Appraisal of the Pediatric Crohn's Disease Activity Index on Four Prospectively Collected Datasets: Recommended Cutoff Values and Clinimetric Properties

OBJECTIVES: The Pediatric Crohn's Disease Activity Index (PCDAI) is the outcome measure of choice in clinical trials of pediatric Crohn's disease. The aim of this study was to provide knowledge on its performance and accuracy of different cutoff scores. METHODS: Longitudinal data prospectively generated from four sources were used, including the REACH and budesonide trials, a North-American inflammatory bowel diseases (IBD) registry, and a cohort aimed at evaluating growth. Cutoff values of disease activity were determined by physician global assessment from the pooled cohort using serial receiver operator characteristic curves and area under the curve (AUC) as well as comparing the overall accuracy. Test-retest reliability and responsiveness were ascertained by comparing the baseline and follow-up scores, using an external anchor. RESULTS: A total of 437 children were included (268 (61%) males, mean age 12.9 +/- 2.6 years). To define remission, a composite definition of < 10 points or < 7.5 points without the height item had the highest accuracy; this addressed the limitation that height is not a responsive item. The best cutoff of 10-27.5 was determined for mild disease, 30-37.5 for moderate disease, 40-100 for severe disease, and a change of > 12.5 points for response (AUC 0.8-0.9; P < 0.001). Ninety children whose disease remained unchanged showed fair test-retest reliability (intraclass correlation coefficient = 0.74-0.8; P < 0.001). The PCDAI showed good responsiveness, as reflected from the correlational (r = 0.7; P < 0.001), distributional (Guyatt's responsiveness statistics = 0.9), and diagnostic utility analysis (AUC 0.85 (95% confidence interval 0.81-0.88). CONCLUSIONS: The clinimetric properties of the PCDAI are sufficient to support its use in clinical research. Cutoff values suggested by this study differ slightly from those previously published on much smaller cohorts.