Predicting Complicated Crohn's Disease and Surgery

Phenotypes, Genetics, Serology and Psychological Characteristics of a Population-Based Cohort

J. D. Ryan; M. S. Silverberg; W. Xu; L. A. Graff; L. E. Targownik; J. R. Walker; R. Carr; I. Clara; N. Miller; L. Rogala; C. N. Bernstein


Aliment Pharmacol Ther. 2013;38(3):274-283. 

In This Article

Abstract and Introduction


Background Predictors of complicated Crohn's disease (CD), defined as stricturing or penetrating behaviour, and surgery have largely been derived from referral centre populations.

Aim To investigate whether serological markers, susceptibility genes or psychological characteristics are associated with complicated CD or surgery in a population-based cohort.

Methods One hundred and eighty-two members of the Manitoba IBD Cohort with CD phenotyped using the Montreal classification underwent genetic and serological analysis at enrolment and after 5 years. One hundred and twenty-seven had paired sera at baseline and 5 years later and their data were used to predict outcomes at a median of 9.3 years. Serological analysis consisted of a seven antibody panel, and DNA was tested for CD-associated NOD2 variants (rs2066845,rs2076756,rs2066847), ATG16L1 (rs3828309, rs2241880) and IL23R (rs11465804). Psychological characteristics were assessed using semi-structured interviews and validated survey measures.

Results Sixty-five per cent had complicated CD and 42% underwent surgery. Multivariate analysis indicated that only ASCA IgG-positive serology was predictive of stricturing/penetrating behaviour (OR = 3.01; 95% CI: 1.28–7.09; P = 0.01) and ileal CD (OR = 2.2; 95% CI: 1.07–4.54, P = 0.03). Complicated CD behaviour was strongly associated with surgery (OR = 5.6; 95% CI: 2.43–12.91; P < 0.0001), whereas in multivariate analysis, only ASCA IgG was associated (OR = 2.66; 95% CI, 1.40–5.06, P = 0.003). ASCA titre results were similar at baseline and follow-up. Psychological characteristics were not significantly associated with disease behaviour, serological profile or genotype.

Conclusions ASCA IgG at baseline was significantly associated with stricturing/penetrating disease at 9–10 years from diagnosis. Stricturing/penetrating disease was significantly associated with surgery. In a model including serology, the genotypes assessed did not significantly associate with complicated disease or surgery.


Crohn's disease (CD) is an idiopathic disorder that is classically characterised by relapsing and remitting inflammation that may affect any region of the gastrointestinal tract. The natural history of CD is often characterised by progression to complications such as strictures, abscesses and fistulas that may result in surgery for as many as half of all patients.[1] Distinct clinical patterns of disease have long been recognised and remain the basis upon which CD phenotypes are defined using the Montreal classification.[2] Several large referral centre studies have identified phenotypic characteristics such as younger age at diagnosis, ileal involvement and perianal fistulas as risk factors for progression to complicated CD, defined as stricturing or penetrating behaviour, and the need for surgical intervention.[3–5] No consensus has emerged, however, on when or how phenotyping should be undertaken and early identification of those individuals at risk of CD complications remains elusive using the Montreal classification alone. This has limited the widespread adoption of formal phenotyping into everyday clinical practice. It has also bred enthusiasm for the potential that a serological marker or genotype or combination of both might be predictive of disease outcomes. In fact, a recent paediatric study assessing CD outcomes suggested that clinical parameters at diagnosis were insufficient to predict a disabling course of paediatric CD, and that more complex models including serological and genetic biomarkers should be tested.[6]

As knowledge of the complex interaction among the immune system, gut microflora and genetic susceptibility to CD grows, so too does the number of potential biological markers of disease. This has prompted the authors of the Montreal classification to call for further studies on the integration of new biomarkers of CD to better characterise those individuals at risk of disease complications.[2,7] In recent years, numerous genome-wide association studies (GWAS) have identified a wide array of single nucleotide polymorphisms (SNPs) and novel susceptibility loci for CD, including oligomerisation domain protein 2 (NOD2), lymphocyte signalling, including IL23 receptor (IL23R), and autophagy involving autophagy-related 16-like 1 gene (ATG16L1).[8] Furthermore, NOD2 mutations have been associated with early age of onset, ileal involvement, and penetrating and stricturing disease.[9–18] Serological markers including anti-Saccharomyces cerevisiae antibody (ASCA), as well as antibodies to Escherichia coli outer-membrane porin C (anti-OmpC), Pseudomonas fluorescens-related protein (anti-I2), bacterial flagellin CBir1 (anti-CBir1) and perinuclear staining antineutrophil cytoplasmic antibody (p-ANCA) have also been associated with increased disease severity and complications. However, it remains unclear if serological and/or genetic analyses augment the ability to predict the development of complicated CD over and above the knowledge of the disease phenotype at baseline, or whether these factors are associated with a stricturing/penetrating phenotype. Population-based studies are needed to confirm that the same phenotypic risk factors as well as predictive genetic or serological markers present in referral centre cohorts are relevant when characterising CD in the general population.

In addition, there are emerging data indicating an effect of psychological factors, including mood disorders and stress, on disease course in IBD.[19–21] For this study, in addition to previously identified psychological factors, we were particularly interested in whether childhood trauma was related to complicated disease in CD, given prior research that has shown that this chronic early life stressor can predispose to later health issues.[22,23] Rates of childhood trauma and adversity are surprisingly high, with two thirds of the general community reporting this type of experience.[24] Briere et al. reported that as many as one in three girls and one in seven boys will be sexually abused at some point in their childhood.[25] Although the mechanisms remain unclear, experimental research has suggested that those with early abuse experiences may have cortisol suppression as an adaptation to an exaggerated stress response at the pituitary and/or hypothalamus level, which may impact on immune function.[26–28] As well, prospective work by Danese et al. found a relationship between early life stress and later life inflammation.[29] They tested the life-course association between childhood maltreatment and adult inflammation in a birth cohort followed up to age 32 years as part of the Dunedin Multidisciplinary Health and Development Study.[29] Maltreated children showed a significant and graded increase in the risk for clinically relevant C-reactive protein levels 20 years later, in adulthood (relative risk (RR) 1.80, 95% confidence interval (CI): 1.26–2.58). Finally, in a meta-analysis of over 300 empirical articles describing a relationship between psychological stress and parameters of the immune system in human, chronic stressors were associated with suppression of both cellular and humoral measures.[30]

Therefore, we aimed to assess predictors of developing complicated CD and the need for abdominal surgery after several years of disease using clinical phenotype, serological markers and susceptibility genes in a population-based cohort of CD participants. We also performed a novel analysis of the association between psychological characteristics, CD phenotype and clinical outcomes.