Association between Crohn’s disease and pulmonary nodule patterns: an inverse probability-weighted analysis of real-world data

Background: Crohn’s disease (CD) is a systemic inflammatory disorder with potential pulmonary involvement; however, its association with pulmonary nodules and how disease phenotype, medication, and inflammation affect nodule characteristics are unclear. Using real-world data, we aimed to comprehensively assess the burden of pulmonary nodules among patients with CD, characterize their longitudinal dynamic trajectories, and identify key determinants of their imaging phenotypes. Methods: In this retrospective multicenter cohort study of 2,009 patients with CD and 4,002 controls, we used inverse probability weighting (IPW) to compare the nodule burden and its dynamics over time in a longitudinal subcohort (n = 2,956). We assessed the associations of the Montreal classification, biologics, corticosteroids, C-reactive protein, and extraintestinal manifestations with nodule features among patients with CD and explored risk factors for high-risk nodules related to CD. Results: CD was associated with higher nodule counts across all subtypes (e.g., solid nodules: incidence rate ratio [IRR], 1.80; 95% confidence interval [CI], 1.62–1.98; P < 0.001). Longitudinally, compared with controls, patients with CD experienced greater reductions in solid nodules and nodules <6 mm (P < 0.001) despite the slightly faster growth of persistent nodules <6 mm. Disease phenotypes and extraintestinal manifestations were not associated with nodule features. Age was an independent predictor and the strongest predictor of the pulmonary nodule burden. Adalimumab was correlated with fewer non-solid and 6- to 8-mm nodules (IRR, 0.615–0.676; P < 0.01). CD-related clinical factors had limited utility for predicting high-risk nodules. Conclusions: CD confers an increased pulmonary nodule burden with a unique dynamic evolution. Risk stratification should prioritize age and medication exposures instead of disease phenotype because phenotypic markers lack predictive value for nodule outcomes.