anti-endomysium and anti-tTG, are the most useful serological markers to screen for the disease [4]. Recently, two additional autoantibodies, namely anti-deaminated gliadin peptide and anti-neoepitope tTG, were found to be reliable for CD screening [5]. HLA-DQ2 and HLA-DQ8 molecules are the most important, so far known, predisposing genetic factors. A lot is known on the pathogenesis of the disease. The sequential chain of events operating in the disease was recently unraveled and gives the hope for future therapeutic strategies [6]. Furthermore, its epidemiology, prevalence, and clinical presentation are changing constantly, and with time, new clinical presentations are depicted and widen the plethora of clinical variability of CD [7].

It has been shown that the classical intestinal clinical picture is disappearing, and the extra-intestinal presentation is emerging. Skin, endocrine, hepatic, skeletal, hematological, gynecological, infertility, dental, and behavioral abnormalities are often described [8–10]. With the growing awareness of family practitioners, hematologists and gastroenterologists, and now gynecologists and neurologists, the diagnosis of CD is increasingly being made during the whole life-span. In fact, about 20% of newly diagnosed cases occur in patients who are older than 60 years of age.One of the growing domains is the extra-intestinal presentations of CD affecting the bone, presenting as osteopenia or osteoporosis, fractures, and falls [11–18].

It is becoming increasingly clear that vitamin D has a wide range of biological activity in addition to the classic endocrine pathway affecting bone state, thus making this hormone important to many aspects of general health. Ecological and epidemiological studies suggest that vitamin D insufficiency plays a role in the development of colorectal, breast, prostatic, ovarian cancers and lymphoma, autoimmune diseases like multiple sclerosis, type 1 diabetes, systemic lupus erythematosus, rheumatic arthritis, hypertension, infectious diseases, diabetes, cardiovascular disease, musculoskeletal disorders, asthma, as well as several psychiatric conditions such as schizophrenia, depression, and dementia [19–21].

Vitamin D is closely linked to autoimmunity. Many autoimmune conditions are more frequent in geographical areas far from the equator and/or less sunny ones. Low serum vitamin D levels are found in patients with autoimmune diseases, and higher levels are inversely correlated with the incidence of some of them. Anti-vitamin D antibodies were found in a subset of patients with autoimmune diseases. Many animal models of autoimmune diseases identified vitamin D as an immunomodulator of the autoimmune process. Vitamin D suppresses several autoimmune pathways including the Th1, B cells, Th-17, dendritic cell, and co-stimulatory molecule systems. Furthermore, polymorphic classical genes, important in vitamin D functions like VDR and CP27B, are associated with susceptibility to several autoimmune diseases. Those epidemiological, immunological, and genetic data are summarized in Table 1 [22–64]. More so, beneficial effects of vitamin D supplementation or vitamin D agonists' therapy further strengthen this causal relation (Table 1). In a recent study on different clinical subgroups of multiple sclerosis patients, 1,25(OH)(2)D(3)was found to play an important role in T cell homeostasis during the course of the disease, suggesting that correction of vitamin D deficiency may be a useful therapeutical strategy, affecting the clinical course of multiple sclerosis [63].

Table 1 Potential roles and involvement of vitamin D in autoimmunity
Full size table

Unfortunately, it is estimated that as many as one billion people worldwide suffer from vitamin D deficiency or insufficiency (commonly defined as levels below 20 and 20–30 ng/ml, respectively), and this was shown to be prevalent across all age groups, genders, and geographic regions [65–74].

The intestine plays a critical role in bone homeostasis in the normal population as well as in various gastrointestinal disorders [75–77]. An increased prevalence of CD in the osteoporotic population was reported, and serological screening was recommended [78]. Others casted doubt on the association and opposed routine CD screening [79]. On the other hand, numerous studies depicted reduced bone mineral density in untreated CD possibly due to calcium and vitamin D malabsorption, release of proinflammatory cytokines, and misbalanced bone remodeling [13, 80–82]. Only few assessed serum vitamin D metabolites and none compared them along the life span in an interfamilial design and compared two sunny countries with a different genetic background.

The present study was undertaken to fill this gap. The main aims were to investigate vitamin D status in children and adults, in two Mediterranean sunny countries like Spain and Israel, in an inter- and intra-family structures in celiac populations. The secondary aims were to challenge the call to investigate all the celiacs for osteopathy, vitamin D levels determination, and their routine vitamin D supplementation, and additionally, to try to understand the differential role played by the genetic background compared to age in the pathogenesis of celiac bone pathology.

Material and Methods

Study Populations

Serum vitamin D levels were determined in 272 individuals, divided to five groups. Group 1 comprised of 51 Israeli children with definitive CD (age, 6 ± 4 years, M/F ratio 29:32) and group 5 included 59 Spanish CD affected children (age, 4 ± 4 years, M/F ratio 37:22). Group 2 comprised of 56 Israeli children diagnosed with nonspecific abdominal pain (age, 8 ± 5 years, M/F ratio 27:29). The pediatric populations were compared to two adults ones: group 3 include 84 adults, parents of group 2 (age, 39 ± 8 years, M/F ratio 49:35), and groups 4–22 Spanish adults with CD (age, 44 ± 13 years, M/F ratio 16:6).

The following information was collected on the Israeli population: Diet: gluten-containing or free diet. Symptoms: abdominal pain, short stature, vomiting, diarrhea, anemia, failure to thrive, and IgA deficiency. Familial diseases: CD, Diabetes mellitus type1/2, FMF, IBD, thyroid disease. Ethnicity: Ashkenazi (European and North American origin)/Sephardic (African and Asian origin)/mixed (Ashkenazi and Sephardic) Jews, Christians, Moslems and Druze, Laboratory parameters: complete blood count, biochemical profile, IgA levels, CD serology: antiendomysial, anti-tissue transglutaminase antibodies. Degree of small intestinal injury: 0 = normal, 1 = slight inflammation, 2 = moderate inflammation and moderate villous atrophy, 3 = mild inflammation and sub-total villous atrophy, 4 = heavy inflammation and total villous atrophy. Presence of Helicobacter pylori in the antral biopsy.

Celiac disease was diagnosed according to the revised criteria of the European Society for Pediatric Gastroenterology and Nutrition based on specific serology (anti-endomysial and/or anti-tissue transglutaminase antibodies by immunofluorescence and ELIZA, respectively) and duodenal biopsies [83]. All the participants were on a gluten-containing diet and had physical examination, laboratory workup, and celiac serology. Endoscopy with at least four duodenal and one antral biopsies was performed in all the groups, except group 3. On the day of endoscopy, 5 ml of peripheral blood was withdrawn, centrifuge 5,000 c/s for 10 min, and the serum was frozen in −80°C until assayed for vitamin D metabolites. None of the participants received vitamin D supplementation.

The study was approved by the local Helsinki Committee in Spain and Israel and informed consent was obtained from all subject or caregivers.

Vitamin D Measurement

A commercial kit, LIAISON® 25(OH)D Assay (DiaSorin, Italy) was used in order to measure serum concentration of 25(OH)D in sera of subjects. The method for quantitative determination of 25(OH) D is a direct, competitive chemiluminescence immunoassay. Specific antibody to vitamin D is used for coating magnetic particles (solid phase), and vitamin D is linked to an isoluminol derivative. During the incubation, 25(OH) D is dissociated from its binding protein and competes with labeled vitamin D for binding sites on the antibody. After the incubation, the unbound material is removed with a wash cycle. Subsequently, the starter reagents are added and a flash chemiluminescent reaction is initiated. The light signal is measured by a photomultiplier as relative light units and is inversely proportional to the concentration of 25(OH) D present in calibrators, controls, or samples.

As different investigators have defined the normal lower limit in the serum as 25(OH)D levels below 12 ng/ml (30 nmol/l) or below 20 ng/ml (50 nmol/l) [84, 85], in the present study, we defined vitamin D deficiency as levels of 25(OH)D below 20 ng/ml, according to the new definition [85]. Vitamin D insufficiency was considered to be at levels between 20 and 30 ng/ml (50–75 nmol/l).

Statistical Analysis

Data analysis was performed using the SPSS statistical package version15.

Comparison of prevalence rates (i.e., vitamin D deficiency and insufficiency percentages) between groups was preformed by chi-square test or Fisher exact test as was appropriate. Continuous variables are expressed as mean ± standard deviation throughout the manuscript.

ANOVA was performed for comparison of vitamin D levels between groups. The Bonferroni correction was applied in case of multiple comparisons. Pearson correlation coefficients were calculated. All p values are two-sided, with a significance level of p < 0.05.

Results

The Israeli and Spanish CD children (groups 1 and 5, respectively) had the highest levels of serum vitamin D metabolites (30.3 ± 12.3 and 25.6 ± 9.7 ng/ml, respectively), compared to the two adult groups (groups 3 and 4) who had the lowest ones (20.7 ± 10.7 and 20.2 ± 10.5 ng/ml, respectively). Vitamin D levels in the two pediatric Israeli groups (groups 1 and 2, 25.6 ± 9.7 and 24.9 ± 11.4 ng/ml, respectively) were comparable. Concerning 25(OH) D sera levels in the five groups, only the difference between groups 5 and 4 was significant (30.3 ± 12.3 and 20.2 ± 10.5 ng/ml, respectively p < 0.003) (Table 2).

Table 2 Mean serum vitamin D metabolites levels in the five groups
Full size table

When vitamin D was splited above and below (deficiency) 20 ng/ml level, 54.5% of Spanish adult CD had vitamin D deficiency, compared to 16.9% of the local CD children (p < 0.001). In group 2, 29.6% had deficient levels compared to their parents with 50% (p < 0.019). Comparing the two CD pediatric population, groups 1 and 5, vitamin D deficiency was more prevalent in the Israeli CD children (33.3% compared to 16.9%, respectively (p < 0.046; Fig. 1).

Fig. 1
figure 1

Percentage of individuals with vitamin D deficiency (<20 ng/ml) in five groups: (1) CD children from Israel. (2) Recurrent abdominal pain (RAP), Israeli children. (3) Parents of RAP children. (4) CD Spanish adults. (5) CD Spanish children. *p < 0.046 comparing CD group 1 to 5. **p < 0.019 comparing intra-familial groups 2 to 3. ***p < 0.001 comparing Spanish groups 4 to 5

Full size image

Concerning origin and ethnicity in the Israeli populations, a significant correlation was found between serum vitamin D levels in Ashkenazi + mixed compared to Sephardic Jews (27.8 ± 10.3, 22.7 ± 10.0 ng/ml, respectively, p = 0.02), Ashkenazi + mixed to Moslem + Druze subjects (27.8 ± 10.3, 16.5 ± 9.4 ng/ml, respectively, p = 0.0001), Sephardic Jews to Moslem + Druze (22.7 ± 10.0,16.5 ± 9.4 ng/ml, respectively, p < 0.012), Christian to Moslem + Druze groups (31.6 ± 10.9,16.5 ± 9.4, respectively, p = 0.01). Grading mean serum vitamin D levels, from the highest to the lowest levels: Christian, Ashkenazi and mixed Jews, Sephardic Jews, Moslem and Druze (31.6, 27.8, 22.7, 16.5 ng/ml, respectively, p = 0.0001). It can be concluded that there is a west to east gradient in vitamin D status.

No correlation was depicted between serum vitamin D metabolites levels and gender, celiac serology magnitude, degree of intestinal injury, and anemia. However, the following parameters correlated to vitamin D levels: age-negative correlation (−0.332, p < 0.0001), IgA levels-negative correlation (−453, p < 0.006), FTT-positive correlation (vitamin D levels in FTT/no FTT, 29.3/22.1 ng/ml, respectively, p < 0.001). There are more percentage of vitamin D deficiency among the non-FTT children, compared to the FTT ones (43.5%, 19.4% respectively, p < 0.01), H. pylori presence-negative correlation (vitamin D levels in Hp+/Hp−, 21.2/26.5 ng/ml, respectively, <0.03). There are more percentage of vitamin D deficiency among the antral Hp+ patients compared to the Hp− ones (50%, 25%, respectively, p < 0.02).

Discussion

Regardless of the definition of vitamin D insufficiency or deficiency, it is clear that suboptimal levels are a worldwide phenomenon with hardly any region spared. Moreover, the Middle East and populations, despite the abundant sun and favorable attitude of the regions, is the most deficient ones [73, 74]. Recent studies emphasize the high frequency of vitamin D insufficiency, even in sunny countries like Florida, Turkey, Lebanon, Tunisia, Jordan, and Saudi Arabia [73]. In the present study, 95 out of 272 (34.9%) subjects are vitamin D deficient with a serum level <20 ng/ml, and 182 (66.9%) are insufficient, having serum levels <30 ng/ml, thus complying with the world pandemic of hypovitaminosis D.

The present study design allows us to study the effect of age, intestinal inflammation, and familial factors on vitamin D status. Choosing two sunny countries along the same geographical area of the Mediterranean, with comparable latitude and subject with the same intestinal disease, eliminates many confounding factors that may bias the conclusions. It is known that geography, older age, higher latitude, winter season, darker skin pigmentation, less sunlight exposure, dietary habits, and absence of vitamin D fortification or non-adherence to vitamin D recommendations among infants are the main conditions that might attribute to hypovitaminosis D.

In order to minimize the age effect, a pathological control group of age-matched children with RAP (group 2) was chosen and compared to an age-matched CD children, coming from two different countries: Israel (group 1) and Spain (group 5). To challenge the age effects, the non-CD pediatric group (group 2) was compared to their parents, thus equilibrating genetic factors and dietary habits. Additionally, the two Spanish group, suffering from the same disease, living in the same climate, latitude, social and dietary habits, but not intra-family relating, permitted to study the age effect on vitamin D levels. The significant negative correlation between the age and vitamin D levels and the significantly lower vitamin D levels in the adult CD Spanish group compared to their pediatric counterparts, strengthen the age effects. Furthermore, the significantly higher percentage of vitamin D deficiency in the two adults group, compared to the three pediatric groups, not related to CD diagnosis, goes along the same conclusion. It is concluded that the age is the major factor affecting serum vitamin D levels. This conclusion is supported by several recent and old dated studies from US, England, and Israel [65, 73, 86, 87].

Several explanations are suggested for the negative effect of the age on vitamin D levels in the celiac patients: (1) Adults are less compliant to gluten-free diet. Gluten elimination improves bone osteopathy and nutritional deficiencies. (2) Children are more exposed to sunlight, being unaware of its deleterious effects. (3) Preventive vitamin D supplementation is a routine in the first year of life, thus delaying the deficiency. (4) Children consume more milk and milk-based products, sometimes fortified with vitamin D. (5) Unrelated to CD, vitamin D deficiency, osteopenia/porosis are age related. It is worthwhile noting that difference in bone mass between children and adults might be explained by the following factors:

a. Children tend to be more motorically active, thus prevent osteopathy b. Smoking and sedentary life is more prevalent in adults. Both are involved in reduced bone mass. c. Gluten elimination in children is much more effective normalization of bone mineral density, compared to adults [81]. d. In general, gluten-free diet is inadequate in calcium and vitamin D [88].

On the same background, it appears that the intestinal inflammation in CD is not a major contributor to the decreased vitamin D. The comparable vitamin D levels in the three pediatric, CD and non-CD, groups are against intestinal inflammation and/or inflammatory factors affecting serum vitamin D. Moreover, one can conclude that malabsorption and dietary deficiencies, sited as contributing to vitamin D deficiency/osteopenia/osteoporosis in CD [11, 13–15, 75–77, 81] are a minor, if at all contributors.

As the genetic and familial influences are concerned, the comparison of the percentage of vitamin D deficiency between the RAP group with their parents further minimizes their influences on vitamin D and maximizes the age effect.

Vitamin D receptors are found in the crypts of CD intestine and alkaline phosphatase, a vitamin D regulated enzyme, exists on their surface enterocytes [89]. Furthermore, no correlation was found, in the present study, between serum vitamin levels and celiac serology levels and the degree of the intestinal injury. Since the amount of those autoantibodies is positively correlated to the degree of intestinal atrophy and the magnitude of the inflammatory infiltrate [90], the minor role played by vitamin D malabsorption in CD is supported.

The west to east gradient in vitamin D status found in the present study is supported by a recent observation in Israel were Ashkenazi Jews had higher vitamin D levels then the Sephardic ones who had a higher levels than the Arabs[74].

The ethnicity influence can be explained by several raisons: (1) Sephardic Jews, Arabs, and Druze have darker skin pigmentation compared to the Christian and the Ashkenazi populations, resulting in reduced skin synthesis of the vitamin [19, 65, 71, 73, 91]. (2) Arab and Druze communities' lives in villages in Israel are more conservative and wear long-sleeved clothes, thus contributing to reduced light skin exposure. The darker skin complexion and the customary clothing most probably join cultural-related habits and other genetic factors distinguishing those ethnic groups. In view of the multiple roles of vitamin D against cancer, autoimmune diseases, infections, diabetes, hypertension, and cardiovascular events, the question arises if the described ethnical hypovitaminosis D affects morbidity/mortality of those communities. Such a phenomenon was described in the USA [19, 65] and recently in Israel [74].

The positive correlation between FTT and vitamin D serum levels is surprising. Is it because those children are supplemented with multivitamins, more often then their non-FTT counterparts?

Hp gastric presence is negatively correlated to serum vitamin D levels. It is plausible that due to the gastric mucosal atrophy and decreased stomach acidity, associated with Hp antritis, vitamin D absorption is affected. Vitamin D deficiency is potentiated after bariatric surgery and after highly selective vagotomy [92, 93]. Alternatively, Hp induces a plethora of proinflammatory cytokines: IL-8, IL-12, gamma-interferon, TNF-alfa [94, 95]. Lastly, since Hp gastritis is accompanied by abdominal pains, nausea and refusal to eat vitamin D consumption may be decreased. Independently of the specific cause, those events could affect vitamin intestinal absorption.

Several studies on CD and the related osteopathy and its consequences recommend routine serum vitamin D evaluation and bone mineral density determination [14, 15, 85, 96–98] while others challenged this approach [16]. Furthermore, some recommend preventing approach by routine vitamin D fortification of the CD population [85, 99, 100]. The present data are going against the routine preventive treatment or fortification of vitamin D in the celiac patients. It is suggested that the approach should be selective. In the pediatric CD age, vitamin D and bone assessment should be applied to malabsorption or chronic diarrhea children. Since a significant negative correlation exists between age and adult CD, it is recommended to investigate this group of age for vitamin D levels and osteopathy and treat accordingly. An alternative approach might be installed by vitamin D routine fortification.

In summary, vitamin D insufficiency and deficiency are prevalent among the CD affected patients. The phenomenon is age and not intestinal injury/inflammation depended, ethnicity and not gender related, and occurs even in sunny countries like Spain and Israel. It is recommended to routinely check the adult CD patients for vitamin D levels and to act selectively in the pediatric CD population.