Search in the Abstract Database

Abstracts Search 2019

P249 Validation of a Novel Integral Disease Index (NIDI) for evaluating the grade of activity in Mexican patients with ulcerative colitis: a prospective cohort study

J. Yamamoto-Furusho*1, K. Bozada-Gutiérrez1, F. Bojalil-Romano1, R. Barreto-Zuñiga2, B. Martínez-Benitez3

1IBD Clinic, Department of Gastroenterology, Instituto Nacional de Ciencias Medicas y Nutricion, Gastroenterology, Mexico, Mexico, 2Instituto Nacional de Ciencias Medicas y Nutricion, Endoscopy, Mexico, Mexico, 3Instituto Nacional de Ciencias Medicas y Nutricion, Pathology, Mexico, Mexico

Background

ulcerative colitis (UC) is a chronic condition characterised by affecting the colon. Clinical disease activity is a complex concept that relies on clinometric assess tools that can be traduced into a score to be used objectively to predict the risk of an individual patient with variables obtained from routine medical visit. However, clinical evaluation with current index scores could have a poor concordance with objective diagnostic tools such as serological biomarkers, colonoscopy and histology findings. The aim of this study was to validate an integral activity index for UC patients.

Methods

This is a prospective cohort study that included 222 patients with definite diagnosis of UC confirmed by histopathology where 546 evaluations were analysed at basal and at least one follow-up visit for each patient. NIDI index was made up considering six categories: (1) Number of stools with blood per day; (2) haemoglobin; (3) high-sensitive C-reactive protein; (4) albumin; (5) endoscopic findings and (6) histological findings. Each variable was graded on a score from 0 to 3. The range of the NIDI Index is from 0 points to 18 points and it was distributed in four categories: (1) remission (0 to 3 points); mild activity (4 to 6 points); moderate activity (7 to 12 points) and severe activity (13 to 18 points). The validation and reliability analysis was done with the principal components analysis and Cronbach Α coefficient for internal consistency and average correlation of individual items. Finally, we calculated the receiver-operating characteristic (ROC) based analysis to define their sensibility, specificity, positive predictive value (PPV), negative predicted value (NVP) and area under the ROC curve (AUC) with 95% confidence intervals (CI). The STATA SE 11.1 statistical Programme was used.

Results

The main component analysis of the six items included in the NIDI Index revealed that the scale is most likely unidimensional with 49.9% of the variance explained by a unique component. An adequate internal consistency was observed with a Cronbach α of 0.78 and an acceptable average inter-correlation for the typified items (r = 0.47, p < 0.05). A good internal consistency was found with a Cronbach α of 0.78 and an acceptable average inter-correlation for the typified items (r = 0.47, p < 0.05). The overall efficacy of the new score was 87.2% of correctly classified patients with an AUC according to the three scenarios described of 0.93, 0.92 and 0.96, respectively. All items analysed had a good discriminative capacity with specificity range between 0.61 and 0.93 and sensibility from 0.25 to 0.94.

Conclusion

The NIDI provides an integral view of UC activity and it will be useful in the optimisation of medical treatment in UC patients.