Back to Annual Meeting Program

Introduction:
Anastomotic leaks will occur in roughly 3-10% of surgical procedures that include an end-to-end anastomosis (EEA). To prevent leaks, surgeons evaluate the doughnuts removed from the autosutures and perform maneuvers to assess the viability of the anastomosis including air-leak tests, beta-dine leak tests, and rigid sigmoidoscopies. This study will evaluate the efficacy of these tests to determine if they are valid.

Methods:
A cohort of 44 surgical patients (ages 18-92; 25 male) who presented to The Mount Sinai Medical Center between 2005 and 2012 with anastomotic leaks were compared to a case control group of 86 patients without postoperative leaks. Only anastomoses distal to the splenic flexure and created by employing an EEA stapler were included, as those anastomoses are testable via beta-dine, air-leak, rigid sigmoidoscopy, and doughnut observation. Air-tests and beta-dine tests for left hemicolectomies were excluded. Patient demographics, surgical procedures, reoperations, pathologic findings, and outcomes of the intraoperative tests were reviewed. All statistical analyses were performed using SPSS v20 software.

Results:
The procedures included low anterior resection (50.0%), left hemicolectomy (15.9%), sigmoid resection (29.5%), and ilio-anal pull through (13.6%). Procedures of the control group were equivalent ±7%. Beta-dine and air-leak tests, as well as doughnut observation, were more likely in the control group, and rigid sigmoidoscopies less likely. Table 1 displays the true positives, true negatives, false positives, and false negatives. Table 2 has their corresponding sensitivities and specificities.

Conclusions:
A combination of EEA doughnut integrity and an air-leak test show the highest predictive value. However, the high rates of false negatives and corresponding low sensitivities of these tests show their inability to correctly identify the patients who will ultimately have anastomotic leaks.
Table 1

TEST RESULT	POST-OPERATIVE LEAK (NUMBER OF PATIENTS)	NO POST-OPERATIVE LEAK (NUMBER OF PATIENTS)
Positive Air-Leak Positive Beta-dine Incomplete Doughnuts Positive Rigid Sigmoidoscopy	1 0 1 0	2 1 2 0
Negative Air-Leak Negative Beta-dine Complete Doughnuts Negative Rigid Sigmoidoscopy	25 9 43 15	66 29 81 13

Table 2

Test	Sensitivity (%)	Specificity (%)	Positive Predictive Value (%)	Negative Predictive Value (%)
Air-Leak	3.85	97.06	33.33	72.53
Beta-dine	0.0	96.67	0.0	76.32
EEA Doughnut Integrity	2.27	97.59	33.33	65.32
Rigid Sigmoidoscopy	0.0	100	-	46.43