Ankylosing spondylitis (AS), a complex autoimmune inflammatory rheumatic disease, has long been considered the archetype of spondyloarthritis (SpA). Common symptoms of AS include arthritic symptoms (such as inflammatory back pain, muscle spasms, and sacroiliac arthritis), potential extra-articular symptoms (such as uveitis, psoriasis, and inflammatory bowel syndrome), and the involvement of the heart, bone, lung, kidneys, and skin (1, 2). The worldwide prevalence of AS ranges between 0.07% and 0.32% (3). In addition, clinical symptoms of patients with AS usually appear between the ages of 26 and 45 years. Men also are more likely to suffer from AS than women, the prevalence being two to three times higher in men than in women (4–6).

The pathogenesis of AS is complex and multifactorial. Early studies have confirmed that AS is strongly associated with the inheritance of HLA allele B27, which might misfold in the endoplasmic reticulum (ER), leading to the upregulation of interleukin (IL)-23 in dendritic cells (7–9). It may also result in the presentation of intracellular peptides to T cells, which may trigger cross reactions, leading to tissue inflammation (10, 11). Several recent studies have emphasized the critical role of intestinal flora dysregulation in the development and progression of AS, and have suggested that 60% of AS patients are associated with subclinical intestinal inflammation (12–14). This might be related to the imbalance of IL-17 or IL-23 cytokines caused by the activation a Th17-mediated immune response by intestinal dysbiosis (15). In contrast, the role of environmental factors in the etiology of AS is far from clear. One of the most popular theories presume that the onset of AS in susceptible individuals may be caused by infections (16), and that infections have the potential to modulate and attenuate immune responses.

The underlying pathogenic mechanisms for linking infections and AS involve changes in target cells and immune cells, and antigenic cross-reactions between microbial and host determinants (17). Infections might cause the quantitative reduction in specific T cells and the host defense defect against the infections that allows microbial antigens to reach the joint (18). The association between the infections and AS may be via IL-17 or C reactive protein levels that can induce inflammatory response (10, 19). In addition, certain microbial infections may reduce CD4+ T cells, and protein fragments released by dying CD4 lymphocytes may induce autoreactive CD8 lymphocytes (20). There is evidence of significantly elevated levels of IL-6 and TNF-a in AS patients, which might be caused by infections (21).

Numerous studies have investigated AS-related infections, including bacterial (10, 17, 18, 22), viral (17, 19, 20, 23), fungal (11), and those by microorganisms with sizes between those of bacteria and viruses (18, 24, 25). The infected sites include the respiratory (18, 24–26), immune (20, 23, 26, 27), digestive (10, 22, 26, 28), and genitourinary systems (19). However, there is no consensus on the association between infections and the risk of AS. To our knowledge, no systematic review and meta-analysis to date has investigated the effect of infections on the risk of AS. Therefore, in order to obtain a more convincing conclusion, this study aimed to review all the relevant studies and summarise the findings, in order to investigate the association between infections and AS.

The current study was developed according to the guidelines for the Meta-Analysis of Observational Studies in Epidemiology (MOOSE) (29) and Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA, Supplementary Table 1 ) (30). The protocol is presented in Supplementary Data .

To the best of our knowledge, this meta-analysis is the first study to investigate the association between infections and the risk of developing AS. The current meta-analysis included seven case-control studies with 1,296,239 participants and seven cohort studies with 7,618,524 participants. The results of this study confirm that infections play an important role in the risk of AS. Determining the effect of infections on AS would be beneficial for the identification of those who are at higher risk of AS as reasonable preventive intervention can be conducted for this population, enabling a far-reaching significance for the prevention of AS.

The results showed that infections are associated with an increased risk of AS in both the case-control and cohort studies. Our findings are consistent with most of the studies, including the four studies from Asia (19, 22, 24, 27), three studies from Europe (17, 28), and one study from North America (20). However, the study by Bartels et al. showed opposite results (10), in that previous Helicobacter pylori (H.pylori) infection may reduce the risk for developing AS. Another study found that H. pylori infection was eradicated in more than 80% of cases in the same cohort as that of Bartels et al. (35), indicating that H. pylori leaves a protective potential for the development of AS later in life when it is eradicated (10). Furthermore, the microbiota in the gastrointestinal tract changes after the eradication of H. pylori, which may have an impact on AS development (36, 37).

Our subgroup analyses showed that there was an association between infection and the risk of AS after adjusting for comorbidities in the cohort studies, which suggested that the comorbidities are a remarkably important confounding factor in cohort studies, and that we must control and adjust it. However, an association between infection and the risk of AS was found in the case-control studies without adjusted comorbidities. This is due to the nature of the case-control study design. As one of the matching factors of case-control studies, comorbidities may be matched in the design stage, cancelling the need to adjust for comorbidities in the statistical analysis stage.

With regards to the infection types, we did not observe that bacterial infections contribute to the risk of AS. In case-control studies, only the study by Keller et al. showed that there is an association between AS and a prior diagnosis of chronic periodontitis, which is characterized by an oral bacterial infection (22). This may be because rheumatic diseases and chronic periodontitis share pathogenic factors, including a dysfunction of inflammatory mechanisms and an imbalance of proinflammatory and anti-inflammatory cytokines (22, 38–40). In the cohort studies, the study by Bartels et al. showed that H. pylori may be a protective factor for AS (10). The study of Nielsen et al. also showed that bacterial infection is associated with the development of AS in the general population (17). More cohort studies are needed to verify whether bacterial infection causes AS. The result of the subgroup of other infection types was that other infection types are associated with AS in both the case-control and cohort studies. In our analysis, other types of infection included fungal, chlamydia, and mycoplasma. The pathogenesis of AS due to other types of infection is far from clear. For example, one study suggested that AS could be induced after exposure to Candida albicans through a T cell-driven model towards Th17 responses (11). Another study suggested that Mycoplasma pneumoniae has a significant impact on immune cells and the immune system of the host, including polyclonal activation of T and B cells and the secretion of related cytokines (24), leading to a breakdown of immune-tolerance. In addition, subgroup analysis indicated that viruses play an important role in the risk of AS in the cohort studies. One study suggests that viruses (e.g., human papillomavirus) might lead to inflammatory or immune-mediated disease by activating the pathogenic IL-23/IL-17 axis, resulting in elevated serum levels of Th17 cells, IL-17, and IL-23, and the imbalance of IL-17A/IL-23 cytokines (19). In the subgroup analysis of infection sites, we found that the infection of the immune system was significantly associated with the risk of AS in the case-control studies. Some immune organs, such as the tonsils, are involved in allergens tolerance by generating allergen-specific FOXP3⁺ regulatory T cells, suggesting that they are critical in the development of immune-tolerance (41). Some studies postulated that the alteration of immune tolerance in patients with tonsillitis might lead to the inflammatory disorders in autoimmune arthritis, including AS; therefore, tonsillitis might be aggravated by spondylitis, leading to the diagnosis of AS (26, 27). In addition, the higher risks of AS among infected people might be explained by HIV-induced antigen-driven immune responses (42), T cell imbalance (43), and molecular mimicry located between HIV protein and self-antigens (44). For the cohort studies, the infections in other sites were significantly associated with the risk of AS, which indicates that AS might be triggered by respiratory tract infections and genitourinary system infections. The pathogenesis of AS caused by the infection in the genitourinary system is mixed. In one of them, human papillomavirus in the genitourinary system might lead to AS by activating the IL-23/IL-17 axis (19). For the respiratory system, Klebsiella pneumoniae might lead to a decrease in the number of specific T cells, which could reflect an insufficient in the host’s defense against Klebsiella, thereby allowing AS to be affected by bacterial antigens that reach the joint (18). In our study, we found that some design types were meaningful and some were not for the same subgroups, which might be related to the small number of included articles or the large heterogeneity between the included studies.

This meta-analysis has the following strengths: pooled effect values were analysed according to the different study design, and we grouped the studies according to the types and sites of infection to determine whether these factors were associated with the risk of AS. This study included analytical epidemiological studies to determine the risk of AS, and the sample size was large. The included studies were adjusted for potential confounding variables, which improved the accuracy of risk estimates.

However, some limitations have affected the current study. First, although heterogeneity was explored via subgroup analysis, it was still very high, which may be related to age, sex distribution of participants, definition of infection, diagnosis of AS, etc. The subgroup analysis suggested that definition of infection was one of sources of heterogeneity in the cohort studies, which indicated that the possible disagreement between measurement methods might be a source of misclassification. But heterogeneity within subgroups remained high. In addition, although we extracted the definition of AS, only two articles declared that the diagnosis of AS was based on the Amor criteria (18, 25). As most of articles were retrospective, the International Classification of Diseases codes for the diagnoses of AS were based on records made by physicians and hospitals rather than a prospective clinical setting; thus, we could not set uniform criteria for the definition of AS, which also may result in heterogeneity. Second, this meta-analysis included only English- and Chinese-language articles, and qualified articles in other languages were not included in the analysis, which might have affected the pooled estimated value. Third, our pooled effect is affected by the study by Nielsen et al. in the cohort studies. However, the study by Nielsen et al. has the largest weight when synthesizing RRs across studies, because of its large sample size, narrow confidence interval, and high quality. When the study by Nielsen et al. was omitted in the sensitivity analysis, the pooled effect estimate was affected by some low-quality studies due to the increased weights. Thus, more large cohort studies are recommended in the future to assess the impact of infections on the risk of AS.

In conclusion, this meta-analysis confirms that there is an association between infections and an increased risk of AS, although the included studies suffered from high heterogeneity. As much as the mechanism of infection and the effect of bacterial and viral infections on AS has not yet been determined, further studies, particularly more higher quality prospective cohort studies and case-control studies, are required to verify that there is a true cause-and-effect relationship between infections and the risk of developing AS.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

GD and XZ designed the study protocol. XZ, XS, and JY conducted the literature search. XZ, AZ, and GD retrieved and selected the articles. LT, YC, and ZhengS conducted data extraction. XZ and ZheS performed the statistical analysis of the data. XZ, ZheS, and GD wrote the manuscript draft. GD and ZhengS supervised the study. All authors contributed to the article and approved the submitted version.

This work was supported by the Natural Science Foundation of Shandong Province for the General Program (Grant No. ZR2020MH339). The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.