Search in the Abstract Database

Abstracts Search 2019

P143 Measuring the mediating effects of tofacitinib on health status in ulcerative colitis: data from the OCTAVE programme

M. Dubinsky1, A. Bushmakin2, M. DiBonaventura3, J. Cappelleri4, L. Salese5, E. Maller5, A. Armuzzi*6

1Mount Sinai, New York, USA, 2Pfizer Inc., New York, USA, 3Pfizer Inc., Patient and Health Impact, New York, USA, 4Pfizer Inc., Groton, USA, 5Pfizer Inc., Collegeville, USA, 6Fondazione Policlinico Gemelli IRCCS – Università Cattolica del Sacro Cuore, Rome, Italy


The Mayo score (MS) is typically used to calculate ulcerative colitis (UC) clinical trial efficacy endpoints and includes four components: stool frequency, rectal bleeding, endoscopic appearance, and physician assessment. Although generic patient-reported outcome measures (PROMs) like the Short Form-36 (SF-36) are also frequently included in UC trials, it is unclear whether treatment effects on these measures are fully explained by MS changes or whether other unobserved variables are in play. Here, we explored the interrelationship among treatment, SF-36 domains and MS using a mediation modelling framework.


Pooled data at the end (Week 8) of the two double-blind, identically designed induction studies of tofacitinib (OCTAVE Induction 1 and 2, NCT01465763 and NCT01458951) were used. Tofacitinib is an oral, small-molecule Janus kinase inhibitor approved in several countries for the treatment of ulcerative colitis (UC). A mediation model was specified such that the MS components served as mediators between treatments (active treatment vs. placebo) and the eight SF-36 domain scores (bodily pain, general health, mental health, physical functioning, emotional role limitations, physical role limitations, social functioning, and vitality), which served as the outcomes. Our primary interest was the extent to which treatment affects the SF-36 domain outside of any change in MS components (ie the direct path).


In total, 1079 patients with moderately to severely active UC were included. For all SF-36 domains, the indirect path (ie the pathways from treatment to the MS components and then to each SF-36 domain score) was significant (all p < 0.05) and explained 65.6% (bodily pain) to 92.9% (mental health) of the total effect of the treatment on SF-36 domain scores. In other words, the majority of the total effect of treatment on the SF-36 scores was explained by changes in Mayo score components. Yet, for bodily pain (34.4%), physical role limitations (31.2%) and vitality (32.7%), the direct paths (ie, the pathway from treatment directly to each SF-36 domain outside of any effect from changes in MS components) were also significant (all p < 0.05). No other direct effects were observed.


Our study suggests that the MS, while important in capturing disease activity, does not fully mediate treatment effects on all SF-36 domains. Hence, the results indicate that tofacitinib can directly improve certain aspects of general health status—specifically, bodily pain, physical role limitation and vitality—outside of any benefit of improving stool frequency, rectal bleeding, endoscopic assessment, or physician assessment. These results reinforce the value of health status PROMs such as the SF-36 in capturing the full benefit of UC treatment.