Sensitivity and specificity of postoperative upper GI series following gastric bypass

Abstract

Background: Routine postoperative GI series has been common before discharging gastric bypass patients. 78,000 operations were performed in the USA in 2002. At 75 dollars each, the total annual expenditure for the upper GI series approaches 6 million dollars. This study examines the value of performing routine upper GI series.

Materials and methods: From 1996 to 2000, 396 open gastric bypass procedures were performed by one surgeon at the University Medical Center. 242 randomly selected charts were retrospectively reviewed for signs and symptoms possibly related to leak or obstruction. Radiology reports were compared with clinical findings.

Results: 82% of patients (192/242) were discharged following unremarkable postoperative courses and normal x-rays. 18% (44/242) exhibited one or more clinical signs suspicious of leak or obstruction. These included fever, tachycardia, tachypnea, inordinate pain, elevated white cell count or GI hemorrhage. Leak was reported in 5, and obstruction in 5. 4 patients with reported leaks were re-operated: 2 were positive for unconfined leak requiring surgical treatment; 2 had negative laparotomies. The 2 patients (0.82%) with free leakage had dramatic clinical deterioration, and x-rays were confirmatory rather than diagnostic. 1 patient with a minimal confined leak was treated non-operatively. 8 films were misread as showing a leak when none was present: 2 underwent negative laparotomy, the others being correctly interpreted after review. 8 of 10 initial interpretations were falsely positive.

Conclusion: Routine postoperative GI series following gastric bypass is not beneficial. All true leaks are demonstrated when x-rays are indicated. We recommend GI series only when clinically indicated. GI series had low positive predictive value for leak.