Background
Microscopic colitis (MC) is an inflammatory disease of the colonic mucosa that predominantly affects elder women, with an average age of 65 years at diagnosis, and a three- to fourfold increased prevalence in women compared with men [
1]. The most common clinical presentation is chronic, non-bloody diarrhea with normal or close to normal endoscopic findings [
1]. MC can be divided into collagenous colitis (CC) and lymphocytic colitis (LC) [
2,
3]. The histological criteria for CC are a thickened subepithelial collagen layer (> 10 µm) in the extracellular matrix of the mucosa, epithelial damage, and presence of an inflammatory infiltrate in the lamina propria [
4]. The criteria for LC are > 20 intra-epithelial lymphocytes/100 enterocytes [
4].
The pathophysiology to MC is unknown. The most well-documented etiological finding of MC is that present smoking, and to a lesser extent also past smoking, show a clear association with MC [
1,
3,
5], and may be one explanation to the higher incidence of MC in Denmark compared with Sweden [
6]. MC has in several studies been found to be associated with celiac disease, autoimmune thyroid disease, Sjögren’s syndrome, diabetes mellitus, and several conditions of the skin and joints [
1]. The concomitant presence of celiac disease was 4.4% in MC [
7] and 12.9% in CC [
5]. The association of MC with celiac disease and autoimmunity was stronger at younger age in nationwide Swedish and Danish studies [
8,
9]. Medication may be another etiology to MC [
1,
10].
Familial occurrence of CC has been described [
11]. This rendered examination with genome-wide association studies (GWAS), which revealed associations between CC and human leucocyte antigen (HLA) variants, also associated with several autoimmune diseases [
12]. This may explain the observed comorbidity with autoimmunity in patients with CC and their first-degree relatives, and that CC truly represents a disease of immune-mediated nature [
12]. These findings were different from LC, which did not show any significant HLA associations in GWAS [
13].
Pathologists describe that a microscopic inflammation can be found throughout the gastrointestinal mucosa in celiac disease, and therefore, the connection between celiac disease and MC may in some cases represent one entity instead of two different entities [
14]. Since the colon mucosa lack villi, the colonic reaction to gluten intolerance could hypothetically instead lead to infiltration of lymphocytes and other inflammatory cells into the lamina propria. This is supported by highest association between celiac disease and LC [
9] and clinical improvement and healing of the LC after introduction of a gluten-free diet in 4 of 9 patients [
15]. However, both colonic mucosal alterations and intestinal villus atrophy may sometimes be resistant to dietary changes [
14,
15].
Since the hypothesis has been raised that MC may be the colonic reaction to gluten in celiac disease [
9,
14,
15], we wanted to retrospectively examine the association between the debut of celiac disease and MC in a female cohort. The aim of the present extended study was therefore to examine the association between MC and celiac disease, adjusting for smoking habits, when considering both the subtype of MC, gastrointestinal symptoms, and the clinical course of the disease.
Discussion
The main finding in the present study was the consolidation of previous literature on the association between LC and celiac disease [
9,
14,
15]. Furthermore, celiac disease seems to be more prevalent in patients with one episode of MC and past smokers.
MC is considered a chronic disease and is often treated for long time with corticosteroids. Nevertheless, we have previously described that MC may occur and be documented as one episode, although MC is classified as a chronic disease [
27]. Past smoking was associated with one episode of MC, and current smoking was associated with refractory MC, when all cases with celiac disease were excluded [
28]. The current findings that almost half of the patients only had had one single episode of MC, based on the patient reports and medical records, is in accordance with a large prospective European cohort study which demonstrated a chronic course in only 49% of cases, independently of subtype [
29]. In contrast, another study described that 63% of LC only have a single episode of the disease, compared to 30% exhibiting a chronic recurrent course [
15]. One episode of MC was much more uncommon in the CC group (2%) [
15]. The follow-up time from diagnosis was several years both in the former and present studies, which strengthens the findings that the disease may present only as a single episode [
15,
27,
29]. There were no clinical differences between CC and LC, except that LC was associated with celiac disease and CC with treatment of corticosteroids. Therefore, the most important is possibly not to stress the histopathological findings, but to follow the disease course.
The hypothesis that MC may be the colonic reaction to gluten in celiac disease [
9,
14,
15] was supported since 9 out of 11 patients with celiac disease only had had one episode of MC, which in half of the cases had debuted simultaneously. To confirm this hypothesis, a prospective study must be performed with duodenal and colonic biopsies before and after exclusion of gluten. However, we suggest greater caution before the clinical diagnosis of chronic MC is set [
27]. Relapses of the disease must be observed during the follow-up before the definitive diagnosis is set, in similarity to other IBD diseases [
22,
23,
30,
31], and secondary forms of mucosal lymphocytic infiltrates must be excluded, e.g., celiac disease, infections, and drugs [
14]. Duodenal intraepithelial lymphocytosis is a type of inflammation associated with MC [
7], although this is a frequent finding after a transient gastrointestinal infection and in gluten sensitivity [
14,
32]. Furthermore, the subtype of MC may shift over time, from one entity to the other, and from MC to other IBDs [
33,
34].
The present association between MC and celiac disease is well described. An American case–control study found a prevalence of celiac disease in 4.4% of MC patients, and MC was diagnosed in 10.0% of patients with celiac disease [
7]. In a histopathological study of celiac disease, 19.0% of cases also had LC at diagnosis [
35], and as many as 30% of patients with celiac disease had MC according to a review of celiac disease [
36]. The association was strongest between LC and celiac disease in both American and Swedish cohort studies [
7,
9], in similarity with our MC cohort, but in opposite to a Danish cohort [
8]. The prevalence of celiac disease in a Scandinavian CC cohort was 12.9% [
5], in comparison to the prevalence of about 2% in the Swedish general population [
37]. The differences of prevalence between studies may depend on assorted designs. The Scandinavian study [
5] was performed in a University Hospital setting, which means an accumulation of selected cases, whereas the American study was performed in all kinds of health care centers [
7]. In a systematic review with meta-analysis, the prevalence of MC in refractory celiac disease was 4.5%, and the prevalence of celiac disease in refractory MC was 6.7% [
38]. Comparable results were found in a second review [
39]. The prevalence of celiac disease was eight times higher in refractory MC compared to controls [
38], which underlines the importance of excluding gluten intolerance in MC patients who do not respond to conventional therapy. Also, patients with celiac disease and persistent symptoms must be examined to exclude MC [
40]. Analysis of anti-tTG in the current study excluded celiac disease in the cases with refractory MC. The question must be raised whether MC and celiac disease are two different entities when occurring in the same patient, or whether the MC in some circumstances should be considered as a component of the celiac disease with lymphocytic invasion of the mucosa [
14].
The celiac disease diagnosis was set a few years prior to the CC diagnosis in most cases [
5]. In another Scandinavian study of LC, 8 of 17 cases with celiac disease were diagnosed before the LC diagnosis, whereas the diagnoses were set at the same time in 9 of 17 cases [
15]. In an American study, 64% were diagnosed with celiac disease before MC, while 25% were diagnosed with both disease in the same month, with 75% diagnosed as LC [
41]. Those with LC had more severe villus atrophy than those with CC [
41]. The risk of MC diagnosis was far higher during the first year after the celiac disease diagnosis than after 10 years in a nationwide Swedish cohort [
9]. When MC and celiac disease are diagnosed in the same patient, the age is in general 10 years lower compared with the age in patients with pure MC [
9]. Treatment with gluten withdrawal may heal both the villus atrophy and the colonic alterations [
15]. This raises the hypothesis whether gluten-free diet would be a treatment option in MC, especially in LC. At least, diagnosis of MC should always be accompanied with exclusion of celiac disease with antibody screening and/or gastroscopy and duodenal biopsy [
3]. Diagnosis of MC some years after celiac disease may depend on poor compliance to the diet. A higher risk of MC development has been related to an untreated celiac disease or suboptimal adherence to gluten-free diet [
9]. The dietary compliance was not examined in the current cohort. This is an important issue, since patients diagnosed with celiac disease through screening programs in adults is not associated with worse adherence to the gluten-free diet and might protect from complications such as MC or metabolic bone disease [
42].
Constipation and bloating are associated symptoms. The higher degree of these symptoms in MC with only one episode compared to refractory MC, suggest that MC with one episode is another disease with other etiologies than refractory MC. Constipation in MC may depend on varying factors such as comorbidity, drug treatments, use of antidiarrheals, physical inactivity, or difficulties to empty the bowel due to pelvic floor dysfunction. Since IBS-like symptoms are that common in MC, it may be difficult to evaluate disease activity and correct diagnosis, without mucosal biopsies [
19,
27].
Several patients were on continuous treatment with corticosteroids, although they only had had one relapse of the disease and normal biopsies in subsequent examinations. This may depend on that the physician believes that the treatment may prevent relapses. The high prevalence of MC with one episode stresses the need for restrictions in the treatment. Corticosteroid use may lead to dependence and misuse [
43], leading to a challenge to quit corticosteroid treatment. Chronic use of corticosteroid is related to side effects, e.g., increased rates of infection, adrenal suppression, osteoporosis, diabetes, cardiovascular events, mood disorders, and all-cause mortality [
43].
It has previously been described that not only the disease of MC is more common in smokers; the disease also debuts earlier in life in smokers compared with non-smokers [
44,
45], reflected by the longer disease duration in smokers in the current study. The higher degree of working in present smokers probably reflects the younger age in that group. IBS-like symptoms and impaired psychological well-being were most frequently found in present smokers. This is in accordance with findings in systematic reviews and in population-based studies, which have described IBS to be associated with smoking [
46,
47]. In the former study describing association between former smoking and MC with one episode in comparison to current smoking and refractory MC, all cases with celiac disease were excluded not to interact with the MC diagnosis when studying smoking and alcohol habits regarding MC [
28]. When including celiac disease in the present study, we found high prevalence of past smokers in the patients with celiac disease and MC with one episode. Smoking has no association with celiac disease [
48]. Accordingly, past smoking was the most common smoking habit in the current subgroup of MC patients with one episode of the disease and in those with celiac disease. As described previously in US women [
49], current smoking appeared to be stronger associated with CC than with LC, also found in a meta-analysis [
50]. Even if the subject has ceased smoking, the effect of chronic smoking may have affected the mucosa and promoted disease development [
49,
51]. Smoking leads to inhalation of several different compounds, with varying effect on cellular and tissue functions of the intestinal mucosa such as the mucosal barrier, gut microbiota, immune system, and microvasculature [
52]. Further, it may be difficult to remember exactly when the symptoms debuted, especially if the symptoms are starting gradually. Snuff and nicotine chewing gums were of no major importance, since the uses were rather uncommon.
To use the histopathological diagnosis as a clinical diagnosis, without considering secondary causes or one episode of LC [
14], will be misleading for the prevalence of MC. Several associations may be found between MC and concomitant diseases and drug only due to the higher age of the patient group [
9,
53]. The present, liberal attitude to diagnosis of MC after one single episode, weakens the trustworthiness of research and may obscure the true disease through inclusion of several heterogenous patient cohorts. The tendency of over diagnosing symptoms must be questioned, since there is a danger of aggravating people’s perception of their symptoms and over prescription of drugs, thus, increasing the cost to the society of health care. Further, to receive a diagnosis of a chronic life-long disease may be a stigma for the individual and weaken the possibilities to get life insurances. Debut of a new disease, e.g., malignancy, may be neglected since the symptoms may be referred to MC and new examinations leading to a correct diagnosis delayed.
The novelty of the current study is the classification into one episode of MC or refractory MC, considering smoking as a confounder, when calculating the association between celiac disease and MC. There are several limitations in the present study. First, there is no adjustment for drug treatments and no control material of healthy individuals. However, a control population and considerations of drug treatments have been used in other studies [
1,
3,
28]. Only women were included, since the prevalence of MC in men was very uncommon in the area, and thus, no evaluations could be performed in male patients. The retrospectively collected data is another limitation. To verify the hypothesis that MC may be the colonic reaction to gluten in celiac disease, a prospective study of consecutively diagnosed patients with celiac disease must be performed with careful biopsy sampling from both the small and large intestine at the same time points, and sample collection both before and after introduction of gluten-free diets.
In conclusion, a great deal of patients with MC has a single episode of the disease and should not be assigned to a chronic MC disease, as has previously been described in a large prospective study [
18]. The first line of treatment of MC should be exclusion of celiac disease as previously proposed in European guidelines [
3]. The close association between MC and celiac disease must be recognized in cases with symptoms refractory to conventional treatment of either disease. The accurate nomenclature of diagnoses when both LC and celiac disease are diagnosed simultaneously should be further discussed, to define whether it is the same or two different entities, especially when there is only one episode of LC.