Re:베체트 증상을 Trigger하는 요인에 대한 탐구논문

[문형철]

BEYOND REASON

신이시여 감사합니다

Triggers of OU attacks

The responses to the questions exploring the general OU attack triggering factors are in Table 3 and Figure 1. In the open question (question 14), 57 patients (70%) reported ≥1 general factor as a trigger of OU attacks. The complete list of 113 responses given is in Table 3. The most commonly cited factors were stress (32 patients [40%]) and fatigue (28 patients [35%]), with 38 patients (47%) citing any of these 2 factors, and food (28 patients [35%]).

Figure 1

Open in figure viewerPowerPoint

Role of A, predefined general situations, and B, physicochemical food characteristics as oral ulcer triggers. The data are based on self‐reported information provided by 81 patients with Behçet's disease. Items are arranged top‐to‐bottom according to their order of appearance in the questionnaire. For the 6 general situations, “any of all” was calculated for patients declaring ≥1 as “surely” or otherwise ≥1 as “possibly” involved. The values in parentheses are the percentage responses (totals may not equal 100% because of rounding).

Figure 2

Open in figure viewerPowerPoint

Percentages of specific foods identified as oral ulcer triggers. The data were provided by 81 patients with Behçet's disease for 24 preselected foods. The food items are ordered by decreasing frequency.

Dietary and Nondietary Triggers of Oral Ulcer Recurrences in Behçet's Disease

Geoffroy Volle 

 

Jean‐Baptiste Fraison 

 

Delphine Gobert 

 

Tiphaine Goulenok 

 

Robin Dhote 

 

Olivier Fain 

 

Solange Gonzalez‐Chiappe 

 

François Lhote 

 

Thomas Papo 

 

Aurélie Thuillier 

 

Sophie Rivière 

 

Alfred Mahr

First published: 18 November 2016

 

https://doi.org/10.1002/acr.23155

 

Citations: 6

Drs. Volle and Fraison contributed equally to this work.

SECTIONS

PDF

TOOLS

 

SHARE

Abstract

Objective

The nature and impact of food and other external triggers in recurrences of Behçet's disease (BD)‐related oral ulcers (OUs) remain unknown. This survey investigated dietary and nondietary triggers of BD‐related OU recurrences.

Methods

Patients with BD who were followed in 7 French hospital departments completed a self‐administered patient questionnaire. General and specific dietary triggering factors were sought in open questions. The questionnaire also included closed questions, notably to eval‎uate the effect of 6 general triggering situations and 24 selected foods. The results were expressed as number (percentage) of positive responses.

Results

Among the 101 questionnaires distributed, 81 were usable. Among the 81 patients, 96% fulfilled the International Criteria for Behçet's Disease classification criteria, and 53% qualified their OU recurrences during the previous 12 months as very discomforting or discomforting. For the 6 general situations suggested, 50 patients (62%) declared ≥1 as a “sure” trigger of OU recurrences. In both open and closed questions, the most frequent triggers were fatigue/stress (37–47% of patients) and food (32–35%). Among the 24 suggested foods, nuts (48%), pineapple (42%), peanuts (32%), Emmental cheese (30%), almonds (23%), lemons (22%), and other cheeses (21%) were the most frequently reported. The corresponding open question gave consistent findings but with lower frequencies.

Conclusion

Most patients can identify triggers of recurring BD‐related OUs, with fatigue/stress and food representing the most frequent triggers. The management of OU must consider such external factors. The histamine‐rich or ‐liberating properties of the commonly cited OU‐triggering foods suggest a hyperreactivity mechanism.

INTRODUCTION

Behçet's disease (BD) is a chronic multisystem disease mainly affecting young and middle‐aged adults originating from an area that spans from Asia to the Mediterranean Sea. This illness may present highly heterogeneous symptoms involving the mucosa, eye, skin, vessels, joints, gut, and central nervous system. BD risk is closely linked with HLA‐B51, but the etiology also likely includes yet unidentified, nongenetic factors 1, 2.

Recurrent oral ulcers (OUs) are a virtually constant feature of BD and usually precede the diagnosis of BD by many years. Even during periods of otherwise quiescent symptoms, OUs tend to recur and negatively affect the quality of life 3, 4. Clinically, BD‐related OUs are similar to those seen with recurrent aphthous stomatitis (RAS), although they tend to recur more often 5, 6, and BD and RAS may be part of the same disease spectrum. OUs represent a clinical cornerstone of BD, but their underlying etiopathogenesis is elusive. Histopathologic and immunologic studies of OUs in BD and RAS show an inflammatory infiltrate mainly composed of lymphocytes, monocytes, and neutrophils 7.

Exogenous factors can precipitate OU episodes. Data from RAS suggest diet, infection, stress or anxiety, menstrual cycles, or trauma as triggers of OU recurrences 8, 9, but these observations rely on scarce and sometimes conflicting findings. For BD‐related OUs, no such data are yet available, and the proportion of patients able to identify a specific trigger of OU recurrences and the precise nature of these triggers are unknown. A better insight into this question may help enhance our understanding of the natural history and pathogenic mechanisms of BD‐related OUs.

We used a patient survey to investigate the OU‐triggering factors in BD and their relative contributions to the condition. Hypothesizing that the investigation of dietary factors might be particularly revealing in elucidating the pathogenesis, we paid special attention to foods as BD‐related OU triggers.

Box 1. Significance & Innovations

• Two thirds of patients with Behçet's disease (BD) report external triggers for their oral ulcer (OU) recurrences, particularly stress/fatigue and specific food items.
• These results generate a unifying hypothesis of OU as a hyperreactivity response to various dietary and nondietary triggers, in which histamine and mast cell degranulation or activation play a pivotal role.
• These specific foods and the lesser importance of mechanical factors, infection, and menstruation as BD‐related OU precipitators warrant further investigation in other areas and populations.

PATIENTS AND METHODS

Patient selection and ethical considerations

The study investigated patients with BD followed by 7 internal medicine departments in French academic or teaching hospitals. Patients were approached during clinic visits or by telephone and were briefly informed about the study's purpose and implications. Selection criteria included a diagnosis of BD with a history of recurrent OUs. Because BD‐related OU activity decreases with aging 10, and in light of a small subset of patients with BD who do not have OUs 2, the site investigators were told not to include patients who were likely uninformative because of a nonsignificant or long‐ago history of OUs. An institutional review board approved the study protocol (no. 00003835).

Patient self‐administered questionnaire

Patients who agreed to participate were provided with the study material that included a structured questionnaire, a 1‐page information letter explaining the context and purpose of the study, and a stamped return envelope. The questionnaire was designed to be self‐completed, but assistance was offered through a dedicated cell‐phone helpline. In rare circumstances such as illiteracy or poor eyesight, the questionnaire could be completed during a face‐to‐face or telephone interview. The patients were instructed to return the questionnaire within 4 weeks, and nonrespondents received a telephone call. Ambiguous or missing information in questionnaires was clarified or completed by contacting patients unless they specifically declined to provide further information.

The questionnaires (available from the corresponding author upon request) were printed as full‐color booklets consisting of 23 sets of questions divided into 5 sections. As detailed here, the questions included both categorical questions with a dichotomous or multiple‐choice format and qualitative open questions. The questionnaire included illustrations and photographs to enhance the clarity and to increase compliance by making it more user‐friendly. Before the launch of the survey, the questionnaire was pilot‐tested with 3 patients with BD to ensure that it was legible and comprehensible.

General demographic data (section 1)

Questions 1–7 asked the respondents to provide their name, sex, and country of birth, the countries of birth of their grandparents, and details about dental health, dietary habits, and smoking history. Dietary habits were asked as multiple‐choice questions that included 6 prespecified responses and a free‐text other option.

OU history (section 2)

Question 8 asked whether respondents had OUs, and if so, they were asked to provide their age at OU onset (question 9), lifetime total number of episodes (stratified into 5 categories: 1–5, 6–20, 21–50, 51–100 and >100 episodes) (question 10) and the impact of OUs during the last 12 months (4 categories: very discomforting, discomforting, not discomforting, and absent) (question 11). Questions 12 and 13 collected information about the medications received to treat BD in general and those additionally taken to treat OU attacks.

General OU triggering factors (section 3)

An open question (question 14) asked about any general factor identified by patients as triggering OU episodes. The following multiple‐choice question (question 15) specifically asked about the role of 6 general situations as OU triggering factors (asked in the following order): stress or fatigue, dental care, tooth brushing, menstruation, intercurrent infection, and diet. The 4 response options were “Yes, I am sure,” “Yes, that's possible,” “No, that's highly unlikely,” and “I don't know.”

Dietary OU‐triggering factors (section 4)

Patients who declared food as a sure or possible OU‐triggering factor in question 15 were asked to declare up to 4 food items they recognized as OU triggers in an open question (question 16). For each cited food, conditional or subordinate questions asked if the food's ingestion triggered OUs “each time” or “only sometimes” and if the food was “totally eliminated” from the diet or was “still taken occasionally” or “taken without restriction.” Question 17 showed a list of 24 food items we had selected from the literature 8, 11-13 or from personal clinical experience that were potential OU‐triggering factors or products belonging to the same food groups. Patients checked the items recognized as OU‐triggering factors and could check multiple items or a box labeled “none of the listed food items.” Question 18 asked for the role of 9 physical and chemical characteristics of foods as triggers of OUs (categorized and ordered as follows: salty, sweet, bitter, sour, spicy, astringent, hard, hot, and cold), and question 19 investigated whether fast eating affected OU occurrence. For both questions, response options used the same 4‐level scale as for question 15.

Comments and contact information (section 5)

The remaining questions provided space for other comments (question 20) and asked whether the respondent would agree to be contacted again in the context of the study (question 21), and whether and how they wanted to be informed about the study's main results (questions 22 and 23).

Clinical information

The site investigators provided clinical data through review of the patients’ medical records. A study‐specific case report form was used to gather information on the main BD disease manifestations and treatments given cumulatively from the time of diagnosis, and on HLA‐B51 status. The form was also used to collect data to eval‎uate whether the diagnosis satisfied the International Criteria for Behçet's Disease classification criteria 14. Organ or system involvements were defined by clinical features (e.g., mucocutaneous disease or arthralgia/arthritis) or specific examinations (e.g., uveitis). Gastrointestinal disease was defined by endoscopic anomalies consistent with inflammatory bowel disease. Central nervous system involvement was defined by consistent anomalies on brain or spinal cord imaging, cerebrospinal fluid analysis, and/or electroencephalography 15.

Statistical analysis

The research was qualitative and did not test an a priori hypothesis. We initially set an inclusion target of 100 to 150 patients and halted inclusions after meeting the minimum enrollment threshold over a 1‐year period because the proportion of patients declaring food‐triggering factors was much higher than anticipated. The analyses were predefined in a study protocol and the results are expressed as number (%) and means ± SDs. In some cases, responses to open questions were pooled if they reflected similar contents. Subgroup analyses involved patients with the most active BD‐related OUs (>50 lifetime episodes of self‐reported BD‐related OU attacks or declaring that the OUs were very discomforting or discomforting over the previous 12 months).

RESULTS

Selection and characteristics of patients

A total of 101 questionnaires were distributed from January 2015 to January 2016, and 88 were returned. Five questionnaires were discarded because of a large number of missing responses. An additional 2 questionnaires were excluded because the patients declared having no OUs. Eventually, we retained the data for 81 patients with complete questionnaires. The number of patients enrolled per center ranged from 5 to 29.

Table 1 summarizes the main clinical characteristics of the 81 patients. In total, 78 patients (96%) fulfilled the international criteria for BD 14. The 3 patients not satisfying these criteria had recurrent OUs with cutaneous manifestations and 2 also had arthralgia. For the 31 women, mean ± SD age and number (%) <45 years was 21 ± 11 years and 30 (97%) at OU onset, and 41 ± 12 years and 21 (68%) at the time of the study. Ethnic backgrounds (defined as the self‐reported birthplaces of ≥3 grandparents) were non‐French for 61 patients (75%), mainly from Morocco, Algeria, or Tunisia. Self‐reported dietary habits included, in decreasing frequency, French and European‐style food (67 patients [83%]), northern African (32 [40%]), Asian (23 [28%]), vegetarian (14 [17%]), sub‐Saharan African (7 [9%]), South American (2 [2%]), and other food (7 [7%]). Dental health was self‐rated very good by 12 patients (15%), good by 38 patients (47%), poor by 23 patients (28%), and very poor by 8 patients (10%). There were 23 current smokers (28%), 21 past smokers (26%), and 37 never smokers (46%).

Table 1. Characteristics of 81 patients with Behçet's disease included in the main analyses of the studya

Variable	Values
No. of patients	81 (100)
Age, mean ± SD years	41 ± 11
Male	50 (62)
Geographic backgroundb
French	20 (25)
Non‐French European	7 (9)
Maghreb (Morocco, Algeria, Tunisia)	29 (36)
Sub‐Saharan Africa	6 (7)
Asia	13 (16)
Other	4 (5)
Mixed	2 (2)
Followup, mean ± SD months	67 ± 63
Cumulative disease manifestationsc
Genital ulcer	62 (77)
Ocular involvement	38 (47)
Anterior uveitis (only)	14 (17)
Posterior uveitis, retinal vasculitis, or panuveitis	24 (30)
Cutaneous involvement	63 (78)
Erythema nodosum	25 (31)
Folliculitis or papular lesions	52 (64)
Arthralgia/arthritis	50 (62)
CNS involvement	15 (19)
Parenchymal disease	12 (15)
Aseptic meningitis	2 (2)
Vascular involvement	34 (42)
Superficial thrombophlebitis	17 (21)
Deep venous thrombosis	19 (23)
Arterial disease	9 (11)
Gastrointestinal involvement	7 (9)
ICBD criteria fulfilled	78 (96)
HLA‐B51d	20 (48)
Cumulative systemic therapy
Colchicine	81 (100)
Glucocorticoids (≥3 months)	55 (68)
Thalidomide	12 (15)
Azathioprine	20 (25)
Cyclophosphamide	10 (12)
Methotrexate	11 (14)
Anti‐TNF monoclonal antibodies	20 (25)
Anticoagulants (≥3 months)	24 (30)

• a Values are the number (%) unless indicated otherwise. CNS = central nervous system; ICBD = International Criteria for Behçet's Disease; TNF = tumor necrosis factor.
• b Defined as the place of birth of ≥3 grandparents (self‐reported in the questionnaire).
• c For definitions, see Patients and Methods.
• d Data were available for 42 patients.

History of OU attacks

Self‐reported information on the history of OUs and specific OU treatments are in Table 2. Mean ± SD age at OU onset was 23 ± 11.2 years, 53 (65%) reported >50 lifetime OU attacks, and 42 (53%, 1 missing response) qualified their OU recurrences during the previous 12 months as very discomforting or discomforting. Overall, 45 patients (56%) reported the use of ≥1 local therapy to treat OU attacks.

Table 2. Self‐reported characteristics of oral ulcer (OU) history in 81 patients with Behçet's diseasea

Variable	Values
Duration since start of recurrent OU history, mean ± SD years	18.8 ± 12.0
Lifelong episodes
1–5	5 (6)
6–20	9 (11)
21–50	14 (17)
51–100	18 (22)
>100	35 (43)
Severity of OUs during the previous 12 monthsb
None	16 (20)
Little discomforting	22 (28)
Discomforting	16 (20)
Very discomforting	26 (33)
Use of local therapies for OU attacks	45 (56)
Unspecified oral pastes or mouth rinses	20 (25)
Topical anesthetics, antiseptics, or analgesics	19 (23)
Topical aspirin or salicylates	9 (11)
Topical corticosteroids	3 (4)
Others	2 (2)

• a Values are the number (%) unless indicated otherwise.
• b Data were available for 80 patients.

Triggers of OU attacks

The responses to the questions exploring the general OU attack triggering factors are in Table 3 and Figure 1. In the open question (question 14), 57 patients (70%) reported ≥1 general factor as a trigger of OU attacks. The complete list of 113 responses given is in Table 3. The most commonly cited factors were stress (32 patients [40%]) and fatigue (28 patients [35%]), with 38 patients (47%) citing any of these 2 factors, and food (28 patients [35%]).

Figure 1

Open in figure viewerPowerPoint

Role of A, predefined general situations, and B, physicochemical food characteristics as oral ulcer triggers. The data are based on self‐reported information provided by 81 patients with Behçet's disease. Items are arranged top‐to‐bottom according to their order of appearance in the questionnaire. For the 6 general situations, “any of all” was calculated for patients declaring ≥1 as “surely” or otherwise ≥1 as “possibly” involved. The values in parentheses are the percentage responses (totals may not equal 100% because of rounding).

Table 3. Responses to open question for 57 of 81 patients with Behçet's disease who reported at least 1 general factor as triggering their oral ulcer attacksa

Triggering event	No. of patients (%)	Comments (no. of responses)
Stress	32 (40)	Stress and/or fatigue 38 Stress and fatigue 22
Fatigue	28 (35)	–
Food	28 (35)	–
Infection	5 (6)	Includes fever 3
Climate	4 (5)	–
Menstrual cycles	3 (10)	–
Tooth brushing	3 (4)	Includes toothpaste 2
Trauma and bites	2 (2)	–
Dental care	2 (2)	–
Medications	1 (1)	–
Miscellaneousb	5 (6)	–

• a Wordings used are those originally provided. Responses were grouped if they were considered to reflect similar contents (and detailed in the comments).
• b Includes dental pain 1, withdrawal of colchicine 1, orodental hygiene 1, alcohol intake 1, and uveitis 1.

For the 6 OU‐triggering situations suggested in the closed question (question 15), 50 patients (62%) reported ≥1 as “surely” involved, most commonly stress/fatigue (30 patients [37%]) and food (26 [32%]). The detailed breakdown of responses is shown in Figure 1. For the 12 women (38%) reporting menstrual cycles as a sure or possible triggering factor, responses to the subordinate question indicated that OU occurred before and/or during the menses (3 missing responses). Sensitivity analysis of patients with the most severe OUs gave results comparable to those for all study participants (see Supplementary Figure 1, available on the Arthritis Care & Research web site at http://onlinelibrary.wiley.com/doi/10.1002/acr.23155/abstract).

The specific dietary factors identified as OU triggers are summarized in Table 4 and Figure 2. In the open question (question 16), 53 patients (65%) reported ≥1 dietary triggering factor. The most commonly cited foods were nuts (22%), pineapple (14%), various citrus fruits (14%), other cheeses (excluding Emmental cheese) (14%), Emmental cheese (12%), strawberries (10%), and tomatoes (9%) (Table 4). Among the 159 cited foods, 81 of 153 foods (53%) and 72 of 153 foods (47%) (6 responses with missing information) led to OU attacks “every single time” and “only sometimes,” respectively, and 49 of 148 foods (33%) and 82 of 148 foods (55%) had been totally or partially excluded from the diet, respectively (11 responses with missing information).

Figure 2

Open in figure viewerPowerPoint

Percentages of specific foods identified as oral ulcer triggers. The data were provided by 81 patients with Behçet's disease for 24 preselected foods. The food items are ordered by decreasing frequency.

Table 4. Responses to open question on dietary oral ulcer (OU)‐triggering factors for 53 of 81 patients with Behçet's disease who declared at least 1 factor as triggering their OU attacksa

Food item	No. of patients (%)	Comment (no. of responses)
Nuts	18 (22)	–
Pineapple	11 (14)	–
Citrus fruits	11 (14)	Includes lemons 5, oranges 3, orange juice 2, clementines 1, grapefruits 1 and sour fruits 1 Three patients entered 2 citrus fruits.
Cheeses (other than Emmental)	11 (14)	Includes hard cheese 1, Roquefort 1, Comté 1, and Parmesan cheese 1
Emmental cheese	10 (12)	–
Strawberries	8 (10)	Includes strawberry ice cream 1
Tomatoes	7 (9)	–
Spices	6 (7)	Includes hot pepper 3 and spiced dishes 1
Bananas	5 (6)	Includes green bananas 1
Dried fruits	5 (6)	Includes dates 1
Kiwis	5 (6)	–
Alcohol	4 (5)	Includes pure whisky 1
Peanuts	4 (5)	–
Melon	4 (5)	Includes watermelon 4 and honeydew melon 1
Meat	4 (5)	Includes beef 1 and mutton 1
Almonds	3 (4)	–
Candies	3 (4)	–
Chocolate	3 (4)	Includes milk chocolate 1
Hazelnuts	3 (4)	–
Peaches	3 (4)	–
Soda	3 (4)	Includes Coke 3

• a Wordings used are those originally provided. Responses were grouped if they were considered to reflect similar contents (and detailed in the comments). Only items quoted by at least 3 patients are listed in the table, and 25 additional response items given by 1 or 2 patients were coffee (2), fruits (2), oil (2), fish (2), paprika (2), sour food (2), apricots (1), eggplants (1), chewing gum (1), chips (1), pickles (1), nectarines (1), onions (1), apples (1), vinegar (1), sour juices (1), sweet food (1), salty food (1), and hot beverages (1).

Among the 24 foods listed in the closed question (question 17), ≥1 were reported as OU triggers by 66 patients (81%). The most commonly cited factors were nuts (39 patients [48%]), pineapple (34 [42%]), peanuts (26 [32%]), Emmental cheese (24 [30%]), almonds (19 [23%]), lemons (18 [22%]), other cheeses (17 [21%]), tomatoes (15 [19%]), strawberries (14 [17%]), and oranges (14 [17%]) (Figure 2). Sensitivity analysis of patients with the most active OUs did not substantially alter these results (detailed results not given).

The responses about physicochemical food characteristics linked with OU attacks (question 18) are in Figure 1. No food characteristic clearly emerged as a triggering factor, apart from sourness, in both the main and sensitivity analyses (see Supplementary Figure 1, available on the Arthritis Care & Research web site at http://onlinelibrary.wiley.com/doi/10.1002/acr.23155/abstract). Fast eating was noted by only 1 patient (1%) as surely and by 17 patients (21%) as possibly related to OU recurrences (question 19). These numbers were virtually unchanged when analyzing patients with the most active OUs (detailed results not given).

DISCUSSION

From self‐reported information for 81 patients with BD obtained by open and closed questions, OU recurrences appear to be triggered by an external event in 62–70% of patients, with stress/fatigue and food being the most common specific factor (cited by 37–47% and 32–35% of patients, respectively). In contrast, mechanical and hormonal factors were less frequently cited triggers. The robustness of these results is supported by the consistency of the information collected from open and closed questions and from patients with the most active OUs.

Our results suggest psychological factors as predominant precipitators of OU recurrences in BD. Stress was also an OU trigger in 51–63% of patients with RAS 16, 17, and studies indicated greater psychological stress in patients with active RAS than controls 18-21 and a temporal relationship between stressful events and OU episodes 22. In addition, our BD patients declaring fatigue as a major OU‐triggering factor may reflect the fact that long‐term stress can be perceived as fatigue. The mechanisms by which psychological factors trigger OU episodes may involve a dysfunctional immune system. Increased salivary cortisol levels were found in patients with active RAS 18, 19, 23, and the triggering effect of immunosuppression may explain the link of OU recurrences with intercurrent infection, cited by some of the patients in our study and 14% of patients with RAS 13.

Food was the other predominant trigger of BD‐related OU recurrence. In particular, tree nuts, cheese, citrus fruits, pineapple, and strawberries were spontaneously reported as OU triggers, each reported by at least 10% of the respondents. The corresponding closed question revealed similar triggering food items but with twice as high frequencies as compared to the responses given to the open questions. The reasons for the higher yield of closed questions is unclear, but similar observations were reported from other surveys studying behavioral aspects 24. The questions seeking a causative physicochemical characteristic of OU‐triggering foods failed to reveal any shared gustatory, mechanical, or thermal properties, except for sourness. The role of dietary factors in RAS is not well studied. Several studies noted that patients with RAS had food sensitivity 12, 16, 25 and estimated that food triggered OUs for 15–22% of them 16, 17. Some data on specific food items potentially involved in RAS attacks suggested a triggering role for citrus fruits, nuts, fruits, figs, tomatoes, cheese, chocolate, apples, and acidic food 12, 13, 26, 27.

Histamine, a mediator with proinflammatory properties that modulates the activity of T and B lymphocytes, neutrophils, eosinophils, basophils, macrophages, dendritic cells, and endothelial cells 28, 29, is the most compelling link between the various food products declared as OU precipitators in BD. Most of the commonly cited food items are rich in histamine (i.e., Emmental and other hard cheeses) or are histamine releasers (e.g., tree nuts, pineapple, citrus fruits, and strawberries) 30. Patients also mentioned other food products known for their histamine‐rich or ‐releasing properties (e.g., tomatoes, peanuts, alcohol, spices, eggplants, and vinegar), although less frequently. Noteworthy, pathology studies consistently found a several‐times higher number of mast cells, the main source of histamine 28, 29, in OU biopsies of patients with BD or RAS than in the mucosa of controls 31-34. Moreover, when challenged by foods, patients with RAS showed increased histamine release 12, and the mast‐cell stabilizer cromoglycate sodium has had some benefit for OU attacks in RAS 35, 36.

Only 2% of patients spontaneously reported mechanical trauma or biting as OU triggers, and relatively few reported tooth brushing, dental care, hard food, chewing gum, or fast eating, which could result in wounding the oral mucosa, as triggers in the closed questions. For unclear reasons but sampling fluctuations or regional differences, this finding contrasts with data indicating trauma as a leading precipitating factor in 38–74% of patients with RAS 13, 16 and the observation that pricking the lower lip mucosa with a relatively large blunt needle led to OU formation at the puncture site in 47% of patients with BD 37. Also, menstrual cycles played a role in BD‐related OU attacks only in a few women as compared with 20–62% of women with RAS 13, 16. As was described for RAS 8, 9, women with BD who cited a link with menstrual cycles reported OU episodes exclusively during the luteal phase.

Collectively, our data suggest a potential central role of mast cells in the pathogenesis of BD‐related OUs. Beyond dietary factors, known triggers are emotional stress, intercurrent infection, physical stimuli 30, 38, and the premenstrual period 39 to bring on non‐IgE‐mediated mast‐cell degranulation or activation. Interestingly, abundant mast cells are also seen in skin pathergy tests 40, 41, a needle puncture‐induced cutaneous hyperreactivity that is a hallmark of BD 2, 42. In turn, BD‐related OUs could be conceptualized as a mucosal pathergy reaction in response to histamine‐rich food or to various dietary and nondietary triggers of mast cell degranulation or activation.

Our study has limitations and still unanswered questions. Compared to in‐person interviews, self‐completed questionnaires minimize within‐investigator or social desirability biases but bear the risk of poor adherence 43. Thus, considering the high 80% return rate of analyzable questionnaires, our findings were not likely substantially biased by nonresponse. Caution is warranted in the interpretation of our observed percentages of external factors perceived as OU triggers because these numbers were calculated for the overall population and not for the subset of patients actually exposed to a given dietary or nondietary factor. Therefore, these percentages may vary across populations with distinct exposures. With regard to studies indicating that oral health was impaired in patients with BD and RAS 44, our survey collected only self‐rated dental health information, and no eval‎uations by oral health care professionals were available. Therefore, our study was not designed to thoroughly describe the oral health characteristics of the analyzed patients and address the question of a putative enhancing effect of poor oral health on mechanical trauma as a trigger of OU recurrence. Also, our study design was not suitable for adequately investigating the potential effect of immunosuppressive or other medications.

To conclude, our findings suggest that many patients with BD can identify factors triggering their OU episodes. OU‐triggering factors should be sought in patients, although we cannot affirm that preventive measures, such as psychological support, exclusion of foods, or hormonal replacement therapy, provide significant relief. The generated hypothesis of a pivotal role of histamine and mast cells in OU recurrences contributes to the understanding of the pathogenesis and suggests avenues for treatment of recurrent OUs in BD. Additional studies from populations with different characteristics and environmental exposures are warranted to further explore these observations.

ACKNOWLEDGMENTS

The authors thank Drs. C. Bachmeyer (Paris, France) and T. Daikeler (Basel, Switzerland) for providing clinical information for patients recruited for this study.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Mahr had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design

Volle, Fraison, Papo, Thuillier, Rivière, Mahr.

Acquisition of data

Volle, Fraison, Gobert, Goulenok, Dhote, Fain, Lhote, Papo, Rivière, Mahr.

Analysis and interpretation of data

Volle, Fraison, Gonzalez‐Chiappe, Lhote, Mahr.