P310 Development of a new clinico-biological score associated with disease activity in patients with inflammatory bowel disease treated with thiopurine

    loading  Checking for direct PDF access through Ovid


Background: To date, there is no simple surrogate biomarker for monitoring thiopurine efficacy in patients with inflammatory bowel diseases (IBD).

Aims. To develop a clinico-biological score that would be associated with disease activity in IBD patients treated with thiopurine.

Methods: This was a multicenter translational study (NCT02367326) performed in 2015 that included all patients with Crohn's disease (CD) or ulcerative colitis (UC), naive of biologics, treated for at least three previous months with a thiopurine. Additional corticotherapy was allowed. At inclusion, patients were defined in remission with HBI <5 and CDEIS <4 for CD, and with total Mayo score ≤2 for UC. Blood tests performed at inclusion were prospectively collected. Best clinico-biological score model was reached based on cut-offs that would be associated with disease activity for each numeric variables determined by ROC curves, and on weights of each variables in the association of disease activity using random Forests and logistic regression analyses. Diagnostic performances of this score for active disease were then assessed by ROC curve in the overall population.

Results: A total of 339 eligible patients was recorded (median age: 37 years [27–48.5]; female: 58.4%; CD: 68.4%). Clinical characteristics were similar for patients with active or inactive disease at inclusion. Statistical significant differences were observed between the two groups for median values of inflammation-related parameters (including MCV, CRP), and liver function test (ALP, AST). In contrast, excluding MCV, there was no difference between the two groups regarding median values of thiopurine metabolism-related parameters (lymphocytes, 6-TGN (n=76), 6-MMP (n=61)). The following model of clinico-biological score reached by the addition of attributed points (pts) per variable was assessed: BMI<30, ALT ≤50 UI/L (6 pts), RBC >3 (10 pts), MCV<90fL, CRP >3mg/L (5 pts), lymphocytes ≥1200/mm3, GGT ≥120 UI/L, white leucocytes ≥6000/mm3 (3 pts), age <52 years, AST ≤50 UI/L, ALP >100 UI/L, neutrophils ≥3700/mm3 (1 pt). The area under the curve (AUC) was 80% [95% CI: 74.2%-85.7%], with sensitivity and specificity of 84.7%, and 65.6%, respectively, for a score over 30 points. Positive (PPV) and negative (NPV) predictive values were 48.4%, and 91.6%, respectively. This low PPV may be compensated by that of CRP assay for a cut-off at 4mg/L (PPV: 82%; AUC: 71% [95% CI: 64.6%-77.4%]. However, diagnostic performances of this new clinico-biological score were overall better than CRP for disease activity (p=0.041).

Conclusions: This new simple clinico-biological score appears associated with disease activity in IBD patients with ongoing thiopurine therapy, and should be validated in an independent prospective cohort.

Related Topics

    loading  Loading Related Articles