Anastomotic leaks – Is stapler brand a risk factor?

Bruce A Orkin, MD, Daniel A Popowich, MD

Thufts Medical Center

BACKGROUND

Anastomotic leaks are one of the most feared complications of intestinal surgery. The vast majority of colonic anastomoses are created using mechanical staplers. A string of recent anastomotic leaks prompted analysis of risk factors. We hypothesized that the brand of stapler was the most significant factor in the rate of leaks.

METHODS

A retrospective review of one surgeon’s experience over 18 years. All anastomoses of the colon or rectum were included. Ileoanals were excluded. The cohort consisted of 807 anastomoses in 794 patients. Mean age was 55 years (range 10-93); 55% were female. Group 1 – Ethicon staplers, 684 anastomoses (1994-2008). Group 2 – Covidien staplers, 124 anastomoses (2009-2011). Risk factor analysis included stapler brand, age, gender, diagnosis, obstruction, abscess, fistula, metastases, prior Hartmann’s, prior leak, operation type (RHC, segmental resection, sigmoid, STC, IRA, LAR, CAA), lap v open, anastomosis type (sewn v single stapled v double), number of anastomoses, diversion, and pre-op radiation.

RESULTS

Overall, 49 of 807 anastomoses leaked (5.6%) – IO 21 (2.6%), PO 28 (3.5%). 4 (0.5%) patients had an IO leak identified & repaired but leaked PO. RHCs and segmental resections had low leak rates (1.7%), while rectal anastomosis rates were markedly higher (all 14.4%, IO 6.5% & PO 9.8%, P<.001). CAA had the highest rates (28.6%, P<.001), however most were found IO and once repaired they did well (1 PO leak, NSD). Group 1 v Group 2 leak rates were – all 4.5 v 11.3%, IO 2.2 v 4.8% and PO 2.9 v 6.5% (P<.05). None of the 65 (8.1%) sewn anastomoses leaked (P<.01). In univariate analysis factors predictive of a higher leak rate included operation type (LAR, CAA, IRA v all others), absence of loop ileostomy, diagnoses (rectal cancer & dysmotility), anastomotic technique, prior leak, anastomotic type (EEA vs all others), and Covidien stapler use (all P<.05). Multivariate analysis showed prior leaks and operation type to be the most significant factors (P<.01). Although leak rates were higher with Covidien staplers, this failed to reach significance.

CONCLUSIONS

Although the overall leak rate with Covidien staplers was significantly higher than with Ethicon staplers, detailed analysis suggested that other risk factors play a larger role than stapler brand.

Session: Poster Presentation

Program Number: P616