Article Text
PDF
Abstract
Objective The possible aetiological link between appendicitis and UC remains unclear. In order to investigate the hereditary component of the association, we studied the risk of UC in family members of individuals with appendicitis.
Design A cohort of 7.1 million individuals was established by linkage of national registers in Denmark with data on kinship and diagnoses of appendicitis and UC. Poisson regression models were used to calculate first hospital contact rate ratios (RR) for UC with 95% CIs between individuals with or without relatives with a history of appendicitis.
Results During 174 million person-years of follow-up between 1977 and 2011, a total of 190 004 cohort members developed appendicitis and 45 202 developed UC. Individuals having a first-degree relative with appendicitis before age 20 years had significantly reduced risk of UC (RR 0.90; 95% CI 0.86 to 0.95); this association was stronger in individuals with a family predisposition to UC (RR 0.66; 95% CI 0.51 to 0.83).
Conclusions Individuals with a first-degree relative diagnosed with appendicitis before age 20 years are at reduced risk of UC, particularly when there is a family predisposition to UC. Our findings question a previously hypothesised direct protective influence of appendicitis on inflammation of the large bowel. Rather, genetic or environmental factors linked to an increased risk of appendicitis while being protective against UC may explain the repeatedly reported reduced relative risk of UC in individuals with a history of appendicitis.
- ULCERATIVE COLITIS
- APPENDICITIS
- EPIDEMIOLOGY
Statistics from Altmetric.com
Significance of this study
What is already known on this subject?
An inverse association between early appendicitis (before age 20 years) and subsequent risk of UC in the same individual is well established.
It remains unclear whether the association is explained by direct effects of early appendicitis on the large bowel mucosa or whether it is due to common genetic or environmental factors that increase the risk of early appendicitis while at the same time conveying protection against UC.
What are the new findings?
The study shows that the patients with early appendicitis are at low risk of subsequent UC themselves. The reduced risk is also present in first-degree relatives of patients with early appendicitis.
The reduced risk of UC in first-degree relatives of patients with early appendicitis was most pronounced in those with a family predisposition to UC.
How might it impact on clinical practice in the foreseeable future?
The findings are not compatible with a direct mucosal role of appendicitis. Rather, genetic or environmental factors linked to an increased risk of appendicitis while being protective against UC may explain the repeatedly reported reduced relative risk of UC in individuals with a history of appendicitis.
Based on the findings, it seems unlikely that appendectomy would improve the disease course in patients with existing UC or protect against UC in family members to patients with UC.
Introduction
UC is a chronic, autoimmune condition characterised by inflammation of the large bowel. The aetiology is multifactorial and involves a complex interaction between genetic and environmental factors, but the complete picture remains uncertain.1 An inverse association between appendicitis and UC has been established in both case–control studies2–7 and large prospective cohort studies,8 ,9 but the underlying mechanism has remained a mystery.10 According to findings in cohort studies, the reduced risk of UC appears to be confined to individuals who have undergone appendectomy before the age of 20 years, and the association has been found to be due to appendicitis, not to the appendectomy per se.8
Attempts to increase our knowledge about the underlying biological mechanisms behind the inverse association are ongoing, because they may improve our understanding of the aetiology of UC.8 ,11 ,12 So far, it remains unexplored whether the association reflects a protective effect of appendicitis per se against future development of UC or, alternatively, whether it reflects the influence of genetic or environmental factors that increase the risk of appendicitis while at the same time conveying protection against UC.
The aim of the present study was to examine whether the reduced risk of UC was present only in individuals with a history of appendicitis, or also in relatives of individuals with such a history, and thereby examine whether shared genes and/or environmental factors could be the underlying mechanism behind the association.
Materials and methods
Data sources
The study population was established from data obtained in the Danish Civil Registration System,13 a national database containing continuously updated kinship information, demographic variables and vital status on all Danish citizens since April 1968. A unique personal 10-digit identification number provided to each Danish resident at birth or immigration permits tracking of individual-level information over time and accurate linkage between nationwide health registries. The National Patient Registry,14 which contains information on all hospitalisations in Denmark since 1977 and all outpatient visits and emergency room contacts since 1995, provided information on diagnoses of appendectomy, appendicitis and UC. Diagnoses of UC and appendicitis were based on the International Classification of Diseases (ICD) versions 8 (period 1977–1993) and 10 (period 1994–2011), whereas operation codes for appendectomy were based the Danish Classification of Surgical Procedures and Therapies (period 1977–1995) and the Danish version of the Nordic Medico-Statistical Committee (period 1996–2011). Data on kinship were obtained from the Danish Family Relations Database, a database that acts as a virtual pedigree tool based on parent–child links registered in the Danish Civil Registration System. Nearly all individuals born in Denmark since the early 1950s have known links to their parents, thus allowing identification of parents, siblings and half-siblings.
Study cohort, exposure and outcome
The cohort consisted of all individuals living in Denmark between 1 January 1977 and 31 December 2011. Diagnoses of UC (ICD-8: 56319 and 56904; ICD-10: DK51), appendicitis (ICD-8: 540, 54199; ICD-10: DK35-DK37) and appendectomy (1977–1995: 43000, 43001; 1996–2011: KJEA00, KJEA01, KJEA10) were identified in the National Patient Registry. To increase diagnostic specificity, we only considered patients as having appendicitis if they had both an operation code for appendectomy and a diagnosis of appendicitis. Because the association with appendicitis differs significantly between Crohn's disease and UC, all patients with a diagnostic code for Crohn's disease (ICD-8 codes 56300-02 and 56308-09, and ICD-10 code K50) were censored on the date of their diagnosis.
By linkage to the Danish Family Relations Database, we identified kinships within the cohort, and these were divided into first-degree, second-degree and third-degree relations. First-degree relatives were parents/children and siblings; second-degree relatives comprised half siblings, grandparents/grandchildren, uncle/aunt, nephew/niece and third-degree relatives as cousin, half uncle/half aunt/half nephew/half niece.
Statistical analysis
Poisson regression analyses were performed to obtain rate ratios (RRs) of first hospital contact incidence rates of UC with 95% CIs. First, we calculated RRs of UC according to cohort members' own history of appendicitis, treating appendicitis status as a time-dependent variable starting as ‘no appendicitis’ at start of follow-up and shifting to ‘appendicitis before age 20 years’ or ‘appendicitis at age 20 years or older’, when relevant. Next, for each degree of kinship, we calculated RRs of UC, comparing rates for exposed cohort members (defined as individuals with one or more relatives with a history of appendicitis) with rates for unexposed cohort members (defined as individuals with one or more relatives of the same degree of kinship, none of whom had a recorded history of appendicitis). For each degree of kinship, an individual was regarded as having a relative from the time of birth (in case of older relatives) or from the date of birth of the relative (in case of younger relatives); before that date the person was regarded as having no relative and did not contribute to the analysis for the particular degree of kinship.
Cohort members were considered exposed from the date of their relative's appendicitis; hence, a person could enter the analysis as unexposed and later shift to the exposed group and, from then on, permanently contribute person-time as exposed. Because individuals might have more than one type of relative with a history of appendicitis, each cohort member could contribute follow-up time as exposed in analyses for all three degrees of kinship. For each degree of kinship, we made separate analyses for appendicitis before 20 years of age and at 20 years or later, which for simplicity will be designated, respectively, as early and late appendicitis in the following.
Cohort members were followed from the date of birth or 1 January 1977, whichever came later, and until the date of first recorded hospital contact for UC, censoring on the date of emigration, death, diagnosis of Crohn's disease or end of the study (31 December 2011), whichever occurred first.
To examine whether a family predisposition to UC had any impact on the RRs associated with own or family members' history of appendicitis, all analyses were repeated in a subcohort comprising all cohort members with a family predisposition to UC, which was operationally defined as all individuals in our original cohort who had at least one first-degree, second-degree or third-degree relative with UC.
All statistical analyses were adjusted for sex, age (5-year intervals) and calendar period (5-year intervals). Analyses were carried out using SAS V.9.4 (Cary, North Carolina, USA).
Results
The cohort consisted of altogether 7 132 317 individuals (3 553 337 females; 3 578 980 males) followed for a total of 174 million person-years between 1977 and 2011. According to the National Patient Registry, 190 004 individuals were appendectomised and diagnosed with appendicitis and 45 202 were diagnosed with UC during follow-up.
RR of UC in individuals with a personal history of appendicitis
The RR of UC in individuals who had a history of early appendicitis themselves was significantly decreased compared with individuals with no such history (RR 0.52; 95% CI 0.47 to 0.59). In patients with a history of late appendicitis, the risk of UC was slightly increased (RR 1.09; 95% CI 1.01 to 1.18) (table 1).
RRs (with 95% CIs) of UC according to personal history of appendicitis
RR of UC in individuals with a family history of appendicitis
In individuals with a first-degree relative with early appendicitis, the RR of UC was significantly reduced (RR 0.90; 95% CI 0.86 to 0.95), whereas this was not the case in individuals with second-degree (RR 1.00; 95% CI 0.94 to 1.06) or third-degree (RR 0.91; 95% CI 0.83 to 1.01) relatives with early appendicitis. Relatives of patients with late appendicitis were not at reduced risk of UC (table 2).
RRs (with 95% CIs) of UC according to history of appendicitis in family members
Impact of family history of UC
The following analyses of the impact of appendicitis on risk of UC were restricted to cohort members with a family history of UC. Among these, individuals who had a history of early appendicitis themselves had a RR of UC that was more than halved (RR 0.41; 95% CI 0.24 to 0.64); for those with late appendicitis, the RR was also diminished (RR 0.57; 95% CI 0.35 to 0.88) (table 1). Also, a significant reduction in RR of UC was observed in individuals with a first-degree relative with early appendicitis (RR 0.66; 95% CI 0.51 to 0.83) as compared with other UC predisposed individuals without a first-degree relative with early appendicitis (table 2). The association was weaker and not statistically significant in individuals with second-degree or third-degree family members with early appendicitis (RR 0.84; 95% CI 0.67 to 1.03 and RR 0.94; 95% CI 0.74 to 1.18, respectively; table 2). A history of late appendicitis in relatives of individuals with a family predisposition to UC had no impact on the RR of UC for any degree of kinship.
Robustness analyses
Additional analyses were performed based on the entire cohort to test the robustness of our findings. Some cohort members had more than one relative with appendicitis, and such relatives may have been of different degrees of kinships and may represent both early and late appendicitis. For instance, the RR observed for late appendicitis in a second-degree relative might be influenced by the effect of having a first-degree relative with early appendicitis or of having an own history of early appendicitis. To account for such potential multiple exposure effects, we performed a robustness analysis in which all RRs were additionally adjusted for the presence of relatives with early and/or late appendicitis in other degrees of kinship, as well as own history of early or late appendicitis. This additional adjustment produced RR estimates similar to those shown in table 2 (see eTable 1 in supplemental material).
Supplemental material
Tobacco smoking is associated with both appendicitis15 and UC,16 so unmeasured smoking habits in the study cohort might act as a confounder of the observed associations of appendicitis in the family with risk of UC. Consequently, a sensitivity analysis was conducted, in which we used published estimates of the association of tobacco smoking with appendicitis and UC to examine the potential influence of unmeasured confounding by smoking habits (see eMethod in supplemental material). The RR for individuals with a first-degree relative with early appendicitis changed from 0.90 (95% CI 0.86 to 0.95) in the main analysis to 0.92 (95% CI 0.88 to 0.97) after control for unmeasured confounding by smoking, and the corresponding RR for the subcohort with a family history of UC changed from 0.66 (95% CI 0.51 to 0.83) to 0.67 (95% CI 0.52 to 0.85). Varying the parameters, and using the most extreme values in this robustness analysis, the RR corrected by unmeasured confounding by tobacco smoking was 1.01 (95% CI 0.97 to 1.07) for individuals with a first-degree relative with early appendicitis, and the corresponding RR was 0.74 (95% CI 0.59 to 0.97) for the subcohort with a family history of UC.
To examine for any possible distortion of our results stemming from the fact that some unknown proportion of cohort members had a diagnosis of UC or appendicitis established before 1977 (and were therefore not registered in the National Patient Registry), we conducted an additional robustness analysis restricted to cohort members born in 1962 or later (resulting in a maximum age at cohort entry of 15 years in 1977), which produced similar results (see eTable 2 in supplemental material). Also, we performed an analysis in which all patients with a record of Crohn's disease at any time between 1977 and 2011 were excluded from the study, as one might be concerned that some patients with Crohn's disease were initially incorrectly diagnosed with UC or appendicitis. This analysis also produced essentially unchanged results (see eTable 3 in supplemental material). Lastly, as the National Patient Registry does not contain data on outpatient and emergency visits diagnoses of UC prior to 1995, leaving non-hospitalised patients with UC to be unregistered before 1995, a sensitivity analysis was performed stratifying the analyses according to a diagnosis of UC before and after 1995. In this additional analysis, the RR of UC before and after 1995 in patients with a first-degree relative with appendicitis before age 20 years was 0.79 (95% CI 0.68 to 0.90) and 0.92 (95% CI 0.87 to 0.97), respectively (see eTable 4 in supplemental material). Thus, a significantly reduced risk of UC is present regardless of year of UC diagnosis.
Discussion
This cohort study used nationwide kinship data for more than 7 million individuals to show that the long-established inverse association of early appendicitis with subsequent risk of UC is present in individuals with a personal history of early appendicitis and also in individuals with a history of early appendicitis in their first-degree relatives. The protective effect of having a first-degree relative with early appendicitis was most pronounced in individuals with a family history of UC. Our novel findings suggest that the repeatedly observed inverse association between appendicitis and UC is unlikely to reflect a direct impact of appendicitis on the colonic mucosa. Rather, shared genetic and/or environmental factors, which increase the risk of appendicitis while at the same time reducing the risk of UC, need consideration.
The inverse association between appendicitis and risk of UC within the same individuals has been demonstrated repeatedly in studies of various size and quality in design. In 2001, Andersson et al performed a matched cohort study assessing the risk of UC in more than 200 000 patients who had undergone appendectomy.9 The study confirmed previous reports of a seemingly protective effect of appendectomy on the development of UC (incidence RR 0.73; 95% CI 0.62 to 0.87). The study further showed that the reduced risk was restricted to patients undergoing appendectomy for an inflammatory condition (appendicitis or lymphadenitis) before age 20 years. A subsequent binational cohort study from Denmark and Sweden confirmed these findings, reporting a reduced risk of UC in patients who underwent appendectomy for appendicitis before age 20 years (standardised incidence ratio 0.53; 95% CI 0.47 to 0.60).8 The study further showed that the relative risk reduction associated with appendectomy for appendicitis was particularly pronounced in individuals with a family disposition to UC. However, neither study8 ,9 was able to distinguish between the possible mechanisms underlying the inverse association.
Our study addresses the link between appendicitis and UC using, for the first time, a nationwide cohort design that includes relatives of patients with appendicitis. We confirm previous findings of a reduced risk of UC in individuals with a personal history of early appendicitis. We also present the novel finding that the reduced risk of UC applies even to family members of patients with appendicitis in childhood or adolescence.
Therefore, it may be questioned whether it is relevant to study any direct mechanism underlying the inverse association between appendicitis and UC, as it has been done in a few animal studies.17–20 These studies have focused on immunological changes in mice following appendectomy for appendicitis, although it has been questioned how well murine models may mimic human inflammatory diseases.21
The observed protective effect of appendicitis on the risk of UC in relatives is not compatible with the theory that immunological changes of the colonic mucosa following appendicitis confers protection against UC. Rather, our findings suggest that yet unidentified genetic traits that confer protection against UC may be closely linked to susceptibility genes for appendicitis (the linkage disequilibrium hypothesis22). Alternatively, environmental exposures that are shared among family members that are correlated to both appendicitis and UC, for example, the hygiene hypothesis, smoking, air pollution, antibiotic use or the microbiome,23–25 may explain our findings. These possibilities warrant further investigation.
The primary strength of this study is its historically prospective study design, which provides information on kinship in a national cohort with long follow-up and a large number of exposed cohort members. Further, the registration of UC diagnoses in the National Patient Registry has been found to be highly complete and accurate.26 Additionally, we identified family members and their diagnoses through nationwide registers and not through self-reports, thus minimising the risk of biased information about exposures and outcomes in the cohort.
The lack of data on tobacco smoking is a limitation and merits discussion. Smoking has been suggested as a possible confounder of the association between appendicitis and UC,27 but the strong inverse association was also seen in studies that took smoking into account.5 ,28 A sensitivity analysis examining the potential influence of unmeasured smoking habits showed that our findings were far from fully explained by our lack of data about smoking. When using rather extreme parameter estimates, the corrected RR for UC became inconspicuous among individuals with a first-degree relative with early appendicitis, but even in this unlikely scenario the RR remained significantly below 1 (RR 0.74) for individuals with a first-degree relative with early appendicitis among individuals with a family disposition to UC. Thus, while we cannot exclude that unmeasured smoking may have influenced our RR estimates to some extent, our findings suggest that a family history of early appendicitis is genuinely associated with reduced risk of UC, at least among individuals with a family disposition to IBD.
Another potential limitation is that some cohort members, notably those born early in the cohort, may have had a diagnosis of appendicitis or UC established prior to 1977 when the National Patient Registry was initiated. Such pre-1977 diagnoses might either be missed or wrongly included as incident diagnoses upon new hospital contacts in 1977 or later. Reassuringly, a robustness analysis restricted to cohort members born 1962 or later revealed unchanged results, suggesting that any possible impact on our findings of prevalent cases of UC diagnosed before 1977 was limited. Lastly, as outpatient and emergency room visits diagnoses of UC were not captured in the National Patient Registry before 1995, those diagnosed before that year with mild UC (in no need of hospitalisation due to their disease) may have been missed, which could have influenced our results. However, a sensitivity analysis stratified according to a diagnosis of UC before and after 1995 showed a reduced RR of UC in first-degree family members to patients with early appendicitis regardless of year of UC diagnosis, thus supporting our main findings.
Conclusion
In conclusion, our study shows that appendicitis in childhood or adolescence is associated with reduced risk of UC later in life in the affected individual and early appendicitis is also associated with reduced risk of UC in first-degree relatives. This association is most pronounced in individuals with a family predisposition to UC. These novel findings question the theory of a direct mucosal role of appendicitis on the future risk of UC; rather our findings suggest that genes and/or shared environmental factors may explain the repeatedly observed inverse association of appendicitis with risk of UC.
References
Footnotes
Contributors NNA had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. She is the guarantor. Study concept and design: all authors. Acquisition of data: SG, MF, TJ. Analysis and interpretation of data: all authors. Drafting of the manuscript: NNA. Critical revision of the manuscript for important intellectual content: SG, MF, TJ. Statistical analysis: SG. Obtained funding: TJ. Study supervision: MF, TJ.
Funding TJ holds a Female Research Leader grant (no. 09-066323) from the Danish Council of Independent Research.
Competing interests All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author). TJ has received funding for travel and speakers fee from AbbVie. NNA has received funding for travel and speakers fee from MSD.
Ethics approval The study was approved by the Danish Data Protection Agency (approval no. 2008-54-0472).
Provenance and peer review Not commissioned; externally peer reviewed.