P682 Development of anti-ustekinumab antibodies is an independent factor associated to loss of response to ustekinumab therapy in IBD patients
Roblin, X.(1)*;Nancey, S.(2);Bastide, L.(3);Veyrard, P.(3);Bischetti, G.(4);Berger, A.E.(5);Barrau, M.(6);Paul, S.(5);
(1)CHU Saint Etienne - Hospital Nord, Department of Gastroenterology and Hepatology- Inserm- CIC1408, Saint Priest en Jarez, France;(2)CHU Lyon Sud, Gastroenterelogy, Lyon, France;(3)CHU Saint Etienne, Gastroenterology, Saint Etiennne, France;(4)CHU Lyon Sud, Gastroneterology, Lyon, France;(5)CHU Saint Etienne, Immunology, Saint Etiennne, France;(6)CHU Saint Etienne, Gastronetrology, Saint Etiennne, France;
Background
Monitoring of antibodies (ADA) under ustekinumab therapy in routine practice is not recommended in IBD patients.
Aim of study : to investigate the relationship between ADA detected by a drug-tolerant assay and loss of response to therapy in a cohort of IBD patients treated with ustekinumab.
Methods
This was a retrospective study enrolling consecutively all adult patients with moderate to severe active IBD and having a follow up period of at least 2 years after ustekinumab induction. Loss of clinical response was defined as CDAI > 220 or HBI > 4 for CD and partial Mayo subscore > 3 for UC associated with the decision to modify disease management or treatment.
Results
Ninety patients were included (78 CD and 12 UC, mean age 37 years). Median levels of anti-ustekinumab antibodies (AUA) were significantly higher in patients with LOR compared with those in responder patients (15.2 µg/mL-eq CI (7.9-21.5) and 4.7 µg/mL-eq CI (2.1-10.5), respectively; p=0.04). The area under the ROC curve (AUROC) for AUA in predicting LOR was 0.76. The optimal cut-off point capable to accurately identify patients with LOR was 9.5 µg/mL-eq with a sensitivity of 80 % and specificity of 85 %. In multivariate analysis, serum AUA ≥ 9.5 µg/mL-eq (HR=0.84), prior vedolizumab exposition (HR=0.78) and azathioprine (HR=3.78) exposures were the only independent factors associated with subsequent clinical response to ustekinumab therapy.
Table 1 : univariate and multivariate analysis to isolate factors associtaed with subsequent clinical response to ustekinumab therapy
| Univariate analysis | Multivariate analysis | |||
| p | HR [95% CI] | p | HR [95% CI] | |
| Gender (male vs female) | 0.87 | 0.95 [0.29-4.02] | ||
| Age | 0.76 | 0.88 [0.56-3.55] | ||
| TLU < or > 4.5 µg/mL | 0. 45 | 1.18 [0.74-4.23] | ||
| AUA ≥ 9.5 μg/mL-eq | 0.045 | 0.81 [0.35-0.91] | 0.013 | 0.84[0,80;0,97] |
| UC vs CD | 0.3 | 1.4 [0.74-2.65] | ||
| Vedolizumab before | 0.06 | 0.86 [0.06-0.89] | 0.015 | 0.78 [0.075; 0,734] |
| AZA before | 0.028 | 3.86 [1.12-6.32] | 0.014 | 3.78 [1.201;12.256] |
| Active smoking | 0.65 | 1.12 [0.40-3.55] | ||
| Age (≤ vs > 40 years) | 0.32 | 1.41 [0.48-2.32] | ||
| Duration of disease | 0.45 | 0.67 [0.26-9.21] | ||
| Two lines of anti-TNF | 0.67 | 1.12 [0.32-8.72] | ||
Conclusion
In our real-life cohort, AUA was identified as an independent predictor of LOR and subsequent drug failure to treatment