Redefining the intraepithelial lymphocytes threshold to diagnose gluten sensitivity in patients with architecturally normal duodenal histology

Aliment Pharmacol Ther 2011; 33: 697–706 Summary Background Accuracy of intraepithelial lymphocytes counts for diagnosing mild enteropathy coeliac disease in absence of villous atrophy can be limited by inappropriate controls included in the studies. Aim To determine the diagnostic accuracy of intraepithelial lymphocytes counts utilising controls lacking HLA coeliac disease-associated alleles. Methods Intraepithelial lymphocytes counting at villus tip and per 100 enterocytes was performed at haematoxylin and eosin (H&E) and CD3-stainings in: 29 cases (21 with potential coeliac disease and 8 affected by latent coeliac disease) representing the patient population and 14 noncoeliac controls lacking HLA-DQ2/DQ8 alleles. Results Threshold (mean + 2 s.d.) of duodenal intraepithelial lymphocytes at villus tip and per 100 enterocytes in noncoeliac controls was respectively: 3.5 and 18 at H&E, 3.2 and 17 following CD3-staining. Considering the whole patient population, the sensitivity of tip intraepithelial lymphocytes in detecting mild enteropathy coeliac disease was 90% (95% CI = 72.6–97.8) both at H&E and CD3-stainings. The sensitivity of intraepithelial lymphocytes per 100 enterocytes was 93% (95% CI = 77.2–99.2) both at H&E and CD3-staining. Specificity of both intraepithelial lymphocytes counts was 100% (95% CI = 76.8–100). Using a threshold of 25 intraepithelial lymphocytes per 100 enterocytes could miss 59% of cases at H&E and 48% following CD3-staining. Conclusions Intraepithelial lymphocytes counts are diagnostic feasible tools to detect mild enteropathy coeliac disease. Threshold of duodenal intraepithelial lymphocytes may be lower than currently accepted.