107
DEVELOPING AN ULCERATIVE COLITIS ENDOSCOPIC INDEX OF SEVERITY (UCEIS): RESULTS OF PILOT PHASE
C Bernhardt3, W Sandborn1, S Travis2, The UCEIS Study Group2, D Schnell3, P Krzeski3
1Mayo Clinic, Rochester, Minnesota, USA; 2John Radcliffe Hospital, Oxford, United Kingdom; 3Procter and Gamble Pharmaceuticals Inc., Mason, Ohio, USA
Aims: There are 9 different endoscopic indices of activity for ulcerative colitis (UC) developed for clinical trials; none have been validated. All 9 indices are subject to interobserver variation (IOV), which has been incompletely quantified. IOV potentially exerts an influence on trial outcome and patient management. The results of the pilot phase of an international initiative to validate an endoscopic index of activity are reported. The purpose of this pilot phase was to confirm the extent of IOV; compare four-point (normal, mild, moderate and severe) and visual analogue scales (VAS, 0-100mm); short (5min) vs. complete video clips; and a pre-study evaluation of individual index components.
Methods: Ten specialists in UC were asked to assess 16 sigmoidoscopies from a pool of 24 digitally recorded videos from a treatment trial that had predefined method and assessment of sigmoidoscopy and mucosal friability. Videos (from a total 334) were selected to represent normal to severe activity, based on the blinded, independent assessment of a central reader (CR). To avoid index-selection bias, the 10 assessors were asked to score severity according to their custom, without reference to different indices. The individual terms used in the Baron score were identified as present or absent allowing use of Baron’s score as the reference index. All assessors were blinded to symptomatic and histologic activity.
Results: Activity scores calculated from individual assessors differed from the CR Baron’s score by >= 2 grades in 3/157 assessments (2%); the median % of assessments (across assessors) that differed by >= 1 grade was 50% (range 31-63). 10 videos had a mean VAS < 20mm and 7/10 were rated normal by the CR; 3 scored 20-40mm and all were rated mild; 8 scored 41-70mm and 3/8 rated moderate; 3 scored > 70mm and 0/3 were rated severe by the CR. There was a trend towards an increase in VAS when evaluating a complete procedure vs. a short version. There was no consistency regarding the relative contribution of individual terms (erythema, granularity, friability, mucopus, ulceration and bleeding), as assessors are determining severity.
Conclusions: There is IOV in endoscopic scoring of UC activity among specialists in UC, either by Baron’s score or VAS. Nevertheless, mean VAS provided a reasonable indication of UC activity. IOV needs to be considered when assessing trial outcome.