DOP038 Faecal calprotectin measurement and infliximab trough levels predict therapeutic evolution CD patients in clinical remission
X. Roblin*1, G. Duru2, L. Clavel1, M. Rinaudo3, M. Cuilleron4, C. Jarlot1, J.M. Phelip1, L. Peyrin Biroulet5, S. Paul3
1University Hospital, Gastroenterology, Saint Etienne, France, 2University Hospital, Statistic, Lyon, France, 3University of Saint Etienne, Immunology, Saint Etienne, France, 4University hospital, Radiology, Saint Etienne, France, 5CHU de Nancy, Department of Gastroenterology, Vandoeuvre-les-Nancy, France
The deep remission notion (clinical remission and mucosal healing) is an important objective for patients under treatment. The appearance of inflammation and pharmacological biomarkers could be a non-harmful way of predicting the evolution of Crohn's disease (CD). The aim of our study was to offer a predictive model for relapse in CD patients presenting clinical remission undergoing infliximab (IFX) treatment.
It was a prospective monocentric study that included all CD patients on IFX maintenance treatment (5mg/kg) and in clinical remission (CDAI< 150) for at least 16 weeks, between 2011 and 2014. On the day of the IFX infusion, all of these patients underwent a faecal calprotectin assay (Buhlmann technique), a CRP assay and pharmacological assays of IFX (ELISA , Theradiag). TLI ( > 2µg/ml) were considered therapeutic as well as CRP levels < 5mg/l and faecal calprotectin levels < 250 mg/g of stools. All of the patients included were followed up for a minimum of nine months. A CDAI score was calculated at each IFX infusion. A patient was defined in loss of response to IFX (LOR) when the CDAI was above 220, resulting in a change of treatment deemed necessary by the physician (IFX optimisation, change of medical treatment including the use of corticosteroids, surgery).
119 patients (mean age: 34 years, M:F sex ratio 1.2, mean duration of the disease 7.8 years) were included. The mean follow-up period was 20.4 months. 17% of the patients were on combotherapy (IFX and azathioprine). During follow-up, 37 patients (31.1%) out of the 119 relapsed, 78% within the first 6 months (mean period: 4.6 months). While the clinical characteristics of the relapsed and non-relapsed patients were similar, a univariate analysis isolated four significant factors predicting LOR: (CRP > 5mg/l (p=0.043), ATI > 20ng/ml (p< 0.001), LTI > 2 µg/ml (p< 0.001) and calprotectin > 250 µg/g stools (p<0.001)). After logistic regression, two independent factors were linked to a loss of clinical response: LTI < 2µg/ml (OR: 4,34 ; 95% CI: 1.28-10.7; p=0.001) and faecal calprotectin > 250µg/g stools (OR: 3.5; 95% CI: 1.5-8.7; p=0.001). In light of these results, a training cohort of 55 patients was isolated randomly in order to implement a predictive model for LOR in patients on IFX and in clinical remission. The combination of calprotectin > 250µg/g stools and TLI < 2µg/ml enabled to be predicted LOR in 95% of the cases within 6 months. This model was validated on the test cohort of 64 patients with a PPV of 95% and an NPV of 95%.
In IFX-treated CD patients and in clinical remission, a combination of TLI (< 2µg/ml) and faecal calprotectin (>250µg/g of stools) enable the prediction of LOR within 6 months in 95% of cases.