Respiratory Viral Infections and the Risk of Rheumatoid Arthritis

Young Bin Joo; Youn-Hee Lim; Ki-Jo Kim; Kyung-Su Park; Yune-Jung Park


Arthritis Res Ther. 2019;21(199) 

In This Article


The idea of infections acting as a trigger for the development of RA has been suggested for a quite long time without much clarification. In this study, the number of weekly incident RA cases exhibited an inverted U-shaped seasonal tendency throughout each year. In addition, the detection rate for ambient respiratory viral infections in the population was associated with an increased number of incident RA cases, which suggests a possible role for respiratory infections as a trigger for the development of RA.

Seasonal tendencies of RA onset or relapse have been evaluated in a small number of studies. In an Italian study including 44 RA patients, there was no seasonal tendency.[29] In the UK between 1957 and 1963, 43 of the 100 patients had reported that RA occurred during the winter.[7] In a more recent study using data from the year 2000 in Israel, RA relapse occurred mostly during the summer.[8] These studies were conducted many years ago with small study groups. Advances in the understanding of RA and in diagnostic tools have since changed the reported characteristics of patients with RA. Thus, there is a need to reevaluate seasonal patterns of RA using recent data derived from large groups.

In the present study, population-based incident RA diagnoses in 2012 and 2013 were used, which is more reliable. Notably, however, observations pertaining to the seasonality of RA in the study should be interpreted with caution. The index date of incident RA is not the date of onset, rather it is the date that RA is diagnosed by physicians. The onset of RA may precede its diagnosis by several weeks or months.[30] Nevertheless, a distinct seasonal pattern of incident RA determined based on the date of diagnosis could be explained by the role of respiratory viral infections in inflammation. Respiratory viral infections may play a role in the exacerbation of inflammation involving joints in patients with subclinical or early-stage RA, prompting patients to visit the hospital.

Several mechanisms of virus-induced initiating or triggering of autoimmune disease have been suggested:[31,32] (1) "Molecular mimicry" is the most widely proposed mechanism and occurs when a virus antigen mimics a host antigen and activates cross-reactive T cells. (2) "Epitope spreading" is another potential mechanism. Tissue damage resulting from virus-specific T cell activation or direct virus-mediated host tissue destruction causes de novo activation of autoreactive T cells and releases self-antigens into the inflammatory environment. (3) "Bystander activation" is the activation of autoreactive T cells as a result of the release of cytokines during a virus-targeted immune response. (4) Encrypted host antigens are released from certain tissues during virus-targeted tissue damage. (5) "Superantigens" activate a wide range of nonspecific T cell clones regardless of their specificity. In this study, however, we could only reveal the association between respiratory viruses and incident RA, but not investigate the possible mechanisms as this is an ecological study.

In our study, parainfluenza, coronavirus, and metapneumovirus were significantly associated with the number of incident RA. The patterns of virus seasonality, the severity or virulence of infections, and the parameters such as peak age of viral infections have differential effects on the associations between each respiratory virus and incident RA. The similarity of the three significant respiratory viruses in the study has not been elucidated by one of the factors described above. More complex interactions among these virus factors or other environmental factors may be involved.

Regarding the confounding factors, we adjusted for meteorological factors and air pollution data, which could affect the seasonality of the number of incident RA cases. Considering the unmeasured confounders associated with seasonality, we also adjusted the time trend in the generalized linear model to avoid overestimation of the effects of respiratory viral infection on the number of incident RA cases. Nevertheless, other factors that could explain why the number of incident RA cases increased in July could exist. This point indicates we should be cautious in interpreting these results.

It was interesting that the effects of ambient respiratory viral infections in the population on the number of incident RA differed according to sex, although men and women showed the same seasonal pattern for the number of incident RA cases. Only the parainfluenza virus was significant in men while all three viruses were significant in women. The relatively small sample size of men might have affected this difference. Also, a stronger immune response in women could affect this difference.[33] Stronger immunity to pathogens in women is associated with lower viral loads and lower prevalence of infections than in men, but it may also be associated with increased severity of disease symptoms.[34] For example, HIV-positive women tend to exhibit less circulating HIV RNA than HIV-positive men, but they are reportedly at a 1.6-fold higher risk of developing AIDS.[35] Fatality following exposure to pathogenic influenza A viruses is higher in women.[36,37] It may be that respiratory viral infections result in stronger inflammation in women than in men, and trigger stronger immune responses, and this results in a differential proportion of incident RA in women and men. However, the virus data in relation to sex was not available for analysis in the present study.

Viral arthritis is distinct from autoimmune disease-associated polyarthritis. Viral arthritis is usually self-limiting, and treatment with immunosuppressants is usually not required.[38] To exclude viral arthritis in the present study, only individuals who underwent treatment for 3 years were included. Additionally, treatment was required to include immunosuppressants. Thus, the seasonal tendency of RA diagnosis and the association between ambient respiratory viral infections in the population and incident RA in the study are unlikely to be an effect of transient viral arthritis.

The current study had some limitations. A causal association between ambient respiratory viral infections in the population and incident RA could not be proven because this is an ecological study where it is not known whether the individuals with RA also had viral infections. A clear causal association between respiratory viral infections and RA development remains to be proven via future individual-level data study. In addition, many respiratory viral infections, especially in the case of mild symptoms, would not have been detected at hospitals, hence would not have been listed in the nationwide database. Further study including all the patients with acute respiratory infection symptoms, not based on PCR, confirmed infection would also be valuable when investigating possible associations between respiratory infections and RA development.