Exabis Library

Prior data have suggested that 5-aminosalicylates (5-ASA) may be associated with an increased risk of severe COVID-19 among inflammatory bowel disease (IBD) patients. We aimed to evaluate the association of 5-ASA with severe COVID-19 in a large cohort of IBD patients.

We analyzed data from the Surveillance Epidemiology of Coronavirus Under Research Exclusion for Inflammatory Bowel Disease (SECURE-IBD) registry, a large, international database of IBD patients with confirmed COVID-19. The primary outcome was severe COVID-19, defined as intensive care unit admission, ventilator use, and/or death. Hospitalization due to COVID-19 was a secondary outcome. We performed multivariable regression modeling with a generalized estimating equation accounting for country as a random effect to analyze the association of 5-ASA with severe COVID-19. Models a priori included age, sex, race, disease phenotype (CD or UC/IBD-U), corticosteroid use, azathioprine/6-mercaptopurine use, TNF antagonist use, disease activity by physician global assessment, number of comorbidities, and days from SECURE-IBD inception to reporting. We constructed three models examining 5-ASA use as binary covariate using 1) all patients, 2) only patients on any biologic, and 3) only patients on TNF antagonists.

5,174patients were included with 212 (4.1%) severe COVID-19 events. At the time of COVID-19 infection, 1,504 patients were taking 5-ASA. 5-ASA patients were older (mean age 44 vs. 38.3 years, p<0.001), more likely to have UC (70.7% vs. 27.7%, p<0.001), less likely to be in remission (49.6% vs. 57.2%, p<0.001), and more likely to have at least one comorbidity (33.6% vs. 26.7%, p<0.001) compared to patients not on 5-ASA. 3,325 patients were on any biologic and 2,216 were on a TNF antagonist. Among all patients, 5-ASA was not associated with severe COVID-19 (adjusted OR [aOR] 1.14, 95% confidence interval [CI] 0.86-1.52) (Table 1). Prior associations of age, comorbidities, TNF antagonists, and corticosteroids with severe COVID-19 were similar to prior analyses (Table 1). In analyses restricting to those on any biologic or only TNF antagonists, there was also no significant association between 5-ASA and severe COVID-19 (aOR 0.76, 95% CI 0.38-1.50 and aOR 0.99, 95% CI 0.43-2.32, respectively). Use of 5-ASA was not associated with risk of COVID-19 related hospitalization in any analysis.

In an analysis of updated data from the SECURE-IBD registry, 5-ASA use was not associated with worse outcomes among IBD patients with COVID-19.

Evaluation of endoscopic disease severity is a key component in the management of ulcerative colitis (UC) patients.  However, endoscopic assessment suffers from substantial intra- and interobserver variation, up to 75 %, thereby limiting the reliability of individual assessments. Our aim was to develop an artificial intelligence (AI) model capable of distinguishing active from healed mucosa as well as to differentiate different levels of endoscopic disease activity.

1484 unique endoscopic images from 467 patients were extracted for classification. Two experts classified all images independent of each other according to the Mayo endoscopic subscore (MES). In case of disagreement, a third expert classified the images.

Different convolutional neural network architectures were implied in the development of the AI model. Five-fold cross-validation was employed to select the best model. Unseen test data were used for evaluation.

The final model was evaluated on its performance for distinguishing MES 0 from 1–3, MES 0–1 (i.e. mucosal healing) from 2–3, and distinguish between all MES.

The accuracy, sensitivity, specificity, positive and negative predictive value, and Cohen’s Kappa were used to evaluate the final models.

Our final model achieved at the most difficult task (distinguishing between all 4 categories of MES) a mean accuracy of 0.82, mean AUC of 0.99, test accuracy of 0.84, a sensitivity of 0.88, and a specificity of 0.81 and a weighted Cohens Kappa of 0.83 (p<0.001 compared to the experts).

The results from the other tasks are shown in table 1.

We propose a new standardised way of evaluating endoscopic images from UC patients for both clinical and academic purposes. The proposed AI model demonstrated a very good capability of distinguishing between all 4 MES levels of activity. This will optimize and unify the evaluation of the disease severity measured by the Mayo endoscopic subscore across all centres and hospitals no matter the level of medical expertise.