Significance of acid-mucin-positive nongoblet columnar cells in the distalesophagus and gastroesophageal junction

Acidic mucin-positive nongoblet columnar cells (NGCC) have recently been ob served in the surface epithelium of the gastroesophageal junction (GEJ) and distal esophagus in resections from patients with traditional long segment (> 3 cm) Barrett's esophagus (BE). However, the significance of finding ac idic mucin-positive NGCC in the surface epithelium of biopsy specimens from the distal esophagus/GEJ region in the absence of goblet cells (GC) remain s unknown. Therefore, to determine the significance of mucin histochemical changes in the distal esophagus/GEJ region, we analyzed and compared the ty pes, prevalence, and distribution of neutral and acidic mucins in biopsy sp ecimens obtained from 2 groups of patients: those with (32 patients) and th ose without (107 patients) GC identified in this area. Various mucin histoc hemical stains (PAS-Ab pH 2.5, HID-Ab pH 2.5, PB/KOH/PAS) were used to iden tify neutral mucins, acidic mucins (sialomucins and sulphomucins), and o-ac etylated sialomucins. The results were compared between the 2 patient group s and correlated with the clinical, endoscopic, and pathologic features. Co mpared with patients without GC, patients with GC had a significantly highe r male/female ratio and a higher proportion of patients with greater than 3 cm of columnar epithelium within the esophagus. Acidic mucin (sialomucin a nd sulphomucin)-positive NGCC in the surface, foveolar, and glandular epith elium did not show any correlation with any of the clinical, endoscopic, or pathologic features, such as esophagitis, carditis, antritis, Helicobacter pylori infection, or length of columnar epithelium in the distal esophagus . However, acidic mucin-positive NGCC correlated strongly with the presence of GC (P < .001). For example, sialomucin-positive NGCC were present in 28 of 32 (88%) patients with GC compared with 31 of 107 (29%) patients withou t GC (P < .001). Similary, sulphomucin-positive NGCC were present in 20 of 32 (62%) patients with GC, compared with 11 of 107 (10%) patients without G C (P < .001). Of the non-GC cases, all biopsy specimens that stained positi vely for sulphomucin in surface NGCC (11 specimens), except one, showed sta ining restricted to the surface of multilayered epithelium, a distinctive t ype of epithelium that shows morphological, ultrastructural, and cytochemic al features of both squamous and columnar epithelium. Sialomucin positivity in surface NGCC from the distal esophagus/GEJ region is a sensitive (sensi tivity 88%), but nonspecific (specificity 71%), indicator of GC metaplasia. In contrast, sulphomucin expression in NGCC from the same anatomic area is a less sensitive (sensitivity 62%), but more specific (specificity = 90%) marker for the presence of metaplastic epithelium, of either the GC or the multilayered epithelial cell type and thus may represent an early or incomp lete form of intestinal metaplasia. Copyright (C) 1999 by W.B. Saunders Com pany.