Abstract Background and aims Vedolizumab (VDZ) blocks integrin α4β7-MAdCAM-1 interactions, limiting leukocyte entry into the gut, yet peripheral-blood immune correlates of clinical response in ulcerative colitis (UC) remain unidentified. Methods Peripheral blood mononuclear cells from healthy controls (HCs) and patients with UC, stratified by VDZ exposure and clinical response, were analyzed using mass cytometry (cytometry by time-of-flight). A 38-marker panel enriched for integrins, chemokine receptors, and activation markers was applied. Major immune populations were interrogated using lineage-specific opt-SNE and FlowSOM clustering. Results The dataset included 26 HCs, 30 VDZ-naive UC (Nancy index ≥ 2), 8 VDZ-treated non-responders (Nancy index ≥ 2), and 8 VDZ-treated responders (Nancy index <2). No significant differences were detected in myeloid, NK-cell, or TCRγδ+ T-cell compartments. Integrin α4β7+ cells were detected across multiple immune lineages and were enriched in the B-cell compartment. Between-group differences were observed in the CD4+ T-cell and B-cell compartments. VDZ responders showed enrichment of an α4β7+ effector memory CD4+ T-cell subset compared with VDZ-naive UC. In parallel, responders exhibited reduced frequencies of class-switched plasmablast and activated B-cell populations relative to VDZ-naive UC. Integrin-pairing analysis demonstrated that these B-cell populations contained both α4β7+ and α4β1+ cells. In contrast, VDZ non-responders showed enrichment of an α4β7+ atypical-like IgA+ memory B-cell subset relative to responders, VDZ-naive UC, and HCs. Conclusions Clinical response and non-response to VDZ in UC are associated with distinct circulating T- and B-cell immune phenotypes. Peripheral-blood immunophenotyping may identify correlates of response and non-response. Prospective studies are needed to determine whether these signatures are predictive before VDZ initiation or reflect pharmacodynamic changes after treatment.
Abstract Background and Aims Infliximab (IFX) trough concentrations are routinely measured to guide dosing and optimize efficacy. Bayesian forecasting tools may enable earlier prediction of trough concentrations, while patient-friendly capillary self-collection devices could improve access to precision-guided dosing (PGD). We evaluated the performance of the PredictrPK IFX platform in forecasting IFX trough concentrations by using serum collected with an upper-arm self-collection device and compared its performance with standard venous blood obtained by phlebotomy. Methods Patients with inflammatory bowel disease (IBD) receiving IFX maintenance therapy were recruited from a single IBD center. Midcycle samples collected via an upper-arm self-collection device and via traditional venipuncture were compared. The IFX clearance was estimated by use of a nonlinear mixed-effects model with Bayesian priors and the predicted IFX levels were compared with measured trough concentrations obtained prior to infusions. Deming regression was used for statistical analysis. Results The study included 57 patients (61% males, median age, 34 years). The median IFX dose was 10 mg/kg with a median interdose interval of 6 weeks. A total of 97 capillary and venous blood pairs were tested. Concentrations of IFX from venous and capillary blood samples demonstrated strong correlations (R2 = 0.987, Demings = 1.08). There was also a strong association between predicted IFX concentrations and measured IFX trough concentrations. Conclusion PredictrPK accurately predicted IFX trough concentrations. Capillary self-collection using an upper-arm device yielded results comparable to those from venous blood measurement, which may facilitate increased implementation and accessibility of PGD in clinical practice.
Abstract Background Physiologic pregnancy changes can alter commonly monitored laboratory values and mimic active inflammatory bowel disease (IBD), complicating prenatal disease monitoring. Biomarker trajectories in women with IBD in remission are not well established. We aimed to characterize trimester-specific changes in key biomarkers among women with IBD in remission before and throughout pregnancy. Methods We performed a retrospective review of pregnant women with IBD in remission enrolled in a prospective preconception/pregnancy registry. Eligible patients had laboratory data from the preconception period and at least 1 trimester (first trimester [T1], second trimester [T2], or third trimester [T3]). Parameters included complete blood counts, albumin, C-reactive protein, erythrocyte sedimentation rate (ESR), iron studies, and fecal calprotectin (FC). Changes from preconception were assessed using Wilcoxon signed rank tests and McNemar’s test when applicable. Results Sixty-eight pregnancies from 61 women were included (68.9% Crohn’s disease, 27.9% ulcerative colitis; median age 33 years; 83.8% on biologics). Compared with preconception, ESR and white blood cell count rose across trimesters (ESR PT2T3 < .001; white blood cell PT1-T3 < .001). Hemoglobin, hematocrit, platelets, and albumin declined significantly, particularly in T2/T3 (all P ≤ .008). C-reactive protein rose during all trimesters (PT1 = .003; PT2, T3 < .001), whereas FC remained stable and within normal range throughout pregnancy (all P > .05). Most biomarkers trended toward baseline postpartum. Conclusions In pregnant women with IBD in remission, most laboratory changes appear physiologic, requiring cautious interpretation. FC was the only stable biomarker throughout pregnancy and remains the most reliable marker of inflammation during this period. These trimester-specific data may improve interpretation of biomarkers in pregnant patients with IBD.
Abstract Background Patients with inflammatory bowel disease (IBD) frequently experience anxiety, pain, and reduced quality of life. Cognitive behavioral therapy (CBT) is effective for anxiety, but access may be limited. We evaluated the feasibility and explored the potential clinical impact of an artificial intelligence–powered virtual reality CBT-based application (XAIA [eXtended Reality Artificially Intelligent Ally]) in patients with IBD. Methods This was a multicenter, mixed-methods pilot study recruiting adults with IBD from the University of Chicago and Cedars-Sinai medical centers. Participants completed a single 30-minute session with the XAIA conversational avatar. Pain and anxiety were measured before and after the session using a visual analog pain scale and State–Trait Anxiety Inventory, respectively. Semi-structured interviews were conducted after the session to assess acceptability. Quantitative analysis was completed using paired t tests, and interview transcripts underwent thematic analysis. Results Thirty-four participants (median age 34 years, 71% women, 59% with Crohn’s disease) participated in this study. Qualitative analysis revealed high acceptability, with participants reporting feelings of safety, validation, and improved coping. Participants valued the immersive environment, conversational avatar, and ability to discuss sensitive topics without judgment. As an exploratory outcome, participants reported a statistically significant improvement in pain (−0.46; P = .011) and anxiety (−10.65; P < .001) immediately following a single session. Conclusions A single artificial intelligence–driven virtual reality CBT-based session in patients with IBD was feasible and well accepted and demonstrated preliminary evidence for acute reductions in pain and anxiety scores. Immersive digital behavioral health interventions may represent a scalable adjunct to traditional mental health support among adults with IBD.
Abstract Background Persistent symptoms in quiescent Crohn’s disease (qCD + S) lack clear mechanistic explanations. Building on prior evidence linking qCD + S to sulfidogenic microbial metabolism, this study characterizes the underlying site-specific mucosal transcriptional programs and host-microbe interactions associated with qCD + S. Methods We integrated paired metagenomic and transcriptomic data from a cross-sectional cohort of patients with qCD. Host-microbe interactions in the ileum and rectosigmoid colon were analyzed using differential expression, penalized regression, and sparse canonical correlation analysis. Results qCD + S was associated with transcriptional differences in separate analyses of ileal and rectosigmoid samples. The ileum showed reduced expression of genes related to immune surveillance, whereas the rectosigmoid colon exhibited reduced chemokine-related and increased ribosomal gene expression. Multi-omic integration identified associations between previously identified sulfidogenic microbes and mucosal genes related to mitochondrial metabolism, epithelial homeostasis, and neurosensory signaling. Exploratory sparse canonical correlation analysis identified multivariate patterns involving oral-associated taxa and mucosal genes related to B-cell receptor signaling and RNA-processing. Conclusions Persistent symptoms in qCD are mucosal transcriptional differences and fecal microbial features. These findings suggest a model in which altered immune regulation, microbial sulfur metabolism, and downstream epithelial and metabolic stress pathways contribute to symptom persistence despite quiescent inflammation, but this model requires mechanistic validation.
