Introduction

Immuno-mediated colitis is one of the most common gastrointestinal side effects associated with immune checkpoint inhibitors (ICIs). ICIs-colitis shares with inflammatory bowel disease (IBD), including both Crohn’s Disease (CD) and Ulcerative Colitis (UC), the clinical presentation, macroscopic, microscopic, and serological findings, making challenging the differentiation between the two conditions. The gut microbiota has been suggested as an important driver in the pathogenesis of ICIs-colitis as well as in the pathogenesis of IBD. We aimed to assess whether a specific microbiota profile, as measured by a machine learning approach, can be associated with ICIs-colitis and whether it differs from the one assessed in patients with IBD and in healthy controls (HCs).

Aims & Methods

In this prospective pilot study, consecutive patients with ICIs-colitis, patients with IBD and HCs were enrolled. Stool samples were collected for fecal microbiota assessment analysis by 16S rRNA gene sequencing approach. The raw reads underwent a filtering procedure performed within QIIME2 analysis framework (version 2022.8). Alpha diversity was evaluated on rarefied counts (Richness, Shannon, and Pielou indices; rarefaction level: 31,556), while beta diversity was calculated on normalized counts (Bray-Curtis, Jaccard, Canberra, Weighted and Unweighted Unifrac; counts normalized with GMPR). The diversity analysis was conducted in R (version 4.1.0). A permutational analysis of variance (PERMANOVA) test on Bay-Curtis dissimilarity was used to test for differences in the microbiota composition between disease status groups (vegan and ecole packages). The MaAsLin2 package was then used to perform differential abundance analysis at all taxonomic levels. Supervised and unsupervised machine learning algorithms were applied.

Results

Nineteen patients with ICIs-colitis, 40 patients with UC (20 active and 20 inactive), 34 with CD (14 active and 20 inactive) and 36 HC were enrolled. Alpha diversity was not significantly different between ICIs-colitis and HCs (p=0.94), while it was statistically significant between IBD and HCs (p=0.02) and between IBD and ICIs-colitis (p=0.03). At phylum levels, we found high levels of Tenericutes in HCs and high levels of Proteobacteria in ICIs-colitis. Moreover, we found high levels of Actinobacteria in IBD compared to the other group. Interestingly, at genus levels, we found very high levels of Enhydrobacters in ICIs-colitis compared to patients with IBD. While at species levels, high levels of Bifidobacterium longum in patients with UC were observed. A specific microbiota profile was found for each group (IBD, ICIs, HC) and was confirmed with sparse partial least squares discriminant analysis, a machine learning-supervised approach. The latter allowed us to observe a perfect class prediction and group separation using the complete information (full Operational Taxonomic Unit table), with a minimal loss in performance when using only 5% of features.

Conclusion

A machine learning approach to 16S rRNA data identifies a bacterial signature characterizing ICIs-colitis from IBD and HCs. Future research will clarify whether such microbiota profiling is useful for prediction, early diagnosis and management of patients with ICIs-colitis.

Disclosure

None

Introduction

Janus kinase (JAK) inhibitors have emerged as an important therapeutic class for the management of several immune-mediated conditions, including inflammatory bowel disease (IBD), but they are associated with several side effects. A renewed attention to the safety profile of this drug class has been driven by increasing reports of JAK inhibitor-induced acneiform eruptions seen in clinical trials and real-life practice.

Aims & Methods

We aimed to investigate the epidemiology, clinical characteristics, and treatment outcomes of this condition in patients with IBD. This was an international, multicenter, retrospective cohort study that consecutively enrolled patients with IBD who were treated with JAK inhibitors and subsequently developed acne. The primary objective was to evaluate the clinical characteristics of acne presentation. Secondary objectives included determining acne prevalence, assessing its impact on quality of life, and evaluating the effectiveness of treatments.

Results

Of the 2,183 JAK inhibitor-treated patients with IBD screened, 272 developed acne and were included in the study. The crude prevalence rates of acne were 15.9% for upadacitinib, 4.3% for tofacitinib, and 1.9% for filgotinib, with dose-dependent relationships observed for upadacitinib and tofacitinib. Acne predominantly affected patients aged 30–50 years, with most cases being mild to moderate in severity. Acneiform eruptions most frequently involved the face (89%), followed by the back (33%) and chest (27%). Lesion characteristics were documented in 159 patients; papules (82%), pustules (52%), and comedones (42%) were the most frequently reported. In contrast, nodules (19%) and cysts (18%) – features associated with more severe disease – were less commonly observed. Among the 213 patients with available time-to-onset data, the median time to acne onset following JAK inhibitor initiation was 47 days (IQR: 28–98 days). Patients with a previous history of acne vulgaris were significantly more likely to develop severe drug-induced acne (p<0.0001). Although corticosteroids are known to contribute to acne, our data did not demonstrate an association between concomitant steroid use and increased acne severity. There was also no difference in time to onset or resolution of acne between the steroid-treated and non-steroid treated groups. A third of patients reported a negative impact on psychosocial well-being, and 40% required pharmacological intervention. Topical therapies were required for 106 patients (39%), with 52 (49%) showing acne improvement or resolution. Systemic oral agents were required for 38 patients (14%), including 25 patients who did not respond or had a partial response to topical therapies. Antibiotics were prescribed to 29 patients (most commonly from the tetracycline class), leading to acne improvement or resolution in 14 patients (48%). Oral isotretinoin was required for 8 patients (3%) and was effective in all but one patient. Eighteen percent of patients who developed acne (2.2% of total cohort) required a JAK inhibitor dose reduction or discontinuation, although most did not have severe disease.

Conclusion

This study is the first to characterise this adverse event in JAK inhibitor-treated patients with IBD and presents the largest cohort of JAK inhibitor-induced acne cases to date across all immune-mediated diseases. JAK inhibitor-induced acne is a common adverse effect resulting in significant psychological burden. Early identification, proactive counselling, and timely interventions, such as dose reduction or referral to dermatology, are crucial in managing this side effect.

Introduction

Proper and timely risk stratification of inflammatory bowel diseases (IBD), including ulcerative colitis (UC) and Crohn’s disease (CD), is essential to tailor personalized treatment.

Aims & Methods

The aims of the current study were to report the early disease course of newly diagnosed UC and CD and identify risk factors for a severe disease course. Patients were identified from an ongoing prospective population-based inception cohort at Hvidovre University Hospital and Herlev University Hospital, IBD Prognosis Study, which includes all patients with newly diagnosed IBD since May 2021 and forth within a catchment population of 1.050.000 (~20% of the Danish population). We defined severe disease course as the need for surgery or initiation of biological therapies within six months of diagnosis. A Cox proportional hazard model with Bonferroni correction adjusting for multiple comparisons was constructed to identify predictors hereof.

Results

As of April 15^th, 2023, the cohort comprised 501 patients, including 287 with UC,180 with CD, and 34 with IBD unclassified (IBDU). The mean age (interquartile range (IQR)) at diagnosis was 39 years (27-56), 36 (25-54), and 36 (27-52), respectively. The cumulative risk of a severe disease course within six months after the onset of UC, CD and IBDU was 3.8% (95% confidence interval 1.9-6.8), 13.9% (95% CI 9.2-19.8) and 11.4% (95%CI 3.2-26.7).
In a Cox multivariable regression model adjusting for age and gender, a severe initial disease course for UC was predicted by extensive disease distribution, severe disease endoscopically at diagnosis, and affected biochemical inflammation parameters (Table 1). In patients with CD, perianal disease but no other phenotypic characteristics were predictive for the initial course. However, in a sensitivity analysis of luminal CD, stricturing phenotype at diagnosis (adjusted hazard ratio (aHR) = 4.6 (95% confidence interval (CI) 1.5-13.6), the platelet count (HR= 1.3 (95% CI 1.2-10.6), and CRP > 10 mg/L (aHR= 14.2 (95% CI 1.9-108.5) were independently associated with a severe initial course, while initial use of systemic steroids was associated with a reduced risk. In patients with IBDU, only low hemoglobin was predictive for an initial severe disease course.

Table 1: Predictors of early severe disease course in newly diagnosed IBD in multivariate Cox proportional hazard model with Bonferroni correction (hazard ratio (95% confidence interval))

Conclusion

In this population-based study, we found 4%, 14%, and 11% of patients with newly diagnosed UC, CD, and IBDU, respectively, to experience an early severe disease course. Furthermore, systemic inflammation was predictive for a severe initial course in UC. In CD, perianal disease or stricturing disease phenotypes identify the subgroup of patients at high risk of early severe disease course.

Disclosure

Mohamed Attauabi: None
Gorm Roager Madsen: None
Flemming Bendtsen: Tillotts, Ferring
Klaus Theede: Ferring, Janssen, Takeda, Pfizer, and Tillotts
Jakob Benedict Seidelin: has received research grants from Takeda and Janssen and is national coordinator of studies from AbbVie, Arena Pharmaceuticals, Ely Lilly, and Boehringer Ingelheim. None of these pertain to the research submitted here.
Johan Burisch: reports grants and personal fees from AbbVie, grants and personal fees from Janssen-Cilag, personal fees from Celgene, grants and personal fees from MSD, personal fees from Pfizer, grants and personal fees from Takeda, grants and personal fees from Tillots Pharma, personal fees from Samsung Bioepis, grants, and personal fees from Bristol Myers Squibb, grants from Novo Nordisk, personal fees from Pharmacosmos, personal fees from Ferring, personal fees from Galapagos, outside the submitted work.

Introduction

A significant proportion of patients with inflammatory bowel disease (IBD) have concomitant rheumatological manifestations. On the other hand, patients with spondyloarthritis (SpA) often report gastrointestinal symptoms, some of them being associated to IBD. However, the exact prevalence of this disorder is still controversial, and the optimal approach to its diagnosis remains unknown.

Aims & Methods

Our main was to determine the prevalence of undiagnosed IBD in patients with axial SpA (radiographic and non-radiographic) and psoriatic arthritis (PA). EISER is a multicenter, nationwide, observational, and cross-sectional study where 13 Spanish hospitals participated. Patients ≥18 years-old diagnosed with PA or SpA according to the CASPAR or ASAS criteria, respectively, were selected. Patients receiving biologics or systemic steroids at baseline or within the previous 30 days were not included. Regarding patients older than 50 years, only those without valid colonoscopy in the previous 3 years were included. Rheumatologist performed a baseline visit and collected demographic, clinical, and treatment data. Fecal calprotectin (FC) was determined by the Quantum Blue rapid test and this was followed by a gastroenterologist consultation. A complete colonoscopy was indicated on all patients with symptoms, FC≥80 µg/g without NSAIDs, or 160 µg/g if they were on NSAIDs. Capsule endoscopy or magnetic resonance enterography were performed in those with no findings on colonoscopy.

Results

A total of 559 patients were included (51% men; mean 52.2 years; mean disease duration of SpA/PA was 12.3 years). The most frequent rheumatological disease was radiographic SpA (37%), followed by peripheral APs (36%). Abnormal FC was observed in 47% and 53% of patients with PA and EspAx (80% radiographic-SpA vs 20% non-radiographic-SpA), respectively. Mean FC levels were higher among radiographic-SpA (395 µg/g) compared to the rest of the groups (306 µg/g for non-radiographic-SpA and 306 µg/g in PA). Overall, 11,5% had gastrointestinal symptoms, with the most common being abdominal pain (16%) and chronic diarrhea (11%). In total, 189 colonoscopies were performed, 40% demonstrating at least one abnormal finding, mostly aphthous ulcers (66%), superficial ulcers (21%) and erythema (14%), and they were mainly located in the terminal ileum (41%). Finally, 23 patients were diagnosed with IBD (4.4%), mostly with radiographic-SpA (83%), followed by peripheral PA (8.7%) and non-radiographic-SpA and axial PA (4.3% each). All patients but two were diagnosed with CD (91.3%), with 1 case of unclassified IBD (4.3%) and 1 with nonspecific ileitis (4.3%).

Conclusion

FC levels in patients with SpA and PA, followed by an appropriate endoscopic and/or radiological diagnostic algorithm, allows the identification of a relevant subgroup of patients with previously undiagnosed IBD