DOP76 Real-world multicenter comparison of effectiveness between tofacitinib and vedolizumab in patients with Ulcerative Colitis after failure to at least one anti-TNF agent

Buisson, A.(1);Nachury, M.(2);Fumery, M.(3);Guilmoteau, T.(1);Leclerc, E.(1);Altwegg, R.(4);Serrero, M.(5);Mathieu, N.(6);Treton, X.(7);Vuitton, L.(8);Pereira, B.(9);Amiot, A.(10);Bouguen, G.(11);

(1)University Hospital Estaing, Department of Gastroenterology- IBD Unit, Clermont-Ferrand, France;(2)CHRU Lille, IBD Unit, Lille, France;(3)CHU Amiens, IBD Unit, Amiens, France;(4)CHU Montpellier, IBD Unit, Montpellier, France;(5)AP-HM, IBD Unit, Marseille, France;(6)CHU Grenoble, IBD Unit, Grenoble, France;(7)AP-HP Beaujon Hospital, IBD Unit, Paris, France;(8)CHU Besançon, IBD Unit, Besançon, France;(9)University Hospital Estaing, DRCI- Biostatistics Unit, Clermont-Ferrand, France;(10)AP-HP Creteil Hospital, IBD Unit, Paris, France;(11)CHU Rennes, IBD Unit, Rennes, France;


Several therapeutic options are now available in ulcerative colitis after anti-TNF failure, but no data compared hitherto tofacitinib and vedolizumab.

We compared the effectiveness of tofacitinib and vedolizumab in UC patients with prior exposure ≥ 1 anti-TNF.


In this multicentre retrospective study, we consecutively included all adult UC patients with partial Mayo score > 2, with ≥ 1 prior anti-TNF agent and started either tofacitinib (10mg b.i.d ± decreased to 5 mg b.i.d from week 8 (W8)) or vedolizumab (300 mg IV at W0-W2-W6 -W14 [± additional W10]) between January 2019 and June 2021.

The primary endpoint was corticosteroid-free clinical remission or CFREM (partial Mayo score ≤ 2) at W16. Secondary endpoints were endoscopic improvement (CFREM + endoscopic Mayo score ≤ 1) and mucosal healing (CFREM + endoscopic and histological remission defined as Nancy index ≤ 1).

All the comparisons were performed using propensity score analyses (inverse probability of treatment weighting) adjusted on gender, smoking, UC duration and extent, number of prior biologics or prior primary failure to biologics, concomitant 5-ASA, steroids or immunosuppressive agents, and disease severity.


Overall, 400 patients will be included. Among the 200 first patients, 87 and 112 received tofacitinib and vedolizumab, respectively (one missing patient). Except for more pancolitis (54.0% vs 38.4%, p=0.028), less immunosuppressive therapies (4.6% vs 27.7%, p < 0.001), and higher rate of prior exposure ≥ 2 biologics (87.4% vs 37.5%, p < 0.001) in tofacitinib arm, baseline characteristics were similar across the two groups including concomitant 5-ASA (10.3% vs 18.8%) and steroids (23.0% vs 31.2%). Vedolizumab infusion at W10 was performed in 34.3% while 42.5% received tofacitinib 10 mg b.i.d until W16.

CFREM was achieved in 54.2% and 42.5% in tofacitinib and vedolizumab, respectively p=0.089). The rate of CFREM at W16 was 57.4% vs 51.1% (p=0.77) after one biologic, 55.4% vs 41.8% (p=0.61) after 2 biologics, 56.9% vs 6.3% (p=0.007) after at least 3 biologics, and 59.0% vs 33.3% (p=0.17) in the subgroup with partial Mayo score ≥ 6, in tofacitinib and vedolizumab groups, respectively.

Tofacitinib was more effective than vedolizumab to achieve CFREM at W16 in patients with primary failure to at least biologic (71.6% vs 30.8%, p=0.049).

Among 177 patients, endoscopic improvement was higher in patients treated with tofacitinib (33.6 % vs 7.1%, p=0.048). Mucosal healing was observed in 6.4% vs 3.8% in tofacitinib and vedolizumab arms, respectively (p=0.27).


Tofacitinib and vedolizumab are effective after failure to anti-TNF agents. Tofacitinib seems to be more effective in case of primary failure to biologics and multiple therapeutic failure.