Assessment of anastomotic reliability with pulse oximetry in graded intestinal ischemia: an experimental study in dogs

BACKGROUND/PURPOSE: Pulse oximetry has been proposed as an appropriate and feasible technique in the assessment of intestinal ischemia in recent years. In this study the authors aimed to assess the reliability of anastomoses in the dog small intestine in which there is graded irreversible ischemia as measured by pulse oxymeter. METHODS: In a control group of four dogs, without any devascularization, three small bowel anastomoses were formed in each dog. The study group consisted of 12 dogs. In each animal three intestinal segments with different levels of ischemia were created by ligating the marginal vessels proximally and distally in sequence beginning from the midpoint of the segmental vascular arcade. Preanastomotic pulse oximeter readings between 80% and 90% were assigned to mild ischemia, 70% and 80% to moderate, and 60% and 70% to severe ischemia group. Pulse oximetry measurements were obtained from probes applied to the antimesenteric serosal surfaces at the midpoint of small intestinal segments. A total of 48 intestinal segments (12 nonischemic in the control group and 36 with three different levels of ischemia in the study group) were transected in the midpoint and anastomosed in double layers. Postanastomotic SaO2 values were also noted. The anastomoses were evaluated 48 hours later macroscopically if there was any leakage, and biopsy specimens were obtained for histopathologic ischemic gradings. All results were studied statistically. RESULTS: Histopathologic grades between each group were statistically different (P < .01 for each comparison) except for control and mild ischemia groups (P > .05), worsening as the level of ischemia increased. Pre- and postanastomotic pulse oximetry measurements correlated very well with the histological gradings (r = -0.90, P < .001 and r = -0.93, P < 0.001 respectively). Number of anastomotic leakages were none in control, one in mild, nine in moderate, and 12 (all of the anastomoses) in severe ischemia groups. In the moderate ischemia group with an average preanastomotic pulse reading of 76.75%, each of the leaking anastomoses had a postanastomotic pulse measurement of lower than 70%. The finding that the difference between histopathologic grades of control and mild ischemia groups with average preanastomotic pulse measurements of 96% and 85%, respectively is not statistically significant enables us to suggest that a saturation of at least 85% is necessary for a reliable anastomosis. CONCLUSION: These results suggest clearly that anastomotic reliability can be predicted objectively with pulse oximetry.