Intracorporeal versus extracorporeal anastomosis in laparoscopic total gastrectomy: a systematic review and...

To evaluate outcomes of intracorporeal (IOJ) versus extracorporeal (EOJ) oesophagojejunostomy following laparoscopic total gastrectomy (LTG) for the treatment of gastric cancer.

A comprehensive search of various electronic databases was conducted. Comparative studies of IOJ versus EOJ following LTG in patients with gastric malignancy were included. Primary outcomes were anastomotic leak, anastomotic bleeding, and anastomotic stricture formation. Secondary outcomes included operative time, length of hospital stay (LOS), volume of intra-operative haemorrhage, number of harvested lymph nodes, time to flatus, time to soft diet, intra-abdominal infection, pulmonary infection, surgical site infection (SSI), duodenal stump leak, pancreatic fistula occurrence, postoperative ileus, re-operation, and mortality. Combined overall effect sizes were calculated using the random-effects model, and the Newcastle-Ottawa Scale was used to assess risk of bias.

Seventeen non-randomised studies enrolling 2,960 patients divided between an IOJ ( n = 1430) and EOJ ( n = 1530) group were included. IOJ was associated with significantly lower risk of anastomotic stricture ( P = 0.01), volume of intra-operative bleeding ( P = < 0.001), and SSI (P = 0.04) compared to EOJ. No difference was found in anastomotic leak ( P = 0.93); anastomotic bleeding ( P = 0.35); operative time ( P = 0.63); LOS ( P = 0.30); lymph node yield ( P = 0.17); time to first flatus ( P = 0.77); time to resumption of soft diet ( P = 0.32); intra-abdominal infection ( P = 0.22); pulmonary infection ( P = 0.45); duodenal stump leak ( P = 0.46); pancreatic fistula occurrence ( P = 0.16); and paralytic ileus ( P = 0.59), re-operation ( P = 0.50), and mortality ( P = 0.23) between the two groups.

LTG for gastric malignancy with IOJ may be associated with lower risk of anastomotic stricture and SSI compared to the extracorporeal approach. However, future adequately powered randomized studies are needed to compare the two techniques.