BIOLOGIC USE IN PATIENTS WITH INFLAMMATORY BOWEL DISEASE AND RISK OF GASTROINTESTINAL INFECTIONS: A PROPENSITY-MATCHED NATIONAL DATABASE COHORT ANALYSIS

Background: Cancer is a major cause of morbidity and mortality in persons with inflammatory bowel disease (IBD). We evaluated temporal trends and relative risks of intestinal and extra-intestinal cancers among persons with IBD and matched controls.

Methods: Using population-level administrative data housed at IC/ES (Ontario, Canada), we identified persons with IBD (2002-2017) using a validated algorithm, which also identified diagnoses of Crohn's disease (CD) and ulcerative colitis (UC). Each case was age and sex matched to 10 non-IBD controls and all persons were followed to 2020 for the development of cancer, as identified in the Ontario cancer registry. Only the first organ-specific cancer was counted in each analysis, after which an individual was censored. Individuals were excluded in the analyses of colorectal cancers (CRC) if they underwent proctocolectomy before their first IBD contact and censored at the time of proctocolectomy thereafter. Individuals were censored in all analyses when they died or lost provincial health insurance coverage. Annual cancer incidence rates were age- and sex-standardized to the July 2015 Ontario IBD population. Log-transformed standardized incidence rates for each cancer were analyzed using linear regression, with a first-order autoregressive term, to derive the average annual percentage change (AAPC). Quasi-Poisson regression was used to calculate rate ratios between persons with IBD and controls (2017-2020).

Results: Of the 19,984,538 individuals in the Ontario population from 2002 to 2017, we identified 110,919 persons with IBD. Among persons with IBD, the AAPC significantly decreased for colorectal cancers (CRC) but significantly increased for small bowel cancers, non-Hodgkin’s lymphoma, melanoma and cervical cancers (Table 1). Rates of lung, breast and prostate cancers did not change in IBD, but declined significantly in controls. These trends were similar in persons with UC and matched controls (excluding small bowel cancers). Among persons with CD, the AAPC was not significant for any cancer. During 2017-2020, rates of intestinal cancers, lymphomas, biliary tract cancers and lung cancers were significantly higher in persons with IBD than controls, whereas they did not differ for melanoma or for breast, cervical or prostate cancer.

Conclusions: Rates of CRC have declined, whereas rates of small bowel cancers, NHL, melanoma and cervical cancers have risen, over the past 25 years among persons with IBD . Trends for small bowel, lung, breast, cervical and prostate cancer, as well as for NHL, are discrepant from persons without IBD. Rates of intestinal cancers, lymphomas, biliary tract cancers and lung cancers are higher in persons with IBD than those without IBD. These data should be used inform discussions regarding cancer risks between IBD patients and their practitioners.

Background
Previous studies have shown that offspring of Crohn’s disease (CD) patients have a 6-8 fold risk of incident CD than the general population. Also, newborns of women with CD (vs healthy women) have altered stool microbiome composition and elevated fecal calprotectin (FCP). The "peri-natal period" (pregnancy and first year postpartum) is critical for the development and maturation of the gut microbiome as well as immune responses. However, it is unclear if exposure to parental CD diagnosis during the peri-natal period is associated with increased risk of CD. We aimed to investigate whether peri-natal exposure to parental CD (compared to exposure later in life) has an impact on offspring’s gut barrier function, microbiome composition, and the offspring’s risk of incident CD.
Methods
We assessed 1252 healthy offspring (6-35 years of age) of patients with CD who were recruited in the Crohn’s Colitis Canada Genetic, Environmental, Microbial (CCC-GEM) Project, a large prospective cohort study following healthy first-degree relatives (FDR) of CD patients. We classified offspring into "peri-natally exposed" or "exposed later in life" to parental CD diagnosis. We measured baseline FCP and urinary fractional excretion ratio of lactulose to mannitol (marker of intestinal permeability, LMR). Fecal microbiome composition was determined by 16S rRNA gene sequencing. General estimating equation regression, MaAsLin2, and Cox-proportional hazard modeling was used to assess if peri-natal exposure to parental CD is associated with LMR, FCP, altered microbial composition, and future CD onset. Offspring’s age and sex and family clusters were accounted for in the models.
Results
Offspring exposed to a CD parent peri-natally have a 4.73 fold higher risk (adjusted HR 95%CI 1.28-17.46) of developing CD compared to those exposed to a parent who developed CD later in life (11/586 vs 3/666). Baseline LMR was significantly higher in the offspring exposed to parental CD peri-natally (p=0.003) while no significant difference was observed for FCP at recruitment (p=0.94). We identified five bacterial taxa that were significantly increased in the peri-natally exposed group (false discovery rate-adjusted p<0.05). In a subgroup of offspring of mothers with CD, a dose effect between exposure age and risk was seen; the earlier the offspring was exposed to mother’s CD diagnosis peri-natally, the higher the offspring’s risk of CD development; this trend was not seen in offspring of fathers with CD.
Conclusion
Offspring exposed to parental CD peri-natally had a higher risk of developing CD than those exposed to parental CD later in life. Early life exposure to parental CD was associated with higher baseline LMR and altered bacterial taxa. These results suggest that environmental exposure during the critical peri-natal period may determine the offspring’s future risk of developing CD.

Background: Bowel urgency is a common and disruptive symptom in patients with ulcerative colitis (UC).¹ Fatigue is also common and has been associated with greater disease activity and lower quality of life in these patients.² However, little is known regarding the relationship between bowel urgency remission, bowel urgency clinically meaningful improvement, rectal bleeding remission, or stool frequency remission and improvement in fatigue. We examined the association between bowel urgency remission or bowel urgency clinically meaningful improvement and Fatigue Numeric Rating Scale (NRS) scores in the presence of rectal bleeding remission or stool frequency remission using pooled data from LUCENT-1 (NCT03518086) and LUCENT-2 (NCT03524092) phase 3 trials.

Methods: Bowel urgency was assessed by the Urgency NRS (UNRS), ranging from 0 (no urgency) to 10 (worst possible urgency). Bowel urgency remission was defined as a UNRS score of 0 or 1; bowel urgency clinically meaningful improvement was defined as a ≥3-point decrease in UNRS score from baseline.³ Rectal bleeding remission was defined as a modified Mayo subscore of 0; stool frequency remission was defined as a subscore of 0 or 1 with ≥1-point decrease from baseline.⁴ Fatigue was assessed by the Fatigue NRS, ranging from 0 (no fatigue) to 10 (worst possible fatigue), using change from baseline in Fatigue NRS score. Mediation analyses were performed to separately examine the relative association between the direct effects of bowel urgency remission and bowel urgency clinically meaningful improvement (predictor variables) and Fatigue NRS scores while adjusting for rectal bleeding remission and stool frequency remission (mediation variables). Analyses were treatment agnostic and combined mirikizumab and placebo treatment groups from LUCENT-1 (N=1162) and LUCENT-2 (N=544).

Results: The direct effect of bowel urgency remission at Week 12 and Week 40 (52 weeks continuous treatment) accounted for 68.3% and 62.1%, respectively, of the improvement in change from baseline in Fatigue NRS score (Fig. 1). Substantially greater proportions of the Fatigue NRS score improvements were related to bowel urgency remission versus rectal bleeding remission or stool frequency remission. Results were similar for bowel urgency clinically meaningful improvement at Week 12 (91.8%) and Week 40 (87.4%) (Fig. 2).

Conclusion: These findings provide strong evidence for a direct association between improvement or remission in bowel urgency and improvement in fatigue in patients with moderately-to-severely active UC.

References
1. Hibi T, et al. Inflamm Intest Dis. 2020;5(1):27.
2. Regueiro MD, et al. Adv Ther. Online Nov 12, 2022. doi: 10.1007/s12325-022-02364-2.
3. Dubinsky MC, et al. J Patient Rep Outcomes. 2022;6(1):114.
4. ClinicalTrials.gov. Updated October 27, 2022. Accessed December 1, 2022. clinicaltrials.gov/ct2/show/NCT03518086.

Background: Post-operative recurrence (POR) occurs in ~30% of patients on post-operative prophylaxis. Drug persistence may serve as a real-world proxy for safety and efficacy and an important consideration for therapeutic positioning, but factors involved in post-operative prophylaxis persistence is unknown. We aim to establish clinical, serologic, and genetic factors associated with post-operative prophylaxis persistence.

Methods: We conducted a single center, retrospective study of adult and pediatric surgically naïve ileal or ileocolonic CD subjects undergoing ileocecal and/or small bowel resection between 1/1/2000-12/31/2021 and prescribed post-operative prophylaxis (5-aminosalicylate, immunomodulator, and/or biologic). Subjects who stopped prophylaxis for non-medical reasons were excluded (e.g. self-discontinued, insurance denial). The primary outcome was time to prophylaxis failure, which was defined as requiring recurrent surgery or changing therapy due to POR despite optimized dose/drug level, immunogenicity, and/or adverse event. IBD serologies (ELISA) and genetics (Immunochip) were generated. Quartile sum scores (QSS) were calculated from IBD serologies. Cox proportional hazard analyses were adjusted for demographic and disease characteristics and principal components. Canonical pathway analysis was generated in Ingenuity Pathway Analysis.

Results: Of 552 subjects, 57.6% (n=318) had prophylaxis failure. Table 1 summarizes cohort demographics. Median length of follow-up was 79.8 months. Number of prior biologic exposure (adjusted hazard ratio (aHR)1.40, 95%confidence interval (CI)[1.15-1.70]), length of resection (aHR1.01, 95%CI[1.00-1.01]), and lymph node granulomas (aHR1.45, 95%CI[1.03-2.04]) were associated with risk of prophylaxis failure (Table 1). ANCA seropositive subjects with low QSS had increased risk of prophylaxis failure (aHR 2.26, p=0.03) while IgA ASCA seropositivity (aHR0.72, p=0.021) and level (aHR0.99, p=0.005), IgG ASCA seropositivity (aHR0.72, p=0.02), and ANCA seronegative/ASCA seropositive status (aHR0.64, p=0.001) were protective (Table 2). No SNPs achieved genome-wide significance. Significant enriched pathways involved Th1 and Th2 activation (p=1.58e^-12), TREM1 (p=2.09e^-10), IL-10 (p=2.40e^-10), IL-6 signaling (p=1.66e^-9), pathogen induced cytokine storm (p=1.95e^-9), and hepatic fibrosis signaling (p=4.68e^-9).

Conclusion: Number of prior biologic exposure, resection length, lymph node granuloma, and ANCA seropositive subjects with low QSS are associated with decreased post-operative prophylaxis persistence. TREM1, cytokine storm, and hepatic fibrosis signaling pathways are also associated with prophylaxis persistence. If validated by larger studies, these factors can be useful prognostic markers to inform post-operative management of CD and/or facilitate identifying novel therapeutic targets for POR.

Background: Compared to those without inflammatory bowel disease (IBD), women with IBD may have increased health-care utilization during pregnancy and postpartum. This may lead to significant morbidity and decrease in quality of life. Characterizing this health-care use is important for health-policy purposes to determine methods to shift care to the ambulatory setting.

Aim: To compare health-care utilization of women with and without IBD during preconception, pregnancy and postpartum.

Methods: We accessed administrative databases and validated algorithms at the Institute of Clinical Evaluative Services (ICES) in Ontario to identify women (age 18-55) with and without IBD who had a completed live, singleton pregnancy between 2003 and 2018. The primary outcome was to characterize differences in emergency department (ED) visits and hospitalizations between women with and without IBD during the 12 months preconception, pregnancy, and in the 12 months postpartum. The secondary outcome was to assess differences in prenatal care between women with and without IBD. Multivariable negative binomial regression with generalizing estimating equations, accounting for multiple pregnancies for each patient, was performed to report incidence rate ratios (IRR) with 95% confidence intervals (95% CI). Covariates included maternal age at conception, location of residence at conception (rural vs. urban), socioeconomic status (using surrogate marker of neighborhood income quintile), and maternal comorbidity.

Results: 9158 pregnancies in 6163 women with IBD and 1,729,411 pregnancies in 1,091,013 women without IBD were included. Women with IBD were older at time of delivery and had greater pre-pregnancy comorbidities. During pregnancy, women with IBD were more likely to visit the ED (IRR 1.13, 95% CI,1.08-1.18) and be hospitalized (IRR 1.11, 95% CI,1.01-1.21) for non-IBD specific reasons. Similarly, during postpartum, women with IBD were more likely to visit the ED (IRR 1.21, 95% CI, 1.15-1.27) and be hospitalized (IRR 1.18, 95% CI, 1.05-1.32) for non-IBD specific reasons. Venous thromboembolic events accounted for 7.0% of all postpartum hospitalizations in women with IBD compared to 2.7% in those without IBD (p<0.0001). There was no difference in ED visits and hospitalizations between women with and without IBD in preconception. Women with IBD had greater number of prenatal visits with obstetricians and were more likely to receive a first trimester prenatal visit.

Conclusions: Compared to those without IBD, women with IBD are more likely to visit the ED and be hospitalized during pregnancy and postpartum, particularly for venous thromboembolic events. Efforts should be made from a health policy perspective to increase access to ambulatory care for patients with IBD during the peripartum period which in turn may reduce acute setting health-services utilization.

Background: Gastrointestinal (GI) infections are known triggers of inflammatory bowel disease (IBD) flares; however, the differential effects of biologics with different mechanisms of action (systemic immunosuppression vs gut-specific targeting) on risk of GI infections is not well-characterized. We evaluated the future risk of GI infections in patients with IBD on biologics using a large national database.
Methods: We analyzed adult patients (18 years or older) who had a diagnosis of IBD (ulcerative colitis or Crohn’s disease) and no prior GI infections based on International Classification of Diseases (ICD) codes in a retrospective, nationally representative database (Healthjump database of about 36 million patients from 2000-2021). Our GI infection outcomes included C. difficile, cytomegalovirus (CMV), bacterial (E. coli, Campylobacter, Yersinia, Proteus, etc), and viral (Rotavirus, Norovirus, Adenovirus, etc) infections. We compared these patients with IBD on anti-TNF (infliximab, adalimumab, cetrolizumab, golimumab) versus anti-integrin (vedolizumab, natalizumab) vs no biologic controls (mesalamine/sulfasalazine comparators) using propensity score (PS) matching (for demographics and co-morbidities).
Results: We identified 1085 patients with IBD on anti-TNF agents and 2377 controls matched to a follow-up time of 9.9 years. Anti-TNF use was associated with increased risk of any GI infection, C. diff, and viral infections in unmatched analysis, but lost significance after PS matching. Anti-TNF use was not associated with CMV infections. Anti-TNF use was only associated with increased risk of bacterial infections (OR 2.8, 95% CI 1.24-6.32, P=0.01) after PS matching. We identified 765 patients with IBD on anti-integrin agents and 8630 controls matched to a follow-up time of 2.4 years. Anti-integrin use was associated with increased risk of any GI infection (OR 1.57, 95% CI 1.13-2.19, P<0.01) in unmatched analysis, but lost significance after PS matching. Anti-integrin use was not associated with risk of CMV or bacterial infections. Anti-integrin use was associated with increased risk of C. diff infections in unmatched and PS matched (OR 2.19, 95% CI 1.15-2.16, P=0.01) analyses and with decreased risk of viral infections in PS matched analysis (OR 0.18, 95% CI 0.01-0.83, P=0.02). There were no differences in GI infection outcomes in PS matched analyses comparing 3758 anti-TNF users vs 326 anti-integrin users.
Conclusions: In this PS-matched analysis of a large, nationally representative dataset, anti-TNF use resulted in increased risk of GI bacterial infections, whereas anti-integrin use was associated with increased risk of C. difficile infections but decreased risk of GI viral infections. Biologics of different mechanisms of action may confer distinct GI infection risk profiles.