OBJECTIVE

To identify predictors of peripheral arthropathy (PA) in patients with mild-to-moderate ulcerative colitis (UC) over a 12-month treatment period.

MATERIALS AND METHODS

A total of 118 patients with mild-to-moderate UC, as defined by the Mayo index and without baseline PA, were included prospectively and retrospectively. At enrollment, patients underwent evaluation of clinical complaints, medical history, extraintestinal manifestations, laboratory parameters, and endoscopic and histopathological findings. Univariate and multivariate analyses were performed to identify predictors of PA development within 12 months. A predictive model was constructed using binary logistic regression.

RESULTS

During the 12-month follow-up, PA developed in 55.9% of UC patients. Multivariate analysis demonstrated that baseline basal plasmacytosis on histology increased the risk of PA development 3.75-fold (95% CI: 1.07—13.18; p=0.04), whereas crypt distortion reduced the risk by 6.71-fold (OR: 0.15; 95% CI: 0.05—0.43; p<0.001). Longer UC duration at baseline was associated with a modest reduction in risk per additional month (OR: 0.990; 95% CI: 0.979—1.000; p=0.049). A history of complications (intestinal bleeding, toxic dilatation) was not significantly associated with PA development (p=0.999).

The final predictive model (p<0.001) yielded an area under the ROC curve of 0.874±0.034 (95% CI: 0.808—0.940), with a sensitivity of 80.4%, specificity of 80.9%, and overall accuracy of 80.6%.

CONCLUSION

The developed model demonstrated good discriminative ability with moderate sensitivity and specificity. Following external validation, it may serve as a useful adjunct prognostic tool in gastroenterology practice.