A retrospective review of gastroesophageal leaks complicating antirefl
ux operations was conducted to determine the incidence, predisposing f
actors, optimal treatment, and outcome in such patients. Twelve postop
erative gastroesophageal leaks occurred in a series of 1,005 antireflu
x procedures (1.2%). Four of the 12 patients had undergone a previous
hiatal operation, and this was a significant risk factor for postopera
tive leak (p < 0.001). Ten of the 12 patients had undergone ''incomple
te'' wraps that involved suturing of the gastric fundus to the esophag
us, and this was a significant risk factor for postoperative leak (p <
0.04). Five patients had peritoneal contamination and 7 had mediastin
al or pleural soilage. Patients with peritoneal perforations were less
likely to require intensive care unit admission than were patients wi
th thoracic perforations (p < 0.05). Six of the 12 perforations were e
ither well contained or well drained at the time of the diagnostic con
trast study. All 6 of these patients responded to conservative treatme
nt. The remaining 6 perforations were not contained at the time of dia
gnosis. Two of the affected patients initially received conservative t
reatment (1 death and 1 late empyema) and 4 were treated by operation
(1 death). The mortality associated with gastroesophageal perforation
was 17%. Contained perforations can be treated conservatively but nonc
ontained perforations require early and aggressive surgical interventi
on.