This is the first nationwide population-based study to investigate the relationship between EV infection and the risk of ADs. Our results indicated that, compared with children without EV infection, children with EV infection had a higher risk of Henoch-Schönlein purpura, type 1 diabetes mellitus, ankylosing spondylitis, and systemic vasculitis.
The mechanism through which EV increases the risk of individual ADs is poorly understood. However, several potential mechanisms have been investigated. Some studies have demonstrated that EV infections are associated with the appearance of beta-cell autoantibodies. Insulin-producing beta cells destroyed by the immune system lead to the manifestation of clinical type 1 diabetes mellitus.30, 31 Type 1 diabetes mellitus is considered a typical organ-specific autoimmune disease. The risk of the disease is modulated by genetic and environmental factors, with both required to trigger the process. Initial genetic susceptibility, mainly genetic coding for human leukocyte antigens, and several environmental factors, including infection, are considered to be part of several mechanisms underlying the development of autoimmune disease. Therefore, preventing exposure to infections is vital for avoiding ADs.32
Several studies have linked viral or bacterial infections with the onset of autoimmune disorders. RA,33–35 MS,36–39 type 1 diabetes mellitus,40,41 antiphospholipid syndrome,42–46 SLE,47–51 polymyositis, dermatomyositis,52–55 systemic sclerosis,56–58 Sjögren’s syndrome,33,34,59 and vasculitis60–62 were associated with viral or bacterial infection. Severe infection with complications increased the risk of autoimmune disorder. Enterovirus infection was associated with an increased risk of attention-deficit/hyperactivity disorder,20 allergic rhinitis and atopic dermatitis,21 tic disorders,22 and depression23 and a lower risk of leukemia.64 Coxsackievirus, which is an EV, was associated with type 1 diabetes mellitus,13–19 polymyositis, dermatomyositis,53,54 and Sjögren’s syndrome.58 Coxsackievirus B3 infection induces cytokine-regulation, which leads to the progression of viral myocarditis to autoimmune myocarditis.64–67 Hammerstad et al. detected EV RNA and EV protein in the thyroid tissue of patients with Graves' disease but not in individuals in a control group.68
Our study used the NHIRD database for a large-scale retrospective study to investigate the association between EV infection and ADs. Our results indicated that EV infection in children increased the risk of Henoch-Schönlein purpura, ankylosing spondylitis, and systemic vasculitis. Type 1 diabetes mellitus, which was often mention in other articles,13–19 had an IRR 1.18 (P > 0.05, 0.95–1.46) and an aHR of 1.30 (P < 0.05, 1.05–1.62). However, Addison’s disease, autoimmune hemolytic anemia, Graves’ disease, Hashimoto’s thyroiditis, myasthenia gravis, inflammatory bowel disease, psoriasis, polymyositis, dermatomyositis, RA, Sjögren’s syndrome, SLE, and systemic sclerosis had no significant association between EV infection and increased risk of ADs.
Strengths and limitations
The strengths of the present study include its large sample size, large validation cohort, and verified long-term data on the patient’s medication. However, our study has some limitations. First, although the NHI Bureau routinely and randomly checks patient charts to ensure the quality of claims from medical institutions, the possibility of miscoding or misclassification could not be completely ruled out. For example, juvenile ankylosing spondylitis (juvenile AS) indicate children (1) who fulfill the modified New York criteria before they are 16 years old or (2) who only have symptoms before they are 16 years old and fulfill the modified New York criteria after they are 16 years old.69 Due to the lack of juvenile AS code in the health insurance system, coding error may have lead patients with juvenile AS to be included as AS (720.0, 720.2, 720.8,720.9, See Supplementary Table 1) in our study. Second, the relationship between disease activity, comorbidity, and the severity of EV infection could not be analyzed and may have interfered with our results. Third, several unmeasured confounders, including body mass index, congenital disease, and other over-the-counter drug use, which are associated with EV infection, were not included in our database. Fourth, we could not contact patients directly about their drug use because of their data were anonymized. We presumed that all prescribed medications were taken by patients as prescribed, which may give rise to an overestimation of the actual ingested dosage. Fifth, laboratory and clinical data were not readily available in the administrative database. We had no method to evaluate the exact inflammatory level or to compare the severity of EV infection. Recurrent infections and infection severity may produce differences of ADs risk. Finally, because of limited follow-up, children who have been infected with EVs may only encounter the more common ADs in their age group, such as Henoch-Schönlein purpura, type 1 diabetes mellitus, AS, and systemic vasculitis. Sjögren's syndrome, RA, and systemic sclerosis are not prevalent in children, which may lessen the findings of our observational study and require longer-term follow-up.
In conclusion, EV infection in children increases their risk of Henoch-Schönlein purpura, type 1 diabetes mellitus, ankylosing spondylitis, and systemic vasculitis. Further investigations should clarify the effects of EV infection on the immune system in humans.