Bidirectional association between infectious gastroenteritis and inflammatory bowel disease: a population-based study

Inflammatory bowel disease (IBD), including ulcerative colitis (UC), Crohn’s disease (CD), and unclassified IBD, results from genetic susceptibility and environmental factors that cause an imbalance in immune response and gut microbiota [1]. IBD has been more prevalent in Western countries. However, IBD incidence has been rising in newly industrialized countries in the past decades, such as in Africa, South America, and Asia including Taiwan (annual percentage change: 4.8% and 4.0% for UC and CD, respectively) [2]. Its increasing morbidity also worsens social economic burden [3‐5].

Infectious gastroenteritis (IGE), which causes dysbiosis, plays an important role in the pathogenesis of IBD [6]. An association of increased IBD risk with enteric infections has been reported in high IBD prevalence populations [7‐13], particularly with bacterial gastroenteritis [8‐11, 13]. An increasing number of previous gastroenteritis episodes may further elevate IBD risk, and the relevant time interval between an episode of gastroenteritis and IBD development may be greater than 10 years [15].

Theoretically, the use of glucocorticoids, immunosuppressants, and biological agents in diagnosed IBD patients may increase the risk of IGE episode [16, 17]. The prevalence of intestinal superinfections by enteropathogens among patients with active IBD ranged from 6% to 26.8% [18, 19]. Clostridium difficile, Campylobacter species, E.coli, and Salmonella species have been commonly reported in the flare-up IBD in previous studies [18‐26]. However, disparities in methodology, including limited sample sizes, the lack of a comparison group, and varied proportions of CD and UC in the population, have limited a valid conclusion to elucidate the relationship between intestinal infections and diagnosed IBD. In addition, population-based studies that compare the IGE incidences before and after IBD diagnosis are warranted. Knowledge of the bidirectional relationship between IGE and IBD might help us understand the entire disease course and elucidate the intertwined causal-relationship between IGE and IBD

To our best knowledge, no previous study has retrospectively investigated the role of IGE in IBD development and, at the same time, prospectively examined the subsequent risk of IGE after IBD diagnosis. Therefore, we used a population-based claims dataset to determine whether patients with incident IBD were more likely to experience antecedent IGE episodes and have an even greater risk of developing subsequent IGE after IBD diagnosis

We used the national health claims database to identify IBD cases in Taiwan, most of whom had UC (71.4%). To our best knowledge, this is the first longitudinal study to suggest a bidirectional association between IGE and IBD. IBD patients had higher odds of experiencing IGE within five years before IBD diagnosis than controls, the OR increased with an increased cumulative frequency of IGE episodes. However, IBD cases had an even higher IRR of subsequent IGE after IBD diagnosis.

We found that the excess incidence and mean number of IGE episodes in IBD cases were magnified after IBD diagnosis. Compared with the controls, the IRR increased from 2.42 before IBD diagnosis to 5.74 after IBD diagnosis. In the aftermath of IBD, various factors such as gut dysbiosis, compromised repairment of the gut epithelial barrier, dysregulated immune responses, and persistent intestinal inflammation may collectively contribute to the rising incidence of IGE [28]. Retrospective cohort studies found that patients with IBD have a higher likelihood of experiencing various common infections, including gastrointestinal infections [28‐30]. Dysbiosis of intestinal microbiota which had been described in the patients with IBD may activate innate immune system and is associated with mucosa damages [31]. Impaired mucosal barrier and increased paracellular permeability may enable penetration by bacteria and increase the risks of intestinal infections [32, 33]. In addition, the use of steroid and multiple immunosuppressant agents on the vulnerable mucosa of the chronic inflamed bowel may increase the risk of IGE after IBD diagnosis [16, 17]. Several studies have investigated the prognostic role of specific pathogens identified in IBD patients, particularly Clostridium difficile, Cytomegalovirus, Campylobacter spp., and other pathogenic bacteria [18, 19, 21, 24, 26]. However, most of these studies were cross-sectional and only included inpatients during IBD flare-up, or lacked a healthy control group, limiting them from providing a comprehensive picture of the excessive IGE incidence before and after IBD diagnosis.

Our study revealed an association between previous IGE episodes and IBD development. A history of IGE occurring within 1 year before the index date was significantly associated with an increased IBD risk. This association diminished after excluding IGE within 6 months before IBD diagnosis, at which point the diagnosed IGE might be the initial presentation of IBD, according to the estimation from CHAD trial [12]. The association (aOR 1.89) between IGE and IBD diminished over time but remained statistically significant through the five-years before IBD diagnosis. A similar trend was observed in both the CD and UC cases. The study of Axelrad et al. had similar findings which demonstrated a previous gastroenteritis significantly associated with IBD at ≥10 years after the episodes (aOR 1.26, 95% CI 1.19–1.33) [15].

Our study also revealed that IBD risk increased with cumulative IGE episodes during > 6 months to 5 years before the index date. Compared to the control group, aOR increased from 1.64 to 4.57 in people experiencing 1 to ≥5 IGE episodes. These findings are similar to those from the study of Axelrad et al. [15], which reported an aOR of 1.61 to 4.64 corresponding to IGE frequency from 1 to > 4 before IBD diagnosis. Previous epidemiologic studies have linked specific intestinal pathogens to the increased risk of IBD development, the most reported included Salmonella, Campylobacter, and Clostridium difficile, while specific pathogens could only be identified in quite limited IGE episodes in IBD cases [7, 8, 10, 15], and potential detection bias resulted from more stool tests in IBD group may compromize the causal relationships between sepecific pathogens and IBD development [14].

Intestinal microbiota are important and necessary for humans, as they participate in the process of intestinal homeostasis, intestinal immune development, host metabolism, and protection of the host from pathogenic infection via colonization resistance [34‐36]. The coexistence of intestinal microbiota is maintained by various immune mechanisms, including intestinal mucus secretion, immunoglobulin A, and antimicrobial peptides, which shape the gut microbiota and prevent direct contact with the epithelium [37]. IGE disrupt the equilibrium of the gut microbiota and elicit a robust immune response against invasive pathogens. As a consequence, this immune reaction can disturb the microbial balance, resulting in dysbiosis that has been associated with the initiation and progression of IBD [38]. Numerous animal models have indicated that colitis led to a compromised immune response (loss of tolerance to commensals and microbiota-specific T cell differentiation) and chronic inflammation of the intestine, corroborating the notion of dysregulated immunity as a key driver of IBD [39, 40]. The imbalance between commensal and pathogenic microorganisms may promote IBD development [34, 35]. In-vitro studies have suggested the role of enteric infections in promoting gastrointestinal microbial dysbiosis and subsequently causing characteristics of intestinal inflammation of IBD [41‐44].

The present study had certain limitations. First, this study lacked sufficient culture data or specific pathogen diagnoses for most gastroenteritis events, which may limit our study to further associate IBD risk with specific pathogens and distinguish between IGE episodes and IBD-related symptoms. Second, there are also some limitations that may cause misclassification of IGE in our study. The IGE diagnosis was based on ICD codes in our study. Without information on the clinical presentations of each IGE episode, there may be misclassification of IGE in both the IBD group and the non-IBD group. In addition to that, the number of IGE cases may be underestimated because many people with IGE did not seek medical care. Moreover, many people who were eventually diagnosed with IBD initially presented with some digestive symptoms and were diagnosed as having IGE. Therefore, we compared the frequency of IGE between two groups in several time-windows as early as possible before IBD was diagnosed (i.e., IGE occurred in > 6 months to 1 year, > 1–3 years, > 3–5 years before the index date). This minimized the possibility of unequal frequency of seeking medical care for IGE between IBD cases and controls, it also largely excluded the IGE-like symptoms which were essentially early manifestations of IBD. As a result, the non-differential misclassification of IGE between the two groups could only lead to the association between IBD and IGE being underestimated and towards the null.

In conclusion, our study suggested a bidirectional association between IGE and IBD. The patients with incident IBD had experienced more IGE previously and were more likely to have subsequent IGE after IBD development. These findings highlight the potential impact of IGE and alert physicians to be aware of the increased risk of subsequent IGE in IBD patients, especially in newly industrialized countries where IBD incidence has been markedly increased in the past decades.