Rotavirus-Associated Hospitalization in Children With Subsequent Autoimmune Disease

This cohort study assesses the association of rotavirus infection with the risk of subsequent autoimmune disease in children and adolescents in South Korea.


Introduction
Rotavirus infection is a leading cause of mortality and morbidity in infants and young children, and these infections are also linked to autoimmunity. [1][2][3] While some individuals recover without complications, others develop manifestations affecting various organs, and viral dissemination may occur. [4][5][6][7][8] Moreover, some infected children experience extraintestinal complications that influence specific organs or whole body systems. Several lines of evidence suggest that rotavirus infection can trigger autoimmune diseases.
There are 9 species of rotavirus, each with a genome comprising double-stranded RNA segments encoding structural and nonstructural proteins. Vaccinations have reduced the prevalence of rotavirus disease, but postinfection illnesses persist. There is also evidence linking rotavirus infection to autoimmune diseases, such as celiac disease, 4,6,9 type 1 diabetes, 7,10,11 pancreatitis, 12 and neuropathy. 13,14 However, previous studies had limitations in terms of controls, 4,6,11 small sample sizes, [12][13][14] and study design. 7,10,13 While there is supporting evidence from humans and animals, clinical epidemiology studies have only provided limited support. We performed a systematic literature search and analyzed a nationwide population-based database of nearly 2 million individuals to investigate the association between rotavirus-associated hospitalization and subsequent autoimmune disease, while accounting for confounding factors.

Methods
The study received approval from the institutional review board and ethics committees of Hallym University Kangnam Sacred Heart Hospital and the Korea National Institute for Bioethics Policy.
Informed consent was waived by these committees because the study used deidentified data that were publicly available in Korea. We followed the Reporting of Studies Conducted Using Observational Routinely-Collected Data (RECORD).

Systematic Search and Identification of Published Literature
We performed a literature search in PubMed using the following Medical Subject Heading (MeSH) terms: rotavirus, rotavirus infection, and autoimmune disease. All relevant studies conducted after 1999 were eligible for inclusion in the review. Studies were included if they (1) examined the incidence of any autoimmune disease-related hospitalizations or outpatient visits after rotavirus infection; (2) obtained fecal samples to test for rotavirus in patients with autoimmune diseases; and (3) examined the distribution of specific types of rotavirus related to autoimmunity in vivo or in vitro.
We included all studies conducted among children up to 18 years of age. Editorials, narrative, or subject reviews, and symposium proceedings were excluded. Studies that investigated rotavirus vaccination or were conducted before 1990 were excluded. human studies reporting rotavirus-associated autoimmune diseases or autoimmunity (eTable 1 in Supplement 1) and 4 studies were nonhuman studies (eTable 2 in Supplement 1). Rotavirus infection was related to various groups of autoimmune diseases, including diseases of the digestive system (eg, inflammatory bowel disease, 20 celiac disease, 4,6 nonceliac gluten sensitivity [NCGS], 3 pancreatitis 21 ) and endocrine system (eg, type 1 diabetes 11,21,22  Tenth Revision [ICD-10]), prescriptions, and procedures. Children who required hospitalization due to a rotavirus infection were defined as the exposed group (Figure 1). The date of the diagnosis was set as the index date, and all included participants were younger than 19 years on that date. Patients were excluded if they had a history of autoimmune disease before the index date or had records with conflicting information. The exposed group was then compared with a population-matched unexposed group that was established using 1:1 matching by density sampling. Each participant in the unexposed group was randomly and individually matched by birth year and sex, and had no history of rotavirus-associated hospitalization or autoimmune disease at the date of diagnosis of the index patient.

Exposure
The specific exposure of interest was hospitalization for rotavirus infection as the main or secondary diagnosis, identified using an ICD-10 code of A08.0X (rotavirus enteritis) (eTable 3 in Supplement 1).

Follow-Up
All participants were followed up from the index date until the first diagnosis of an autoimmune disease, death, or the end of the study (December 31, 2017), whichever happened first. The follow-up of individuals in the unexposed group involved assessing rotavirus-associated hospitalization at a later time. If an individual experienced a rotavirus-associated hospitalization during the follow-up period, they were censored and transferred to the exposed group. The first year of follow-up was excluded from the analyses to reduce the probabilities of reverse causality and surveillance bias.

Autoimmune Diseases
Information on autoimmune diseases was from the NHIS database and corresponding ICD-10 codes (eTable 3 in Supplement 1). 17 In total, 41 autoimmune diseases were considered because previous studies suggested these were related to rotavirus infection. The outcome and time frame were based on previous human and nonhuman studies (eTables 1, 2, 4, and 5 in Supplement 1).  15 The analysis considered the latest available information before the index date. Medical resource usage, including inpatient and outpatient visits, was collected during follow-up.

Covariates
Additionally, antibiotic, systemic steroid use, and hospitalization duration were assessed.

Statistical Analysis
We used a conditional Cox proportional hazard model to estimate hazard ratios (HRs) and 95% CIs of autoimmune disease in relation to rotavirus-associated hospitalization using time after the index date as the underlying time scale. The visual inspection of covariates using Schoenfeld residuals in the Cox proportional hazard regression indicated no violations of the assumption. The analysis did not examine a specific autoimmune disease with fewer than 100 cases separately, but incorporated it in the HR calculations for its corresponding major group. All analyses were stratified by matching identifiers and adjusted for birth residence, household income, and perinatal history. HRs were separately calculated for sex, age at the index date (24 months or less or more than 24 months), calendar year of birth (2002-2003 or 2004-2005), calendar year of birth at the index date (2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009) or 2010-2017), birth residence (Seoul/metropolitan, city, or rural area), household income (low or high), season of birth (spring, summer, fall, or winter), disorders related to any perinatal status (yes or no), history of asthma (yes or no), history of hospital admission (yes or no), and number of outpatient visits (15 or fewer or more than 15, based on the median) during the first year after study entry. Data collected during the first month after study entry were not considered, because participants in the exposed group used more health care services during the first month after diagnosis of infection. The first year of follow-up was excluded from all analyses. Differences in HRs were assessed by introducing an interaction term into the Cox models. The absolute rate differences and 95% CIs were determined for all associations. The difference in log (HR) between strata was used to calculate the z score and P value for this comparison. We examined the risk of a diagnosis of any autoimmune disease as 9 major groups of autoimmune diseases (inflammatory arthritis; vasculitis; connective tissue disorders; and diseases of the digestive, endocrine, skin, hematologic, and nervous systems, among others).
Sensitivity analyses were performed to examine the association using a more stringent definition of the exposure (rotavirus enteritis as the main diagnosis of hospitalizations were included in the exposure group, and cases of rotavirus enteritis as a secondary diagnosis were not considered as the exposure group), 2 or more autoimmune syndromes or 3 or more autoimmune syndromes, individual autoimmune diseases, any of the major groups of autoimmune diseases (eTable 2 in

JAMA Network Open | Pediatrics
Rotavirus-Associated Hospitalization in Children With Subsequent Autoimmune Disease  . In these analyses, the cumulative incidence curves of autoimmune diseases among participants with more than 1 year of follow-up were analyzed. All statistical analyses were conducted using SAS version 9.4 (SAS Institute), and a 2-sided P < .05 was considered statistically significant. there were 86 517 individuals in the exposed group, and the same number in the unexposed group after 1:1 incidence density sampling. The median (IQR) age at rotavirus-associated hospitalization was 1.5 (0.92-2.67) years and 49 072 (57.0%) of all patients were male ( Table 1). Compared with the unexposed group, the exposed group had more inpatient and outpatient hospital visits during the first year after study entry and had increased use of prescription medications, especially antibiotics and systemic steroids. During the mean (SD) follow-up time of 12.05 years (3.17), we identified newly diagnosed autoimmune diseases in 7589 of 86 517 individuals (8.8%) in the exposed group (incidence rate, 73.1 per 10 000 person-years) and in 6095 of 86 517 individuals (7.0%) in the a Missing information in weighted data: control group, 1; exposed group, 13.

Results
b Income was classified into 2 groups based on the regional median income. Missing information in weighted data: control group, 2528; exposed group, 3234. c The first month after study entry was removed from this calculation.   After controlling for confounders, the exposed group had a significantly greater risk for autoimmune disease (HR, 1.24; 95% CI, 1.19-1.28) ( Table 2). Rotavirus-associated hospitalization was associated with higher rates for each of the 9 major groups of autoimmune diseases (Figure 2). The population-matched analysis indicated these associations were notable for the following groups: We performed a sensitivity analysis to examine the robustness of the results from the primary analysis ( Table 3). The results indicated the association of rotavirus-associated hospitalization with autoimmune disease was stronger for those with 2 or more autoimmune syndromes (HR, 1.51; 95%

JAMA Network Open | Pediatrics
CI, 1.31-1.73) and with 3 or more autoimmune syndromes (HR, 1.79; 95% CI, 1.18-2.72). This sensitivity analysis also indicated greater risk when excluding cases diagnosed 2 years after study entry or 5 years after study entry, when exposure was defined more stringently (rotavirus enteritis as the main diagnosis of hospitalization), and when the outcome was defined by the diagnosis of an autoimmune h The first month after study entry was removed from this calculation. Cox models were stratified by matching identifiers (birth year and sex) and adjusted for birth residence (Seoul/metropolitan, city, or rural area), household income (low, middle, or high), and perinatal history (disorder related to length of gestation and fetal growth, birth trauma, infections specific to the perinatal period, congenital malformation/ deformation, and chromosomal abnormalities). The first year of follow-up was excluded from all analyses. HR indicates hazard ratio; IgA, immunoglobin A. a Autoimmune diseases with fewer than 100 cases were not analyzed separately, but were included in calculations of HRs for the main categories.

JAMA Network Open | Pediatrics
b Includes autoimmune vascular disorders, such as polyarteritis nodosa and Churg-Strauss syndrome.

Discussion
The results of this large nationwide cohort study of Korean children demonstrated that exposure to rotavirus-associated hospitalization increased the risk for subsequent autoimmune disease. This association remained significant after further sensitivity analysis, in groups with different demographic and perinatal status, and in subgroup analysis that considered medication records.
Additionally, our findings demonstrated that patients with longer hospitalization and more frequent hospitalization events due to rotavirus infection had increased risk of autoimmune disease. To the best of our knowledge, this is the first study to examine the association of rotavirus infection with major autoimmune diseases in children using a population-based comparison. The findings of our study provide important insights into the possible etiological relationship of rotavirus infection and autoimmune diseases. In addition, the greater incidence of autoimmune disease in our exposed group (73.1 vs 58.0 per 10 000 person-years, respectively) suggests the importance of awareness for children who experience rotavirus infections at an early age. Overall, our finding suggests that rotavirus-associated hospitalization is associated with increased risk for developing autoimmune diseases that affect different organs and for developing multiple autoimmune syndromes. Previous studies examined the possible mechanism by which rotavirus infection triggered autoimmune processes, and proposed several possible explanations. Although studies of antigenic mimicry suggest that humoral and cellular immune responses may be responsible for the autoimmunity reported in studies of humans, animals, and in vitro models. 36,[40][41][42] Gastrointestinal infections from rotavirus can trigger immune responses distinct from those caused by food antigens because of oral tolerance. Gastrointestinal induction of autoimmune responses may disrupt oral tolerance and cause a type 1 T helper cell immune response in predisposed individuals.
The high similarity of rotavirus surface proteins known to have gastrointestinal pathogenicity (eg, VP4 and VP7) with bovine milk, casein, and human T-cell epitopes may contribute to increased autoimmunity. 22,36 A recent study proposed the bystander activation model to explain autoimmune provocation following rotavirus infection, 1 suggesting that inflammation during the infection reduces self-tolerance and activates autoimmune cells. Higher antibiotic usage was observed in the exposed group, which is also a considered as a potential risk factor for childhood autoimmune diseases. 43 The strengths of our study are that we used a population-based cohort, complete follow-up of nearly 50 000 patients with a rotavirus-associated hospitalization for nearly 10 years, and a large sample size enabling subgroup analysis and sensitivity analysis. We also considered age at exposure and multiple outcome measures. Detailed sociodemographic and medical data enabled effective control for numerous potential confounders.

Limitations
This study has limitations. For example, there may have been some surveillance bias. To address this issue, we performed sensitivity analyses that considered extended lag times, different definitions for exposure and outcome, and adjusted for estimated levels of medical surveillance. Additionally, we relied on registered data, which may have led to underestimating cases if infection occurred as a comorbidity, if there were delays in pathogen identification, or if there were missed diagnoses.
Furthermore, because this was an observational study, we cannot make causal inferences regarding the relationship of rotavirus-associated hospitalization with autoimmune disease. The potential links of other unmeasured factors, such as unreported health conditions, alterations in health-related behaviors, and the use of certain medications, should be addressed in future studies. Moreover, we were not able assess the association between rotavirus-associated hospitalization with autoimmune diseases later in life.
In addition, in the present data set, we were unable to assess whether the matched unexposed group experienced hospitalization due to other causes. Future studies investigating the underlying contributing factors are warranted.

Conclusions
In this nationwide cohort study, rotavirus-associated hospitalization was associated with an increased risk for childhood autoimmune disease overall and multiple specific autoimmune diseases.
Clinicians should be aware of the increased predisposition for autoimmune disease in individuals who experienced rotavirus-associated hospitalization. Further studies are needed to understand the underlying mechanisms.