P262 Decreased infliximab trough levels (TLI) is a moderate predictor of loss of response to IFX in IBD patients.

Roblin, X.(1);Berger, A.E.(2);Bastide, L.(1);Tournier, Q.(1);Veyrard, P.(1);Paul, S.(2);

(1)CHU Saint Etienne - Hospital Nord, Department of Gastroenterology and Hepatology- Inserm- CIC1408, Saint Priest en Jarez, France;(2)CHU Saint Etienne - Hospital Nord, Immunology - Inserm- CIC1408, Saint Priest en Jarez, France;


If the use of anti-TNF trough levels (TDM) are proposed in case of loss of response (reactive attitude), a proactive strategy has not shown in most randomized trials its interest. The aim of this study was to analyze the predictive value of a decrease of TLI in IBD patients in terms of loss of response over time.


This was a retrospective study that included all IBD patient followed in our university hospital. Eligible patients had to have IBD treated with IFX as maintenance therapy (5mg/kg/8weeks). Patients had to be in clinical remission of the disease at inclusion with calprotectin levels < 100 µg/g stool and therapeutic levels of IFX concentration (TLI > 5 µg/ml) on IFX mono- or combination therapy. IFX levels were analyzed by ELISA (Theradiag, France). Every two months, patients were reviewed in the day hospital with fecal calprotectin, TLI and a measure of the clinical activity score (CDAI for CD and Mayo clinical score for UC). Loss of response was defined for CD as a CDAI > 220 with fecal calprotectin > 250µg/g stool and for UC as a full Mayo score > 5 with an endoscopic Mayo score > 1. Low TLI during follow-up was defined as a level below 3 µg/mL.  Patients who had dose optimization based on calprotectin or TLI alone were excluded from the study.


145 patients (65% CD, mean age 34 years, disease duration 4.5 years) were included. The mean follow-up was 66 months. 79 patients (55%) showed a loss of response. A decrease of TLI was found in 65% of patients who relapsed within 6 months versus 33% in patients in long-term clinical remission (p=0.14).  The predictive value of loss of response within 6 months was moderate in the presence of low IFX levels (sensitivity: 71%, specificity: 73%; PPV: 73%, NPV: 68%). Conversely, fecal calprotectin (> 250 µg/g stools) increase was strongly predictive of loss of response within 6 months (sensitivity: 85%, specificity: 85%).  The association of the two markers (decrease of TLI and/or increase of fecal calprotectin) improved the predictivity of relapse in the follow-up (sensitivity: 90%, Specificity: 89%). For the 79 patients who relapsed, for 35% (28 patients) of the cases, the rise of fecal calprotectin preceded the fall of TLI but for 25% of the cases (25 patients), apparition of low TLI preceded the rise of fecal calprotectin.


The decrease of TLI is n moderately  predictive of a loss of clinical response in the short or medium term under IFX for IBD. These results may explain the negative results of a proactive attitude based on TDM alone. A proactive strategy based on TDM and fecal calprotectin could be very interesting but needs to be confirmed.