Management of anastomotic leakage after nondiverted large bowel resection

BACKGROUND: The purpose of this study was to determine the natural history of anastomotic leakage after elective colorectal resection and supraperitoneal anastomosis without temporary stoma. STUDY DESIGN: Medical records from 1990 to 1997 were studied; 655 consecutive patients underwent colonic or rectal resection (without stoma). Patients were divided into two groups: those with clinical anastomotic leakage confirmed by laparotomy (group 1) and those without anastomotic leakage (group 2). Postoperative clinical and biologic findings were compared between the two groups. RESULTS: Anastomotic leakage occurred in 39 of 655 patients (6%). Clinically suspected anastomotic leakage was only confirmed by contrast radiography in 13 of 24 patients (54%), and by CT in 8 of 9 patients (89%). Significantly more patients in group 1 than group 2 had the following: fever (> 38 degrees C) on day 2 (p < 0.001); absence of bowel action on day 4 (p < 0.001); diarrhea before day 7 (p < 0.001); collection of more than 400 mL of fluid through abdominal drains from day 0 to day 3 (p < 0.01); renal failure on day 3 (p < 0.02); and leukocytosis after day 7 (p < 0.02). Among the 39 patients in group 1, 28 (71%) had at least one of these clinical or biologic manifestations before day 5, but the mean delay for reoperation was only 8 days. The combination of signs observed before day 5 was associated with an increased risk of anastomotic leakage, from 18% with two signs to 67% with three signs. Overall mortality rate was 2% (13 of 655) and was significantly higher in group 1 than group 2: 5 of 39 (13%) versus 8 of 616 (1%, p < 0.001). In patients with anastomotic leakage, death occurred in 5 of 23 patients (22%) reoperated on after day 5, versus 0 of 11 patients (0%) reoperated on before day 5 (NS). Univariate analysis showed that three clinical characteristics were associated with a significantly high risk of mortality after reoperation for anastomotic leakage: age greater than 65 years (p < 0.01), American Anesthesiologist Association score greater than 3 (p < 0.05), and blood transfusions during the first operation (p < 0.02). CONCLUSIONS: In our study, some postoperative clinical and biologic signs were associated with a higher risk of anastomotic leakage. The knowledge of these findings might help in the early diagnosis and management of patients with anastomotic leakage after large bowel resection.