Search in the Abstract Database

Search Abstracts 2015

* = Presenting author

P164 Evaluation of an interim Crohn's disease outcome measure (PRO-2) based on two patient-reported components (stool frequency, abdominal pain) of the Crohn's Disease Activity Index (CDAI) in the ustekinumab CERTIFI Study

C. Gasink1, J. Friedman1, L.L. Gao1, D. Chan*1, W.J. Sandborn2, B. Feagan3

1Janssen Research and Development, LLC, Immunology, Spring House, United States, 2University of California San Diego, Division of Gastroenterology , La Jolla, United States, 3Robarts Clinical Trials Inc., Robarts Research Institute, University of Western Ontario, London, Ontario, Canada


The CDAI has been the standard outcome measure in CD trials for ~40 years, but its complexity, high PBO response, and inconsistency with FDA guidance on patient-reported outcome (PRO) measures are disadvantages. This has generated interest in exploring a PRO based on the stool frequency(SF) and abdominal pain(AP) components of the CDAI, known as PRO-2. A potential PRO-2 definition for remission (mean daily score of AP ≤ 1 AND[also] SF ≤ 1.5) was proposed based on data from two CD populations, one with mild-moderate disease and one with steroid-dependent disease. Whether this PRO-2 definition is optimal in patients with moderate-to-severe disease, and/or refractory to TNF α antagonists is unknown.


Using the previously reported 8-week induction data from the ustekinumab(UST) Phase 2 CERTIFI study (in patients previously failing and/or intolerant to TNF-antagonists; mean baseline CDAI=324), the contributions of SF/AP to baseline CDAI and subsequent improvements post-treatment were assessed. Sensitivity(SENS)/specificity(SPEC) analyses that defined optimal cutoffs for detecting remission(CDAI<150) at week 8 were used to generate remission definitions based on unweighted mean daily scores for AP alone, SF alone, or both: Either requiring maximums for each(e.g. AP ≤ 1 AND SF ≤ 1.5)or, applying maximum thresholds to the sum of AP+SF, using standard CDAI weightings. The treatment effects of UST in CERTIFI were then evaluated and compared using the selected definitions.


The unweighted mean daily score cutoffs of AP ≤ 1/SF ≤ 3/combined AP ≤ 1 AND SF ≤ 3,respectively, provided better balance of SENS(87.1%;84.7%;72.9%)and SPEC(77.1%;73.9%;94.1%) compared with the previously proposed cutoff of combined AP ≤ 1 AND SF ≤ 1.5(55.3% SENS;96.8% SPEC) in the detection of remission at week 8. A weighted total AP+SF score cutoff of <75 best optimized SENS(91.8%)&SPEC(85.3%) compared with the unweighted mean daily AP/SF cutoffs. Candidate remission definitions all demonstrated significant treatment effects(UST vs. PBO):total AP+SF<75 (29.4% vs.16.7%, p=0.004),combined AP ≤ 1 AND SF ≤ 3(19.3% vs.9.1%, p=0.007), combined AP ≤ 1 AND SF ≤ 1.5(13.7% vs.5.3%, p=0.009).


Definitions of PRO-2 remission based on SF&AP CDAI components are sensitive and specific in an anti-TNF refractory/intolerant population with high CD activity. In this population,the PRO-2 definition of AP ≤ 1 AND SF ≤ 3 better reflects clinical remission than the previously proposed AP ≤ 1 AND SF ≤ 1.5 threshold, identified in the mild-to-moderate patient populations. A weighted total AP+SF score <75 may be a good PRO-2 definition of remission, because it provides the best overall balance of SENS/SPEC and facilitates comparisons with studies utilizing the full CDAI.