The link between ankylosing spondylitis and oral health conditions: two nested case-control studies using data of the UK Biobank

Abstract Ankylosing Spondylitis (AS) is an inflammatory rheumatic disease that affects the axial skeleton and the sacroiliac joints. Recent studies investigated the link between AS and oral diseases, particularly periodontitis. Others suggested that periodontitis may have a role in the pathogenesis of rheumatic diseases. Objective: The aim of this study is to investigate the association between AS and oral conditions. Material and Methods: This research was conducted using the UK Biobank Resource under Application Number 26307. The UK Biobank recruited around 500000 participants throughout Great Britain. Clinical records were available for 2734 participants. Two case-control studies were conducted based on whether AS was self-reported or clinically diagnosed. Oral conditions were identified using self-reported reports of oral ulcers, painful gums, bleeding gums, loose teeth, toothache, and dentures. The association between AS and oral conditions was assessed using logistic regression adjusted for age, gender, educational level, smoking status, alcohol consumption, and body mass index. Results: A total of 1307 cases and 491503 control participants were eligible for the self-reported AS study. The mean age was 58 years for the cases [7.5 standard deviation (SD)] and 57 years for the control groups (8.1 SD). Also, 37.1% of the cases and 54.2% of the control participants were females. Among the oral conditions, only oral ulcers were strongly associated with AS [1.57 adjusted odds ratio (OR); 95% confidence interval (CI) 1.31 to 1.88]. For the study of clinically diagnosed AS, 153 cases and 490351 control participants were identified. The mean age for both cases and control groups was 57 years; 7.6 SD for the cases and 8.1 for the control group. Females corresponded to 26.1% of the cases, and 54.2% of the control participants. Clinically diagnosed AS was associated with self-reported oral ulcers (2.17 adjusted OR; 95% CI 1.33 to 3.53). Conclusion: Self-reported and clinically diagnosed AS populations have increased risk of reporting oral ulcers. Further investigations are required to assess the link between a specific type of oral condition and AS.


Introduction
reporting of the number of natural teeth, presence or absence of prosthetic appliance, and root canal filling by the patients was agreeable with the clinical findings for these patients. 14 The aims of this study are: 1. To report the findings of two case-control studies based on the UK Biobank database; 2. Identify any associations between AS and oral conditions identified in the database.

UK Biobank
The UK Biobank is a large population-based Detailed cohort protocol, scientific rationale, and study design are available online. 15 Study design Two nested case-control studies were conducted based on the method for defining AS; self-reported AS case-control study, and clinically diagnosed AS case-control study. The first study investigated whether individuals who self-reported AS have higher prevalence of oral health problems than those who did not. The outcome is self-reported AS and the exposure is self-reported oral conditions. The second case-control study investigated the association between clinically diagnosed AS and self-reported oral conditions.

Identifying AS cases
In the self-reported AS case-control study, AS was identified by asking the participants whether they had been told by a doctor that they have some sort of severe non-cancer illness or disability. If the participant was uncertain of the type of illness they had, they The link between ankylosing spondylitis and oral health conditions: two nested case-control studies using data of the UK Biobank J Appl Oral Sci. 2019;27:e20180207 3/10 described it to the interviewer (a trained nurse) who attempted to place it within the coding tree. If the illness could not be located in the coding tree, then the interviewer entered a free-text description of it. These free-text descriptions were subsequently examined by a doctor and, where possible, matched to entries in the coding tree. Free-text descriptions which could not be matched with very high probability were marked as "unclassifiable" In the clinical case-control study, AS was identified based on clinical records using the International Classification of Diseases, 10 th Revision (ICD-10).
Clinical records were available for 2734 participants only.

Definition of oral conditions
Oral conditions were defined according to the selfreported data on mouth/teeth or dental problems.
The participants used a touchscreen to answer the question "Do you have any of the following? (You can select more than one answer)''. The possible answers were: Mouth ulcers, painful gums, bleeding gums, loose teeth, toothache, and dentures.

Exclusion criteria
Self-reported RA, and other back problems which are not classified as AS such as spine arthritis/ spondylosis were excluded from the analysis plan of self-reported AS.
During the analysis of clinically diagnosed AS, the participants with clinically diagnosed RA and selfreported RA were excluded from both the cases and control groups. Similarly, self-reported AS participants were excluded from the control group only to avoid including participants who might have delayed clinical AS diagnosis.

Data analysis
Descriptive analysis was used to describe the characteristics of the study population including total number and percentage. Mean and standard deviation (SD) were used to describe the age of the participants.
In order to assess the relationship between AS status and oral conditions, a logistic regression model was used to calculate the odds ratio (OR) and 95% confidence interval (CI) adjusted for age, gender, educational qualification level, smoking status, alcohol consumption, and body mass index. Statistical significance level was defined at p=0.05. All data were processed using the IBM SPSS Statistics Software Package, version 24.

Results
As we have conducted two case-control studies, we will describe the findings of each study separately.

Population description
Following the exclusion criteria, 1307 participants reported having AS and 491503 were considered as part of the non-AS or control groups. The mean age of the AS cases group was 58 years (SD 7.5), and 57 years for the control group(SD 8.1). The age group between 55 and 64 years old predominated both in the AS and non-AS groups (49.8% and 42%, respectively).
Over a third of the AS (37.1%) and half of the control group (54.2%) were female. The education level for both AS and non-AS participants was roughly the same: 39.1% of the AS cases and 39.8% of the non-AS participants had college or university degree.
Current smokers were more prevalent among the AS cases compared to the non-AS participants (14.9% and 10.5%, respectively). On the other hand, a lower percentage of AS participants reported having never smoked compared to the non-AS cases (44.6% vs. 55%). Please see Table 1 for more details.

Prevalence of oral conditions and their association with self-reported AS
Generally, the AS participants reported a higher prevalence of oral conditions compared to the control group (Table 2). Also, they reported a higher prevalence of oral ulcers than the control group (14.5% vs. 9.9%). There was more than 50% increase in the risk of reporting AS among those who reported having oral ulcers (1.54 crude OR; 95% CI 1.32 to 1.80). Furthermore, no change was observed after adjusting it for confounding factors (1.57 adjusted OR; 95% CI 1.31 to 1.88). Further analysis was conducted by excluding participants who reported having dentures, to exclude oral conditions related to tooth loss and denture wear ( show any significant change in OR (Tables 2 and 3

Population description
In this study, 153 participants clinically diagnosed with AS and 490351 serving as control were identified after excluding clinically diagnosed RA, self-reported AS (from the control group only), and self-reported RA.
The mean age was the same for both AS and non-AS participants (57 years; 7.6 SD and 8.1 SD for AS and non-AS, respectively). The distribution of age group between AS and non-AS participants was roughly the same: 9.2% of AS and 10.3% of non-AS participants had less than 45 years of age, and 16.3% of the AS and 19.1% of the non-AS participants were older than 64.
Females consisted of around a quarter of the AS cases (26.1%), and half of the non-AS participants (54.2%).
In broad terms, the distribution of educational qualification level in the AS and non-AS participants was roughly the same. The number of AS participants who had a college or university degree was relatively lower than the number for non-AS participants (35.7% and 39.8%, respectively) ( Table 4).    The participants who reported having dentures represent nearly one fourth of the AS cases (22.2%), whereas only 16.3% of the non-AS participants reported having dentures. An association was found between AS and reporting having dentures (1.47 crude OR; 95% CI 1.00 to 2.14). However, taking confounding factors into account during the regression analysis led to a non-statistically significant association between AS and denture reports (1.31 adjusted OR; 95% CI 0.78 to 2.18).

Discussion
Our findings demonstrate a link between AS and oral ulcers. Other oral conditions, such as painful gums, bleeding gums, loose teeth, toothache and reporting having dentures were not associated with AS; however, their prevalence was generally higher in the AS than in the control group.

Method critique
Generally, the AS and control groups were not different in relation to mean age. This supports the homogeneity of the two populations. The AS population was significantly different from the non-AS population in terms of male/female ratio, smoking status, and education level. In the AS cases, the number of males was three times higher than the number of females. This gender ratio was previously shown globally, 17 as AS affects males more than females. Recently, the effect of smoking was shown as a risk factor for AS. 18 It has been shown that smoking might be associated with the development and progression of AS. This is in line with our study which showed a high rate of current smokers in the AS compared to the non-AS population. In the clinically diagnosed AS cases, those who had a higher education degree were less prevalent when compared to the non-AS population.
These findings corroborate a previous study 19   or necrotising periodontitis which are related diseases.
Therefore, painful gums might not be the right measure to identify periodontitis.
To the best of our knowledge, no study has investigated the link between toothache and AS.
However, one study showed an association between self-reported arthritis and toothache 32 which might oppose our findings. Toothache could result from acute periodontitis (i.e., periodontal abscess). 28 However, it is impossible to diagnose the actual cause of toothache based on the self-reported questionnaire only.
Toothache has previously been shown to be related to poor oral health and thus to quality of life, as it affects the daily performance of the individual. 33,34 Therefore, associating toothache with arthritis could be a sign of low health-related quality of life.
Those who have dentures have lost their teeth partially completely. Our findings suggest that losing teeth increases the risk of AS, but these results were found only among the self-reported AS patients.
Losing natural teeth could be due to a range of causes, mostly including advanced periodontitis, dental caries, and trauma. It is impossible to predict the number and the cause of natural teeth loss by asking about the presence or absence of dentures.
It is generally accepted that losing natural teeth can predict bad oral health and has an impact on the patients' quality of life. 35 In addition, losing natural teeth is associated with age, which might explain the faded association with AS once common risk factors were considered in the analysis.

Conclusion
Within the limitations of this study, the AS population showed increased risk of reporting oral ulcers. There was an increase in the prevalence of reports of painful gums, toothache, loose teeth and loss of natural teeth in the AS compared to the non-AS population. However, when common risk factors between AS and oral conditions were taken into account, this association was impaired. Therefore, further investigations are required in this field by designing large population-based studies, with clear and concise definitions of oral diseases. In addition, designing a prospective study to identify the timecourse of oral ulcers related to AS would be of interest to identify the temporal association between the two conditions.