Introduction
Irritable bowel syndrome (IBS) is a common gastrointestinal functional disease characterized by the presence of chronic abdominal pain or discomfort associated with changes in bowel habits, consisting predominantly of diarrhoea, constipation or alternating patterns; defecation urgency; tenesmus; bloating; and abdominal distension [
1,
2]. Diagnosis of IBS is based on a positive history of gastrointestinal symptoms according to the Rome III criteria in the absence of obvious alarm signs [
3].
Several comorbidities which may occur more often than expected by chance in IBS patients have been identified, including fibromyalgia syndrome (FMS), chronic fatigue syndrome (CFS), gastro-oesophageal reflux disease, headache, backache, genitourinary symptoms, temporomandibular joint disorder, anxiety and depression [
4,
5]. Specifically, FMS, a prevalent, chronic, widespread pain disorder affecting mainly females, classified according to the 1990 American College of Rheumatology (ACR) criteria [
6], occurs in 20% to 32% of people with IBS, and, in turn, 32% to 70% of people with FMS also meet the criteria for IBS [
7‐
9]. It has been reported that the incidence of coeliac disease (CD) in patients with IBS is much higher than that expected to be found in the general population [
10,
11].
In spite of the fact that IBS, FMS and CD are three prevalent disorders in Spain, no study has been carried out to date to assess their association rate in this country. Therefore, with the aim of estimating the prevalence of CD masquerading as IBS and/or FMS, a specific protocol for active case-finding of CD was applied to an IBS/FMS population from Asturias (Cantabrian coast, northern Spain) and to a sex-, age- and size-matched cohort of IBS patients without FMS to establish comparisons.
Methods
During the 6-year period from 2007 through 2012, the Rome III criteria for IBS diagnosis [
3] and the 1990 ACR criteria for FMS classification [
6] were prospectively applied to 442 consecutive patients upon their first visit to an outpatient gastroenterology clinic at the Central University Hospital of Asturias (HUCA; Oviedo, Spain). Most patients were referred from the rheumatology and internal medicine departments of the same hospital for the study of a variety of long-standing gastrointestinal symptoms.
Diagnosis of IBS was based on a positive history of abdominal discomfort or pain associated with disturbed defecation (according to the Rome III criteria) in the absence of obvious alarm features such as rectal bleeding, iron deficiency anaemia, weight loss, fever, onset after age 40 years, family history of colon cancer, nocturnal symptoms and faecal soilage. When these symptoms are present, a more extensive evaluation must be carried out in the differential diagnosis. The evaluation should include testing for coeliac sprue, structural colon lesions such as polyps or cancer, parasites, endocrine disease, bacterial overgrowth or carbohydrate malabsorption.
Despite the fact that IBS is no longer a diagnosis of exclusion, since there is good evidence that a positive clinical diagnosis is reliable enough [
12], a battery of tests was performed in all patients before they were entered into the study, including a comprehensive medical history, a thorough physical examination and complete laboratory haematological and biochemical broad screening.
In patients who did not respond to the usual therapies, as well as in IBS cases in which other associated organic illnesses, a specific hydogren breath test was performed in order to exclude possible lactose intolerance or small-bowel bacterial overgrowth. Appropriate faecal cultures were done in some patients to exclude the presence of parasitic infections. Furthermore, in patients with persistent diarrhoea, a total colonoscopy was performed and random colonic biopsies were taken to rule out microscopic colitis.
An immunological faecal occult blood test (iFOBT) was done in patients over 50 years of age and in those with a positive familial history of colon cancer in first-degree relatives. If this iFOBT was positive, the study was completed with a total colonoscopy.
The acceptance criteria for IBS patients included (1) older than 18 and younger than 65 years of age, (2) meeting the Rome III criteria, (3) absence of any other associated organic gastrointestinal disease and (4) not meeting the 1990 ACR criteria for FMS. These selection criteria were also used to select IBS patients with associated FMS, except the last one, because by definition they had to fulfil the 1990 ACR criteria for FMS classification. The exclusion criteria were (1) incomplete or doubtful Rome III criteria for IBS cases and incomplete or doubtful 1990 ACR criteria for IBS/FMS cases, (2) any abnormal finding in the analytical screening or in the colonoscopy or colonic biopsies and (3) unwillingness to participate in the study.
Only 263 (59%) of 442 individuals were eligible to participate in the study, and 34 of the eligible people did not agree to sign a written consent form, thus excluding them from participation. Thus a total of 229 individuals agreed to take part in this study. Participants were then assigned to two groups: 104 comprising the IBS plus FMS group, who fulfilled both the Rome III criteria for IBS diagnosis and the 1990 ACR criteria for FMS classification, and 125 constituted the IBS group, who were unrelated age- and sex-matched patients from the same Asturias population who met the Rome III criteria for IBS diagnosis but did not have FMS widespread pain or skin tender points (TPs) and did not meet the 1990 ACR criteria.
These 229 participants were invited to participate on a voluntary basis after signing a specific informed consent form. The study was approved by the HUCA Research and Ethics Committee according to the principles included in the modified Declaration of Helsinki.
Outline of the study protocol
Initially, every selected participant underwent an updated medical history, a quality of life (QoL) battery of tests and a thorough physical examination.
Tender points
The FMS TPs were identified by digit pressure on the 18 locations recommended by the 1990 ACR criteria [
6]. This test was routinely carried out in the clinical setting. Digital palpation with the thumb pad was performed on the standard TP sites on a copy of the body figure (front and back) with location of all TPs that we found in each patient included in the clinical record. An approximate force of 4 kg, sufficient to blanch the nail bed of the thumb, was applied over each of these TPs. The TPs were palpated with constant palpation for about 4 seconds. Patients responded ‘yes’ or ‘no’ if they had any pain. When they responded ‘yes’ , the examiner asked them to rate their pain on a scale of 0 (no pain) to 10 (worst pain) and recorded each response on the body figure.
Physical, mental, psychological, social functioning and quality of life questionnaires
Each participant filled out the self-administered Spanish version form of the Fibromyalgia Impact Questionnaire (FIQ) [
13], the Stanford Health Assessment Questionnaire (HAQ) [
14] and the 36-Item Short Form Health Survey (SF-36) [
15].
The FIQ is a 10-item instrument that measures, in a range from 0 to 80 points, the physical functioning, work status and degree of depression, anxiety, sleep, pain, stiffness, fatigue and well-being, with scores of 0 to 39, 40 to 59 and 60 or greater evaluated as mild, moderate and severe FMS, respectively.
The 20-item disability scale portion of the HAQ measures, in a range from 0 to 3 points, the patient’s difficulty with activities of daily living and concomitant need for help and assistive devices, with the highest scores representing the maximum impairment (that is, 0 = able to do without any difficulty, 1 = some difficulty, 2 = much difficulty and 3 = unable to do).
The SF-36 is has two components: the Physical Component Summary (PCS) and the Mental Component Summary (MCS). Scores in the range from 0 to 100 punctuate each of these aspects, with the lowest scores indicating the poorest health status. No cutoff values on the SF-36 for classifying the impact of disease have been established yet. For guidance, the reported values for adults in the Spanish general population (expressed as mean and standard deviation) [
16] were PCS 73.0 (27.8) and MCS 74.4 (24.4).
Laboratory tests
Complete blood cell counts were performed with the CELL-DYN 3500 automated haematology analyser (Abbott Cientifica, Madrid, Spain). Coagulation panels were carried out with the IL ACL 3000 Coagulation Analyzer (Beckman Coulter, Brea, CA, USA). Analytical biochemistry tests were performed on a Roche Hitachi automated modular chemistry analyser (Roche Diagnostic Systems, Indianapolis, IN, USA) for SXA-phosphoprotein-binding domains using enzymatic or kinetic methods, including urea, glucose, total proteins, albumin, C-reactive protein, calcium, folate, vitamin B12, creatinine, creatine kinase, lipid profile, liver function tests (LFTs), iron metabolism tests, immunoglobulin G (IgG), IgA and IgM, rheumatoid factor, thyroid function test and urinalysis with microscopic examination of sediment.
Antinuclear antibodies (ANAs) and antithyroid peroxidase (anti-TPO) antibodies were measured in each participant, and in those individuals with altered LFTs antimitochondrial antibodies (AMAs) were also assessed. Determination of these biomarkers was performed by indirect immunofluorescence assay on the HEp-20-10 cell line according to the manufacturer’s instructions (EUROIMMUN, Lübeck, Germany). Anti-IgA tissue transglutaminase subtype 2 (tTG-2) was measured by using a commercially available enzyme-linked immunosorbent assay kit (Phadia Diagnostics, Uppsala, Sweden).
To assess the participants’ genetic susceptibility to CD, the major histocompatibility complex class II human leucocyte antigen (HLA) marker HLA-DQ2 was characterized by means of a polymerase chain reaction (PCR) with a commercially available kit (PROTRANS HLA Celiac Disease Domino System; PROTRANS, Hockenheim, Germany). Characterization of the HLA-DQ8 haplotype was not systematically assayed in this explorative study, but it was carried out in isolated HLA-DQ2-negative cases showing villous atrophy in duodenal biopsy.
Duodenal biopsy studies
An upper gastrointestinal endoscopy with at least four duodenal biopsies was performed in all patients following the usual methodology employed in our service for CD diagnosis [
17]. Samples were routinely stained with haematoxylin and eosin (H & E) and with anti-CD3 immunohistochemical monoclonal antibodies to count the number of intraepithelial lymphocytes (IELs) and in turn to quantify them per 100 epithelial cells. Samples were studied by two expert pathologists at HUCA and classified into the following types according to the histological classification for CD described by Marsh [
18] and later modified by Oberhüber
et al. [
19]: stage 0: histological normal duodenum; stage 1: increased IEL infiltration with a total count of 25% or greater; stage 2: crypt hyperplasia and diffuse chronic inflammatory infiltrate at the lamina propria; and stage 3: villous atrophy, which was subdivided into three categories: (a) mild, (b) moderate and (c) severe.
Helicobacter pylori was systematically investigated by taking endoscopic biopsies from the antrum (
n = 2) and the corpus of the stomach (
n = 2). Antibiotics, proton pump inhibitors (PPIs) and drugs containing bismuth and H2 blockers were stopped within the previous 2 weeks. One antral biopsy was immediately used for a quick urease test (Pronto Dry Kit; Pentland Medical Ltd, Edinburgh, UK). The rest of the samples were used for histopathological examination (that is, routine H & E staining, Giemsa staining and immunohistochemistry using polyclonal anti-
H. pylori antibody) and microbial cultures. Positive cases received triple therapy comprising 14-day treatment with PPI (standard dose twice daily), clarithromycin (500 mg twice daily) and amoxicillin (1,000 mg twice daily). After treatment was completed,
H. pylori eradication was confirmed 4 to 6 weeks later with a rapid urease breath test [
20,
21].
Statistical analysis
Descriptive statistics (means calculations, standard deviations and observed ranges) were used on continuous parameters. For qualitative variables, percentages were used. Kruskal–Wallis analysis of variance (ANOVA) contingency tables were analysed. If the continuous variables followed a normal distribution, Student’s t-test was used. Differences between groups were evaluated by ANOVA followed by post hoc analysis using Fisher’s test. The statistical calculations were performed using SPSS 15.0 software (SPSS Inc, Chicago, IL, USA), and P values less than 0.05 were considered significant.
Discussion
The most remarkable finding of this pilot study was our finding of seven cases with solid histological, laboratory and clinical evidence of CD in a group of FMS patients (mostly females) whose clinical symptoms started with gastrointestinal relapsing disorders mimicking IBS when they were in their 20s. They developed multisystem complaints about two decades later, with very negative effects in their health-related QoL. Remarkably, almost 30% of these CD-associated IBS and FMS patients had a family history of CD, and a significant number of them also had high serum titres of ANAs, anti-TPOs and AMAs. Interestingly enough, these seven patients were started on a gluten-free diet (GFD), which led to significant improvement in their digestive and systemic symptoms upon follow-up examinations.
A recent systematic review concluded that biopsy-proved CD in cases meeting the diagnostic criteria for IBS was more than fourfold higher than that in controls without IBS [
23]. Although further work is needed to definitively confirm that CD is a significant problem in terms of misdiagnosis in IBS, it has been estimated that testing IBS patients for CD can be cost-effective in clinical practice [
24].
In recent years, a possible causal link between IBS, FMS and multisymptom forms of adult CD has been suspected on the basis of some cases with overlapping symptoms involving gastrointestinal, musculoskeletal and other body systems that have had similar clinical features [
25,
26]. In addition, a few case reports describing isolated patients with FMS and CD simultaneously, all of whose symptoms dramatically resolved after the removal of gluten from their diet, have been reported [
27,
28].
Studies of the prevalence of CD in European patients with IBS are scant. Nevertheless, two separate prospective studies carried out in Sheffield, UK [
29], and in Krakow, Poland [
30], respectively, concluded that IBS patients were seven times more likely than matched controls to have biopsy-proven CD.
As previously stated, it is widely accepted that IBS and FMS are interrelated diseases, with IBS occurring in about 50% of FMS patients [
4,
7]. It is also well-known that the association of CD with IBS is not infrequent, because the prevalence of CD in IBS is four- to sevenfold higher than the prevalence of CD alone expected to be found in the general population [
23,
29,
30].
CD is a multisystem autoimmune disorder related to a permanent intolerance of gluten stored in wheat, rye, barley and other cereals. It affects 1% to 2% of individuals (mainly females) worldwide. CD patients generally are carriers of one of the two major histocompatibility complex class II HLA genotypes, HLA-DQ2 or HLA-DQ8. In these patients, gliadin peptides trigger an aberrant immune response that results in the production of tTG autoantibodies and an immune-mediated chronic inflammation of the small-bowel mucosa. This immune-mediated enteropathy is characterized by villous atrophy, intraepithelial lymphocytosis and crypt hyperplasia. CD clinical manifestations may appear at any age together with gastrointestinal and/or extraintestinal systemic symptoms, although some diagnoses can be made in asymptomatic individuals. Notably, adhering to a GFD results in complete clinical remission and full intestinal mucosa recovery in the vast majority of CD patients [
31,
32].
The available level of scientific evidence supporting the association of CD with FMS is mostly limited to a low-powered, cross-sectional study in which a 2% incidence of CD was detected among 50 American adolescents with FMS [
27], an American nationwide survey reported that FMS was the physician’s initial diagnosis in 9% of 134 patients later diagnosed with CD [
25]. Two case reports in the literature have described, respectively, a child [
27] and three women [
28] with comorbid FMS and CD, all of whose symptoms resolved very well once gluten was removed from their diet. In the present study, we found that 6.7% patients with the triad of IBS, FMS and CD showed remarkable symptom improvement when placed on a GFD.
FMS is a complex chronic pain syndrome affecting 1% to 3% of people worldwide. FMS often clusters in families and affects mainly females (85% or more) between their 20s and 50s, although the incidence rises with age, reaching 7% in women over 70 years of age. Clinically, FMS is characterized by widespread soft-tissue pain; generalized TPs; abnormal fatigue; sleep disturbance; and skin, gastrointestinal, urinary, cognitive and various other symptoms. FMS symptoms vary from person to person in both number and severity and tend to fluctuate in response to emotional or physical stress, lack of sleep, exertion, injuries, infections, menstruation and weather changes, among other stressors. Because no effective treatment for FMS control is currently available, its personal, familial, labour-related and social impacts are very negative, usually leading to excessive use of healthcare services [
6,
33,
34]. The pathogenesis of FMS remains elusive, although it is believed that genetic, immunologic and environmental factors contribute to its complex, multifactorial pathological process, and it is accepted that a combination of increased peripheral impulse input and increased central pain sensitivity, with aberrant pain facilitation and impaired inhibition, may be responsible for this disorder. The relative role of peripheral and central factors and ascending and descending pathways are not known, but probably all of the neural pathways, in variable proportions and depending of each case, may be involved in an interactive way [
35,
36]. No laboratory or imaging techniques for an accurate diagnosis of FM are currently available, thus necessitating an exclusion-based differential diagnosis to rule out other, similarly typified diseases after appropriate evaluation screening in patients fulfilling the 1990 ACR criteria for FMS [
6]. Interestingly, FMS is frequently associated with other rheumatic disorders, infections and systemic illnesses. For example, 30% of females with ankylosing spondylitis, 16% with SLE, 15% with rheumatoid arthritis, 24% with psoriatic arthritis, 11% with osteoarthritis, 9.2% with Behçet’s disease and up to 40% with joint hypermobility syndrome meet the ACR criteria for FMS. FMS has also been documented in women with hyperprolactinaemia (71%), hypothyroidism (34%), Crohn’s disease (26%), ulcerative colitis (11%), diabetes mellitus (17%) and endometriosis (6%). An increased prevalence of FMS has also been reported in patients with hepatitis C virus (57%) and HIV (29%), as well as following physical trauma in 22% of patients, mostly women, with neck injuries and in 1.7% patients with lower-extremity fractures [
33,
34,
36,
37]. FMS is frequently found to be associated with other central sensitivity syndromes, such as: post-traumatic stress disorder (57%), CFS (55%), multiple chemical sensitivities (55%), IBS (41%), tension, migraine and mixed headaches (26%), TMJ disorder (24%), bulbar vestibulitis syndrome (23%), Gulf War syndrome (18%) and interstitial cystitis (15%) [
33,
36].
A separate issue is the importance of the significant number of IBS/FMS patients with related CD in our series who exhibited an excessive prevalence of ANA, anti-TPO and AMA autoantibodies. Indeed, there are two populations included in the present study: one constituted 104 IBS/FMS patients and the other 125 patients with only IBS. The IBS/FMS group had more frequent positive ANAs than those with only IBS (28% vs. 3%; P < 0.001), and the same was true for anti-TPO antibodies (11% vs. 0%), AMAs (5% vs. 0%) and anti-tTG autoantibodies (mean serum levels = 0.4 vs. 4.7; P < 0.001). However, none of the ANA-positive participants met the 1982 ACR criteria necessary for the diagnosis of SLE, none with increased anti-TPO antibodies showed clinical manifestations of thyroid dysfunction and none of the AMA-positive individuals had clinical manifestations of liver disease.
Because FMS cannot be considered an autoimmune disease, despite the fact that it is common in patients with autoimmune disease and may be the source of many of the symptoms and much of the disability in these patients [
38], the findings of our study are more concordant with CD, an autoimmune disease often associated with a variety of autoimmune diseases, including type 1 diabetes mellitus, autoimmune thyroid disease, autoimmune adrenal disease, Sjögren syndrome, rheumatoid arthritis, SLE, primary biliary cirrhosis, primary sclerosing cholangitis, autoimmune cholangitis, autoimmune hepatitis and IBS [
32].
A secondary finding worthy of comment is the large number of patients (20 in the IBS-only group and 58 in the IBS/FMS group) with increased infiltration of IELs (above 25% of enterocytes) with normal villous architecture in the duodenal mucosa, a pathologic finding known as
lymphocytic duodenosis (LD). This condition, also named
lymphocytic duodenitis,
nonatrophic lymphocytic enteritis or (in the gluten-sensitivity context)
Marsh stage 1, is an unspecific diagnosis of gluten intolerance by itself. Therefore, such patients should not be diagnosed with CD solely on the basis of histological studies. A recent prospectively designed study demonstrated that only 16% of patients with LD will develop CD and that the IEL count even becomes normal on repeated biopsies in up to 76% of patients [
39]. In addition, different studies have reported a variety of LD and non-gluten-related sensitivity associations, such as drug intake (for example, nonsteroidal anti-inflammatory drugs (NSAIDs)), nongluten food protein intolerance (for example, cow’s milk, eggs, peanuts and soy), autoimmune disorders (for example, thyroiditis, type 1 diabetes mellitus, rheumatoid arthritis, psoriasis, multiple sclerosis and SLE), several inflammatory and/or infectious digestive disorders (for example, Crohn’s disease,
H. pylori, bacterial overgrowth, tropical sprue,
Giardia lamblia,
Cryptosporidium and viral infections), IgA deficiency and T-cell intestinal lymphoma [
40,
41].
About 80% of the IBS/FMS patients in our series were taking NSAIDs on an irregular basis, and more than 40% from both cohorts were infected with H. pylori at the beginning of the study. Therefore, it appears reasonable to attribute the high prevalence of NSAID chronic intake and H. pylori infection found in our series to these two known confounder causes of LD. However, it should be taken into account that ten (17%) of these IBS-LD/FMS patients who tested positive for HLA-DQ2 A1/B1 and were also reported to have first-degree relatives with CD, which are two laboratory and clinical features compatible with a CD diagnosis.
The comorbid triad of IBS, chronic fatigue and musculoskeletal pain is striking [
26], and the fact that our patients reported a long history of digestive complaints before the appearance of generalized soft-tissue pain, multiple TPs, weakness and other multisystemic symptoms resembling FMS is concordant with the increased prevalence of FMS reported by other series of women with different chronic processes within the gastrointestinal tract, such as IBS, Crohn’s disease and ulcerative colitis [
33,
36] and makes it feasible to hypothesize that a gluten-related autoimmune inflammatory process initiated within the gastrointestinal tract may contribute, in some gluten-sensitive patients with CD, to the well-documented central nervous system sensitivity (with aberrant pain facilitation and descending inhibitory pain impairment) responsible for FMS.
Despite the fact that our study included a relatively small number of nonrandomized patients comprising individuals referred to a gastroenterological clinic for symptoms indicative of a risk for CD, making it difficult to draw unequivocal conclusions, it appears evident that our findings may provide insightful contributions into the complex clinical picture of IBS/FMS CD-associated disorder, adding new clinical evidence in favour of the presence of a possible relationship between CD (and gluten sensitivity in general), IBS and some cases of FMS. Unfortunately, there are not any other available published case-finding studies of adult patients with FMS to establish comparisons.
Authors’ contributions
LR designed the study and collected the patients and the clinical information. IB conceived the study, participated in its design and coordination and helped to draft the manuscript. JB performed all the clinical health-related QoL tests and calculated and interpreted the information derived from them. FdS participated in the design of the study, made interesting suggestions about the study design and contributed to the analysis and interpretation of the results. All authors read and approved the final manuscript.