Phenotype of Crohn’s disease according to the Montreal classification in relation to dental health status

Abstract

Objective

The relationship between oral health and Crohn’s disease is uncertain. Previous studies have yielded contradictory results, reflecting perhaps the different phenotypes of the disease. The aim of the present study was to describe and analyse the dental status of a group of patients with Crohn’s disease (CD), considering the positions of the inflammatory loci and disease phenotype.

Methods

In total, 47 patients with Crohn’s disease (18 males and 30 females; mean age. 48.7 years; range, 23–61 years) were consecutively recruited to this study. Interviews and clinical examinations were performed to assess dental status, medication, smoking history, heredity of inflammatory bowel disease (IBD), duration of disease, oral mucosal manifestations of Crohn’s disease. Furthermore, data on subjective health assessments and family status, along with medical histories from the patients were obtained through questionnaires. The disease phenotypes were assessed and classified according to the Montreal classification. The data on oral health status were first correlated with the Montreal classifications of IBD, and, thereafter, all the collected data were included in a multivariate generalised linear model.

Results

The dental status of the patients was comparable to that of the Swedish average. No statistically significant associations were found between oral status and the different CD phenotypes. However, within the Montreal classification, there were significantly fewer teeth in those patients with perianal lesions than in those without such lesions, and there was a significant correlation between deeper pocket depth and problems with strictures and penetrations. No significant differences (p = .074) between the patients with CD (N = 47) and controls (N = 38) were found regarding the presence of oral mucosal lesions.

Conclusion

Dental health may be adversely affected in severe cases of CD whereas most of the remaining patients with CD appear to have a level of dental health that is comparable to that in the general population.

Keywords:

• Crohn’s disease
• inflammatory bowel disease
• dental status
• oral mucosa
• Montreal classification

Introduction

Crohn’s disease (CD), which is a chronic inflammatory bowel disease (IBD) [1,2], is often associated with significant morbidity related to recurring relapses of disease activity. While the aetiology of CD is not known [3], genetic and environmental factors and the microbiota have been implicated, with associated modulation of the regulation of the immune system [3,4].

There are several clinical phenotypes of CD and in 2000 a classification was introduced to classify these subgroups of the disease [5]. This classification, the Vienne classification, was later on replaced by the more detailed Montreal classification in 2008. The Montreal classification categorizes the subgroups according to factors, such as age, distribution of inflammation, and disease behaviour. The Montreal system of classification is currently the most recognized classification system [6].

CD can affect the entire GI tract, from the oral cavity to the anus [7]. The distal small intestine (ileum), colon, or both (ileo-colonic) [3], are most commonly affected by the disease [7]. In addition, some studies have indicated poorer dental health among patients with CD [8] or in subgroups of such patients [9]. The consequences of this are a greater need for improved oral hygiene and more intensive dental care, leading to higher dental care costs.

Oral lesions are present in 4–20% of patients with CD [8,10]. Children with CD appear to be more prone to develop oral mucosal lesions [11]. It has been suggested that recurrent aphthous stomatitis and non-specific ulcerations appear in the oral cavity several years before the intestinal manifestations of CD [12,13]. Oro-facial granulomatosis (OFG) is also linked to CD, especially in children [14,15], and may herald a more severe form of CD [16]. Furthermore, in other IBDs, such as ulcerative colitis (UC) and celiac disease, there may be manifestations in the oral mucosa [17,18]. However, in a recent prospective study conducted in Sweden, no statistically significant association between CD and oral lesions was found [19].

The dietary intake of people affected by CD often differs from standard diets in that it includes less protein and fat, as well as an increased amount of rapidly digestible carbohydrates, combined with an increased intake frequency [20]. The microflora associated with dental caries is mainly mutans streptococci, the populations of which may be boosted by the availability of these readily fermentable carbohydrates [21]. Thus, patients with CD may have a higher risk of dental caries [21]. However, the results of recent studies are ambiguous in relation to whether the self-reported dental health status of patients with CD differs from that of the general population [22,23].

Other factors, such as the quality and amount of the saliva, intake of fluoride, oral hygiene, systemic diseases, and specific genetic factors, also interact during the development of dental caries [24]. Some of these factors, e.g., saliva [9], can be affected by CD. The saliva is the primary line of defence [25,26] for preventing tooth decay, mainly due to its buffering effect on the oral cavity fluids [27]. The volume of secreted saliva is reduced by many types of medication. The reduced volume of saliva, together with a more cariogenic diet, may increase the risk of poor dental health in patients with CD [21].

Another dental disease that may be affected or correlated with CD is periodontitis. In a recent study of a Swedish group of patients with CD, the patients received more periodontal treatment than the controls [28]. Periodontitis, in similarity to CD, is an inflammatory disease, appearing in both chronic and acute forms, whereas dental caries is not.

Contrary to these hypotheses of a detrimental effect of CD on oral health, a Swedish prospective study found that early tooth loss and negligent oral hygiene seemed to protect against future CD [19]. The aim of this study was to evaluate the association between the disease phenotype of Crohn’s, as designated according to the Montreal classification, and the level of oral health.

Patients and methods

Patients

The medical records of 380 patients with CD admitted to the Department of Internal Medicine at Norra Älvsborgs Hospital were reviewed. All the patients were scheduled to undergo an examination of the gastro-intestinal (GI) tract. The inclusion criteria are listed in Table 1. Patients eligible for the study received a letter of invitation to participate in the study. A second letter was sent to those who did not respond to the first letter. Sixty-three (16%) patients with CD consented to participate in the study. Nine patients could not be reached subsequently for appointments, and five patients did not show up to the agreed appointments.

Table 1. Inclusion and exclusion criteria for participants in the study.

Inclusion

1: Treated for a diagnosis of Crohn’s Disease K50.0 & K50.9 (ICD 10)

2: Between 20 and 60 years of age at time of inclusion to the study

Exclusion

1: Autoimmune disease that causes dry mouth (e.g., Sjögrens syndrome, Mb Bechterew’s disease, Celiac disease)

2: Receiving immunosuppressive therapy (other than corticosteroids) - not to be included in the biopsy group

3: Anti-cholinergic medication

4: Radiotherapy to the head and neck region.

5: Haemophilia (not in the biopsy group)

The following exclusion criteria were used: edentulous patients, patients who had undergone radiation therapy for cancer of the head and neck region, patients who had received anti-cholinergic agents, patients with autoimmune diseases associated with dry mouth (e.g., Sjögren’s syndrome, Bechterew’s disease, and celiac disease). Furthermore, patients who were treated with immunosuppressive drugs (excepting corticosteroids) and patients with haemophilia were excluded from the biopsy group. Only one patient was excluded, an edentulous patient and a total of 47 (12%) patients with CD (mean age, 48.7 years; range 23–61 years) was finally audited due to the strict inclusion criteria of the study.

The 47 patients with CD were compared with a Swedish study conducted by Lapidus in 2006, which included a patient group of similar age (median age, 49.7 years) but with a higher proportion of women (ratio 1:2 vs. 1:1 total prevalence in Sweden) [26].

Data collection

A protocol was developed in collaboration with specialists in both medicine and odontology, including a dietician. At the first appointment, all the subjects were asked to complete a questionnaire. The patients with CD where then interviewed and examined. This was followed by oral and medical examinations by an oral and maxillofacial surgeon.

Questionnaire

The questionnaire included questions regarding the patients’ background and subjective health status, ongoing medication, smoking, and history of familial IBD. The answers were supplemented with data from the medical records of the patients, making it possible to categorise the patients according to the Montreal classification. This was done by a specialist in gastroenterology.

Collection of dental data

A specially trained dental hygienist together with an oral and maxillofacial surgeon examined all the participants. The radiographs were taken by the hygienist but reviewed by an oral and maxillofacial radiologist. Various tests of saliva samples were carried out, including buffering capacity, the volume of saliva, and detection of caries-associated bacteria.

Radiography

Two bilateral X-ray images (bite-wings) of the premolar region and two bilaterally in the molar region and one apical X-ray image of the premolar region of the lower jaw were obtained. The apical X-ray images were used to study the bone density of the patients with CD, and the results of these examinations have been presented previously [25]. The bite-wing images were used to diagnose dental caries and periodontitis. The X-radiographs were used subsequently to confirm the dental status of the premolar and molar regions. Missing teeth, restorations, periodontitis, and dental caries in the frontal region (regions 13–23 and 33–43) were only examined directly in the mouth. D + F% scores were calculated from the data. A filled surface with dental caries was registered as decayed in the DMFT scoring system.

Saliva

Saliva collection: A 5-min paraffin-stimulated saliva test was performed to obtain saliva specimens for analysis. The amount and rate of saliva flow were measured in ml/min. Using the same saliva sample, the buffering capacity was measured with buffering strips (GC Cooperation standard saliva kit), giving three possible outcomes: red (pH 5.0–5.8); yellow (pH 6.0–6.6); or green (pH 6.8–7.8). Since there was a risk of saliva contamination, all subjects were asked not to eat, drink or smoke at least 2 h before the sample collection.

Collection of oral mucosa data

Visual inspection of the oral mucosa was performed, and mucosal lesions were registered in the records. Photographs of the lesions were taken. In patients with oral mucosal lesions, biopsies were taken.

Furthermore, from the first 25 patients in our CD group (from list 1–25), a standard 5-mm stans mucosal biopsy was acquired from a determined area, buccally from tooth 38. The same procedure was repeated for a healthy control group (CG; matched for gender and age) of 25 patients referred to our clinic (Department of Oral & Maxillofacial Surgery, Norra Älvsborgs County Hospital, Sweden) for surgical extraction of tooth 38. The mucosal samples were divided into two. One half was embedded in paraffin and coloured with TX to visualise acanthosis, vacuolar degeneration, and the inflammatory status. The remaining half of each sample was frozen in liquid nitrogen (−198 °C) for immunohistological analyses.

Collection of medical data

The patients gave consent for their medical files to be examined. A specialist in internal medicine extracted the relevant information. Exclusions of subjects were made if the inclusion criteria were not met.

Finally, a blood sample was collected from each patient, to measure the levels of HB, TPK, albumin, CRP, SR, orosomucoid, plasma cobalamin, and plasma folate.

The study was approved by the Ethics Committee of the Medical Faculty of the University of Gothenburg, Sweden (Dnr. 693-08). All participants gave their informed consent before the start of the examination.

Statistical analysis

The Pearson’s Chi-square test and Fisher’s exact test were used for testing a hypothesis, including only nominal and ordinal variables. A t-test or ANOVA with Tukey’s post-hoc test was used to compare nominal or ordinal variables with scale variables. A multivariate generalised linear model (GLM) was conducted using the Montreal classifications as nominal variables. The IBM SPSS Statistics ver. 20.0 software was used. A p-value <.05 was considered statistically significant.

Results

In total, 48 consecutively treated patients (seven males and 31 females) consented to participate in the study. The patients were classified according to the Montreal classification (Table 2). The mean age ± SD was 48 ± 10 years and the mean age at diagnosis was 32 ± 12 years. Six of the patients (13%) were aged <16 years at diagnosis (A1) (all females), 29 (60%) were between 17 and 40 years of age at diagnosis (A2) (19 women and 10 men), and 13 (27%) were aged >40 years (A3) (6 women and 7 men) (Table 2).

Table 2. Description of data at baseline in the group clinico-pathological characteristics of the study participants.

	Age at diagnosis—Montreal classification
	A1	A2	A3	Total N	p-Value^a
Gender
Male	1	10	7	18	.09
Female	6	18	6	30
Age (mean ± SD) in years	42 ± 13	46 ± 9	57 ± 2	47
Disease duration (years)	23 ± 11	20 ± 11	9 ± 6	47
Family history of IBD (N)	2	8	1	11	.3
Smoker
No	3	12	5	20
Quit	2	7	4	15	.95
Yes	1	10	4	15
Number of teeth	26 ± 6	26 ± 6	26 ± 5	47	.89
BMI	25 ± 3	27 ± 3	26 ± 2	47	.56
Saliva buffer strip blue (N)	4	18	7	29	.77
Green (N)	2	5	4	11
Yellow (N)	0	4	1	5
Stimulated saliva
Secretion units	1.0 ± 0.6	1.4 ± 0.8	1.4 ± 0.9	47	.51
Location
L1 (L1,4)	0	2 (0)	1 (0)	3 (0)
L2 (L2,4)	0	5 (0)	2 (0)	7 (0)
L3 (L3,4)	6 (0)	21 (0)	10 (0)	37 (0)
L4	0	0	0	0	.92
Crohn’s disease behaviour (perianal involvement)
B1 (B1p)	2 (1)	12 (1)	5 (0)	19 (2)
B2 (B2p)	2 (0)	9 (1)	3 (0)	14 (1)
B3 (B3p)	0 (0)	1 (1)	2 (1)	3 (2)
B2 + B3 ((B2 + B3) + p)	2 (2)	6 (1)	3 (1)	11 (4)	.88
Patient treated with (currently using)
Steroids	2 (2)	8 (6)	5 (3)	15 (11)	.75
5-ASA	3 (3)	13 (13)	6 (6)	22 (22)	1.00
Immunosuppressive drugs	2 (2)	17 (8)	8 (5)	27 (15)	.54
Abdominal surgery	4	14	6	24	.71
Anti-diuretics	4 (4)	9 (9)	4 (4)	17 (17)	.13

A1; age <16 years; A2; age 17–40 years; A3; age >40 years.

Location of Crohn’s disease: L1, ileal; L2, colonic; L3, ileo-colonic; L4, upper GI tract disease.

Behaviour of Crohn’s disease: B1, non-stricturing; non-penetrating; B2, stricturing; B3, penetrating; p, perianal.

^aFisher’s exact test.

Medical data

We were able to extract data on the location of the disease for 47 of the patients. The most common location was ileo-colonic (38%), followed by the colon (29%), ileon (20%), rectum (19%), and other locations (2.5%). Apart from these locations, 15 (32%) of the patients had extra-intestinal lesions. Of the 47 patients, 13 (28%) had the disease in one location, 11 (23%) in two locations, 9 (19%) in three locations, 10 (21%) in four locations, and 4 (8%) in five locations.

Twenty-two of the patients (47%) had not undergone any surgery, 12 (26%) had undergone surgery once, eight patients (17%) had undergone surgery twice, four patients (8%) had undergone surgery three times, and one patient had undergone surgery 25 times.

Seven patients (15%) were not receiving any medication, 22 patients (47%) were treated with a cocktail of 5-ASA 17 patients (37%) were treated with anti-diuretics, 10 patients (22%) were treated with corticosteroids, 15 patients (32%) were treated with immunosuppressive medication, two patients (4%) were treated with biological agents, and 36 patients (80%) had different vitamin substitutions (not differentiated).

Oral mucosa results

Two patients (4%) in the CD group presented with oral mucosal lesions: one patient with reticular oral lichen planus and one patient with erythematous oral lichen planus. In both patients, the clinical diagnosis was biopsy-verified. None of the patients in the CD group displayed aphthous ulcers at the time of examination. In the oral mucosa CG, no oral mucosal lesions were registered. No significant difference was detected between the CD group and the CG (p = .495).

The biopsies from the buccal region of 38 subjects were graded into four classes (Table 2). The CD group showed almost the same numbers as the CG. In the CD group, no inflammation was found in nine patients, as compared with four in the control group. Inflammation grade 1 was detected in 17 patients in the CD group vs. 20 in the CG. Grade 2 was not present in the CD group and in two persons in the CG. Grade 3 was not detected in either group.

Vacuolar degeneration was negative in 22 samples from both the patients with CD and the CG. Grade 1 was found in four patients in the CD group and in four in the CG. Grades 2 and 3 were not found in either group.

Acanthosis was not found in five patients in the CD group and six patients in the CG. Among the patients with aconthosis, grade 1 was found among 10 in the CD group and 13 in the CG, grade 2 among nine in CD and six in the CG, and grade 3 among three in the CD group and three in the CG.

The immuno-histopathology results will be described in a forthcoming study.

Dental data

Three variables were used as measures of dental health and disease: the number of teeth, decayed and/or filled teeth as a percentage of all present teeth, and mean pocket depth (Table 3). As all of these variables are correlated with age and some with gender and smoking history, a final multivariate GLM was used to analyse these factors.

Table 3. Association between the different Montreal classes and the three dental variables.

Montreal class (N)	No. of teeth	Decayed plus filled %	Mean pocket depth
A1 (6)	25.8 ± 5.9	34.4 ± 31.5	2.18 ± 0.46
A2 (28)	26.5 ± 4.4	46.7 ± 27.7	2.19 ± 0.36
A3 (13)	25.9 ± 4.6	57.1 ± 23.3	2.66 ± 0.77
p-Value
ANOVA	.893	.231	.026
After controlling*	.703	.733	.378
L1 (3)	27.3 ± 2.1	47.0 ± 14.7	3.04 ± 1.20
L2 (7)	29.6 ± 3.0	42.1 ± 25.6	2.14 ± 0.30
L3 (37)	25.5 ± 4.7	49.2 ± 28.8	2.30 ± 0.49
p-Value
ANOVA	.085	.825	.047
After controlling*	.061	.754	<.001
B1 (19)	27.1 ± 2.6	43.0 ± 24.2	2.22 ± 0.62
B2 (14)	27.4 ± 2.8	45.0 ± 31.0	2.21 ± 0.25
B3 (3)	27.0 ± 7.8	62.5 ± 23.0	2.76 ± 0.65
B2 + 3 (11)	23.1 ± 6.8	56.4 ± 28.9	2.52 ± 0.60
p-Value
ANOVA	.070	.457	.195
After controlling*	.318	.830	.619
B non-p (39)	27.2 ± 3.8	44.1 ± 25.3	2.28 ± 0.55
B p (8)	21.8 ± 5.5	67.0 ± 30.8	2.49 ± 0.51
p-Value
ANOVA	.001	.029	.334
after controlling*	.007	.039	.750

Values shown are mean and SD for each measure. ANOVA and multivariate GLM were used to test if any classification was statistically significantly associated with any of the dental variables. Significant p-values are presented in bold letters.

*Multivariate GLM; independent variables are sex, age, smoking, and either Montreal class, and dependent variables are the number of teeth, D + F %, and mean pocket depth.

Of the patient group, 15 were smokers, 13 were former smokers, and 20 had never smoked, and smoking was correlated to mean pocket depth. Although both Montreal class A and Montreal class L were associated with mean pocket depth, when analysed with multivariate GLM only the association for the L class remained significant. The L1 class had a deeper pocket depth than the other two classes. The blood pressure group had fewer teeth and a higher proportion of decayed and filled teeth. However, when these factors were run through the multivariate GLM analysis only the higher proportion of decayed and filled teeth remained significant. No interaction was found regarding Montreal classes A and L, but there was an interaction between Montreal class B and smoking regarding the number of teeth, and one between Bp and gender regarding the proportion of D + F %.

When stimulated saliva secretion flow rate and buffer capacity were included as dependent variables the results retained statistical significance. However, when they were included as independent variables the statistical significance of the results regarding Bp disappeared.

Montreal classes B and L showed a statistically significant correlation with each other (Spearman’s rho 0.33, p = .02). All but one of those who had undergone surgery were classified as Montreal L3 (p < .002) and most of those who were classified as B2, B3, or both (25 of 28) underwent surgery (p < .001).

Table 2 lists some of the oral variables assessed. The average number of teeth was: in group B1, 27.4 (N = 17); in group B2, 27.6 (N = 13); in group B3, 31 (N = 1); and among those with both stricturing and penetrative features (B2 + B3), 27.3 (N = 7). Among those with perianal involvement, fewer teeth were present: B1p, 24.5 (N = 2); B2p, 24 (N = 1);0 B3p, 25 (N = 2); and B2 + B3p, 12.8 (N = 4). The difference between those groups with and without perianal involvement was statistically significant (p = .014), but not when considering the different distributions within the different groups (B1 to B2 + 3) (Table 4).

Table 4. Correlation between degree severity? of Crohn’s disease and the number of missing teeth.

Average number of teeth in each disease behavior group
	B1 (N)	B2 (N)	B3 (N)	B2 + B3 (N)	N
No perianal involvement	27.4 (17)	27.6(13)	31 (1)	27.3 (7)	38
Perianal involvement	24.5 (2)	24 (1)	25 (2)	12.8 (4)	9
n	19	14	3	11	47 p = .074^a

Univariate GLM: Montreal class B, p = .034; Montreal class p, p = .008. Interaction effect, p = .55.

Montreal classification in relation to Crohn’s disease behaviour: B1, non-stricturing; non-penetrating; B2, stricturing; B3, penetrating; p, perianal.

Shown are the average numbers of teeth for each group, with a maximum of 32 teeth.

^aFisher’s exact test.

Discussion

This study reveals an association between patients with anal fistulas (Bp) and a higher caries rate (percentage decayed plus filled and also a lower number of teeth. Furthermore, an association was also found between the location of the intestinal lesion and mean pocket depth. This is partly in accordance with the results from two studies conducted in Stockholm, Sweden, which found more pronounced symptoms among patients with the more severe disease [9] and with a more extensive dental treatment [25]. However, the associations are not uniform and it seems that some groups are affected, whereas others are not. This may explain the differences in results presented by other studies.

When salivary secretion rate and buffer capacity were included as independent variables within the GLM the significance of the associations between Bp and the dental variables disappeared. This suggests some influence of Bp on salivary conditions, which may explain, at least in part, the higher caries rate and greater loss of teeth. The effect may be a direct one or a mediated one (e.g., a consequence of the diet being changed, which may cause a deteriorated salivary condition).

An invitation to participate in our study was sent by ordinary mail to all patients with Crohn’s disease in the north-western part of Västra Götaland County. As this region is quite large and the examination was not coordinated with routine visits to the hospital, a low rate of participation (13%) was to be expected. This might have given a skewed result, as can be seen in the age and sex distributions. Participating patients were more likely to be those patients with a good general dental status and those that maintained their oral health throughout the course of the disease. The data from the CD group night have been more reliable if we had more participants, although the main focus was to have a varied study group and representation of almost all the Montreal classes. Therefore, the study is valid as an explorative study.

The study results are from a small area of Västra Götaland County in Sweden. It is not possible to conclude that the results reflect all of Västra Götaland County or even Sweden. There are no major cities in the study area, and it differs in various other aspects from both Västra Götaland County and the overall country of Sweden.

Specific lifestyle changes, including dietary adjustments, elemental diet, proper hydration, and cessation of smoking, may reduce symptoms. Smaller but more frequent meals are recommended. To manage symptoms, a balanced diet with adequate intake control is essential, and lower fat intake is also recommended. The natural fuel substitute for fat is readily fermentable carbohydrates, which give quick energy. However, as mentioned earlier, this is the recipe for dental caries.

Earlier studies have shown that many patients with CD have a low body mass index (BMI). However, our study group had a significantly (p < .02) higher BMI (26.6 ± 3.5) than the average Swedish person (25.4) [29]. This indicates that the investigated group is not affected severely by the disease, meaning that they can eat more and perhaps have a more varied diet. Another indication of a sample that has a well-treated disease is the absence of oral mucosal manifestations of the disease. Interestingly, 31% of the group still smoked despite information on the linkages between tobacco usage and CD.

Oral mucosal diseases were rarely seen in our adult study population. This is in line with data from Greenstein and co-workers, who reported mucosal lesions in 5% of patients with CD [17]. In contrast, Pittock and co-workers found that 48% of paediatric patients with CD had disease-specific abnormalities in the mouth [30]. It is possible CD does not occur in the oral cavity of adults to the same extent as in children. This study indicates that oral mucosal manifestations of CD are not a significant problem in adult patients with the disease.

We have not focused directly on the cause and effect of tooth loss, although we have tried to analyse the associations between CD using the Montreal classification and oral health. We plan to investigate the sample diet further to examine the association between diet and CD and dental health.

Conclusion

Our study shows a weak association between patients with CD classified according to the Montreal classification, and their oral health and disease variables. Patients with anal fistulas (Bp) exhibit slightly poorer oral health in the form of fewer teeth, whereas patients with the less-severe disease have an oral health status that resembles the general population.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.