Abstract Background Medications limiting the adhesion of ‘adherent and invasive E. coli’ (AIEC) represent potential strategies to treat Crohn’s disease (CD). However, the ileal AIEC identification is a time-consuming procedure, and the number of AIEC strains which colonise ileal CD mucosa remains unknown. There is an unmet need for non-invasive biomarkers to identify patients colonised by AIEC. We aimed to evaluate non-invasive biomarker of ileal AIEC colonisation in patients with CD. Methods This prospective and multi-centre study included CD patients requiring ileocoloscopy. Saliva, serum, stools and ileal biopsies were collected. Abundance and global invasive ability of ileal or faecal E. coli were performed. Isolated E. coli were characterised as AIEC or non-AIEC on I407 epithelial cells and THP1 macrophages. The ERIC-PCR profiles of ileal E. coli were performed. Ileal E. coli/CEACAM6 interaction was analysed by a yeast aggregation test and T84 assays (CEACAM6 protein expression, adhesion inhibition test with D-mannose). Quantification of serum anti-E. coli and ileal or salivary CEACAM6 was realised by ELISA. Results Overall, 102 CD patients were enrolled in this study and 25.8% of them exhibited ileal AIEC colonisation (AIEC+). The abundance and global invasive ability of ileal mucosa-associated E. coli were higher in AIEC+ CD patients compared with CD patients without AIEC (AIEC−) (p = 0.0065 and p = 0.0007, respectively). There was no difference between faecal abundance and invasive ability of E. coli between AIEC+ and AIEC− patients. The ERIC-PCR profiles of ileal E. coli showed that CD AIEC+ were for 78% of them colonised by not more than 2 clonal AIEC strains. In addition, AIEC were able to interact with CEACAM6 by binding D-mannose residues and to induce CEACAM6 expression in T84 cells (p = 0.0009 and p = 0.0185, vs. non-AIEC; respectively). This was also supported by adhesion inhibition test. Serum anti-E. coli level was higher for CD AIEC+ (vs. CD AIEC-). Ileal CEACAM6 level were positively correlated with abundance of ileal associated E. coli in AIEC+ patients (r = 0.4000; p = 0.0362) and with salivary CEACAM6 level (r = 0.4690; p < 0.0001). The non-invasive biomarker ‘serum anti-E.coli/salivary CEACAM6’ index was higher for CD AIEC+ (p = 0.0174; vs. CD AIEC-). A cut-off value < 1.34 × 10−6 eliminated the presence of ileal AIEC with a high negative predictive value (90% CI95% [69%–95%]). Conclusion Our study reported that identification of faecal AIEC cannot replace identification of AIEC from ileal biopsies, most of AIEC infection are mono or bi-clonal (≤ 2 strains) and that non-invasive biomarker such as ‘serum anti-E.coli/salivary CEACAM6’ index could be helpful to screen CD patients for AIEC infection.