Oral health in children and adolescents with juvenile idiopathic arthritis – a systematic review and meta-analysis

Background Observational studies examining the association between oral health and juvenile idiopathic arthritis (JIA) among children and adolescents have reported inconsistent findings. The aims of this systematic review and meta-analysis were to ascertain a potential difference in oral health and oral health-related quality of life (OHRQoL) among children and adolescents with JIA and healthy peers, and to assess the association of prevalence of oral diseases/conditions, temporomandibular disorders (TMD), including temporomandibular joint (TMJ) diseases, in relation to activity and severity of JIA. Method Medline Ovid, Embase, CINAHL, SweMed+ and Cochrane Library were searched up to 25 November 2018. All articles published in English, German and Scandinavian languages focusing on children and adolescents with JIA and without JIA in relation to oral health measures, were considered. Two authors independently evaluated observational studies for inclusion. The study quality was assessed using modified Newcastle Ottawa Scale. Meta-analysis was performed for studies focusing on dental caries as an outcome. Results Nineteen articles met the inclusion criteria, covering a range of oral diseases/conditions and OHRQoL. Eighteen studies had cross-sectional design. No mean difference of dmft/DMFT indices (decayed/missed/filled teeth) was observed between the JIA - and healthy group. None of the oral health measures including dental erosive wear, enamel defects, dental maturation and OHRQoL, indicated better oral health among children and adolescents with JIA compared to healthy group. However, periodontal conditions and TMD were more predominant among children and adolescents with JIA compared to healthy peers. Conclusions Based on the cross-sectional studies, periodontal diseases and TMD were found to be more frequent in children and adolescents with JIA compared to healthy peers. Furthermore, more high-quality studies with large sample size are needed before we infer any concrete conclusion regarding the association between the prevalence of oral and TMJ diseases or oral conditions in relation to activity and severity of JIA.


Background
Juvenile idiopathic arthritis (JIA) is a common chronic rheumatic condition, affecting around 1 in 1000 children under the age of 16 years [1,2]. The incidence and prevalence of JIA varies across different studies globally, but by pooling data from several studies, it is estimated that around 60,000 children below 16 years are affected in Europe, with an estimated incidence of around 7000 new cases each year [3]. The incidence in the Nordic countries including Norway is among the highest in the world [4]. JIA comprises a group of distinct clinical entities of unknown aetiology, characterized by joint inflammation with symptoms persisting for more than six weeks and onset before 16 years of age [5]. Currently, it is classified according to the International League of Associations of Rheumatology (ILAR) as systemic arthritis, polyarthritis (Rheumatoid factor (RF) negative or positive), oligoarthritis (persistent or extended), enthesitisrelated arthritis, psoriatic arthritis and undifferentiated arthritis [5].
Long-term inflammation and use of anti-inflammatory drugs, such as corticosteroids, may cause disturbances in growth and pubertal development, overall bone maturation, and eventually the development of osteopenia with low bone mineral content and low mineral density. These consequences are found to be associated with the duration of active JIA and severity [6] and are more frequent in individuals with early-onset JIA [7].
Overload of bacteria is considered as a possible trigger of rheumatic arthritis (RA) in adults [8]. This means that the oral cavity, one of the most bacteria colonised parts of the body and hosting nearly 800 species of bacteria [9], should be kept free from oral diseases. When the oral microbiota [10] shifts from balance to imbalance (dysbiosis), e.g. during rapid caries development, bacteria might pass through exposed dentine, pulp or periapical bone to the bloodstream [11]. In case of plaque accumulation at gingival margins or during ongoing gingivitis or periodontitis, bacteria might pass the blood stream through periodontal pockets, or through the oral mucosa directly if there is oral mucositis or ulcer. In patients with RA, dysbiosis has been detected in the gut and oral microbiomes (dental and saliva microbiome) and has been found to be correlated with clinical measures of RA status and to be altered compared with healthy individuals [12].
Individuals with JIA may be subjected to unfavourable underlying oral health determinants. If JIA reduces functional ability of the upper limbs, effective tooth brushing and plaque removal will be difficult. Plaque removal might also be impeded in children with JIA with restriction in mouth opening [13]. When JIA is accompanied by impaired masticatory function, consumption of softer and more sugary foods in small amounts might be more convenient [14]. Frequent and long-term intake of liquid oral medication with sugary or acidic content has previously been reported in children with JIA [15], but today sugar-free alternatives exist [16], and there is reason to believe that such intake is more rare. Knowledge of intra-oral adverse effects and frequency of side effects of modern long-term administration of anti-rheumatic drugs, is hitherto scarce.
Temporomandibular disorder (TMD) is an umbrella term including Temporomandibular joint (TMJ) involvement as well as localized pain in the masticatory muscles, decreased mouth opening and chewing ability, pain associated with mandibular movement during eating, chewing or yawning, and comorbidities such as earache and headache [7]. A high proportion of children with JIA might have involvement of the TMJ during disease course [17]. The consequences of local inflammation in the TMJs may involve local growth disturbances and as a consequence impaired mandibular growth [18]. Development of malocclusion and facial deformities such as micro-or retrognathia, are later scenarios associated with established permanent sequelae in the TMJ [7]. To identify TMJ arthritis early enough to prevent permanent growth disturbances and joint damage, it is important to recognize all clinical symptoms associated with JIA involvement. A challenge is that TMJ arthritis might evolve without or with TMD symptoms, especially in the youngest children who are unable to communicate and localize their pain adequately [7]. Thus, early detection of TMD by imaging signs of inflammation in the joints is essential.
In 2016, the key part of Vision 2020 [19] was approved, including an upgraded definition of oral health which is estimated to be multifaceted and to include different attributes of oral health. The new definition not only includes disease and condition status, but also underlying determinants, moderating factors, overall health and well-being. Thus, the ability to speak, smile, smell, taste, touch, chew, swallow and express emotions, functioning without feeling pain or discomfort, are integrated components in oral health. Children and adolescents with TMJ arthritis, may experience reductions in one, some or all these abilities, resulting in both reduced quality of life (QoL) [20] and reduced oral health-related quality of life (OHRQoL) [21]. For the group of children and adolescents with JIA, documentation of reduced OHRQoL due to oral diseases restricted to the oral cavity, e.g. dental caries, dental erosion, and not including jaw symptoms, is sparse.
Whether children and adolescents with JIA have a heavier burden of oral conditions and as a consequence, experience reduced OHRQoL, is not clearly established. For this reason, the aims of this systematic review were to gain reliable information on the following research questions;

Inclusion and exclusion criteria
This review primarily reports articles restricted to peerreviewed journal articles published in English, German, Norwegian, Swedish or Danish during the period 1998 through 25 Nov 2018 covering children and adolescents' age groups. Randomised controlled trials (RCTs), controlled clinical trials (CTs), cohort studies, crosssectional studies or case-control studies were included. The exclusion criteria were systematic reviews, metaanalyses, case reports, conference publications and grey literature. Grey literature was excluded as this type can vary considerably and often be affected by low standard of quality, review and production. In addition, studies lacking comparing groups (i.e. groups without JIA) were excluded for the analyses purpose. Lastly, as another systematic review is planned, articles mainly addressing saliva variables and orthodontic considerations in children and adolescents with JIA, were excluded.

Search strategy
PRISMA [41] was followed as a guide for reporting this systematic review and meta-analysis. The levels followed in the literature search were as follows: 1) title and authors, 2) abstracts, and 3) full text. For abstracts decided to be within the scope of interest, full-text articles were read. Two reviewers (MSS and AB) independently evaluated studies for inclusion, and studies were selected after reading abstracts, and selected full-text articles. When abstract selection was not straightforward and the reviewers were in doubt, full-text articles were re-read by both reviewers and resolved by discussion. A flow diagram is presented as Additional file 5: Figure. S1.

Outcomes
Oral health and OHRQoL were assessed among children and adolescents with JIA and among those without JIA, and these examination data constituted the outcomes. Any outcome measures with information outside the scope of this review, but within included articles, are not mentioned. An overview of key information from the final evaluation is shown in Tables 1 and 2 in a similar way as a previous systematic review [74].

Qualitative assessment
Central themes and topics from the PICOS (participants, interventions, comparators, and study design) approach were only to some extent extracted in Tables 1 and 2 as an intervention was not the focus. The characteristics considered important for the evaluation of reliability and validity, were study design, level of control matching, exclusion criteria, non-respondents, sample size, calibration procedures, number of examiners, documentation of JIA history (activity assessment, laboratory evaluation, medication), applied imaging type, and oral health diagnostic tools. MSS conducted the data extraction and checked by AB for accuracy. Assessment of risk of bias was performed based on an adapted version of the Newcastle -Ottawa Scale (NOS) [75], which was further modified in support of this systematic review (Additional file 2: Table S2). Scoring was performed by two authors (MSS and AB), but in case of discrepancies, a third author (AS) was consulted. The range of the scores was from 0 to 10 (low risk of bias = overall scores were 9-10, medium risk of bias = 6-8, high risk of bias = 0-5). Summarized scores of each study are presented in Additional file 3: Table S3.

Statistical analysis
It was not possible to perform meta-analysis of oral health outcomes regarding oral hygiene (dental plaque and calculus accumulation), periodontal disease (gingivitis included), enamel defects, tooth calcification (dental maturation) disorders, TMJ arthritis, TMJ involvement, TMD, oral ulcerations, and OHRQoL. The reasons include inadequate sample size, poor study quality, use of inconsistent definitions of outcomes (e.g. periodontitis assessment) or studies that failed to report number of children and adolescents with JIA. Nevertheless, metaanalyses for dental caries was performed. Two separate meta-analyses were conducted using continuous outcomes: dmft score (decayed/missed/filled primary teeth) and DMFT score (decayed/missed/filled permanent teeth). We used random-effect model [76] to calculate pooled mean differences between dmft / DMFT scores of children and adolescents with JIA and those without JIA. The articles that did not report dmft or DMFT score or standard deviation were excluded from this meta-analysis. The heterogeneity between the studies were quantitatively assessed by the Q-test and I 2 statistics [77]. I 2 is the proportion of total variation explained by between-study variation. I 2 -values of 0, 25, 50% and ≥ 75% indicates no, low, moderate and high heterogeneity, respectively. Publication bias was assessed by inspection of funnel plots for asymmetry and using Egger's test [78] and Begg-Mazumdar test [79].

Results
Nineteen articles met the inclusion criteria, ten from Europe and nine from countries outside Europe with Brazil in the lead, see Table 1. The age range of children and adolescents with JIA was from two to four years in two studies [22,31] and up to 20 years of age in one study [30]. Altogether, the included articles covered topics such as dental caries, oral hygiene (dental plaque and calculus accumulation), periodontal disease (gingivitis included), enamel defects, tooth calcification (dental maturation) disorders, TMJ arthritis, TMJ involvement, TMD, oral ulcerations, and OHRQoL. Beyond these topics, information about inflammatory mediator measurements in blood samples and gingival crevicular fluid, was reported.
Eighteen studies were cross-sectional in nature, only a study by Miranda et al. [32,33] had a prospective cohort study design. At baseline, adolescents with and without JIA were examined for clinical and immunological variables of periodontal inflammation, and two years later a subgroup, eighteen adolescents with JIA and fourteen without JIA, were re-examined. Another study of Lehtinen et al. [24], distributed coded radiographs at random between different examination sessions, so the only examiner was blinded for the information of whether the radiographs belonged to participants with JIA or healthy controls.
All included studies reported age of children and adolescents with JIA and of those without JIA. However, the degree of matching varied. Although no study of true case-control design was included in the review, two studies showed controls matched for age, gender and ethnicity [22,39]. Another characteristic of included studies was a distinct variation of sample sizes. In some studies, the sample size was too low to justify any result evidence. Beforehand sample size calculation was uncommon, as only one article [29] described this. In most studies the number of examiners was low, usually one. With some exceptions [22,24,31,33], no description of calibration of examiners or reliability values were included. Bitewing radiographs were reported by only two research groups [25,26,32].

Dental caries
Eight of the included articles described dental caries, but with divergent results. Both Ahmed et al. and Welbury et al. [22,31] documented that a significantly larger proportion of children with JIA had untreated caries compared with healthy peers. Welbury et al. [31] also documented that individuals with JIA, had a higher burden of caries than individuals without JIA; among children, more primary teeth decayed, filled or extracted and among adolescents, predominantly more dental decay (D: decayed component in the DMFT). In contrast, the study of Santos et al. [29] revealed caries in primary teeth to be more frequent among healthy children than among children with JIA. Five articles [23,25,27,30,39] did not show any significant difference between the children and adolescents with and without JIA when subgroups were not included. The way of reporting caries varied from untreated caries, dmft, DMFT, D to caries prevalence of affected individuals. The use of diagnostic tools also varied. Both the World health Organization (WHO) criteria [45] and the British Association for the Study of Community Dentistry (BASCD) standardized criteria [80] were used, while some studies did not report the caries diagnostic tool used. Only one study by Leksell et al. [25], reported enamel caries.

Quantitative synthesis
Three cross-sectional studies (three publications) were included in the analysis to evaluate the association between caries in primary teeth and JIA (71 children with JIA and 141 total participants). We observed no difference in summary mean dmft scores between JIA and those who did not experience JIA (− 1.16, 95%CI, − 3.02-0.71, I 2 = 87.9%, p heterogenity = < 0.0001) (Fig. 1).
Six cross-sectional studies (three same publications as used above as they comprised of data from both primary-and of permanent teeth, and three other publications) were included in the analyses to evaluate the association between caries in permanent teeth and JIA (162 children and adolescents with JIA and 320 total participants). We observed no difference in summary mean DMFT score between children and adolescents with JIA and those who did not experience JIA (− 0.08, 95% CI, − 0.42 to 0.26, I 2 statistic = 0.0%, %, p heterogenity = 0.95) (Fig. 2).
No evidence of publication bias with Egger's test (P dmft = 0.27, P DMFT = 0.78) or with Begg's test was found (P dmft = 0.98, P DMFT = 0.45) (Additional file 6: Figure. S2 and Additional file 7: Figure. S3). However, because of the small number of studies and small sample size of included studies, the results from these formal tests should not be inferred with great reliability.

Plaque, gingivitis and periodontitis
The oral health descriptors most often reported were dental plaque and signs of periodontal inflammation (gingival bleeding and bleeding on probing, probing depth ≥ 2 mm, clinical attachment loss, pocket depths etc.). There were studies focusing on oral hygiene and dental plaque [23,25,28,31] showing a statistically higher Plaque Index (PI) or Simplified Oral Hygiene Index (OHI-S) in the JIA group compared with those without JIA. Other studies [22,26,27,29,30,39] did not find this association. Additionally, Leksell et al. found calculus to be significantly more prevalent in individuals with JIA compared with those who did not experience JIA [25]. Many articles also documented poorer periodontal status among children and adolescents with JIA; more gingival inflammation and gingival bleeding [22,29,31], bleeding upon probing [25], deeper probing depth [25,26] and periodontal attachment loss [26,28]. However, not all articles documented differences in periodontal status when comparing individuals with JIA with healthy counterparts [23,30,39].

Developmental enamel defects and ulcers
The only study reporting developmental enamel defect [22], found the condition more frequent among children with JIA than among healthy peers, but the sample size was too small to draw any reliable conclusion. Another study focused on oral ulceration [25] and found five out of forty-one children with JIA to be affected, but only one out of forty-one children in the group without JIA.

Dental maturation
Two of the included studies investigated the status of dental maturation and found divergent results. By examining orthopantomograms (OPG), Lehtinen A et al. [24] in 2000 documented more advanced dental development in children with juvenile rheumatoid arthritis (JRA) compared with healthy peers, while Ley et al. [40] nine years later assessed dental maturity in children and adolescents with JIA to be within the norms of healthy peers.
TMD TMJ arthritis (active inflammation of the joint), TMJ involvement (osteoarthritis or growth disturbances as a result of TMJ arthritis) [81] and TMD were coherently reported more frequently among children with JIA than among healthy peers [35][36][37][38].

OHRQoL
In the study of Leksel et al. [38], orofacial symptoms influenced more often the daily life in a group of children with JIA compared to the healthy individuals. Santos et al. [29] also compared oral health-related quality of life in children and adolescents with JIA and healthy peers. The instrument Brazilian SF:13-B-PCPQ instrument was used and consisted of thirteen items related to oral symptoms, functional limitations and well-being. In the different groups, most caregivers indicated that the oral health status of their children and adolescents had little or no effect on their well-being, and no significant differences between the groups were found. For the present review Additional file 3: Table S3 presents the critical appraisal of the included studies while Additional file 4: Table S4 shows a completed 2009 PRISMA check list.

Influence by JIA activity and severity
The majority of the studies contained some clinical information about the JIA status of the participants. Examples of descriptors were JIA category, disease activity, anti-rheumatic medication, JIA onset and functional impairment. Pugliese et al. [27] showed that wellestablished JIA disease and validated activity scores (Juvenile Arthritis Disease Activity Score (JADAS), physician global assessment of disease activity visual analogue scale (PhysglobVAS) and parent/patient global assessment of well-being VAS (PglobVAS) were positively correlated with the DMT score. Other scores; Escola Paulista de Medicina Range of Motion Scale (EPM-ROM) and Child Health Assessment Questionnaire (CHAQ) were positively correlated with Gingival Index (GI), PhysglobVAS was correlated with PI, and Pediatric Quality of Life Inventory 4.0 (PedsQL) parents was correlated with Gingival Bleeding Index (GBI). Savioli et al. [30] found that a subgroup of children with polyarticular RF negative JIA, had a statistically higher GBI and DMFT index than controls. Also a subgroup of children with three to eight affected joints in upper extremities, had significantly higher bleeding index than controls [30]. Self-reported pain or weakness in the hand when tooth brushing was documented in the study of Leksell et al. [25] as a problem among children with JIA. A significantly higher proportion of children with JIA compared with children without JIA, also answered that they did not brush their teeth when they did not feel well. Additionally, Miranda et al. reported a significantly higher mean number of joints with limitation of movements (LOM) in adolescents with two mm or more attachment loss (AL) than in adolescents without AL [26]. It is also worth mentioning the findings of Miranda et al. [32] of increased serum IL-18 and IL-1β in adolescents of JIA subgroups with AL, suggesting that AL might be associated with systemic inflammatory response. Low sample sizes, however, made it difficult to draw conclusions.
For children and adolescents with JIA, medication constitutes a substantial part of life which in turn may impact on oral health. Leksell et al. showed that children taking anti-TNFα had a higher frequency of sites with increased probing depth compared to children not taking this medicine [25]. Reichert et al. [28], comparing adolescents with JIA who took non-steroidal antiinflammatory drugs (NSAIDs) with other JIA peers who did not take drugs, found a significant decreased mean value for modified sulcular bleeding index in the NSAID group. The frequency of cyclosporine medication, assessed by Pugliese et al. [27], was found to be higher in JIA patients with gingivitis compared with those without gingivitis. It is important to bear in mind that it is very difficult to differentiate the effect of single drugs from the effect of the disease activity with associated systemic inflammatory response in these studies. Children on anti-TNFα or cyclosporine probably have a more severe disease than children without these drugs, and differences in oral health between groups with or without a certain drug may be due to the disease severity and not the drug itself. Miranda et al. [33] in a follow-up study of adolescents with JIA documented that anti-rheumatic treatment resulting in reduction in disease activity clearly and positively influenced the periodontal status. After two years, no clinical or laboratory differences in periodontal inflammation could be documented between the adolescents with and without JIA. Pugliese et al. [27] documented that adolescents with an increased C -reactive protein (CRP) showed a higher mean clinical attachment loss (CAL) compared with those with normal CRP values.
All these reports about medication shared the previously reported problem of lacking adequate sample sizes for evidence. The comparisons were also hampered by the fact that the disease status of children and adolescents with JIA taking a certain drug were not the same as children and adolescents with JIA who did not take the drug.

Discussion
The aim of this systematic review was to investigate the relationship between oral health measures and OHRQoL among children and adolescents with JIA compared with peers without JIA. The present systematic review and meta-analysis includes mostly studies with crosssectional design and the overall qualitative assessment of these studies was found to be low. As a wealth of information was reported, interpretation of the data needed a clear and thorough reporting of methodology, quality and bias [82]. Tables 1 and 2 constituted the quality evidence basis when answering the research questions of the article. For dental caries, also caries meta-analysis (a quantitative method to combine data) was feasible.
Reviews published more than a decade ago have concluded that oral health in children with JIA is poor [13,14]. However, the articles examined in these reviews were from the 1970s and the 1980s. Conclusions from this review based on more recent research on the caries situation among children and adolescents with JIA, have not been easy to draw. Due to insufficient sample size in the study of Santos et al. [29], the statistically lower dmft value among children with JIA compared with healthy peers, is not compact. The opposite conclusion drawn by Welbury et al. [31], showing a higher mean dmft among 0-11-year-olds with JIA and a higher D component among 12-17-year-olds with JIA, is probably a more reliable finding due to higher sample sizes and a calibrated examiner. Nevertheless, as the total sample included many subgroups, the exact number in the two subgroups reported was not reassuring; the youngest subgroup of both individuals with and without JIA included 46 individuals, the older subgroup 32. Additionally, bitewing radiographs were not included in the caries examination, which actually meant an underscoring of approximal caries lesions and of total caries experience [83]. However, as both cases and controls were examined without bitewing radiographs, it was not necessary to take into account any bias in the comparison.
Although the present review evaluated eight articles with caries as subject, we could not conclude that caries was more prevalent among children and adolescents with JIA than among healthy peers. The findings from this meta-analysis on dental caries suggest no significant mean difference in dmft or DMFT between JIA affected individuals and not. One of the explanations for a possible improved caries status in individuals with JIA during the later years, might be the development of a more effective overall treatment of JIA [84]. Another explanation is the increased focus on oral health in JIA, including the development of other sweeteners and sugar alternatives used in medicines e.g. in NSAID mixtures [16].
The finding that plaque, gingivitis and periodontitis were more common among children and adolescents with JIA than among those without JIA, constituted a consistent trait in the present review. The fact that so many studies drew this conclusion increased the quality of evidence supporting this result. Unfortunately, it was not possible to perform meta-analysis due to inconsistency of outcome definitions for periodontitis across the studies.
The present review lacked studies with focus on erosive wear, a condition which in later years has been reported to be as commonly distributed as caries in some groups of adolescents [85,86]. Only one study reported on enamel defects [22], but it had very small sample size, so reliable information about the prevalence of this oral condition is still lacking. To the question of whether dental maturation was more advanced among patients with JIA than among healthy peers, there was no clear answer. The OPG radiographs in the study of Lethinen et al. [24], dating from the late 1960s to the early 1980s, were therefore too old to represent today's patients, and the study of Ley et al. [87] instead of matched controls, compared the findings in children with JIA with normative values obtained from healthy Canadian, German and Dutch children. Conclusively, for dental erosive wear, enamel defects and dental maturation, there is no scientific evidence to answer the posed research questions.
Concerning TMJ arthritis and TMJ involvement, the present review consolidated the literature reporting these conditions to be more common in children and adolescents with JIA than in healthy counterparts [88]. However, not all the five included articles that described this topic had sufficient sample size to give reliable results [22,[35][36][37][38], but a higher frequency of surface flattening of the condylar head in children with JIA versus those without JIA, seemed to be a valid radiological feature, reported by Shwaikh et al. [35]. Furthermore, TMD and structural TMJ changes were found to be more prevalent in children with JIA than in healthy peers [38], and when comparing OHRQoL in the two groups, these were poorest among the children with JIA [38]. This was not a surprising result, taking into account that oral health definition includes all functioning without feeling pain or discomfort.
In order to answer the research question, whether OHRQoL only due to oral diseases/conditions restricted to the oral cavity is more common among those with JIA than those without, more studies related to this topic are needed. Only one study was included in the review concerning this topic: Santos et al. [29] documented that oral health status had little or no effect on well-being among both individuals with JIA and those without JIA.
To be able to respond to the second research question, whether the activity and severity of JIA had any impact on the prevalence of oral and TMJ diseases or oral conditions, larger sample sizes are needed before reliable answers can be given.

Strengths and limitations
The strength when comparing children and adolescents with JIA with those without JIA, is that the overall oral health outcomes were taken into account and discussed elaborately. Another strength of the study was the adoption of PRISMA protocol [41] and the use of modified Newcastle-Ottawa Scale to comprehensively evaluate and assess the methodological quality of the selected studies [75]. Additionally, meta-analysis was performed for studies focusing on dental caries as an outcome. However, the present. systematic review was not without limitations. Firstly, majority of the included studies were cross-sectional in nature which is tied to high risk of bias. Secondly, due to inadequate studies and inconsistency of outcome definitions, only meta-analyses regarding dental caries, not regarding other oral diseases or conditions, could be performed. Thirdly, as grey literature was excluded in the present systematic review, the comprehensiveness of the search might have been reduced and therefore should be considered as a limitation of the review [89].