Background and Study Aims: The visible vessel is an important endoscop
ic sign for predicting rebleeding in ulcers. Freeman has described a v
isible vessel with a high rate of rebleeding as a ''pearl'' color (whi
tish) compared with a darkly colored ''sentinel'' clot with a low rebl
eeding rate. Clarifying the color of visible vessels helps to distingu
ish high-risk bleeding ulcers. We conducted a retrospective study that
compared pathological findings with endoscopic pictures to determine
the significance of the visible vessel's color. Patients and Methods:
From January 1986 to December 1992, 110 patients who underwent endosco
py and received subtotal gastrectomies for ulcer bleeding were include
d in this study, Of these, 24 received endoscopic therapy before the o
peration. There were 94 males and 16 females, ranging in age from 30 t
o 90 years, with a mean age of 62.5 years. According to Freeman's repo
rt, a subgroup of visible vessels (II av +) was defined as having eith
er a (''pearl-colored'' collar around a red or black protruding mount
or a ''pearl-colored'' mount on the ulcer base. The endoscopic finding
s were compared with the findings of the pathological specimens. Resul
ts: The endoscopic findings on the stigmata of recent hemorrhage in th
e 110 patients revealed that 31 were of type II a (including 18 type I
I av + and 13 type II av -), 56 of type II b, 18 of type II c, and 5 o
f type III. Fifty-four patients (49,1%) were found from their patholog
ical specimen to have an eroded vessel on their ulcer base. Type II a
patients had a higher percentage of eroded vessels. The percentages of
eroded vessels in types II a, II b, II c, and III were 67.7%, 46.4%,
33.3%, and 20%, respectively (P < 0.05, Fisher's exact test). Of the 5
4 patients with an eroded vessel in their pathological specimen, 13 (2
4 %) were found to have some vessel wall above the ulcer base (six in
type II a, four in type II b, two in type II c, and one in type III).
Among the six with vessel wall above the ulcer base in type II a, five
patients (83 %) were identified as type II av + under endoscopy. Ther
e was a greater frequency of having a vessel wall above the ulcer base
in type II av + than in type II av -(38.5% vs. 12.5%) among the 31 en
doscopic type II a patients. Conclusion: From this study, we determine
d that the wall of an eroded vessel on a gastric ulcer may protrude ei
ther above or below the ulcer base. A vessel wall on the ulcer base wi
ll appear pearl-colored under endoscopic view.