No detectable remission-specific remodeling of intestinal monocyte and macrophage metabolic transcriptional programs after adalimumab in Crohn’s disease: A patient-level reanalysis of a longitudinal single-cell atlas

We examined whether remission after adalimumab in Crohn’s disease (CD) was associated with differential change in intestinal monocyte and macrophage metabolic transcription. We reanalyzed the corrected public TAURUS atlas using 19 site-matched biopsy pairs from 15 patients (10 remission, 5 nonremission). Remission required at least two of Harvey–Bradshaw Index <5, Nancy index ≤1 and no ulceration; post-treatment sampling occurred at least eight weeks after treatment began. Seven programs were evaluated using an outcome-blind continuum, patient-level paired changes and exact permutation tests. Twelve eligible pairs, including five of six nonremission pairs, came from terminal ileum. No primary contrast survived correction (14 tests; minimum P = 0.397; all adjusted P = 0.987). The glycolysis-level difference in change was −0.005 score units (Welch 95% CI −0.052 to 0.042). Alternative scoring, trajectories and composition analyses also yielded no corrected program contrast. Published TAURUS benchmarks were reproduced within the discovery data. Batch explained 41.5% of patient outcome variation; within-batch inference relied on nine patients and 40 label allocations (minimum adjusted P = 0.350). Donor pseudobulk analysis tested 7,092 expressed genes with paired edgeR models allowing batch-specific time effects; none survived correction (minimum adjusted P = 0.894). None of 17 metabolic transcript modules survived correction of 34 directional and mixed tests (minimum adjusted P = 0.134). Including all 26 CD pairs and varying cell thresholds yielded no corrected program contrast. Empirical 80%-power detection thresholds were approximately 1.4–1.6 SD per primary endpoint, describing detection sensitivity without establishing equivalence or excluding effects of that size. Sparse colonic remission biopsies, batch imbalance and missing individual treatment metadata constrain interpretation. These findings concern RNA programs; they do not establish absence of metabolic remodeling, treatment causality or metabolic-flux effects.