Is intercellular space different among layers in normal esophageal mucosa? An electron microscopic study.

Background/introduction: Dilatation of intercellular space (IS) of esophageal epithelial cells is described as a sensitive early marker for epithelial damage by refluxate. Esophageal epithelia are morphologically subdivided into three layers according to the shape of the cells and nuclei. Meanwhile, ten transmission electron microscopy (TEM) photographs and ten randomly selected measurements per photo from gastroesophageal reflux disease (GERD) patients have been widely accepted without any theoretical criticism. We assumed that the IS differs among each layer and thus studied IS according to subdivided layers of normal esophageal epithelium. We also evaluated an optimal number of IS measurements per photograph.

Materials And Methods: Esophagogastroduodenoscopy was performed in 15 healthy adults without any symptom of GERD, taking biopsies from mucosa above 5 cm from the squamo-columnar junction. Tissues were handled and prepared for TEM, verifying three layers of esophageal mucosa, i.e., squamous cell layer, prickle cell layer, and basal layer. Ten digital photomicrographs were taken from each three layers by TEM, and ISs were measured with a computerized image analysis program. For the method of measuring IS, 5, 10, 20, 30, and 40 measurements per photomicrograph were respectively performed by four different examiners. Mean value and intraclass correlation coefficient (ICC) was also yielded.

Results: Mean IS of lower esophagus irrelevant to three epithelial layers were 0.39 ± 0.30 μm. When subdivided into three layers, however, mean IS of squamous cell layer was 0.62 ± 0.23 μm, prickle cell layer 0.23 ± 0.19 μm, and basal layer 0.55 ± 0.36 μm, with their difference statistically significant (p < 0.05). On the other hand, ICC of 5, 10, 20, 30, and 40 measurements were 0.688, 0.917, 0.837, 0.790, and 0.765, respectively.

Conclusions: Mean IS values of each three layers of esophageal epithelium in normal subjects were significantly different, and reconsideration of the standard measurement method is needed. Ten measurements per photo had an adequate inter-observer agreement.