P435 Real-world comparison of effectiveness between ustekinumab and a second anti-TNF agent in patients with symptomatic stenosing Crohn's disease after failure of a first anti-TNF agent: results of the USTEKNOSIS study

Buisson, A.(1)*;Jammet, A.(1);Pereira, B.(2);Fumery, M.(3);

(1)CHU Clermont-Ferrand, IBD Unit, Clermont-Ferrand, France;(2)CHU Clermont-Ferrand, Biostatistics Unit, Clermont-Ferrand, France;(3)CHU Amiens, IBD Unit, Amiens, France;


While surgery was considered as the reference for stricturing Crohn's disease (CD), anti-TNF agents are now the first-line treatment thanks to the CREOLE study.
However, there is no efficacy data after anti-TNF failure.

We aimed to compare the effectiveness of ustekinumab and a second anti-TNF agent after failure of a first anti-TNF in symptomatic stricturing Crohn’s disease (CD).


In this multicenter study, we retrospectively included all adult patients with CD treated with ustekinumab or anti-TNF for symptomatic stricture (confirmed on imaging or endoscopy) after prior exposure ≥ one anti-TNF for the current stricture, without surgery between the failure of the last anti-TNF and the study.

The primary endpoint was clinical remission (composite endpoint) at 6 months was defined as no pain, no vomiting, no food restriction, no sub-occlusive episode, no steroid, no surgery or discontinuation of treatment.

The long-term endpoints were discontinuation of treatment for failure, bowel damage progression and surgery.

The comparisons were performed after using propensity score analysis adjusted on potential confounders.


Overall, 70 patients were analyzed, including 34 in the ustekinumab group and 36 in the anti-TNF group (29 on infliximab and 7 on adalimumab). The two groups were similar for age (38.5 vs 37.6 years), CD duration (12.2 vs 13.5 years), female gender (41.2% vs 51.8%), CD location (p=0.43) and CD phenotype (p=0.38). Concomitant immunosuppressant was observed in 63.9% with an anti-TNF versus 8.8% with ustekinumab. The number (p=0.56) and the length (p=0.10) of stricture were also similar. The proportion of patients ≥ 2 prior biologics was higher in the ustekinumab group (41.2% vs 8.3%; p=0.001). The rate of primary failure to anti-TNF agent was comparable (16.7% vs 20.6%).

After adjustment based on propensity scores, the rate of clinical remission at 6 months was 73.9% and 42.7% (p = 0.24), under ustekinumab and anti-TNF, respectively. The predictive factors of remission in patients receiving ustekinumab were prior bowel resection (p=0.001) and length of stricture < 12 cm (p=0.042), while no predictor was found in those treated with anti-TNF agent. The risk of treatment discontinuation for failure (HR =2.86 [1.33-6.15]; p=0.008) or bowel damage progression (HR=3.90 [1.64-9.24]; p=0.003) were significantly greater in patients receiving another anti-TNF agent compared to those on ustekinumab. We did not observe any significant difference concerning the risk of surgery (HR=2.60 [0.70-9.61]; p=0.15).


Ustekinumab seems to be more effective than a 2nd anti-TNF to treat symptomatic stricturing CD after failure of a first anti-TNF. However, these data need to be confirmed by independent prospective data.