Substantial underreporting of anastomotic leakage after anterior resection for rectal cancer in the Swedish Colorectal Cancer Registry

Background: The causes and effects of anastomotic leakage after anterior resection are difficult to study in small samples and have thus been evaluated using large population-based national registries. To assess the accuracy of such research, registries should be validated continuously. Material and methods: Patients who underwent anterior resection for rectal cancer during 2007–2013 in 15 different hospitals in three healthcare regions in Sweden were included in the study. Registry data and information from patient records were retrieved. Registered anastomotic leakage within 30 postoperative days was evaluated, using all available registry data and using only the main variable anastomotic insufficiency. With the consensus definition of anastomotic leakage developed by the International Study Group on Rectal Cancer as reference, validity measures were calculated. Results: Some 1507 patients were included in the study. The negative and positive predictive values for registered anastomotic leakage were 96 and 88%, respectively, while the κ-value amounted to 0.76. The false-negative rate was 29%, whereas the false-positive rate reached 1.3% (the vast majority consisting of actual leaks, but occurring after postoperative day 30). Using the main variable anastomotic insufficiency only, the false-negative rate rose to 41%. Conclusions: There is considerable underreporting of anastomotic leakage after anterior resection for rectal cancer in the Swedish Colorectal Cancer Registry. It is probable that this causes an underestimation of the true effects of leakage on patient outcomes, and further quality control is needed.