To determine the significance of the endoscopic classification of gast
ritis proposed by a working party at the World Congress of Gastroenter
ology in Sydney 1990, 167 patients undergoing upper alimentary endosco
py were prospectively assessed by comprehensive endoscopic and histolo
gical methods. Ninety eight patients had endoscopic mucosal changes of
gastritis according to the Sydney classification. Twenty six (27%) of
these had histologically normal biopsy specimens. This was not statis
tically significantly different to the 26 (38%) of 69 with normal endo
scopies whose biopsy specimens were histologically normal (chi 2=1.857
, p>0.1). Forty three (62.5%) patients with normal endoscopies had his
tological gastritis. No histological counterpart was found for the mac
roscopic appearances of the gastric mucosa said to show inflammation p
roposed by the Sydney classification of gastritis. These findings conf
irm the inappropriateness of an endoscopic diagnosis of gastritis and
it is suggested such a term should be reserved for the histological fi
ndings.