Micronutrients are associated with endoscopic postoperative recurrence in Crohn's disease: a multicenter prospective...

Diet may influence the disease course in inflammatory bowel disease, but its role in postoperative outcomes for Crohn's disease (CD) remains unclear. This study aimed to assess the association of macro- and micronutrient intake with endoscopic postoperative recurrence (ePOR) in a prospective multicenter cohort.

Patients with CD following ileocolic resection were prospectively recruited from six North American centers. Primary study outcome was ePOR (modified Rutgeerts' score ≥i2a) during follow-up. Nutritional intake was assessed through 2-day food diaries verified by dietitian-delivered interview. Associations between nutrient intake and ePOR were evaluated. Random forest models with 10-fold cross-validation assessed the predictive value of nutrients and clinical factors for ePOR.

A total of 520 food diaries from 103 patients were analyzed; 37 patients (36%) experienced ePOR. Univariate analysis identified 8 nutrients associated with ePOR including lower intake of isoflavones (genistein, daidzein, glycitein; P < .01), inositol (P < .01), pinitol (P = .02), provitamin-A carotenoid (P < .01), xylitol (P = .03) and parinaric Acid (P = .03). Sensitivity analysis confirmed the association of lower intake of all isoflavones (P < .01) and pinitol (P = .04) with ePOR at first postoperative colonoscopy. Random forest models showed poor discrimination for clinical factors alone (area under the curve [AUC] 0.59) but an acceptable discrimination for nutrients alone (AUC 0.67), which improved when combining nutrients and clinical factors (AUC 0.71).

Lower intake of specific micronutrients is associated with ePOR in CD patients. A machine-learning model combining nutrient intake and clinical factors enhanced the prediction of ePOR. These findings highlight the importance of postoperative nutritional assessment and suggest dietary interventions may help prevent postoperative recurrence in CD.