Z Gastroenterol 2014; 52 - KG232
DOI: 10.1055/s-0034-1386254

Prediction of clinical outcome in Crohn's disease with confocal laser endomicroscopy: a prospective, observational, follow-up study

GE Tontini 1, J Mudter 1, M Vieth 2, R Atreya 1, C Günther 1, M Vecchi 3, MF Neurath 1, H Neumann 1
  • 1Universitätsklinik Erlangen, Erlangen, Germany
  • 2Klinikum Bayreuth, Bayreuth, Germany
  • 3IRCCS Policlinico San Donato, Mailand, Italy

Background: Assessment of prognostic factors in Crohn's disease (CD) patients is of important importance for early intervention and “treat to target” strategies. Confocal Laser Endomicroscopy (CLE) enables on demand in vivo characterization of architectural changes during endoscopy.

Aim: Prospective evaluation of CLE, endoscopic index of severity (CDEIS) and serum C-reactive protein (CRP) for prediction of clinical outcomes in CD.

Methods: Consecutive CD patients undergoing colonoscopy with fluoresceine-aided confocal imaging were enrolled in a blind, observational, follow-up study.

Results: Thirty-two CD patients were included (14 men; median age/range 37/18 – 60 years; mean distance from diagnosis 11 years). Baseline CRP was ≥5 mg/L in 46% and CDEIS ≥3 in 73% of patients. Mean follow-up period was 2.5 years (range = 6 – 48 months). Focal cryptitis and discontinuous crypt architectural abnormality were observed in 63% of patients. This finding showed a weak correlation with CDEIS (P= 0.068, RR = 1.6) and no correlation with CRP (P= 0.4, RR = 1.6). Focal cryptitis and discontinuous crypt architectural abnormality were significantly associated with an increased risk of medical treatment escalation with biologics, immunosuppressant or systemic steroids within 6 months (P= 0.045, RR = 2.0), showing 75% sensitivity and 70% specificity. This finding was also confirmed at 12 month follow up (P= 0.012, RR = 2.1; sensitivity = 76%, specificity = 78%). Patients with positive CLE findings developed significantly more transmural complications such as stenosis or perianal disease during the first 12 months (P= 0.023, RR = 6.0; sensitivity = 91%, specificity = 52%). Conversely, basal CDEIS ≥3 was only associated with treatment escalation at month 12th (P= 0.022, RR = 2.27; sensibility = 85%, specificity = 37%). CRP was not correlated with prognostic clinical outcomes.

Conclusions: In vivo characterization of CD-related signs of acute inflammation by means of CLE showed moderate correlation with CDEIS but not with CRP. CLE appeared as a good predictor of relevant clinical outcomes such as treatment escalation and transmural complications, performing better than CRP and CDEIS. This finding was substantial in the short term but disappeared after one year follow up.