Subclinical or latent cases of gluten-sensitive enteropathy (GSE) are difficult to diagnose, and serology-positive, histology-negative (minimal morphologic change) and serology-negative, histology-positive patients have been identified. Both, particularly the histology-negative group, require the correct diagnosis for proper management, especially because the concept of minimal histologic change GSE has escaped attention in standard textbooks. We assessed the numbers and distribution of intraepithelial T cells and their subsets with CD3, CD8, and CD4 immunostaining and examined for crypt hyperplasia with mitotic and Ki-67 proliferative indices with the aim of refining the criteria for the diagnosis of minimal change GSE. Duodenal biopsies from 46 clinically suspected cases of GSE tested for antigliadin, antiendomysium, and antitissue transglutaminase antibodies were divided into four groups: serology-positive, histology-positive (S+H+, n = 20); serology-positive, histology-negative (S+H-, n = 22), representing the minimal morphologic change group; serology-negative, histology-positive (S-H+, n = 4); and serology-negative, histology-negative (S-H-, n = 28), controls with histologically normal duodenal biopsies obtained for unrelated reasons. The numbers of CD3+ and CD8+ intraepithelial T cells (IETCs) were significantly higher in histology-positive biopsies with (mean, 40.3/100 and 39.3/100 enterocytes, respectively) and without positive serology (mean, 33.3/100 and 35/100 enterocytes, respectively) compared with all other groups (S+H-, mean, 26.5/100 and 24.3/100 enterocytes, respectively; S-H-, mean, 23.3/100 and 17.9/100 enterocytes, respectively). Values for Ki-67 index in crypt enterocytes were also significantly different between histology-positive and histology-negative groups (P = 0.000). The distribution of CD3+ and CD8+ IETCs was mostly even along the surface enterocytes in histology-positive cases compared with the controls, which showed an uneven distribution. The 2 parameters that significantly discriminated between minimal morphologic change GSE (S+H-) and controls (S-H-) were Ki-67 index (P = 0.007) and the distribution pattern of CD8+ IETCs (P = 0.049). CD4 IETC counts were generally low, with no significant difference between all groups. The few S-H+ cases seen most likely represented false-negative serology, because the assessed parameters of this group and S+H+ cases were indistinguishable.